Research www.AJOG.org

OBSTETRICS

Birth by vacuum extraction delivery and school performanceat 16 years of ageMia Ahlberg, RNM; Cecilia Ekeus, PhD; Anders Hjern, PhD

OBJECTIVE: The aim of the present study was to investigate cognitive mean mathematics test scores than children born vaginally without

competence, as indicated by school performance, at 16 years of age,in children delivered by vacuum extraction.

STUDY DESIGN: This was a register study of a national cohort of126,032 16 year olds born as singletons, with a vertex presentation, at agestational age of 34 weeks or older, with Swedish-born parents,delivered between 1990 and 1993 without major congenital malfor-mations. Linear regression was used to analyze mode of delivery inrelation to mean scores from national tests in mathematics (40.2; scale,10-75; SD, 14.9) and mean average grades (223.8; scale, 10-320; SD,52.3), with adjustment for perinatal and sociodemographic confounders.

RESULTS: Children delivered by vacuum extraction (e0.51; 95%confidence interval [CI], e0.76 to 0.26) as well as by nonplannedcesarean section (e0.51; 95% CI,e0.82 toe0.20) had slightly lower

From the Division of Reproductive Health,Department of Women’s and Children’sHealth, (Ms Ahlberg and Dr Ekéus), and theDivision of Clinical Epidemiology, Departmentof Medicine, Karolinska Institute, and the Centerfor Health Equity Studies, Stockholm, Sweden(Dr Hjern).

ReceivedMay 20, 2013; revised Sept. 25, 2013;accepted Nov. 6, 2013.

The study was supported by grants from theSwedish Research Council.

The authors report no conflict of interest.

Presented in oral format at the First EuropeanCongress on Intrapartum Care, Amsterdam,The Netherlands, May 23-25, 2013.

Reprints: Mia Ahlberg, RNM, Departmentof Women’s and Children’s Health, Divisionof Reproductive Health, Karolinska Institutet,Stockholm, Sweden 17176.mia.ahlberg@ki.se.

0002-9378/$36.00ª 2014 Mosby, Inc. All rights reserved.http://dx.doi.org/10.1016/j.ajog.2013.11.015

instruments, after adjustment for major confounders. Mean averagegrades in children delivered by vacuum extraction were e1.05 (95%CI,e1.87 toe0.23) ande1.20 (95% CI,e2.24 toe0.16) in childrendelivered by nonplanned cesarean section compared with childrenborn vaginally.

CONCLUSION: Children delivered by vacuum extraction had slightlylower grades at age 16 years compared with those born by nonin-strumental vaginal delivery but very similar to those delivered bynonplanned cesarean. This suggests that vacuum extraction andnonplanned cesarean are equivalent alternatives for terminating de-liveries with respect to cognitive outcomes.

Key words: cesarean section, cognitive development, mode ofdelivery, school grades, vacuum extraction

Cite this article as: AhlbergM, Ekeus C, Hjern A. Birth by vacuum extraction delivery and school performance at 16 years of age. Am J Obstet Gynecol 2014;210:361.e1-8.

elivery by vacuum extraction is a

D common obstetrical procedure inthe Western world, and in many coun-tries it has replaced the use of forceps.1

Delivery by vacuum extraction hasbeen associated with both mild and

severe acute complications for the infant.Mild complications include scalp edema,retinal hemorrhage, and cephalic he-matoma. More serious complicationsinclude subgaleal hematoma, skull frac-tures, asphyxia, and intracranial hem-orrhage.1-5

Despite the risk of cerebral compli-cations, few studies have focused onlong-term consequences for childrendelivered by vacuum extraction. Thestudies conducted have not shown acorrelation between vacuum extractiondelivery and impaired development, ce-rebral sequel, vision tests, and intelli-gence.6-9 However, a majority of thesestudies used small sample sizes andmight be unrepresentative of the gen-eral population. Most studies also suf-fer from a short follow-up time and alack of adjustment for major con-founders. Large population-based studiesinvolving long-term follow-ups are thusrequired.The aim of the present study was to

investigate whether delivery by vacuumextraction influences the child’s cogni-tive competence in terms of school per-formance at 16 years of age.

APRIL 2014 Ameri

MATERIALS AND METHODS

This study was based on routinelycollected data from Swedish nationalregisters held by Statistics Sweden andthe National Board of Health andWelfare. All Swedish residents areassigned a unique 10 digit identifica-tion number at birth or immigration.These identification numbers wereused to link information from differentregister sources. The study was ap-proved by the Regional Ethics Com-mittee in Stockholm.

Study populationThe study cohort was identified in theMedical Birth Register and included allfirstborn singleton infants of nativeSwedish parents, born after gestationalweek 34þ0 during the years 1990-1993.In 1990, Sweden introduced pregnancydating by ultrasound as a routine, andchildren born after 1993 did not havetheir school grades forwarded to theNational School Register. Children de-livered before gestational week 34 wereexcluded because this is consideredinappropriate use of vacuum extractionin Sweden. For the child to be included in

can Journal of Obstetrics & Gynecology 361.e1

Research Obstetrics www.AJOG.org

the study, his or her mother had tobe alive and a resident of Swedenwhen the child reached 15 years of age(n ¼ 148,881).

From this population we excluded in-fants with a malformation with possibleclinical significance (n ¼ 4558), infantsdelivered by forceps (n ¼ 1369), andthose born in breech presentations (n ¼4399). Furthermore, we excluded chil-dren without a registered grade pointaverage (n ¼ 4691) or mathematics testscore (n ¼ 7832) at the age of 16 years.The definition of serious malformationwas based on the diagnoses made bythe attending pediatrician. It includedall malformations with the exceptionof nondescended testicles, preauricularappendage, and minor malformations offingers and toes.

The study population of 126,032children was followed up at 15-16 yearsof age, at graduation from compulsoryschool (at the end of ninth grade) be-tween 2006 and 2009.

Exposure variablesThe exposure variables, modes of de-livery, were collected from the MedicalBirth Register and categorized intononinstrumental vaginal delivery, mean-ing a vaginal delivery without the useof any instrument, vacuum extraction,nonplanned cesarean section, and plan-ned cesarean section. A cesarean sectionwas defined as nonplanned if the opera-tion was conducted after the onset of la-bor, either spontaneous or by induction,and defined as planned if it was per-formed before the onset of labor.

Pregnancy, birth, and perinatalvariablesInformation about the mother’s age atbirth, mode of delivery, smoking habitsin early pregnancy, maternal diseases,and complications during pregnancyand labor was collected from the SwedishMedical Birth Register, a national regis-ter with high-quality data on more than99% of all births in Sweden.10 From thesame register, perinatal data werecollected about sex, gestational age,birthweight, and head circumference,whether the child was small for gesta-tional age (SGA; <2 SD) or large for

361.e2 American Journal of Obstetrics & Gynecol

gestational age (LGA;>2 SD11); cephalichematoma, intracranial bleeding, centralnervous system complications, convul-sions after birth, and Apgar scores at5 minutes. Maternal diseases and preg-nancy and labor complications, as well asneonatal diagnoses, were coded accord-ing to the International Classification ofDiseases, ninth revision (ICD-9).12

Socioeconomic variablesData on parental education were col-lected from the Swedish EducationalRegister and defined as the highestformal education attained in the house-hold 1 year before the child’s graduation.The educational level was categorized byyears of education into less than 10 years(compulsory), 10-14 years (high school),and more than 14 (university). Data onsocial welfare recipiency and householddisposable income were collected fromthe Total Enumeration Income Surveyand was categorized into quintiles inwhich quintile 1 included the 20% ofchildren in families with the lowest in-comes. Residency was retrieved from theRegister of the Total Population andcategorized into big city, small city, andrural.

Maternal morbidityData from the Swedish HospitalDischarge Register 1973-2008 were usedto create dichotomized variables associ-ated with psychiatric and addictive dis-orders in the mother. Illicit drug andalcohol abuse (yes/no) indicates whetherthe mother had been discharged from aSwedish hospital with a diagnosis asso-ciated with use or abuse of alcohol orillicit drugs in ICD-9.12 Psychiatric dis-order (yes/no) shows whether themother had been discharged at least oncefrom a Swedish hospital with a diagnosisin the psychiatric chapter in ICD-9.12

Outcome variablesThe study population was followed up at15-16 years of age using 2 outcome var-iables: mean grade point average and themean score on national mathematicstests. Data were collected from the Na-tional School Register kept by StatisticsSweden. The quality of the data in theNational School Register is high, and

ogy APRIL 2014

summary statistics are published on aregular basis.13 The majority of all chil-dren (more than 90%) attend compul-sory school, which lasts 9 years, usuallybetween the ages of 7 and 16 years.

The register contains information onall school grades at the final examinationsand graduation grades for each pupilleaving Swedish compulsory school.From 1998 onward, the Swedish schoolsystem has used alphabetic grades with 4levels: fail, pass, pass with distinction,and pass with special distinction. Gradesare awarded on a 3-level scale: a pass iscounted as 10 merit points, a pass withdistinction as 15 merit points, and a passwith special distinction as 20 meritpoints. Themaximummerit rating is 320points. Before selection for upper sec-ondary schools, the pupil’s final school-leaving grade (ie, mean grade pointaverage) is calculated as the sum of thepoints of the pupil’s 16 best grades on theleaving certificate.

The second outcome, the mathe-matics test score, was based on the sumof 4 subsample national tests in mathe-matics that all pupils attending the finalyear of compulsory school write. Themaximummathematics test scores in theyears 2003-2006 ranged from 68 to 75.

Statistical methodsWe conducted linear regression analysesusing SPSS version 18.0 for Windows(SPSS Inc, Chicago, IL) to estimate crudeand adjusted mean differences in meangrade points and mean mathematicsgrades with 95% confidence intervals inrelation to mode of delivery. Nonin-strumental vaginal delivery served as thereference group in the analyses.

To control for potential confounders,we used the following 3 models: model 1was adjusted for year of graduation andsex only. For model 2 we added thefollowing socioeconomic, demographic,and maternal morbidity variables to theconfounders in model 1; maternal age,highest educational level in the household1 year before graduation, urban/ruralresidency 1 year before graduation, single-parent household, maternal smoking,maternal drug abuse, maternal alcoholabuse, and maternal psychiatric diagnosiswere added inmodel 2. In the finalmodel,

TABLE 1Perinatal, maternal, and sociodemographic characteristics in relation to mode of delivery

Study population (n [ 126,032) nNIVD, %(n [ 100,038)

VE, %(n [ 14,494)

NPCS, %(n [ 8753)

PC, %(n [ 2386)

VE plus EC, %(n [ 361)

Sex

Male 64,577 49.9 57.1 56.3 51.6 58.4

Gestational age, wks

34-36 5406 4.0 2.2 6.6 20.6 1.1

37-41 109,655 88.8 84.4 74.5 73.0 82.0

>42 10,971 7.2 13.4 18.9 6.5 16.9

�32 9522 7.9 4.4 7.4 14.4 2.8

Head circumference, cm

33-35 78,330 65.2 50.4 50.4 51.6 38.2

>36 34,408 24.4 41.4 36.5 27.5 48.8

Missing 3772 2.5 3.7 5.8 6.5 10.2

Neonatal complicationsa

Yes 3209 1.6 9.7 1.1 0.7 21.1

Maternal psychiatric morbidityb

Yes 13,314 10.5 9.9 11.7 13.0 9.1

Single household

Yes 30,156 23.8 23.8 25.1 25.5 19.1

Pregnancy complicationsc

Yes 10,938 6.9 9.2 20.8 34.3 10.5

Maternal age, y

<19 6301 5.5 3.1 3.2 2.8 2.2

20-29 76,535 76.5 68.4 65.9 58.5 69.5

30-39 25,121 17.7 27.6 29.3 33.4 27.4

>40 746 0.3 0.9 1.6 5.2 0.8

Smoking

Yes 25,669 20.2 21.1 21.4 21.8 16.3

Missing 6447 5.1 4.9 5.5 5.9 5.8

Highest educational leveld

Compulsory 12,722 10.1 9.9 10.5 11.2 8.6

High school 88,835 70.7 69.7 69.5 67.9 70.4

University 22,165 17.3 18.7 18.1 19.0 19.4

Incomed

First quintile 15,069 12.4 11.3 10.8 12.0 11.8

Second quintile 20,209 16.3 16.4 16.2 16.6 15.2

Third quintile 24,794 20.1 19.5 20.2 18.3 18.3

Fourth quintile 30,613 24.9 24.0 24.1 23.6 27.8

Fifth quintile 33,115 26.2 28.7 28.6 29.6 27.0

Ahlberg. Mode of delivery and cognitive function. Am J Obstet Gynecol 2014. (continued)

www.AJOG.org Obstetrics Research

APRIL 2014 American Journal of Obstetrics & Gynecology 361.e3

TABLE 1Perinatal, maternal, and sociodemographic characteristics in relation to mode of delivery (continued)

Study population (n [ 126,032) nNIVD, %(n [ 100,038)

VE, %(n [ 14,494)

NPCS, %(n [ 8753)

PC, %(n [ 2386)

VE plus EC, %(n [ 361)

Residencyd

Big city 50,695 39.8 41.9 42.0 40.1 37.4

Small city 53,687 42.7 42.3 40.8 43.9 46.3

Rural 21,650 17.4 15.8 17.2 16.0 16.3

NIVD, noninstrumental vaginal delivery; NPCS, nonplanned cesarean section; PC, planned cesarean section; VE, vacuum extraction; VE plus EC, vacuum extraction with a subsequent cesareansection.

a Cephalic hematoma, intracranial bleeding, convulsions, and other central nervous system disorders; b Drug abuse, alcohol abuse, and psychiatric diagnosis; c Gestational diabetes, maternaldiabetes, preeclampsia, oligohydramnion, polyhydramnion, small for gestational age, and large for gestational age; d Registered at the child’s age of 14 years.

Ahlberg. Mode of delivery and cognitive function. Am J Obstet Gynecol 2014.

Research Obstetrics www.AJOG.org

we added potential medical confoundersto the variables in model 2: SGA, LGA,oligohydramnion, polyhydramnion, mat-ernal diabetes, preeclampsia, head cir-cumference, and gestational age.

To further control for residual peri-natal confounding, subgroup analyseswere conducted on a selected populationof term children (gestational week>36),with Apgar score (>6) at 5 minutes, whowere born without a diagnosis of fetaldistress. The same linear regressionmodels were used to test for the associ-ation mode of delivery and school grades(not shown in a table).

Our main analyses are based on chil-dren who graduated from compulsoryschool with registered mean grade pointaverages and mathematics test results.The children with no available graderecords were analyzed separately toassess possible selection bias.

RESULTS

The study population included 126,032children: 11.5%were delivered by vacuumextraction, 7.0% by nonplanned cesarean,1.9% by elective cesarean, 0.3% by vac-uum extraction and subsequent cesarean,and 79.4% by noninstrumental vaginaldelivery. From 1990 to 1993, deliveries byvacuum extraction increased from 10.5%to 12.6%, whereas nonplanned cesareansections ranged between 6.9% and 7.5%.Mean mathematics test scores remainedrelatively stable over the years 1990-1993,ranging from 40.0 to 40.4. The meangrade point average increased from 218.8in 1990 to 226.0 in 1993.

361.e4 American Journal of Obstetrics & Gynecol

Table 1 showsmaternal, perinatal, andsociodemographic characteristics of thestudy population in relation to mode ofdelivery. Women whose children weredelivered by vacuum extraction weremore likely to be older, to have a highereducation, and to have a higher house-hold income thanwomenwho gave birthvaginally without instruments. Of allchildren delivered by vacuum extraction,2% had Apgar scores lower than 7 at 5minutes. More boys (57.1%) than girls(42.9%) were delivered by vacuumextraction.Table 2 shows maternal, perinatal, and

sociodemographic characteristics in re-lation to the crude mean mathematics testscore and the crude mean grade pointaverage. Differences in crude meanmathematics scores as well as in crudemean grade point averages were small inrelation to mode of delivery. Childrendelivered by vacuum extraction had thehighest crudemeanmathematics test score(40.5), whereas the lowest score (40.2) wasfound among children delivered bynoninstrumental vaginal delivery andnonplanned cesarean section. The highestcrude mean grade point average wasfound in children delivered by plannedcesarean section (225.4), and the lowestmean grade point average was foundamong children delivered by a vacuumextraction attempt followed by an acutecesarean section (222.9). Crude meanmathematics test scores and crude meangrade point averages rose increasinglywith higher maternal age, higher edu-cation, and higher income level.

ogy APRIL 2014

Table 3 presents a linear regressionanalysis of the association between themode of delivery and the mean mathe-matics test score. In model 1, afteradjustments for the year of graduationand sex, children delivered by vacuumextraction scored 0.34 points higher thanchildren born by noninstrumental vag-inal delivery. In model 2, after weadded all maternal and socioeconomicconfounders, the children delivered byvacuum extraction showed lower math-ematics test scores (e0.35) than childrenborn by noninstrumental vaginal delivery.

Maternal age was the confounder thatmost influenced the negative association.In the last model, when all perinatal,social, and maternal confounders wereadded, children delivered by vacuumextraction had lower scores (e0.51)than children born by noninstrumentaldelivery.

Table 4 presents a linear regressionmodel of the association between modeof delivery and mean grade pointaverage. In model 1, after we adjusted foryear of graduation and sex, the meangrade point average among childrendelivered by vacuum extraction was 2.12points higher than that of children bornby noninstrumental delivery.

In model 2, after adding all maternaland socioeconomic confounders, chil-dren delivered by vacuum extraction hada lower mean grade point average(e0.77) than children born by nonin-strumental delivery.

In the last model, when all perinatal,social, and maternal confounders were

TABLE 2Sociodemographic, maternal, and perinatal characteristics in relation to mean mathematics test scores andmean grade point averageDemographic, n [ 126,032 Mathematic scores, Mean (SD) 95% CI Grade point average, Mean (SD) 95% CI

NIVD

n ¼ 100,038 40.2 (14.9) 40.1e40.3 223.8 (52.4) 223.4e224.1

VE

n ¼ 14,494 40.5 (14.9) 40.2e40.7 224.3 (52.3) 223.4e225.2

NPCS

n ¼ 8753 40.2 (14.8) 39.9e40.5 223.1 (52.1) 222.01e224.2

PC

n ¼ 2386 40.4 (14.8) 39.7e41.1 225.4 (52.0) 223.3e227.5

VE plus EC

n ¼ 361 40.3 (15.0) 38.8e41.8 222.9 (52.0) 220.5e228.3

Sex

Male 40.0 (14.9) 39.9e40.1 212.4 (50.5) 211.7e212.8

Female 40.5 (14.9) 40.4e40.6 235.8 (51.5) 235.4e236.2

Gestational age, wks

34-36 39.6 (14.9) 39.2e40.0 220.9 (52.4) 219.5e222.3

37-41 40.2 (14.9) 40.1e40.3 223.9 (52.4) 223.6e224.2

>42 40.5 (14.8) 40.2e40.8 224.0 (51.7) 223.0e225.0

Head circumference, cm

�32 38.3 (14.8) 38.0e38.6 220.3 (52.7) 219.2e221.4

33-35 39.9 (14.9) 39.8e40.0 224.1 (52.4) 223.7e224.5

>36 41.5 (14.9) 41.3e41.7 224.4 (52.0) 223.9e224.9

Missing 39.9 (14.8) 39.4e40.4 221.4 (52.5) 219.7e223.1

Low Apgar 5 minutes

Yes 39.8 (15.0) 38.9e40.7 220.4 (52.4) 217.3e223.5

Missing 39.8 (15.0) 39.0e40.7 222.5 (52.7) 219.5e225.5

Neonatal complicationsa

Yes 39.3 (14.8) 38.8e39.8 219.5 (50.8) 217.7e221.3

Maternal psychiatric morbidityb

Yes 36.4 (14.6) 36.2e36.7 206.2 (56.5) 205.2e207.2

Single-parent household

Yes 37.3 (14.6) 37.1e37.5 209.7 (54.7) 209.1e210.3

Missing 36.9 (14.1) 36.3e37.5 212.3 (55.8) 210.0e214.6

Pregnancy complicationsc

Yes 39.2 (15.0) 38.9e39.5 219.8 (52.7) 218.8e220.8

Maternal age, y

>19 32.1 (13.1) 31.8e32.4 186.3 (52.6) 185.0e187.6

20-29 39.7 (14.8) 39.6e39.8 221.9 (51.3) 221.6e222.2

Ahlberg. Mode of delivery and cognitive function. Am J Obstet Gynecol 2014. (continued)

www.AJOG.org Obstetrics Research

APRIL 2014 American Journal of Obstetrics & Gynecology 361.e5

TABLE 2Sociodemographic, maternal, and perinatal characteristics in relation to mean mathematics test scores andmean grade point average (continued)

Demographic, n [ 126,032 Mathematic scores, Mean (SD) 95% CI Grade point average, Mean (SD) 95% CI

30-39 44.2 (14.9) 44.0e44.4 239.8 (50.2) 239.2e240.4

>40 44.2 (14.9) 43.2e45.3 241.4 (51.0) 237.7e245.1

Smoking

Yes 35.3 (14.1) 35.1e35.5 203.1 (53.7) 202.5e203.7

Missing 39.3 (14.8) 39.0e39.7 220.2 (51.8) 218.9e221.5

Highest educational leveld

Compulsory 35.1 (13.9) 34.9e35.3 202.4 (52.4) 201.5e203.3

High school 39.1 (14.5) 39.0e39.2 220.2 (50.7) 219.9e220.5

University 47.9 (14.4) 47.7e48.1 251.8 (47.6) 251.2e252.4

Incomed

First quintile 37.1 (14.3) 36.9e37.3 211.5 (52.9) 210.7e212.3

Second quintile 37.5 (14.3) 37.3e37.7 213.9 (53.0) 213.2e214.6

Third quintile 38.5 (14.6) 38.3e38.7 217.1 (52.3) 216.4e217.8

Fourth quintile 40.0 (14.7) 39.8e40.2 223.3 (49.9) 222.7e223.9

Fifth quintile 45.1 (14.9) 44.9e45.3 241.7 (48.9) 241.2e242.2

Residencyd

Big city 42.0 (15.0) 41.9e42.1 230.3 (53.3) 229.8e230.8

Small city 39.4 (14.8) 39.3e39.5 220.3 (51.3) 219.9e220.7

Rural 38.0 (14.6) 37.8e38.2 217.2 (52.3) 216.5e217.9

NIVD, noninstrumental vaginal delivery; NPCS, nonplanned cesarean section; PC, planned cesarean section; VE, vacuum extraction; VE plus EC, vacuum extraction with a subsequent cesareansection.

a Cephalic hematoma, intracranial bleeding, convulsions, and other central nervous system disorders; b Drug abuse, alcohol abuse, and psychiatric diagnosis; c Gestational diabetes, maternaldiabetes, preeclampsia, oligohydramnion, polyhydramnion, small for gestational age, and large for gestational age; d Registered at the child’s age of 14 years.

Ahlberg. Mode of delivery and cognitive function. Am J Obstet Gynecol 2014.

Research Obstetrics www.AJOG.org

added, children delivered by vacuumextraction had (e1.05) lower scores,compared with children born bynoninstrumental delivery. Overall, chil-dren born by noninstrumental deliveryhad the highest mean grade pointaverage and mean mathematics testscore compared with all other modes ofdelivery.

When the analyses (using the samemodels) were restricted to term childrenand with an Apgar score of 7 or greater at5 minutes and without a diagnosis offetal distress (not shown in a table), theresults showed the same pattern as thatin Tables 3 and 4. Compared with chil-dren born by noninstrumental vaginaldelivery, those delivered by vacuum ex-traction had a slightly higher mean

361.e6 American Journal of Obstetrics & Gynecol

mathematics test score (þ0.47; 95%confidence interval [CI], 0.14e0.80) andmean grade point average (þ1.57; 95%CI, 0.43e2.73), after adjusting for yearof graduation and sex. When the peri-natal, maternal, and social confounderswere added, children delivered by vac-uum extraction had e0.47 (95% CI,e0.78 to e0.15) mean mathematics testscore and e0.74 (95% CI, e1.77 toe0.30) mean grade point average.The odds of having no registered

mathematics test scores or no regis-tered merit grade were 0.99 (95% CI,0.91e1.08) and 0.70 (95% CI,0.58e0.86) for children delivered byvacuum extraction compared with thosedelivered without a vacuum extraction,respectively.

ogy APRIL 2014

COMMENT

In this national cohort study of 126,032children, we found that both vacuumextraction delivery and nonplannedcesarean section were associated with aslight decrease in both mean mathe-matics test score and mean grade pointaverage at 16 years of age, comparedwith noninstrumental vaginal delivery.Expressed as standard mean differ-ences, the effect sizes of vacuumextraction compared with noninstru-mental vaginal delivery, however, wereonly 0.08 for mathematics test and0.09 for mean grade average, whichare marginal effects according to thecriteria suggested by Cohen,14 in whicheffects between 0.20 and 0.40 areconsidered small.

TABLE 3Linear regression for mean mathematics test score in relation to mode of delivery

MOD Mean

Model 1 Model 2 Model 3

b (95% CI) b (95% CI) b (95% CI)

NIVD 40.2 Referent Referent Referent

VE 40.5 0.34 (0.08e0.60) e0.35 (e0.60 to e0.11) e0.51 (e0.76 to e0.26)

VE plus EC 40.3 0.26 (e1.28 to 1.80) e0.83 (e2.27 to 0.61) e1.08 (e2.52 to 0.36)

NPCS 40.2 0.02 (e0.30 to 0.35) e0.50 (e0.81 to e0.19) e0.51 (e0.82 to e0.20)

PC 40.4 0.27 (e0.33 to 0.87) e0.54 (e1.11 to e0.03) e0.51 (e1.10 to 0.08)

Model 1: year of graduation and sex; model 2: same as model 1, adding maternal age, educational level and residency, single-parent household, maternal smoking, maternal drug abuse, maternalalcohol abuse, and maternal psychiatric diagnosis; model 3: same as previous 2 models, adding small for gestational age, large for gestational age, oligohydramnion, polyhydramnion, maternaldiabetes, preeclampsia, head circumference, and gestational age.

CI, confidence interval; MOD, mode of delivery; NIVD, noninstrumental vaginal delivery; NPCS, nonplanned cesarean section; PC, planned cesarean section; VE, vacuum extraction; VE plus EC,vacuum extraction and subsequent cesarean section.

Ahlberg. Mode of delivery and cognitive function. Am J Obstet Gynecol 2014.

www.AJOG.org Obstetrics Research

These differences, which were verysimilar for vacuum extraction and non-planned cesarean, cannot be expectedto have anything but a minimal impacton the life course of these individuals.Furthermore, our results suggest thatvacuum extraction and nonplanned ce-sarean section are equivalent alternativesfor terminating deliveries with respect tocognitive outcomes.

No previous studies have shown an as-sociation between vacuum extraction de-livery and various adverse long-termhealth outcomes.5-8One largepopulation-based study comparable with ours foundno effects of mode of delivery on intelli-gence scores among children at 17 yearsof age.15 Considering that the observeddifferences in our study were small, we

TABLE 4Linear regression for mean grade po

MOD Mean

Model 1

b (

NIVD 223.8 Referent

VE 224.3 2.12 (1

VE plus EC 222.9 1.03 (e

NPCS 223.1 0.73 (e

PC 225.4 2.08 (0

Model 1: year of graduation and sex; model 2: same as model 1,alcohol abuse, and maternal psychiatric diagnosis; model 3: sadiabetes, preeclampsia, head circumference, and gestational ag

CI, confidence interval; MOD, mode of delivery; NIVD, noninstruvacuum extraction and subsequent cesarean section.

Ahlberg. Mode of delivery and cognitive function. Am J Obs

believe that these 2 studies could beinterpreted in a similar vein. Thus, ourresults suggest that vacuum extraction isquite a safe method of delivery withrespect to cognitive outcomes.It should be noted that this study was

conducted in a country with a relativelyhigh rate of vacuum extraction from aninternational perspective, and therefore,most operators have a high level ofexperience. It has been shown that a lackof operator experience is related toadverse neonatal outcomes in vacuumextractions.16,17 Our results can there-fore not automatically be generalized tosettings in which vacuum extraction is arare choice for terminating a delivery.A major strength of this study is

its nationwide population-based design

int average in relation to mode of delive

Model 2

95% CI) b (95% CI)

Referent

.23e3.01) e0.77 (e1.59 to 0.0

4.24 to 6.29) e3.96 (e8.75 to 0.8

0.39 to 1.84) e1.47 (e2.50 to e0

.01e4.15) e1.43 (e3.33 to 0.4

adding maternal age, educational level and residency, single-parent hme as models 1 and 2, adding small for gestational age, large for ge.

mental vaginal delivery; NPCS, nonplanned cesarean section; PC, pla

tet Gynecol 2014.

APRIL 2014 Ameri

along with the possibility of acquiringinformation on confounders throughdifferent registers. Information on ex-posure, outcomes, and potential con-founders was collected independentlyand without involving the study subjects,thus minimizing various types of bias(eg, selection, recall).

We excluded 12,523 children fromthe study because they had not gradu-ated from compulsory school. No in-dication arose in separate analyses ofthis group that the exclusion biased ourresults.

Although many potential perinatalconfounders were present in the register,many of these have to be consideredlow-precision proxies for the underlyingrisk factors, such as asphyxia, thus

ry

Model 3

b (95% CI)

Referent

4) e1.05 (e1.87 to e0.23)

3) e4.30 (e9.10 to 0.49)

.45) e1.20 (e2.24 to e0.16)

7) e0.90 (e2.85 to 1.05)

ousehold, maternal smoking, maternal drug abuse, maternalestational age, oligohydramnion, polyhydramnion, maternal

nned cesarean section; VE, vacuum extraction; VE plus EC,

can Journal of Obstetrics & Gynecology 361.e7

Research Obstetrics www.AJOG.org

making residual perinatal confoundingan important concern in the interpreta-tion of our results.

ConclusionBoth delivery by vacuum extraction anddelivery by nonplanned cesarean sectionare associated with a marginally lowerschool performance at 16 years of age,compared with noninstrumental vaginaldelivery. The differences, however, areso small that it probably has no impactor only a minimal impact on the lifecourse of a child. This study indicatesthat delivery by vacuum extraction isquite safe for children from a long-termperspective. -

REFERENCES

1. Miksovsky P, Watson WJ. Obstetric vac-uum extraction: state of the art in the newmillennium. Obstet Gynecol Surv 2001;56:736-51.2. Edozien LC. Towards safe practice in instru-mental vaginal delivery. Best Pract Res ClinObstet Gynaecol 2007;21:639-55.

361.e8 American Journal of Obstetrics & Gynecol

3. Simonson C, Barlow P, Dehennin N, et al.Neonatal complications of vacuum-assisteddelivery. Obstet Gynecol 2007;109:626-33.4. Doumouchtsis SK, Arulkumaran S. Head in-juries after instrumental vaginal deliveries. CurrOpin Obstet Gynecol 2006;18:129-34.5. Wen SW, Liu S, Kramer MS, et al. Compari-son of maternal and infant outcomes betweenvacuum extraction and forceps deliveries. Am JEpidemiol 2001;153:103-7.6. Ngan HY, Miu P, Ko L, Ma HK. Long-termneurological sequelae following vacuumextractor delivery. Aust N Z J Obstet Gynaecol1990;30:111-4.7. Blennow G, Svenningsen NW, Gustafson B,Sunden B, Cronquist S. Neonatal and pro-spective follow-up study of infants delivered byvacuum extraction (VE). Acta Obstet GynecolScand 1977;56:189-94.8. Johanson RB, Heycock E, Carter J,Sultan AH, Walklate K, Jones PW. Maternal andchild health after assisted vaginal delivery: five-year follow up of a randomised controlledstudy comparing forceps and ventouse. Br JObstet Gynaecol 1999;106:544-9.9. Carmody F, Grant A, Mutch L, Vacca A,Chalmers I. Follow up of babies delivered in arandomized controlled comparison of vacuumextraction and forceps delivery. Acta ObstetGynecol Scand 1986;65:763-6.

ogy APRIL 2014

10. Cnattingius S, Ericson A, Gunnarskog J,Kallen B. A quality study of a medical birth reg-istry. Scand J Soc Med 1990;18:143-8.11. Marsal K, Persson PH, Larsen T, Lilja H,Selbing A, Sultan B. Intrauterine growth curvesbased on ultrasonically estimated foetal weights.Acta Paediatr 1996;85:843-8.12. World Health Organisation. Internationalclassification of disorders, 9th ed. Geneva(Switzerland): World Health Organisation; 1986.13. Swedish National Agency for Educationweb site. Available at: http://www.skolverket.se/statistik-och-utvardering/statistik/grundskola/betyg/betyg-i-grundskolan-lasar-2010-11-1.160176. Accessed Aug. 30, 2013.14. Cohen JA. power primer. Psychol Bull1992;112:155-9.15. SeidmanDS, Laor A, Gale R, StevensonDK,Mashiach S, Danon YL. Long-term effects ofvacuum and forceps deliveries. Lancet1991;337:1583-5.16. Murphy DJ, Liebling RE, Patel R, Verity L,Swingler R. Cohort study of operative delivery inthe second stage of labour and standard ofobstetric care. BJOG 2003;110:610-5.17. Cheong YC, Abdullahi H, Lashen H,Fairlie FM. Can formal education and trainingimprove the outcome of instrumental delivery?Eur J Obstet Gynecol Reprod Biol 2004;113:139-44.