Seminars in Fetal & Neonatal MedicineAmsterdam • Boston • London • New York • Oxford • Paris • Philadelphia • San Diego • St. Louis

Aims and ScopeSeminars in Fetal & Neonatal Medicine (formerly Seminars in Neonatology) is a bi-monthly journal which publishes topic-based issues, including current ‘Hot Topics’ on the latest advances in fetal and neonatal medicine. The change in title relates to the growing interest amongst obstetricians, midwives and fetal medicine specialists.

The Journal commissions review-based content covering current clinical opinion on the care and treatment of the neonate and draws on the necessary specialist knowledge, including that of the respiratory physician, the infectious disease physician, the surgeon, as well as the paediatrician and obstetrician.

Each topic-based issue is edited by an authority in their field and contains 8–10 articles.

Current and forthcoming events can be viewed on the Internet at: http://www.elsevier.com/locate/siny

Seminars in Fetal & Neonatal Medicine provides: •coverageofmajordevelopmentsinneonatalcare; •value to practising neonatologists, consultant and trainee paediatricians,

obstetricians, midwives and fetal medicine specialists wishing to extend their knowledgeinthisfield;

•up-to-dateinformationinanattractiveandrelevantformat.

Editorial BoardEditor-in-ChiefSteven M DonnUniversity of Michigan Health System

Ann Arbor, MI, USA

Associate EditorsM Blennow, Huddinge, Sweden D Peebles, London, UKL Cornette, Brugge, Belgium S Sinha, Middlesbrough, UKD J Field, Leicester, UK A M Weindling, Liverpool, UKK Maršál, Lund, Sweden

Advisory Board

F A Chervenak, USA J M Perlman, USAN Evans, Australia E Saliba, FranceV Fellman, Sweden N E Vain, ArgentinaN N Finer, USA M Vento, SpainP C NG, Hong Kong L de Vries, The Netherlands

Emeritus Editor-in-ChiefM I LeveneUniversity of Leeds, Leeds England, UK

lable at ScienceDirect

Seminars in Fetal & Neonatal Medicine 19 (2014) 71

Contents lists avai

Seminars in Fetal & Neonatal Medicine

journal homepage: www.elsevier .com/locate/s iny

Editorial

Long term outcome for the tiniest or most immature babies

Health professionals working with high-risk babies and theirmothers, the families, and the children themselves all have a vestedinterest in long-term health outcomes for those born very tiny orimmature. Such information assists with clinical decision making,including in some cases the decision not to pursue active treatment.Consequently, accurate prognostic data are vital, and need to becontinually updated.

Survival rates for the tiniest (<1000 g birth weight) or mostimmature (<28 weeks’ gestational age) infants have increaseddramatically in recent decades with the advent of modern perinataland neonatal intensive care [1,2]. As the tiniest or most immaturesurvivors have more adverse health outcomes than larger ormore mature infants, they contribute disproportionately to theburden of illness in later life, including into adulthood. A better un-derstanding of the possible adverse outcomes and their causes cancontribute to strategies designed to prevent long-term ill-health,either by eliminating some of the causes of adverse outcomes, orby early interventions for higher risk groups that are designed toimprove long-term health.

This issue provides a comprehensive overview of the long-termoutcomes for children born very small or early including mortality,child impairments and quality of life, as well as parent and familyoutcomes. Lucy Smith, Elizabeth Draper, and David Field discussthe various influences on reported mortality rates for the mostimmature infants, and particularly how mortality rates can varywith selection filters, such as different categories of births or admis-sions to intensive care nurseries. Betty Vohr reports on speech andlanguage outcomes, and some of the causes for impairments inthese areas, in particular brain injury and hearing impairment. Ali-cia Spittle and Jane Orton describe the higher rates of cerebral palsyand other motor impairments consistent with developmental coor-dination disorder, and their associated causes and co-morbidities,mainly with poorer cognitive and academic function, as well aseffects on daily living. They make recommendations for follow-up, and suggest interventions to improve motor function. PeterAnderson examines neuropsychological outcomes, particularlywith respect to general cognitive ability, processing speed, atten-tion, visual and perceptual skills, memory and learning, language,executive function, and educational outcomes. Samantha Johnsonand Neil Marlow explore mental health outcomes throughoutchildhood, including behavioural and emotional problems in thepre-school years, ADHD and Autism Spectrum Disorder at schoolage, and psychological and psychiatric disorders in adulthood.Anne-Marie Gibson and Lex Doyle synthesise data on respiratory

1744-165X/$ e see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.siny.2013.11.015

health and respiratory function, including hospital readmissions,respiratory flow rates and lung volumes, and exercise tolerance,emphasising the respiratory problems of survivors of bronchopul-monary dysplasia. Eero Kajantie and Petteri Hovi ask if very pre-term birth is a risk factor for adult cardiovascular or metabolicdisease. They highlight the evidence supporting higher risks forpreterm survivors for hypertension, abnormal glucose tolerance,and reduced activity and fitness levels, and the potential roles ofdiet, sleep, and smoke and alcohol exposure. Gehan Roberts andJeanie Cheong investigate long-term growth and general health,including changes in weight and height, chronic health conditions,vision and hearing, bone health, and health care utilisation. SarojSaigal considers functional outcomes of very premature infantsinto adulthood, including neurosensory impairments, vision andhearing, educational achievement, employment, social functioning,reproduction, and quality of life. Finally, Karli Treyvaud reviewsparent and family outcomes, including mental health of mothersand fathers, parenting stress, family functioning, and the impacton the family of having a very preterm birth.

We are indebted to all authors for their wonderful contributionsto this issue of Seminars in Fetal and Neonatal Medicine, which as awhole provides an accurate and up-to-date review of the conse-quences of being born very small or early, as well as highlightingareas for future research.

References

[1] Doyle LW, Roberts G, Anderson PJ. Changing long-term outcomes for infants500e999 g birth weight in Victoria, 1979e2005. Arch Dis Child Fetal NeonatalEd 2011;96:F443e7.

[2] Doyle LW, Roberts G, Anderson PJ. Outcomes at age 2 years of infants < 28weeks’ gestational age born in Victoria in 2005. J Pediatr 2010;156:49e53.

Lex W. Doyle*

Department of Obstetrics and Gynaecology, The Royal Women’sHospital, 20 Flemington Road, Parkville, Victoria 3052, Australia

Peter J. AndersonVictorian Infant Brain Studies, Murdoch Childrens Research Institute,

Royal Children’s Hospital, Flemington Road, Parkville,Victoria 3052, Australia

E-mail address: peter.anderson@mcri.edu.au.

*Corresponding author.E-mail address: lwd@unimelb.edu.au (L.W. Doyle).

lable at ScienceDirect

Seminars in Fetal & Neonatal Medicine 19 (2014) 72e77

Contents lists avai

Seminars in Fetal & Neonatal Medicine

journal homepage: www.elsevier .com/locate/s iny

Review

Long-term outcome for the tiniest or most immature babies: survivalrates

Lucy K. Smith*, Elizabeth S. Draper, David FieldDepartment of Health Sciences, University of Leicester, Leicester, UK

Keywords:Birth registrationExtremely preterm birthInfant mortalitySurvival

* Corresponding author. Address: Department of HLeicester, 22e28 Princess Road West, Leicester LE1 65468; fax: þ44 (0) 116 252 3272.

E-mail address: lks1@le.ac.uk (L.K. Smith).

1744-165X/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.siny.2013.11.002

s u m m a r y

This article focuses on the survival rates of the most immature babies considered viable from around theworld. It discusses the various factors in terms of definition, inclusion criteria and policy which can resultin marked differences in the apparent rates of delivery (all births), live birth, admission to neonatalintensive care and ultimately survival seen between otherwise similar countries, different regions of thesame country, and even adjacent hospitals. Such variation in approach can result in major differences inreported survival and consequentially have large effects on apparent rates of adverse long-term outcome.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The aim of this edition of Seminars in Fetal and Neonatal Medicineis to discuss the long-term outcome of the tiniest and mostimmature babies. These babies represent the ‘cutting edge’ ofneonatal medicine throughout the developed world, and theappropriateness, or not, of offering active intensive care to thesebabies is a matter of great debate and polarised views (http://justice4jayden.webs.com/). It is clear that babies who do surviveafter delivery at 22, 23 or 24 weeks are likely to need at least 4e5months of inpatient care after birth (sometimes much longer) andare at high risk of death or disability [1,2]. Having a clear under-standing of these risks is important for clinicians in counsellingparents, and in enabling families to make informed decisionsregarding how such babies should be managed. However, pub-lished data can be confusing, with apparently very different rates ofsurvival, and these differences inevitably have an effect on reportedrates of longer-term adverse outcome as the child grows. In thispaper we review the range of influences that can result in both realand artefactual differences in survival and, as a result, affect re-ported rates of adverse long-term outcome.

2. Defining the cohort: gestation or birth weight?

One of the most important issues when studying extreme pre-term birth is the need for a clear definition of the cohort being

ealth Sciences, University ofTP, UK. Tel.: þ44 (0) 116 252

All rights reserved.

studied. Including the tiniest (defined by birth weight) and mostimmature (defined by gestational age) babies inevitably leads to alack of standardisation of the cohort. Gestational age is a fixedmeasure at any one point in time, independent of both birth weightand fetal growth (an important factor in studies of the preterminfant), whereas birth weight is biologically dependent upon bothgestational age and fetal growth [3]. So a group defined by weightwill include a mixture of immature babies and more mature babieswith a degree of growth restriction, particularly those from mul-tiple births. Studies of preterm birth indicate that unless a combi-nation of both extreme immaturity and severe growth restrictionco-exist then it is the influence of gestation which has the great-est influence on survival [4].

3. Reported survival rates

Mortality rates of the most immature babies have beenreviewed previously, identifying significant variation betweensettings [1,2,5e8]. Reported rates of survival in studies based ongeographical or quasi-geographical cohorts have not to datereached double figures at 22 weeks, but at 23, 24 and 25 weeks ofgestation reported survival rate has ranged from 0 to 53%, 3 to 70%and from 29 to 85%, respectively. These survival data have beenextracted from studies that span over 20 years but few haveexamined changes in the same population over time. Studies in theUK have identified improved survival for babies at 24 and 25 weeksof gestation between the 1990s and the second half of the firstdecade of the 21st century. Nevertheless for babies born before 24weeks of gestation there was little change [9,10]. Studies in Sweden[5,11] are a little more difficult to examine in terms of like-for-likecomparison, but again outcome generally improved with time and

L.K. Smith et al. / Seminars in Fetal & Neonatal Medicine 19 (2014) 72e77 73

the reported survival rates are substantially higher than in the UK.By contrast, data from Australia for this same group of babiesshowed no change in survival between 1997 and 2005; however,survival rates in the latter cohort were comparable with contem-poraneous European cohorts [12].

4. Scientific advances over time

One of the greatest influences on reported survival rates of themost immature babies is that of changes over time and the defi-nition of ‘the tiniest and most immature babies’ at different pointsin the development of neonatal care. This change primarily reflectsdevelopments in neonatal care and the viewof those providing careabout what is feasible at that time (in terms of technology andresources). Some advances such as the availability of exogenoussurfactant and the widespread use of antenatal corticosteroidsclearly have had a very major influence on what is possible. It maybe obvious, but nonetheless worth emphasising, that the percep-tion of ‘what is possible’ still varies enormously around the world,and for this reason it is easier to consider survival of the mostimmature babies in the broadly similar settings provided in thedeveloped world.

5. Type of study

In general, rates of survival or long-term outcome are reportedeither for a geographically defined population (such as births frommothers living in a whole country or a particular region) or a singlehospital or group of hospitals where the catchment area is gener-ally less well defined. Population-based studies that report theoutcome of all babies alive at the onset of labour (or ‘potential livebirths’) can be considered a gold standard. Rates of survival in thesestudies will generally be lower as they use total births (live birthsand intrapartum stillbirths) as a denominator. Bt contrast, studiesbased on hospitals generally have higher survival rates as they tendto use either live births or neonatal unit admissions as their de-nominator, since the total number of births is usually not known orcannot be calculated (as some babies are transferred into theneonatal unit from elsewhere). The use of these different de-nominators for calculating survival will strongly influence the ratesof reported survival. For example, a hospital reporting the outcomeof babies born at 22e24 weeks of gestation will produce verydifferent rates of survival depending on whether the denominatoris all births, all those alive at the onset of labour, live births, orneonatal unit admissions at that gestation e the latter will clearlyproduce the highest reported rate of survival [13]. The inclusion ofbabies transferred from elsewhere (particularly postnatal transfers)will further influence the results since, in order for the transfer tobe possible, the baby will generally have to have been in goodcondition. Since these different types of study populations varyfundamentally in their make-up, in wider comparisons with otherpublished studies only those that include all births alive at theonset of labour as a denominator can really be used as the basis ofreliable comparisons.

Whereas it is generally easy to predict the influence of decisionson which babies to include when reporting mortality among themost immature babies, estimating the influence on neuro-developmental outcome is far more difficult. Geographical studiesminimise selection bias, but, with hospital-based series, the antic-ipated better survival rates due to the transfer into tertiary centresof those most likely to survive may lead to a larger proportion ofinfants surviving with long-term disability. Readers should keepthese potential influences in mind when reading articles reportingoutcomes of this type.

6. Inclusion of terminations

Among the many factors that affect the apparent rates of sur-vival for babies born at the lowest gestation is the issue of whetheror not babies born after a termination of pregnancy are included inthe denominator and/or the numerator. This is an issue that pri-marily affects population-based studies, especially comparisonsbetween countries, and is dependent on a variety of legislativedifferences in respect of under what circumstances and up to whatstage of pregnancy termination may be offered. Therefore in com-parisons of survival of the most immature babies between coun-tries, it is important to identify whether or not terminations ofpregnancy have been included. Those countries where termina-tions are included in the numerator will have poorer survival ratessince inevitably they are all deaths [14].

In the wider context, the approach to termination in a particularcountry has consequences for other early life outcomes such asincreasing neonatal mortality due to congenital anomalies incountries where these are either not detected and/or termination isnot offered early in pregnancy. However, this issue affects mainlymature infants.

7. Congenital anomalies

Although generally not an issue for national comparisons, aca-demic studies of very immature babies will often choose to excludebabies with a ‘major’ congenital anomaly. Fortunately majoranomalies are rare among this group and hence they generally havelittle impact on reported rates of survival. However, in these cir-cumstances different interpretations of what constitutes a major/lethal anomaly (and hence the babies included or excluded) willhave an impact on survival.

8. Impact of variation in rates of prematurity

For each of the known influences on reported survival discussedso far, it is possible to have an appreciation of the type of ‘bias’introduced e for example, as a result of the study type and exclu-sion/inclusion criteria. However, the scale of the impact of otherinfluences is much more difficult to estimate.

TheMarchofDimes regularlypublishesdataon itswebsite (http://www.marchofdimes.com/peristats/Peristats.aspx) demonstratingmarkeddifferences in the rate of verypretermbirth between states inthe USA. Similar data exist which show wide variations in very pre-termbirth rates between countries of the EU (Fig.1) and globally [17].Whereas these apparent ethnographic differences have a huge in-fluence on the need for neonatal services in these states/countries,therewill almost certainly be a parallel effect on the number of babiesbornat the lowest gestations. There are nodata to determinewhetherthis variation translates into differences in how these babies aremanaged, cared for, or registered, and their subsequent outcomeseither in terms of survival or longer-term health status.

It is well established that socio-economic deprivation is a majorinfluence on rates of very preterm delivery. It seems likely thatdifferent background rates of deprivation, with their associatedexposures and lifestyle influences, are responsible for a significantproportion of the differences seen in the rate of very prematuredelivery between countries and regions. Interestingly there aredata to demonstrate that, in the short term, the neonatal course ofchildren born to mothers from socio-economically deprived areasis no different from that for babies of the same gestation born tomothers from less deprived areas [18]. However, in stark contrast,there is good evidence that the factors that constitute socio-economic deprivation are often associated with poorer neuro-developmental outcome in childhood.

Fig. 1. International rates of very preterm birth and neonatal mortality per 1000 live births. Data derived from three separate sources Field et al. [15,16], and March of Dimes (http://www.marchofdimes.com/peristats/Peristats.aspx).

L.K. Smith et al. / Seminars in Fetal & Neonatal Medicine 19 (2014) 72e7774

9. National policy and ethos

It is clear that views of how to approach the care of the mostimmature babies varies around theworld. Atoneendof this spectrumwas the viewadopted inTheNetherlands in the early 2000s of havinganabsolutegestationalagecut-off forofferingneonatal intensivecare,initially fixed at 25 weeks of gestation [19e21]. In the UK, manage-ment practice is based on recommendations by the Nuffield Councilon Bioethics (Box 1), which are broadly typical of similar guidelinespublished by organisations or countries around the world. In generalthe guidelines donot prescribe particular courses of action and hencelocal views and policies heavily influence what takes place [22,23].

Despite these guidelines, it is clear that over time there has beenan increase in centres offering intensive care more regularly tobabies of 22 weeks of gestation [10,24]. Inevitably these changes inapproach will have had some effect on both reported survival ratesand rates of long-term disability since management practice willinfluence whether babies at these early gestations are reported as alive birth or stillbirth (see below).

10. Issues of reporting and registration: comparing countriesand regions

10.1. Criteria for registration of births

All the major developed countries produce national statisticsrelating to late pregnancy losses and deaths in infancy. Published

rates of stillbirth, perinatal death and deaths in infancy are oftenplaced in the public arena by the media and attract the attention ofthe public and politicians. These comparisons of population-levelsurvival are frequently used as surrogates to assess the effective-ness of local antenatal or neonatal services, with calls for publichealth interventions to lower rates in areas perceived to be per-forming poorly. Themost immature babies are amajor component ofthese overall rates of mortality in developed countries. As a conse-quence international variations in birth registration practices havebeen shown to have a major influence on infant mortality rankings[25e27], and this represents another situation in which there is aneed to distinguish between ‘real’ variations and those arising fromartefactual differences. There is increasing evidence of systematicdifferences in how births are reported and of misclassification ofstillbirths and live births at the threshold of viability, occurring notonly between countries but also between hospitals within countries[28,29], resulting in misleading comparisons of survival rates.

Registration disparities mainly relate to births of uncertainviability when definitions of late-fetal death and live birth are opento variations in interpretation depending upon personal, culturaland policy differences. Globally the World Health Organization(WHO) definition of a live birth is generally accepted, being ‘thecomplete expulsion or extraction from its mother of a product ofconception, irrespective of the duration of pregnancy, which, aftersuch separation, breathes or shows any other evidence of life, suchas beating of the heart, pulsation of the umbilical cord, or anydefinite movement of voluntary muscles, whether or not the

Box 1Guidelines on giving intensive care to extremely prematurebabies.a

� At 25 weeks and above:Intensive care should be initiated and the babyadmitted to a neonatal ICU, unless he or she is knownto be affected by some severe abnormalityincompatible with any significant period of survival.

� Between 24 weeks, 0 days and 24 weeks, 6 days:Normal practice should be that a baby will be offeredfull invasive intensive care and support from birth andadmitted to a neonatal ICU, unless the parents and theclinicians are agreed that in the light of the baby’s con-dition it is not in his or her best interests to startintensive care.

� Between 23 weeks, 0 days and 23 weeks, 6 days:It is very difficult to predict the future outcome for anindividual baby. Precedence should be given to thewishes of the parents. However, where the conditionof the baby indicates that he or she will not survivefor long, clinicians should not be obliged to proceedwith treatment wholly contrary to their clinical judge-ment, if they judge that treatment would be futile.

� Between 22 weeks, 0 days and 22 weeks, 6 days:Standard practice should be not to resuscitate thebaby. Resuscitation should only be attempted andintensive care offered if parents request resuscitation,and reiterate this request, after thorough discussionwith an experienced paediatrician about the risks andlong-term outcomes, and if the clinicians agree thatit is in the baby’s best interests.

� Before 22 weeks:Any intervention at this stage is experimental.Attempts to resuscitate should only take placewithin a clinical research study that has beenassessed and approved by a research ethicscommittee and with informed parental consent.

a Source: Nuffield Council for Bioethics statement on intensive care for

extremely premature babies (http://www.nuffieldbioethics.org/neonatal-

medicine/neonatal-medicine-guidelines-intensive-care-extremely-

premature-babies).

L.K. Smith et al. / Seminars in Fetal & Neonatal Medicine 19 (2014) 72e77 75

umbilical cord has been cut or the placenta is attached.’ (http://www.who.int/healthinfo/statistics/indunder5mortality/en/).Conversely a fetal death or stillbirth is defined as not displaying anyof these signs of life. These WHO definitions are ‘irrespective of theduration of pregnancy’ and so do not include thresholds for birthweight or gestational age. Despite these definitions being globallyaccepted, and as discussed earlier, the local anticipation of viabilityof babies varies widely and this leads to a more pragmatic localapproach to the reporting of live births and stillbirths.

In low- andmiddle-income countries where there is little accessto neonatal intensive care and babies are predominantly born athome with no health professional present to assess signs of life,babies born before 32 weeks rarely survive [30]. Consequentlybirths, even at 30 weeks of gestation, are rarely recorded as livebirths or stillbirths but viewed as miscarriages. By contrast, in high-income countries with wide access to neonatal intensive care, assurvival rates improve for extremely preterm births, births aredeemed viable at much earlier gestations (see Box 1). However,even between high-income countries with similar access toneonatal intensive care there is variation in the recording of livebirths and stillbirths [31]. For fetal deaths, registration criteria vary

considerably with some countries registering deaths as early as 12weeks of gestation, whereas others only report those from 28weeks of gestation (Table 1). Similarly registration of live births alsovaries with all live births being registered in England andWales, theUSA and Canada irrespective of gestation, but in some Europeancountries including France, The Netherlands and the Czech Re-public there are thresholds for reporting live births based ongestation, birth weight or survival [31].

10.2. Perceptions of viability

Even within countries with national definitions of stillbirthand little variation in access to neonatal care, there is evidence ofvariation in the perceptions of viability between regions. Studiesin the USA have shown that despite the implementation of na-tional reporting requirements, clinical decision-making is basedon definitions of viability [29] and consequently variation existsin state registration practices [32]. Much of this variation mayarise from practical difficulties in interpreting true signs of lifeand subjective differences in judgements about the best outcomefor parents. In the USA, where the WHO definition is accepted,research shows that nearly a third of physicians include gesta-tional age in their personal criteria for defining a live birth [32],with definitions ‘open to subjective interpretation’ [33]. Similarlyin the UK, there is evidence of regional variation in the reportingof births before 24 weeks as live born or stillborn, with rates oflive births at this gestation ranging from 20% to 80% betweenhealth regions [28]. This suggests that differences in manage-ment strategies and decisions regarding viability by a range ofhealth care professionals present at delivery are likely to beresponsible for the wide variations in delivery outcome ofextremely preterm infants. Such variation in decisions overwhether or not to register a birth as live born may also affectparents greatly. This is evident in the UK, where births registeredas live-born at any gestation resulting in a neonatal death (even ifdeath is just a few seconds later) or as stillbirths after 24 weeksof gestation are eligible for statutory benefits. However, since latefetal deaths (<24 weeks of gestation) are not officially registered,these parents are not eligible for statutory benefits and there isno requirement for a funeral. This has a major impact ondeprived populations where extreme preterm birth rates areconsiderably higher. The difference in how the baby is viewedadministratively (was the baby ever alive after birth, i.e. did thebaby exist as a person?) may also impact greatly on the parents’grieving process. It is clear that the classification of babies bornat very low gestation has both administrative consequences anda very practical impact on the family.

These systematic reporting and registration differences aroundthe threshold of viability result in important differences in nationaland regional published data concerning the reported proportion ofextremely preterm births. As these babies are at the highest risk ofpoor outcome, countries or regions with a lower reported propor-tion of extremely preterm births (<28 weeks of gestation) mayreport better survival rates for these births since they exclude thosewith the poorest prognosis. This variation has been shown to have aconsiderable impact on a variety of outcome measures includingthe survival of extremely preterm infants and the overall infantmortality rate. Studies of both inter- and intra-national compari-sons have shown that these variations lead to inappropriate rank-ings of mortality [28,31].

11. Practical consequences

It will be clear from the issues discussed that a wide range offactors relating to the tiniest and most immature babies affect

Table 1Gestational age at delivery and birth weight criteria for stillbirth registration bycountry for Europe, North America, Australia and New Zealand.a

Basis of limit Limit Country

No limit e LuxembourgSpain

Birth weight �500 g AustriaBelgium: FlandersCanada: QuébecGermanyPolandSlovenia

Gestational age �12 weeks Norway�20 weeks USA�22 weeks Czech Republic

DenmarkItalyLatviaLithuaniaNorthern IrelandScotlandSlovak RepublicSpain: Valencia

�24 weeks England and WalesHungaryPortugal

�28 weeks GreeceSweden

Birth weight or gestational age �20 weeks or >400 g AustraliaNew Zealand

�20 weeks or >500 g Canada�22 weeks or >500 g Belgium: Brussels

EstoniaFinlandFranceMaltaNetherlands

�24 weeks or >500 g Ireland

a Only selected countries with data used by Joseph et al. [31].

L.K. Smith et al. / Seminars in Fetal & Neonatal Medicine 19 (2014) 72e7776

their apparent survival or whether they are even included inpublished data regarding late pregnancy and early life mortality.However, because of the relatively high mortality of these in-fants, in comparison to more mature babies, their effect onoverall survival rates relating to all births is disproportionatelylarge.

For those involved in perinatal care the issues are somewhatdifferent. The most important factor in counselling familiesappropriately is the interpretation of data about survival rates forthis high-risk group of babies. It has been suggested [34] that riskadjustment should be used when considering these most imma-ture babies and advising on their management. Whereas this musthave potential merits, any risk adjustment may not be robustoutside its country of derivation (because of the issue raisedearlier in this article) and in many countries the low overallnumbers will make the development of an adjustment tool verychallenging.

When reviewing published outcome data about the mostimmature babies as a basis for counselling families it isimportant to try and identify reports that are derived eitherfrom the same country or from one with a similar degree ofeconomic development and ethos in relation to extreme pre-maturity. Nonetheless some of the difficulties that can arise inattempting to make sound interpretations from what arefundamentally different studies can be seen from the followingexamples:

� Sugiura et al. [24] report the outcome of seven babies of 22weeks of gestation inborn at a single Japanese hospital over 4

years, six of whom survived. We know nothing of the widercontext of the population fromwhich these babies were drawn,but the work does demonstrate that, in Japan at least, survival ispossible at this gestation.

� Riley et al. [35] report on survival at very low gestation in asingle centre over a number of years but including inborn andoutborn babies, i.e. there is no real denominator. Again,although this paper demonstrates that survival at lowgestation is possible, extrapolating the findings, say, to asingle couple whose baby is not yet born is fraught withdifficulty.

� Doyle et al. [36] report survival and long-term outcome using abirth weight cut-off of 1000 g. The report is from a geographicalpopulation, excludes terminations for congenital anomaly, andhas the advantage of having data about long-term outcome.However, the use of live births and birth weight as entry criteriamakes it more difficult to compare to similar studies andcertainly limits its usefulness when counselling a couple whosebaby is not yet born.

� The EPICure studies [10] have dealt with the majority of thepotential confounders of reported survival for the mostimmature babies. However, they demonstrate two other factorsthat should be considered: (a) mortality rates are likely tochange over time, so data used for counselling should becontemporaneous; (b) attitudes and management practiceschange over time, with, for example, a much higher proportionof babies admitted to neonatal units in the second time-periodcovered in these studies.

12. Possible solutions

All of these issues relating to the interpretation of mortalitydata are also potentially present in studies which look at thelong-term outcome of the tiniest and most immature babies. Ifanything there is a greater tendency to report on either thepopulation of live births or survivors to the age of assessment.Again these are all perfectly valid approaches to reportingoutcome, but from the point of view of counselling it isimportant to consider the context, i.e. how the generic chanceof normal survival changes when considering a 23-weekgestation baby whose mother is predicted to deliver withinthe next 6 h versus the same baby born alive and in goodcondition versus the same baby established on a ventilator inthe neonatal unit.

Where underlying regional or international differences overdecisions regarding the viability of preterm births exist, thenstandard published mortality statistics will be prone to bias andthe validity of comparisons will be compromised. Excludingbirths less than a certain cut-off such as 24 or 28 weeks ofgestation from mortality rates may be an initial quick-fix tominimise these problems in certain situations. However, thiswould mean the exclusion of up to a third of infant deaths insome areas and would prevent the comparison of survival inbirths before this gestation. Standardisation of data collection inhigh-income countries to record all births from 22 weeks ofgestational age would improve the ability to investigate withincountry differences and also allow for direct international com-parisons (of survival and long-term outcome). Such an approachwould seem sensible for all developed countries. However, since95% of all births occur in countries with incomplete registrationof births [37] it is unlikely that these issues affecting the inter-national comparison of survival rates will be resolved in the shortterm.

Learning points

� Survival of babies aged 24 weeks of gestation and greaterhas improved over time.

� Babies of less than 24 weeks still have extremely pooroutcomes.

� However, there is wide variation in reported survivalrates.

� Reported survival rates are affected by several key issues:^Study type: population-based or hospital-based cohorts.^Selection of denominator: inclusion/exclusion of still-births, births alive at onset of labour, terminations ofpregnancy, congenital anomalies.

^Management practices: policies on cut-offs for offeringneonatal intensive care.

^Reporting and registration: variation in the definition of,and cut-offs for, reporting of late fetal losses.

^Perceptions of viability and definition of prematurity:access to and provision of care for extremely pretermbirths.

� When comparing survival between hospitals, regions andcountries it is vital that the impact of these influentialfactors is assessed.

� There is a need for standardisation of definitions inter-nationally to overcome some of these issues.

Research directions

� Investigation of the effect on parents resulting from theinconsistent classification of babies on the margins ofviability in terms of being alive or dead at birth.

� Investigation of the factors resulting in genuine differ-ences in survival for the most immature and smallestbabies cared for in otherwise comparable settings.

L.K. Smith et al. / Seminars in Fetal & Neonatal Medicine 19 (2014) 72e77 77

Conflict of interest statement

None declared.

Funding sources

None.

References

*[1] Saigal S, Doyle LW. An overview of mortality and sequelae of preterm birthfrom infancy to adulthood. Lancet 2008;371:261e9.

*[2] Milligan DW. Outcomes of children born very preterm in Europe. Arch DisChild Fetal Neonatal Ed 2010;95:F234e40.

*[3] Blair E. The undesirable consequences of controlling for birth weight in peri-natal epidemiological studies. J Epidemiol Community Health 1996;50:559e63.

[4] Kamoji VM, Dorling JS, Manktelow BN, Draper ES, Field DJ. Extremely growth-retarded infants: is there a viability centile? Pediatrics 2006;118:758e63.

[5] EXPRESS Group, Fellman V, Hellström-Westas L, Norman M, Westgren M,Källén K, et al. One-year survival of extremely preterm infants after active peri-natal care in Sweden. JAMA 2009;301:2225e33.

[6] Lundqvist P, Kallen K, Hallstrom I, Westas LH. Trends in outcomes for verypreterm infants in the southern region of Sweden over a 10-year period. ActaPaediatr 2009;98:648e53.

*[7] Field DJ, Dorling JS, Manktelow BN, Draper ES. Survival of extremely pre-mature babies in a geographically defined population: prospective cohortstudy of 1994e9 compared with 2000e5. BMJ 2008;336:1221e3.

*[8] Seaton SE, King S, Manktelow BN, Draper ES, Field DJ. Babies born at thethreshold of viability: changes in survival and workload over 20 years. ArchDis Child Fetal Neonatal Ed 2013;98:F15e20.

[9] Costeloe K, Hennessy E, Gibson AT, Marlow N, Wilkinson AR. The EPICurestudy: outcomes to discharge from hospital for infants born at the thresholdof viability. Pediatrics 2000;106:659e71.

*[10] Costeloe KL, Hennessy EM, Haider S, Stacey F, Marlow N, Draper ES. Shortterm outcomes after extreme preterm birth in England: comparison of twobirth cohorts in 1995 and 2006 (the EPICure studies). BMJ 2012;345:e7976.

[11] Hakansson S, Farooqi A, Holmgren PA, Serenius F, Hogberg U. Proactive man-agement promotes outcome in extremely preterm infants: a population-basedcomparisonof twoperinatalmanagement strategies. Pediatrics2004;114:58e64.

*[12] Victorian Infant Collaborative Study G. Outcomes at age 2 years of infants<28 weeks’ gestational age born in Victoria in 2005. J Pediatr 2010;156.49.e1e53.e1.

[13] Evans DJ, Levene MI. Evidence of selection bias in preterm survival studies: asystematic review. Arch Dis Child Fetal Neonatal Ed 2001;84:F79e84.

[14] Papiernik E, Zeitlin J, Delmas D, Draper ES, Gadzinowski J, Künzel W, et al.Termination of pregnancy among very preterm births and its impact on verypreterm mortality: results from ten European population-based cohorts inthe MOSAIC study. Br J Obstet Gynaecol 2008;115:361e8.

[15] Field D, Bajuk B, Manktelow BN, Vincent T, Dorling J, Tarnow-Mordi W, et al.Geographically based investigation of the influence of very-preterm births onroutine mortality statistics from the UK and Australia. Arch Dis Child FetalNeonatal Ed 2008;93:F212e6.

*[16] Field D, Draper ES, Fenton A, Papiernik E, Zeitlin J, Blondel B, et al. Rates ofvery preterm birth in Europe and neonatal mortality rates. Arch Dis ChildFetal Neonatal Ed 2009;94:F253e6.

[17] Blencowe H, Cousens S, Oestergaard MZ, Chou D, Moller AB, Narwal R, et al.National, regional, and worldwide estimates of preterm birth rates in the year2010 with time trends since 1990 for selected countries: a systematic analysisand implications. Lancet 2012;379:2162e72.

*[18] Smith LK, Draper ES, Manktelow BN, Field DJ. Socioeconomic inequalities insurvival and provision of neonatal care: population based study of verypreterm infants. BMJ 2009;339:b4702.

[19] Draper ES, Zeitlin J, FentonAC,Weber T, Gerrits J, Martens G, et al. Investigatingthe variations in survival rates for very preterm infants in 10 European regions:the MOSAIC birth cohort. Arch Dis Child Fetal Neonatal Ed 2009;94:F158e63.

[20] NeoNed Study GGroup LNFS. Mortality, neonatal morbidity and two yearfollow-up of extremely preterm infants born in The Netherlands in 2007. PloSOne 2012;7:e41302.

[21] Verloove-Vanhorick SP. Management of the neonate at the limits of viability:the Dutch viewpoint. Br J Obstet Gynaecol 2006;113(Suppl 3):13e6.

[22] Pignotti MS, Donzelli G. Perinatal care at the threshold of viability: an in-ternational comparison of practical guidelines for the treatment of extremelypreterm births. Pediatrics 2008;121:e193e8.

[23] Kattwinkel J, Perlman JM, Aziz K, Colby C, Fairchild K, Gallagher J, et al.Neonatal resuscitation: 2010 American Heart Association guidelines forcardiopulmonary resuscitation and emergency cardiovascular care. Pediat-rics 2010;126:e1400e13.

[24] Sugiura T, Kouwaki M, Togawa Y, Sugimoto M, Togawa T, Koyama N. Neu-rodevelopmental outcomes at 18 months’ corrected age of infants born at 22weeks of gestation. Neonatology 2011;100:228e32.

[25] EURO-PERISTAT Project with SCPE and EUROCAT. The European PerinatalHealth Report: Health and Care of Pregnant Women and Babies in Europe in2010; 2013.

[26] Draper ES. Evaluating and comparing neonatal outcomes. Arch Dis ChildFetal Neonatal Ed 2010;95:F158e9.

[27] Kramer MS, Platt RW, Yang H, Haglund B, Cnattingius S, Bergsjo P. Registra-tion artifacts in international comparisons of infant mortality. Paediat Peri-natal Epidemiol 2002;16:16e22.

*[28] Smith L, Draper ES, Manktelow BN, Pritchard C, Field DJ. Comparing regionalinfant death rates: the influence of preterm births <24 weeks of gestation.Arch Dis Child Fetal Neonatal Ed 2013;98:F103e7.

[29] Ramsay SM, Santella RM. The definition of life: a survey of obstetricians andneonatologists in New York City hospitals regarding extremely prematurebirths. Matern Child Health J 2011;15:446e52.

[30] Yasmin S, Osrin D, Paul E, Costello A. Neonatal mortality of low-birth-weightinfants in Bangladesh. Bull WHO 2001;79:608e14.

[31] Joseph KS, Liu S, Rouleau J, Lisonkova S, Hutcheon JA, Sauve R, et al. Influenceof definition based versus pragmatic birth registration on internationalcomparisons of perinatal and infant mortality: population based retrospec-tive study. BMJ 2012;344:e746.

[32] Ehrenthal DB, Wingate MS, Kirby RS. Variation by state in outcomes classi-fication for deliveries less than 500 g in the United States. Matern ChildHealth J 2011;15:42e8.

[33] Macfarlane PI, Wood S, Bennett J. Non-viable delivery at 20e23 weeksgestation: observations and signs of life after birth. Arch Dis Child FetalNeonatal Ed 2003;88:F199e202.

[34] National Institute of Child Health and Human Development NeonatalResearch Network. Intensive care for extreme prematurity e moving beyondgestational age. N Engl J Med 2008;358:1672e81.

[35] Riley K, Roth S, Sellwood M, Wyatt JS. Survival and neurodevelopmentalmorbidity at 1 year of age following extremely preterm delivery over a 20-year period: a single centre cohort study. Acta Paediatr 2008;97:159e65.

[36] Victorian Infant Collaborative Study G. Changing long-term outcomes forinfants 500e999 g birth weight in Victoria, 1979e2005. Arch Dis Child FetalNeonatal Ed 2011;96:F443e7.

[37] Group GR. Global report on preterm birth and stillbirth (1 of 7): definitions,description of the burden and opportunities to improve data. BMC PregnChildbirth 2010;10(Suppl. 1):S1.

lable at ScienceDirect

Seminars in Fetal & Neonatal Medicine 19 (2014) 78e83

Contents lists avai

Seminars in Fetal & Neonatal Medicine

journal homepage: www.elsevier .com/locate/s iny

Review

Speech and language outcomes of very preterm infants

Betty Vohr*

Department of Pediatrics, Women & Infants Hospital, Alpert Medical School of Brown University, Providence, RI, USA

Keywords:BrainHearing lossLanguagePreterm infantVery low birth weight infant

* Address: Department of Pediatrics, Women & IStreet, Providence, RI 02905, USA. Tel.: þ1 401 274 17571.

E-mail address: bvohr@wihri.org.

1744-165X/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.siny.2013.10.007

s u m m a r y

Speech and language impairments of both simple and complex language functions are common amongformer preterm infants. Risk factors include lower gestational age and increasing illness severityincluding severe brain injury. Even in the absence of brain injury, however, altered brain maturation andvulnerability imposed by premature entrance to the extrauterine environment is associated with brainstructural and microstructural changes. These alterations are associated with language impairments withlasting effects in childhood and adolescence and increased needs for speech therapy and educationsupports. Studies are needed to investigate language interventions which begin in the neonatal intensivecare unit.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Speech and language impairments are common in preterm in-fants with delays in the acquisition of expressive language, recep-tive language processing, and articulation, and deficits inphonological short-term memory [1e6]. Factors which may beassociated include low gestational age, increased illness severity,neonatal morbidities including brain injury, duration of hospitali-zation, hearing status, gender, age of assessment, socio-economicrisk factors and environment [7e10]. Language delay may beassociated with additional neurodevelopmental and neurosensorymorbidities, particularly hearing loss and cerebral palsy. Proficientlanguage skills are critical for the development of appropriatecommunication, joint attention, and social interactions [11]. Thisreview explores possible mechanisms that contribute to languagedelays and impairments and reports on outcomes of contemporarypopulations.

2. Origins of language and the language environment

Auditory input is critical for the development of speech and theauditory cortex [12]. The maturational progression of the auditorysystem occurs in utero with the perception, reaction, and storing ofauditory information including maternal physiologic sounds andvoice at approximately 26 weeks of gestation.

nfants’ Hospital, 101 Dudley122x47435; fax: þ1 401 453

All rights reserved.

Exposure of preterm infants to maternal sounds in a neonatalintensive care unit (NICU) compared to routine nursery sounds hasbeen shown to be associated with reduced frequency of apnea andbradycardia [13]. Both term and preterm infants respond prefer-entially to their mother’s voice, [14e16] the preference beingdemonstrated within hours of birth [15]. So what is the extra-uterine learning environment in a busy NICU for the infant who isdelivered at 26 weeks, intubated, and placed on a ventilator? Thereis an immediate change in the sound environment with loss of theintrauterine controlled sound environment to the noise of the de-livery room and NICU. Several studies [12,13,16] have shown thatexposure to maternal sounds and voice are significantly reduced inthe NICU. Deprivation of maternal sounds in the NICU during thisimportant period of auditory system development has been shownto impact both auditory brain maturation and subsequent speechand language [17,18]. Caskey et al. [16] reported that very preterminfants begin to make vocalizations in the NICU at 32 weeks andincrease their vocalizations between 32 and 36 weeks. Medianpercent exposure times to sounds captured in 16 h recordings were1% and 5% for total language, 19% and 36% for electronic soundsincluding monitors, 39% and 27% for silence, and 25% and 29% fornoise at 32 and 36 weeks, respectively. Adult word counts andmothereinfant conversation turns were significantly higher whenparents were visiting compared to nurses caring for the infant.There were also a higher number of conversational turns (parentand child vocalization within 5 s) at 32 weeks when a parent wasgiving the feeding compared with a staff nurse. This is highlysuggestive that parent interactions with their infants in the NICUplay an important role in early language stimulation. Currently,however, there is limited information about the effects of the lan-guage environment in the NICU on language outcomes.

B. Vohr / Seminars in Fetal & Neonatal Medicine 19 (2014) 78e83 79

3. The immature brain versus the injured brain

Preterm infants have reductions in cerebral tissue volume atterm equivalent compared to term infants [19,20]. Neonatal riskfactors associated with this decrease in cerebral tissue volumeinclude gestational age, dexamethasone therapy, brain injury, in-trauterine growth restriction, and bronchopulmonary dysplasia e

all risk factors associated with adverse outcomes [21e24].The language centres of the brain are located predominantly in

Broca’s area and Wernicke’s area of the left hemisphere [25]. Braininjury in these areas is a strong predictor of subsequent cognitiveand language impairments [26e28]. Children born preterm alsohave increased workingmemory deficits which have been linked tolanguage delays [29]. Working memory deficits at 2 years wereassociated with smaller hippocampal brain volumes at termequivalent in preterm infants<1250 g and persisted after adjustingfor perinatal risk, socio-economic status, and developmental fac-tors [30]. At 12 years both memory scores and cortical volumessubserving language and memory were reduced on magneticresonance imaging (MRI) in a second cohort [31]. Peterson et al.[32] reported decreased cortical sensorimotor, premotor, mid-temporal, parietal, occipital, and subgenual regions in pretermcompared to term controls at school age. Larger sensorimotor areasand mid-temporal cortices were associated with higher full-scale,verbal and performance IQ.

Cortical development of the temporal lobe is differentiallyvulnerable in preterm infants. Preterm children have significantlygreater bilateral temporal lobe gyrification compared to termcontrols [33]. Higher left temporal gyrification index has beencorrelated with lower reading recognition scores, a marker forlanguage skills at 8 years of age. Since gyrification begins during thethird trimester, the extrauterine environment may impact on thisdevelopmental process.

In addition to volumetric differences, there is increasing evi-dence of altered microstructure and connectivity in the brains ofpreterm infants. At 8 years of age preterm subjects in the Indo-methacin Trial had impaired performance on semantic test tasks,and used different pathways than term children [34]. Alterations infunctional connectivity for language tasks were identified in sub-sets of the population [35]. It was speculated that plasticity ofnetwork connections provides the opportunity for improving basiclanguage skills with increasing age among preterm children.

There are direct associations between specific areas of brainmicrostructure and developmental functions. Counsell et al. [36]performed diffusion tensor imaging studies and developmentalassessments at 2 years of age in preterm children without any focalabnormality on conventional MRI and reported that developmentalimpairments were associated with specific brain microstructuralabnormalities, namely lower fractional anisotropy (FA). FA is asummary measure of microstructure as it assesses water diffusivityin tissue which can reflect cell, axonal and myelination integrity.Microstructure abnormalities were also identified in a cohort ofpreterm children at 12 years of age with no major neonatal braininjury and no ventriculomegaly [37]. Diffusion tensor imagingidentified decreased FA in fibre tracts of regions subserving lan-guage. Values in the left anterior uncinate correlatedwith verbal IQ,full-scale IQ and Peabody Picture Vocabulary test (PPVT) scores forpreterm boys. Preterm boys had the lowest FA values in the rightanterior uncinate fasciculus. FA values in this region also correlatedwith verbal IQ and PPVT for preterm boys. These fibres contributeto the temporal stem. Other reports have shown reorganization ofpathways subserving lexical semantic processing [38], languageprocessing [39], phonological tasks [40], auditory language tasks[41], and auditory sentence comprehension [42] in preterm ado-lescents. These studies indicate that preterm birth places the infant

brain at increased risk of gray and white matter injury, and that,even in the absence of injury, brain development is altered withsignificant structural and microstructural changes which are asso-ciated with the neurodevelopmental impairments. The temporallobe and adjacent regions, which are centers for language devel-opment, are particularly vulnerable.

4. Assessment of speech and language

Traditionally, there are two methods of language assessment:simple, which includes vocabulary words and short phrases, andcomplex, which includes an expanded spectrum of languagecomponents including wording, and meaning of concepts, use ofverbs and relational terms, and complex sentences. Early pre-vocalizations heard in preterm infants are a form of simple lan-guage. Language also has subcategories of semantics (meaning),grammar (language structure), phonological awareness (under-standing of sounds), discourse (integrating information in conver-sations), and pragmatics (use of language appropriate to context).

5. Language outcomes

A representative sample of language outcome studies of pre-term infants published since 2000 are shown in Table 1. Publicationof the Bayley III with a separation of cognitive and language com-posite scores resulted in a series of studies reporting early languageskills in preterm infants. A National Institutes of Child Health andHuman Development (NICHD) Neonatal Research Network study[43] reported Bayley III language scores [44] of extremely preterminfants of 401e1000 g who also had a gestational age <27 weeks.The mean language composite score of infants evaluated at 18e22months of corrected age (CA) was 83 � 18, and 20% had a languagecomposite score <70. This rate of language impairment is consis-tent with prior reports of preterm toddlers [3,45]. Duncan et al. [46]examined effects of race and ethnicity on Bayley III language scoresof preterm infants <28 weeks. Children who were black and His-panic had similar cognitive scores but lower language scores thanwhite children. The authors note that the Bayley III has no stan-dardized Spanish version and therefore may provide a bias againstnon-English speaking children. The findings indicate, however, thatminority status suggests vulnerability for language delay. Loweet al. [47] explored language and ethnicity further by comparing thelanguage outcomes of preterm children whose primary languagewas Spanish compared to those whose primary language was En-glish. Although cognitive scores were similar for the two groups,Bayley III language scores were significantly lower for childrenwhose primary language was Spanish. A third study examinedoromotor control in infants �26 weeks [48]. Dysfunctional feedingat 18 months CA was defined as any of the following: physicianorder of no oral feeds, gastrostomy feeds, cough/gag/choking dur-ing oral feeds, aspiration, excessive drooling during feeds, or diffi-culty swallowing. Children with dysfunctional feeding hadsignificantly lower cognitive and language scores compared tothose with normal feeding.

These early language delays indicate a need for support services.Hintz et al. [49] reported that 33.7% of extremely preterm infants at18e22 months CA received speech therapy and 55.8% receivedearly intervention (EI). Rates of receiving speech therapy servicesranged from 41.2% for infants born �24 weeks to 25.6% for infantsborn at 27 weeks.

Meta-analyses consistently report speech and language delaysof preterm children compared to term children [1,6]. A meta-analysis of preterm children aged 3e12 years [6] identified thatpreterm children scored significantly lower than term children onboth simple and complex language function tests and that preterm

Table 1Language outcomes of preterm infants from corrected age (CA) of 18 months to age 12 years.

Study Population Birth years Age Test Outcome

Duncan et al. [46] NICHD<28 weeks

Jan 2008eJune 2009 18e22 m CA Bayley III White Black Hispanic PCognitive composite 91.9 � 15 88.2 � 14 88.2 � 14 0.0009Language composite 89.7 � 17 81.8 � 16 79.2 � 17 0.0001

Lowe et al. [47] NICHD<28 weeks

Jan 2008eJune 2009 18e22 m CA Bayley III Primary languageEnglish Spanish P

Cognitive composite 90.0 � 15 88.3 � 13 0.29Language composite 85.6 � 18 78.7 � 14 0.001Receptive 7.7 � 3 6.1 � 2.5 0.01Expressive 7.8 � 3 6.8 � 2 0.001

Adams-Chapmanet al. [48]

NICHD�26 weeks

Jan 2006eMarch 2008 18e22 m CA Bayley III Normal feeding Dysfunctional PCognitive composite 92 � 13 77 � 8 0.001Language composite 87 � 16 73 � 20 0.001

Wolke et al. [50] NationalCohort<26 weeks

1995 6 years Extreme preterm Comparison PPLS 3 Boys 85.1 � 22* 102 8 � 8 0.001

Girls 93.6 � 17 104.8 � 12 0.01Impaired articulation Boys 13.8%a 4.2% 0.05

Girls 4.8% 1.1%Reidy et al. [51] Australia

<30 weeks2001e2003 7 years NEPSY and CELF scores

below �1.25 SDPreterm Term P

Phonologicalawareness

22.8% 4.3% 0.001

Semantics 20.3% 1.4% 0.001Grammar 23.4% 2.9% 0.001Discourse 30.1% 5.7% 0.001Pragmatics 27.8% 14.6% 0.04

Luu et al. [4] 600e1250 1989e1992 3, 4.5, 6, 8,12 years

PPVT Age (years): 3 4.5 6 8 12Boys indo 87 � 21 87 � 20 97 � 20 95 � 23 97 � 23Boys placebo 78 � 22 79 � 30 87 � 30 90 � 30 92 � 28Girls indo 83 � 22 88 � 22 93 � 25 91 � 25 92 � 25Girls placebo 89 � 17 90 � 22 92 � 22 93 � 22 97 � 22

Luu et al. [52] 600e1250 1989e1992 12 years Preterm Term Adjusted mean differencea

Verbal IQ 90.8 � 19 103.9 � 16 �8.9 (�12.0 to �5.7)CELF receptive 86.9 � 21 102.0 � 6 �11.7 (�15.4 to �7.9)CELF expressive 85.8 � 20 99.5 � 15 �9.9 (�13.4 to �6.4)CELF total 83.4 � 20 100.7 � 16 �11.2 (�14.7 to �7.7)PPVT-R 92.0 � 25 105.4 � 21 �8.7 (�12.8 to �4.5)

Luu et al. [5] 600e1250 1989e1992 16 years Impaired <2 SD Preterm Term Adjusted mean differenceVerbal IQ 12% 3% �9.4 (�15.1 to �3.7)PPVT-R 13% 3% �5.5 (�12.3 to 1.3)CTOPPRapid naming 12% 3% �3.9 (�10.9 to 3.2)Phonological awareness 18% 2% �5.1 (�10.1 to �0.1)Phonemic decoding 14% 2% �5.6 (�10.5 to �0.7)

Northam et al. [53] <33 weeks 1989e1994 16 years Focal oromotor controlimpairment

31% had problems in oromotor and speech-motor control

NICHD, National Institutes of Child Health and Human Development; CA, corrected age; PLS, Preschool Language Scale; NEPSY-A, Developmental NEuroPSYchologicalAssessment; CELF, Clinical Evaluation of Language Fundamentals; PPVT-R, Peabody Picture Vocabulary Test e Revised; indo, indomethacin; CTOPP, Comprehensive Test ofPhonological Processing.

a Brain injury excluded.

B. Vohr / Seminars in Fetal & Neonatal Medicine 19 (2014) 78e8380

children had increasing difficulty with complex language withincreasing age. Differences persisted after exclusion of major dis-abilities and effects of social economic status [6].

Additional school-age outcome studies are shown in Table 1.Assessment of infants <26 weeks of gestation at age 6 yearsidentified significantly lower language scores for both boys andgirls than sex-matched term children, and higher rates of impairedarticulation [50]. In addition, preterm boys had significantly lowerscores than preterm girls and higher rates of impaired articulation(13.8% vs 4.8%, respectively). At age 7 years, children <30 weeks ofgestation performed more poorly than term controls on all lan-guage subdomains [51]. White matter abnormality mediated theeffect of group differences on phonological awareness, and partlymediated this effect for semantics, grammar, and discourse. Thereis some evidence of recovery of simple language function, specif-ically receptive vocabulary with increasing age. One of the advan-tages of the reports of outcomes of preterm children in theIndomethacin Trial is longitudinal data on both preterm childrenand term controls. Improvement of PPVT scores (simple languagefunction) for both boys and girls in both arms of the Indomethacin

Trial was identified between 3 and 12 years of age [4]. Recovery ofreceptive vocabulary was associated with higher maternal educa-tion and non-minority status, whereas severe brain injury wasassociated with slower gains with increasing age.

Specific language deficits in preterm children that can persist atschool age include phonological short-term memory, and prosodicprocessing. At 12 years of age, children who were born at <1250 gin the Indomethacin Trial had fewer differences in lower levellanguage skill tests (phonological processing, phonemic decoding,and sight word reading) compared to term controls, but exhibitedmore difficulty with higher level skills (syntax, semantics, verballanguage memory) [4,52]. Despite improvements, they continuedto have lower scores on the PPVT (92 vs 105) and higher rates ofvocabulary impairment (13% vs 4%), as well as lower scores (85e87vs 100e103) and higher rates of impairment (22e24% vs 3e4%) onthe Clinical Evaluation of Language Fundamentals (CELF) comparedto term controls [4]. Differences in test scores persisted afterexclusion of children with severe brain injury. In multivariate an-alyses severe brain injury was the strongest predictor of test scores.Educational resource needs were analysed by presence or absence

Practice points

� All former preterm infants with and without brain injuryshould be monitored post-discharge for speech and lan-guage delays, impairments, and HL.

� Preterm infants should be referred for early interventionand speech/language services as needed.

� Evidence supports recovery of both language and brainstructural and microstructural systems that subservelanguage in optimal environments.

� Needs for speech language resources should be moni-tored and re-evaluated as needed in preterm school-agechildren.

Research directions

� Longitudinal studies that begin in the NICU and extend toschool age are needed to identify interventions thatcontribute to optimal language outcomes for pretermchildren.

Table 2Hearing.

Author Population Years Age (years) Hearing outcome 22 weeks 23 weeks 24 weeks 25 weeks

Hintz et al. [62] NRN USA 1999e20012002e2004

18e22 Bilateral amplification 3.2%4.9%

1.6%2.9%

Ishii et al. [63] Japan 2003e2005 36e42 Bilateral amplification 0.0% 3.4% 1.2% 1.3%Moore et al. [64]

EPICureUK EPICure 2006 36 Severe: profound/not improved with amplification 3% 0% 0%

Moderate: improved with amplification 5% 5% 5%Morris et al. [60]

NRNNRN USA <1000 g 18e22 Bilateral amplification Phototherapy Mean gestation 26 weeks

Conservative 3%Aggressive 1%

NRN, Neonatal Research Network.

B. Vohr / Seminars in Fetal & Neonatal Medicine 19 (2014) 78e83 81

of severe neonatal brain injury on cranial ultrasound. Rate ofspeech and language therapy was 50% for preterm children withhistory of brain injury, 11% for preterm with no history of braininjury, and 2% for term controls.

Luu et al. [5] reported continued catch-up in receptive vo-cabulary between 8 and 16 years of age. However, preterm ad-olescents continued to have deficits in higher order languageskills (phonological awareness: 18% vs 2%; phonemic decoding:14% vs 2%) compared to term controls, respectively, at age 16years. A subgroup (65%) of very preterm children displayeddevelopmental trajectories of receptive vocabulary similar toterm controls between 8 and 16 years [4]. Absence of neuro-sensory impairments, higher level of maternal education, andresiding in a two-parent household were associated with catch-up for preterm children, suggesting that a more optimallearning environment is beneficial.

At age 16 years, Northam et al. [53] identified persistent prob-lems in oromotor control including precision of individual andcombined movements of the lips, jaw, face and tongue. In multi-variate analyses, neurologic impairment and abnormalities of theprimary motor projections of the corticospinal tract and speech-motor corticobulbar tract were associated with poor speech andoromotor outcome.

Congenital hearing loss (HL) is an important predictor of lan-guage delay. Prior to the 1990s children with permanent HL werenot identified in the newborn period and did not receive earlyintervention services [54]. They steadily fell behind hearing peersin language, cognitive and academic skills. Studies have shown thatearly identification and early access to meaningful language canoffset the detrimental impact of HL on language, socialeemotionalskills, and academics [55e59]. An association between moreoptimal maternal communicative interaction and higher vocabu-lary scores for children with HL at 18e24 months of age wasidentified [59]. In the era of universal newborn hearing screeningand earlier identification, additional factors must be considered,including age of early intervention, age of amplification, and degreeof HL. Preterm infants cared for in a NICU are at greater risk of latediagnosis and intervention because of illness severity and pro-longed hospitalization.

Bilateral hearing impairment requiring amplification in recentreports ranges from 1% to 9% of ELBW infants. Rates are impacted bythe age of assessment since preterm infants are at increased risk ofboth late onset and/or progressive HL. In the NICHD PhototherapyTrial [60], rates of bilateral hearing impairment with amplificationwere 1% in the aggressive phototherapy arm and 3% in the con-servative phototherapy arm. In a Canadian study [61] of preterminfants �800 g, the overall rate of bilateral permanent HL was 9%(50/586 infants). As shown in Table 2, rates of permanent bilateralHL are higher at lower gestational age. There is a need for studiesshowing the language outcomes of early identified very preterminfants with bilateral HL.

6. Conclusions

In summary, speech and language delays are a common form ofdelay seen in very preterm infants. Studies suggest that impair-ments remain evident in school age and adolescence, particularly ofmore complex language functions. Brain injury, altered brainmaturation, illness severity, co-morbidities, and non-optimal earlyenvironment in the NICU all contribute to language delays andimpairments. Early language interventions for infants withcongenital HL have already been shown to be beneficial. MRIstudies indicate that there is a potential for both recovery of brainstructure and microstructure differences, and longitudinal neuro-developmental studies indicate the potential for language recovery[5]. These findings are encouraging and provide evidence for bothbrain structure and microstructure plasticity and developmentalplasticity. Early and continued identification, support, intervention,and remediation are needed to optimize language outcomes ofpreterm infants.

Conflict of interest statement

None declared.

Funding sources

None.

B. Vohr / Seminars in Fetal & Neonatal Medicine 19 (2014) 78e8382

References

[1] Barre N, Morgan A, Doyle LW, Anderson PJ. Language abilities in childrenwhowere very preterm and/or very low birth weight: a meta-analysis. J Pediatr2011;158:766e74. e761.

*[2] Foster-Cohen SH, Friesen MD, Champion PR, Woodward LJ. High prevalence/low severity language delay in preschool children born very preterm. J DevBehav Pediatr 2010;31:658e67.

[3] Ortiz-Mantilla S, Choudhury N, Leevers H, Benasich AA. Understanding lan-guage and cognitive deficits in very low birth weight children. Dev Psycho-biol 2008;50:107e26.

*[4] Luu TM, Vohr BR, Schneider KC, Katz KH, Tucker R, Allan WC, et al. Trajec-tories of receptive language development from 3 to 12 years of age for verypreterm children. Pediatrics 2009;124:333e41.

*[5] Luu TM, Vohr BR, Allan W, Schneider KC, Ment LR. Evidence for catch-up incognition and receptive vocabulary among adolescents born very preterm.Pediatrics 2011;128:313e22.

[6] van Noort-van der Spek IL, Franken MC, Weisglas-Kuperus N. Languagefunctions in preterm-born children: a systematic review and meta-analysis.Pediatrics 2012;129:745e54.

[7] Gaffney M, Eichwald J, Grosse SD, Mason CA. Identifying infants withhearing loss - united states, 1999e2007. Morb Mortal Wkly Rep 2010;59:220e3.

[8] Clark C, Woodward L. Neonatal cerebral abnormalities and later verbal andvisuospatial working memory abilities of children born very preterm. DevNeuropsychol 2010;35:622e42.

[9] Grunau RV, Kearney SM, Whitfield MF. Language development at 3 years inpre-term children of birth weight below 1000 g. Br J Disord Commun1990;25:173e82.

[10] Marston L, Peacock JL, Calvert SA, Greenough A, Marlow N. Factors affectingvocabulary acquisition at age 2 in children born between 23 and 28 weeks’gestation. Dev Med Child Neurol 2007;49:591e6.

[11] Durkin K, Conti-Ramsden G. Language, social behavior, and the quality offriendships in adolescents with and without a history of specific languageimpairment. Child Dev 2007;78:1441e57.

[12] McMahon E, Wintermark P, Lahav A. Auditory brain development in pre-mature infants: the importance of early experience. Ann N Y Acad Sci2012;1252:17e24.

*[13] Doheny L, Hurwitz S, Insoft R, Ringer S, Lahav A. Exposure to biologicalmaternal sounds improves cardiorespiratory regulation in extremely preterminfants. J Matern Fetal Neonatal Med 2012;25:1591e4.

[14] Kisilevsky BS, Hains SM, Brown CA, Lee CT, Cowperthwaite B, Stutzman SS,et al. Fetal sensitivity to properties of maternal speech and language. InfantBehav Dev 2009;32:59e71.

[15] DeCasper AJ, Fifer WP. Of human bonding: Newborns prefer their mothers’voices. Science 1980;208:1174e6.

*[16] Caskey M, Stephens B, Tucker R, Vohr B. Importance of parent talk on thedevelopment of preterm infant vocalizations. Pediatrics 2011;128:910e6.

[17] deRegnier RA. Wewerka S, Georgieff MK, Mattia F, Nelson CA. Influences ofpostconceptional age and postnatal experience on the development ofauditory recognition memory in the newborn infant. Dev Psychobiol2002;41:216e25.

[18] Fifer WP, Moon CM. The role of mother’s voice in the organization of brainfunction in the newborn. Acta Paediatr Suppl 1994;397:86e93.

[19] Maalouf EF, Duggan PJ, Counsell SJ, Rutherford MA, Cowan F, Azzopardi D,et al. Comparison of findings on cranial ultrasound and magnetic resonanceimaging in preterm infants. Pediatrics 2001;107:719e27.

[20] Inder TE, Wells SJ, Mogridge NB, Spencer C, Volpe JJ. Defining the nature ofthe cerebral abnormalities in the premature infant: a qualitative magneticresonance imaging study. J Pediatr 2003;143:171e9.

[21] Inder TE, Warfield SK, Wang H, Huppi PS, Volpe JJ. Abnormal cerebralstructure is present at term in premature infants. Pediatrics 2005;115:286e94.

[22] Nosarti C, Al-Asady MH, Frangou S, Stewart AL, Rifkin L, Murray RM. Ado-lescents who were born very preterm have decreased brain volumes. Brain2002;125:1616e23.

[23] Thompson DK, Warfield SK, Carlin JB, Pavlovic M, Wang HX, Bear M, et al.Perinatal risk factors altering regional brain structure in the preterm infant.Brain 2007;130:667e77.

[24] Murphy BP, Inder TE, Huppi PS, Warfield S, Zientara GP, Kikinis R, et al.Impaired cerebral cortical gray matter growth after treatment with dexa-methasone for neonatal chronic lung disease. Pediatrics 2001;107:217e21.

[25] Armstrong E, Schleicher A, Omran H, Curtis M, Zilles K. The ontogeny ofhuman gyrification. Cereb Cortex 1995;5:56e63.

[26] Nosarti C, Walshe M, Rushe TM, Rifkin L, Wyatt J, Murray RM, et al. Neonatalultrasound results following very preterm birth predict adolescent behav-ioral and cognitive outcome. Dev Neuropsychol 2011;36:118e35.

[27] Skiold B, Vollmer B, Bohm B, Hallberg B, Horsch S, Mosskin M, et al. Neonatalmagnetic resonance imaging and outcome at age 30 months in extremelypreterm infants. J Pediatr 2012;160:559e66. e551.

*[28] Ment LR, Vohr B, Oh W, Scott DT, Allan WC, Westerveld M, et al. Neuro-developmental outcome at 36 months’ corrected age of preterm infants inthe multicenter indomethacin intraventricular hemorrhage prevention trial.Pediatrics 1996;98:714e8.

[29] Caravale B, Tozzi C, Albino G, Vicari S. Cognitive development in low riskpreterm infants at 3e4 years of life. Arch Dis Child Fetal Neonatal Ed2005;90. F474e9.

[30] Beauchamp MH, Thompson DK, Howard K, Doyle LW, Egan GF, Inder TE, et al.Preterm infant hippocampal volumes correlate with later working memorydeficits. Brain 2008;131:2986e94.

*[31] Fraello D, Maller-Kesselman J, Vohr B, Katz KH, Kesler S, Schneider K, et al.Consequence of preterm birth in early adolescence: the role of language onauditory short-term memory. J Child Neurol 2011;26:738e42.

[32] Peterson BS, Vohr B, Staib LH, Cannistraci CJ, Dolberg A, Schneider KC, et al.Regional brain volume abnormalities and long-term cognitive outcome inpreterm infants. JAMA 2000;284:1939e47.

[33] Peterson BS, Vohr B, Kane MJ, Whalen DH, Schneider KC, Katz KH, et al. Afunctional magnetic resonance imaging study of language processing and itscognitive correlates in prematurely born children. Pediatrics 2002;110:1153e62.

[34] Ment LR, Peterson BS, Vohr B, Allan W, Schneider KC, Lacadie C, et al. Corticalrecruitment patterns in children born prematurely compared with controlsubjects during a passive listening functional magnetic resonance imagingtask. J Pediatr 2006;149:490e8.

[35] Kesler SR, Vohr B, Schneider KC, Katz KH, Makuch RW, Reiss AL, et al.Increased temporal lobe gyrification in preterm children. Neuropsychologia2006;44:445e53.

[36] Counsell SJ, Edwards AD, Chew AT, Anjari M, Dyet LE, Srinivasan L, et al.Specific relations between neurodevelopmental abilities and white mattermicrostructure in children born preterm. Brain 2008;131:3201e8.

[37] Constable RT, Ment LR, Vohr BR, Kesler SR, Fulbright RK, Lacadie C, et al.Prematurely born children demonstrate white matter microstructural dif-ferences at 12 years of age, relative to term control subjects: an investigationof group and gender effects. Pediatrics 2008;121:306e16.

[38] Schafer RJ, Lacadie C, Vohr B, Kesler SR, Katz KH, Schneider KC, et al. Alter-ations in functional connectivity for language in prematurely born adoles-cents. Brain 2009;132:661e70.

[39] Gozzo Y, Vohr B, Lacadie C, Hampson M, Katz KH, Maller-Kesselman J, et al.Alterations in neural connectivity in preterm children at school age. Neuro-Image 2009;48:458e63.

[40] Mullen KM, Vohr BR, Katz KH, Schneider KC, Lacadie C, Hampson M, et al.Preterm birth results in alterations in neural connectivity at age 16 years.NeuroImage 2011;54:2563e70.

[41] Frye RE, Malmberg B, McLean 3rd J, Swank P, Smith K, Papanicolaou A, et al.Increased left prefrontal activation during an auditory language task in ad-olescents born preterm at high risk. Brain Res 2010;1336:89e97.

[42] Barde LH, Yeatman JD, Lee ES, Glover G, Feldman HM. Differences in neuralactivation between preterm and full term born adolescents on a sentencecomprehension task: implications for educational accommodations. DevCogn Neurosci 2012;2(Suppl. 1):S114e28.

[43] Vohr BR, Stephens BE, Higgins RD, Bann CM, Hintz SR, Das A, et al. Areoutcomes of extremely preterm infants improving? Impact of bayleyassessment on outcomes. J Pediatr 2012;161. 222e228.e3: 222e228.e223.

[44] Bayley N. Bayley scales of infant and toddler development. third ed. SanAntonio, TX: Harcourt Assessment, Inc; 2006.

[45] Herold B, Hohle B, Walch E, Weber T, Obladen M. Impaired word stresspattern discrimination in very-low-birthweight infants during the first 6months of life. Dev Med Child Neurol 2008;50:678e83.

[46] Freeman Duncan A, Watterberg KL, Nolen TL, Vohr BR, Adams-Chapman I,Das A, et al. Effect of ethnicity and race on cognitive and language testing atage 18-22 months in extremely preterm infants. J Pediatr 2012;160:966e71.e962.

[47] Lowe JR, Nolen TL, Vohr B, Adams-Chapman I, Duncan AF, Watterberg K.Effect of primary language on developmental testing in children bornextremely preterm. Acta Paediatr 2013;102:896e900.

*[48] Adams-Chapman I, Bann CM, Vaucher YE, Stoll BJ. Association betweenfeeding difficulties and language delay in preterm infants using Bayley scalesof infant development e third edition. J Pediatr 2013.

[49] Hintz SR, Kendrick DE, Vohr BR, Poole WK, Higgins RD. Community supportsafter surviving extremely low-birth-weight, extremely preterm birth: specialoutpatient services in early childhood. Arch Pediatr Adolesc Med 2008;162:748e55.

*[50] Wolke D, Samara M, Bracewell M, Marlow N. Specific language difficultiesand school achievement in children born at 25 weeks of gestation or less. JPediatr 2008;152:256e62.

*[51] Reidy N, Morgan A. Thompson DK, Inder TE, Doyle LW, Anderson PJ. Impairedlanguage abilities and white matter abnormalities in children born verypreterm and/or very low birth weight. J Pediatr 2013;162:719e24.

[52] Luu TM, Ment LR, Schneider KC, Katz KH, Allan WC, Vohr BR. Lasting effectsof preterm birth and neonatal brain hemorrhage at 12 years of age. Pediatrics2009;123:1037e44.

[53] Northam GB, Liegeois F, Chong WK, Baker K, Tournier JD, Wyatt JS, et al.Speech and oromotor outcome in adolescents born preterm: Relationship tomotor tract integrity. J Pediatr 2012;160:402e8. e401.

[54] Thompson DC, McPhlllips H, Davis R, Lieu T, Homer CJ, Helfand M. Universalnewborn hearing screening: summary of evidence. JAMA 2001;286:24e31.

[55] Watkin P, McCann D, Law C, Mullee M, Petrou S, Stevenson J, et al. Languageability in children with permanent hearing impairment: the influence of

B. Vohr / Seminars in Fetal & Neonatal Medicine 19 (2014) 78e83 83

early management and family participation. Pediatrics 2007;120:e694e701.

[56] Yoshinaga-Itano C, Sedey AL, Coulter DK, Mehl AL. Language of early- andlater-identified children with hearing loss. Pediatrics 1998;102:1161e71.

[57] Moeller MP. Current state of knowledge: psychosocial development in chil-dren with hearing impairment. Ear Hear 2007;28:729e39.

[58] Vohr B, Jodoin-Krauzyk J, Tucker R, Johnson MJ, Topol D, Ahlgren M. Earlylanguage outcomes of early-identified infants with permanent hearing lossat 12 to 16 months of age. Pediatrics 2008;122:535e44.

[59] Vohr B, Pierre LS, Topol D, Jodoin-Krauzyk J, Bloome J, Tucker R. Associationof maternal communicative behavior with child vocabulary at 18e24 monthsfor children with congenital hearing loss. Early Hum Dev 2010;86:255e60.

[60] Morris BH, Oh W, Tyson JE, Stevenson DK, Phelps DL, O’Shea TM, et al.Aggressive vs. Conservative phototherapy for infants with extremely lowbirth weight. N Engl J Med 2008;359:1885e96.

[61] Synnes AR, Anson S, Baum J, Usher L. Incidence and pattern of hearingimpairment in children with </¼ 800 g birthweight in british columbia,canada. Acta Paediatr 2012;101:e48e54.

[62] Hintz SR, Kendrick DE, Wilson-Costello DE, Das A, Bell EF, Vohr BR, et al.Early-childhood neurodevelopmental outcomes are not improving for in-fants born at <25 weeks’ gestational age. Pediatrics 2011;127:62e70.

[63] Ishii N, Kono Y, Yonemoto N, Kusuda S, Fujimura M. Outcomes of infants bornat 22 and 23 weeks’ gestation. Pediatrics 2013;132:62e71.

[64] Moore T, Hennessy EM, Myles J, Johnson SJ, Draper ES, Costeloe KL, et al.Neurological and developmental outcome in extremely preterm childrenborn in England in 1995 and 2006: the epicure studies. BMJ 2012;345.e7961.

lable at ScienceDirect

Seminars in Fetal & Neonatal Medicine 19 (2014) 84e89

Contents lists avai

Seminars in Fetal & Neonatal Medicine

journal homepage: www.elsevier .com/locate/s iny

Review

Cerebral palsy and developmental coordination disorder in childrenborn preterm

Alicia Jane Spittle a,b,c,*, Jane Orton c

aVictorian Infant Brain Studies, Murdoch Childrens Research Institute, Parkville, AustraliabDepartment of Physiotherapy, School of Health Sciences, University of Melbourne, Carlton, AustraliacNeonatal Services, Royal Women’s Hospital, Parkville, Australia

Keywords:Cerebral palsyDevelopmental coordination disorderInfantMotor developmentPreterm

* Corresponding author. Address: VIBeS, MurdochFlemington Road, Parkville, Victoria 3052, Australia. Te3 93454836.

E-mail address: alicia.spittle@mcri.edu.au (A.J. Spi

1744-165X/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.siny.2013.11.005

s u m m a r y

Children born early (<37 weeks of gestation) are at high risk of a range of motor impairments due to avariety of biological and environmental risk factors. Cerebral palsy occurs more frequently in thosechildren born preterm, with the risk increasing with decreasing gestational age. Mild and moderatemotor impairments, consistent with developmental coordination disorder, occur in almost half of thosechildren born preterm and include difficulties with balance, manual dexterity and ball skills. All forms ofmotor impairment are associated with comorbidities, which may have a greater effect on quality of life,academic achievement and participation in extracurricular activities than the motor impairment itself.Infants at risk of motor impairment can be identified in early infancy with a combination of clinicalassessment tools and perinatal risk factors. However, the reliable diagnosis of motor impairment requiresfollow-up into early childhood and it is important to ensure that the appropriate intervention isimplemented.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Motor difficulties are one of the most frequently reportedadverse neurodevelopment impairments in children born preterm[1,2]. The range of motor impairment can vary from mild, such asdelays in crawling and walking, to the most severe motor disabilityof childhood, cerebral palsy (CP). Children born preterm (<37weeks of gestation) have a different trajectory of motor develop-ment compared with those children born at full term [3]. Earlyexposure to the extrauterine environment results in alteredmovement (e.g. gravity) and sensory (e.g. light and sound) expe-riences on the developing musculoskeletal and central nervoussystems [2,4]. Further, the birth of an infant in the late second orearly in the third trimester during periods of rapid brain develop-ment is thought to disrupt the genetically programmed pattern ofbrain genesis [5]. Postnatally, biological factors that may influencemotor development include insufficient growth (reduced weight,height, and head circumference) and smaller muscle size (with alower proportion of fast-twitch fibres), along with further alter-ations in brain maturation [3,5,6]. In addition, environmental

Childrens Research Institute,l.: þ61 3 8345 3778; fax: þ61

ttle).

All rights reserved.

influences such as parentechild interactions, expectations andexperiences may also be affected by the birth of a child prematurely[7]. A preterm infant at term age presents with different neuro-behavioral and motor strategies than an infant born at term andmay exhibit less flexor muscle tone, use more extended postures,have monotonous spontaneous movements against a backgroundof hypotonicity and have particular difficulty with antigravitymovements [2,8]. A meta-analysis of motor outcomes for childrenvery preterm (VPT defined as<32 weeks of gestational age) or verylow birth weight (VLBW defined as �1500 g) reported that,compared with children born at term, VPT or VLBW children scoredalmost one standard deviation (SD) lower [mean difference: �0.88;95% confidence interval (CI): �0.96 to �0.80] on the psychomotorscale of the Bayley Scales of Infant Development during infancy (6e36 months) [1]. By school age (5e15 years) children born VPT orVLBW were still performing lower than their peers on two of themost widely used measures of motor impairment, the MovementAssessment Battery of Children (MABC) (mean difference: �0.62;95% CI: �0.69 to�0.55) and the BruininkseOseretsky Test of MotorPerformance (BOTMP) (mean difference: �0.53; 95% CI: �0.60to �0.46) [1].

Research on long-term motor outcomes is limited by inconsis-tency in classification of motor impairments, selection of pop-ulations of infants (by birthweight and/or gestational age) and toolsused to determine impairment, making the findings difficult toreadily synthesize [7,9,10]. Whereas it is evident that many children

A.J. Spittle, J. Orton / Seminars in Fetal & Neonatal Medicine 19 (2014) 84e89 85

born preterm have variations in their motor developmentcompared with peers born at term, it is not always clear whetherthese variations result in long-term motor impairments [3,11]. It isessential to understand and distinguish the range of motor im-pairments experienced by children born preterm including mildmotor impairment and developmental coordination disorder (DCD)from CP, so that the most appropriate interventions can be imple-mented early and caregivers have an understanding of long-termmotor function [12].

2. Cerebral palsy

Cerebral palsy is primarily a motor disorder and is an um-brella term ‘to describe a group of disorders of the developmentof movement and posture, causing activity limitations, which areattributed to non-progressive disturbances that occurred in thedeveloping fetal or infant brain’ [13]. The prevalence of CP in thegeneral population varies slightly between countries and is about0.1e0.2% of live births [10,14]. A recent review of CP registries inAustralia from 1970 to 2004 reported that the prevalence of CPincreased in the 1970s and 1980s due to the increasing survivalof extremely preterm (EP: defined as <28 weeks of gestationalage) infants that occurred without a concomitant improvementin neurological outcomes, but since the early 1990s CP rateseither stabilized or decreased [14]. The risk of developing CPincreases with decreasing gestational age. A comprehensivemeta-analysis by Himpens et al. [10] including 25 studies re-ported the prevalence as 14.6% (95% CI: 12.5e17) at 22e27weeks, 6.2% (4.9e7.8) at 28e31 weeks and 0.7% (0.6e0.9) at 32e36 weeks of gestation.

Children with CP have a variety of clinical presentations and arange of functional abilities from walking independently torequiring full physical assistance for all activities [15,16]. Given thevariety of motor presentations of a child with CP, several classifi-cation systems are used to describe a child’s CP including motortype, topography, functional severity and comorbidities [12,13].Preterm childrenwith CP differ from their term-born peers with CPwith respect to motor type and topography. Traditionally, motortype is referred to as spastic, dyskinesia, ataxia and hypotonia, andtopography as hemiplegia, diplegia and quadriplegia [12]. However,due to the low inter-rater reliability of these classifications, manyregisters now describe topographies as two easily distinguishablegroups, unilateral versus bilateral, and motor type as spastic versusnon-spastic [17]. The meta-analysis by Himpens et al. [10] of CPpresentation in term and preterm children reported that the spasticform of CP was the predominant motor type in preterm and termchildren (96% and 82% respectively). Preterm children were morelikely to present with bilateral CP at a rate of 73% versus 48% interm-born children. Spastic diplegiawas themost frequent subtypeand distribution in the preterm child (60%) and occurred atapproximately three times the rate of spastic quadriplegia, the nextmost common CP type in preterm children [10].

When describing functional outcomes of children with CP, theGross Motor Function Classification System (GMFCS) is consideredthe gold standard, as it is a reliable and validated measure toclassify motor function in terms of activity levels in those with CP[18]. It describes the movement ability of children across fiveordinal levels with a focus on self-initiated movement of sitting,transferring and mobilising. Children classified as level I are able towalk without restrictions but may have difficulty in the speed,balance and coordination of higher-level skills, whereas childrenclassified as level V have significant difficulty with any voluntarycontrol of movement and limited mobility, even with assistedtechnology. The proportion of preterm children with CP with themildest motor impairment, i.e. GMFCS level I, increases with

gestational age [19]. Between one-half and two-thirds of childrenborn preterm with CP are level I and II, and therefore can walkindependently [19e21]. However, most studies reporting on long-term outcomes of preterm birth do not use the GMFCS andinstead use the classification system of mild, moderate or severe,which has not been operationally defined and has poor inter-raterreliability [10,11]. Using this description, the majority of researchreports that approximately half of children born preterm with CPare classified as mild CP, with the remaining having moderate orsevere CP.

3. Motor impairments in children without CP

Motor deficits in coordination, balance, gross and fine motorcontrol and visual motor integration are reported in pretermchildren without CP [9,22]. These motor deficits may be classifiedas DCD if the child meets four criteria according to the DSM-V,including: (a) motor coordination and performance are belowthat expected for the child’s chronological age and intelligencelevel; (b) the motor disorder interferes with activities of dailyliving or academic achievement; (c) it is not due to generalmedical or neurological condition; and (d) if intellectualdisability is present, the motor difficulties are in excess of thoseassociated with it. The World Health Organization (WHO) alsohave criteria which are similar including that the motorimpairment interferes with academic performance and/or ac-tivities of daily living and that the child should not have adiagnosable neurological disorder [23]. The WHO provides morespecific criteria for classification of motor impairment defined asa child scoring 2 SD below the mean on a standardized test offine or gross motor coordination and exclusion of children withIQ < 70. In the general population the incidence of DCD forschool-age children is reported to be 5e6%, whereas the rate ismuch higher in those children born preterm [24]. In a large studyof more than 20 000 Danish children, gestational age at birth wasinversely associated with the risk of probable DCD, assessedusing the DCD-Questionnaire, with a decline in gestational ageby a week associated with a 19% (95% CI: 14e25) increased risk ofpositive DCD screening among children delivered before 40weeks [25]. It is unclear in the literature whether the rates ofDCD have changed over time for children born preterm. This ispartly due to the various descriptions of these children includingclumsy, having developmental dyspraxia, handwriting problems,handeeye coordination problems, sensory integration dysfunc-tion, deficits in attention, motor control and perception, minorneurological dysfunction, among many others [26]. Since 1994,when an international consensus meeting agreed to use the termDCD to classify these children, there has been an increase inknowledge and awareness of DCD, but a standardized approachhas not yet been achieved [26,27].

To meet the criteria for DCD it is necessary to assess not onlymotor function but also activities of daily living and intelligence.The majority of research reports on motor impairment, classifiedby the child falling below a certain cut-off point on a standard-ized motor assessment; however, the effect of the impairmenton activities of daily living and relationship with intelligencelevel are often not described. For these reasons there is variationin the literature on the use of the term ‘DCD’ and ‘non-CP motorimpairment’ when reporting on long-term motor outcomes ofpreterm children. Further, there is also debate as to whetherchildren born preterm should be diagnosed with DCD as theyhave a medical and/or neurological deficit that can explain theirmotor impairment (criterion ‘c’ of DSM-V criteria) [28]. However,the majority of literature on DCD does not exclude children whowere born preterm and therefore in this review we refer to the

A.J. Spittle, J. Orton / Seminars in Fetal & Neonatal Medicine 19 (2014) 84e8986

literature on DCD and non-CP motor impairment. In childrenwithout CP, mild motor impairment often refers to children whoscore between the 5th and 15th centile or 1 SD below the mean,and moderate as <5th centile or 2 SD below the mean onstandardized motor assessments. A meta-analysis on motoroutcomes of preterm children born <37 weeks of gestation atschool age reported the rate of mildemoderate motor impair-ment to be 41% (95% CI: 32.1e48.9) and moderate motorimpairment to be 19% (14.2e23.8) [24]. Another meta-analysisfound that children born VPT or VLBW are six times (OR: 6.3;95% CI: 4.4e9.1) more likely than their term-born peers to have amoderate motor impairment and almost nine times (8.7; 3.4e22.1) more likely to have a mild motor impairment in theabsence of CP [22].

Clinicians often overlook mildemoderate motor impairmentsin children born preterm, but there is extensive evidence that themotor skill difficulties experienced by these children affect howthey function [27,29e31]. The types of motor problems includedifficulties in gross and/or fine motor skills including balanceskills, ball skills and manual dexterity, consistent with the motorimpairments seen in children with DCD [1]. Balance skills are re-ported to be the most impaired, and, in combination with theother motor problems, preterm children with mild or moderatemotor impairment can be challenged with to respect to crawling,walking, running and jumping in the early years. At school age,difficulties with activities ranging from running to drawing canaffect physical, social and academic performance [31,32]. Impair-ments continue into adolescence and adulthood, highlighting thatthese motor problems are not merely a delay that a child willoutgrow [31,32].

4. Comorbidities associated with motor impairment

Although CP and DCD are primarily motor disorders they areoften associatedwith comorbidities, which can have a greater effecton function and quality of life than themotor impairment itself. Thehigh rate of comorbidities is of clinical importance when counsel-ling families about the long-term implications of motor impair-ments and planning interventions. The rate of impairments isrelated to gestational age at birth and severity of motor impair-ment, with those children born EP with CP having the highestpercentages of all impairments [18]. A recent meta-analysis of CPregisters reported on all children (term and preterm) with CP that75% (95% CI: 72e78) had pain, 49% (34e64) had an intellectualdisability, 16% (16e40) could not ambulate, 23% (19e27) were non-verbal, 28% (21e34) had a hip displacement, 26% (24e28) had abehaviour disorder, 24% (17e29) had epilepsy at some point, 24%(95% CI unable to assess as only one study) had a sleep disorder, 23%(95% CI unable to assess as only one study) had problems withcontinence, 22% (95% CI unable to assess as only one study) drib-bled, 11% (5e17) were blind, 6% (3e9) were tube-fed and 4% (2e6)were deaf [33].

DCD is considered a mild motor impairment by comparisonwith CP; however, the effect on gross motor coordination, bal-ance, and motor planning, as well as fine motor skills that un-derlie manipulation, self-care skills, and handwriting have long-term consequences [1,31]. The Avon Longitudinal Study of Par-ents and Children, a large population-based cohort of 6902 ofwhich 5% met the criteria for probable DCD (8.2% born �37weeks and 9.3% � 2500 g), reported that children with probableDCD had an increased risk of difficulties in attention (OR: 1.94;95% CI: 1.17e3.24), non-word repetition (1.83; 1.26e2.66), socialcommunication (1.87; 1.15e3.04), reading (3.35; 2.36e4.77),spelling (2.81; 2.03e3.90), self-reported depression (2.09; 1.36e3.19) and parent-reported mental health difficulties (4.23; 3.10e

5.77) compared with term-born peers at school age [34,35].Similar findings have been reported in studies including only EPor ELBW children, with motor impairment being associated withpoorer cognitive function and academic test scores, externalizingand adaptive behavior problems and attention deficit hyperac-tivity disorder in children without CP [29,36]. Not surprisingly,problems with motor skills are associated with poorer fitness[37], reduced physical activity [38] and obesity (especially inmales) [39].

5. Neural mechanisms

The neural mechanisms involved in CP are varied but alwaysinvolve an alteration to the developing brain, whereas themechanisms for other motor impairments in children bornpreterm are not clear. Brain injury during the perinatal period isthe most frequent cause of morbidity for preterm infants [5]. Theincidence of cystic periventricular leukomalacia (PVL), whichconsists of focal necrotic lesions evolving to cysts, is highlypredictive of CP, and has decreased in the past two decades,occurring in w3e4% of surviving VP infants [40,41]. On the otherhand, diffuse non-cystic central white matter abnormalities(WMA), also referred to as non-cystic PVL, which consists prin-cipally of magnetic resonance imaging (MRI) signal change andoften accompanied by ventricular dilation, occurs in 30e50% ofVP infants [42,43]. Severity of WMA in VP children has beenshown to correlate with later motor function, with rates of motorimpairment increasing with severity of WMA [44]. MRI is rec-ommended for all children with a suspected diagnosis of CP toestablish the likely aetiology, provide information about thetiming of the injury and exclude other diagnoses such as atumour, but is not required for a diagnosis. A systematic reviewof neuroimaging in children with CP found that 83% have someabnormal neuroradiological findings, with white matter abnor-malities the most common [45]. Combined gray and whitematter abnormalities are more common among children withhemiplegia, whereas isolated white matter abnormalities aremore common with bilateral spasticity e the most common formof CP in the preterm child. The understanding of neural mech-anisms of DCD is improving with the use of advanced MRItechniques, although research is still in the early stages andlimited to small groups. Compared with controls, children withDCD are reported to have significantly lower mean diffusivity ofthe posterior corticospinal tract and posterior thalamic radia-tions, which are associated with lower scores on a clinical test ofmotor abilities [29]. Children with DCD have also been shown tooveractivate specific brain regions during motor performancetasks and underactivate other regions during motor learningtasks [29].

6. Perinatal risk factors

The mechanisms involved in motor impairment are not fullyunderstood but are most likely to be multifactorial, involving ge-netic, biomedical, personal, and environmental factors [12,29].Whereas WMA is one of the most common risk factors for CP, otherrisk factors in the perinatal period for the preterm infant includelower gestational age, infection, multiple births, male gender, andpostnatal corticosteroids [12]. Studies exploring the relationshipbetween DCD in preterm children and perinatal risk factors haveshown variable results. White matter abnormality, male sex, post-natal corticosteroids, low birthweight and gestational age, pro-longed rupture of membranes and retinopathy of prematurity haveall been associated with motor impairment in preterm childrenwithout CP [1,44,46].

A.J. Spittle, J. Orton / Seminars in Fetal & Neonatal Medicine 19 (2014) 84e89 87

Magnesium sulphate given to mothers when preterm labour isimminent has been shown to have a neuroprotective role and re-duces the risk of CP in their child [relative risk (RR): 0.68; 95% CI:0.54e0.87] and a reduction in the rate of substantial gross motordysfunction (0.61; 0.44e0.85) [47]. Caffeine for the treatment ofapnoea of prematurity has also been found to reduce the incidenceof CP (adjusted OR: 0.58; 95% CI: 0.39e0.87), and at 5 years relatedto improvement in GMFCS classification and MABC scores [48,49].These two neuroprotective interventions, along with the reducedusage of postnatal corticosteroids to treat bronchopulmonarydysplasia, may explain the diminishing rates of CP reported bysome studies in the 2000s [21].

7. Recommendations for follow-up

The American Academy of Pediatrics has published recom-mendations for follow-up of VLBW infants, which include astructured age-appropriate neuromotor examination at leastonce during the first 6 months of life, followed by an examina-tion once during the second 6 months, at the ages of 1e2, 2e3,3e4 and 4e5 years [50]. Some of the motor impairments expe-rienced in children born preterm are apparent very early indevelopment, although follow-up throughout childhood isnecessary to accurately diagnose DCD and the milder forms of CP.Prediction of CP in the early months is now more reliable withthe use of assessment tools such as the general movements(GMs) and MRI which have high sensitivity and specificity[12,44,51]. Both abnormal GMs and MRI findings have beenassociated with milder motor impairments, but the association isnot as strong as that with CP [52,53]. Whereas neither of thesetools should be used for diagnosis of CP in the newborn period,care-givers can be counselled about risk factors, the importanceof ongoing assessment and the infant enrolled in early inter-vention [12,33].

The age at which the diagnosis of CP is made is variable, withmisdiagnosis common prior to the age of 18 months [51]. At 2years the GMFCS can provide accurate prognostic information onfuture independent function for children with CP [12]. The diag-nosis of DCD is best made from the age of 5 years for a number ofreasons according to the European Academy for ChildhoodDisability (EACD) consensus for definition and diagnosis of DCDincluding the ‘catch-up’ that younger children make in the earlyyears and the reliability of assessment tools for both motor anddaily living skills [24]. The EACD recommend using a score �15thpercentile on the MABC (or equivalent assessment) to classifychildren as DCD, whereas previous guidelines published in 2006recommended using a score <5th percentile on the MABC toclassify children as DCD. Children aged 3e4 years can be reliablydiagnosed if they perform below the 5th centile on the MABCaccording to the EACD [24].

8. Interventions to improve motor outcomes

The role of intervention is multidimensional and the focus variesas the child develops and any ongoing delays or disabilities areidentified. Given the high rates of comorbidities and associatedimpairments with both CP and DCD, a multidisciplinary approachto intervention is important. Both CP and delayed developmentmay present similarly in the early months, and delays in diagnosis,while doctors ‘wait and see’ what may evolve or other diagnosesare excluded, may present a missed opportunity for early inter-vention and add to parental distress [12,33]. In the early periodprior to any diagnosis of CP or DCD the primary role of interventionmay be preventive and aimed at ameliorating the effects of pre-maturity to provide the opportunity for the infants to reach their

optimal developmental outcome [54]. The evidence for effective-ness for early developmental interventions in the first year of lifefor preterm infants to improve motor outcomes is increasing withmore randomized controlled trials being published [55]. OurCochrane review on the effectiveness of these programs demon-strated a small improvement in motor outcomes up to 2 years ofage (motor scale developmental quotient: standardized mean dif-ference: 0.10 SD; 95% CI: 0.00e0.19) but this does not extend intoschool age [55].

A diagnosis of CP or DCD allows intervention to become morespecifically targeted and strategies are then aimed at prevention offurther delay and compensating for deficits with interventions topromote the best function and independence for the child [56]. Thediagnosis of CP or DCD may allow families access to funding forintervention that is not available without diagnosis of a disability.For children with CP, aspects of the child’s functional ability, motortype and topography become the focus of intervention in thecontext of other comorbidities and known risks such as hipdisplacement [12]. For children with DCD intervention is also rec-ommended in the context of other comorbidities [24]. A recentreview of therapies for children with DCD identified that the task-oriented approach, which includes cognitive approaches with afocus on specific aspects of a motor skill, was an effective way ofteaching motor skills in DCD [57]. Traditional physiotherapy andoccupational therapy were also effective, particularly when amotortraining approach was taken. Process-oriented therapy, which fo-cuses on aspects of body function such as muscle strength or sen-sory integration to improve a skill performance, was shown to beno more effective than no therapy and was not recommended forchildren with DCD [57].

9. Future directions

There is a need for future randomized controlled trials toinvestigate the most effective interventions for preterm childrenwith motor impairments earlier in development, along withresearch to identify early predictors of milder motor impairments.Without the consistent use of measures for the classification ofmotor impairment there is limited ability to compare outcomesbetween studies, slow progress in understanding causal pathwaysand risk factors, and limited development of intervention andmanagement programs for children with motor impairments bornpreterm [26].

10. Conclusions

With improvements in obstetric and neonatal care the rates ofsevere neurodevelopmental impairments have decreased sincethe 1990s; however, children born preterm remain at increasedrisk of CP compared with their peers born at term. There isincreasing recognition of mildemoderate motor impairments inchildren born preterm, consistent with motor skill problems seenin children with DCD in children born preterm. Whereas thesemotor impairments are considered milder than CP, the long-termeffects on academic achievement, participation in extracurricularactivities and mental health mean that this disorder cannot beattributed to a simple maturational delay, and in most cases it isnot outgrown. Perinatal and neonatal medicine have an impor-tant role not only in protecting the developing central nervoussystem at biomedical level with the use of pharmaceutical in-terventions including antenatal magnesium sulphate andcaffeine after birth, but also with long-term follow-up and sup-port for children with motor impairments and their families,including a multidisciplinary team of pediatricians and alliedhealth workers.

Research directions

� Consistent classification of motor impairment is neededin the reporting of long-term outcomes of children bornpreterm.

� Further understanding of the causal pathways and pre-dicting those children most at risk of motor impairmentare needed to inform interventions.

Practice points

� Motor impairment is common in children born pretermand can range in severity from mild to severe, includingDCD and CP.

� MRI and GMs assessments in the newborn period arepredictive of later motor impairment, including DCD andCP.

� DCD is highly prevalent in children born preterm andtherefore follow-up needs to continue to at least 5 years ofage.

A.J. Spittle, J. Orton / Seminars in Fetal & Neonatal Medicine 19 (2014) 84e8988

Conflict of interest

None declared.

Funding sources

None.

References

*[1] de Kieviet JF, Piek JP, Aarnoudse-Moens CS, Oosterlaan J. Motor developmentin very preterm and very low-birth-weight children from birth to adoles-cence: a meta-analysis. JAMA 2009;302:2235e42.

[2] Bracewell M, Marlow N. Patterns of motor disability in very preterm children.Ment Retard Dev Disabil Res Rev 2002;8:241e8.

[3] van Haastert IC, de Vries LS, Helders PJ, Jongmans MJ. Early gross motordevelopment of preterm infants according to the Alberta Infant Motor Scale.J Pediatr 2006;149:617e22.

[4] Fallang B, Hadders-Algra M. Postural behavior in children born preterm.Neural Plasticity 2005;12:175e82.

[5] Volpe JJ. Neurology of the newborn brain. 5th ed. Philadelphia: Saunders;2002.

[6] Cooke RW, Foulder-Hughes L. Growth impairment in the very preterm andcognitive and motor performance at 7 years. Arch Dis Child 2003;88:482e7.

[7] Saigal S, Doyle LW. An overview of mortality and sequelae of preterm birthfrom infancy to adulthood. Lancet 2008;371(9608):261e9.

[8] Brown NC, Doyle LW, Bear MJ, Inder TE. Alterations in neurobehavior at termreflect differing perinatal exposures in very preterm infants. Pediatrics2006;118:2461e71.

*[9] Williams J, Lee KJ, Anderson PJ. Prevalence of motor-skill impairment inpreterm children who do not develop cerebral palsy: a systematic review.Dev Med Child Neurol 2010;52:232e7.

*[10] Himpens E, Van den Broeck C, Oostra A, Calders P, Vanhaesebrouck P. Prev-alence, type, distribution, and severity of cerebral palsy in relation togestational age: a meta-analytic review. Dev Med Child Neurol 2008;50:334e40.

[11] Rosenbaum P. Classification of abnormal neurological outcome. Early HumDev 2006;82:167e71.

*[12] McIntyre S, Morgan C, Walker K, Novak I. Cerebral palsy e don’t delay. DevDisabil Res Rev 2011;17:114e29.

[13] Bax M, Goldstein M, Rosenbaum P, et al. Proposed definition and classifica-tion of cerebral palsy, April 2005. Dev Med Child Neurol 2005;47:571e6.

[14] Reid SM, Carlin JB, Reddihough DS. Rates of cerebral palsy in Victoria,Australia, 1970 to 2004: has there been a change? Dev Med Child Neurol2011;53:907e12.

[15] Vohr BR, Msall ME, Wilson D, Wright LL, McDonald S, Poole WK. Spectrum ofgross motor function in extremely low birth weight children with cerebralpalsy at 18 months of age. Pediatrics 2005;116:123e9.

[16] Ancel P, Livinec F, Larroque B, et al. Cerebral palsy among very pretermchildren in relation to gestational age and neonatal ultrasound abnormal-ities: the EPIPAGE Cohort Study. Pediatrics 2006;117:828e35.

[17] Surveillance of Cerebral Palsy in E. Surveillance of cerebral palsy in Europe: acollaboration of cerebral palsy surveys and registers. Surveillance of cerebralpalsy in Europe (SCPE). Dev Med Child Neurol 2000;42:816e24.

[18] Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi B. Develop-ment and reliability of a system to classify gross motor function in childrenwith cerebral palsy. Dev Med Child Neurol 1997;39:214e23.

[19] Himmelmann K, Beckung E, Hagberg G, Uvebrant P. Gross and fine motorfunction and accompanying impairments in cerebral palsy. Dev Med ChildNeurol 2006;48:417e23.

[20] Nordmark E, Hagglund G, Lagergren J. Cerebral palsy in southern Sweden II.Gross motor function and disabilities. Acta Paediatr 2001;90:1277e82.

[21] Victorian Infant Collaborative Study G. Outcomes at age 2 years of infants<28 weeks’ gestational age born in Victoria in 2005. J Pediatr 2010;156. 49e53 e41.

*[22] Edwards J, Berube M, Erlandson K, et al. Developmental coordination disor-der in school-aged children born very preterm and/or at very low birthweight: a systematic review. J Dev Behav Pediatr 2011;32:678e87.

[23] World Health Organization. Classification of mental and behavioural dis-orders: clinical descriptions and diagnostic guidelines. Geneva: WHO;1992.

*[24] European Academy for Childhood D. European Academy for ChildhoodDisability (EACD): recommendations on the definition, diagnosis and inter-vention of developmental coordination disorder (long version). Dev MedChild Neurol 2012;54:54e93.

[25] Zhu JL, Olsen J, Olesen AW. Risk for developmental coordination disordercorrelates with gestational age at birth. Paediatr Perinat Epidemiol 2012;26:572e7.

[26] Magalhaes LC, Missiuna C, Wong S. Terminology used in research reports ofdevelopmental coordination disorder. DevMed ChildNeurol 2006;48:937e41.

[27] Kirby A, Sugden DA. Children with developmental coordination disorders. J RSoc Med 2007;100:182e6.

*[28] Zwicker JG, Missiuna C, Harris SR, Boyd LA. Developmental coordinationdisorder: a review and update. Eur J Paediatr Neurol 2012;16:573e81.

[29] Victorian Infant Collaborative Study G. Developmental coordination disorderat 8 years of age in a regional cohort of extremely-low-birthweight or verypreterm infants. Dev Med Child Neurol 2007;49:325e30.

[30] Majnemer A. Motor incoordination in children born preterm: coordinatedefforts needed. Dev Med Child Neurol 2007;49:324.

[31] Msall ME. Optimizing neuromotor outcomes among very preterm, very low-birth-weight infants. JAMA 2009;302:2257e8.

[32] Zwicker JG, Harris SR, Klassen AF. Quality of life domains affected in childrenwith developmental coordination disorder: a systematic review. Child CareHealth Dev 2013;39:562e80.

[33] Novak I, Hines M, Goldsmith S, Barclay R. Clinical prognostic messages from asystematic review on cerebral palsy. Pediatrics 2012;130:e1285e312.

[34] Lingam R, Jongmans MJ, Ellis M, Hunt LP, Golding J, Emond A. Mental healthdifficulties in children with developmental coordination disorder. Pediatrics2012;129:e882e91.

[35] Lingam R, Golding J, Jongmans MJ, Hunt LP, Ellis M, Emond A. The associationbetween developmental coordination disorder and other developmentaltraits. Pediatrics 2010;126:e1109e18.

[36] Marlow N, Hennessy EM, Bracewell MA, Wolke D, Group EPS. Motor andexecutive function at 6 years of age after extremely preterm birth. Pediatrics2007;120:793e804.

[37] Burns YR, Danks M, O’Callaghan MJ, et al. Motor coordination difficulties andphysical fitness of extremely-low-birthweight children. Dev Med ChildNeurol 2009;51:136e42.

[38] Cairney J, Hay JA, Faught BE, Wade TJ, Corna L, Flouris A. Developmentalcoordination disorder, generalized self-efficacy toward physical activity,and participation in organized and free play activities. J Pediatr 2005;147:515e20.

[39] Cairney J, Hay JA, Faught BE, Hawes R. Developmental coordination disorderand overweight and obesity in children aged 9e14 y. Int J Obes 2005;29:369e72.

[40] Inder TE, Wells SJ, Mogridge NB, Spencer C, Volpe JJ. Defining the nature ofthe cerebral abnormalities in the premature infant: a qualitative magneticresonance imaging study. J Pediatr 2003;143:171e9.

[41] Volpe JJ. Cerebral white matter injury of the premature infant e morecommon than you think. Pediatrics 2003;112:176e80.

[42] Woodward LJ, Anderson PJ, Austin NC, Howard K, Inder TE. Neonatal MRI topredict neurodevelopmental outcomes in preterm infants. N Engl J Med2006;355:685e94.

[43] Pavlova MA, Krageloh-Mann I. Limitations on the developing preterm brain:impact of periventricular white matter lesions on brain connectivity andcognition. Brain 2013;136(Pt 4):998e1011.

*[44] Spittle AJ, Boyd RN, Inder TE, Doyle LW. Predicting motor development invery preterm infants at 12 months’ corrected age: the role of qualitativemagnetic resonance imaging and general movements assessments. Pediatrics2009;123:512e7.

[45] Korzeniewski SJ, Birbeck G, DeLano MC, Potchen MJ, Paneth N.A systematic review of neuroimaging for cerebral palsy. J Child Neurol2008;23:216e27.

[46] Zwicker JG, Yoon SW, Mackay M, Petrie-Thomas J, Rogers M, Synnes AR.Perinatal and neonatal predictors of developmental coordination disorder invery low birthweight children. Arch Dis Child 2013;98:118e22.

A.J. Spittle, J. Orton / Seminars in Fetal & Neonatal Medicine 19 (2014) 84e89 89

[47] Doyle LW, Crowther CA, Middleton P, Marret S, Rouse D. Magnesium sulphatefor women at risk of preterm birth for neuroprotection of the fetus. CochraneDatabase Syst Rev 2009;(1):CD004661.

[48] Schmidt B, Anderson PJ, Doyle LW, et al. Survival without disability to age 5years after neonatal caffeine therapy for apnea of prematurity. JAMA2012;307:275e82.

[49] Schmidt B, Roberts RS, Davis P, et al. Long-term effects of caffeine therapy forapnea of prematurity. N Engl J Med 2007;357:1893e902.

*[50] Wang CJ, McGlynn EA, Brook RH, et al. Quality-of-care indicators for theneurodevelopmental follow-up of very low birth weight children: results ofan expert panel process. Pediatrics 2006;117:2080e92.

[51] BosanquetM, Copeland L,Ware R, Boyd R. A systematic reviewof tests to predictcerebral palsy in young children. Dev Med Child Neurol 2013;55:418e26.

[52] Spittle AJ, Spencer-Smith MM, Cheong JL, et al. General movements in verypreterm children and neurodevelopment at 2 and 4 years. Pediatrics2013;132:e452e8.

[53] Spittle AJ, Cheong J, Doyle LW, et al. Neonatal white matter abnormalitypredicts childhood motor impairment in very preterm children. Dev MedChild Neurol 2011;53:1000e6.

[54] Hadders-Algra M. Challenges and limitations in early intervention. Dev MedChild Neurol 2011;53(Suppl 4):52e5.

[55] Spittle A, Orton J, Anderson P, Boyd R, Doyle LW. Early developmentalintervention programmes post-hospital discharge to prevent motor andcognitive impairments in preterm infants. Cochrane Database Syst Rev2012;(12):CD005495.

[56] Majnemer A. Benefits of early intervention for children with developmentaldisabilities. Semin Pediatr Neurol 1998;5:62e9.

[57] Smits-Engelsman BC, Blank R, van der Kaay AC, et al. Efficacy of interventionsto improve motor performance in children with developmental coordinationdisorder: a combined systematic review and meta-analysis. Dev Med ChildNeurol 2013;55:229e37.

lable at ScienceDirect

Seminars in Fetal & Neonatal Medicine 19 (2014) 90e96

Contents lists avai

Seminars in Fetal & Neonatal Medicine

journal homepage: www.elsevier .com/locate/s iny

Review

Neuropsychological outcomes of children born very preterm

Peter J. Anderson a,b,*

aClinical Sciences, Murdoch Childrens Research Institute, Melbourne, AustraliabDepartment of Paediatrics, The University of Melbourne, Melbourne, Australia

Keywords:AttentionCognitionDevelopmentNeuropsychologyVery low birth weightVery preterm

* Address: Victorian Infant Brain Studies, MurdochRoyal Children’s Hospital, Flemington Road, Parkville,3 99366704.

E-mail address: peter.anderson@mcri.edu.au.

1744-165X/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.siny.2013.11.012

s u m m a r y

Considerable research has investigated the consequences of being born very preterm (VP; <32 weeks ofgestation), especially in relation to cognitive functioning. While numerous cognitive and neuropsycho-logical outcome studies have been published, it is important to consider methodological issues whenreviewing this research, as the generalizability of the studies varies greatly. This article describes thenature of cognitive difficulties confronting VP children, both in terms of the frequency and severity ofdeficits. The breadth of cognitive difficulties reported in this population implies a generalized cognitiveimpairment; however, the presence of selective or primary cognitive deficits is discussed. It is concludedthat whereas mortality and neonatal morbidity rates have decreased significantly in VP infants in recentdecades, these children continue to be at significant risk for cognitive impairments and need to be closelymonitored throughout childhood.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

This review article describes the nature of the cognitive im-pairments exhibited by very preterm (VP) children. As well asexamining general outcomes such as IQ and academic achieve-ment, functioning in specific cognitive domains are reviewed,including processing speed, attention, visualespatial abilities, lan-guage, memory and learning, and executive function.

2. Methodological considerations

The quality of VP outcomes studies varies greatly and studydesign requires careful considerationwhen reviewing the cognitiveoutcome literature. Before the 1990s, the inclusion criteria foroutcome studies tended to be based on birth weight [e.g. very lowbirth weight (VLBW),<1500 g; extremely low birth weight (ELBW),<1000 g] rather than on gestational age due to the lack of certaintyof obstetric estimation. Whereas birth weight and gestational ageare related, they are not interchangeable measures, with birthweight-selected cohorts having a proportion of children born laterin gestation who are small for gestational age (SGA). Selectioncriteria for studies may include the entire VP population (<32weeks of gestation) or may restrict selection to a subgroup such as

Childrens Research Institute,VIC 3052, Australia. Tel.: þ61

All rights reserved.

extremely low gestational age infants (ELGA; <26 weeks of gesta-tion). In addition to expecting the severity of deficits to be greaterwith increasing immaturity, it is reasonable to speculate that theELGA infants exhibit a unique profile of cognitive impairments. Avariety of exclusion criteria are implemented by studies, some ofwhich can influence the generalizability of the findings. Forexample, excluding children with cranial ultrasound abnormality[e.g. grade 3/4 intraventricular haemorrhage (IVH), cystic periven-tricular leukomalacia (PVL)] or with IQ <80 may be appropriate forspecific research questions, but the findings are likely to underes-timate the population’s true level of impairment.

Marked improvement in survival rates was observed betweenthe 1970s and the 2000s with advancements in perinatal medicine[1]. This reduction inmortality has been predominantly observed inthe most immature, tiniest and sickest infants [1], which are alsothose infants considered at highest risk for later cognitive prob-lems. Accordingly, caution is needed when comparing outcomesacross eras. For instance, although it may be disheartening to seelittle change in cognitive outcomes for VP children born in the mid-2000s compared with those born in the early 1990s, one may arguethat this really represents an improvement in outcome since themid-2000s population includes a higher proportion of high-riskinfants.

Sample composition is another important consideration whenreviewing the VP literature. Prospective geographic cohorts whichrecruit all infants born in a specific region is the preferred approach,although studies reporting regional samples are rare and tend tofocus on broader cognitive and academic outcomes [2,3]. Hospital-based or network cohorts are regularly reported, but are less

P.J. Anderson / Seminars in Fetal & Neonatal Medicine 19 (2014) 90e96 91

representative due to possible ascertainment bias and local man-agement policies (Smith et al., this issue, Chapter 1). Samplesrecruited using other strategies such as attendees to follow-upclinics have questionable generalizability. Participation rate andfollow-up attrition are also important as non-participation is notrandom but is more common in socially disadvantaged families andfor children with impairments [4]. Therefore, studies with lowparticipation rates are likely to under-report the severity ofcognitive deficits in VP children.

VP outcome studies need a baseline to judge the level of func-tioning. Relying on test norms is less than ideal as the socio-demographic characteristics of the standardization sample willdiffer from that of the VP cohort. Compounding this issue is theFlynn effect (phenomena in which normative mean scores forcognitive measures such as IQ increase with time) [5], which canmask to some extent the severity of impairments. Numerousstudies have demonstrated that the rate of impairment is greatlyunderestimated if judged according to test norms rather than alocally matched control group [6e8]. Whereas numerous ap-proaches have been used to recruit a local control group, a matchedcomparison group is preferable.

Early outcome studies tended to focus on IQ. Over the past 20years more studies have applied a neuropsychological approach inorder to better characterize the cognitive profiles of VP children,which is critical for informing surveillance and intervention pro-grams. Neuropsychological studies have tended to focus on a spe-cific cognitive domain (e.g. attention), and in many cases haverelied on hospital-based or follow-up clinic samples. Prospectivelongitudinal neuropsychological studies are lacking.

3. General cognitive ability

The foundation of neuropsychological assessments is anassessment of general cognitive ability (IQ). Numerous IQ measuresare reported in the literature, and in recent times there has been amovement towards abbreviated measures, which allows additionaltime to evaluate specific cognitive domains. However, caution isneeded when interpreting IQ scores from abbreviated measures asthey are based on fewer tasks and assess fewer abilities. Even morecaution is needed when abbreviated scales are based on a personalselection of subtests from established batteries as these compositescores lack validation.

Systematic reviews have been performed evaluating the IQ ofpreterm children (<37 weeks of gestation). The first review pooledthe results of 15 caseecontrol studies, including 1556 preterm and1720 term children born between 1975 and 1988 [9]. The meangroup difference for these studies ranged from 7 to 23 points infavour of the term controls, with a mean difference of 10.9 points[95% confidence interval (CI): 9.2, 12.5]. To put this finding incontext, the preterm population had an IQ of 0.7 SD below theirterm peers. Mean cognitive scores were significantly related togestational age (R2 ¼ 0.49), but the association with age at assess-ment was weak, suggesting that this level of deficit remainedrelatively stable across childhood. As expected, the group differencein IQ for high-quality studies was marginally higher than that forlow-quality studies (11.2 vs 9.4).

Given that survival rates and management practices haveimproved over the past 30 years, it is possible that more contem-porary preterm children exhibit better outcome in terms of IQ thanprevious generations. To address this issue, an updated meta-analysis was recently reported which included caseecontrolstudies published between 1980 and 2009 [10]. This review iden-tified 27 eligible studies including 3504 preterm and 3540 termchildren born between 1975 and 2000. The meta-analysis revealeda mean difference of 11.9 points (95% CI: 10.5, 13.4), with the

preterm children performing 0.8 SD below term controls on mea-sures of IQ. As with the Bhutta et al. study [9], IQ was associatedwith gestational age. For example, the mean difference was 8.4(95% CI: 6.6, 10.2) for children with a mean gestational age �32weeks, 11.4 (95% CI: 9.7, 13.2) for children with a mean gestationalage between 28 and 31 weeks, and 13.9 (95% CI: 11.5, 16.2) forchildren with gestational age <28 weeks. There was no associationwith year of birth, suggesting no gain in IQ for preterm children inthis 25-year period.

In summary, there is convincing evidence that IQ is reduced inpreterm children, and there is no evidence to indicate that this hasimproved in more contemporary eras or that the gap with termpeers reduces with increasing age. There is evidence that impair-ment severity increases with decreasing gestational age, such thatIQ is estimated to decrease 1.5 points per week for those born <33weeks [11]. Whereas IQ scales provide a reliable assessment ofgeneral cognitive ability, they are not ideal for detecting mild def-icits, specific cognitive strengths and weaknesses, or subtle brainabnormalities [12]. Specific neuropsychological measures areneeded for these purposes.

4. Processing speed

Processing speed refers to the time required to interpret andrespond to incoming stimuli or information, and is assessed bymeasures of reaction time and decision time. Processing speed is an‘elementary’ cognitive process [13], as it is critical to the func-tioning of other cognitive domains. Processing speed developsrapidly in childhood [14], and its developmental trajectory mimicsthat of working memory and fluid intelligence, leading to specu-lation that increasing efficiency in information processing is asso-ciated with enhanced working memory and intelligence [15].

Slower processing speed has been reported in VP/VLBW chil-dren. Rose et al. [16] studied processing speed in preterm infants(<1750 g birth weight) in their first year of life and reported thatthey needed nearly 30% more inspection time than term infants at5, 7 and 12months of age. In middle childhood VLBW children havebeen found to perform similarly to term children on simple reactiontime tasks, but their decision time slowed more steeply than con-trols with increasing complexity on choice reaction time tasks [13].These findings are supported by a Dutch study of VP 7-year-olds[17], although they found that slower processing speedwas due to agreater proportion of extremely slow responses and not related tolower average processing speed. Thus, it may be speculated that VPchildren, although capable of exhibiting age-appropriate responsetimes, have difficulty maintaining a high level of efficiency whenthe complexity of the task increases. Whereas no long-term lon-gitudinal studies have been reported, it seems that reduced pro-cessing efficiency persists into adulthood [18,19].

5. Attention

Attention is another core cognitive ability, critical for theacquisition of new skills and knowledge [20]. Attention is amultifaceted construct [21,22], consisting of the capacity to selec-tively focus (i.e. focus on relevant stimuli and ignore distractingstimuli), sustain (i.e. maintain alert state for extended period),encode (i.e. hold information in temporary store), shift (i.e. fluentlytransfer focus from one activity to another), and divide attention(i.e. focus on multiple competing stimuli simultaneously) [20,23].

Numerous studies have investigated attention domains in pre-term infants and preschoolers, as highlighted in a review by Van deWeijer-Bergsma et al. [24]. The review found that all attentiondomains are delayed in young preterm children compared withterm controls, and that these differences tend to increase with

P.J. Anderson / Seminars in Fetal & Neonatal Medicine 19 (2014) 90e9692

increasing age. In addition to this review, Mulder et al. [25] per-formed meta-analyses exploring differences in selective, sustainedand shifting attention domains between preterm and term childrenaged �2 years. The selective attention analysis, which involved1196 preterm and 1008 term children, found that preterm childrenperformed 0.4 SD below term peers. As expected, this deficit wasgreater (0.6 SD) for those born <26 weeks of gestational age. Morevariability was observed for sustained attention, but a moderateeffect size of 0.5 SDwas still revealed by themeta-analysis in favourof term children, and increased to 0.7 SD when restricted to chil-dren born<26weeks of gestation. Conflicting results were found inthe meta-analysis for shifting attention, with a marked deficit seenin preterm children on the Trails B task (0.5 SD), but no groupdifference was observed on sorting tasks.

More recently some large-scale studies incorporating acomprehensive assessment of attention have been reported[20,23]. These studies enable the frequency and severity ofimpairment to be contrasted across different attention domains inthe same cohort. The study by Anderson et al. [20] found a gener-alized attention impairment in a large geographic cohort of EP/ELBW 8-year-olds, who performed significantly below matchedterm controls in selective, sustained, encoding, shifting and dividedattention, with the effect size varying from 0.3 to 0.6 SD. Rates ofimpairment were relatively high, and in contrast to term peers EP/ELBW childrenwere 2.4 times more likely to have an impairment inselective and sustained attention, and more than 3 times morelikely to have an impairment on shifting and divided attention.Using a similar methodological approach but with EP/ELBW ado-lescents, Wilson-Ching et al. [23] reported deficits in selective,shifting and divided attention, but performance on a measure ofsustained attention was similar to matched term controls. The EP/ELBW adolescents were more likely than their term peers to havean impairment in selective (36% vs 14%), shifting (41% vs 17%) anddivided attention (15% vs 8%). Without doubt, attention difficultiesare one of the most common issues for VP children.

6. Visual and perceptual skills

Visual sensory, perceptual and motor skills are important foracademic functioning, social interactions and other everyday ac-tivities [26]. Thought to be related to retinopathy of prematurity orperiventricular white matter injury, an elevated rate of visualproblems is often reported in preterm children, including impairedvisual acuity and contrast sensitivity, and strabismus [27e29]. Vi-sual perceptual difficulties have also been reported in VP children[28]. In a recent study of visual outcomes in a large geographiccohort of EP/ELBW adolescents [26], 43% had at least one visualsensory impairment in contrast to 29% of term controls, withsignificantly elevated rates of impairments in stereopsis (26% vs10%) and convergence (16% vs 6%). In relation to visual perceptualskills, this EP/ELBW adolescent cohort performed significantlybelow term peers on all five measures administered, and was 3times more likely to have a visual perceptual impairment [26]. TheBeery Test of Visual Motor Integration (VMI) is often included in VPoutcome studies. This measure involves copying increasinglycomplex geometric designs, and dependent on the integration ofvisual, perceptual, and fine motor abilities. VP/VLBW childrenconsistently perform below controls on the VMI [30], with onestudy reporting that 30% of VLBW children with normal cranialultrasounds scored below the 15th percentile [31].

7. Memory and learning

Memory is a complex systemwhich is fundamental for everydayfunctioning. New information enters immediate memory, and,

while in this temporary storage system, can be processed ormanipulated e a capacity referred to as working memory. Infor-mation in temporary storage can be transferred to long-termmemory. Implicit (procedural) and explicit (declarative) memoryare two forms of long-term memory, with the majority of outcomestudies focusing on explicit memory ability. Explicit memory can befurther divided into episodic memory e the ability to recall per-sonal events and experiences e or semantic memory e the abilityto remember factual information and knowledge.

VP/VLBW children have been reported to have deficits in allmemory domains. Explicit memory has been tested in infants usingelicited imitated tasks, with studies demonstrating poorer recall ofactions after a 15-min delay in children from 12 to 36months of age[32,33]. Immediate/working memory has attracted the mostattention, with preterm samples nearly always scoring lower thanterm controls, with significant deficits reported in both verbal[2,34,35] and visual [36e38] modalities. There is some speculationthat severity of immediate/working memory deficits lessens withincreasing age. For example, studies have found marginal or nodifference in working memory between VP and control groups inadolescence [39,40], with one follow-up study reporting possibledevelopmental catch-up [41]. Studies have also investigatedlearning ability in VP children, such as remembering information orstimuli when given multiple presentations. Using a spatial locationmemory task, Baron et al. [38] found that ELBW 3-year-oldsrecalled fewer locations than term controls on the first presenta-tion, but the ELBW group failed to display the equivalent learningtrajectory when additional presentations of the stimuli were pro-vided and fell further behind. A similar pattern has been reportedby Taylor et al. in middle childhood [42] and adolescence [43] usinga verbal learning task.

A comprehensive study of memory in VP children was recentlypublished by Omizzolo et al. [44], who assessed a large contem-porary cohort of 7-year-olds using visual and verbal measures ofimmediate memory, working memory, and learning ability.Consistent with earlier reports, the VP group performed signifi-cantly below term controls in all memory domains. The VP groupwas 2.1 to 3.5 times more likely than the term controls to have amemory impairment, with the rate of impairment varying from19% to 41% in the VP group and only 10e18% in the controls. Insummary, evidence demonstrating generalized memory andlearning deficits in VP children is accumulating. Whereas there issome indication that the severity of memory deficits lessens inadolescence, memory concerns persist on measures of everydaymemory [45].

8. Language

Language and communications skills are critical for interper-sonal relationships and social interactions, and are highly corre-lated with academic achievement [46]. Whereas language iscommonly divided into expressive and receptive skills, a compre-hensive evaluation assesses semantics (meaning of words andsentences), grammar (sentence structure), phonological awareness(speech sounds), discourse (understanding passages of text orconversation) and pragmatics (use of language in social contexts)[46]. Vohr provides a detailed review of these outcomes in VPchildren in this issue (Chapter 2), and highlights two recent meta-analyses in this area [46,47]. van Noort-van der Spek et al. [47]examined simple and complex language outcomes in pretermchildren, finding that term children outperformed preterm childrenon simple language measures by 0.5 SD (95% CI: 0.3, 0.6) and oncomplex language measures by 0.6 SD (95% CI: 0.4, 0.8). Whereasthe group difference was not associated with assessment age forsimple language outcomes, the group difference increased with

P.J. Anderson / Seminars in Fetal & Neonatal Medicine 19 (2014) 90e96 93

increasing age for complex language outcomes [47]. A review byBarre et al. [46] was restricted to VP/VLBW children and attemptedto examine language domains more specifically. Whereas noeligible studies were identified for phonological awareness,discourse or pragmatics domains, the group difference in favor ofterm controls was significant for expressive language, receptivelanguage, and semantics. These meta-analyses imply a generalizedlanguage impairment in the VP population, a finding consistentwith a comprehensive study of language functioning recentlypublished which reported significant deficits in phonologicalawareness, semantics, grammar, discourse and pragmatics [48].

9. Executive function

Executive functioning refers to a set of inter-related cognitiveskills needed for purposeful, goal-directed behavior [49].Numerous skills have been linked to executive functioning,including (a) anticipation and deployment of attention, (b) impulsecontrol and self-regulation, (c) initiation of activity, (d) workingmemory, (e) mental flexibility and utilization of feedback, (f)planning ability and organization, (g) selection of efficientproblem-solving strategies, and (h) monitoring of performance[49]. These high-level cognitive skills are clearly important foradaptive functioning such as academic performance, everydaybehavior and social interactions. Executive functioning in VP chil-dren has generated considerable interest, with numerous studiesreporting this cognitive domain to be an area of concern for thispopulation [18,34,50,51].

Twometa-analyses examining executive functioning in pretermchildren have been reported [25,35]. Mulder et al. [25] performedseparate meta-analyses for response inhibition, verbal fluency,planning ability, and shifting. For response inhibition the meta-analysis focused on tasks that adopted a GoeNo Go paradigm,pooling the results of 830 preterm and 740 control children. Thegroup difference was 0.3 SD (95% CI: 0.0, 0.5) in favour of controls;however, a significant positive correlation was found with GA(R2 ¼ 0.74) with the group difference reaching 0.5 SD (95% CI: 0.1,0.9) when restricted to studies with an average gestational age atbirth of <26 weeks. Separate meta-analyses were performed forsemantic and phonemic fluency, although both found group dif-ferences of around 0.5 SD. Interestingly, the group difference inphonemic fluency increased with increasing chronological age(R2 ¼ 0.79), suggesting that deficits in this area may become worsewith increasing age. The planning ability meta-analysis revealed agroup difference in favour of term controls of 0.4 SD (95% CI: 0.1,0.7), largely influenced by studies whereby the preterm group had amean gestational age of <26 weeks (0.7 SD; 95% CI: 0.5, 0.9). Themeta-analysis for shifting attention revealed a significant deficit inthe preterm group on the Trails B task (0.5 SD), but no group dif-ference on sorting tasks.

Whereas the systematic review by Mulder et al. focused onstudies of preterm children published between 1990 and 2008,Aarnoudse-Moens et al. [35] restricted their systematic review tostudies of VP/VLBW published between 1998 and 2008 and per-formed meta-analyses for verbal fluency, working memory andcognitive flexibility. Five studies were eligible for the verbal fluencymeta-analysis, revealing a group difference of 0.6 SD (95% CI: 0.3,0.8) in favor of term controls. In contrast toMulder et al. [25], verbalfluency was not associated with chronological age. The workingmemory meta-analysis (studies that reported on Digit Span) had agroup difference of 0.4 SD (95% CI: 0.2, 0.5) and the cognitive flexi-bility meta-analysis (studies that reported on Trails B) had a groupdifference of 0.5 SD (95% CI: 0.3, 0.7) in favour of term controls.

In summary, there is considerable evidence to suggest that VPchildren are at increased risk for executive function impairments.

There are few long-term longitudinal studies of executive function,so it is not yet certain how much these deficits are related todevelopmental lag. However, poorer executive functioning hasbeen reported in adolescence [43] and adulthood [18], suggestingthat VP children do not grow out of these problems. It should benoted that most measures used to assess executive function sufferfrom task impurity as they are dependent on multiple executiveand non-executive abilities [49]. Thus, caution is required wheninterpreting poor performance on these measures as it could relateto difficulties in other domains.

10. Educational outcomes

Research consistently demonstrates that VP children performpoorer on standardized measures of academic attainment [2,52e54]. For example, a large geographic cohort of EP/ELBW 8-year-olds performed 0.5e0.6 SD belowmatched term controls on tests ofreading, spelling and mathematics [2]. The level of academic un-derachievement increases with decreasing gestational age [35,55],and at 11 years of age the EPICure cohort (<26 weeks of gestationalage) performed below term classmates by 1.0 SD for reading and1.7 SD for mathematics, even after excluding children with severecognitive impairment [52]. In the EPICure cohort, the rate ofreading impairment was 52%, compared with 11% for classmates,whereas the rate of mathematics impairment was nearly 70%compared with only 14% of classmates [52]. Meta-analyses confirmthe significant disadvantage of VP children at school, with pooledeffect sizes of 0.5 SD for reading, 0.8 SD for spelling and 0.6 SD formathematics [35].

Some studies have specifically focused on rates of learningdisability by excluding children with below-average IQ andneurosensory impairment. Applying the low-achievement criteriafor learning disability, Litt et al. [56] reported that children withbirth weights <750 g were seven times more likely to have areading learning disability, 6 times more likely to have a mathslearning disability, and 13 times more likely to have a combinedreading/maths disability than term controls. Similarly, Grunau et al.[57] found that 65% of ELBW children aged 8e9 years had a learningdisability in at least one area compared with only 13% of termcontrols. In line with these findings, VP children are more likely torepeat a school grade, to attend a non-mainstream school forchildren with special needs, and to require special educationintervention [52,54,56,58,59].

Acknowledging that VP children have more difficulties attainingbasic educational skills, intervening prior to school is recom-mended, before the problems emerge and become harder toremediate. Underlining the importance for early intervention, VPchildren are reported to be less school ready [60] and display lessmature early linguistics and numerical reasoning skills in the pre-school period [60]. In general, the rates of VP survivors to completehigh school and continue on to college/university are lower thanthat of term peers, although the majority of VP individuals com-plete secondary education [59,61,62].

11. Selective versus generalized cognitive impairment

The VP population is vulnerable to impairments in all cognitivedomains, which would imply a generalized cognitive impairment[2,54]; however, the presence of selective or primary impairmentsis still possible. For example, working memory, attention and pro-cessing speed are core cognitive skills, and deficits in these domainsmay at least partly explain difficulties in higher-order learning,language and executive function domains.

Determining the profile of cognitive vulnerabilities in VP chil-dren has clinical implications, as this knowledge will influence the

Practice points

� Very preterm children are at increased risk for impair-ments across all cognitive domains.

� Cognitive functioning is related to gestational age, withimpairment severity increasing with decreasing maturityat birth.

� The cognitive deficits observed in very preterm childrenare unlikely to reflect developmental delay alone, as im-pairments are also observed in very preterm adolescentsand adults.

Future research directions

� Research is needed to investigate the presence of selec-tive deficits in core cognitive domains such as processingspeed, working memory and attention in very pretermchildren.

� Long-term longitudinal studies are needed to understandhow the nature and severity of cognitive impairmentschanges in very preterm children with development.

� Understanding the mechanisms (risk and resilience fac-tors) associated with cognitive impairments in very pre-term children will inform preventative and earlyintervention programs.

P.J. Anderson / Seminars in Fetal & Neonatal Medicine 19 (2014) 90e9694

focus of surveillance programs and the content of interventionstrategies. However, it is methodologically difficult to untangle theindividual contribution of specific cognitive abilities and to assessthe presence of primary deficits [63]. One approach is to statisti-cally control for the performance on one measure purported tomeasure a distinct cognitive skill (e.g. processing speed) whenexamining level of impairment on another cognitive outcome (e.g.executive function). However, cognitive tasks are not pure mea-sures of a specific cognitive ability, and in most cases tap multipleabilities. Even simple processing speed tasks can be dependent onselective attention, workingmemory andmotor control. Taylor [63]recommends examining differential effects when assessing selec-tive vulnerabilities, which refers to contrasting the magnitude ofimpairment across different tasks that measure relatively distinctskills. This approach also suffers from the imprecise nature ofcognitive tasks in assessing specific cognitive skills, but also doesnot take into account the enormous variability in sensitivity andspecificity across cognitive tasks.

Mulder and colleagues have been particularly interested inidentifying the underlying reasons for underperformance in aca-demic attainment, attention and executive function observed in VPchildren [64e66]. They have reported that verbal processing speedand working memory are significant and independent predictors ofacademic attainment in VP children [64]. In a separate analysis, thisgroup also demonstrated that processing speed mediated the sig-nificant effect of VP birth on selective tests of executive function,including inhibition, working memory, shifting and semanticfluency [65]. Finally, Mulder et al. [66] have reported that symp-toms of impulsivity/hyperactivity and inattention, as noted byparents, correlate significantly with an objective measure of verbalprocessing speed in VP children but not term children. Othergroups have also attempted to isolate the effect of differentcognitive skills to functional outcomes [67].

In summary, deficits in core cognitive domains such as pro-cessing speed, working memory and attention will affect higher-level cognitive domains. While it is quite possible that VP chil-dren have a specific vulnerability in these core cognitive domains,representing a possible VP cognitive phenotype, more evidence isneeded.

12. Summary

The breadth of cognitive difficulties highlighted in this reviewpaints a relatively grim picture for families and health professionalsof VP children. However, in reality, the majority of VP children haverelatively mild impairments or no problems at all and go on to livevery productive lives [62]. Whereas numerous risk and resiliencefactors have been associated with long-term impairments in VPsurvivors, it is clear that outcome is ultimately related to an inter-play of genetic, medical, social and environmental factors. Under-standing the nature of cognitive weaknesses in the VP population,as discussed in this review, will inform intervention and remedia-tion programs, but an understanding of risk and resilience factorshas the potential to prevent cognitive deficits occurring in the firstplace.

In conclusion, VP survivors continue to be at risk of generalizedcognitive and academic impairments and require close surveillancethroughout development. While it is still unclear whether the na-ture and severity of cognitive impairments persists, worsens orimproves with age, enough long-term studies have been reportedto suggest that most cognitive vulnerabilities observed in child-hood are also present in adulthood. Early intervention programs forVP children are effective [68], but traditional modes of delivery ofthese programs are costly and inaccessible to many families.Cognitive training and other forms of intervention have also been

trialled with VP children [69,70], although it is unlikely that anysingle program will resolve the breadth of challenges confrontingthe VP population.

Conflict of interest statement

None declared.

Funding sources

This work was supported by the Australian National Health andMedical Research Council (Senior Research Fellowship ID 628371 toPA), and the Victorian Government Operational InfrastructureSupport Program.

References

[1] Saigal S, Doyle LW. Preterm birth 3 e an overview of mortality and sequelaeof preterm birth from infancy to adulthood. Lancet 2008;371:261e9.

*[2] Hutchinson EA, De Luca CR, Doyle LW, Roberts G, Anderson PJ, the VictorianInfant Collaborative Study Group. School-age outcomes of extremely pretermor extremely low birth weight children. Pediatrics 2013;131:E1053e61.

[3] Marret S, Marchand-Martin L, Picaud JC, Hascoet JM, Arnaud C, Roze JC, et al.Brain injury in very preterm children and neurosensory and cognitive dis-abilities during childhood: the EPIPAGE cohort study. PloS One 2013;8(5):e62683.

[4] CallananC,DoyleLW,RickardsAL,KellyEA, FordGW,DavisNM.Childrenfollowedwith difficulty: how do they differ? J Paediatr Child Health 2001;37:152e6.

[5] Flynn JR. Searching for justice e the discovery of IQ gains over time. AmPsychol 1999;54:5e20.

[6] Anderson PJ, De Luca CR, Hutchinson E, Roberts G, Doyle LW, the VictorianInfant Collaborative Study Group. Underestimation of developmental delayby the new BayleyeIII Scale. Arch Pediatr Adolesc Med 2010;164:352e6.

[7] Marlow N, Wolke D, Bracewell MA, Samara M, the EPICure Study Group.Neurologic and developmental disability at six years of age after extremelypreterm birth. N Engl J Med 2005;352:9e19.

[8] Wolke D, Ratschinski G, Ohrt B, Riegel K. The cognitive outcome of verypreterm infants may be poorer than often reported e an empirical

P.J. Anderson / Seminars in Fetal & Neonatal Medicine 19 (2014) 90e96 95

investigation of how methodological issues make a big difference. Eur JPediatr 1994;153:906e15.

*[9] Bhutta AT, Cleves MA, Casey PH, Cradock MM, Anand KJS. Cognitive andbehavioral outcomes of school-aged children who were born preterm e ameta-analysis. J Am Med Assoc 2002;288:728e37.

[10] Kerr-Wilson CO, Mackay DF, Smith GCS, Pell JP. Meta-analysis of the asso-ciation between preterm delivery and intelligence. J Public Health 2012;34:209e16.

[11] Johnson S. Cognitive and behavioural outcomes following very preterm birth.Semin Fetal Neonatal Med 2007;12:363e73.

[12] Lezak MD, Howieson DB, Bigler RD, Tranel D. Neuropsychological assess-ment. New York: Oxford University Press; 2012.

*[13] Rose SA, Feldman JF. Memory and processing speed in preterm childrenat eleven years: a comparison with full-terms. Child Dev 1996;67:2005e21.

[14] Kail R. Developmental change in speed of processing during childhood andadolescence. Psychol Bull 1991;109:490e501.

[15] Fry AF, Hale S. Relationships among processing speed, working memory, andfluid intelligence in children. Biol Psychol 2000;54:1e34.

[16] Rose SA, Feldman JF, Jankowski JJ. Processing speed in the 1st year of life: alongitudinal study of preterm and full-term infants. Dev Psychol 2002;38:895e902.

[17] de Kieviet JF, van Elburg RM, Lafeber HN, Oosterlaan J. Attention problems ofvery preterm children compared with age-matched term controls at school-age. J Pediatr 2012;161:824e9.

[18] Nosarti C, Giouroukou E, Micali N, Rifkin L, Morris RG, Murray RM. Impairedexecutive functioning in young adults born very preterm. J Int NeuropsycholSoc 2007;13:571e81.

[19] Strang-Karlsson S, Andersson S, Paile-Hyvarinen M, Darby D, Hovi P,Raikkonen K, et al. Slower reaction times and impaired learning in youngadults with birth weight <1500 g. Pediatrics 2010;125:E74e82.

*[20] Anderson PJ, De Luca CR, Hutchinson E, Spencer-Smith MM, Roberts G,Doyle LW, et al. Attention problems in a representative sample of extremelypreterm/extremely low birth weight children. Dev Neuropsychol 2011;36:57e73.

[21] Mirsky AF, Duncan CC. A nosology of disorders of attention. Ann N Y Acad Sci2001;931:17e32.

[22] Posner MI, Dehaene S. Attentional networks. Trends Neurosci 1994;17:75e9.[23] Wilson-Ching M, Molloy CS, Anderson VA, Burnett A, Roberts G, Cheong JLY,

et al. Attention difficulties in a contemporary geographic cohort of adoles-cents born extremely preterm/extremely low birth weight. J Int Neuro-psychol Soc 2013;19:1e12.

[24] van de Weijer-Bergsma EV, Wijnroks L, Jongmans MJ. Attention developmentin infants and preschool children born preterm: a review. Infant Behav Dev2008;31:333e51.

*[25] Mulder H, Pitchford NJ, Hagger MS, Marlow N. Development of executivefunction and attention in preterm children: a systematic review. Dev Neu-ropsychol 2009;34:393e421.

[26] Molloy CS, Wilson-Ching M, Anderson VA, Roberts G, Anderson PJ, Doyle LW.Visual processing in adolescents born extremely low birth weight and/orextremely preterm. Pediatrics 2013;132:1e9.

[27] Powls A, Botting N, Cooke RWI, Stephenson G, Marlow N. Visual impairmentin very low birthweight children. Arch Dis Child 1997;76:F82e7.

[28] Hard AL, Niklasson A, Svensson E, Hellstrom A. Visual function in school-agedchildren born before 29 weeks of gestation: a population-based study. DevMed Child Neurol 2000;42:100e5.

[29] O’Connor AR, Stephenson T, Johnson A, Tobin MJ, Moseley MJ, Ratib S, et al.Long-term ophthalmic outcome of low birth weight children with andwithout retinopathy of prematurity. Pediatrics 2002;109:12e8.

[30] Ornstein M, Ohlsson A, Edmonds J, Asztalos E. Neonatal follow-up of very-low-birth-weight/extremely low-birth-weight infants to school age e acritical overview. Acta Paediatr Scand 1991;80:741e8.

[31] Torrioli MG, Frisone MF, Bonvini L, Luciano R, Pasca MG, Lepori R, et al.Perceptual-motor, visual and cognitive ability in very low birthweight pre-school children without neonatal ultrasound abnormalities. Brain Dev2000;22:163e8.

[32] Rose SA, Feldman JF, Jankowski JJ. Recall memory in the first three years oflife: a longitudinal study of preterm and term children. Dev Med Child Neurol2005;47:653e9.

[33] de Haan M, Bauer PJ, Georgieff MK, Nelson CA. Explicit memory in low-riskinfants aged 19 months born between 27 and 42 weeks of gestation. DevMed Child Neurol 2000;42:304e12.

[34] Anderson PJ, Doyle LW, Victorian Infant Collaborative Study Group. Ex-ecutive functioning in school-aged children who were born very pretermor with extremely low birth weight in the 1990s. Pediatrics 2004;114:50e7.

*[35] Aarnoudse-Moens CSH, Weisglas-Kuperus N, van Goudoever JB,Oosterlaan J. Meta-analysis of neurobehavioral outcomes in very pretermand/or very low birth weight children. Pediatrics 2009;124:717e28.

[36] Rose SA, Feldman JF, Jankowski JJ, Van Rossem R. Basic information pro-cessing abilities at 11 years account for deficits in IQ associated with pretermbirth. Intelligence 2011;39:198e209.

[37] Luciana M, Lindeke L, Georgieff M, Mills M, Nelson CA. Neurobehavioralevidence for working memory deficits in school-aged children with historiesof prematurity. Dev Med Child Neurol 1999;41:521e33.

[38] Baron IS, Erickson K, Ahronovich MD, Litman FR, Brandt J. Spatial locationmemory discriminates children born at extremely low birth weight and late-preterm at age three. Neuropsychology 2010;24:787e94.

[39] Rushe TM, Rifkin L, Stewart AL, Townsend JP, Roth SC, Wyatt JS, et al. Neu-ropsychological outcome at adolescence of very preterm birth and its rela-tion to brain structure. Dev Med Child Neurol 2001;43:226e33.

[40] Saavalainen P, Luoma L, Bowler D, Maatta S, Kiviniemi V, Laukkanen E, et al.Spatial span in very prematurely born adolescents. Dev Neuropsychol2007;32:769e85.

[41] Curtis WJ, Lindeke LL, Georgieff MK, Nelson CA. Neurobehavioural func-tioning in neonatal intensive care unit graduates in late childhood and earlyadolescence. Brain 2002;125:1646e59.

[42] Taylor HG, Klein N, Minich NM, Hack M. Verbal memory deficits in childrenwith less than 750 g birth weight. Child Neuropsychol 2000;6:49e63.

[43] Taylor HG, Minich N, Bangert B, Filpek PA, Hack M. Long-term neuro-psychological outcomes of very low birth weight: associations with earlyrisks for periventricular brain insults. J Int Neuropsychol Soc 2004;10:987e1004.

*[44] Omizzolo C, Scratch SE, Stargatt R, Kidokoro H, Thompson DK, Lee KJ, et al.Neonatal brain abnormalities and memory and learning outcomes at 7 yearsin children born very preterm. Memory 2013 Jun 27. http://dx.doi.org/10.1080/09658211.2013.809765 [Epub ahead of print].

[45] Narberhaus A, Segarra D, Gimenez M, Junque C, Pueyo R, Botet F. Memoryperformance in a sample of very low birth weight adolescents. Dev Neuro-psychol 2007;31:129e35.

[46] Barre N, Morgan A, Doyle LW, Anderson PJ. Language abilities in childrenwhowere very preterm and/or very low birth weight: a meta-analysis. J Pediatr2011;158:766e74. e1.

[47] van Noort-van der Spek IL, Franken M, Weisglas-Kuperus N. Languagefunctions in preterm-born children: a systematic review and meta-analysis.Pediatrics 2012;129:745e54.

[48] Reidy N, Morgan A, Thompson DK, Inder TE, Doyle LW, Anderson PJ. Impairedlanguage abilities and white matter abnormalities in children born verypreterm and/or very low birth weight. J Pediatr 2013;162:719e24.

[49] Anderson PJ, Reidy N. Assessing executive function in preschoolers. Neuro-psychol Rev 2012;22:345e60.

[50] Aarnoudse-Moens CSH, Smidts DP, Oosterlaan J, Duivenvoorden HJ, Weis-glas-Kuperus N. Executive function in very preterm children at early schoolage. J Abnorm Child Psychol 2009;37:981e93.

[51] Baron IS, Kerns KA,Muller U, AhronovichMD, Litman FR. Executive functions inextremely low birth weight and late-preterm preschoolers: effects onworkingmemory and response inhibition. Child Neuropsychol 2012;18:586e99.

[52] Johnson S, Hennessy E, Smith R, Trikic R, Wolke D, Marlow N. Academicattainment and special educational needs in extremely preterm children at 11years of age: the EPICure study. Arch Dis Child Fetal Neonatal Ed 2009;94:F283e9.

[53] Litt JS, Taylor HG, Margevicius S, Schluchter M, Andreias L, Hack M. Academicachievement of adolescents born with extremely low birth weight. ActaPaediatr 2012;101:1240e5.

[54] Anderson P, Doyle LW, the Victorian Infant Collaborative Study Group.Neurobehavioral outcomes of school-age children born extremely low birthweight or very preterm in the 1990s. J Am Med Assoc 2003;289:3264e72.

[55] TaylorHG, EspyKA,AndersonPJ.Mathematicsdeficiencies in childrenwithverylow birth weight or very preterm birth. Dev Disabil Res Rev 2009;15:52e9.

*[56] Litt J, Taylor HG, Klein N, Hack M. Learning disabilities in children with verylow birthweight: prevalence, neuropsychological correlates, and educa-tional interventions. J Learn Disabil 2005;38:130e41.

[57] Grunau RE, Whitfield MF, Davis C. Pattern of learning disabilities in childrenwith extremely low birth weight and broadly average intelligence. ArchPediatr Adolesc Med 2002;156:615e20.

[58] Aarnoudse-Moens CSH, Oosterlaan J, Duivenvoorden HJ, van Goudoever JB,Weisglas-Kuperus N. Development of preschool and academic skills in chil-dren born very preterm. J Pediatr 2011;158:15e20.

[59] Hack M, Flannery DJ, Schluchter M, Cartar L, Borawski E, Klein N. Outcomes inyoung adulthood for very-low-birth-weight infants. N Engl J Med 2002;346:149e57.

[60] Roberts G, Lim J, Doyle LW, Anderson PJ. High rates of school readiness dif-ficulties at 5 years of age in very preterm infants compared with term con-trols. J Dev Behav Pediatr 2011;32:117e24.

[61] Mathiasen R, Hansen BM, Anderson AMN, Greisen G. Socio-economicachievements of individuals born very preterm at the age of 27 to 29 years: anationwide cohort study. Dev Med Child Neurol 2009;51:901e8.

[62] Saigal S, Stoskopf B, Streiner D, Boyle M, Pinelli J, Paneth N, et al. Transition ofextremely low-birth-weight infants from adolescence to young adulthood ecomparison with normal birth-weight controls. J Am Med Assoc 2006;295:667e75.

[63] Taylor HG. Children born preterm or with very low birth weight can haveboth global and selective cognitive deficits. J Dev Behav Pediatr 2006;27:485e6.

*[64] Mulder H, Pitchford NJ, Marlow N. Processing speed and working memoryunderlie academic attainment in very preterm children. Arch Dis Child FetalNeonatal Ed 2010;95:F267e72.

[65] Mulder H, Pitchford NJ, Marlow N. Processing speed mediates executivefunction difficulties in very preterm children in middle childhood. J IntNeuropsychol Soc 2011;17:445e54.

P.J. Anderson / Seminars in Fetal & Neonatal Medicine 19 (2014) 90e9696

[66] Mulder H, Pitchford NJ, Marlow N. Inattentive behaviour is associated withpoor working memory and slow processing speed in very pre-term childrenin middle childhood. Br J Educ Psychol 2011;81:147e60.

[67] Aarnoudse-Moens CSH, Weisglas-Kuperus N, Duivenvoorden HJ, vanGoudoever JB, Oosterlaan J. Executive function and IQ predict mathematicaland attention problems in very preterm children. Plos One 2013;8(2):e55994.

[68] Spittle A, Orton J, Anderson P, Boyd R, Doyle LW. Early developmentalintervention programmes post-hospital discharge to prevent motor and

cognitive impairments in preterm infants. Cochrane Database Syst Rev2012;(12):CD005495.

[69] Grunewaldt KH, Lohaugen GCC, Austeng D, Brubakk AM, Skranes J. Workingmemory training improves cognitive function in VLBW preschoolers. Pedi-atrics 2013;131:E747e54.

[70] Lohaugen GCC, Antonsen I, Haberg A, Gramstad A, Vik T, Brubakk AM, et al.Computerized working memory training improves function in adolescentsborn at extremely low birth weight. J Pediatr 2011;158:555e61.

lable at ScienceDirect

Seminars in Fetal & Neonatal Medicine 19 (2014) 97e104

Contents lists avai

Seminars in Fetal & Neonatal Medicine

journal homepage: www.elsevier .com/locate/s iny

Review

Growing up after extremely preterm birth: Lifespan mental healthoutcomes

Samantha Johnson a,*, Neil Marlow b

aDepartment of Health Sciences, University of Leicester, Leicester LE1 6TP, UKbDepartment of Academic Neonatology, Institute for Women’s Health, University College London, London, UK

Keywords:Attention deficit hyperactivity disorderAutism spectrum disordersComorbidityMental healthOutcomesPreterm birth

* Corresponding author. Address: Department of HLeicester, 22e28 Princess Road West, Leicester LE1 65798; fax: þ44 (0) 116 252 3272.

E-mail address: sjj19@le.ac.uk (S. Johnson).

1744-165X/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.siny.2013.11.004

s u m m a r y

There is growing interest in the long-termmentalhealth sequelaeof extremely pretermbirth. In thispaperwereview literature relating to mental health outcomes across the lifespan. Studies conducted in the preschoolyears, school age and adolescence, and adulthood showcontinuity in outcomes and point to an increased riskfor inattention, socio-communicative problems and emotional difficulties in individuals born extremelypreterm. Both behavioural and neuroimaging studies also provide evidence of a neurodevelopmental originformental health disorders in this population. Herewe summarise contemporary evidence and highlight keymethodological considerations for carrying out and interpreting studies in this field.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Extremely preterm (EP) births, before 28 weeks of gestation,continue to pose the greatest challenge for neonatal medicine.Providing life-sustaining treatment, minimising environmentalstressors and supporting the family through a traumatic life event arekey challenges for neonatologists and other professionals involved inperinatal care. For these babies and their families, however, the caredoes not end there. The biological vulnerability conferred by EP birth,which may be amplified through socio-economic disadvantage, canhave a profound impact on development with consequences thatextendacross the lifespan.AlthoughEPbirths comprise just 0.6% of allbirths, morbidity is highest among these survivors [1,2]. Cognitiveimpairments are themost frequent adverse outcomes [3,4], but thereis growing interest in the impact of preterm birth on mental healthand wellbeing. Here we review literature relating to mental healthoutcomes following EP birth. Although we focus on reports from themost contemporary cohorts, much may be gained through under-standing outcomes for older cohorts now in adult life.

2. Studying mental health following extremely preterm birth

Mental health outcomes are generally evaluated as part of lon-gitudinal studies which have, for the most part, sought to identify

ealth Sciences, University ofTP, UK. Tel.: þ44 (0) 116 252

All rights reserved.

the prevalence of disorders at various ages. Like all outcomestudies, these suffer the inherent problems of selective drop-out.Some of the issues relating to the maintenance of cohorts haverecently been discussed [5]. Key aspects of cohort evaluations are:

� Having a clear denominator in order to evaluate how the find-ings may be extrapolated to other studies and how represen-tative they are of the population from which they are drawn.

� Evaluating the effect of drop-outs and, where necessary, sup-plementing the findings with sensitivity analyses or imputationtechniques.

� Having due regard to these in drawing conclusions.

Single centre studies are more practical to manage, but groupsof babies born in individual hospitals may not be representative ofthe wider population. Further challenges occur by the simple factthat populations change over time, such that more contemporarycohorts comprise a higher proportion of EP children making com-parison with historical reports challenging. When studying out-comes it is important that there is a strong underlying hypothesisand that pre-study power calculations using realistic estimates ofgroup differences are computed.

It is widely considered preferable to use diagnostic criteria forstudying mental health disorders and to facilitate comparison be-tween studies, yet variation may still exist depending on themeasure used [6]. However, the often insurmountable economicand practical challenges of implementing diagnostic interviewsmean that most studies have relied solely on behavioural ques-tionnaires (see Johnson [7] and Arpi and Ferrari [8] for reviews).These typically generate higher rates of individuals that score above

S. Johnson, N. Marlow / Seminars in Fetal & Neonatal Medicine 19 (2014) 97e10498

the cut-off for clinically significant problems than meet the criteriafor disorders. This is illustrated using data from the UK EPICureStudy (Fig. 1) [9]. As part of a follow-up at 11 years of age,emotional, conduct, hyperactivity/inattention and peer relation-ship problems in a cohort of children born EP (<26 weeks) wereassessed using parent and teacher questionnaires; diagnoses ofcorresponding disorders were obtained concurrently [10]. For alldomains, parents reported significantly more problems than dis-orders and teachers reported more attention and peer problems(Fig. 1). This begs the question of who is the most appropriaterespondent for assessing childhood psychopathology. It is well-documented that parent and teacher reports are only modestlycorrelated and that parents report higher rates of problems thanteachers or adolescents themselves, particularly for emotionaldisorders [11�17]. Obtaining multi-informant data is thereforeadvocated for mental health assessment [18,19].

The use of dimensional measures is also advocated for studyingchildhood psychopathology in order to quantify the degree towhich symptoms are manifest in individuals and populations [20e23]. These considerations are particularly important for studyingmental health following EP birth in which there appears to be ageneral population shift in psychopathology and a cluster ofsymptoms that extends across diagnostic boundaries (see Section4) [24e26]. As the expression of childhood psychopathology alterswith development, the nature, severity and frequency of behav-iours that are considered typical at one age may be rated as path-ological at another [20]. It is therefore important to use age- andgender-specific norms and to obtain contemporaneous referencedata from term-born controls. Where possible, control groupsshould be matched, or analyses adjusted, for confounding factorssuch as age, sex and socio-economic status. There is controversyover adjusting for IQ given statistical and theoretical limitations[27,28] and the comorbidity of neurocognitive sequelae in EPchildren (see Section 4.5). Parental mental health may also be aconfounder in light of the higher risk for psychopathology in theoffspring of those with disorders [29]. However, although parentsof preterm children are at risk for parenting stress and poor mentalhealth [30e32] medical, biological and neurodevelopmental vari-ables are stronger predictors of childhood psychopathology inpreterm samples [33e36], and there is inconsistency in studies of

Fig. 1. Prevalence of parent- and teacher-reported emotional, conduct, attention and peer p219 children born extremely preterm (<26 weeks of gestation; EPICure Study) [10]. AsterisksDifficulties Questionnaires and psychiatric diagnoses (P < 0.05). ADHD, attention deficit hy

the relationship between parental mental health and pretermchildren’s socio-emotional development [15,30,37,38]. These asso-ciations are likely to be bidirectional, potentially mediated by thequality of parenteinfant interaction [30,39e41]. The importance ofparental mental health as a causal factor in EP children’s psycho-pathology requires elucidation in longitudinal studies. Parentalmental health is discussed further in this issue by Karli Treyvaud(Chapter 10).

3. The preschool years

3.1. Behaviour and emotional problems in the preschool years

There is a surprising lack of research regarding behaviouraloutcomes during the preschool years. Studies in infancy havefocused on the development of attachment relationships, temper-ament and parenteinfant interaction (see Korja et al. [42] andVanderbilt [43] for reviews). The assessment of early psychopa-thology becomes more refined from the age of 2e3 years whenwell-standardised tools are available to identify clinically signifi-cant difficulties, such as the Child Behavior Checklist (CBCL) [44]and the Strengths and Difficulties Questionnaire (SDQ) [45].

Only a few recent studies have investigated outcomes in EP orextremely low birthweight (ELBW) preschoolers (Table 1). In twostudies, children hadmore problems on all SDQ scales, suggesting ageneric risk for mental health problems following EP birth [46,47].Interestingly, Elgen et al. [47] reported that 38% of EP children hadclinically significant scores yet only 8% had been referred for psy-chiatric follow-up, highlighting the preponderance of subclinicalsymptoms. In a longitudinal study, EP children had significantlypoorer emotional and behavioural regulation than term-bornchildren at both 2 and 4 years of age (Table 1); moreover, theyshowed less developmental gain than full-term and VP children,which is suggestive of a specific vulnerability in the development ofearly regulatory competence in EP children [48]. More recently,Scott et al. [49] obtained multi-informant data on EP/ELBW chil-dren at 5 years of age. Using both a dimensional and diagnosticapproach, the authors reported greater specificity in outcomescompared with earlier studies; only attention deficit/hyperactivitydisorder (ADHD) symptoms were significantly and consistently

roblems and diagnoses of corresponding psychiatric disorders at 11 years of age amongdenote significant between-group differences between informant-rated Strengths andperactivity disorder; ASD, autism spectrum disorders.

Table 1Prospective studies of behavioural and emotional problems and psychiatric disorders in extremely low birthweight (ELBW) or extremely preterm (EP) preschoolers born in the1990s and beyonda.

Publication Sample characteristics Domains with significantly highermean scores in EP/ELBW children

Measure

Age (years) Birth year Preterm Control

Clark et al. [48] 2 1998e2001 N ¼ 39 N ¼ 103 Regulation BRSSingle centre <28 weeks

4 Regulation ERCRegulation BRS

Woodward et al.[46] 4 1998e2000 N ¼ 43 N ¼ 107 Emotional SDQConductSingle centre <28 weeksHyperactivity/inattentionPeer relationshipTotal difficulties

Elgen et al. [47] 5 1999e2000 N ¼ 255 N ¼ 1089 Emotional SDQNational <28 weeks/<1000 g Conduct

Hyperactivity/inattentionPeer relationshipTotal difficulties

Scott et al. [49] 5 2001e2003 N ¼ 148 N ¼ 111 ADHD problems CBCLSingle centre <28 weeks/<1000 g ADHD problems TRF

Affective problems P-ChIPSADHD/inattentiveODD

BRS, Behaviour Rating Scale (abbreviated version) of the Bayley Scales of Infant Development, 2nd ed., rated by a researcher. ERC, Emotion Regulation Checklist (modifiedversion), rated by parents; SDQ, Strengths and Difficulties Questionnaire; CBCL, Child Behavior Checklist; TRF, Teacher Report Form; P-ChIPS, Children’s Interview for Psy-chiatric Syndromes e Parent.

a Studies are included where they report results for comparisons between EP/ELBW children and term-born controls.

S. Johnson, N. Marlow / Seminars in Fetal & Neonatal Medicine 19 (2014) 97e104 99

increased, and there was a 2.5-fold increased risk for ADHD di-agnoses (Table 1). Although parents and teachers rated higherscores for affective and oppositional problems, respectively, noother disorders were increased. These studies highlight an associ-ation between EP birth, regulatory problems and ADHD that isalready evident in the early years.

3.2. Early indicators of autism spectrum disorders (ASD)

An area gathering increasing interest is the risk for ASD in EPinfants (Table 2). In recent reports, 21e41% of EP [50,51] infantsscreened positive for ASD at 18e24 months using the ModifiedChecklist for Autism in Toddlers (M-CHAT) parent questionnaire. Ininfants born at <27 weeks, 10e20% screened positive using a rangeof ASD screening tools [52]. There is warranted concern that thesestudies have substantially overestimated the true risk for ASD in

Table 2Cohort studies that have investigated autism spectrum symptoms and disorders in extre

Publication Sample characteristics

Age (years) Birth year Preterm Cont

The early yearsKuban et al. [50] 2 2002e2004 N ¼ 988 e

Regional <28 weeksMoore et al. [51] 2 2006 N ¼ 523 e

National <26 weeksStephens et al. [52] 2 2008e2009 N ¼ 554 e

Regional <27 weeksSchool age and adolescenceTreyvaud et al. [60] 7 2001e2003 N ¼ 177 N ¼ 6

Single centre <30 weeks/<1250 gHack et al. [24] 8 1992e1995 N ¼ 219 N ¼ 1

Single centre <1000 gJohnson et al. [10] 11 1995 N ¼ 219 N ¼ 1

National <26 weeks

Johnson et al. [25] 11 1995 N ¼ 219 N ¼ 1National <26 weeks

M-CHAT, Modified Checklist for Autism in Toddlers; PDD-2, Pervasive DevelopmentaDevelopment and Well Being Assessment; SCQ, Social Communication Questionnaire; n

this population [53]. Not only are screening tests associated with ahigh rate of false positives, but the high prevalence of cognitive,motor and sensory impairments in this population further con-founds scores on these scales [50e52,54]. Importantly, thesestudies have only used screening questionnaires and have notimplemented the M-CHAT follow-up interview to improve speci-ficity. The predictive validity of positive screens and the trueprevalence of ASD in this population require investigation: studiesat school age may hold the answer (see Section 4).

3.3. Predictive validity of early assessments

There is a paucity of longitudinal studies of behavioural out-comes in EP cohorts. Studies of VLBW children have indicatedmoderate stability in problems over the preschool years [55,56] andthere is evidence that early screening is predictive of later mental

mely preterm and/or extremely low birthweight children.

Prevalence of problems P Measure

rol Preterm Control

21% e e M-CHAT

41% e e M-CHAT

10% e e PDD-2

5 4.5% 0 P ¼ 0.08 DAWBA

76 2% 0.6% ns CSI-4: autistic disorder1% 0 ns CSI-4: Asperger disorder

53 8% 0 P < 0.001 DAWBA: any ASD6.5% 0 P ¼ 0.001 ns DAWBA: autistic disorder1.5% 0 DAWBA: atypical autism

53 15.8% 2.9% P < 0.001 SCQ

l Disorders Screening Test, 2nd ed.; CSI-4, Child Symptom Inventory 4; DAWBA,s, non-significant.

S. Johnson, N. Marlow / Seminars in Fetal & Neonatal Medicine 19 (2014) 97e104100

health [57]. In EP children, internalising behaviour problems at 2.5years were a significant predictor of disorders at 11 years of age,whereas externalising difficulties were not [10]. Scores on the CBCLWithdrawn subscale have also been shown to be a significantpredictor of later ASD symptoms [25], and scores on this and theEmotionally Reactive subscale have also been associated withpositive M-CHAT screens [58]. Using a VP sample, Treyvaud et al.[59] have also shown that internalising difficulties, socialeemotional competence and externalising problems at 2 years pre-dicted emotional symptoms, peer problems and conduct problemsat 5 years, respectively. Socialeemotional problems at age 5 yearsalso predicted psychiatric disorders at age 7 years [60]. With suchspecificity in prediction, it may be assumed that there is stability inoutcomes among EP children and that the early problems high-lighted above will manifest in attention, emotional problems andASD later in childhood.

4. School age and adolescence

4.1. Dimensional studies

EP survivors are at high risk for clinically significant problemsthroughout middle childhood, with prevalence estimates of 18e38% [24,26,61e63]. Studies have also shown that the mental healthproblems of EP children have a greater impact on their daily livingthan those of term-born controls [62,64]. Commensurate with thetemporal stability in neurodevelopmental outcomes [65], the rateof mental health problems remains high despite advances inneonatal care [61,64]. Whereas there is a greater risk for behaviourproblems in EP versus VP children [3,66], there is little evidence fora gestation-related gradient within the EP group itself [62,64].

Compared with the preschool years, there is greater consistencyin findings at school age. In three studies of six EP/ELBW cohorts,the authors identified an excess of attention, social and thought oremotional problems despite using different measures [61,64,67]. Ina study using multi-informant data to identify pervasive problems,EP children had significantly more problems on all scales but oddsratios (ORs) were greatest for these three domains [62]. Conradet al. [14] also found that only depression/anxiety and hyperactiv-ity/inattention were significantly increased in ELBW children aged7e16 years. These findings thus point to a cluster of attention, socialand emotional problems in EP children [26].

4.2. Diagnostic studies

In a recent meta-analysis of five studies [68], the authors re-ported a weighted OR of 3.66 for psychiatric disorders among

Table 3Mental health disorders in children born extremely preterm (EP) or with extremely low

Publication Sample characteristics

Age (years) Birth year Preterm Contro

Scott et al. [49] 5 2001e2003 N ¼ 148 N ¼ 11Single centre <28 weeks/<1000 g

Treyvaud et al. [60] 7 2001e2003 N ¼ 177 N ¼ 65Single centre <30 weeks/<1250 g

Hack et al. [24] 8 1992e1995 N ¼ 219 N ¼ 17Single centre <1000 g

Johnson et al. [10] 11 1995 N ¼ 219 N ¼ 15National <26 weeks

OR, odds ratio; CI, confidence interval; P-ChIPS, Children’s Interview for Psychiatric SyndrBeing Assessment.

preterm children [68]. Three studies of EP/ELBW children reportconsistently elevated risks for disorders with prevalence estimatesranging 23e32% [10,24,60]. These also provide evidence for thespecific risk for attention and social problems in this population(Table 3) [10,24]. In their VP cohort, Treyvaud et al. [60] did notreport a significant increase in these disorders which the authorsacknowledge may be a result of low statistical power; however,there was a trend for a higher rate of ADHD and anxiety disorders.Thus there appears to be a pattern of elevated risk for symptomsand disorders associated with ADHD, peer relationship problemsand emotional disorders, the cluster of which has been termed thepreterm behavioural phenotype [26].

The symptoms and correlates of ADHD and ASD in EP childrenmay indicate a different aetiology that is associated with aberrantbrain development. Indeed, Laucht et al. [29] have shown thatpsychosocial and biological risk factors had independent effects onbehavioural outcomes; specifically, psychosocial risk was morestrongly associated with externalising difficulties whereas biolog-ical risk factors, such as preterm birth, were associated with iso-lated social and attention problems. Other studies have also notedstronger associations of attention and peer problems with neuro-developmental impairments [47]. Such studies provide increasingsupport for an environmental, neurodevelopmental origin forADHD and ASD in EP children. Literature relating to ADHD and ASDis summarised briefly below.

4.3. Attention deficit hyperactivity disorder

As noted earlier, Scott et al. [49] reported a 2.5-fold increasedrisk for ADHD in EP children at 5 years of age; these findings areechoed later in childhood in which authors have reported ORs of2.6e2.7 for ADHD in VP/VLBW children [69,70] and higher ORs of4.2e4.3 for ADHD in those born EP/ELBW [10,24]. Dimensionalmeasures have indicated a generally increased liability to ADHDsymptoms in EP children and an excess of children with problemswho do not meet diagnostic criteria [66,71,72]. Increasing interesthas focused on the expression of ADHD symptoms in this popula-tion with a number of studies highlighting a higher risk for inat-tention relative to hyperactivity/impulsivity in terms of bothsymptoms and disorders [10,34,38,73]. There is also a notableabsence of comorbid conduct disorders that are frequentlyobserved in general population samples (Table 3) [10,61,72]. It hastherefore been suggested that ADHD in preterm populations maybe better described as inattentive subtype disorders with a purerneurodevelopmental origin [26,67,74]. The cognitive profile of EPbirth is characterised by core deficits in working memory andvisuo-spatial skills which have been shown to mediate

birthweight (ELBW) in the 1990s and beyond.

Prevalence of any disorder Disorders significantlyincreased in EP/ELBWchildren

Measure

l Index vs control OR (95% CI)

1 e ADHD/combined P-ChIPSADHD/hyperactive

24% vs 9% 3.13 (1.27e7.71) None DAWBA

6 32% vs 15% 2.7 (1.6e4.5) ADHD (any) CSI-4ADHD/inattentiveADHD/combinedSpecific phobia

3 23% vs 9% 3.2 (1.7e6.2) ADHD (any) DAWBAADHD/inattentiveEmotional disordersASD

omes e Parent; CSI-4, Child Symptom Inventory 4; DAWBA, Development andWell-

Fig. 2. Prevalence of moderate/severe neurodevelopmental disability and psychiatricdisorders in 219 11-year-old children born extremely preterm (<26 weeks of gestation)and a comparison group of 153 classmates born at term (EPICure Study) [4,10].

S. Johnson, N. Marlow / Seminars in Fetal & Neonatal Medicine 19 (2014) 97e104 101

performance on executive function and intelligence tests [75,76].These core deficits may also be implicated in the inattention diffi-culties observed in EP children [73,77].

4.4. Autism spectrum disorders

Studies at school age have confirmed a significantly increasedrisk for ASD in EP children, and, as anticipated, the 4e8% prevalenceis markedly lower than the positive screen rate in infancy [24,25]. Inaddition, almost twice as many children have clinically significantsymptoms than have diagnoses [25]. As with ADHD, ASD may alsohave a different origin in EP children that is again associated withaberrant brain development [10,26,50]. ASD in EP children aremore closely associated with a smaller head circumference,cognitive deficits and neurological injuries than are ASD in thegeneral population [10,50,78,79]. Moreover, EP children havegreater symptoms on the dimensions of impaired social interactionand communication than repetitive or stereotyped behaviour, thelatter of which is a core symptom domain in diagnostic classifica-tions [71,80]. As such, these symptoms may be better characterisedby a social communication disorder. Like ASD, these impairmentsmay be mediated by EP children’s cognitive deficits that adverselyimpact on their processing of social and emotional stimuli.

4.5. Comorbidity

The few studies addressing issues of comorbidity have reportedmixed results regarding the rate of comorbid mental health disor-ders in EP children compared with controls [10,24,49,60]. The mostconsistent finding is the lack of comorbid ADHD and conduct dis-orders.What is perhaps amore pertinent question is towhat extentmental health disorders confer additional morbidity over otherneurodevelopmental sequelae. Numerous studies have shown thatmental health problems are associated with neurosensory, cogni-tive and motor impairments in this population and, in the majorityof cases, these do not account for the excess of mental health dis-orders [3,24,56,62]. In the EPICure Study, mental health disordersconferred only a small degree of additional morbidity over neuro-developmental disabilities in EP children (Fig. 2). However, it isimportant to note that there is a substantial portion of childrenwhose parents and teachers report clinically significant difficultiesbut who do not meet criteria for diagnoses (Fig. 1); thus the truefunctional impact of mental health problems is likely to be fargreater than is suggested by the excess of psychiatric disordersalone.

5. Mental health in adulthood

Emerging data from population registry linkage studies inScandinavia indicate a significant increase in the risk of adultmental health disorders with decreasing gestational age at birth.Because of the relatively low population prevalence of these con-ditions, such large, and necessarily broad, studies are the only onesto reliably report such findings. Moster et al. [81] reported anincreased relative risk (RR) of ASD [9.5; 95% confidence interval(CI): 1.5, 36.2] and other disorders of psychological development,behaviour and emotions (10.5; 5.6, 19.9) in Norwegian adults bornEP. Although the risk of schizophrenia was not significantlyelevated, it is worth noting that the point prevalence was low eventhough the population was censored at 36 years of age. In a furtherstudy of treatment registration among adults in Norway, Halmoyet al. [82] confirmed that EP survivors remain at increased risk forADHD in adulthood (adjusted RR: 5.0; 2.1, 11.8), and a Swedishstudy of psychiatric hospitalisations to 23 years also reportedincreased risk for non-affective psychosis, depressive disorder and

bipolar affective disorder in adults born <32 weeks [83]. Theselatter data confirm the earlier findings of increased admissions forpsychiatric and addictive disorders [84]. Finally, a study by Crumpet al. [85] confirmed that young adults born EP were morefrequently prescribed psychotrophic medications, specifically an-tipsychotics, antidepressants and hypnotics. As each of thesestudies demonstrated a significant ‘dose effect’ of prematurity anda gradation of increased risk with decreasing gestation, there isapparent continuity of risk from the adolescent studies referred toabove.

Demonstrating this in smaller prospective cohort studies ismore difficult. Saigal et al. [86] reported greater internalisedbehaviours in young adults, together with an increased rate ofprescriptions for antidepressants. Several studies of VLBW youngadults have demonstrated a significant reduction in risk takingbehaviours and social interaction [72,87,88]. Most recently, Bur-nett et al. [89] have reported a sustained increase in the preva-lence of ADHD in EP/ELBW adolescents at 18 years of age, but nosignificant excess of anxiety or mood disorders compared withnormal birth weight controls, using both diagnostic anddimensional measures. This is somewhat unexpected given thatthe majority of previous cohort and population registry linkagestudies have identified increased mood disorders in adolescenceand adulthood. This is therefore encouraging but it remains to beseen whether such findings are replicated in other post-1990cohorts. Taken together, these studies show that although chil-dren remain at increased risk of psychiatric disorders in the long

Practice points

� Children born extremely preterm may have attention,emotional or peer relationship problems that do not meetdiagnostic criteria but which may impact on daily func-tion, for which intervention may be beneficial.

� Screening for behaviour problems from 2 years of agemay aid in detecting children with early difficulties andthose at risk of later mental health disorders.

� Researchers and clinicians should consider includingmental health assessments as part of neuro-developmental outcome evaluations.

Research directions

� Longitudinal studies are needed to investigate the evo-lution of mental health sequelae throughout the lifespan.

� The role of parental mental health in the development ofchildhood psychopathology requires elucidation in thispopulation.

� More studies are required to understand the neurologicalbases of mental health disorders in extremely pretermchildren.

S. Johnson, N. Marlow / Seminars in Fetal & Neonatal Medicine 19 (2014) 97e104102

term, the societal consequences of EP birth in young adulthoodappear to be somewhat less than anticipated, with better socialadaptation and quality of life than might once have beenpredicted.

6. A biological basis for psychiatric morbidity

A biological basis for psychiatric disorders in EP survivors isperhaps understandable after consideration of the effect of pre-maturity on brain development, which has been described as acomplex amalgam of destructive and developmental influences[90]. Studies using magnetic resonance imaging and computationaltechniques have identified differences in the brains of EP childrenand adults which may act as biomarkers for these evolving condi-tions [91]. These include not only ongoing adaptation to destructivelesions, but more subtle differences in brain size/surface area [92]and regional volumes such as the frontal and temporal cortex orhippocampus [93], deep gray matter [94] and corpus callosum [95].Reduced complexity of brain folding and abnormalities on func-tional activation imply altered developmental trajectories thatmirror impaired executive functions [96]. As discussed above, suchcognitive deficits may underpin mental health symptoms and dis-orders in EP children [77,97].

Although ex-preterm brain structure and function in relation tospecific psychiatric disorders have not been investigated in largepopulations of EP individuals, data from middle childhood sug-gests that structural alterations may be associated with behav-ioural findings, for example with attention and internalisingbehaviours (fractional anisotropy in a range of overlappingareas),[98] socio-emotional development (hippocampal size) [99]and wellbeing (cerebellar growth) [100]. As yet, such studieshave not been sufficiently large or systematic to develop a pictureof the underlying neural basis for psychiatric disorders associatedwith preterm birth, but the data suggest that there may be abiological basis for the clinically observed excess of psychopa-thology in this population.

7. Conclusions

Studies of mental health outcomes following EP birth havelargely sought to document the prevalence of disorders inmiddle childhood and adolescence, but reports of longer-termoutcomes are beginning to surface as contemporary cohortsreach adulthood. Early attention and regulatory problems areevident in the preschool years and, by childhood, the greaterspecificity in outcomes points to a cluster of inattention, peerrelationship problems and emotional symptoms. Approximately25% have psychiatric disorders but up to double those numbersmay have significant difficulties that impact on function. Thisbehavioural phenotype shows continuity into adult life withincreased risk for disorders, yet quality of life and social adap-tation are better than may have once been anticipated. Mentalhealth symptoms appear to have a strong neurodevelopmentalorigin which may be mediated by core cognitive deficits asso-ciated with EP birth. Targeted screening for the cluster of mentalhealth problems associated with EP birth may be beneficial fordetecting children with subclinical difficulties, particularly dur-ing the preschool years when problems start to become evident.Future studies are needed to identify resilience and risk factorsfor mental health disorders and to further elucidate the role ofparental mental health in the evolution of psychopathology inEP children. As more EP children continue to enter society, agreater understanding of the aetiology and functional impact ofthese disorders will aid in providing appropriate lifelongsupport.

Funding sources

Neil Marlow receives a proportion of funding from the Depart-ment of Health’s NIHR Biomedical Research Centres fundingscheme at UCLH/UCL.

References

[1] Saigal S, Doyle L. An overview of mortality and sequelae of preterm birthfrom infancy to adulthood. Lancet 2008;371:261e9.

[2] MacKay DF, Smith GC, Dobbie R, Pell JP. Gestational age at delivery andspecial educational need: retrospective cohort study of 407,503 school-children. PLoS Med 2010;7:e1000289.

[3] Anderson P, Doyle LW. Neurobehavioral outcomes of school-age childrenborn extremely low birth weight or very preterm in the 1990s. JAMA2003;289:3264e72.

[4] Johnson S, Fawke J, Hennessy E, Rowell V, Thomas S, Wolke D, et al.Neurodevelopmental disability through 11 years in children born before26 weeks of gestation: the EPICure study. Pediatrics 2009;124:e249e57.

[5] Marlow N. Measuring neurodevelopmental outcome in neonatal trials: acontinuing and increasing challenge. Arch Dis Child Fetal Neonatal Ed2013;98:F554e8.

[6] Angold A, Erkanli A, Copeland W, Goodman R, Fisher PW, Costello EJ.Psychiatric diagnostic interviews for children and adolescents: acomparative study. J Am Acad Child Adolesc Psychiatry 2012;51:506e17.

[7] Johnson S. Cognitive and behavioural outcomes following very pretermbirth. Semin Fetal Neonatal Med 2007;12:363e73.

[8] Arpi E, Ferrari F. Preterm birth and behaviour problems in infants andpreschool-age children: a review of the recent literature. Dev Med ChildNeurol 2013;55:788e96.

[9] EPICure Studies. Population based studies of survival and later healthstatus in extremely premature infants. www.epicure.ac.uk; 2013.

*[10] Johnson S, Hollis C, Kochhar P, Hennessy E, Wolke D, Marlow N. Psychiatricdisorders in extremely preterm children: longitudinal finding at age 11years in the EPICure study. J Am Acad Child Adolesc Psychiatry 2010;49.453.e1e463.e1.

[11] Bora S, Pritchard VE, Moor S, Austin NC, Woodward LJ. Emotional andbehavioural adjustment of children born very preterm at early school age.J Paediatr Child Health 2011;47:863e9.

[12] Dahl LB, Kaaresen PI, Tunby J, Handegard BH, Kvernmo S, Ronning JA.Emotional, behavioral, social, and academic outcomes in adolescents bornwith very low birth weight. Pediatrics 2006;118:E449e59.

[13] van Dulmen MHM, Egeland B. Analyzing multiple informant data on childand adolescent behavior problems: predictive validity and comparison ofaggregation procedures. Int J Behav Dev 2011;35:84e92.

[14] Conrad AL, Richman L, Lindgren S, Nopoulos P. Biological and environ-mental predictors of behavioral sequelae in children born preterm. Pedi-atrics 2010;125:e83e9.

S. Johnson, N. Marlow / Seminars in Fetal & Neonatal Medicine 19 (2014) 97e104 103

[15] Spiker D, Kraemer HC, Constantine NA, Bryant D. Reliability and validity ofbehavior problem checklists as measures of stable traits in low birthweight, premature preschoolers. Child Dev 1992;63:1481e96.

[16] Gardner F, Johnson A, Yudkin P, Bowler U, Hockley C, Mutch L, et al.Behavioral and emotional adjustment of teenagers in mainstream schoolwho were born before 29 weeks’ gestation. Pediatrics 2004;114:676e82.

[17] Saigal S, Pinelli J, Hoult L, Kim MM, Boyle M. Psychopathology and socialcompetencies of adolescents who were extremely low birth weight. Pe-diatrics 2003;111:969e75.

[18] Renk K. Cross-informant ratings of the behavior of children and adoles-cents: the “gold standard”. J Child Fam Stud 2005;14:457e68.

[19] Ford T, Goodman R, Meltzer H. The British child and adolescent mentalhealth survey 1999: the prevalence of DSM-IV disorders. J Am Acad ChildAdolesc Psychiatry 2003;42:1203e11.

[20] Hudziak JJ, Achenbach TM, Althoff RR, Pine DS. A dimensional approach todevelopmental psychopathology. Int J Methods Psychiatr Res2007;16(Suppl. 1):S16e23.

[21] Goodman A, Goodman R. Strengths and difficulties questionnaire as adimensional measure of child mental health. J Am Acad Child AdolescPsychiatry 2009;48:400e3.

[22] Nigg JT. Future directions in ADHD etiology research. J Clin Child AdolescPsychol 2012;41:524e33.

[23] Coghill D, Sonuga-Barke EJ. Annual research review: categories versusdimensions in the classification and conceptualisation of child andadolescent mental disorders e implications of recent empirical study. JChild Psychol Psychiatry Allied Discipl 2012;53:469e89.

*[24] Hack M, Taylor HG, Schlichter M, Andreias L, Drotar D, Klein N. Behavioraloutcomes of extremely low birth weight children at age 8 years. J DevBehav Pediatr 2009;30:122e30.

[25] Johnson S, Hollis C, Kochhar P, Hennessy E, Wolke D, Marlow N. Autismspectrum disorders in extremely preterm children. J Pediatr 2010;156:525e31.

*[26] Johnson S, Marlow N. Preterm birth and childhood psychiatric disorders.Pediatr Res 2011;69:11re8r.

[27] Dennis M, Francis DJ, Cirino PT, Schachar R, Barnes MA, Fletcher JM. WhyIQ is not a covariate in cognitive studies of neurodevelopmental disorders.J Int Neuropsychol Soc 2009;15:331e43.

[28] Taylor HG. Children born preterm or with very low birth weight can haveboth global and selective cognitive deficits. J Dev Behav Pediatr 2006;27:485e6. author reply 6e7.

[29] Laucht M, Esser G, Baving L, Gerhold M, Hoesch I, Ihle W, et al. Behavioralsequelae of perinatal insults and early family adversity at 8 years of age. JAm Acad Child Adolesc Psychiatry 2000;39:1229e37.

[30] Treyvaud K, Anderson VA, Lee KJ, Woodward LJ, Newnham C, Inder TE,et al. Parental mental health and early socialeemotional development ofchildren born very preterm. J Pediatr Psychol 2010;35:768e77.

[31] Miles MS, Holditch-Davis D, Schwartz TA, Scher M. Depressive symptomsin mothers of prematurely born infants. J Dev Behav Pediatr 2007;28:36e44.

[32] Glazebrook C, Marlow N, Israel C, Croudace T, Johnson S, White IR, et al.Randomised trial of a parenting intervention during neonatal intensivecare. Arch Dis Child Fetal Neonatal Ed 2007;92:F438e43.

[33] Indredavik MS, Vik T, Heyerdahl S, Kulseng S, Fayers P, Brubakk AM.Psychiatric symptoms and disorders in adolescents with low birth weight.Arch Dis Child 2004;89:F445e50.

[34] Botting N, Powls A, Cooke RWI, Marlow N. Attention deficit hyperac-tivity disorders and other psychiatric outcomes in very low birthweightchildren at 12 years. J Child Psychol Psychiatry Allied Discipl 1997;38:931e41.

[35] Indredavik MS, Vik T, Evensen AI, Skranes J, Taraldsen G, Brubakk A.Perinatal risk and psychiatric outcome in adolescents born preterm withvery low birth weight or term small for gestational age. J Dev BehavPediatr 2010;31:286e94.

*[36] Spittle AJ, Treyvaud K, Doyle LW, Roberts G, Lee KJ, Inder TE, et al. Earlyemergence of behavior and socialeemotional problems in very preterminfants. J Am Acad Child Adolesc Psychiatry 2009;48:909e18.

[37] Huhtala M, Korja R, Lehtonen L, Haataja L, Lapinleimu H, Rautava P.Parental psychological well-being and behavioral outcome of very lowbirth weight infants at 3 years. Pediatrics 2012;129:e937e44.

[38] Jaekel J, Wolke D, Bartmann P. Poor attention rather than hyperactivity/impulsivity predicts academic achievement in very preterm and full-termadolescents. Psychol Med 2013;43:183e96.

[39] Jaekel J, Wolke D, Chernova J. Mother and child behaviour in very pretermand term dyads at 6 and 8 years. Dev Med Child Neurol 2012;54:716e23.

[40] Forcada-Guex M, Pierrehubert B, Borghini A, Moessinger A, Muller-Nix CM. Early dyadic patterns of mothereinfant interactions and out-comes of prematurity at 18 months. Pediatrics 2006;118:107e14.

[41] Assel M, Landry S, Swank P, Steelman L, Miller-Loncar C, Smith K. How domothers’ childrearing histories, stress and parenting affect children’sbehavioural outcomes? Child Care Health Dev 2002;28:359e68.

[42] Korja R, Latva R, Lehtonen L. The effects of preterm birth on motherein-fant interaction and attachment during the infant’s first two years. ActaObstet Gynecol Scand 2012;91:164e73.

[43] Vanderbilt D, Gleason MM. Mental health concerns of the premature in-fant through the lifespan. Pediatr Clin North Am 2011;58:815e32.

[44] ASEBA. Achenbach system of empirically based assessment. http://www.aseba.org/; 2012.

[45] Goodman R. The strengths and difficulties questionnaire: a research note.J Child Psychol Psychiatry 1997;38:581e6.

[46] Woodward LJ, Moor S, Hood KM, Champion PR, Foster-Cohen S, Inder TE,et al. Very preterm children show impairments across multiple neuro-developmental domains by age 4 years. Archs Dis Childh Fetal NeonatalEd 2009;94:F339e44.

[47] Elgen SK, Leversen KT, Grundt JH, Hurum J, Sundby AB, Elgen IB, et al.Mental health at 5 years among children born extremely preterm: a na-tional population-based study. Eur Child Adolesc Psychiatry 2012;21:583e9.

[48] Clark CAC, Woodward LJ, Horwood LJ, Moor S. Development of emotionaland behavioral regulation in children born extremely preterm and verypreterm: biological and social influences. Child Dev 2008;79:1444e62.

*[49] Scott MN, Taylor HG, Fristad MA, Klein N, Espy KA, Minich N, et al.Behavior disorders in extremely preterm/extremely low birth weightchildren in kindergarten. J Dev Behav Pediatr 2012;33:202e13.

[50] Kuban KCK, O’Shea TM, Allred EN, Tager-Flusberg H, Goldstein DJ,Leviton A. Positive screening on the Modified Checklist for Autism inToddlers (M-CHAT) in extremely low gestational age newborns. J Pediatr2009;154:535e40.

[51] Moore T, Johnson S, Hennessy E, Marlow N. Screening for autism inextremely preterm infants: problems in interpretation. Dev Med ChildNeurol 2012;54:514e20.

[52] Stephens BE, Bann CM, Watson VE, Sheinkopf SJ, Perata-Carcelen M,Bodnar A, et al. Screening for autism spectrum disorders in extremelypreterm infants. J Dev Behav Pediatr 2012;33:535e41.

[53] Johnson S, Marlow N. Positive screening results on the modified checklistfor autism in toddlers: implications for very preterm populations. J Pediatr2009;154:478e80.

[54] Luyster RJ, Kuban KCK, O’Shea TM, Paneth N, Aldred EN, Leviton A. Themodified checklist for autism in toddlers in extremely low gestational agenewborns: individual items associated with motor, cognitive, vision andhearing limitations. Paediatr Perinat Epidemiol 2011;25:366e76.

[55] Gray RF, Indurkhya A, McCormick MC. Prevalence, stability, and predictorsof clinically significant behavior problems in low birth weight children at3, 5, and 8 years of age. Pediatrics 2004;114:736e43.

[56] Delobel-Ayoub M, Arnaud C, White-Koning M, Casper C, Pierrat V,Garel M, et al. Behavioral problems and cognitive performance at 5 yearsof age after very preterm birth: the EPIPAGE study. Pediatrics 2009;123:1485e92.

[57] Briggs-Gowan MJ, Carter AS. Socialeemotional screening status in earlychildhood predicts elementary school outcomes. Pediatrics 2008;121:957e62.

[58] Limperopoulos C, Bassan H, Sullivan NR, Soul JS, Robertson RL, Moore M,et al. Positive screening for autism in ex-preterm infants: prevalence andrisk factors. Pediatrics 2008;121:758e65.

[59] Treyvaud K, Doyle LW, Lee KJ, Roberts G, Lim J, Inder TE, et al. Socialeemotional difficulties in very preterm and term 2 year olds predict specificsocialeemotional problems at the age of 5 years. J Pediatr Psychol2012;37:779e85.

[60] Treyvaud K, Ure A, Doyle LW, Lee KJ, Rogers CE, Kidokoro H, et al. Psy-chiatric outcomes at age seven for very preterm children: rates and pre-dictors. J Child Psychol Psychiatry Allied Discipl 2013;54:772e9.

[61] Farooqi A, Hagglof B, Sedin G, Gothefors L, Serenius F. Mental health andsocial competencies of 10- to 12-year-old children born at 23 to 25 weeksof gestation in the 1990s: a Swedish national prospective follow-up study.Pediatrics 2007;120:118e33.

[62] Samara M, Marlow N, Wolke D. Pervasive behavior problems at 6 years ofage in a total-population sample of children born at �25 weeks ofgestation. Pediatrics 2008;122:562e73.

[63] Grunau R, Whitfield MF, Fay TB. Psychosocial and academic characteristicsof extremely low birth weight (<800g) adolescents who are free of majorimpairment compared with term-born control subjects. Pediatrics2004;114:e725e32.

[64] Hutchinson EA, De Luca CR, Doyle LW, Roberts G, Anderson PJ. VictorianInfant Collaborative Study G. School-age outcomes of extremely pretermor extremely low birth weight children. Pediatrics 2013;131:e1053e61.

[65] Moore T, Hennessy EM, Myles J, Johnson S, Draper ES, Costeloe KL, et al.Neurological and developmental outcome in extremely preterm childrenborn in England in 1995 and 2006: the EPICure studies. BMJ 2012;345:e7961.

[66] Taylor HG, Klein N, Minich N, Hack M. Middle school-age outcomes inchildren with very low birth-weight. Child Dev 2000;71:1495e511.

*[67] Hille ETM, den Ouden AL, Saigal S, Wolke D, Lambert M, Whitaker A, et al.Behavioural problems in childrenwho weigh 1000 g or less at birth in fourcountries. Lancet 2001;357:1641e3.

*[68] Burnett AC, Anderson PJ, Cheong J, Doyle LW, Davey CG, Wood SJ. Prev-alence of psychiatric diagnoses in preterm and full-term children, ado-lescents and young adults: a meta-analysis. Psychol Med 2011;41:2463e74.

[69] Bhutta A, Cleves M, Casey P, Cradock M, Anand K. Cognitive and behavioraloutcomes of school-aged children who were born preterm. A meta-analysis. J Am Med Assoc 2002;288:728e37.

S. Johnson, N. Marlow / Seminars in Fetal & Neonatal Medicine 19 (2014) 97e104104

[70] Linnet KM, Wisborg K, Agerbo E, Secher NJ, Thomsen PH, Henriksen TB.Gestational age, birth weight, and the risk of hyperkinetic disorder. ArchDis Child 2006;91:655e60.

*[71] Johnson S, Wolke D. Behavioural outcomes and psychopathology duringadolescence. Early Hum Dev 2013;89:199e207.

[72] Hack M, Flannery DJ, Schluchter M, Cartar L, Borawski E, Klein N. Out-comes in young adulthood for very-low-birth-weight infants. N Engl JMed 2002;346:149e57.

[73] Shum D, Neulinger K, O’Callaghan M, Mohay H. Attentional problems inchildren born with very preterm or with extremely low birth weight at 7e9 years. Arch Clin Neuropsychol 2008;23:103e12.

[74] Martel MM, von Eye A, Nigg JT. Revisiting the latent structure of ADHD: isthere a ‘g’ factor? J Child Psychol Psychiatry 2010;51:905e14.

[75] Mulder H, Pitchford NJ, Hagger MS, Marlow N. Development of executivefunction and attention in preterm children: a systematic review. DevNeuropsychol 2009;34:393e421.

[76] Aarnoudse-Moens CS, Duivenvoorden HJ, Weisglas-Kuperus N, VanGoudoever JB, Oosterlaan J. The profile of executive function in verypreterm children at 4 to 12 years. Dev Med Child Neurol 2012;54:247e53.

[77] Nadeau L, Boivin M, Tessier R, Lefebvre F, Robaey P. Mediators of behav-ioral problems in 7-year-old children born after 24 to 28 weeks ofgestation. J Dev Behav Pediatr 2001;22:1e10.

[78] Hack M, Youngstrom EA, Cartar L, Schluchter M, Taylor HG, Flannery D,et al. Behavioral outcomes and evidence of psychopathology among verylow birth weight infants at age 20 years. Pediatrics 2004;114:932e40.

[79] Skranes J, Vangberg TR, Kulseng S, Indredavik MS, Evensen KA,Martinussen M, et al. Clinical findings and white matter abnormalitiesseen on diffusion tensor imaging in adolescents with very low birthweight. Brain 2007;130:654e66.

[80] American Psychiatric Association. Diagnostic and statistical manual ofmental disorders (DSM-V). 5th ed. Washington DC: APA; 2013.

[81] Moster D, Lie RT, Markestad T. Long-term medical and social conse-quences of preterm birth. N Engl J Med 2008;359:262e73.

[82] Halmoy A, Klungsoyr K, Skjaerven R, Haavik J. Pre- and perinatal riskfactors in adults with attention-deficit/hyperactivity disorder. Biol Psy-chiatry 2012;71:474e81.

*[83] Nosarti C, Reichenberg A, Murray RM, Cnattingius S, Lambe MP, Yin L, et al.Preterm birth and psychiatric disorders in young adult life. Arch GenPsychiatry 2012;69:E1e8.

[84] Lindstrom K, Lindblad F, Hjern A. Psychiatric morbidity in adolescents andyoung adults born preterm: a Swedish national cohort study. Pediatrics2009;123:e46e53.

[85] Crump C, Winkleby MA, Sundquist K, Sundquist J. Preterm birth andpsychiatric medication prescription in young adulthood: a Swedish na-tional cohort study. Int J Epidemiol 2010;39:1522e30.

*[86] Saigal S, Stoskopf B, Streiner DL, Boyle M, Pinelli J, Paneth N, et al. Tran-sition of extremely low-birth weight infants from adolescence to young

adulthood. Comparison with normal birth-weight controls. JAMA2006;295:667e75.

[87] Hallin AL, Stjernqvist K. Follow-up of adolescents born extremely pre-term: self-perceived mental health, social and relational outcomes. ActaPaediatr 2011;100:279e83.

[88] Cooke RW. Health, lifestyle, and quality of life for young adults born verypreterm. Archs Dis Childh 2004;89:201e6.

*[89] Burnett A, Davey CG, Wood SJ, Wilson-Ching M, Molloy C, Cheong JL, et al.Extremely preterm birth and adolescent mental health in a geographicalcohort born in the 1990s. Psychol Med 2013 Aug;28:1e12 [Epub ahead ofprint].

[90] Volpe JJ. Brain injury in premature infants: a complex amalgam ofdestructive and developmental disturbances. Lancet Neurol 2009;8:110e24.

[91] Ment LR, Hirtz D, Huppi PS. Imaging biomarkers of outcome in thedeveloping preterm brain. Lancet Neurol 2009;8:1042e55.

[92] Skranes J, Lohaugen GC, Martinussen M, Haberg A, Brubakk AM, Dale AM.Cortical surface area and IQ in very-low-birth-weight (VLBW) youngadults. Cortex 2013;49:2264e71.

[93] de Kieviet JF, Zoetebier L, van Elburg RM, Vermeulen RJ, Oosterlaan J. Braindevelopment of very preterm and very low-birthweight children inchildhood and adolescence: a meta-analysis. Dev Med Child Neurol2012;54:313e23.

[94] Boardman JP, Counsell SJ, Rueckert D, Kapellou O, Bhatia KK, Aljabar P,et al. Abnormal deep grey matter development following preterm birthdetected using deformation-based morphometry. Neuroimage 2006;32:70e8.

[95] Northam GB, Liegeois F, Tournier JD, Croft LJ, Johns PN, Chong WK, et al.Interhemispheric temporal lobe connectivity predicts language impair-ment in adolescents born preterm. Brain 2012;135:3781e98.

[96] Smyser CD, Snyder AZ, Shimony JS, Blazey TM, Inder TE, Neil JJ. Effects ofwhite matter injury on resting state FMRI measures in prematurely borninfants. PLoS One 2013;8:e68098.

[97] Mulder H, Pitchford NJ, Marlow N. Processing speed mediates executivefunction difficulties in very preterm children in middle childhood. J IntNeuropsychol Soc 2011 Mar;28:1e10 [Epub ahead of print].

[98] Loe IM, Lee ES, Feldman HM. Attention and internalizing behaviors inrelation to white matter in children born preterm. J Dev Behav Pediatr2013;34:156e64.

[99] Rogers CE, Anderson PJ, Thompson DK, Kidokoro H, Wallendorf M,Treyvaud K, et al. Regional cerebral development at term relates to school-age socialeemotional development in very preterm children. J Am AcadChild Adolesc Psychiatry 2012;51:181e91.

[100] Parker J, Mitchell A, Kalpakidou A, Walshe M, Jung HY, Nosarti C, et al.Cerebellar growth and behavioural and neuropsychological outcome inpreterm adolescents. Brain 2008;131:1344e51.

lable at ScienceDirect

Seminars in Fetal & Neonatal Medicine 19 (2014) 105e111

Contents lists avai

Seminars in Fetal & Neonatal Medicine

journal homepage: www.elsevier .com/locate/s iny

Review

Respiratory outcomes for the tiniest or most immature infants

Anne-Marie Gibson a, Lex W. Doyle b,c,*

aRespiratory Research Group, Murdoch Childrens Research Institute, Parkville, Victoria, AustraliabDepartment of Obstetrics and Gynaecology, The Royal Women’s Hospital, Parkville, Victoria, AustraliacClinical Sciences, Murdoch Childrens Research Institute, Parkville, Victoria, Australia

Keywords:Bronchopulmonary dysplasiaExtremely low birth weightExtremely prematureInfantRespiratory function tests

* Corresponding author. Address: Department ofThe Royal Women’s Hospital, 20 Flemington Rd, ParkTel.: þ61 3 8345 3716; fax þ61 3 8345 3702.

E-mail address: lwd@unimelb.edu.au (L.W. Doyle)

1744-165X/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.siny.2013.10.006

s u m m a r y

Extremely low birth weight (<1000 g birth weight) or extremely preterm (<28 weeks of gestation)infants are surviving in greater numbers as neonatal care advances. Many of these survivors, especiallythose who develop bronchopulmonary dysplasia, have more respiratory ill health in the first years afterdischarge home, reduced respiratory function and impaired exercise capacity throughout childhood andinto adulthood compared with term-born controls. It is important to establish the long-term respiratoryoutcomes for the tiniest or most immature survivors as they grow older, since they may contributedisproportionately to rates of chronic obstructive pulmonary disease and respiratory ill-health inadulthood.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Perinatal care for the tiniest and most immature infantsadvanced rapidly towards the end of the twentieth century,including treatments such as antenatal corticosteroids from the1970s to accelerate fetal lung maturation [1], and exogenous sur-factant after birth from the late 1980s and early 1990s to reducerespiratory distress [2,3], combined with an increased willingnessto offer care [4]. Consequently more extremely low birth weight(ELBW; <1000 g at birth) [5] or extremely preterm (EP; <28 weeksof gestation at birth) [6] infants survived, despite many requiringsubstantial respiratory support early in the newborn period. Somesurvivors developed bronchopulmonary dysplasia (BPD), a diseasecharacterised by lung injury and arrested alveolar development [7].Given their early breathing problems, ELBW/EP survivors may bemore prone to respiratory ill-health later in life.

The purpose of this article is to review the long-term respiratoryoutcomes for the tiniest or most immature survivors. We considerrespiratory symptoms and lung function data in ELBW/EP survivorscompared with term controls, and in ELBW/EP survivors who didand did not develop BPD in the newborn period. Not all studieshave included only ELBW or EP survivors, so we have reviewed anystudies of preterm or low birth weight children that have reportedlung function data later in childhood.

Obstetrics and Gynaecology,ville, Victoria 3052, Australia.

.

All rights reserved.

2. Normal lung growth and function with age

Lung volume and function increase in healthy children, reachinga maximum in late adolescence/early twenties, and then begin todecline steadily with age (Fig. 1). Despite these changes, in theabsence of lung disease, the respiratory system remains capable ofmaintaining adequate gas exchange for the entire lifetime [8].However, early lung injury or maldevelopment may mean thatpeak lung growth is reduced, and hence respiratory symptomsmight appear earlier in life, even in the absence of other lung dis-eases. Furthermore, toxic agents that cause lung disease, such ascigarette smoking, may lead to an earlier decline in lung function inELBW/EP survivors.

3. Bronchopulmonary dysplasia e a changing picture

Despite the advances in the intensive care provided to ELBW/EPinfants in the modern era, the fact remains that a small proportionwill develop respiratory failure, and many will go on to developBPD with an continuing requirement for oxygen at and beyond 36weeks of corrected gestational age. Some may even require sup-plementary oxygen at home [9]. Prolonged periods of exposure tooxygen, even in low concentrations, and to mechanical ventilationcan lead to inflammatory and other biochemical and histologicalchanges to the lungs at a time when they should not be exposed tosuch insults [10]. Today the lungs of infants with BPD tend to haveless fibrosis and more uniform inflation than in the past [11e13].However, they have simplified gas exchange structures with fewer,larger alveoli, indicating disruption with alveolarisation in thedeveloping lung [11e13]. Severe cases of BPD are also associated

Fig. 1. Normal lung development and decline in lung function compared with abnormal or impaired development that may occur after mechanical ventilation and oxygen therapy,and development of bronchopulmonary dysplasia (BPD). (Adapted from Tager et al.8).

A.-M. Gibson, L.W. Doyle / Seminars in Fetal & Neonatal Medicine 19 (2014) 105e111106

with pulmonary hypertension and abnormal pulmonary vasculardevelopment [14]. The development of BPD is now widely viewedas a consequence of lung inflammation, potentially arising from theexposure to intrauterine infection, or to mechanical ventilation andsupplemental oxygen after birth [11,15].

In 1967, Northway and colleagues first reported the lung dam-age that occurred in infants with respiratory failure who hadreceived mechanical ventilation [16]. They defined the impaireddevelopment that occurred after exposure of immature andvulnerable lungs to mechanical ventilation as BPD [16]. Therepeated opening and closing of alveoli at higher pressures andvolumes, alongside high concentrations of supplemental oxygen,resulted in cystic and fibrotic changes within the lungs of these tiny,immature infants [16,17]. As neonatal care has advanced since the1960s, there have been improvements in the types and adminis-tration of assisted ventilation, including gentler ventilation tech-niques, antenatal corticosteroids and exogenous surfactant therapythat have led to changes in the epidemiology and definition of BPD.‘Old BPD’ presented predominantly in the pre-surfactant era, andwas characterised by alveolar septal fibrosis and inflammation,whereas ‘new BPD’, observed mainly in the post-surfactant era, ischaracterised by impaired alveolar growth, reflecting increasingprematurity of infants surviving with extremely low birth weightswho have greater exposure to mechanical ventilation and supple-mental oxygen [7,18].

4. Respiratory health problems

Children born very tiny or immature havemore upper and lowerrespiratory illnesses than term-born children over the first fewyears of life [19,20], and more so in those who had BPD [20e23].Asthma or recurrent wheezing are more prevalent later in life inthose born very tiny or immature than in those not born very tiny orimmature in some [24e26] but not all studies [27,28]. Those whohad BPD sometimes have higher rates of asthma than thosewithoutBPD [29].

Children born very tiny or immature are more likely than term-born controls to require readmission to hospital for respiratoryillnesses over the first few years of life, and rates of hospital

readmission have risen as survival rates have increased over time[20]. Overall, respiratory illnesses are the commonest cause ofrehospitalisation in the first few years [30,31], and occur morefrequently in preterm survivors who had BPD in the newbornperiod [30]. However, as the rate of hospital readmission declineslater in childhood, those who had BPD are no more likely to bereadmitted to hospital, for respiratory or other reasons [28].

5. Respiratory function in preterm survivors

Respiratory function testing in common clinical practice pri-marily assesses airflow through spirometry, and lung volumes.Other tests can include diffusing capacity of the lungs for carbonmonoxide, ventilation efficiency of the lung, or cardiopulmonaryexercise testing, which provide additional details of lungpathophysiology.

Spirometry is generally the first clinical option because it iseasily performed by most patients, with appropriate coaching by arespiratory technologist; however, childrenwith amental age of<5years are less likely to produce reliable results [32]. A seated patientinhales maximally from tidal breathing to total lung capacity (TLC),and then rapidly exhales until no further volume is exhaled, whichleaves a residual volume (RV) in the lung (Fig. 2) [32]. The initialhigh flow of gas emanates from the large airways, whereas the laterflow of gas comes more from the smaller airways. If performedforcefully, a forced vital capacity (FVC) is generated. Themost usefulclinical value generated by a spirogram is the forced expired vol-ume in 1 s (FEV1) because it is easy to measure, and it is repro-ducible and sufficiently sensitive in detecting airway obstruction,the commonest lung function abnormality seen in children. Othermeasures of flow from the spirogram include the forced expiratoryflow between 25% and 75% vital capacity (FEF25e75%), which is ob-tained from the slope of the spirogramwhen between 25% and 75%of the FVC has been expired; abnormalities in the FEF25e75% reflectsmall airways disease. Lung volumes (TLC and RV) may bemeasured by body plethysmography, and the functional residualcapacity (FRC) of the lung calculated. In asthma and other forms oflung disease, the RV:TLC ratio and the FRC typically rise, reflectinggas trapping within the lungs.

Fig. 2. Spirogram, with volume on the vertical axis and time on the horizontal axis. FVC, forced vital capacity; FEV1, forced expired volume in 1 s; FEF25e75, forced expiratory flowbetween 25% and 75% vital capacity.

A.-M. Gibson, L.W. Doyle / Seminars in Fetal & Neonatal Medicine 19 (2014) 105e111 107

There is a paucity of data relating to lung function outcome inpreschool-aged ELBW/EP survivors. From the limited data availablefor ELBW/EP preschoolers, those with BPD have persistent re-ductions in airflow and higher airways resistance compared withthose who did not have BPD, and compared with control subjects[25,33,34].

There are many more studies with values for FEV1 reported forpreterm survivors compared with controls from school age throughto early adulthood. The characteristics of some of these studies arelisted in Table 1. Some pre-date the use of surfactant, with cohortsborn before 1990, whereas other studies comprise births frommorerecent years. The ages at which subjects have been studied areyounger for cohorts that are more recent. There are also manyvariations in the birth weights or gestational ages of those studiede few studies have focused solely on ELBW/EP subjects. Some of thestudies range from regional cohorts, with the intent being to assessas many survivors as possible from a defined geographical region,which reduces selection biases and hence makes results morewidely applicable, through to highly selected cohort studies, fromwhich the importance and applicability of the results are less clear.The methods of the studies vary; in particular, results have beenreported either as percentage predicted or as Z-scores adjusted forage, height and sex. Regardless of how the initial results have beenreported, two groups can be compared by calculating standardisedmean differences and 95% confidence intervals between groups,provided that mean, standard deviation (SD) and sample size areavailable for each group, and the overall results can be synthesisedby meta-analysis (Figs. 3 and 4).

From studies of respiratory outcomes at school age through toearly adulthood, it is clear that ELBW/EP survivors have substantialreductions in airflow compared with controls. In Fig. 3 most indi-vidual studies have reported significant reductions in FEV1 in pre-term subjects compared with controls. The results are mostlyconsistent, regardless of the initial differences in the demographiccharacteristics and other features between the various studies,including whether the cohorts were born before or after surfactantwas available. In the pooled analysis, the reduction in standardisedmean difference for the FEV1 in the preterm groups compared withcontrols averaged almost 0.8 SD, regardless of how the initial studyreported the results, and regardless of other characteristics of thestudies, including whether the study was from the presurfactant orthe surfactant era. In the one regional study that focused on sur-vivors born at <26 weeks of gestation, the reduction in FEV1 wasthe largest, at 1.26 SD lower in the preterm group compared with

controls [35]. Therewere no obvious trends with timewithin eithersubgroup.

Within preterm groups, those who had BPD in the newbornperiod had evenmore reductions in FEV1 comparedwith thosewhodid not have BPD in most individual studies, with the differencesapproaching 1 SD between groups overall (Fig. 4). The reductionswith BPD were larger in the presurfactant era than in the surfactantera. There were no obvious trends with time in either subgroup.

In addition to changes in the FEV1, there are studies that reportchanges in other variables consistent with airflow obstruction inthe smaller airways (reduced FEF25e75%) and air trapping (raisedRV/TLC) in ELBW/EP survivors, especially in those who had BPD[36,37]. In addition to overall reductions in airflow variables, ELBW/EP survivors have higher proportions with values in clinicallyimportant ranges; Fawke et al. [35] reported that 32% of the sub-jects born at <26 weeks of gestational age and 66% of those withBPD had results that were clinically important. Despite havingmoreabnormal lung function on average and higher proportions withvalues in clinically important ranges, most ELBW/EP survivors areasymptomatic.

There is no clear answer as to whether lung function impair-ments improve, worsen or track with time. Several of the studiesin Table 1 reported results of FEV1 relative to controls from thesame cohort as the subjects aged. From the Royal Women’s Hos-pital cohort of infants <1501 g birth weight who were studied at14 and 18 years [36], and in two different regional cohorts fromwestern Norway, one studied at both 17 and 25 years, and theother at both 10 and 17 years, there was little change over timerelative to controls [38].

In addition to airway obstruction, there is some evidence thatthe airways of preterm subjects are more reactive than controls, asdemonstrated by increased airway provocation by methacholineand histamine challenges and increased bronchodilator response inpreterm subjects comparedwith controls, particularly in thosewhohad BPD [24,35,39,40]. The increased airway reactivity may beassociated with some of the airflow impairment seen in the pre-term cohorts comparedwith controls. Diffusion impairment and airtrapping within the lungs are other abnormalities that occur morefrequently in preterm survivors, and persist into early adulthood.Northway et al. described increased airways resistance, increasedair trapping and reduced ventilation efficiency in a preterm cohortwith BPD at 18 years of age [41]. Vrijlandt et al. demonstratedpersistent airflow obstruction, air trapping and reduced lungdiffusion in 19-year-old preterm survivors [42].

Table 1Features of studies reporting forced expired volume in 1 s (FEV1) of preterm children compared with controls, or within preterm groups comparing those with and withoutBPD, from school age onwards.

Study Years of birth Origin Birth weight or gestational age Age studied (years) No. preterm No. of controls

Presurfactant eraNorthway et al. [41] 1964e73 Single hospital 1936 (756) g; 33.8 (3.7) weeks 18 52 53Kitchen et al. [54] 1977e82 Single hospital <1501 g 8 170 0Doyle et al. [28] 1977e82 Single hospital <1501 g 14 169 39Doyle [37] 1977e82 Single hospital <1501 g 18e22 147 37Hakulinen et al. [55] 1978e85 Single hospital <1250 g 7e11 31 20Narang et al. [40] 1979e80 Single hospital <2000 g and <37 weeks 19e25 58 48Anand et al. [56] 1980e81 Regional <1500 g 15 128 128Kennedy et al. [57] 1981e82 Single hospital <1501 g 11 102 82Halvorsen et al. [24] 1982e85 Regional <1001 g or <29 weeks 17.7 46 46Vollsaeter et al. [38], a 1982e85 Regional <1001 g or <29 weeks 24.9 40 43Vrijlandt et al. [42] 1983 Regional <1500 g or <32 weeks 19 42 48Pelkonen et al. [58] 1983e87 Single hospital;

RCT surfactant<30 weeks 7e12 40 20

Kilbride et al. [43] 1983e89 Single hospital <801 g 9e15 50 25McLeod et al. [59] ,b 1984 Regional <1500 g 8e9 300 590Gross et al. [60] 1985e86 Single hospital 24e31 weeks 7 96 108Siltanen et al. [26] 1987e88 Single hospital <1501 g 10 50 54Malmberg et al. [61] 1989e91 Single hospital <1500 or <30 weeks 8 49 18Surfactant eraKorhonen et al. [62] 1990e94 Selected <1500 g 7e8 68 34Doyle [37] 1991e92 Regional <1000 g or <28 weeks 8e9 240 208Vollsaeter et al. [38] 1991e92 Regional <1001 g or <29 weeks 10.5 35 35Vollsaeter et al. [38], a 1991e92 Regional <1001 g or <29 weeks 17.8 29 31Danks et al. [63] 1992e94 Single hospital <1000 g 11e13 48 55Smith et al. [44] 1992e94 Regional <1000 g or <32 weeks 10.1 126 34Fawke et al. [35] 1995 Regional <26 weeks 11 182 161Hacking et al. [64] 1997 Regional <1000 g or <28 weeks 8e9 150 149

BPD, bronchopulmonary dysplasia; RCT, randomised controlled trial.a Same data as Halvorsen et al., [24] and hence only Vollsaeter et al. [38] included in the figures.b No data on FEV1 reported alone, and hence not included in the figures.

Standardised Mean Difference

-1 0 1-2Favours control Favours preterm

Presurfactant era birthsNorthway 1964-73; 18 yearsDoyle 1977-82; 14 yearsDoyle 1977-82; 18 yearsNarang 1979-80; 21 yearsAnand 1980-81; 15 yearsKennedy 1981-82; 11 yearsVollsaeter 1982-85; 17 yearsVollsaeter 1982-85; 25 yearsKilbride 1983-89; 11 yearsVrijlandt 1983; 19 yearsGross 1983-86; 7 yearsSiltanen 1987-88; 10 yearsMalmberg 1989-91; 8 years

Surfactant era birthsKorhonen 1990-94; 7 yearsDoyle 1991-92; 8 yearsVollsaeter 1991-92; 10 yearsVollsaeter 1991-92; 17 yearsDanks 1992-94; 12 yearsFawke 1995; 11 yearsDoyle 1997; 8 years

Subtotal -0.73; 95% CI -0.97, -0.50

Subtotal -0.87; 95% CI -1.07, -0.66

Total -0.77; 95% CI -0.94, -0.60

Fig. 3. Standardized mean differences (SMDs), with 95% confidence intervals (CIs), in forced expired volume in 1 s (FEV1) between preterm subjects compared with controls. Studiesdivided into presurfactant and surfactant eras, and then ordered chronologically relative to the years of birth of the subjects. Individual studies with first-named author, years ofbirth, and ages of subjects when tested are shown. SMDs (central box) and 95% CIs (line through box) are shown for individual studies, then subgroups (subtotal), then for all studiescombined (total).

A.-M. Gibson, L.W. Doyle / Seminars in Fetal & Neonatal Medicine 19 (2014) 105e111108

Standardised Mean Difference

-1 0 1-2Favours no BPD Favours BPD

Presurfactant era birthsNorthway 1964-73; 18 yearsKitchen 1977-82; 8 yearsDoyle 1977-82; 14 yearsDoyle 1977-82; 18 yearsHakulinen 1978-85; 7-11 yearsKennedy 1981-82; 11 yearsJacob 1981-87; 11 yearsVollsaeter 1982-85; 17 yearsVollsaeter 1982-85; 25 yearsKilbride 1983-89; 11 yearsGross 1983-86; 7 yearsMalmberg 1989-91; 8 years

Surfactant era birthsKorhonen 1990-94; 7 yearsDoyle 1991-92; 8 yearsVollsaeter 1991-92; 10 yearsVollsaeter 1991-92; 17 yearsFawke 1995; 11 yearsDoyle 1997; 8 years

Subtotal -0.98; 95% CI -1.25, -0.72

Subtotal -0.58 95% CI-0.73, -0.42

Total -0.83; 95% CI -1.01, -0.64

Fig. 4. Standardized mean differences (SMDs), with 95% confidence intervals (CIs), in forced expired volume in 1 s (FEV1) between preterm subjects who had bronchopulmonarydysplasia (BPD) compared with those who did not have BPD. Studies are divided into presurfactant and surfactant eras, and are then ordered chronologically relative to the years ofbirth of the subjects. Individual studies, years of birth, and ages of subjects when tested are shown. SMDs (central box) and 95% CIs (line through box) shown for individual studies,then subgroups (subtotal), then for all studies combined (total).

Practice points

� The tiniest and most immature survivors present morefrequently to the physician with respiratory ill health inthe first few years after discharge home.

� Some children and young adults who were tiny or veryimmature at birth have reduced respiratory function anddiminished exercise capacity compared with term-bornchildren.

� Despite their reduced exercise and respiratory function,most tiny or immature survivors are asymptomatic.

� Those who had BPD have more respiratory ill-health andreduced lung function than those who did not have BPD.

A.-M. Gibson, L.W. Doyle / Seminars in Fetal & Neonatal Medicine 19 (2014) 105e111 109

6. Exercise capacity

There are several studies in which reduced exercise capacity,as reflected in maximal oxygen capacity, was reported in ex-preterm survivors compared with controls [43,44]. Kilbrideet al. [43] reported a reduced mean maximal oxygen consump-tion of 7.3 (95% CI: 4.4, 10.2) ml/kg/min in survivors <801 g birthweight at a mean age of 11 years compared with controls. Smithet al. [44] found a reduced mean maximal oxygen consumptionof 3.9 (95% CI: 2.6, 5.2) ml/kg/min in survivors <1000 g birthweight and <32 weeks of gestation at a mean age of 10 yearscompared with controls. In the study of survivors <26 weeks ofgestation and controls at a mean age of 11 years of age, Welshet al. [45] reported a mean reduction of peak oxygen consump-tion of 253 (95% CI: 147, 359) ml/min in the EP group, afteradjustment for body weight. However, not all studies have re-ported significant reductions in oxygen consumption in ELBW/EPcompared with term subjects [42,46].

7. Smoking exposure

Maternal smoking during pregnancy is associated withabnormal lung function in childhood [47e49], and with LBW,which in turn is associated with later respiratory morbidity[50,51]. After birth, passive smoking has been associated withreduced airflow and air trapping in a cohort of 120 VLBW sur-vivors at 11 years of age [52]. Active smoking in teenagers hasalso been associated with reduced airflow in a cohort of 44 ELBWsurvivors at a mean age of 20 years [53]. As ELBW/EP survivorswill probably not reach their expected peak of lung function by20e25 years of age, it is probable that both passive and activesmoking will lead to an earlier presentation of smoking-relateddiseases in their adult life than would occur in normal birthweight or term controls.

8. Conclusions

Over the past 40 years there have been important advances inneonatal care that have markedly improved survival of ELBW/EPinfants. Despite the advances in care, BPD remains a major cause ofrespiratory morbidity and mortality. The rates of BPD, the mostcommon chronic respiratory condition affecting preterm infants,increased alongside the improved survival of these smaller, lessmature infants.

ELBW/EP survivors have more respiratory ill-health and aremore likely to require rehospitalisation for respiratory illnessesthan their term-born counterparts in the first few years afterdischarge. However, respiratory symptoms decrease with time,although they can persist into adulthood in some preterm survi-vors. In terms of pulmonary function, abnormal lung functionpersists in ELBW/EP survivors comparedwith controls, especially invariables related to airflow and air trapping. Cardiopulmonary ex-ercise capacity has also been shown to be abnormal in some pre-term survivors.

Research directions

� Respiratory function of the tiniest and most immaturesurvivors, particularly those who had BPD, must bedetermined later into adult life, at ages when normalrespiratory function is in decline.

� The long-term respiratory effects of more gentle resusci-tation and respiratory support techniques used today,and of exposure to cigarette smoke, must also be deter-mined well into adulthood.

A.-M. Gibson, L.W. Doyle / Seminars in Fetal & Neonatal Medicine 19 (2014) 105e111110

Conflict of interest statement

None declared.

Funding sources

None.

References

[1] Roberts D, Dalziel S. Antenatal corticosteroids for accelerating fetal lungmaturation for women at risk of preterm birth. Cochrane Database Syst Rev2006;(3):CD004454.

[2] Soll RF. Prophylactic natural surfactant extract for preventing morbidity andmortality in preterm infants. Cochrane Database Syst Rev 2000;(2):CD000511.

[3] Soll RF. Natural surfactant extract versus synthetic surfactant for neonatalrespiratory distress syndrome. Cochrane Database Syst Rev 2000;(2):CD000144.

[4] Gultom E, Doyle LW, Davis P, Dharmalingam A, Bowman E. Changes over timein attitudes to treatment and survival rates for extremely preterm infants(23e27 weeks’ gestational age). Aust NZ J Obstet Gynaecol 1997;37:56e8.

[5] Doyle LW, Roberts G, Anderson PJ. Changing long-term outcomes for infants500e999 g birth weight in Victoria, 1979e2005. Arch Dis Child FetalNeonatal Ed 2011;96:F443e7.

[6] Doyle LW, Roberts G, Anderson PJ. Outcomes at age 2 years of infants <28weeks’ gestational age born in Victoria in 2005. J Pediatr 2010;156:49e53.e1.

[7] Jobe AH, Bancalari E. Bronchopulmonary dysplasia. Am J Respir Crit Care Med2001;163:1723e9.

[8] Tager IB, Weiss ST, Rosner B, Speizer FE. Effect of parental cigarette smokingon the pulmonary function of children. Am J Epidemiol 1979;110:15e26.

[9] Greenough A, Dimitriou G, Bhat RY, et al. Lung volumes in infants whohad mild to moderate bronchopulmonary dysplasia. Eur J Pediatr2005;164:583e6.

[10] Hayes Jr D, Meadows Jr JT, Murphy BS, Feola DJ, Shook LA, Ballard HO. Pul-monary function outcomes in bronchopulmonary dysplasia through child-hood and into adulthood: implications for primary care. Prim Care Respir J2011;20:128e33.

[11] Westover AJ, Moss TJ. Effects of intrauterine infection or inflammation onfetal lung development. Clin Exp Pharmacol Physiol 2012;39:824e30.

[12] Evensen KA, Steinshamn S, Tjonna AE, et al. Effects of preterm birth and fetalgrowth retardation on cardiovascular risk factors in young adulthood. EarlyHum Dev 2009;85:239e45.

[13] Wong PM, Lees AN, Louw J, et al. Emphysema in young adult survivors ofmoderate-to-severe bronchopulmonary dysplasia. Eur Respir J 2008;32:321e8.

[14] Abman SH. Pulmonary vascular disease and bronchopulmonary dysplasia:evaluation and treatment of pulmonary hypertension. NeoReviews 2011;12:e645e50.

[15] Baker CD, Accurso FJ. Preterm birth and airway inflammation in childhood.J Pediatr 2012;161:979e80.

[16] Northway Jr WH, Rosan RC, Porter DY. Pulmonary disease following respi-rator therapy of hyaline-membrane disease. Bronchopulmonary dysplasia.N Engl J Med 1967;276:357e68.

[17] Dargaville PA, Tingay DG. Lung protective ventilation in extremely preterminfants. J Paediatr Child Health 2012;48:740e6.

[18] Baraldi E, Carraro S, Filippone M. Bronchopulmonary dysplasia: definitionsand long-term respiratory outcome. Early Hum Dev 2009;85(10 Suppl):S1e3.

[19] O’Callaghan MJ, Burns Y, Gray P, et al. Extremely low birth weight and controlinfants at 2 years corrected age: a comparison of intellectual abilities, motorperformance, growth and health. Early Hum Dev 1995;40:115e28.

[20] Doyle LW, Ford G, Davis N. Health and hospitalisations after discharge inextremely low birth weight infants. Semin Neonatol 2003;8:137e45.

[21] Yu VY, Orgill AA, Lim SB, Bajuk B, Astbury J. Bronchopulmonary dysplasia invery low birthweight infants. Aust Paediatr J 1983;19:233e6.

[22] Tammela OK. First-year infections after initial hospitalization in low birthweight infants with and without bronchopulmonary dysplasia. Scand J InfectDis 1992;24:515e24.

[23] Korhonen P, Koivisto AM, Ikonen S, Laippala P, Tammela O. Very low birth-weight, bronchopulmonary dysplasia and health in early childhood. ActaPaediatr 1999;88:1385e91.

[24] Halvorsen T, Skadberg BT, Eide GE, Roksund OD, Carlsen KH, Bakke P. Pul-monary outcome in adolescents of extreme preterm birth: a regional cohortstudy. Acta Paediatr 2004;93:1294e300.

*[25] Vrijlandt EJ, Boezen HM, Gerritsen J, Stremmelaar EF, Duiverman EJ. Respi-ratory health in prematurely born preschool children with and withoutbronchopulmonary dysplasia. J Pediatr 2007;150:256e61.

[26] Siltanen M, Savilahti E, Pohjavuori M, Kajosaari M. Respiratory symptomsand lung function in relation to atopy in children born preterm. PediatrPulmonol 2004;37:43e9.

[27] Steffensen FH, Sorensen HT, Gillman MW, et al. Low birth weight and pre-term delivery as risk factors for asthma and atopic dermatitis in young adultmales. Epidemiology 2000;11:185e8.

*[28] Doyle LW, Cheung MM, Ford GW, Olinsky A, Davis NM, Callanan C. Birthweight <1501 g and respiratory health at age 14. Arch Dis Child 2001;84:40e4.

[29] Ng DK, Lau WY, Lee SL. Pulmonary sequelae in long-term survivors ofbronchopulmonary dysplasia. Pediatr Int 2000;42:603e7.

[30] Cunningham CK, McMillan JA, Gross SJ. Rehospitalization for respiratoryillness in infants of less than 32 weeks’ gestation. Pediatrics 1991;88:527e32.

[31] Kitchen WH, Ford GW, Doyle LW, Rickards AL, Kelly EA. Health and hospitalreadmissions of very-low-birth-weight and normal-birth-weight children.Am J Dis Child 1990;144:213e8.

[32] Miller MR, Hankinson J, Brusasco V, et al. Standardisation of spirometry. EurRespir J 2005;26:319e38.

[33] Talmaciu I, Ren CL, Kolb SM, Hickey E, Panitch HB. Pulmonary function intechnology-dependent children 2 years and older with bronchopulmonarydysplasia. Pediatr Pulmonol 2002;33:181e8.

[34] Kairamkonda VR, Richardson J, Subhedar N, Bridge PD, Shaw NJ. Lungfunction measurement in prematurely born preschool children with andwithout chronic lung disease. J Perinatol 2008;28:199e204.

*[35] Fawke J, Lum S, Kirkby J, et al. Lung function and respiratory symptoms at 11years in children born extremely preterm: the EPICure study. Am J Respir CritCare Med 2010;182:237e45.

*[36] Doyle LW, Faber B, Callanan C, Freezer N, Ford GW, Davis NM. Broncho-pulmonary dysplasia in very low birth weight subjects and lung function inlate adolescence. Pediatrics 2006;118:108e13.

*[37] Doyle LW. Respiratory function at age 8e9 years in extremely low birth-weight/very preterm children born in Victoria in 1991e1992. Pediatr Pul-monol 2006;41:570e6.

*[38] Vollsaeter M, Roksund OD, Eide GE, Markestad T, Halvorsen T. Lung functionafter preterm birth: development from mid-childhood to adulthood. Thorax2013;68:767e76.

[39] Pelkonen AS, Hakulinen AL, Turpeinen M. Bronchial lability and respon-siveness in school children born very preterm. Am J Respir Crit Care Med1997;156(4 Pt 1):1178e84.

[40] Narang I, Rosenthal M, Cremonesini D, Silverman M, Bush A. Longitudinalevaluation of airway function 21 years after preterm birth. Am J Respir CritCare Med 2008;178:74e80.

*[41] Northway Jr WH, Moss RB, Carlisle KB, et al. Late pulmonary sequelae ofbronchopulmonary dysplasia. N Engl J Med 1990;323:1793e9.

*[42] Vrijlandt EJ, Gerritsen J, Boezen HM, Grevink RG, Duiverman EJ. Lung func-tion and exercise capacity in young adults born prematurely. Am J Respir CritCare Med 2006;173:890e6.

[43] Kilbride HW, Gelatt MC, Sabath RJ. Pulmonary function and exercise capacityfor ELBW survivors in preadolescence: effect of neonatal chronic lung dis-ease. J Pediatr 2003;143:488e93.

[44] Smith LJ, van Asperen PP, McKay KO, Selvadurai H, Fitzgerald DA. Reduced ex-ercise capacity in children born very preterm. Pediatrics 2008;122:e287e93.

[45] Welsh L, Kirkby J, Lum S, et al. The EPICure study: maximal exercise andphysical activity in school children born extremely preterm. Thorax 2010;65:165e72.

[46] Clemm H, Roksund O, Thorsen E, Eide GE, Markestad T, Halvorsen T. Aerobiccapacity and exercise performance in young people born extremely preterm.Pediatrics 2012;129:e97e105.

[47] Le Souef PN. Pediatric origins of adult lung diseases. 4. Tobacco related lungdiseases begin in childhood. Thorax 2000;55:1063e7.

[48] Cunningham J, Docy DW, Speizer FE. Maternal smoking during pregnancy asa predictor of lung function in children. Am J Epidemiol 1994;139:1139e52.

[49] Janson C. The effect of passive smoking on respiratory health in children andadults. Int J Tuberc Lung Dis 2004;8:510e6.

[50] Li YF, Gilliland FD, Berhane K, et al. Effects of in utero and environmentaltobacco smoke exposure on lung function in boys and girls with and withoutasthma. Am J Respir Crit Care Med 2000;162:2097e104.

[51] Rizzi M, Sergi M, Andreoli A, Pecis M, Bruschi C, Fanfulla F. Environmentaltobacco smoke may induce early lung damage in healthy male adolescents.Chest 2004;125:1387e93.

[52] Doyle LW, Ford GW, Olinsky A, Knoches AM, Callanan C. Passive smoking andrespiratory function in very low birthweight children. Med J Aust 1996;164:266e9.

A.-M. Gibson, L.W. Doyle / Seminars in Fetal & Neonatal Medicine 19 (2014) 105e111 111

[53] Doyle LW, Olinsky A, Faber B, Callanan C. Adverse effects of smoking onrespiratory function in young adults born weighing less than 1000 grams.Pediatrics 2003;112(3 Pt 1):565e9.

[54] Kitchen WH, Olinsky A, Doyle LW, et al. Respiratory health and lung functionin 8-year-old children of very low birth weight: a cohort study. Pediatrics1992;89(6 Pt 2):1151e8.

[55] Hakulinen AL, Jarvenpaa AL, Turpeinen M, Sovijarvi A. Diffusing capacity ofthe lung in school-aged children born very preterm, with and withoutbronchopulmonary dysplasia. Pediatr Pulmonol 1996;21:353e60.

[56] Anand D, Stevenson CJ, West CR, Pharoah PO. Lung function and respi-ratory health in adolescents of very low birth weight. Archs Dis Child2003;88:135e8.

[57] Kennedy JD, Edward LJ, Bates DJ, et al. Effects of birthweight and oxygensupplementation on lung function in late childhood in children of very lowbirth weight. Pediatr Pulmonol 2000;30:32e40.

[58] Pelkonen AS, Hakulinen AL, Turpeinen M, Hallman M. Effect of neonatalsurfactant therapy on lung function at school age in children born verypreterm. Pediatr Pulmonol 1998;25:182e90.

[59] McLeod A, Ross P, Mitchell S, et al. Respiratory health in a total very lowbirthweight cohort and their classroom controls. Arch Dis Child 1996;74:188e94.

[60] Gross SJ, Iannuzzi DM, Kveselis DA, Anbar RD. Effect of preterm birth onpulmonary function at school age: a prospective controlled study. J Pediatr1998;133:188e92.

[61] Malmberg LP, Mieskonen S, Pelkonen A, Kari A, Sovijarvi AR, Turpeinen M.Lung function measured by the oscillometric method in prematurely bornchildren with chronic lung disease. Eur Respir J 2000;16:598e603.

[62] Korhonen P, Laitinen J, Hyodynmaa E, Tammela O. Respiratory outcome inschool-aged, very-low-birth-weight children in the surfactant era. ActaPaediatr 2004;93:316e21.

[63] Danks M, Burns YR, Gibbons K, et al. Fitness limitations in non-disabledextremely low birthweight adolescents. J Paediatr Child Health 2013;49:548e53.

*[64] Hacking DF, Gibson AM, Robertson C, Doyle LW. Respiratory function at age8e9 after extremely low birthweight or preterm birth in Victoria in 1997.Pediatric Pulmonol 2013;48:449e55.

lable at ScienceDirect

Seminars in Fetal & Neonatal Medicine 19 (2014) 112e117

Contents lists avai

Seminars in Fetal & Neonatal Medicine

journal homepage: www.elsevier .com/locate/s iny

Review

Is very preterm birth a risk factor for adult cardiometabolic disease?

Eero Kajantie a,b,c,*, Petteri Hovi a,b

aNational Institute for Health and Welfare, Diabetes Prevention Unit, Helsinki, FinlandbChildren’s Hospital, Helsinki University Central Hospital and University of Helsinki, Helsinki, FinlandcDepartment of Obstetrics and Gynaecology, Oulu University Central Hospital and University of Oulu, Oulu, Finland

Keywords:CardiovascularExerciseExtremely low birth weightPrematureProgrammingRisk factorVery low birth weight

* Corresponding author. Address: National InstituDiabetes Prevention Unit, Helsinki, Finland. Tel.: þ355248661.

E-mail address: eero.kajantie@helsinki.fi (E. Kajan

1744-165X/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.siny.2013.11.006

s u m m a r y

The first infants to experience modern pre- and neonatal care are now in their thirties, an age at whichthe incidence of cardiometabolic disease is low. However, data from cohorts born preterm prior to theintroduction of modern care suggest an increased risk of type 2 diabetes. For young adult cohorts offormer very small or very preterm infants, there is accumulating evidence of increased risk factors forlater cardiovascular disease, including higher blood pressure, lower lean body mass, impaired glucoseregulation, and perhaps a more atherogenic lipid profile. Regarding lifestyle, adults born very small orvery preterm undertake less non-conditioning physical activity and may have a lower intake of fruit andmilk products. Any intervention reducing risk factors, in particular blood pressure and low physicalactivity, would have a substantial potential to reduce the lifetime disease burden in small preterm in-fants. There are now enough data to warrant an expert evaluation of the level of evidence for car-diometabolic disease in individuals born very small or very preterm, which has possible public healthimplications.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Along with improvements in prenatal and neonatal care fromthe 1970s onwards, the long-term outcomes of the tiniest and mostimmature infants have received increasing attention. A number ofneonatal centres have established follow-up programs and cohortsof survivors born very small or very preterm who are now in theirtwenties or thirties. Much of the initial focus of these follow-upstudies was on outcomes in which the increased risks wereimmediately obvious, such as neurodevelopmental, respiratory andgrowth-related outcomes. Parallel to this, early life origins ofchronic adult disease have attracted considerable research interestin the public health domain. This interest originally arose fromobservations linking small body size at term birth with increasedrates of cardiovascular disease and type 2 diabetes in adult life.Together with experimental animal research, this led to theDevelopmental Origins of Health and Disease (DOHaD) theory,which suggests that conditions at specific sensitive developmentalperiods ‘program’ an individual’s cells, tissues and organs in waysthat lead to altered function throughout life to adjust to prevailing

te for Health and Welfare,8 29 5248610; fax: þ358 29

tie).

All rights reserved.

environmental conditions [1,2]. Altered function, then, may appearas increased risk for chronic disease. Although birth weight canonly be thought of as a surrogate of the early environment, it doespredict chronic disease, shown in cohorts born mostly at termwitha birth weight distribution similar to that in the general population.Regarding the lower extremes of birth weight, it is natural to as-sume that the highly abnormal environment small preterm infantsexperience in neonatal intensive care, and some also in utero,would predict altered risk for disease. Recent research bringingtogether the preterm follow-up and DOHaD fields has shown thatthis indeed is true; risk factors for cardiometabolic disease arehigher in subjects born very preterm.

Our aim is to review current knowledge on risk factors for keycardiometabolic diseases in people who were born very small orvery preterm. Those diseases and conditions include cardiovasculardisease (e.g. coronary heart disease and stroke), type 2 diabetes andthe metabolic syndrome. Bone health is reviewed elsewhere in thisissue (Roberts and Cheong, Chapter 8). We begin by discussing thescanty evidence on hard disease outcomes; our focus is then on riskfactors because these diseases manifest only late in life. Our style isnarrative: systematic accounts exist in recent meta-analyses onsystolic blood pressure [3] and components of the metabolic syn-drome [4,5] and systematic reviews on the overall adult outcomeafter preterm birth [6,7]. We discuss possible mechanisms that mayunderlie the findings, their meaning in the prevention of disease,and future research directions. We focus on physiological and

E. Kajantie, P. Hovi / Seminars in Fetal & Neonatal Medicine 19 (2014) 112e117 113

lifestyle risk factors, as psychological and mental health outcomesare reviewed elsewhere in this issue (Anderson, Chapter 3; John-son, Chapter 5). However, it is of note that although some psy-chological characteristics may confer increased risks of disease,other characteristics such as the personality trait of conscien-tiousness [8] may actually be associated with a healthy lifestyle. Wefocus on adults born either very small or very preterm, which indifferent studies may be defined as very low birth weight (VLBW;<1500 g), extremely low birth weight (ELBW; <1000 g), very pre-term (VP; <32 weeks) or extremely preterm (EP; <28 weeks).Studies on adults born moderately preterm are scarce and remainoutside the scope of this review. However, the few publishedstudies suggest similar but moderately higher cardiometabolic riskfactors among these adults [9e15]. As moderately preterm birth ismuch more common than very preterm birth, even weaker risksmay be equally important on a population level.

2. Preterm birth and adult cardiometabolic disease e is thereany hard evidence?

The first generations of very small or very preterm infants thathave benefited from modern perinatal intensive care have not yetreached old age and there are so far few published data on ‘hard’cardiometabolic disease outcomes. Circumstantial evidence isprovided by a number of studies in older adults. Most of thesestudies focus on type 2 diabetes (Table 1). Although each study usesa different method to identify subjects with diabetes, the increasedrisk of diabetes in those born preterm is consistently seen and in-dependent of the association between slow fetal growth and dia-betes [9e12]. A specific risk for VP infants is seen in a remarkableSwedish cohort that includes 986 subjects born before 32 weeks ofgestation between 1925 and 1949 [9]. Interestingly, a large registerstudy, also based on Swedish data, suggests that, in addition to type2 diabetes, the risk of type 1 diabetes at age 25e37 years is alsohigher [12].

By contrast, evidence on coronary heart disease and stroke e

key manifestations of atherosclerosis e is inconclusive. We areaware of only two studies focusing on coronary heart diseaseamong those born preterm. Neither showed an increased risk,suggesting that the well-replicated association between low birthweight and coronary heart disease could solely be attributed toslow fetal growth [14,16]. As to stroke, a Swedish study showed anincreased risk among those born before 35 weeks [14], in particularfor thrombotic stroke. This was not seen in the Finnish HelsinkiBirth Cohort Study [17]. The absence of association is surprising inthe light of increased risk factors for atherosclerosis. Possible ex-planations may include inaccuracies in the estimation of gesta-tional age by the last menstrual period, or survivor bias: when the

Table 1Studies on type 2 diabetes in adults born preterm.

Study Age (years) Diabetes indicated by: N

Lawlor et al. [11] 44e50 Self-report 57Kaijser et al. [9] 57e81 Hospital diagnosis 64

Kajantie et al. [10] 58e68 Special reimbursement formedication, granted after age 40 years

12 7

Pilgaard et al. [13] 30e60 OGTT 46Crump et al. [12] 25e37 Purchase of any diabetes medication 630 0

OR, odds ratio; CI, confidence interval; N/A, not available; OGTT, oral glucose tolerance tAll models have people born at term as a reference group and (except for the study bystatus.

a Model also includes birth weight SD score and whether first born.b Covariates in the model not reported.c Model also includes mother’s age, marital status and diabetes.

subjects were born, only a minority of very small or very preterminfants survived and theymay not be representative of infants bornat similar gestational ages today. However, should the result bereplicated, it would imply potentially important protective factorsthat would counteract the risk factors reviewed in detail below.

3. Body composition

As discussed by Roberts and Cheong in this issue (Chapter 8),adults born very preterm are shorter and have lower bone mineraldensity than their counterparts born at term. Lower body massindex is seen in some [18,19] but not all [20,21] studies. Relativelyfew adult studies have assessed body composition. The HelsinkiStudy of Very Low Birth Weight Adults (HeSVA) used dual X-rayabsorptiometry, and found that VLBW adults had lower lean bodymass than controls [18]. This is consistent with studies in adoles-cents [22] and is potentially important because low lean body massand excess fat mass are both strong independent predictors for all-cause mortality [23]. In HeSVA there was no difference in fat per-centage between VLBW subjects and controls, and an uncertaindifference in truncal:peripheral fat ratio [18], an indicator of centraladiposity which is considered metabolically more harmful. Bycontrast, increased central adiposity in VLBW adults was suggestedby higher subscapular:triceps skinfold ratio in the Trondheimcohort [24]. An important gap in knowledge is that no studies havespecifically assessed visceral fat.

4. Blood pressure

The association between VP birth and higher blood pressure inadult life has been extensively replicated [3,4,18,24e30]. A recentmeta-analysis, which included studies focusing on survivors bornsmall or preterm, with any degree of prematurity, concluded thatthe mean difference between adults born preterm and controls was4.2 mmHg for systolic and 2.6 mmHg for diastolic blood pressure[4]. Another meta-analysis reported a comparison of VLBW or VPadults with controls and showed a mean difference of 3.3 mmHg insystolic pressure [3]. This difference was present in both sexes,although it was stronger among women (systolic/diastolic 4.9/2.9 mmHg) than among men (2.0/1.3 mmHg) [4]. There is somedegree of heterogeneity between studies, a notable example beinga difference of 8.6/4.3 mmHg in a study in Melbourne, Australia[26]. These differences are considerable given that, at the popula-tion level, a 2 mmHg reduction in diastolic pressure is estimated toresult in a 6% reduction in the risk of coronary heart disease and a15% reduction in risk of cerebrovascular events [31].

It is likely that multiple mechanisms contribute to the higherblood pressure in people born VP. Prime candidates include current

No. preterm (%) No. diabetes (%) OR (95% CI) for diabetes in thoseborn preterm

93 N/A 113 (2.0%) 2.10 (1.23e3.61)25 2931 (45.6%) 508 (7.9%) 1.67 (1.33e2.11) for<32 weeks;

1.29 (1.05e1.58) for 33e36 weeks31 796 (6.3%) 652 (5.1%) 1.59 (1.00e2.58)a for<35 weeks;

0.67 (0.42e1.07)a for 35e36 weeks44 443 (9.5%) 223 (4.8%) 1.79 (1.17e2.74)b

90 27 953 (4.4%) 7751 (1.2%) 1.16 (1.04e1.28)c

est.Pilgaard et al. [13]) include sex, year/period of birth and childhood socio-economic

E. Kajantie, P. Hovi / Seminars in Fetal & Neonatal Medicine 19 (2014) 112e117114

lifestyle, discussed below in a separate section, and early pro-gramming of renal function, endothelial function, arterial stiffness,peripheral vascular resistance or sympathetic activity. Two studieshave assessed renal function in VLBW adults. In the Dutch POPScohort, smaller kidney volume and lower glomerular filtration rateand effective renal plasma flow were present in VP small forgestational age (SGA) subjects (but not VP appropriate for gesta-tional age subjects) compared with controls born at term [32]. Nodifference was found in a Swedish study in women, although itspower was limited [27]. The vulnerability of the kidney is furthersupported by a caseecontrol study that suggested increased risks offocal segmental glomerulosclerosis in people born at VLBW [33]. Asto vascular function, increased peripheral vascular resistance issuggested by an observation of abnormal retinal vascularization inVLBW adults [34]. However, no association was seen with endo-thelial function as measured by flow-mediated arterial dilatation inthe HeSVA [35] or in another study of 13e16-year-olds [36]. Aninteresting potential mechanism is arterial stiffness, which is anindependent predictor of cardiovascular events: it was elevated in11-year-olds born before 26 weeks [37].

As to sympathetic activity, VLBW adults have higher restingheart rate than their peers born at term [18], and higher diastolicblood pressure during the Trier Social Stress Test [38]. Consistentwith this, differences in blood pressure between VLBW adults andcontrols are stronger for clinic (often a mildly stressful situation)than for ambulatory blood pressures [26,30].

Preterm birth may also affect the development of the myocar-dium. A study using cardiac MRI compared adults born <1850 gwith controls and found increased mass and small internal di-ameters of both ventricles, and reduced systolic and diastolicfunction [39,40]. The differences were stronger for subjects born atearlier gestational ages and larger for the right ventricle, which alsohad an impaired ejection fraction [40].

5. Glucose tolerance

We showed in the HeSVA that young adults born at VLBW havehigher blood glucose and insulin concentrations after an oralglucose tolerance test (OGTT) than their peers born at term [18].The difference was not attributable to the VLBW adults’ lower leanbody mass. Together with previous findings in prepubertal childrenborn at �32 weeks, our preliminary findings by an intravenousglucose tolerance test suggested that the difference in glucosetolerance is due to reduced insulin sensitivity, which is compen-sated by increased but possibly suboptimal insulin release [41,42].Accordingly, reduced insulin sensitivity was also found in 57 youngadults born at VLBW/very low gestational age and 30 controls byhyperinsulinemiceeuglycemic clamp [43], and compensatory in-sulin release also by another study of 26 VLBW adults and 24controls who underwent a five-sample OGTT [44]. By contrast, onestudy showed no difference in intravenous glucose tolerance testbetween adults born preterm and controls born at term, but theterm controls were a heterogeneous group, some of whom wererecruited for being born SGA [45]. Taken together, these findingssuggest that adults born VLBW may be more liable to develop type2 diabetes in later life, which is consistent with observations inolder adults born less preterm [9e13].

6. Plasma lipids

A recent meta-analysis concluded that adults born preterm havelow-density lipoprotein (LDL) cholesterol concentrations0.15 mmol/l higher than controls [4]. This result was driven bypreviously unpublished data from one of the cohorts. Althoughmany of the individual studies in VLBW adults have reported no

differences in standard lipid measurements [18], higher LDL andtotal cholesterol were reported in adolescent boys but not girlsborn at <34 weeks of gestation (M. Sipola-Leppänen et al., un-published data). However, standard lipid measurements may notcapture all essential variation in atherogenicity. In a detailedanalysis of lipoprotein fractions we found that adults born at VLBWhave higher concentrations of triglycerides both in the largest verylow-density lipoprotein particles and in small high-density lipo-protein particles, both of which are potentially atherogenic [46].

7. Lifestyle

Healthy lifestyle is central in preventing cardiometabolic dis-ease. Although we are unaware of any lifestyle intervention studiesspecific to adults born preterm, there is little reason to expect thatthe benefits of lifestyle changes would not be similar to the generalpopulation with a similar lifestyle. There is, however, increasingevidence that lifestyle at baseline is different between adults bornvery preterm and their peers born at term. As individuals with ahigh-risk lifestyle have the highest potential to benefit from in-terventions, it seems probable that adults born very preterm couldpotentially benefit from specific lifestyle interventions, for examplefrom increasing physical activity. Many of these interventions canbe efficient from childhood onwards, and they could benefit thewhole family. In the following we review the current knowledge onphysical activity, diet, sleep, smoking and use of alcohol in adultsborn preterm.

8. Physical activity and fitness

Physical activity is effective in reducing the risks associated withhigh blood pressure and impaired glucose regulation. Accumu-lating observations, summarized in Table 2, suggest lower degreesof physical activity and physical self-confidence in adolescents andadults born VP [20,47e51]. The differences are substantial, with upto 68% lower energy expenditure in VLBW as compared with con-trol adults [49], and specific to conditioning physical activity, whichmeans actively undertaking exercise. A limitation of these findingsis that they have thus far not been confirmed by objective mea-surement. Objective measurement by accelerometry has been usedin one study in 11-year-olds born before 25 weeks, who, despitelower exercise capacity and self-perceived ability, had similar ac-tivity counts. However, the study was relatively small, and may notcapture differences specific to conditioning, health-enhancingphysical activity [52].

Studies on fitness in adolescents or adults are limited [20,50](Table 2) and suggest substantially lower cardiorespiratory fitnessand muscle strength, at least among those born ELBW.

Reasons for these substantial differences are poorly understood.They may be a result of a vicious cycle starting from poor motor co-ordination [20] and perhaps pulmonary function, leading to lowerphysical self-confidence and perceived physical ability [47,50],possibly further aggravated by poor visual acuity, and togethermaking physical activity less rewarding. This may lead to lowerdegrees of physical activity and may also aggravate the lower ex-ercise capacity [20] and lower lean body mass [18]. Therefore, webelieve that promotion of health-enhancing physical activity de-serves attention in the healthcare of children and adults bornpreterm.

9. Diet

Although nutrient intake is closely related to cardiometabolicrisk factors, it has been relatively little studied in adults born pre-term. In HeSVA we showed by 3-day food diary that VLBW adults

Table 2Physical activity and fitness in adolescents and adults born very preterm.

Study Age(years)

Measure Inclusion No. preterm/control

Result Adjusted for:

Physical activityRogers et al. [20] 16e19 Sports participation <800 g 53/31 Lower previous (62 vs 92%) and current

(34 vs 74%) sports participationHack et al. [47] 20 PA questions of CHIP-AE VLBW 241/232 0.24 SD lower PA score Sex, SESKajantie et al. [48] 18e27 Questionnaire Unimpaired VLBW 163/188 Conditioning LTPA: adjusted OR for

low frequency 1.73 (95% CI: 1.07e2.78),low intensity 2.49 (1.43e4.68) and shortsession duration 3.37 (1.45e7.85)

Sex, age, SES height,smoking, maternalsmoking

Kaseva et al. [49] 21e30 KIHD 12-monthquestionnaire

Unimpaired VLBW 94/101 Conditioning LTPA: energy expenditure68.4% lower (adjusted; 95% CI: 37.2e84.0%)

Sex, age, SES, BMI,smoking, personality

Saigal et al. [50] 23 Physical self-efficacy scale ELBW 142/133 Lower perceived PA, physical self-presentationconfidence (differencesw0.5 SD) and regularsports participation (38 vs 59%),

Unadjusted

Andersen et al. [51] 15e69 Various questionnaires(meta-analysis)

1260e1750 ga 194/16 356 OR for adequate leisure-time physical activity0.67 (0.47e0.94)

Meta-regression,stratified for sex,cohort and age

Cardiorespiratory or muscle fitnessRogers et al. [20] 16e19 Step test, comprehensive

assessment of muscle fitness<800 g 53/31 Lower cardiorespiratory and muscle fitness

(SD not reported and effect size not possibleto calculate)

Unadjusted

Saigal et al. [50] 23 Handgrip Unimpaired ELBW 109/130 32 vs 38 kg; mean difference 0.64 SD Unadjusted

PA, physical activity; CHIP-AE, Child Health and Illness Profile, Adolescent Edition; VLBW, very low birth weight; SD, standard deviation; SES, socio-economic status; LTPA,leisure-time physical activity; OR, odds ratio; CI, confidence interval; KIHD, Kuopio Ischaemic Heart Disease Study (a detailed questionnaire of monthly frequencies andintensities of different physical activity types); ELBW, extremely low birth weight.

a Meta-analysis also including other categories.

E. Kajantie, P. Hovi / Seminars in Fetal & Neonatal Medicine 19 (2014) 112e117 115

consume less fruit, vegetables and berries than controls [53]. Lowconsumption of fruit is associated with high blood pressure andcardiovascular disease and was ranked by the Global Burden ofDisease Study as the fifth most important factor causing worldwidedisease burden e more important than overweight and obesity orphysical inactivity [54]. VLBW adults also consumed less milkproducts, resulting in lower calcium and vitamin D intakes, whichmay predispose to high blood pressure or osteoporosis; macronu-trient and sodium intakes were similar [53]. Reasons for thesefindings are not clear: as reviewed in detail [55], they may dependon the VLBW subject’s individual taste preferences for food, or oralmotor capabilities, or on eating habits of the family. Whatever thereasons, if the result is replicated, nutritional interventions mighthave a role in reducing cardiometabolic risk in adult life.

10. Sleep

The amount and quality of sleep is closely related to car-diometabolic risk factors [56]. VLBW adults report more sleep-disordered breathing [57]. This may be related to adenotonsillarhypertrophy or the anatomy of the pharynx, as they are also two-fold more likely to report a history of adenoidectomy [58]. Amore detailed assessment of sleep-disordered breathing wouldrequire polysomnography which has to our knowledge not beenused in this context. However, a study using accelerometry re-ported no difference in the amount or quality of sleep betweenVLBWadults and controls [59]. Interestingly, VLBW adults reportedearlier bedtimes and wake-up times than controls, which was alsoconfirmed by accelerometry [59]. Accordingly, in a standardizedquestionnaire they had higher scores indicating a morning prefer-ence than controls [60]. This is potentially important as a morningpreference is associated with good health, including lower bloodpressure and lower rates of type 2 diabetes [56].

11. Smoking and alcohol use

Some [61,62], although not all [63], studies suggest that adultsborn VLBW smoke less often than their peers born at term. This

difference is at least in part driven by lower smoking rates in VLBWadults with a history of bronchopulmonary dysplasia [64]. Use ofalcohol, in particular getting drunk, is also less frequent [61e63].These lower rates may be one sign of a general propensity forcautious and less risk-taking behaviour, which among other bene-fits may also protect from cardiometabolic disease [8,63].

12. Are risk factors for cardiometabolic disease differentaccording to the aetiology of preterm birth?

Frequent and often overlapping reasons for medically indicatedpreterm delivery include intrauterine growth restriction, usuallyindicated by being born SGA, and maternal pre-eclampsia. Bothconditions are, per se, associated with cardiometabolic risk factorsamong offspring born at term [1,2,65]. A general conclusion of most[5], but not all [9,24], studies of preterm survivors is that theseconditions have remarkably little effect on risk factors for disease,including body composition [18], blood pressure [18,25,26,28,30],glucose regulation [18,66], physical activity [48,49] and diet [53],over and above effects of very preterm birth. The prenatal envi-ronment of VLBW subjects exposed to maternal pre-eclampsia orborn SGA is very different from that of unexposed VLBW subjects,whereas postnatal challenges are to a great extent similar. There-fore, the similar cardiometabolic risk factors in these groups implythat the differences in these factors are a result of postnatal eventsor prematurity itself, rather than prenatal conditions.

13. Neonatal nutrition and growth

Postnatal events during infancy may play a role in the devel-opment of cardiometabolic risk. As recently reviewed, there are nopublished studies assessing the effects of neonatal nutrition onadult cardiometabolic risk [5], and the few existing studies inchildren and adolescents show mixed results. In the absence ofnutritional data, growth has been used as a surrogate measure ofneonatal morbidity and nutrition. These associations are alsoinconsistent [5]. Blood pressuremay, if anything, be associatedwithslow growth between birth and term [30], whereas there is some

Research directions

� Mechanisms underlying increased cardiometabolic riskfactors. In adults born very preterm, candidate mecha-nisms include visceral fat and arterial stiffness.

� Ways to promote health-enhancing physical activity inchildren and adults born very preterm.

� Identification of subgroups with high levels of risk factorsor with protective factors through analysis of pooleddatasets.

E. Kajantie, P. Hovi / Seminars in Fetal & Neonatal Medicine 19 (2014) 112e117116

evidence that impaired glucose regulation might be associatedwith more rapid growth [18,36,45], although in one study this wasconfined to those born SGA [18]. Two studies have assessed theeffects of weight gain during the first 2 weeks on endothelialfunction (flow-mediated dilatation): one study in adolescents born<1850 g showed more rapid weight gain to predict poor function[36], whereas the other in VLBW adults found the opposite [35].Weight gain in later childhood may be important; it is discussed byRoberts and Cheong in this issue (Chapter 8).

14. Identification of risk and protective factors requiresanalysis of pooled data

In terms of prevention, it would be important to identify sub-groups of adults born very preterm with particularly high risks ofcardiometabolic disease, or subgroups characterized by protectivefactors. The subgroups could be characterized by intrauterine orpostnatal growth, as outlined above, or by complications of pre-maturity or treatments such as neonatal nutrition or antenatalcorticosteroids. Knowledge is thus far very limited. Single follow-up studies usually have limited power to assess subgroup effects,whichmay require analysis of pooled data fromdifferent studies. Toachieve this, we have established the APIC (Adult Born PretermInternational Collaboration) that serves as a structured frameworkto facilitate pooling of data across cohorts and also co-ordination ofmeasures in future follow-up studies. Currently the APIC includes12 cohorts and a range of interest groups focusing on specificoutcomes: blood pressure, growth and body composition, pulmo-nary, reproductive, psychopathology, executive functioning, neu-roimaging and quality of life. We welcome new collaborations withgroups working with existing data or planned data collections inadults born very preterm.

15. Conclusions

There is now compelling evidence that adults born very small orvery preterm have increased risk factors for later cardiometabolicdisease. These risk factors include physiological ones such as higherblood pressure, lower lean body mass, impaired glucose regulation,and perhaps a more atherogenic lipid profile, and also lifestyle fac-tors such as low rates of physical activity and a less healthy diet. Anyintervention reducing these risk factors would have a substantialpotential to reduce the lifetime disease burden in small preterminfants. There are also protective factors such as lower rates ofsmoking, lower alcohol intake, and personality traits such asconscientiousness which is associated with good health and betteradherence to lifestyle changes. Althoughmore research is needed toidentify additional risk and protective factors, there is now enoughinformation towarrant an expert evaluation of the level of evidencefor cardiometabolic disease in individuals born very small or verypreterm, and its possible public health implications.

Practice points

� Adults born very preterm have higher levels of car-diometabolic risk factors.

� Risk factors can be reduced with a healthy lifestyle.� Measurement of blood pressure and appropriate treat-ment are important when encountering adults and chil-dren born very preterm.

� Adults and children born preterm and their familiesshould bemade aware of prevailing recommendations onhealth-enhancing physical activity. Ways to increasephysical activity should be discussed when possible.

Conflict of interest statement

None declared.

Funding sources

None.

References

[1] Barker DJ. Mothers, babies and health in later life. Edinburgh: ChurchillLivingstone; 1998.

[2] Gluckman PD, Hanson MA, Cooper C, Thornburg KL. Effect of in utero andearly-life conditions on adult health and disease. N Engl J Med 2008;359:61e73.

*[3] de Jong F, Monuteaux MC, van Elburg RM, Gillman MW, Belfort MB. Sys-tematic review and meta-analysis of preterm birth and later systolic bloodpressure. Hypertension 2011;59:226e34.

*[4] Parkinson JR, Hyde MJ, Gale C, Santhakumaran S, Modi N. Preterm birth andthe metabolic syndrome in adult life: a systematic review and meta-analysis.Pediatrics 2013;131:e1240e63.

[5] Lapillonne A, Griffin IJ. Feeding preterm infants today for later metabolic andcardiovascular outcomes. J Pediatr 2013;162:S7e16.

*[6] Saigal S, Doyle LW. An overview of mortality and sequelae of preterm birthfrom infancy to adulthood. Lancet 2008;371:261e9.

[7] Hack M. Adult outcomes of preterm children. J Dev Behav Pediatr 2009;30:460e70.

[8] Pesonen AK, Räikkönen K, Heinonen K, Andersson S, Hovi P, Järvenpää AL,et al. Personality in young adults with very low birth weight e Helsinki studyof very low birth weight adults. J Child Psychol Psychiatry 2008;49:609e17.

[9] Kaijser M, Bonamy AK, Akre O, Cnattingius S, Granath F, Norman M, et al.Perinatal risk factors for diabetes in later life. Diabetes 2009;58:523e6.

[10] Kajantie E, Osmond C, Barker DJP, Eriksson JG. Preterm birth e a risk factor fortype 2 diabetes? The Helsinki birth cohort study. Diabetes Care 2010;33:2623e5.

[11] Lawlor DA, Davey Smith G, Clark H, Leon DA. The associations of birthweight,gestational age and childhood BMI with type 2 diabetes: findings from theAberdeen Children of the 1950s cohort. Diabetologia 2006;49:2614e7.

[12] Crump C, Winkleby MA, Sundquist K, Sundquist J. Risk of diabetes amongyoung adults born preterm in Sweden. Diabetes Care 2011;34:1109e13.

[13] Pilgaard K, Færch K, Carstensen B, Poulsen P, Pisinger C, Pedersen O, et al. Lowbirthweight and premature birth are both associated with type 2 diabetes ina random sample of middle-aged Danes. Diabetologia 2010;53:2526e30.

[14] Koupil I, Leon DA, Lithell HO. Length of gestation is associated withmortality from cerebrovascular disease. J Epidemiol Community Health2005;59:473e4.

*[15] Lawlor DA, Hübinette A, Tynelius P, Leon DA, Smith GD, Rasmussen F. Asso-ciations of gestational age and intrauterine growth with systolic bloodpressure in a family-based study of 386,485 men in 331,089 families. Circu-lation 2007;115:562e8.

*[16] Kaijser M, Bonamy AK, Akre O, Cnattingius S, Granath F, Norman M, et al.Perinatal risk factors for ischemic heart disease: disentangling the roles ofbirth weight and preterm birth. Circulation 2008;117:405e10.

[17] Osmond C, Kajantie E, Forsén T, Eriksson JG, Barker DJ. Infant growth andstroke in adult life: the Helsinki birth cohort study. Stroke 2007;38:264e70.

*[18] Hovi P, Andersson S, Eriksson JG, Järvenpää AL, Strang-Karlsson S, Mäkitie O,et al. Glucose regulation in young adults with very low birth weight. N Engl JMed 2007;356:2053e63.

[19] Hack M, Schluchter M, Cartar L, Rahman M, Cuttler L, Borawski E. Growth ofvery low birth weight infants to age 20 years. Pediatrics 2003;112:e30e8.

[20] Rogers M, Fay TB, Whitfield MF, Tomlinson J, Grunau RE. Aerobic capacity,strength, flexibility, and activity level in unimpaired extremely low birthweight (< or ¼ 800 g) survivors at 17 years of age compared with term-borncontrol subjects. Pediatrics 2005;116:e58e65.

E. Kajantie, P. Hovi / Seminars in Fetal & Neonatal Medicine 19 (2014) 112e117 117

[21] Saigal S, Stoskopf B, Streiner D, Paneth N, Pinelli J, Boyle M. Growth trajec-tories of extremely low birth weight infants from birth to young adulthood: alongitudinal, population-based study. Pediatr Res 2006;60:751e8.

[22] Weiler HA, Yuen CK, Seshia MM. Growth and bone mineralization of youngadults weighing less than 1500 g at birth. Early Hum Dev 2002;67:101e12.

[23] Bigaard J, Frederiksen K, Tjønneland A, Thomsen BL, Overvad K, Heitmann BL,et al. Waist circumference and body composition in relation to all-causemortality in middle-aged men and women. Int J Obes (Lond) 2005;29:778e84.

[24] Evensen KA, Steinshamn S, Tjønna AE, Stølen T, Høydal MA, Wisløff U, et al.Effects of preterm birth and fetal growth retardation on cardiovascular riskfactors in young adulthood. Early Hum Dev 2009;85:239e45.

[25] Hack M, Schluchter M, Cartar L, Rahman M. Blood pressure among very lowbirth weight (<1.5 kg) young adults. Pediatr Res 2005;58:677e84.

[26] Doyle LW, Faber B, Callanan C, Morley R. Blood pressure in late adolescenceand very low birth weight. Pediatrics 2003;111:252e7.

[27] Kistner A, Celsi G, Vanpee M, Jacobson SH. Increased blood pressure butnormal renal function in adult women born preterm. Pediatr Nephrol2000;15:215e20.

[28] Keijzer-Veen MG, Finken MJ, Nauta J, Dekker FW, Hille ET, Frolich M, et al. Isblood pressure increased 19 years after intrauterine growth restriction andpreterm birth? A prospective follow-up study in The Netherlands. Pediatrics2005;116:725e31.

*[29] Norman M. Preterm birth e an emerging risk factor for adult hypertension?Semin Perinatol 2010;34:183e7.

[30] Hovi P, Andersson S, Räikkönen K, Strang-Karlsson S, Järvenpää AL,Eriksson JG, et al. Ambulatory blood pressure in young adults with very lowbirth weight. J Pediatr 2010;156:54e9.

[31] Cook NR, Cohen J, Hebert PR, Taylor JO, Hennekens CH. Implications of smallreductions in diastolic blood pressure for primary prevention. Arch InternMed 1995;155:701e9.

[32] Keijzer-Veen MG, Kleinveld HA, Lequin MH, Dekker FW, Nauta J, de Rijke YB,et al. Renal function and size at young adult age after intrauterine growthrestriction and very premature birth. Am J Kidney Dis 2007;50:542e51.

[33] Hodgin JB, Rasoulpour M, Markowitz GS, D’Agati VD. Very low birth weight isa risk factor for secondary focal segmental glomerulosclerosis. Clin J Am SocNephrol 2009;4:71e6.

[34] Kistner A, Jacobson L, Jacobson SH, Svensson E, Hellström A. Low gestationalage associated with abnormal retinal vascularization and increased bloodpressure in adult women. Pediatr Res 2002;51:675e80.

[35] Hovi P, Turanlahti M, Strang-Karlsson S, Wehkalampi K, Järvenpää AL,Eriksson JG, et al. Intima-media thickness and flow-mediated dilatation inthe Helsinki study of very low birth weight adults. Pediatrics 2011;127:e304e11.

[36] Singhal A, Cole TJ, Fewtrell M, Deanfield J, Lucas A. Is slower early growthbeneficial for long-term cardiovascular health? Circulation 2004;109:1108e13.

[37] McEniery CM, Bolton CE, Fawke J, Hennessy E, Stocks J, Wilkinson IB, et al.Cardiovascular consequences of extreme prematurity: the EPICure study.J Hypertens 2011;29:1367e73.

[38] Pyhälä R, Räikkönen K, Feldt K, Andersson S, Hovi P, Eriksson JG, et al. Bloodpressure response to psychosocial stress in young adults with very low birthweight e the Helsinki study of very low birth weight adults. Pediatrics2009;123:731e4.

[39] Lewandowski AJ, Augustine D, Lamata P, Davis EF, Lazdam M, Francis J, et al.Preterm heart in adult life: cardiovascular magnetic resonance revealsdistinct differences in left ventricular mass, geometry, and function. Circu-lation 2013;127:197e206.

[40] Lewandowski AJ, Bradlow WM, Augustine D, Davis EF, Francis J, Singhal A,et al. Right ventricular systolic dysfunction in young adults born preterm.Circulation 2013;128:713e20.

[41] Hofman PL, Regan F, Jackson WE, Jefferies C, Knight DB, Robinson EM, et al.Premature birth and later insulin resistance. N Engl J Med 2004;351:2179e86.

[42] Kajantie E, Strang-Karlsson S, Hovi P, Wehkalampi K, Lahti J, Räikkönen K,et al. Preterm birth and insulin sensitivity in young adult life e the Helsinkistudy of very low birth weight adults. In: Developmental origins of healthand disease 7th World Congress, Portland, OR, USA 2011. J Dev Orig HealthDis 2011;2:S103,P12–240.

[43] Rotteveel J, van Weissenbruch MM, Twisk JW, Delemarre-Van de Waal HA.Infant and childhood growth patterns, insulin sensitivity, and blood pressurein prematurely born young adults. Pediatrics 2008;122:313e21.

[44] Smith CM, Wright NP, Wales JK, Mackenzie C, Primhak RA, Eastell R, et al.Very low birth weight survivors have reduced peak bone mass and reducedinsulin sensitivity. Clin Endocrinol (Oxf) 2011;75:443e9.

[45] Willemsen RH, Leunissen RW, Stijnen T, Hokken-Koelega AC. Prematurity isnot associated with reduced insulin sensitivity in adulthood. J Clin Endo-crinol Metab 2009;94:1695e700.

[46] Hovi P, Kajantie E, Soininen P, Kangas AJ, Järvenpää AL, Andersson S, et al.Lipoprotein subclass profiles in young adults born preterm at very low birthweight. Lipids Health Dis 2013;12:57.

[47] Hack M, Cartar L, Schluchter M, Klein N, Forrest CB. Self-perceived health,functioning and well-being of very low birth weight infants at age 20 years.J Pediatr 2007;151:635e41. 41 e1e2.

[48] Kajantie E, Strang-Karlsson S, Hovi P, Räikkönen K, Pesonen AK, Heinonen K,et al. Adults born at very low birth weight exercise less than their peers bornat term. J Pediatr 2010;157:610e6.

*[49] Kaseva N, Wehkalampi K, Strang-Karlsson S, Salonen M, Pesonen AK,Räikkönen K, et al. Lower conditioning leisure-time physical activity in youngadults born preterm at very low birth weight. PLoS One 2012;7. e32430.

[50] Saigal S, Stoskopf B, Boyle M, Paneth N, Pinelli J, Streiner D, et al. Comparisonof current health, functional limitations, and health care use of young adultswho were born with extremely low birth weight and normal birth weight.Pediatrics 2007;119:e562e73.

[51] Andersen LG, Angquist L, Gamborg M, Byberg L, Bengtsson C, Canoy D, et al.Birth weight in relation to leisure time physical activity in adolescence andadulthood: meta-analysis of results from 13 nordic cohorts. PLoS One 2009;4.e8192.

[52] Welsh L, Kirkby J, Lum S, Odendaal D, Marlow N, Derrick G, et al. The EPICurestudy: maximal exercise and physical activity in school children bornextremely preterm. Thorax 2010;65:165e72.

*[53] Kaseva N, Wehkalampi K, Hemiö K, Hovi P, Järvenpää AL, Andersson S, et al.Diet and nutrient intake in young adults born preterm at very low birthweight. J Pediatr 2013;163:43e8.

[54] Lim SS, Vos T, Flaxman AD, Danaei G, Shibuya K, Adair-Rohani H, et al.A comparative risk assessment of burden of disease and injury attributable to67 risk factors and risk factor clusters in 21 regions, 1990e2010: a systematicanalysis for the Global Burden of Disease Study 2010. Lancet 2012;380:2224e60.

*[55] Portella AK, Kajantie E, Hovi P, Desai M, Ross MG, Goldani MZ, et al. Effects ofin utero conditions on adult feeding preferences. J Dev Origins Health Dis2012;3:140e52.

[56] Merikanto I, Lahti T, Puolijoki H, Vanhala M, Peltonen M, Laatikainen T, et al.Associations of chronotype and sleep with cardiovascular diseases and type 2diabetes. Chronobiol Int 2013;30:470e7.

[57] Paavonen EJ, Strang-Karlsson S, Raikkonen K, Heinonen K, Pesonen AK,Hovi P, et al. Very low birth weight increases risk for sleep-disorderedbreathing in young adulthood: the Helsinki study of very low birth weightadults. Pediatrics 2007;120:778e84.

[58] Parviainen A, Hovi P, Strang-Karlsson S, Järvenpää AL, Eriksson JG,Andersson S, et al. Two-fold rates of adenoidectomy after preterm birth: theHelsinki study of very low birth weight adults. In: Pediatric Academic Soci-eties Annual Meeting, Washington, DC, USA. http://www.pas-meeting.org/past/2013DC/default.asp; 2013.

[59] Strang-Karlsson S, Räikkönen K, Kajantie E, Andersson S, Hovi P,Heinonen K, et al. Sleep quality in young adults with very low birthweight e the Helsinki study of very low birth weight adults. J PediatrPsychol 2007;33:387e95.

[60] Strang-Karlsson S, Kajantie E, Pesonen AK, Räikkönen K, Hovi P, Lahti J, et al.Morningness propensity in young adults born prematurely: the Helsinkistudy of very low birth weight adults. Chronobiol Int 2010;27:1829e42.

[61] Strang-Karlsson S, Räikkönen K, Pesonen AK, Kajantie E, Paavonen EJ, Lahti J,et al. Very-low-birth-weight and behavioral symptoms of ADHD in youngadulthood e the Helsinki study of very low birth weight adults. Am J Psy-chiatry 2008;165:1345e53.

[62] Hille ET, Dorrepaal C, Perenboom R, Gravenhorst JB, Brand R, Verloove-Vanhorick SP. Social lifestyle, risk-taking behavior, and psychopathology inyoung adults born very preterm or with a very low birthweight. J Pediatr2008;152:793e800. e1e4.

[63] Hack M, Flannery DJ, Schluchter M, Cartar L, Borawski E, Klein N. Outcomes inyoung adulthood for very-low-birth-weight infants. N Engl J Med 2002;346:149e57.

[64] Doyle LW, Faber B, Callanan C, Freezer N, Ford GW, Davis NM. Broncho-pulmonary dysplasia in very low birth weight subjects and lung function inlate adolescence. Pediatrics 2006;118:108e13.

[65] Davis EF, Lazdam M, Lewandowski AJ, Worton SA, Kelly B, Kenworthy Y, et al.Cardiovascular risk factors in children and young adults born to preeclampticpregnancies: a systematic review. Pediatrics 2012;129:e1552e61.

[66] Finken MJ, Keijzer-Veen MG, Dekker FW, Frolich M, Hille ET, Romijn JA, et al.Preterm birth and later insulin resistance: effects of birth weight and post-natal growth in a population based longitudinal study from birth into adultlife insulin resistance 19 years after preterm birth. Diabetologia 2006;49:478e85.

lable at ScienceDirect

Seminars in Fetal & Neonatal Medicine 19 (2014) 118e124

Contents lists avai

Seminars in Fetal & Neonatal Medicine

journal homepage: www.elsevier .com/locate/s iny

Review

Long-term growth and general health for the tiniest or most immatureinfants

Gehan Roberts a,b,c,*, Jeanie L.Y. Cheong a,c,d

a Premature Infant Follow-up Program at the Royal Women’s Hospital, Melbourne, AustraliabDepartment of Paediatrics, University of Melbourne, Melbourne, AustraliacMurdoch Childrens Research Institute, Melbourne, AustraliadDepartment of Obstetrics and Gynaecology, University of Melbourne, Melbourne, Australia

Keywords:Extremely pretermGrowthHealth

* Corresponding author. Address: Centre for ComChildren’s Hospital, Flemington Rd, Parkville 3052, Au

E-mail address: gehan.roberts@rch.org.au (G. Robe

1744-165X/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.siny.2013.11.003

s u m m a r y

Given the improving survival rates of extremely preterm (EP, gestational age <28 weeks) infants, there isa need to understand their general growth and health outcomes not only in childhood, but also intoadulthood. EP children are shorter and lighter compared with term children at term-equivalent age; withtime, the weight disadvantage diminishes but the height disadvantage remains relatively unchanged. EPchildren and young adults also have higher rates of reported health concerns, medical conditions andvisual impairment. Hospital readmissions are higher in early childhood, mostly attributed to respiratoryillness. Individuals born EP have reduced bone health and are at increased risk for metabolic disorders.Increased rates of conditions such as diabetes or pathological fractures are not reported in the literature,although follow-up studies so far have only tracked EP individuals into young adulthood. Consequently,health care utilization and costs are increased in EP children and young adults. A thorough knowledge ofthe health risks related to EP birth is essential in planning surveillance and intervention strategies tooptimize their health and wellbeing. Despite the increased risk of health problems, EP young adultsgenerally report their quality of life to be similar to that reported in their term counterparts.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The survival of infants born extremely preterm (EP, gestationalage <28 weeks) has been steadily increasing over the last severaldecades [1,2]. More of these children are reaching adulthood thanever before [3]. Research into better understanding the outcomesfor these individuals has progressed from crude estimates of mor-tality and major morbidity to a more nuanced understanding oftheir increased risk for many chronic conditions [3].

This review focuses on growth and the general health outcomesof individuals whowere born EP and survived to discharge from thenursery. Gestational age and birthweight are continuous variables,and the gradient of outcomes across these variables providesimportant context to understanding the outcomes of EP in-dividuals. Therefore, we will also review research that describesoutcomes for cohorts defined according to birthweight. As cardio-vascular and respiratory outcomes and cerebral palsy are covered

munity Child Health, Royalstralia. Tel.: þ61 3 9345 5356.rts).

All rights reserved.

elsewhere in this issue, they are not the focus of this review, and areonly discussed briefly in the context of other health outcomes. It isimportant to note that most EP individuals will not have serious,life-limiting health conditions, and tend to self-report robustquality-of-life and health status [4]. Quality-of-life outcomes arediscussed more fully in this issue by Saigal (Chapter 9).

The Millennium Cohort Study, a nationally representative, pro-spective UK cohort study, has clearly shown a gradient effect ofgestation on adverse health outcomes, with children who are bornat the lowest gestational ages at the greatest risk for poorer healthoutcomes, even after controlling for potential confounders such assex, ethnicity, social risk, breast feeding and in-utero exposure topotentially harmful insults [5]. Therefore, the EP population isclearly a high-risk group for a variety of adverse health outcomes,with implications for counselling families in the nursery, screeningand surveillance throughout childhood and early intervention,where possible, to prevent adult morbidity. Well-established pop-ulation-level prevention theory suggests that the EP population islikely to benefit not only from the universal screening provided tothe whole population, but also from selective prevention strategies,targeted to specific health issues that occur at a greater prevalencein this population [6].

G. Roberts, J.L.Y. Cheong / Seminars in Fetal & Neonatal Medicine 19 (2014) 118e124 119

2. Growth

Children who are born preterm begin life both shorter andlighter than their term-born peers at term equivalent age [7].However, this weight disadvantage tends to diminish during theadolescent years, although height disadvantages persist, comparedwith term controls [8]. A limitation of several studies in this field isthe use of birthweight as the selection criterion: this can introducebias, as more mature growth-restricted infants (who may be athigher risk of future growth disorders) are included with infantswho are more premature, but appropriate weight for gestation [9].Most studies in this field have reported data from the pre-surfactant era [8], when the survival rates from EP individualswere much lower than those seen in the modern era of neonatalintensive care [10]. Surfactant for treatment of respiratory distress

Fig. 1. Changes in weight and height standard deviation (SD) scores between birth and age 1Australia, in 1991e1992. The size of the circles in the figure is proportional to the numbdischarge.

syndrome in preterm infants was introduced in the late 1980s andearly 1990s, and resulted in decreased mortality worldwide [10].

Roberts et al. [9] recently published data from a large,population-based cohort of EP adolescents and matched termcontrols, born in the early 1990s, when surfactant use wasconsidered standard care for these tiny infants. These young peoplewere recruited at birth, and assessed at 2, 5, 8 and 18 years of age.As seen in Fig. 1, the EP adolescents had lower weight Z-scores thancontrols at birth, but the greatest difference was seen at hospitaldischarge [EP weight Z-score: 1.16 SD (95% confidence interval (CI):0.99e1.33) lower than term controls]. This difference decreasedprogressively until age 18 years [EPweight Z-score: 0.38 SD (95% CI:0.09e0.67) lower than controls]. The EP children were shorter thancontrols at all ages, and this difference of w0.7 SD did not altergreatly over time. The results previously reported in historical

8 years in extremely preterm (EPT) and term control (Con) individuals born in Victoria,er of individuals. Mid-parental height SD score is shown in the height panel. Disch.,

G. Roberts, J.L.Y. Cheong / Seminars in Fetal & Neonatal Medicine 19 (2014) 118e124120

cohorts, therefore, remained consistent with those seen for thismore contemporary cohort, despite higher survival rates of thetiniest and most vulnerable infants.

Mid-parental height (MPH) is a widely used measure to predictadult height in the general population [11]. This association wassupported in a cohort of term-born adolescents by Roberts et al. [9];however, for the EP individuals, height at age 2 years was a strongerpredictor of height at age 18 years than MPH. These findings sug-gest that early health and growth have a greater influence on finalheight than genetic potential in those born EP. Interestingly, theMPH of parents of the EP adolescents was slightly lower than that ofthe parents of the term controls. The EP adolescents’ height Z-scores at age 18 years were, on average, slightly lower than theirMPH, and the term control adolescents’ height Z-scores weregreater than their MPH [9].

Height and weight, therefore, should be closely monitored in EPindividuals, and if growth failure can be prevented in the earlyyears, final growthmay be predicted to be optimal, according to theindividual’s genetic potential. Although physical growth is relatedto overall health and self-esteem, it is encouraging to know that theheight disadvantage seen in EP adolescents, compared with termcontrols, is not associated with any differences in self-reportedhealth status or self-esteem [4].

3. Overview of general health conditions

3.1. General health

The most frequently occurring medical conditions contributingto poorhealth inpretermchildren and young adults include asthma,recurrent bronchitis and epilepsy [12e14]. Health status is generallydetermined by questionnaires concerning chronic medical, neuro-logic or mental health conditions, with ‘poorer health‘ status re-ported to be higher in early childhood in those born very preterm(<32 weeks of gestation) compared with children born between 39and 41 weeks [13,14]. Boyle et al. [5] demonstrated a gradient effectof poorer health outcomes with decreasing gestational age,including general health, hospital admissions and longstandingillness. For children born<32weeks of gestation, the odds ratio (OR)of having a longstanding illness limiting activities at age 5 years was3.9 (95% CI: 2.4e6.3) compared with term children [5]. Parentalperception of ill-health was also higher (OR: 2.3; 95% CI: 1.0e4.5).Although the rates of chronic illness in preterm children are higherthan in term controls, studies looking at changes with age have notshown a change in prevalence. For example, Hack et al. reported onthe trajectory of rates of chronic conditions including asthma, andfound steady rates between the ages of 8 and 14 years [15].

Higher rates of ‘poor health’ compared with term controls havebeen reported in adolescence and young adulthood [13,14,16]. In acohort of young adults aged 20 years, Hack et al. reported higherrates of chronic medical or psychiatric disorders in those born verylow birthweight, i.e. 19e23% compared with 16% in controls [13].No sex differencewas noted. In a Norwegian registry study of youngadults (age range: 20e36 years), the relative risk of non-psychiatricmedical disorders (including blindness, low vision, hearing loss andepilepsy) was higher with decreasing gestational age at birth. Therelative risk, compared with young adults born �37 weeks ofgestation, was 19.6 (95% CI: 11.9e32.2) for those born<28 weeks ofgestation, which decreased to 9.3 (95% CI: 6.6e13) for those bornbetween 28 and <31 weeks [16].

3.2. Vision and hearing

Visual sequelae associated with extremely preterm birth ismostly a consequence of retinopathy of prematurity (ROP) and

cerebral visual impairment [17]. With the availability of magneticresonance imaging of the newborn brain, there has been increasedrecognition of cerebral involvement of the retrogeniculate visualpathways, related to preterm brain injury, that may be present evenin the absence of significant ROP [17].

With the introduction of cryotherapy for treatment of severeROP in the 1990s, blindness rates decreased from around 8e10% to<2% [18e20]. A gestational age gradient for blindness and severevisual impairment has been reported, with rates of around 4e8% inthose born at <25 weeks compared with 1e2% in those born be-tween 26 and 27 weeks [14,21,22]. These rates remained un-changed between the ages of 8 and 14 years, as reported in alongitudinal study of extremely low birth weight (ELBW: birth-weight <1000 g) children [15].

Refractive error and visual accommodation problems are moreprevalent in extremely preterm children compared with termcontrols [17]. Myopia and hypermetropia rates of around 20e30%have been reported in early childhood [21,23], and rates of stra-bismus are in the range of 10e30%, with the lowest rates reportedin children born in the late 1990s compared with those born adecade earlier [22,24]. EP or ELBW children and young adults alsohave higher rates of prescription glasses and requirement for visualaids. In one study, up to 64% of ELBW young adults required pre-scription glasses compared with 37% term controls; and visual aidsincluding Braille equipment were used by 7% of ELBW young adultscompared with zero in term controls [14].

The risk of retinal detachment is highest in the neonatal periodin preterm infants with severe ROP, and close surveillance is stan-dard care during this period. Higher than expected rates of lateretinal detachment have also been reported in early adulthood, asdiscussed in this issue by Saigal (Chapter 9).

Recently, Molloy et al. reported impaired visual processing in acohort of 18-year-old ELBWor EP subjects, with worse visual acuity,poorer stereopsis, convergence and visual perception after habitualcorrection (i.e. wearing the participant’s own prescription glasses)[25], which lends further support that visual impairment in pre-term children and adolescents not only have ocular but also cere-bral involvement.

In general, hearing impairment rates in preterm children andyoung adults are low, in the range of 1e4% [13,20,23]. Doyle et al.[20] reported similar rates of hearing impairment (as defined byhearing loss requiring amplification) in ELBW or EP children at age2 years, in cohorts born in 1979e1980 compared with those born in2005 in Victoria, Australia. In a Canadian cohort of young adults, 7%of the ELBW group were reported to have ‘hearing difficultiesresulting in functional limitations’ (as determined by a standard-ized questionnaire) compared with 2% in term controls [14].However, there was no difference in requirement for hearing aidsbetween the groups.

Auditory processing deficits have been reported in ELBW ado-lescents compared with term controls, and these deficits areassociated with poorer intellectual, academic and behaviouralprogress [26]. Altered patterns of speech and sound discriminationhave been identified in preterm infants as young as term age up to 6months [27,28] and 4 years. Although these deficits are likely toplay a role in learning difficulties at school, causal pathways havenot been clearly established. One study on learning disability sub-types could not establish significant differences in immediateauditory memory performances between language and perceptual-motor learning disability groups [29].

3.3. Bone health

Individuals who are born EP are at high risk of reduced bonemass and osteopenia, which will, in turn, increase the risk of

G. Roberts, J.L.Y. Cheong / Seminars in Fetal & Neonatal Medicine 19 (2014) 118e124 121

fractures during adult life. [30,31] Up to 80% of the body calcium ofa term infant accrues during the third trimester, so infants who areborn early commonly have suboptimal bone development [32].Historical studies have reported poor bone mineralization in 57% ofELBW infants [33]. Longitudinal cohort studies have shown thatthis increased risk persists over time. For example, a population-based cohort study of all very low birth weight (VLBW: birthweight <1500 g) births born within the larger Helsinki area from1978 to 1985 reported reduced femoral neck bone density [Z-score:e0.56 (95% CI: e0.78 to e0.34) SD] and reduced lumbar spine bonedensity [Z-score: e0.51 (95% CI: e0.75 to e0.28) SD] in adulthood,compared with controls. This did not translate to a higher rate ofpathological fractures, although the participants were youngadults, between 18 and 27 years of age at follow-up [31].

With improved monitoring and nutrition as standard care inneonatal nurseries, pathological fractures and rickets are now veryrare in infancy. However, there is as yet insufficient evidence tosuggest the best method of optimizing bone health beyond thenursery. Current evidence-based guidelines do not identify EP in-dividuals as a specific high-risk group for bone mineral densityscreening [34,35], and therefore the decision to screen or treat re-mains an individual decision, based on family history, neonatalcourse and current health.

3.4. Metabolic risk

Body composition and metabolic risk are of interest in EP chil-dren, given Barker’s developmental origins of adult disease hy-pothesis [13,36]. There is an increasing interest in the role ofmetabolic programming in preterm children predisposing them tometabolic disorders such as insulin resistance [37]. Low birthweight and growth restriction have been shown to be risk factorsfor adverse adult cardiovascular outcomes [38e40]. Whereas goodweight gains are certainly a desirable goal for the parents andcarers of EP children, weight gain that is too rapid may be associ-ated with an increased long-term risk of metabolic disorders. Thisincreased risk has been suggested by historic cohort studies, suchas that conducted by Eriksson et al. [39], who reported that childrenborn between 1924 and 1933 who had a lower body mass index(BMI; kg/m2) in early childhood, but whose BMI rose rapidly to theaverage or above average range in later childhood, had a greaterrisk of death due to coronary disease in adulthood, compared withthose who had a slower rise in BMI.

Hovi et al. [41] carried out glucose tolerance tests with VLBWyoung adults. Compared with the control group of normal birthweight controls, the VLBW adults had impaired glucose tolerance[6.7% increase in the 2 h glucose concentration (95% CI: 0.8e12.9)]and increased fasting insulin [40.0% increase in the 2 h insulinconcentration (95% CI: 17.5e66.8)]. These findings were not, as yet,associated with a higher risk of clinical diagnoses of diabetes [41].Similar to the findings of Eriksson et al., this increased risk may bedue to rapid weight gain in early life secondary to initial inadequatenutrition followed by increased post-discharge nutrition [39].Glucose tolerance is discussed further in the article by Kajantie andHovi in this issue (Chapter 7).

Roberts et al. [9] reported BMI Z-scores at multiple time-points in EP individuals, compared with term controls, in apopulation-based longitudinal cohort study. BMI Z-scores werelowest, compared with controls, at age 5 years (e0.78 SD; 95% CI:e1.03 to e0.53), but by age 18 years they were not different fromthe BMI Z-scores of the controls (e0.03 SD; 95% CI: e0.31 to 0.25)[9]. Euser et al. [8] previously reviewed this literature in 2008,and reported that, during childhood, preterm children tend tohave reduced fat mass compared with their peers, but havecatch-up weight gain during adolescence, although their final

BMI is not greater than the reference range in most studies. EPindividuals, therefore, have an accelerated BMI gain in middlechildhood, which may increase the risk of metabolic disorders inadulthood.

Growth and body composition therefore need to be carefullymonitored in EP individuals over time, and standardized growthcharts can be used to track growth and BMI. Primary care practi-tioners who provide a medical home for EP individuals are wellplaced to carry out this surveillance, and to initiate furtherscreening such as lipid profiles, fasting glucose and liver functiontests, and to introduce and monitor intervention strategies asneeded [6].

4. Health care utilization

4.1. Hospital readmissions

Hospital readmissions in the first few years following dischargeare greater in very preterm children compared with term-bornchildren [5,42,43]. In a large population cohort in the UnitedKingdom, children born<32 weeks of gestation had greater ORs forhospital readmissions compared with children born between 39and 41 weeks [5]. The adjusted OR (95% CI) for at least three hos-pital readmissions in the first 9monthswas 13.7 (6.5e29.2), and 6.0(3.2e11.4) between 9 months to 5 years respectively. The com-monest cause for readmission of very preterm-born children wasrespiratory illness, which has been reported to account for up to40% of hospital readmission in the first 2 years following discharge[42,44]. Other causes of readmission were gastrointestinal disor-ders, including gastroenteritis and gastro-esophageal reflux; andviral illness or fever [5]. However, despite the higher rates of hos-pital readmissions in very preterm children, it is important to keepin perspective that, at a population level, it is the early term babies(37e38 weeks of gestation) that contribute the greatestpopulation-attributable fraction to these outcomes, given theirmuch greater numbers. This metric refers to the reduction inmorbidity in the general population that would occur if thegestational age of these children was >38 weeks of gestation. Forexample, the population-attributable fractions for having at leastthree hospital admissions between 9 months and 5 years were 3.8%(95% CI: 1.3e6.5%) for children born <32 weeks, 5.7% (2.0e10.0) forchildren born 32e36 weeks and 7.2% (1.4e13.6) for birth at 37e38weeks.

In a Spanish multi-centre study of preterm children born at �32weeks of gestation, respiratory syncytial virus (RSV) was respon-sible for 66.3% of all the hospital readmissions in the first year afterdischarge [44]. Risk factors identified for respiratory readmissionsincluded lower gestational age at birth, bronchopulmonarydysplasia, parental smoking and living with school-age siblings[42,44].

Studies of preterm young adults have reassuringly reportedsimilar rates of hospitalizations or surgery compared with term-born peers, despite a higher prevalence of chronic health condi-tions in the preterm group [14]. Rates of acute hospitalizationswere decreased in adolescence when compared with earlierchildhood [45].

4.2. Health care costs and resource utilization

The higher rates of chronic health problems in EP individualslead to increased utilization of health resources and higher healthcare costs. As the absolute number of these individuals is small, it isdifficult to generate precise estimates, and studies often presentdata relevant to individuals born at slightly greater gestational ages.Recent data from a large, population-based longitudinal Australian

Practice points

� Extremely preterm individuals start life smaller andshorter than their term peers; the weight difference di-minishes over time, although the height difference per-sists over time.

� There are higher rates of medical conditions, includingrespiratory problems, epilepsy and visual impairment inchildren and young adults born extremely preterm.

� Hospital readmissions are greatest in early childhood, butappear to decrease to a similar rate by young adulthood.

� Perception of quality of life in young adulthood isgenerally similar to term-born peers.

G. Roberts, J.L.Y. Cheong / Seminars in Fetal & Neonatal Medicine 19 (2014) 118e124122

study reported that children born at <32 weeks with a birthweight<1500 g and/or extremely small for gestational age (<5thpercentile) had higher mean costs in the range of A$450e500 perchild per year of non-hospital healthcare attendances and pre-scription medication in the first 10 years compared with childrenwho were born at term with normal birth weight [46]. Similarfindings have been reported in other countries. A Finnish study thatestimated costs resulting from hospitalizations, visits to health careprofessionals and therapists, and the use of other social welfareservices in the fifth year after birth found that very preterm chil-dren had a 1.4-fold higher health cost compared with term chil-dren, and this increased to 4.4-fold in children with morbiditiessuch as cerebral palsy, seizures, obstructive airways disease, visualand hearing problems [12]. As can be expected given the gesta-tional age gradient for morbidities in childhood, the use of specialhealth care resources was highest in the most preterm anddecreased with increasing gestational age at birth. A large Frenchcohort study reported that at 5 years of age special health careresources were used by 42% of children born at 24e28 weeks,compared with 31% of those born at 29e32weeks, and 16% of thoseborn at 39e40 weeks [47].

The rates of prescription medication use is also higher in verypreterm children and young adults compared with term controls.Asthma medication is commonest in the younger age group [5],but increased use of psychotropic medication is also reported inyoung adults [14,48]. One study that looked at the change incompensatory dependencies (which included prescribed medica-tion, physician-ordered special diets), in a cohort of ELBW childrenbetween ages 8 and 14 years, reported that the significant in-creases from 22% to 34% were due to increases in prescribedmedication [15].

5. Screening and early intervention

As discussed above, EP individuals are a high-risk group whoshould receive surveillance and screening to identify and treat earlysigns of chronic health conditions. Despite a rich literaturedescribing this increased risk, there are, as yet, no consensusguidelines for clinical care. A few specific conditions e bloodpressure, RSV prophylaxis and osteopenia of prematurity e havesome evidence to guide prevention efforts.

5.1. Blood pressure (BP)

Blood pressure in ex-preterm in adulthood is discussed morefully in this issue by Kajantie and Hovi (Chapter 7). Children who‘have a history of prematurity, VLBW or other neonatal compli-cation requiring intensive care’ have been identified by the Na-tional Institutes of Health as a special risk group who should havetheir BP measured during routine health care visits after dischargefrom the nursery [49]. Auscultation is the preferred method inchildhood, but oscillometric devices are preferred in infants due toease of use. An elevated BP should be confirmed with a repeatmeasurement, and ambulatory BP monitoring can be used if thereare specific concerns, such as the need to rule out ‘white-coathypertension’. Hypertension is diagnosed based on standard ta-bles, usually using the 95th percentile as a cut-off, based on age,sex and height [49]. Further evaluation and treatment is beyondthe scope of this article.

5.2. Respiratory syncytial virus prophylaxis

As discussed above, much of the morbidity experienced by EPchildren during early childhood is due to respiratory illness. Hu-man RSV is a common pathogen, that can lead to serious illness

and even death in EP infants [50]. Palivizumab is a humanizedmonoclonal antibody that has been approved for the prevention ofRSV disease in high-risk groups. Bentley et al. [50] carried out arecent cost-effectiveness analysis of Palivizumab in the UK, usingthe metric of the incremental cost-effectiveness ratio (ICER, thechange in cost associated with a gain in health). The ICER forPalivizumab for infants born <29 weeks of gestation was £3,845per quality adjusted life-year (QALY) gained, and this rose rapidlyto £30,205 per QALY gained for infants born at 29e32 weeks ofgestation.

5.3. Osteopenia of prematurity

Nutritional supplementation and monitoring metabolic pa-rameters in the nursery are well established as standard care for EPinfants. A Cochrane review has concluded that passive range ofmotion exercises of the extremities for preterm infants during theinitial nursery admission resulted in a significant increase of bonemineral content and bone formation markers in the short term,although the data were inadequate to assess long-term effects [51].Whether long-term interventions or screening are effective is yet tobe established.

6. Summary and conclusions

Ever-increasing numbers of individuals who are born extremelypreterm are surviving to adulthood. Given the considerable in-vestments in their health care during the first months of their lives,it is important that their long-term health outcomes are well un-derstood by their families, their clinicians, and especially the EPindividuals themselves, as they assume responsibility for their ownhealth care over time. As gestation is a continuous variable, thehealth outcomes of EP individuals should be understood in thecontext of a gradient of adverse effects that increases as gestationalage decreases. As we have described, EP individuals experiencemore acute health conditions, including respiratory illness andhospitalization, compared with term peers. They also experiencemore chronic health conditions, such as vision, hearing and respi-ratory problems, comparedwith term peers. They are also at higherrisk of adverse metabolic, cardiovascular and bone health out-comes. Reassuringly, however, most EP individuals experience goodhealth and self-report health status similar to their term peers.These long-term outcomes should be taken into account whencounselling families faced with impending EP birth. Looking ahead,the challenge to clinicians involved in the long-term care of EPindividuals is to maintain constant vigilance, with screening, sur-veillance and intervention informed by the ever-increasingknowledge of the long-term health and growth consequences ofEP birth.

Research directions

� Follow-up beyond young adulthood is vital to under-standing the general health impact of very preterm birth.

� Identification of potentially modifiable lifestyle factorsthat may help inform intervention and preventative stra-tegies to improve long-term health of very pretermsurvivors.

G. Roberts, J.L.Y. Cheong / Seminars in Fetal & Neonatal Medicine 19 (2014) 118e124 123

Conflict of interest statement

None declared.

Funding sources

None.

References

*[1] Saigal S, Doyle LW. An overview of mortality and sequelae of preterm birthfrom infancy to adulthood. Lancet 2008;371(9608):261e9.

[2] Doyle LW, Roberts G, Anderson PJ. Outcomes at age 2 years of infants <28weeks’ gestational age born in Victoria in 2005. J Pediatr 2010;156. 49e53 e1.

*[3] Doyle LW, Anderson PJ. Adult outcome of extremely preterm infants. Pedi-atrics 2010;126:342e51.

[4] Roberts G, Burnett AC, Lee KJ, Cheong J, Wood SJ, Anderson PJ, et al. Quality oflife at age 18 years after extremely preterm birth in the post-surfactant era.J Pediatr 2013;163. 1008e13.e1.

*[5] Boyle EM, Poulsen G, Field DJ, Kurinczuk JJ, Wolke D, Alfirevic Z, et al. Effectsof gestational age at birth on health outcomes at 3 and 5 years of age:population based cohort study. BMJ (Clin Res Ed) 2012;344:e896.

[6] Mrazek PJ, Haggerty RJ. Reducing risks for mental disorders: frontiers forpreventive intervention research. Washington DC: National Academy Press,Institute of Medicine; 1994.

[7] Bracewell MA, Hennessy EM, Wolke D, Marlow N. The EPICure study: growthand blood pressure at 6 years of age following extremely preterm birth. ArchDis Child Fetal Neonatal Ed 2008;93:F108e14.

*[8] Euser AM, de Wit CC, Finken MJ, Rijken M, Wit JM. Growth of preterm bornchildren. Horm Res 2008;70:319e28.

*[9] Roberts G, Cheong J, Opie G, Carse E, Davis N, Duff J, et al. Growth ofextremely preterm survivors from birth to 18 years of age compared withterm controls. Pediatrics 2013;131:e439e45.

[10] Schwartz RM, Luby AM, Scanlon JW, Kellogg RJ. Effect of surfactant onmorbidity, mortality, and resource use in newborn infants weighing 500 to1500 g. N Engl J Med 1994;330:1476e80.

[11] Tanner JM, Goldstein H, Whitehouse RH. Standards for children’s height atage 2 to 9 years allowing for height of parents. Arch Dis Child1970;45(244):819.

[12] Korvenranta E, Linna M, Rautava L, Andersson S, Gissler M, Hallman M, et al.Hospital costs and quality of life during 4 years after very preterm birth. ArchPediatr Adolesc Med 2010;164:657e63.

[13] Hack M, Flannery DJ, Schluchter M, Cartar L, Borawski E, Klein N. Outcomes inyoung adulthood for very-low-birth-weight infants. N Engl J Med 2002;346:149e57.

[14] Saigal S, Stoskopf B, Boyle M, Paneth N, Pinelli J, Streiner D, et al. Comparisonof current health, functional limitations, and health care use of young adultswho were born with extremely low birth weight and normal birth weight.Pediatrics 2007;119:e562e73.

[15] Hack M, Schluchter M, Andreias L, Margevicius S, Taylor HG, Drotar D, et al.Change in prevalence of chronic conditions between childhood and adoles-cence among extremely low-birth-weight children. JAMA 2011;306:394e401.

[16] Moster D, Lie RT, Markestad T. Long-term medical and social consequences ofpreterm birth. N Engl J Med 2008;359:262e73.

[17] Dutton GN. The spectrum of cerebral visual impairment as a sequel to pre-mature birth: an overview. Doc Ophthalmol 2013;127:69e78.

[18] Doyle LW, Anderson PJ. Improved neurosensory outcome at 8 years of age ofextremely low birthweight children born in Victoria over three distinct eras.Arch Dis Child Fetal Neonatal Ed 2005;90:F484e8.

[19] Hack M, Fanaroff AA. Outcomes of extremely-low-birth-weight infants be-tween 1982 and 1988. N Engl J Med 1989;321:1642e7.

[20] Doyle LW, Roberts G, Anderson PJ. Changing long-term outcomes for infants500e999 g birth weight in Victoria, 1979e2005. Arch Dis Child FetalNeonatal Ed 2011;96:F443e7.

[21] O’Connor AR, Stephenson T, Johnson A, Tobin MJ, Moseley MJ, Ratib S, et al.Long-term ophthalmic outcome of low birth weight children with andwithout retinopathy of prematurity. Pediatrics 2002;109:12e8.

[22] Darlow BA, Clemett RS, Horwood LJ, Mogridge N. Prospective study of NewZealand infants with birth weight less than 1500 g and screened for reti-nopathy of prematurity: visual outcome at age 7e8 years. Br J Ophthalmol1997;81:935e40.

*[23] Marlow N, Wolke D, Bracewell MA, Samara M. Neurologic and developmentaldisability at six years of age after extremely preterm birth. N Engl J Med2005;352:9e19.

[24] Haugen OH, Nepstad L, Standal OA, Elgen I, Markestad T. Visual function in 6to 7-year-old children born extremely preterm: a population-based study.Acta Ophthalmol 2012;90:422e7.

[25] Molloy CS, Wilson-Ching M, Anderson VA, Roberts G, Anderson PJ, Doyle LW.Visual processing in adolescents born extremely low birth weight and/orextremely preterm. Pediatrics 2013;132:e704e12.

[26] Davis NM, Doyle LW, Ford GW, Keir E, Michael J, Rickards AL, et al. Auditoryfunction at 14 years of age of very-low-birthweight. Dev Med Child Neurol2001;43:191e6.

[27] Herold B, Hohle B, Walch E, Weber T, Obladen M. Impaired word stresspattern discrimination in very-low-birthweight infants during the first 6months of life. Dev Med Child Neurol 2008;50:678e83.

[28] Jansson-Verkasalo E, �Ceponien R, Valkama M, Vainionpää L, Laitakari K,Alku P, et al. Deficient speech-sound processing, as shown by the electro-physiologic brain mismatch negativity response, and naming ability in pre-maturely born children. Neurosci Lett 2003;348:5e8.

[29] McCoy TE, Conrad AL, Richman LC, Nopoulos PC, Bell EF. Memory processesin learning disability subtypes of children born preterm. Child Neuropsychol2013;19:173e89.

[30] Fewtrell MS, Williams JE, Singhal A, Murgatroyd PR, Fuller N, Lucas A. Earlydiet and peak bone mass: 20 year follow-up of a randomized trial of earlydiet in infants born preterm. Bone 2009;45:142e9.

[31] Hovi P, Andersson S, Järvenpää A, Eriksson JG, Strang-Karlsson S, Kajantie E,et al. Decreased bone mineral density in adults born with very low birthweight: a cohort study. PLoS Med 2009;6. http://dx.doi.org/10.1371/journal.pmed.1000135.

[32] Kovacs CS, Fuleihan Gel H. Calcium and bone disorders during pregnancy andlactation. Endocrinol Metab Clin North Am 2006;35:21e51 [v].

[33] Brooke OG, Lucas A. Metabolic bone disease in preterm infants. Arch DisChild 1985;60:682e5.

[34] Rauch F, Plotkin H, DiMeglio L, Engelbert RH, Henderson RC, Munns C, et al.Fracture prediction and the definition of osteoporosis in children and ado-lescents: the ISCD 2007 Pediatric Official Positions. J Clin Densitometry2008;11:22e8.

[35] Crandall CJ, Newberry SJ, Diamant A, Lim Y, Gellad WF, Suttorp MJ, et al.Treatment to prevent fractures in men and women with low bone density orosteoporosis: update of a 2007 report; 2012. Rockville, MD.

[36] Barker DJ, Eriksson JG, Forsen T, Osmond C. Fetal origins of adult dis-ease: strength of effects and biological basis. Int J Epidemiol 2002;31:1235e9.

[37] Hofman PL, Regan F, Cutfield WS. Prematurity e another example of perinatalmetabolic programming? Horm Res 2006;66:33e9.

[38] Forsen T, Eriksson JG, Tuomilehto J, Osmond C, Barker DJ. Growth in uteroand during childhood among women who develop coronary heart disease:longitudinal study. BMJ (Clin Res Ed) 1999;319(7222):1403e7.

*[39] Eriksson JG, Forsen T, Tuomilehto J, Winter PD, Osmond C, Barker DJ. Catch-up growth in childhood and death from coronary heart disease: longitudinalstudy. BMJ 1999;318(7181):427e31.

[40] Ravelli ACJ, van der Meulen JHP, Michels RPJ, Osmond C, Barker DJP, Hales CN,et al. Glucose tolerance in adults after prenatal exposure to famine. Lancet1998;351(9097):173e7.

*[41] Hovi P, Andersson S, Eriksson JG, Järvenpää A, Strang-Karlsson S, Mäkitie O,et al. Glucose regulation in young adults with very low birth weight. N Engl JMed 2007;356:2053e63.

[42] Ralser E, Mueller W, Haberland C, Fink F, Gutenberger K, Strobl R, et al.Rehospitalization in the first 2 years of life in children born preterm. ActaPaediatr 2012;101:e1e5.

[43] Doyle LW, Ford G, Davis N. Health and hospitalisations after discharge inextremely low birth weight infants. Semin Neonatol 2003;8:137e45.

[44] Carbonell-Estrany X, Quero J, Bustos G, Cotero A, Domenech E, Figueras-Aloy J, et al. Rehospitalization because of respiratory syncytial virus infectionin premature infants younger than 33 weeks of gestation: a prospectivestudy. IRIS Study Group. Pediatr Infect Dis J 2000;19:592e7.

[45] Saigal S, Stoskopf BL, Streiner DL, Burrows E. Physical growth and currenthealth status of infants who were of extremely low birth weight and controlsat adolescence. Pediatrics 2001;108:407e15.

[46] Westrupp EM, Lucas N, Mensah FK, Gold L, Wake M, Nicholson JM. Com-munity-based healthcare costs for children born low birthweight, pretermand/or small for gestational age: data from the Longitudinal Study ofAustralian Children. Child Care Health Dev 2013 Mar 5 [Epub ahead ofprint].

[47] Larroque B, Ancel P, Marret S, Marchand L, André M, Arnaud C, et al. Neu-rodevelopmental disabilities and special care of 5-year-old children bornbefore 33 weeks of gestation (the EPIPAGE study): a longitudinal cohortstudy. Lancet 2008;371(9615):813e20.

G. Roberts, J.L.Y. Cheong / Seminars in Fetal & Neonatal Medicine 19 (2014) 118e124124

[48] Crump C, Winkleby MA, Sundquist K, Sundquist J. Preterm birth and psy-chiatric medication prescription in young adulthood: a Swedish nationalcohort study. Int J Epidemiol 2010;39:1522e30.

[49] National Institutes of Health; National Heart, Lung, and Blood Institute. Thefourth report on the diagnosis, evaluation, and treatment of high bloodpressure in children and adolescents. Bethesda, MD: NIH; May 2005.

[50] Bentley A, Filipovic I, Gooch K, Busch K. A cost-effectiveness analysis ofrespiratory syncytial virus (RSV) prophylaxis in infants in the UnitedKingdom. Health Econ Rev 2013;3:18.

[51] Schulzke SM, Trachsel D, Patole SK. Physical activity programs for promotingbone mineralization and growth in preterm infants. Cochrane Database SystRev 2007;(2):CD005387.

lable at ScienceDirect

Seminars in Fetal & Neonatal Medicine 19 (2014) 125e130

Contents lists avai

Seminars in Fetal & Neonatal Medicine

journal homepage: www.elsevier .com/locate/s iny

Review

Functional outcomes of very premature infants into adulthood

Saroj Saigal*

Neonatal Follow-up Program, McMaster University, Hamilton, Ontario, Canada

Keywords:AdulthoodFunctional abilitiesQuality of lifeVery low birth weightVery preterm

* Address: 1280 Main St West, Room HSC 4F, HamilTel.: þ1 905 521 2100x76959; fax: þ1 905 521 5007.

E-mail address: saigal@mcmaster.ca.

1744-165X/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.siny.2013.11.001

s u m m a r y

The outcomes of very low birth weight survivors born in the early post-neonatal intensive care era havenow been reported to young adulthood in several longitudinal cohort studies, and more recently fromlarge Scandinavian national databases. The latter reports corroborate the findings that despite disabil-ities, a significant majority of very low birth weight survivors are leading productive lives, and arefunctioning better than expected. This is reassuring, but there are still concerns about future psycho-pathology, cardiovascular and metabolic problems as they approach middle age. Although these findingsmay not be directly applicable to the current survivors of modern neonatal intensive care, they doprovide a yardstick by which to project the outcomes of future survivors until more contemporaneousdata are available.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Neonatal intensive care started in the late 1960s in mostindustrialised countries. The next decade was a transitional periodwhen neonatal intensive care units (NICUs) were being established.It was not until the 1980s that survival of very premature infantsstarted to improve and approached 50%. Reports of the outcomes atadulthood of very low birth weight (VLBW) and extremely lowbirth weight (ELBW) infants started to emerge in the early 21stcentury [1,2].

An important reason to determine the life course of these high-risk infants is that children are moving targets, and their outcomes,experiences and expectations change over time [3]. Many limita-tions stabilise, or improve, and newer problems may emergedepending on the academic and social challenges that they mayface. Further, with increasing age, there are fears that they mayencounter a higher prevalence of cardiovascular and metabolicproblems than the normal term population [4e9]. Thus, theemergence of problems is age-dependent and not necessarily cu-mulative. Physicians who will assume their subsequent care needto be aware of the special challenges that the ‘aging’ prematureinfants may face in the future, some of which might still beunknown.

This issue is devoted to the long-term outcome of the tiniest ormost immature babies. Several eminent international investigatorsare presenting a broad array of outcomes. Althoughmany outcomes

ton, Ontario, Canada L8S 4K1.

All rights reserved.

are interrelated, in the interest of avoiding overlap, this chapter willfocus on the residual limitations and challenges, adult role func-tioning, life achievements, social functioning, and self-perceivedquality of life (QoL). Since there are very few studies of exclu-sively ELBW infants, this chapter will also include former VLBW orvery preterm (VP) infants <29 weeks of gestation who havereached adulthood.

2. Methodological limitations

Due to the lack of ultrasound confirmation to determine theaccuracy of gestational age, cohort studies from the earlier era havereported the outcomes mainly by birth weight categories of<1001 g or 1000e1500 g. The bias in this artificial cut-off is wellrecognised, and the assumption is that themajority of ELBW infantswere<29 weeks of gestation. The incidence of small for gestationalage varied between the diverse populations and by the growthcurves used, and ranged between 18% and 24%. The longitudinalcohort studies reported are from a single hospital, or from regionalcentres, and may or may not include children with disabilities.Because of the low survival in that era, there are limited studies ofexclusively ELBW infants to adulthood. The current studies ofextremely preterm infants are gestational-age-based, and the old-est will soon be reaching adulthood. Outcomes from the earlierstudies will therefore be reported by birth weight, unless specifiedotherwise.

In the last decade, there have been several studies of prematureinfants to adulthood using two different designs. Until recently, themajority of studies to adulthood were descriptive cohort studieswith matched controls that were followed longitudinally. These

S. Saigal / Seminars in Fetal & Neonatal Medicine 19 (2014) 125e130126

studies provide meticulously collected information using validatedstandardised tests, self-completed questionnaires, and interviews,and allow us to observe changes in outcomes over time. On thenegative side, they are time-consuming and very expensive toconduct, attrition rates can be quite high, and they are limited bysmall sample sizes that do not permit further exploratory analyses[2,10]. However, the information collected is rich, and provides amore accurate account of the diagnoses and severity of disabilities,behavioral and emotional problems, and self-perception of thehealth-related quality of life (HRQL).

Recently, the Scandinavians have taken a more innovativeapproach by using data from large epidemiological National Reg-istries that have resulted in an explosion of publications to adult-hood [11e14]. These databases have unique identifiers that linkbirth data with subsequent vital statistics, allowing a wealth ofinformation to be collected on various aspects of the individual. Notonly are they cheaper than longitudinal cohort studies, but alsothere are minimal losses to follow-up. These large sample sizesprovide the power for subgroup analyses, exploration of con-founders, and the impact of gestational age gradients on outcomes(the gestational ages in these large databases before the use ofroutine ultrasound may also be imprecise, particularly for infants<30 weeks of gestation) [14,15]. The survival rates of very prema-ture infants in Europe in the 1970s and 1980s was lower than forthe corresponding gestational ages in North American cohorts,partly because of differences in intervention policies for offeringintensive care at lower gestational ages [16]. This strategymay haveresulted in both lower rates of survival and lower prevalences ofdisabilities.

Additionally, some concerns have been expressed about theaccuracy of the diagnosis of cerebral palsy (CP) in the Danish Na-tional Patient Register [14]. Other limitations include the inabilityto elicit the personal perspectives of individuals in terms ofemotional and mental health, and quality of life. Also, with largesample sizes, it is necessary to temper interpretations of statisticalsignificance with questions about clinical relevance [2,10]. Never-theless, these studies provide valuable information that can becomplementary to support, or to refute, the findings in the smallersample studies.

Interestingly, social scientists are questioning the very definitionof adulthood, and it appears that these days adulthood does notnecessarily begin where adolescence ends. Arnett [17] has coined anew phase in life, termed ‘emerging adulthood’, for the periodbetween the late teen years through to the early 20s, when in-dividuals may have shed dependency, but have not yet assumed thefull responsibilities of adulthood according to societal expectations.In fact, the age of ‘young adulthood’ has shifted to beyond 25 years.It is well accepted that even when compared to 30 years ago, theresponsibilities of adulthood in terms of employment, marriage,and parenthood are considerably delayed. This poses a problem asseveral published follow-up studies have reported adult outcomesstarting from age 18 years upwards. For the purpose of the currentreport, only studies above a mean age of 19 years will beconsidered.

3. Neurodevelopmental outcomes

3.1. Neurosensory impairments

Overall, there appears to be stability in the incidence of neuro-logical impairments from school age to adulthood [18]. Data fromboth cohort studies and National Registries confirm that there is agradient in all neurosensory impairments based on gestational age:higher impairments are found at lower gestational ages [11,12].These impairments include CP, blindness, deafness, or cognitive

deficits, and some also include seizures/epilepsy. CP is one of themost common disabilities among premature survivors in adult-hood. Fortunately, despite the improved survival of very immatureinfants, the prevalence of CP has been declining in recent years[16,19]. In addition, the severity of functional limitations has alsoimproved, reportedly due to the reduction of grade III cystic peri-ventricular leukomalacia [16]. It remains to be seen if these currentimprovements in brain pathology translate into better outcomes forthe future.

Using the Norwegian Medical Birth Registry of births between1967 and 1983, Moster et al. [11] conducted a large-scale study ofpremature infants with a wide range of gestational ages, whowere followed to a minimum age of 20 years. They reported therate of CP for infants between 23 and 27 completed weeks ofgestation of 9.1% compared to 0.1% in those born �37 weeks ofgestation [relative risk (RR): 78.9; 95% confidence interval (CI):56.1e110.0] [11].

Saigal et al. [20] reported an overall rate of 27% with neuro-sensory impairments among ELBW young adults (10% had multipleimpairments) at a mean age of 23 years, compared to 2% in normalbirth weight (NBW) controls. Of these, 13.4% had CP (2.6% werenon-ambulatory), compared to 0.8% of controls. In another Cana-dian study of ELBW young adults at age 19 years, the overall rate ofneurosensory impairments was 12%, of which 7% had CP [21]. Hacket al. [22] reported a CP rate of 6.2% among VLBW adults at age 20years. The prevalence of CP was 6.9% in VLBW males at the medicalexamination for the Swedish National Service [23].

3.2. Ophthalmological problems

Most studies report a high prevalence of various ophthalmo-logical problems after preterm birth that is dependent on both thesurvival rates and the level of immaturity. Before the imple-mentation of the National Screening Programs for ROP and theadvent of laser therapy, the prevalence of blindness among ELBWyoung adults ranged between 0% [21] and 13% [24] (of these, 7% hadbilateral blindness). Moster et al. [11] reported a combined inci-dence of blindness, low vision and hearing loss of 4.1% for thoseborn between 23 and 28weeks of gestation. Blindness was lower inVLBW infants: 1.6% (0.4% bilateral) in the Cleveland cohort [22], andseverely reduced vision was present in 7% of Swedish VLBW 19-year-old males [23]. Errors of refraction, however, are very com-mon, and at young adulthood 64% of ELBW versus 37% NBW sub-jects required prescription glasses [24].

A serious finding that bears highlighting is the risk of late retinaldetachment (RD) reported in 4% of 149 ELBW (95% CI: 0.87, 7.19)young adults in their early 20s [24]. In addition, three other youngadults had retinal tears that required surgery. Further follow-upinto their 30s in the ongoing study elicited a few more cases ofRD as well as development of cataracts in two cases (S. Saigal et al.,unpublished data). These findings are limited by the small samplesize, but have not been reported in other longitudinal cohortstudies. However, a recent Swedish population-based registry of 3million children and young adults found late RD in 42 cases out of20 470 subjects (0.2%) born before 32 weeks of gestation [25].Compared to subjects born at term, the adjusted hazard ratio (HR)for RD for infants born <28 weeks of gestation was 19.2 (95% CI:10.3e35.8); it was lower for those born between 28 and 31weeks ofgestationwith HR of 4.3 (95% CI: 2.7e6.9). The gestational age of thecohort followed by Saigal et al. [24] was likely lower than that in theSwedish report, accounting for the higher prevalence of RD in theformer study. These findings suggest a causal relationship betweenshort gestational age and the risk of RD. Despite careful monitoring,there is still a 3e9-fold increase in the risk of RD in the currentsurvivors born extremely prematurely. As the risk of retinal

S. Saigal / Seminars in Fetal & Neonatal Medicine 19 (2014) 125e130 127

detachment increases with age, it is imperative that these youngadults be followed through life.

3.3. Audiological problems

The reported incidence of deafness is quite variable and rangesfrom 0% [24] to 7% [21] in ELBW young adults. The incidence ofdeafness in VLBW young adults was 1.2% in the Cleveland study[22], 7.4% in the Dutch study [26], and 5.8% in Swedish VLBWmales[23].

4. Adult functioning

4.1. Educational achievements

In terms of school completion, a gestational age gradient wasagain observed in the Norwegian study [11]: at 23e27 completedweeks of gestation 68% had completed high school, and 25% hadcompleted a bachelor’s degree: the corresponding figures for 28e30 weeks of gestation were 70% and 28%, compared with 75% and35% for those born at term. Although Lefebvre et al. [21] reportedthat fewer ELBW subjects had obtained a secondary schooldiploma compared to normal birth weight group (56% vs 85%),this was related to the fathers’ socio-economic score. Somestudies have reported no statistically significant differences be-tween ELBW and controls in the proportion that graduated fromhigh school (82% vs 87%) [20]. Sex differences were observedwithin this ELBW group, with a higher proportion of male par-ticipants with less than high school education (25% vs 10%,P ¼ 0.01), and fewer males enrolled in, or graduated from, post-secondary education (49% vs 71%, P ¼ 0.01). Also, a significantlylower proportion of ELBW young adults was enrolled in post-secondary education (23% vs 37%, P ¼ 0.01). A similar findingwas reported in the Swedish National cohort study [12], with 26%of those born with a gestational age of �28 weeks who hadcompleted university education, compared to 38% of those born atNBW; there were no differences in high school attainment bygestational age. In a much smaller, geographically defined Swed-ish cohort born in the late 1980s, informationwas elicited throughmailed questionnaires at age 20 years. No differences were notedin education, occupation, or social situation compared to controls[27]. Similar to the findings by Saigal et al. [20], Hack et al. [22]reported a female advantage, with 66% of VLBW males and 81%of VLBW females who completed high school, compared with 75%of term males and 90% of term females. Ironically, the rates ofcompletion of high school education among ELBW young adults inthe Ontario study [20] and in a Minnesota single hospital report ofbirths between 23 and 26 weeks of gestation [28] were similar tothat of the NBW controls in the Cleveland study [22], demon-strating the powerful effect of social class on educationalattainment.

4.2. Employment

As a group, adults born very preterm or very low birth weighthave higher rates of unemployment and lower net income, partly asa consequence of being disabled. Moster et al. [11] showed thatalthough unemployment rates were not different, Norwegianyoung adults born between 23 and 27 completed weeks of gesta-tion had lower job-related income (23%) than those born at term(20%) and a higher proportion received social security benefits(19.9% versus 17.6%), but the differences were not statistically sig-nificant (RR: 1.2; 95% CI: 0.9e1.5). When people with disabilitieswere excluded, there was still a significant, though weaker, asso-ciation between lower gestational age overall and the proportions

who attained a higher education, obtained a better-paying job, orwho received social benefits. Similar findings were reported fromthe Swedish National Cohort study [12]. A higher proportion ofadults born between 24 and 28 weeks of gestation compared withthose born at term were receiving social welfare (5% vs 1.8%) anddisability assistance (0.6% vs 0.1%). However, overall a significantproportion of all those born between 24 and 28 weeks of gestationwere employed and more were paying taxes than receivingbenefits.

At a mean age of 23 years, 48% of Ontario ELBW subjects vs 57%of controls had permanent employment, and there were no dif-ferences in occupational prestige between the groups [20].

However, in a sub-analysis, a higher proportion of ELBW thanNBW participants were neither employed, nor at school (26% vs15%, P ¼ 0.02), largely due to chronic illnesses or permanentdisability (46% vs 15%, P ¼ 0.03). Female ELBW subjects were moredisadvantaged in terms of employment. In the Dutch study [26],three times as many VLBW 19-year-olds than controls were neitheremployed nor at school (7.6% and 2.6%, respectively).

4.3. Living arrangements, dating and marriage

A higher proportion of 24e28-week gestation young adults thancontrols were living at home with their parents between 23 and 29years of age in the Swedish National Study (18.3% vs 15.0%,P ¼ 0.001).12 In the Helsinki study, VLBW young adults who werefree of major disabilities (age range: 18e27 years) were 1.7-foldmore likely to be living in the parental home than controls (95%CI: 1.01e3.05; P ¼ 0.05); this association was significant only formen [29]. Interestingly, although Finnish and Canadian cohorts hadsimilar socio-demographic background and access to health care,no statistically significant differences were observed in the pro-portion of ELBW young adults compared with controls still living athome (55% vs 47%) in the latter study [20].

Moster [11] reported that a lower proportion of Norwegianyoung adults born between 23 and 27 weeks of gestation thancontrols were either married or cohabiting (10.0% vs 18.3%,respectively; RR: 0.7; CI: 0.5e1.0). This was also true for theSwedish [12] and the Finnish cohorts [29]. Further, several studieshave reported that VLBW young adults were less likely to expe-rience sexual intercourse, and they had fewer sexual partnerscompared with controls [22,29,30]. In another Swedish regionalcohort study of infants born between 1987 to 1988, no differenceswere observed at age 20 years between VLBW and NBW youngadults in employment, living independently or cohabiting, aselicited through postal questionnaires [27]. Cooke [31] also re-ported no differences in sexual experiences or intimate re-lationships between the VLBW and control subjects for eithergender.

4.4. Reproduction

Who would have thought that one day we would be discussingreproduction as an outcome in our extremely premature survivors?Biological parenthood was reported in 28.7% of subjects born 23e27weeks, but was lower than the rate of 43.1% in term controls (RR:0.8; 0.6e1.0) in the Norwegian study [11]. Hack et al. [22] reportedfewer pregnancies in women, but not in the partners of men, at age20 years. Saigal et al. [20] reported no differences in reproductiverates or parenthood in a longitudinal cohort study of ELBW infantscompared to term-born controls. However, the power to detectsuch differences was limited by the small sample sizes and theyoung age of the participants.

The Norwegian Medical Birth Registry study [13] of births be-tween 1967 and 1976 reported strong evidence of lower

S. Saigal / Seminars in Fetal & Neonatal Medicine 19 (2014) 125e130128

reproductive rates for both men and women who had been bornvery preterm. Only 25% of women and 13.9% of men who had beenborn between 22 and 27 weeks of gestation had subsequentlyreproduced, in contrast to 68% of women and 50% of men born atterm. Interestingly, only female premature participants were atincreased risk of recurrent preterm birth, with a dose responsebased on the degree of maternal prematurity. The risk of having apreterm offspring was 14% for women born between 22 and 27weeks of gestation compared to 6.4% for NBW women. A subse-quent Swedish population-based registry [32] has confirmed theabove finding of reduced probability of reproduction by very pre-mature males and females (HR: 0.78; 95% CI: 0.70e0.86 for males;HR: 0.81; 95% CI: 0.75e0.88 for females). Without these large da-tabases, the evidence for reduced reproductive rates would havebeen very weak.

Apart from the biological and physiological factors for thereduced reproductive rates, psychosocial, nutritional, and eco-nomic factors may affect the ability to reproduce. As Swamy et al.[13] point out, survivors born pretermmay have more difficulties infinding a partner because of medical factors, disabilities andcognitive deficits. Moreover, several studies have shown that fewerpremature infants, and particularly those with disabilities, experi-ence sexual intercourse [22,29,30,33]. Also, in western societies themean age of parenting is increasing, and this may additionallycontribute to decreased fertility and possible childlessness in thefuture.

4.5. Social relationships and risk-taking behaviours

A Swiss study [34] reported poorer social relations in VLBWyoung adults: there were fewer visits from friends and family(P ¼ 0.04); VLBW young adults spent less time with friends(P ¼ 0.001), and had lower mean number of friends than controls.Cooke [31] reported that VLBW young adults in Liverpool partici-pated similarly to controls in social activities. Saigal et al. [33] re-ported similar peer, partner and family relationships, includingmean number of friends, and involvement in clubs and social ac-tivities, among ELBW young adults and term controls. The rates ofoverall criminality in the Norwegian National Study were similar inthose born 23e27 weeks (9.6%) than in those born at term (8.7%;RR: 1.1; 0.8e1.6) [11].

Most studies, however, are consistent in the finding of lowerrisk-taking behaviours among premature young adults comparedto their term-born peers [12,22,30,31,33]. These include a lowerproportion that used drugs or consumed alcohol, smoked cigarettesor marijuana, exhibited delinquent behaviours, rates of crimeconviction or incarceration, or contact with police. It is not entirelyclear, but the reasons for these behaviours may be due to increasedparental monitoring [22], shy personality [35,36], and possiblyfewer social opportunities.

5. Functional status and quality of life

How do we define functional outcomes, and what variables andmeasurement tools should we consider in assessing the same? Inthe past, the ability of a person to perform the routine activities ofdaily living, as well as leisure and socially allocated roles, wasconsidered as an acceptable functional outcome. Functional statusis therefore a way of reporting the limitations resulting from adisease or illness in an ‘objective’ manner.

Although most studies show that the general health of formerpremature infants improves by adulthood, they are still left withsome residual functional limitations. These include visual andhearing deficits (described above), dexterity and clumsiness; and,in a minority, reduced self-care abilities [24]. These limitations

remained significant, even when young adults with disabilitieswere excluded. Consistent with this study, several investigatorshave reported that VLBW young adults lead a more sedentarylifestyle, and have limited participation in strenuous physical ac-tivities [22e24,37].

The above traditional biomedical model of reporting functionallimitations per se is no longer considered sufficient. With the cur-rent broader approach in defining patient outcomes, this definitionhas been expanded to combine biomedical factors with social sci-ence approaches to obtain a more holistic picture of an individual’sfunctioning. It is therefore recommended that the ‘subjective’views e ‘how a person feels’ e should be elicited directly from theperson most affected by the process. Yet, there are few studies thataddress the personal perspectives of the individual in question atadulthood.

Several reviews of the definition and conceptual framework ofQoL have been published [38,39]. However, in the context of healthcare, most studies report the health-related quality of life (HRQL),that allows an individual to implicitly weigh aspects of their healthand provide a personal valuation of the same. It is this personalvaluation that distinguishes HRQL from other measures of healthand function [38,39].

Many different techniques and measures have been used toassess the HRQL of premature infants at adulthood. Using astructured questionnaire that included both objective and sub-jective measures, Danish investigators [40,41] reported that onboth these variables, non-impaired VLBW 18e20-year-olds weresimilar to their NBW peers. However, the QoL was lower in thosewith impairments. Three studies used the SF36, a mailed healthquestionnaire. In the British study [31], VLBW 19e22-year-oldmales rated themselves lower only in physical functioning andhealth perception compared to their NBW peers. The Swiss study[37] reported no differences in both physical and psychologicalfunctioning in their 23-year-old young adults <1250 g birth weightcompared to their peers born at NBW. In another Swiss study [42]of 55 ELBW young adults, lower self-perceived mental health andlower socio-emotional functioning was noted compared to Euro-pean norms.

HRQL was assessed longitudinally in Canadian ELBW subjectsat adolescence [43] and at young adulthood [44], using directlymeasured preferences with the standard gamble method. Con-trary to expectations, there were no differences in the meanutility scores between the overall ELBW young adults and theirNBW counterparts; nor were there any differences betweenthose with and without disabilities [44]. The CollaborativeProject on Preterm and Small for Gestational Age Infants (POPS)in The Netherlands used the HUI3 utility formula applied to theself-reported health status of VLBW young adults at age 19 years[45]. They reported fairly high HRQL (0.87, on a scale of 0e1),and remarkable stability from adolescence to young adulthood.HRQL was negatively associated with internalising problems andwith physical disabilities. Further follow-up to age 28 yearsshowed that the HRQL increased from 0.83 at age 19 years to0.85 at age 28 years [46]. A major limitation was that only one-third of the cohort participated at age 28 years, and data for theremaining two-thirds were imputed. Nor did they have a controlgroup.

Thus, despite impairments and varying methods of elicitingQoL, it appears that former premature young adults reporttheir QoL to be fairly high d comparable to their term-borncounterparts. This phenomenon of discordance between func-tional limitations and high valuation has been termed ‘disabilityparadox’ by Albrecht and Devleiger [47], and has been reportedin many non-premature adult populations. The explanationfor the discordance remains elusive. Self-reported QoL

Practice points

� Residual neurodevelopmental disabilities and functionallimitations persist to adulthood.

� Most studies report lower educational attainment andlower income.

� The proportions married or cohabiting and reproductiverates are lower.

� The quality of life is similar to term-born peers.� There is early evidence of more psychopathology, higherblood pressure and insulin resistance.

Research directions

� Longer-term follow-up to monitor the cardiovascular andmetabolic sequelae until middle age.

� Monitor and treat psychopathology, particularly infemales.

� Collect data tomiddle adulthood on the current survivors.

S. Saigal / Seminars in Fetal & Neonatal Medicine 19 (2014) 125e130 129

provides complementary information to traditional biomedicaloutcomes, and should be elicited from individuals to obtain theirpersonal perceptions, and to tailor their care to their perceivedneeds.

6. Prematurity, aging, and mortality in young adulthood

In a Norwegian meta-analysis, an inverse relationship wasfound between birth weight and mortality in adulthood [48].There was a 6% lower risk of deaths per kilogram among higherbirth weights (adjusted HR: 0.94; 95% CI: 0.92e0.97). However,the first large study to gauge the effects of gestational age onmortality at young adulthood was derived from the National Birthand Death Registry from Sweden [15]. Included in this study weresingletons born between 1973 and 1979 and who survived thefirst year of life. Mortality rates (per 1000 person-years) bygestational age at birth were 0.94 for 22e27 weeks and 0.86 for28e33 weeks of gestation. There was an independent, stepwiseinverse relationship between gestational age and mortality, and itaffected both sexes similarly. Preterm birth was associated withincreased mortality even among those born late preterm. Theunderlying mechanisms are unknown, but may be related to fetaland postnatal nutritional abnormalities, hormonal alterations orgenetic factors. Prematurity rates in Sweden are only 5%, andmortality risk will no doubt have a larger impact in countrieswhere the prematurity rates are higher. This may be furthercompounded by the current survival of more and more immatureinfants.

7. Conclusion

Although most VLBW infants go through significant diffi-culties in childhood and adolescence, by and large, by the timethey reach adulthood, they do better than expected in terms of‘adult functioning’. Many young adults may still have chronichealth conditions and some functional limitations, but despitethat, they seem to be fairly resilient and lead relatively normallives [1e3,11,29]. A most rewarding finding is that a significantmajority of VLBW and ELBW participants rated their QoL equiv-alent to that of their peers. Certainly, although there is a gesta-tional age gradient in the incidence and severity of disabilities,there is no such difference in terms of subjective valuation oftheir QoL. Whether these findings are generalisable to the morerecent survivors remains to be seen. One might anticipate thattheir outcomes might be substantially better, as these infants arethe beneficiaries of technological and social advances in care.However, there are also some reservations as increasing numbersof extremely immature infants are surviving, and disability rateshave not declined correspondingly.

The mean age of follow-up of the studies reported to adulthoodso far is still fairly young. It is likely that the disparity in occupa-tional prestige and income may widen significantly when they arein their 30s or 40s. There is also some suggestion of higher rates ofpsychopathology as they become older [49,50]. In addition, thereare concerns that other problems, such as diabetes, hypertension,atherosclerosis and cardiovascular diseases may manifest as theyreach middle-age [4,8]. There is already some evidence of insulinresistance [5e7], and higher blood pressure among VLBW youngadults [7,9,51,52].

Thus, we need to be vigilant regarding the long-term outlookfor the former preemies in their middle age, and ensure thatfurther follow-up of the above populations is continued. Inaddition, the current cohorts who are now adolescents from therecent era of neonatal intensive care need to be monitored toadulthood.

Conflict of interest statement

None declared.

Funding sources

None.

References

[1] Hack M. Adult outcomes of preterm children. J Dev Behav Pediatr 2009;30:460e70.

*[2] Doyle LW, Anderson PJ. Adult outcome of extremely preterm infants. Pedi-atrics 2010;126:342e51.

[3] Saigal S, Rosenbaum P. What matters in the long term: reflections on thecontext of adult outcomes versus detailed measures in childhood. SeminFetal Neonatal Med 2007;12:415e22.

[4] Barker DJ, Winter PD, Osmond C, Margetts B, Simmonds SJ. Weight in infancyand death from ischaemic heart disease. Lancet 1989;2(8663):577e80.

[5] Hovi P, Andersson S, Eriksson JG, Järvenpää AL, Strang-Karlsen S. Glucoseregulation in young adults with very low birth weight. N Engl J Med2007;356:2053e63.

[6] Finken MJ, Keijzer-Veen MG, Dekker FW, Frolich M, Hille ETM, Romijn JA,et al. Preterm birth and later insulin resistance: effects of birth weight andpostnatal growth in a population based longitudinal study from birth intoadult life. Diabetologia 2006;49:478e85.

*[7] Rotteveel J, van Weissenbruch MM, Twisk JW, Delemarre-Van de Waal HA.Infant and childhood growth patterns, insulin sensitivity, and blood pressurein prematurely born young adults. Pediatrics 2008;122:313e21.

[8] Eriksson JG, Forsen T, Tuomilehto J, Winter PD, Osmond C, Barker DJ. Catch-up growth in childhood and death from coronary heart disease: longitudinalstudy. BMJ 1999;318:427e31.

[9] Dalziel SR, Parag V, Rodgers A, Harding JE. Cardiovascular risk factors at age30 following pre-term birth. Int J Epidemiol 2007;36:907e15.

[10] Saigal S, Streiner D. Commentary on “Socio-economic achievements of in-dividuals born very preterm at the age of 27e29 years”. Dev Med ChildNeurol 2009;51:848e50.

*[11] Moster D, Lie RT, Markestad T. Long-term medical and social consequences ofpreterm birth. N Engl J Med 2008;359:262e73.

*[12] Lindstrom K, Windbladh B, Haglund B, Hjern A. Preterm infants as youngadults: a Swedish national cohort study. Pediatrics 2007;120:70e7.

*[13] Swamy GK, Ostbye T, Skjaerven R. Association of preterm birth with long-term survival, reproduction, and next-generation preterm birth. JAMA2008;299:1429e36.

[14] Mathiasen R, Hansen BM, Nybo Anderson AM, Greisen G. Socio-economicachievements of individuals born very preterm at the age of 27 to 29 years: anationwide cohort study. Dev Med Child Neurol 2009;51:901e8.

[15] Crump C, Sundquist K, Winkleby MA. Gestational age at birth and mortalityin young adulthood. JAMA 2011;306:1233e40.

S. Saigal / Seminars in Fetal & Neonatal Medicine 19 (2014) 125e130130

[16] Van Haastert IC, Groenendaal F, Uiterwaal CS, Termate JUM, van der Heide-Jalving, Elsermans MJC, et al. Decreasing incidence and severity of cerebralpalsy in prematurely born children. J Pediatr 2011;159:86e91.

[17] Arnett JJ. Emerging adulthood: a theory of developments from the late teensto the twenties. Am Psychol 2000;55:469e80.

[18] Saigal S, Doyle LW. An overview of mortality and sequelae of preterm birthfrom infancy to adulthood. Lancet 2008;371:261e9.

[19] Robertson CM, Watt MJ, Yasui Y. Changes in the prevalence of cerebral palsyfor children born very prematurely within a population-based program over30 years. JAMA 2007;297:2733e40.

[20] Saigal S, Stoskopf B, Streiner D, Pinelli J, Paneth N, Goddeeris J. Transition ofextremely low-birth-weight infants from adolescence to young adulthood.JAMA 2006;295:667e75.

[21] Lefebvre F, Mazurier E, Tessier R. Cognitive and educational outcomes in earlyadulthood for infants weighing 1000 grams or less at birth. Acta Paediatr2005;94:733e40.

*[22] Hack M, Flannery D, Schluchter M, Cartar L, Borawski E, Klein N. Outcomes inyoung adulthood of very low birth weight infants. N Engl J Med 2002;346:149e57.

[23] Ericson A, Källé B. Very low birthweight boys at the age of 19. Arch Dis ChildFetal Neonatal Ed 1998;78:F171e4.

[24] Saigal S, Stoskopf B, Boyle M, Streiner D, Paneth N, Pinelli J, et al. Comparisonof current health, functional limitations and health care utilization of youngadults born with extremely low birthweight and normal birthweight. Pedi-atrics 2007;119:e562e73.

[25] Bonamy AK, Holmström G, Stephansson O, Ludvigsson J, Cnattingius S. Pre-term birth and later retinal detachment: a population-based cohort study ofmore than 3 million children and young adults. Ophthalmology 2013. pii:S0161-6420(13)00320-00325.

[26] Hille ET, Weisglas-Kuperus N, van Goudoever JB, Jacobusse GW, Ens-Dokkum MH, de Groot L, et al. Dutch Collaborative POPS 19 Study Group.Functional outcomes and participation in young adulthood for very pretermand very low birth weight infants; the Dutch project on preterm and smallfor gestational age infants at 19 years of age. Pediatrics 2007;120:e587e95.

[27] Gaddlin P-O, Finnström O, Sydsjö G, Leijon I. Most very low birth weightsubjects do well as adults. Acta Paediatr 2009;98:1513e20.

[28] Hoekstra RE, Gilmore LT, Ferrara TB, Payne NR. Early adult outcomesfollowing delivery at 23e26 weeks. J Neonatal Perinat Med 2011;4:93e9.

*[29] Kajantie E, Hovi P, Raikkonen K, Pesonen AK, Heinonen K, Järvenpää AL, et al.Young adults with very low birth weight: leaving the parental home andsexual relations e Helsinki study of very low birth weight adults. Pediatrics2008;122:e62e72.

[30] Hille ET, Dorrepaai C, Perenboom R, Gravenhorst JB, Brand R, Verloove-Vanhorick SP. Dutch POPS-19 Collaborative Study Group. Social lifestyle, risk-taking behavior, and psychopathology in young adults born very preterm orwith a very low birthweight. J Pediatr 2008;152:793e800.

[31] Cooke RW. Health, lifestyle, and quality of life for young adults born verypreterm. Arch Dis Child 2004;89:201e6.

[32] deKeyser N, Josefsson A, Bladh M, Carstensen J, Finnström O, Sydsjö G. Pre-mature birth and low birthweight are associated with a lower rate ofreproduction in adulthood: a Swedish population-based registry study. HumReprod 2012;27:1170e8.

[33] Saigal S, Stoskopf B, Pinelli J, Boyle M, Streiner D. Social functioning peer,partner and family relationships, and satisfaction with life among former

extremely low birthweight (ELBW) and normal birthweight (NBW) subjectsat young adulthood. PAS 2005 (Abstract 2125);013:209e13.33.

[34] Lund LK, Vik T, Lydersen S, Lohaugen GCC, Skranes J, Brubakk AM, et al.Mental health, quality of life and social relations in young adults born withlow birth weight. Health Qual Life Outcome 2012;10:145e56.

[35] Schmidt LA, Miskovic V, Boyle MH, Saigal S. Shyness and timidity in youngadults who were born at extremely low birth weight. Pediatrics 2008;122:e181e7.

[36] Pesonen AK, Räikkönen K, Heinonen K, Andersson S, Hovi P, Järvenpää AL,et al. Personality of young adults born prematurely: the Helsinki study ofvery low birth weight adults. J Child Psychol Psychiatry 2008;49:609e17.

[37] Baumgardt M, Bucher HU, Mieth RA, Fauchère JC. Health-related quality oflife of former very preterm infants in adulthood. Acta Paediatr 2012;101:e59e63.

[38] Saigal S, Tyson J. Measurement of quality of life of survivors of neonatalintensive care: critique and implications. Semin Perinatol 2008;32:59e66.

[39] Saigal S. Quality of life of former premature infants during adolescence andbeyond. Early Hum Dev 2013;89:209e13.

[40] Bjerager M, Steensberg J, Greisen G. Quality of life among young adults bornwith very low birthweights. Acta Paediatr 1995;84:1339e43.

[41] Dinsen S, Greisen G. Quality of life in young adults with very low birthweight. Arch Dis Child Fetal Neonatal Ed 2001;85:F165e9.

[42] Natalucci G, Becker J, Becher K, Bickle GM, Landoit MA, Bucher HU. Self-perceived health status and mental health outcomes in young adults bornwith less than 1000g. Acta Paediatr 2013;102:294e9.

[43] Saigal S, Feeny D, Rosenbaum P, Furlong W, Burrows E, Stoskopf B. Self-perceived health status and health-related quality of life of extremely lowbirthweight infants at adolescence. JAMA 1996;276:453e9.

*[44] Saigal S, Stoskopf B, Pinelli J, Streiner D, Hoult L, Paneth N, et al. Self-perceived health-related quality of life of former extremely low birthweightinfants at young adulthood. Pediatrics 2006;118:1140e8.

[45] Verrips G, Brouwer L, Vogels T, Taal E, Drossaert C, Feeny D, et al. Longterm follow-up of health-related quality of life in young adults born verypreterm or with a very low birth weight. Health Qual Life Outcome2012;10:49e56.

[46] van Lunenburg A, van der Pal SM, van Dommelen P, van der Pal-de Bruin KM,Gravenhorst J, Verrips GH. Changes in quality of life into adulthood after verypreterm birth and/or very low birth weight in the Netherlands. Health QualLife Outcome 2013;11:51.

[47] Albrecht G, Devleiger PJ. The disability paradox: high quality life against allodds. Soc Sci Med 1999;48:977e88.

[48] Risnes KR, Vatten LJ, Baker JL, Jameson K, Sovio U, Kajantie E, et al. Birth-weight and mortality in adulthood: a systematic review and meta-analysis.Int J Epidemiol 2011;40:647e61.

*[49] Boyle M, Saigal S, Stoskopf B, Pinelli J. Psychopathology in young adults bornat extremely low birth weight. Psychol Med 2011;41:1763e74.

[50] Räikkönen K, Pesonen AK, Heinonen K, Kajantie E, Hovi P, Järvenpää AL, et al.Depression in young adults with very low birth weight: the Helsinki study ofvery low birth-weight adults. Arch Gen Psychiatry 2008;65:290e6.

[51] Hack M, Schluchter M, Cartar L, Rahman M. Blood pressure among very lowbirth weight (<1.5 kg) young adults. Pediatr Res 2005;58:677e84.

*[52] Hovi P, Andersson S, Räikkönen K, Strang-Karlsson S, Järvenpää AL,Eriksson JG, et al. Ambulatory blood pressure in young adults with very lowbirth weight. J Pediatr 2010;156:54e1e9el.

lable at ScienceDirect

Seminars in Fetal & Neonatal Medicine 19 (2014) 131e135

Contents lists avai

Seminars in Fetal & Neonatal Medicine

journal homepage: www.elsevier .com/locate/s iny

Review

Parent and family outcomes following very preterm or very low birthweight birth: A review

Karli Treyvaud a,b,*

aMurdoch Childrens Research Institute, Parkville, Victoria, AustraliabDepartment of Paediatrics, University of Melbourne, Melbourne, Victoria, Australia

Keywords:FamilyOutcomesParentPreterm birthVery low birth weight

* Corresponding author. Address: Murdoch ChildRoyal Children’s Hospital, Flemington Road, ParkviTel.: þ61 3 9345 4756.

E-mail address: karli.treyvaud@mcri.edu.au.

1744-165X/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.siny.2013.10.008

s u m m a r y

Parents and the family environment have a pronounced influence on child development. For children atincreased risk such as those born very preterm (VPT) or with very low birth weight (VLBW), parent andfamily functioning can influence the child’s level of risk or resilience. This review describes parent andfamily outcomes after VPT/VLBW birth, specifically parental mental health, parenting stress and theimpact of the child on the family. Factors associated with these outcomes are examined, as well as thespecific outcomes for fathers. Overall the influence of VPT/VLBW birth on parents and the family appearsto be more pronounced in early childhood, with less influence seen by the time of adolescence. Emergingevidence suggests that fathers experience high rates of psychological distress in the first months afterVPT birth. Whereas characteristics of the VPT/VLBW child are strongly associated with parent and familyoutcomes, parent and social factors are also important influences.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Sameroff and Seifer [1] proposed that a child’s developmentalrisk relates to the ability of parents and the environment to mod-erate the effects of risk factors on the child. This statement reflectsthe widely accepted assertion that parents and the family envi-ronment have a pronounced influence on child development [2],and have the potential to provide a buffer against poor outcomesfor children at high risk. This statement is directly relevant tochildren born very preterm (VPT, born <32 weeks of gestationalage), who have an increased risk for impairments in neurosensory,physical, socialeemotional, and academic functioning [3e5].Although the research is limited, a more optimal home environ-ment has been associated with improved cognitive and socialeemotional development for VPT children, supporting this bufferingeffect [6e8].

Within an ecological framework, the daily activities and socialinteractions between family members in the home environmentare the most immediate and proximal influences on child devel-opment, but these relationships are also reciprocal [9]. Thus in or-der to understand how parents and the family may be able to buffer

rens Research Institute, Thelle, Victoria 3052, Australia.

All rights reserved.

against risk and promote optimal outcomes for VPT children, weneed to know the levels and patterns of distress and adaptivefunctioning of parents and families of VPT infants. It is importantthat parent and family outcomes are understood over the course oftime from birth to early adulthood, because in addition to the sig-nificant and stressful event of the preterm birth itself and subse-quent stay in the neonatal intensive care unit (NICU), theconsequences for the child and parents associated with VPT birthcan be ongoing, such as impairments in child neurodevelopment.

This review examines the relationship between preterm birthand parent and family outcomes over time. The focus of the reviewis on the most vulnerable infants, such as those born VPT orextremely preterm (<28 weeks of gestation). Given that for manyearlier studies (prior to the introduction of routine prenatal ultra-sound in the 1990s) birth weight was used to classify infants ratherthan gestational age, studies reporting parent and family outcomesfor very low birth weight (VLBW; <1500 g) or extremely low birthweight infants (ELBW; <1000 g) are also discussed.

2. Parental mental health outcomes

Understanding parental mental health outcomes following VPT/VLBWbirth is particularly important due to potential negative effectson children’s health and development. In the general population,maternal postnatal depression has been associated with increasedmaternal negativity, unresponsive or negative maternalechildinteraction, impairment in ability to recognize infant cues, providing

K. Treyvaud / Seminars in Fetal & Neonatal Medicine 19 (2014) 131e135132

fewer learning opportunities, more child behavior and emotionalproblems, and poorer language and cognitive development acrosschildhood and adolescence [10]. Similar associations betweenparentalmental health problems andpoorer child development havealso been specifically identified for VPT/VLBW children [6,11e13].

The majority of research in this area has focused on parental(particularly maternal) depression, with limited research on otherspecific mental health disorders such as anxiety after VPT/VLBWbirth. When studies have not focused on depression, they tend toassess overall parental psychological distress using measures suchas the Brief Symptom Inventory [14]. Meta-analyses and largecohort studies have demonstrated that obstetrical stressors (suchas preterm birth) have a modest association with postnataldepression [10,15]. In a review, Vigod et al. [16] reported thatmothers of VPT/VLBW infants had a higher risk for depressionsymptoms within the first year after preterm birth compared withmothers of term-born infants, and that higher depression scorestended to be sustained throughout this first year. One month afterbirth, rates of clinically significant depression symptoms in VPTmothers have been reported to be as high as 40% [17]. Althoughothers have reported lower rates of clinically significant depressionor psychological distress symptoms within the first 2 months afterVPT/VLBW birth (between 13% and 17%) [18,19], these rates are stillhigher than population prevalence rates of postnatal depression(w10%) [15]. This elevated level of psychological distress appears tocontinue over the early years after VPT/VLBWbirth, but less distresstends to be reported by the time the child reaches adolescence.Specifically, parents of 2-year-old VPT/VLBW children report higherrates of clinically significant symptoms of psychological distress(including symptoms of anxiety and depression) compared withparents of term-born children [13,18], withw26% of parents of VPTchildren reporting clinically significant symptoms of mental healthproblems [13]. By contrast, a few studies have followed parents ofVLBW/ELBW children through adolescence to early adulthood, andall reported no significant differences in parent psychologicaldistress for these parents compared with parents with term-bornchildren [20e22].

Although there is currently little research in the area, there isincreasing interest in parental post-traumatic stress symptoms af-ter VPT/VLBW birth, with the birth and subsequent hospitalizationof the infant experienced as a significantly traumatic event by someparents. Research suggests that mothers of VLBW infants reportsignificantly higher scores on screeningmeasures of post-traumaticstress symptoms within the initial days and weeks after the birthcompared with mothers of term-born infants [23]. This patternseems to persist over time, with parents of VPT/VLBW childrenreporting more post-traumatic stress symptoms on screeningmeasures 14e36 months after the birth than parents of term-bornchildren [23e25]. The rate of significant traumatic reactions ap-pears to be high, with Pierrehumbert et al. [25] finding that 34% (17/50) of mothers of 18-month-old VPT children demonstrated clini-cally significant post-traumatic stress symptoms comparedwith 4%(1/25) of mothers of term-born children. Further research is neededto understand the traumatic responses experienced by some par-ents after VPT/VLBW birth, and to identify the progression andinfluence of these symptoms on child and family outcomes.

Overall these results suggest that parents of VPT/VLBW childrenhave higher rates of mental health problems in the early yearscompared with mothers of term-born children; however, it isimportant to note that many of these studies used screeningmeasures of depression and distress such as the Edinburgh Post-natal Depression Scale [26] or the Hospital Anxiety and Depres-sion Scale [27], rather than clinical diagnostic tools [16]. Furtherresearch using well-validated clinical diagnostic measures is rec-ommended to confirm and extend these previous results.

There is research suggesting that depression during pregnancyis predictive of preterm birth and LBW [28], indicating that theremay be a bidirectional relationship between maternal depressionand preterm birth. As previous history of mental health problems isone of the strongest predictors of postnatal depression for allmothers [10,15], in order to estimate the relative influences of VPT/VLBW birth and prenatal mental health conditions on mentalhealth outcomes after the birth, it is important for future studies toobtain a thorough prenatal psychiatric history from parents.Several other factors have been associated with poorer mentalhealth outcomes for parents after VPT/VLBW birth. Although onereview reported that level of illness in VPT infants was not a clearpredictor of maternal depression within the first 12 months afterbirth [16], individual medical complications, such as broncho-pulmonary dysplasia in VLBW infants, have been associated withhigher levels of parental anxiety [18]. Further, parents of ELBWadolescents with a history of neurosensory impairment are morelikely to report poorer emotional well-being compared with par-ents of term-born normal birth weight (NBW) adolescents [29].Parental and environmental factors are also important, with lowersocio-economic status associated with higher psychologicaldistress in parents of ELBW adolescents [20], and reduced socialsupport associated with depression symptoms for parents of pre-term infants [16].

3. Parenting stress outcomes

The Parenting Stress Index (PSI) [30] is the most commonly usedmeasure of stress associated with parenting that has been usedwith families with VPT/VLBW children, and includes domains ofparent-related stress, child-related stress, parentechild relation-ship stress, and overall stress. The interest in stress associated withthe parenting role is based on research suggesting a negative in-fluence of high parenting stress on child outcomes for VPT children[12]. There is also a measure of parent stress associated with thephysical and psychosocial environment of the NICU (ParentalStressor Scale: Neonatal Intensive Care Unit; PSS:NICU) [31], butthe use of this measure is limited to the time during hospitalizationof the infant and is not focused on parenting the child, but ratherparent stressors from multiple sources within the NICUenvironment.

A recent meta-analysis of the relationship between pretermbirth and parenting stress reported a small but significant effectfor parents of preterm children to report higher levels of parentingstress than parents of term-born children (effect sizes rangedbetween 0.18 and 0.33) [32]. In VPT/VLBW populations, resultssuggest a pattern of higher parenting stress in the first 18 months[33e35], with a dip to similar levels to parents of term-bornchildren around 2e3 years [36]. Whereas two studies reportedlittle effect of VLBW on stress associated with being a parent at 3and 4 years [18,37], Ong and Boo [37] found that parents of 4-year-old VLBW children had higher overall parenting stress scores andhigher parenting stress associated with raising the particular childthan parents of NBW children. This suggests that VPT/VLBW mayinfluence different aspects of parenting stress (e.g. child-relatedversus parent-related stress). Specifically, the largest group dif-ferences in parenting stress associated with preterm or VPT/VLBWbirth are generally reported in child-related domains of parentingstress (such as child acceptability and demandingness) [32], orstress within the parentechild relationship, as opposed to parent-related stress such as social isolation or physical health.

Overall these results suggest that, for many parents of VPT/VLBW children, the primary source of their increased parentingstress is related to characteristics of their child that make it morechallenging to parent them, rather than personal parental distress.

K. Treyvaud / Seminars in Fetal & Neonatal Medicine 19 (2014) 131e135 133

Parenting is influenced by characteristics of both the parent andchild [38], and the increased difficulties with socialeemotional andcognitive development in VPT children [3] are likely to influenceinteraction with their parents and may lead to increased parentingstress. Research supports this theory, with many of the factorsfound to predict higher parenting stress for parents of VPT childrenidentified as child-related factors. For instance, for parents of VPTchildren, the following have been associated with higher parentingstress: surgery for retinopathy of prematurity, ‘difficult’ childtemperament [35], male infant sex, discharge home on oxygen [34],higher neonatal medical risk [18], lower birth weight [22,32,34],and poorer toddler cognitive development for boys [33]. However,it is important to note that others have found less evidence thatparenting stress for parents of VPT children is associated with childfactors such as neurodevelopmental disability [36], neonatal infantmedical illness severity [33], or child age [32]. Furthermore,maternal and environmental factors also appear to be importantpredictors of parenting stress, and include maternal depressionsymptoms [35], lower marital satisfaction [34], having twins [39],and lower maternal education [33]. Thus whereas characteristics ofthe VPT child may be important predictors of parenting stress,maternal and environmental factors are also likely to be significantinfluences. The parenting stress associated with VPT children mayalso change over time, and result from interactions between otherfactors. For example, the relationship between child temperamentand parenting stress for VLBWmothers was found to be moderatedby mothers’ child rearing attitudes, suggesting that the fit betweenmother and child was more important than characteristics of thechild alone [40]. Further, Brummelte et al. [33] suggested that thesource of stress changes for parents of VPT children over time, sothat the direct influence of medical factors such as illness severitydecline over time whereas functional outcomes (such as develop-mental delay) become more stressful for parents.

4. Family functioning outcomes and impact on the family

Some studies have looked to the family unit more broadly toexamine the impact that having a VPT/VLBW child has on familysystems. Many of these studies have used the Impact on Familyquestionnaire (IOF [41]), which is based on parental report of theinfluence the child has (whether chronically ill or not) on thesocial systems of the family, such as disruption of social relation-ships, parental coping, and financial impact. In the preschoolyears, research suggests an increased negative impact on familiesor poorer family functioning overall for those with VPT/VLBWchildren compared with term-born controls [18,36,42]. Singeret al. [43] reported that the negative impact of having a VLBWchild on the family decreases over time from infancy to adoles-cence. Consistent with this, by early adulthood, family functioningis reported to be similar between families with an ELBW or NBWchild [21].

Some studies have attempted to look more closely at specificaspects of family functioning, such as effects on the marital rela-tionship. Whereas ELBW/VLBW have been associated with maritaldiscord and separation during preschool and adolescence [29,42],some families with ELBW children also reported that their child’shealth had brought parents and the family closer [29]. By earlyadulthood, no differences in marital harmony could be attributed toELBW status [21]. Looking to family structure, parents of ELBW/VLBW children have been found to be less likely to have subsequentchildren compared with parents with NBW children [29,42], whichparents of ELBW adolescents attributed directly to their child’shealth [29]. Although no differences in financial difficulties havebeen reported between families with ELBW and NBW adolescents[29], ELBW appears to influence parental education advancement,

with mothers of ELBW children increasing their education level at aslower rate than mothers of NBW children by the time their chil-dren reached adolescence [43]. Further, parents of young adultswho were born ELBW reported that their child’s health had morenegative impact on the parents’ employment compared withfamilies with NBW young adults [21].

There appears to be an important gestational age/birth weightand medical risk gradient related to the impact on the family ofVPT/VLBW birth [22], suggesting that families with infants born atlower gestational ages [44], with lower birth weights or increasedneonatal medical risk, report a more negative impact on the family.Another study found an increased negative impact on the family forfamilies with VLBW adolescents compared with families with NBWchildren, but only for those with a birth weight <750 g (e.g. not forthose with birth weight between 750 and 1500 g) [22]. Singer et al.[43] found that compared with families with children born at term,families with VLBW infants with higher neonatal medical risk (BPD)reported a higher negative impact on the family during infancy,early childhood and adolescence, whereas families with a VLBWchild with low medical risk reported a similar impact to termfamilies. Neurodevelopmental disability or developmental delay[29,36,42,44] has also been associated with an increased negativeimpact of the VPT/VLBW child on the family, as have parent andenvironmental factors including low parental education or incomeand lower socio-economic status [22,42].

5. Outcomes for fathers after VPT/VLBW birth

The majority of research on mental health and family outcomesfor families after VPT/VLBW birth has been based on maternal self-report, with fathers rarely included. The few studies that haveattempted to capture the experiences of fathers after VPT birth havebeen based on small samples, primarily from the newborn or NICUperiod, and have used qualitative measures. However, there ispreliminary evidence that fathers also experience high rates ofdistress in the first months after birth. In 35 fathers of VPT infants,Mackley et al. [45] measured depressive symptoms on day 7, day 21and day 35 after the birth. They found that a high proportion offathers initially scored in the clinically significant range fordepression symptoms (60%), which decreased over time, with 36%reporting clinically significant symptoms by day 35. Rather thandepression symptoms in fathers being associated with their child’sillness severity, other life stressors and marital status were signif-icant predictors of depression symptoms.

Several studies have compared maternal and paternal mentalhealth and perceptions of the child and family after VPT/VLBWbirth. Overall it seems that mothers report higher levels of psy-chological distress and parenting stress, but fewer differences areobserved on parental perceptions of the family. Specifically, Doer-ing et al. [46] reported that mothers of VLBW infants reportedsignificantly more symptoms of anxiety and depression while theirinfant was in the NICU compared with fathers, with no real dif-ferences between mothers’ and fathers’ perception of family func-tioning. Furthermore, a meta-analysis reported higher levels ofparenting stress in mothers of preterm infants compared with fa-thers [32]. Interestingly, during the preschool years there is evi-dence that fathers of VLBW children rate their child has having amore positive impact on the family compared with mothers [42].

Little is known about the factors predicting paternal outcomesfollowing VPT/VLBW birth, although having a VLBW child withpoorer cognitive development has been associated with higherlevels of paternal depression symptoms and parenting stress [47]. Ithas also been suggested that, since many fathers continue paidemployment after the birth of their child, the social support theyreceive from workmates may help to buffer against poor mental

Practice points

� Parents of VPT/VLBW children report higher levels ofmental health problems, increased parenting stress andan increased negative impact on family systemscompared with parents of term-born/NBW children in theearly years.

� These initially high levels of distress, stress and negativefamily impact appear to diminish over time.

� Parents and families with VPT/VLBW children need early,specific and targeted support for themselves, and tosupport their child’s development.

� Evidence-based home and hospital early interventionprograms for parents of VPT/VLBW children exist.

Research directions

� Follow families with VPT/VLBW infants over a longer timeperiod (e.g. until early adulthood) to better understandoutcomes for these families and how this influences childdevelopment.

� Focus on risk and resilience factors associatedwith parentand family outcomes following VPT/VLBW birth.

� Include fathers in studies examiningmental health, stressand family functioning outcomes.

� Evaluate alternative ways to increase access for familiesto evidence-based early intervention programs.

K. Treyvaud / Seminars in Fetal & Neonatal Medicine 19 (2014) 131e135134

health, resulting in more positive outcomes for fathers after VPT/VLBW birth [46].

6. Conclusions and clinical implications

Within the first 2 years after VPT/VLBW birth, parents reporthigher levels of mental health problems, higher levels of stressassociated with being a parent, and an increased negative impacton family systems compared with parents of term-born or NBWchildren. Rates of psychological distress and depression appear tobe particularly high, with between 13% and 40% of mothers of VPT/VLBW infants reporting significant clinical symptoms within thefirst few months after the birth [18,48], and 26% reporting sig-nificant symptoms two years after the birth [13]. Although fewstudies to date have followed parents and families of VPT/VLBWchildren past early childhood, it appears that parent mental health,parenting stress and family functioning improve with time.However, evidence for some longer-lasting effects of VPT birth onfamilies exists, such as lower rates of mothers continuing educa-tion [43] and a more negative influence on parental employmentin later years for these families [21]. Further research followingmore recently born cohorts of children born VPT across laterchildhood and adolescence is needed, particularly as comparisonsbetween previous studies are challenging due to cohort differ-ences in gestational age, birth weight, era of birth, and smallsample sizes.

From the limited evidence available, fathers of VPT infantsexperience high rates of psychological distress in the first monthsafter birth [45], but slightly lower than for mothers [32,46]. Morestudies are needed to understand paternal outcomes not only tosupport fathers themselves, but also because fathers play animportant role in supporting mothers and promoting childdevelopment. High involvement in the family by fathers has beensuggested to increase family cohesion, reduce maternal distress,and provide a buffer for infants if the mother is experiencingclinically significant mental health problems [49], which may beimportant after VPT/VLBW birth where maternal mental health isparticularly at risk.

The characteristics of the VPT/VLBW child (such as level ofmedical illness, neurodevelopmental disability and lower birthweight) appear to be particularly strong predictors of parent andfamily outcomes [21,22,32,43]. However, this is not consistentacross all studies [16,33,36], and parental and environmental fac-tors such as lower socio-economic status or parent education[20,22,33,42], lower marital satisfaction [34], and reduced socialsupport [16] are also important predictors of parent and familyoutcomes after VPT/VLBW birth. Understanding these risk factorsand the interactions between them will assist with the develop-ment and provision of targeted and specific support for familiesafter VPT/VLBW birth.

Support for families of VPT/VLBW children is needed due to thenegative effect that high early levels of parental distress anddysfunction in families can have on the developing VPT/VLBWchild [11,12], and conversely the protective effect that a positivehome and family environment can have on the VPT/VLBW child[6e8,43]. Home- and/or hospital-based early interventions forpreterm children and their parents decrease parental stress andanxiety, and improve parenting behaviours such as sensitivity andresponsiveness, the quality of the home environment, and out-comes for infants [50e53]. Future research directed at improvingaccess to and specifically targeting evidence-based interventionsfor parents of VPT/VLBW children is needed, particularly for thosewho appear to be at highest risk for poor outcomes, such asfamilies with VPT/VLBW infants with higher medical risk or lowersocio-economic status.

Conflict of interest statement

None declared.

Funding sources

Dr Treyvaud is supported by the NHMRC Centre for ClinicalResearch Excellence in Newborn Medicine (546519) and the Mur-doch Childrens Research Institute Clinical Sciences Theme.

References

[1] Samerfoff A, Seifer R. Familial risk and child competence. Child Dev 1983;54:1254e68.

[2] Hilferty F, Redmond G, Katz I. The implications of poverty on children’sreadiness to learn. Canberra: Australian Research Alliance for Children andYouth; 2009.

[3] Anderson PJ, Doyle LW. Neurobehavioral outcomes of school-age childrenborn extremely low birth weight or very preterm in the 1990s. JAMA2003;289:3264e72.

[4] Johnson S, Marlow N. Preterm birth and childhood psychiatric disorders.Pediatr Res 2011;69:11Re8R.

[5] Milligan DW. Outcomes of children born very preterm in Europe. Arch DisChild Fetal Neonatal Ed 2010;95:F234e40.

[6] McCormick MC, Workman-Daniels K, Brooks-Gunn J. The behavioral andemotional well-being of school-age children with different birth weights.Pediatrics 1996;97:18e25.

[7] Siegel LS. Reproductive, perinatal, and environmental factors as predictors ofthe cognitive and language development of preterm and full-term infants.Child Dev 1982;53:963e73.

*[8] Treyvaud K, Inder TE, Lee KJ, Northam EA, Doyle LW, Anderson PJ. Can thehome environment promote resilience for children born very preterm in thecontext of social and medical risk? J Exp Child Psychol 2012;112:326e37.

[9] Bronfenbrenner U. Ecology of the family as a context for human develop-ment: research perspectives. Dev Psychol 1986;22:723e42.

[10] O’Hara MW, McCabe JE. Postpartum depression: current status and futuredirections. Annu Rev Clin Psychol 2013;9:379e407.

K. Treyvaud / Seminars in Fetal & Neonatal Medicine 19 (2014) 131e135 135

*[11] Gray RF, Indurkhya A, McCormick MC. Prevalence, stability, and predictors ofclinically significant behavior problems in low birth weight children at 3, 5,and 8 years of age. Pediatrics 2004;114:736e43.

[12] Huhtala M, Korja R, Lehtonen L, et al. Parental psychological well-being andbehavioral outcome of very low birth weight infants at 3 years. Pediatrics2012;129:e937e44.

[13] Treyvaud K, Anderson VA, Lee KJ, et al. Parental mental health and earlysocialeemotional development of children born very preterm. J PediatrPsychol 2010;35:768e77.

[14] Derogatis LR. The brief symptom inventory manual. Baltimore: Clinical Psy-chometric Research; 1992.

[15] Blom EA, Jansen PW, Verhulst FC, et al. Perinatal complications increase therisk of postpartum depression. The Generation R Study. Br J Obstet Gynaecol2010;117:1390e8.

*[16] Vigod SN, Villegas L, Dennis CL, Ross LE. Prevalence and risk factors forpostpartum depression among women with preterm and low-birth-weightinfants: a systematic review. Br J Obstet Gynaecol 2010;117:540e50.

[17] Davis L, Edwards H, Mohay H, Wollin J. The impact of very premature birth onthe psychological health of mothers. Early Hum Dev 2003;73:61e70.

[18] Singer LT, Salvator A, Guo S, Collin M, Lilien L, Baley J. Maternal psychologicaldistress and parenting stress after the birth of a very low-birth-weight infant.JAMA 1999;281:799e805.

[19] Pritchard MA, Colditz PB, Cartwright D, Gray PH, Tudehope D, Beller E. Six-week postnatal depression predicts parenting stress profiles in mothers ofpreterm children. J Reprod Infant Psychol 2012;30:303e11.

*[20] Moore M, Taylor HG, Klein H, Minich N, Hack M. Longitudinal changes infamily outcomes of very low birth weight. J Pediatr Psychol 2006;31:1024e35.

*[21] Saigal S, Pinelli J, Streiner DL, Boyle M, Stoskopf B. Impact of extreme pre-maturity on family functioning and maternal health 20 years later. Pediatrics2010;126:e81e8.

*[22] Taylor HG, Klein N, Minich NM, Hack M. Long-term family outcomes forchildren with very low birth weights. Arch Pediatr Adolesc Med 2001;155:155e61.

[23] Kersting A, Dorsch M, Wesselmann U, et al. Maternal posttraumatic stressresponse after the birth of a very low-birth-weight infant. J Psychosom Res2004;57:473e6.

[24] Ahlund S, Clarke P, Hill J, Thalange NKS. Post-traumatic stress symptoms inmothers of very low birth weight infants 2e3 years post-partum. ArchWomen’s Ment Health 2009;12:261e4.

[25] Pierrehumbert B, Nicole A, Muller-Nix C, Forcada-Guex M, Ansermet F.Parental post-traumatic reactions after premature birth: implications forsleeping and eating problems in the infant. Arch Dis Child Fetal Neonatal Ed2003;88:F400e4.

[26] Cox JL, Holden JM, Sagovsky R. Detection of postnatal depression: develop-ment of the 10-item Edinburgh Postnatal Depression Scale. Br J Psychiatry1987;150:782e6.

[27] Zigmond AS, Snaith RP. The hospital anxiety depression scale. Acta PsychiatrScand 1983;67:361e70.

[28] Grote NK, Bridge JA, Gavin AR, Melville JL, Iyengar S, Katon WJ. A meta-analysis of depression during pregnancy and the risk of preterm birth, lowbirth weight, and intrauterine growth restriction. Arch Gen Psychiatry2010;67:1012e24.

[29] Saigal S, Burrows E, Stoskopf BL, Rosenbaum PL, Streiner D. Impact of extremeprematurity on families of adolescent children. J Pediatr 2000;137:701e6.

[30] Abidin RR. Parenting stress index. 3rd ed. Odessa, FL: Psychological Assess-ment Resources; 1995.

[31] Miles MS, Funk SG, Carlson J. Parental Stressor Scale: neonatal intensive careunit. Nurs Res 1993;42:148e52.

*[32] Schappin R, Wijnroks L, Uniken Venema MMAT, Jongmans MJ. Rethinkingstress in parents of preterm infants: a meta-analysis. PLoS One 2013;8:e54992.

[33] Brummelte A, Grunau RE, Synnes AR, Whitfield MF, Petrie-Thomas J.Declining cognitive development from 8 to 18 months in preterm chil-dren predicts persisting higher parenting stress. Early Hum Dev 2011;87:273e80.

[34] Gray PH, Edwards DM, O’Callaghan MJ, Cuskelly M. Parenting stress inmothers of preterm infants during early infancy. Early Hum Dev2012;88:45e9.

[35] Gray PH, Edwards DM, O’Callaghan MJ, Cuskelly M, Gibbons K.Parenting stress in mothers of very preterm infants d influence ofdevelopment, temperament and maternal depression. Early Hum Dev2013;89:625e9.

[36] Treyvaud K, Doyle LW, Lee KJ, et al. Family functioning, burden and parentingstress 2 years after very preterm birth. Early Hum Dev 2011;87:427e31.

[37] Ong LC, Chandran V, Boo NY. Comparison of parenting stress betweenMalaysian mothers of four year-old very low birthweight and normalbirthweight children. Acta Paediatr 2001;90:1464e9.

[38] Bell RQ. A reinterpretation of the direction of effects in studies of social-isation. Psychol Rev 1968;75:81e94.

[39] Beer C, Israel C, Johnson S, Marlow N, Whitelaw A, Glazebrook C. Twin birth:an additional risk factor for poorer quality maternal interactions with verypreterm infants? Early Hum Dev 2013;89:555e9.

[40] Halpern LF, Brand KL, Malone AF. Parenting stress in mothers of very lowbirth weight (VLBW) and full term infants: a function of infant andbehavioral characteristics of child-rearing attitudes. J Pediatr Psychol2001;26:93e104.

[41] Stein REK, Jones Jessop D. Tables documenting the psychometric propertiesof a measure of the impact on chronic illness on a family. New York: AlbertEinstein College of Medicine; 1985.

*[42] Cronin CMG, Shapiro CR, Casiro OG, Cheang MS. The impact of very low-birth-weight infants on the family is long lasting. Arch Pediatr AdolescMed 1995;149:151e8.

[43] Singer LT, Fulton S, Kirchner L, et al. Longitudinal predictors of maternalstress and coping after very low-birth-weight birth. Arch Pediatr AdolescMed 2010;164:518e24.

[44] Lee SK, Penner PL, Cox M. Impact of very low birth weight infants on thefamily and its relationship to parental attitudes. Pediatrics 1991;88:105e9.

[45] Mackley AB, Locke RG, Spear ML, Joseph R. Forgotten parent. NICU paternalemotional response. Adv Neonatal Care 2010;10:200e3.

[46] Doering LV, Dracup K, Moser D. Comparison of psychosocial adjustment ofmothers and fathers of high-risk infants in the neonatal intensive care unit.J Perinatol 1999;19:132e7.

[47] Huhtala M, Korja R, Lehtonen L, et al. Parental psychological well-being andcognitive development of very low birth weight infants at 2 years. ActaPaediatr 2011;100:1555e60.

[48] Davis L, Edwards H, Mohay H, Wollin J. The impact of very preterm birth onthe psychological health of mothers. Early Hum Dev 2003;3:61e70.

*[49] Feldman R. Maternal versus child risk and the development of parentechildand family relationships in five high-risk populations. Dev Psychopathol2007;19:293e312.

[50] Brecht CJ, Shaw RJ, St John NH, McCue Horwitz S. Effectiveness of therapeuticand behavioral interventions for parents of low-birth-weight premature in-fants: a review. Inf Ment Hlth J 2012;33:651e65.

[51] Brett J, Staniszewska S, Newburn M, Jones N, Taylor L. A systematic mappingreview of effective interventions for communicating with, supporting andproviding information to parents of preterm infants. BMJ Open 2011;1:e000023.

[52] Spencer-Smith MM, Spittle AJ, Doyle LW, et al. Long-term benefits of home-based preventive care for preterm infants: a randomized trial. Pediatrics2012;130:1e8.

[53] Spittle AJ, Anderson PJ, Lee KJ, et al. Preventive care at home for very preterminfants improves infant and caregiver outcomes at 2 years. Pediatrics2010;126:171e8.