Armelagos et al., 2009

8/11/2019 Armelagos et al., 2009

1/11

Enamel Hypoplasia and Early Mortality:Bioarcheological Support for the Barker HypothesisGEORGE J. ARMELAGOS, ALAN H. GOODMAN, KRISTIN N. HARPER, AND MICHAEL L. BLAKEY

The Barker hypothesis asserts that stressful events early in the life history of

an individual have negative health consequences later in adulthood. The hypoth-

esis initially focused on prenatal stressors as indicated by birth weight and

related outcomes. This initial concern with the fetal phase of development led to

its description as the fetal programming or fetal origins hypothesis. The real-

ization that stressors in the postnatal phase had similar impacts on adult health

has led to its latest characterization as the Developmental Origins of Health andDisease Hypothesis (DOHaD). In this paper, we review the history and evidence

in support of the DOHaD hypothesis. We then introduce an untapped source of

information on early life stress: enamel hypoplasias and other developmental

defects of enamel. Enamel defects are nearly indelible records of physiological

perturbations, or stress, to developing ameloblasts (enamel-forming cells). Fur-

thermore, the location of the defects translates to specific periods of growth,

providing a permanent temporal record of early life stressors from in utero to

approximately twelve years of age. As we discuss, a handful of studies of differ-

ent populations reveals that individuals with enamel defects that developed

in utero and early in infant-childhood development tend to be subject to earlier

adolescent or adult mortality.

The Barker hypothesis was coinedin an editorial in the British MedicalJournal1 to describe the innovativeand controversial idea that manyadult diseases have their origins infetal development. Named after

David J. P. Barker, an epidemiologistat the University of Southampton,this hypothesis was initially met withskepticism. However, after a greatdeal of replication and testing, theBarker hypothesis has expanded intoa widely accepted phenomenon. Inthis review, we add evidence frombioarcheology that supports theBarker hypothesis. The story of howit gained acceptance is revealed in aquarter-of-a-century journey inwhich many intellectual skirmisheswere fought. We offer additional sup-

port for the Barker hypothesis from

George J. Armelagos is Goodrich C.White Professor of Anthropology atEmory University, Atlanta, Georgia. Hisresearch has focused on diet and dis-

ease in prehistory. He was the VikingMedal Medalist (Wenner-Gren Founda-tion) in 2005, received The Franz Boas

Award for Exemplary Service to Anthro-

pology from the American Anthropologi-cal Association in 2008, and TheCharles Darwin Award for Lifetime

Achievement to Biological Anthropologyfrom the American Association of Physi-cal Anthropologist in 2009. E-mail: [email protected]

Alan Goodman is the immediate pastpresident of the American Anthropological

Association (AAA) and the vice presidentof academic affairs, dean of faculty, andprofessor of biological anthropology atHampshire College in Amherst, Massa-chusetts. His research focuses on devel-oping measures of health and nutritionand the biological consequences of rac-ism, poverty, and inequality. Goodman

co-directs the AAAs public education pro-

ject/Understanding Race/including an

award winning museum exhibit and web-

site (understandingrace.org). E-mail:[email protected]

Kristin Harper is a Robert Wood John-

son Health & Society Scholar atColumbia University. Her research cen-

ters on the role that common factors

such as malnutrition and exposure to

pollution play in molding our suscepti-bility to disease. Currently, she is inves-

tigating the relationship between trace

metal levels, mutation rates, and therisk of developing diseases such as

schizophrenia and cancer. E-mail:

[email protected]

Michael L. Blakey is National Endow-

ment for the Humanities Professor of

Anthropology, Professor of American

Studies, and Founding Director of the

Institute for Historical Biology at the

College of William and Mary. Blakey is a

key advisor on prominent museum proj-

ects including the Race Exhibition of

the American Anthropological Associa-

tion and the National Museum of AfricanAmerican History and Culture of the

Smithsonian Institution, where he heldseveral prior research positions at the

Natural History Museum. He was Scien-tific Director of New York Citys colonial

African Burial Ground archaeologicalsite that has become a National Monu-ment. He has numerous publications in

flagship journals on bioarchaeology,race, racism, ethics, and the history of

anthropology. Michael Blakey held pro-

fessorships at Spelman College, Colum-bia, Brown, Rome and Howard Univer-sity where he founded the W. MontagueCobb Biological Anthropology Labora-

tory. E-mail: [email protected]

Key words: fetal programming; maternal stress;

infancy; DOHaD; childhood stress; adult mortality

VVC 2009 Wiley-Liss, Inc.DOI 10.1002/evan.20239Published online in Wiley InterScience(www.interscience.wiley.com).

ARTICLES

Evolutionary Anthropology 18:261271 (2009)


2/11

bioarcheological evidence that dentalenamel defects occuring in utero,infancy, and childhood have a signifi-cant impact and, indeed, are associ-ated with decreases in longevity.

The story begins when David Barkerand Clive Osmond were examiningthe newly publishedAtlas of Mortalityfrom Selected Diseases in England andWales, 196819782 and were struck bymaps of the distribution of cardiovas-cular diseases, which revealed a re-markable geographic pattern. Barkerand Osmond expected to find thehighest level of cardiovascular diseasein the most affluent areas in GreatBritain, since that would mirror theepidemiological trends for heart dis-ease in the United States and other

high-income nations. Instead, themaps presented them with a conun-drum. TheAtlasshowed that in Walesand England, the highest ischemicheart disease mortality rates in 19681978 were in areas that had experi-enced high infant mortality rates sixdecades earlier (19211925).3 To testthis observation, they used the 212 ge-ographic areas in Wales and Englandto evaluate the association betweenearly infant mortality and later adult

health.3 They found respective corre-lations of 0.69 and 0.68 between ratesof ischemic heart disease and rates ofneonatal and postnatal deaths.

Barker, Osmond, and Law4 thentested the earlier findings with addi-

tional data from the 212 localities inEngland and Wales, this time sepa-rately studying ischemic heart diseaseand stroke. The completeness anddetail of infant mortality records from1911 to 19681978 allowed them toexamine rates of infant mortality atdifferent ages and from different

causes. For example, they could dis-tinguish neonatal mortality (before 1month of age) from postneonatalmortality (from 1 month to 1 year).Using these data, they were able toshow two factors that were associated

with cardiovascular disease (CVD). Inthe first group, CVD was associatedwith lower standards of living inwhich the fetus and infant were sub-

jected to negative stressors; in the sec-ond group, CVD arose later in life andwas linked to an affluent lifestyle.

While these associations were in-triguing, Baker and his team neededa cohort to more rigorously testthese earlier findings. Baker andcoworkers5 decided to use a nationalcohort sample of 9,921 ten-year-oldsand 3,259 adults from Wales and

England to show that systolic bloodpressure was inversely related tobirth weight, controlling for gesta-tional age. The critical evidence that

the findings were independent of ges-tational age confirmed that theobserved relationships were relatedto compromised fetal growth.

Barkers group then began a relent-less search for archival data that

could provide an even more refinedanalysis. After reporting many falsestarts, in which he uncovered incom-plete archives in hospitals, countyhalls, and other places, Barkeruncovered the records of Ethel Mar-garet Burnside (Fig. 1), a midwifefrom Hertfordshire County, England,located 40 miles north of London.6

Beginning in early 1911, Burnsidehad recorded birth information,weight at birth, weight at age one,method of feeding, and illnesses ofthose born in the county6 (Fig. 2).

Barker and colleagues were able tolink the birth records in the Burnsidedatabase with the subsequent death

records of these individuals. Theircohort study7 of linked individual-level birth and death data was com-pleted by the Medical Research CentreEnvironmental Epidemiology Unit,which computerized Burnsides Hert-fordshire County records. The result-ing dataset contained information onmortality outcomes as of 1992 formore than 15,000 men and womenborn between 1911 and 1930. In the

first study based on individuals ratherthan ecological data, they found thatdeath rates from cardiovascular dis-ease among women and men fell pro-

Figure 1. Ethel Margaret Burnside, the mid-

wife whose records were the basis for testing

the Barker hypothesis. Source: Hertoford-

shire Cohort Study. Southampton School of

Medicine.

Figure 2. Ledger showing the format for Ethel Margaret Burnsides records that were used

in testing the Barker hypothesis. Source: Hertofordshire Cohort Study. Southampton School

of Medicine.

262 Armelagos et al. ARTICLES


3/11

gressively between the low and highbirth weights groups (v2 4.3, p 0.04 for women; v2 8.5, p < 0.005for men). In addition, they demon-strated that high levels of weight gainduring infancy decreased mens risk

of CVD while rapid weight gain afterage two increased CVD risk.8

This led to the next step, in whichBarker and colleagues examined acohort of individuals from Hertford-shire born between 1920 and 1930who were still alive.7 This cohortprovided an adequate sample of agroup that was still living. Studies onthis group demonstrated a linkbetween low birth weight and a vari-ety of health issues. The investigatorsfound a relationship between stressduring fetal development and a vari-

ety of adult diseases,4,9

includinghypertension,10 respiratory disease,11

type 2 diabetes,12 insulin resistance

and metabolic syndrome,13 osteopo-rosis,14 and sacropenia15 (loss ofmuscle mass) later in life.7

In a subsequent extension of thisstudy, known as the HertfordshireCohort Study (HCS), 3,000 men andwomen born between 1931 and 1939from Burnsides data were included,allowing greater power to detect andcharacterize statistical associations.This cohort, which was a subsample

of Burnsides database, has been sub-jected to an array of medical testing,including dual energy X-ray absorpti-ometry, which measures bone den-sity. Subjects proximal femora, lum-bar spines, and knees were scanned.7

The HCS again found that higherbirth weight reduced the risk of cir-culatory disease mortality, decreasedrisk of mortality from falls,16 anddecreased risk of mortality frompneumonia, diabetes, and musculo-skeletal disease in women.17

DEVELOPMENTAL ORIGINS

OF HEALTH AND

DISEASE HYPOTHESIS

The Barker hypothesis18,19 has beencalled the fetal programming hy-pothesis,2023 the intrauterine growthrestriction hypothesis,24,25 and theprenatal programming hypothesis.26

However, since many of the physio-logical features of adaptation occur in

the postnatal phase of development,27

the model has been reformulated asthe DOHaD hypothesis.28,29

A growing body of animal researchis beginning to elucidate themechanism by which developmental

changes in the fetus and during earlylife have an impact on adult morbid-ity and mortality. Animal studies30,31

provide a model in which maternalphysiology can be manipulated,32 inways that are not possible withhumans. Such studies have revealedthat feeding pregnant mice a protein-restricted diet played a decisive rolein fetal programming, mediated by

serum-and glucocorticoid-induciblekinase SGK-1, of hypertension in off-spring.33 There is also evidence thatexcessive exposure to glucocorticoidslinks fetal maturation and adultpathology, with transgenerationaleffects.34 In sum, animal studies of fe-tal programming have supported theproposed pathways in humans forsuch conditions as hypertension,35

vascular function,36 metabolic syn-drome,37 and hyperphagia.38

EFFECTS: BEYOND

CIRCULATORY DISEASE

While the effects of early stresseson diseases such as hypertension,coronary heart disease, and diabetesare best characterized and reinforce

the relationship found in the originalBarker hypothesis, increasing evi-dence suggests that other adulthealth outcomes are also profoundlyinfluenced by experiences duringearly development. For example,

multiple independent studies ofinfants whose mothers were exposedto famine early in pregnancy haveshown an increased risk of develop-ing schizophrenia.3941 Similarly,mothers subjected to extreme psy-chological stress early in pregnancybear children with an increased riskof developing schizophrenia.42,43

Animal studies have demonstratedthat early dysregulation in fetalprogramming of the hypothalamic-pituitary-adrenal (HPA) axis duringespecially sensitive periods of fetal

development results in permanentneurological changes in offspring thataffect behavior in complex ways.44

The importance of HPA program-ming in response to insults is cur-rently being investigated in humans.

There is also evidence that devel-opment of the immune system isaffected by early life stressors, withlasting effects. For example, an asso-ciation has been found betweenmaternal respiratory infections dur-ing pregnancy and the risk of asthmain offspring.45 In another sample,

this relationship was also found forprenatal vaginal infections, as well asother types of febrile infection.46

Various other disorders may also berelated to the DOHaD hypothesis,including earlier onset of puberty,47

osteoporosis,48 small kidney size,49,50

and kidney disease.51

WHY THE RELATIONSHIP EXISTS:

A THRIFTY PHENOTYPE

The concept of the thrifty pheno-

type52,53

has been developed toexplain the impact of early life stres-sors on adult health. Supporters ofthis hypothesis propose that meta-bolic syndromes result from theimpact of poor nutrition during thefetal, infant, and childhood periods,producing permanent changes in me-tabolism that affect adult health.52

The thrifty phenotype hypothesisseems to correct unnecessary ass-umptions of the thrifty genotype hy-

As the concern of the

DOHaD moves beyond

circulatory diseases,explanations other thanthe thrifty phenotype willhave to be considered.

Associations between

prenatal or earlychildhood insults andmany conditions may

reflect disruption ofnormal developmental

processes rather thanany sort of adaptation.

ARTICLES Enamel Hypoplasia and Early Mortality 263


4/11

pothesis,54,55 that during the Paleo-lithic period there was selection for asuite of alleles well-adapted to thefeast-or-famine environment. Suchgenetic adaptation would minimizemetabolism in times of plenty to

enhance the storage of fat that couldbe used in time of famine. However,the benefits of this once-favored ge-notype were rendered obsolete if itshost lived in a modern, high-incomeenvironment characterized by anoverabundance of nutrients. Thethrifty phenotype hypothesis assertsthat selection for genetic changes isnot necessary to explain the meta-bolic syndromes proposed by JamesNeels thrifty genotype hypothesis.Instead, it posits that early-life meta-bolic changes in uterine and child-

hood development aid survival of thefetus and child while having a nega-tive impact on adult health.56,57,62

The thrifty phenotype hypothesishas been used to understand the de-

velopment of type 2 diabetes, whichoriginates in intrauterine life and isexacerbated by rapid childhoodgrowth, resulting in a biphasic nutri-tional insult.58 These fetal insults,followed by subsequent catch-up, areimplicated in obesity56,59 and higherbody mass indexes (BMI). In aneight-year follow-up study, Meas and

coworkers60

showed that adults bornsmall for gestational age exhibited alarger gain in BMI than did thosewho had an appropriate size for ges-tational age. As adults, they hadgreater fat mass with more abdomi-nal fat, indicating that the effects offetal growth restriction on body com-position continue long beyond theearly postnatal catch-up growth.

As the concern of the DOHaDmoves beyond circulatory diseases,explanations other than the thriftyphenotype will have to be consid-

ered. Associations between prenatalor early childhood insults and manyconditions may reflect disruption ofnormal developmental processes61

rather than any sort of adaptation.

CRITICISMS OF THE

BARKER HYPOTHESIS

While the studies that stemmedfrom Barkers hypothesis capturedthe interest of the medical commu-

nity, they were also met with skepti-cism.6163 Joseph and Kramer,61 forexample, questioned whether possi-ble selection bias might be at work,since Barker and colleagues wereonly able to link birth and early

weights with adult outcomes for5,700 of the 15,664 individuals in theBurnside database. In support of thiscritique, they observed that individu-als traced later, because of difficultyin linking data, differed in importantrespects from those traced earlier inthe study.

In addition, they described themuch-repeated argument that there isthe possibility of confounding factorsin linking coronary heart disease withchildhood lifestyle and subsequent

poverty,62,63 as well as outlininginconsistencies within and betweenstudies, conflicting evidence, andinability to replicate findings in ani-mal studies. Subsequent meta-analy-ses have echoed some of these con-cerns. One, for example, concluded

that claims of strong inverse associa-tions between weight at birth andsubsequent blood pressure may be at-tributable, in large part, to randomerror, the selective presentation ofresults, and inappropriate adjustmentsfor current weight and confounders.64

In fact, while the Barker hypothesis isintriguing, it is difficult not to viewcontinued poverty as an overarchingcause of both fetal growth retardationand adult health deprivations.

Studies of select cohorts exposedto famine during gestation haveyielded mixed results. These resultsare of special interest because, typi-cally, the whole population wasexposed to the stress, eliminating

some of the problems in controllingfor SES and genotype. Studies ofbabies born during the Dutch Hun-ger Winter revealed that exposureduring early gestation increased theodds of developing coronary heartdiseases65 and, in females only,increased BMI and waist circumfer-ence at the age 50 years.65 In addi-tion, babies exposed during late ges-tation experienced decreased glucosetolerance as adults.66

However, the results found in theDutch cohort have not been repli-

cated in other cohorts exposed tofamine. A study of babies born dur-ing the nineteenth-century Finnishfamine found no effect on mortalityrates.67 Similarly, those who wereborn during the siege of Leningraddisplayed no increased glucose intol-erance, hypertension, or cardiovascu-lar disease as adults, though they didexhibit increased levels of markers ofendothelial damage, as well as astronger association between obesityand blood pressure.68

In twin studies, the effects of re-

stricted in utero growth for babies inmultiple births can be compared tosingleton siblings, thus controlling formothers socioeconomic status (SES)and genotype. One study analyzedwithin-pair differences in birth weightand adult blood pressure, finding asignificant inverse association bet-ween the two.69 However, a studyusing similar methods found no suchassociation,70 and another examiningtwins that were discordant for acutemyocardial infection found no differ-ence in birth measurements (weight,

length, or head circumference) bet-ween affected individuals and theirunaffected siblings or control twin-pairs.71 It was noted that the numberof twins available for the latter studysignificantly limited the power of thisanalysis to detect differences.72 Thislimitation is true of most twin studies,in which statistical power is rarelydiscussed. Studies of twins versus sin-gleton births have also been per-formed, as twins tend to experience

Hypoplasias are caused

by disruptions ofameloblast function

during the secretaryphase in whichsuccessive layers of

enamel matrix are layed

down. In a sense,enamel hypoplasia

provides a kymographicrecord that is a windowinto the past.



5/11

growth retardationin utero. One suchstudy on a large Danish sample foundno relationship between being a twinand higher mortality rate after theage of six years.73

Because of the conflicting results

arising from more rigorous tests ofthe Barker hypothesis, many skep-tics74 argue that much more researchis needed in both humans and ani-mals. At a minimum, it seems clearthat while there is a strong generalassociation between early life stressesand adult diseases, the results are

variable. Certainly, on an individuallevel, early deprivation may be over-come and, at times, appears to leaveno evidence. To the ample data on theimpact of DOHaD on mortality, weadd evidence from prehistory and an-

cient populations that supports theobservation that early lifetime stres-sors have an influence on earlier mor-tality in these groups.

EVIDENCE FROM

BIOARCHEOLOGY

To expand the range of studies ofthe relationship between early lifestresses and adult health, as well asto introduce new methods of track-

ing early life stress, we next consider

bioarcheological evidence. We sum-marize the bioarcheological evidencefor the Barker hypothesis. Currentstudies focus on the relationshipbetween early development andchronic diseases such as coronaryheart disease and diabetes. However,in most of human history and inmany areas of the world today, thesechronic adult conditions are not themost important threats to health.Rather, combinations of infectionand endemic undernutrition havebeen and are the challenges that end

the majority of human lives.A wide variety of studies have dem-

onstrated that earlier stressors have animpact on infection-related adult im-munity, which affects life spans.75

These studies have been able to linkimmunity and lifespan to months ofbirth.76,77 Other factors that affect im-munity,78 including thymic function79

and the impact of non-nutritionalmaternal illness on fetal growth80 playa role in adult health, immunity, and

life span. We ask, then, whether in an-cient populations a relationshipbetween early life challenges and mor-tality can be found and, if so, what itmeans.

Dental enamel permanently re-cords an organisms response tophysiological and chemical condi-

tions from the second trimesterthough early childhood. Further-more, because teeth are abundant inthe bioarcheological record and skel-etons can be aged at the time ofdeath, it is possible to link early con-ditions to longevity.

DEVELOPMENTAL DEFECTS

OF ENAMEL

Dental enamel hypoplasia (Fig. 3),a well-studied class of developmentaldefects of enamel, provides a nearly

indelible record of evidence of physi-ological disruption. In order for thistype of defect to affect the perma-nent teeth, an insult must occurbetween birth, when enamel apposi-tion begins in some permanent teeth,and around the beginning of adoles-cence, when third-molar appositionends.81 The deciduous teeth areaffected earlier, from around the fifthfetal month to the tenth to twelfthpostnatal month.81 Since teeth are

not subject to remodeling, theyremain a permanent record of theseearly events.82 Only the loss of teethdue to dental disease or some otherunfortunate incident, dental attri-tion, loss in height of enamel crownsand, in some cases, dental abrasion,a light wearing away of labial and

lingual surfaces, would result in lossof this information.In their discussion of linear enamel

hypoplasia, Goodman and Rose83

have described the wide variety ofstressors that can initiate the physio-logical mechanisms that lead tochanges in ameloblastic behavior andcause these enamel defects.84 Nutri-tional deficiencies,85 disease,86 con-genital abnormalities,86 and trauma87

are all reported to cause disruptionsin matrix formation. For this reason,linear enamel hypoplasias (LEH)

must be considered nonspecific indi-cators of stress.88 In modern popula-tions, for example, LEHs have beenfound in 18%62% of the deciduousdentition of rural Guatemalan chil-dren,89 but had a prevalence of 6% inthe population of Iowa.90 The nonspe-cific etiology of LEHs is the precisereason why they are often used as partof a toolkit of multiple stress indica-tors to assess physiological disruptionin past populations.

Figure 3. Enamel hypoplasia from the Chokepukio site, Cuzco region of Peru. Source: Val-

erie Andrushko.

ARTICLES Enamel Hypoplasia and Early Mortality 265


6/11

A description of how enamel hypo-

plasias occur may be helpful in under-standing why they are indeliblemarkers of stress that can be dated toprecise periods of development.Because enamel is secreted in a regu-lar ring-like fashion, the tooths crownprovides a permanent chronologicalrecord of any disruption in enamel de-

velopment. Hypoplasias are caused bydisruptions of ameloblast functionduring the secretary phase in whichsuccessive layers of enamel matrix arelayed down.91 In a sense, enamel hy-

poplasia provides a kymographic re-

cord that is a window into the past.Since the chronology of enamel devel-opment in humans is well known,92 itis possible to determine the age atwhich a physiological disruptioncaused the enamel hypoplasia.

Enamel hypocalcification is an-other type of defect. It occurs duringthe second stage of enamel develop-ment, when mineralization occurs. Itproduces a chalky band as extrinsicpigments are incorporated into theenamel. Hypocalcified enamel seg-ments involve the deeper enamel,

leaving the outer enamel skinintact, without the characteristic in-dentation of hypoplasias caused bymatrix apposition.91

It is not necessary to analyze allteeth in order to gain informationabout an individuals health. Good-man, Rose, and Armelagos93:526 sug-gested a best tooth analysis thatincludes two maxillary incisors andtwo mandibular canines; this app-roach reveals about 95% of the infor-

mation that could be obtained if all

teeth were examined. Researchers areexplicit in reporting whether theyused the total sample of dentition,performed best tooth analysis, or

just examined a specific segment ofthe teeth that developed during a par-ticular phase. For example, Goodmanand Armelagos82 used the segments ofthe teeth that mineralized from 3.5 to7 years to assess the impact of enamelhypoplasia on adult mortality.

BIOARCHEOLOGICAL TESTS OF

THE BARKER HYPOTHESIS

White94 provided the first archeo-logical evidence suggesting that child-hood linear enamel hypoplasia isassociated with lower age at mortal-ity. White examined enamel hypopla-sias on permanent maxillary firstmolars of Australopithecines from theSouth African Pleistocene (ca. 2.01.0myr). He determined that individualsfrom the Swartkrans site with enamelhypoplasia on the first maxillarymolar (N 6) died earlier (betweenthe ages of four and ten years) than

did those without enamel defects. Theindividuals without enamel defects ontheir first molars (N 100) diedbetween 8 and 31 years of age. Subse-quently, Goodman95 used Whites dataand computed the mean age at deathfor those with LEHs as 7.8 years andthose without LEHs as 19.6 years.Unlike subsequent studies, the differ-ence in mean ages at death could beexplained in part by the fact thatolder individuals had less enamel

available for study due to enamelattrition and abrasion. However, evenwith this methodological possibility,small sample size, and problems in-herent in analyzing age at death infragmentary material, the nearly 12-

year decrease in mean life expectancyassociated with LEHs is remarkable.Rose, Lallo, and Armelagos,96 used

histological techniques to study areasof disturbed enamel formation(Wilson bands) (Fig. 4) in MiddleWoodland, Mississippian Accultu-rated Late Woodland, and MiddleMississippian samples from DicksonMounds, Illinois. Remarkably, Wil-son bands are not affected by surfaceabrasion. In this group who under-went a transition from horticultureto intensive agriculture, they found

that individuals with Wilson bands,histological measures of stress, hadan earlier mean age at death (26.7years, N 210) than did those with-out Wilson bands (42.1 years, N 66). That is, individuals with a re-cord of early life stress lived 15.4years less than did those without thedefect.

Cook and Buikstra97 were the firstto use a large sample to demonstratethat early mortality was greater inindividuals who experienced stressevents as measured by defects in

enamel development during uterinedevelopment. In their benchmarkstudy,97 they analyzed LEHs in thedeciduous teeth in pooled Middle(ca. 100 BCE to 500 CE) and LateWoodland populations from theLower Illinois Valley. The MiddleWoodland sample included individu-als from the Gibson, Pete Klunk,Lawrence Gay, and the Joe Gaymound groups; the Late Woodlandsample incorporated materials fromthe Gibson, Pete Klunk, Joe Gay, andLedders mounds. The 170 children

represented died during the intervalfrom birth to seven years of age. Theage at which these individuals suf-fered the stressors that produced thehypoplasias was determined to befrom the fourth fetal month to a yearafter birth. We have pooled the dataand constructed a survivorship curvecomparing those with and withouthypoplastic events (Fig. 5). It is evi-dent that the survivorship of thosewithout hypoplasias is greater.

Figure 4. Hypoplasia beneath the double arrow and a Wilson band leading the formation

of the hypoplasia labeled with an arrow (3 20). A second Wilson band located at the

base of the arrow follows and is associated with an undulation in the floor of the hypo-

plasia. From Goodman and Rose.83



7/11


8/11

Duray98 also found support forthe hypothesis that developmentalenamel defects represent stress-in-duced growth disruptions with long-term health effects. In a sample fromthe extensively studied Libben popula-tion from Ottawa County, Ohio, hedemonstrated a relationship betweenenamel defects and age at death. Heexamined defects in the permanentdentitions of 143 individuals using theDevelopmental Defects of Enamel

Index.99

The population age at deathwas determined using multifactorialmethods. Duray found a significantlylower mean age at death for individu-als with enamel defects than amongthose without defects (Table 1).

The age-at-death distribution forindividuals with enamel defectsshowed two peaks: one in the 1520-year age class and one in the 3035-year age class. Duray suggested thatthe earlier mortality of individualswith enamel defects may be due to bi-ological damage to the immune sys-

tem during prenatal or postnatal de-velopment. Among individuals withenamel defects, the mean age at deathwas 5.37 years lower (26.92 yearscompared to 32.29 years) than thatamong individuals with no enameldefects.98 This result was significant

at the 0.01 level (pooled t-test).Duray also found a dose

response effect between the severityof the hypoplastic events and age atdeath. Individuals with grade A or B

linear enamel hypoplasias died, onaverage, 5.88 years younger (26.41years) than did individuals withoutenamel defects (32.29 years) (P

Armelagos et al., 2009

Documents

Transcript of Armelagos et al., 2009