j25 davis 2011 prenatal stress & hpa.pdf

Prenatal maternal stress programs infant stressregulation

Elysia Poggi Davis,1,2 Laura M. Glynn,1,3 Feizal Waffarn2 and Curt A. Sandman11Department of Psychiatry and Human Behavior, University of California, Irvine, Orange, CA, USA; 2Department of

Pediatrics, University of California, Irvine, Orange, CA, USA; 3 School of Health and Life Sciences, ChapmanUniversity, Orange, CA

Objective: Prenatal exposure to inappropriate levels of glucocorticoids (GCs) and maternal stress areputative mechanisms for the fetal programming of later health outcomes. The current investigationexamined the influence of prenatal maternal cortisol and maternal psychosocial stress on infantphysiological and behavioral responses to stress. Methods: The study sample comprised 116 womenand their full term infants. Maternal plasma cortisol and report of stress, anxiety and depression wereassessed at 15, 19, 25, 31 and 36 + weeks gestational age. Infant cortisol and behavioral responses tothe painful stress of a heel-stick blood draw were evaluated at 24 hours after birth. The associationbetween prenatal maternal measures and infant cortisol and behavioral stress responses was examinedusing hierarchical linear growth curve modeling. Results: A larger infant cortisol response to the heel-stick procedure was associated with exposure to elevated concentrations of maternal cortisol during thelate second and third trimesters. Additionally, a slower rate of behavioral recovery from the painfulstress of a heel-stick blood draw was predicted by elevated levels of maternal cortisol early in pregnancyas well as prenatal maternal psychosocial stress throughout gestation. These associations could not beexplained by mode of delivery, prenatal medical history, socioeconomic status or child race, sex or birthorder. Conclusions: These data suggest that exposure to maternal cortisol and psychosocial stressexerts programming influences on the developing fetus with consequences for infant stress regula-tion. Keywords: Pregnancy, stress, cortisol, development, glucocorticoids, prenatal, anxiety, depres-sion, infancy, fetal programming.

The prenatal period is a time of enormous changeduring which organs and organ systems are formingand are susceptible to both organizing and disorga-nizing influences. These influences on the fetus havebeen described as programming: the process bywhich a stimulus or insult during a vulnerabledevelopmental period has a long-lasting or perma-nent effect. The effects of programming are depen-dent on the timing (i.e., the developmental stage oforgan systems and the changes in maternal andplacental physiology) and the duration of exposure(Davis & Sandman, 2010; Nathanielsz, 1999).Compelling evidence from epidemiological studiesindicates that adverse birth outcomes such as lowbirth weight and preterm birth are associated with anumber of diseases of adulthood including heartdisease and obesity (Barker, 1998, 2002) as well aspsychological dysfunction (Bohnert & Breslau, 2008;Costello, Worthman, Erkanli, & Angold, 2007). Theinfluence of prenatal exposure to maternal stresssignals on the developing fetal hypothalamicpitui-taryadrenal (HPA) axis has been proposed as onemechanism that underlies fetal programming ofadult health outcomes (Kapoor, Petropoulos, &Matthews, 2008; Seckl, 2008).Glucocorticoids are steroid hormones that exert

influences on nearly every organ and tissue in thebody (Drake, Tang, & Nyirenda, 2007). Regulation of

the maternal HPA axis changes dramatically duringthe course of normal human pregnancy. Maternalcortisol increases two- to four-fold over the course ofnormal gestation (Sandman et al., 2006). Fetalexposure to the increasing concentrations of mater-nal cortisol is regulated by a placental enzyme, 11b-hydroxysteroid dehydrogenase type 2 (11b-HSD2),which oxidizes cortisol to its inactive form, cortisone(Brown et al., 1996). Levels of placental 11b-HSD2also rise as pregnancy advances, providing partialprotection for the fetus from maternal cortisol duringcritical stages of development (Murphy & Clifton,2003). Because placental 11b-HSD2 is only a partialbarrier, active maternal cortisol passes through theplacenta, and fetal cortisol levels are significantlycorrelated with maternal levels (Gitau, Cameron,Fisk, & Glover, 1998; Gitau, Fisk, Teixeira, Camer-on, & Glover, 2001). Exposure to cortisol that passesthe placental barrier can influence the fetal nervoussystem because cortisol easily passes through thebloodbrain barrier and targets GC receptors thatare present throughout the central nervous system(Jacobson & Sapolsky, 1991; Sanchez, Young, Plot-sky, & Insel, 2000).Results from animal models indicate that fetal

exposure to both maternal stress and glucocortic-oids disrupts the regulation of physiological andbehavioral stress responses in the offspring (Abeet al., 2007; Kapoor et al., 2008; Maccari & Morley-Fletcher, 2007; Schneider, 1992). Similarly, inConflict of interest statement: No conflicts declared.

Journal of Child Psychology and Psychiatry 52:2 (2011), pp 119129 doi:10.1111/j.1469-7610.2010.02314.x

2010 The Authors. Journal of Child Psychology and Psychiatry 2010 Association for Child and Adolescent Mental Health.Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main St, Malden, MA 02148, USA

humans, prenatal treatment with synthetic gluco-corticoids shapes the development of the fetal HPAaxis (Ashwood et al., 2006; Davis, Townsend et al.,2004; Davis et al., 2006). Few human studies haveevaluated the joint role of maternal cortisol andpsychosocial stress for infant development(Sandman & Davis, 2010). There is evidence thathigh levels of prenatal psychosocial stress (Bergman,Sarkar, OConnor, Modi, & Glover, 2007; Davis,Snidman et al., 2004; Van den Bergh, 1990) andprenatal maternal cortisol (Davis et al., 2007; C. deWeerth, van Hees, & Buitelaar, 2003) are associatedwith fearful and reactive behaviors in the offspring.Less is known about the influence of prenatalmaternal stress signals on the developing HPA axis(Sandman & Davis, 2010). Several studies haveshown that maternal psychosocial stress alters cir-cadian regulation and laboratory levels of cortisol inthe offspring (Grant et al., 2009; OConnor et al.,2005; Van den Bergh, Van Calster, Smits, Van Huf-fel, & Lagae, 2007). To our knowledge, only Guttelingand colleagues (2004, 2005) have evaluated theinfluence of prenatal maternal cortisol on HPA axisfunctioning in the offspring. In samples of 24 and 29children respectively, they found that elevatedmaternal cortisol and psychosocial stress measuredat one time during gestation (15 to 17 gestationalweeks) independently predicted higher cortisol levelson the day of an inoculation and on the first day of anew school year. These data suggest that the devel-opment of the fetal HPA axis is shaped by prenatalmaternal stress signals.The present study is unique in the evaluation of

both prenatal maternal psychosocial stress andcortisol, at multiple gestational periods, in associ-ation with infant physiological and behavioralstress regulation. Multiple assessments of stressare essential for examining timing or critical periodeffects on the developing fetus and for determiningthe trajectory of stress measures over the course ofpregnancy. Evaluation of infant behavioral andphysiological stress regulation in the immediateneonatal period provides a unique opportunity toidentify the consequences of prenatal programmingindependent of postnatal influences. This studywill determine whether prenatal maternal psycho-social stress and cortisol exert a joint or indepen-dent influence on infant stress regulation andwhether there are critical periods in gestationduring which the fetus is more susceptible to theseinfluences.

Methods

Study overview

Study participants included motherinfant pairs froman ongoing longitudinal study of prenatal stress anddevelopment. Women with singleton pregnancies lessthan 16 weeks gestational age were recruited from

obstetric clinics in Southern California and have beenfollowed longitudinally.

Participants

The current sample comprised 116 full term infants (55girls and 61 boys) and their mothers. Initial prenatalrecruitment criteria were as follows: English speaking,non smoker, over the age of 18, no steroid medication,and no evidence for drug or alcohol use during preg-nancy. Additional inclusion criteria for the currentstudy were full term at birth, delivery at the Universityof California Irvine, Medical Center, admission to theWell-baby Nursery and complete data for the heel-stickprocedure (data collection was attempted for 126infants. However, six infants did not provide sufficientsaliva samples for cortisol analysis and four weremissing behavioral data). Mothers gave informed con-sent for all aspects of the protocol, which was approvedby the Institutional Review Board for protection ofhuman subjects. Descriptive information for the studysample is shown in Table 1.

Maternal data collection

Maternal plasma samples were collected for cortisolanalysis and maternal psychological state (state anxi-ety, perceived stress and depression) at five intervalsduring pregnancy (Time 1: 15.1 .9, Time 2: 19.3 1.0,Time 3: 25.4 .9, Time 4: 30.9 .69; Time 5:36.5 1.2 weeks of gestation). The sampling intervalswere selected based upon evidence that biologicalmarkers of stress (e.g., cortisol) increase over the courseof human pregnancy (e.g., Sandman et al., 2006). Wehave adopted roughly equal intervals that repre-sent early, mid and late pregnancy to capture the tra-jectories of these markers. The initial period is theearliest that we could reliably recruit women into thestudy. Prenatal medical history and risk for adverseobstetric outcomes were obtained from extensivestructured medical interviews at each visit in combi-nation with a thorough review of prenatal and hospitalmedical records.

Table 1 Demographic information for the study sample

Study sample (N = 116)

Infant gestational age at birth(weeks)

M = 39.1, SD = 1.1

Infant birth weight (grams) M = 3448, SD = 434Maternal age at delivery (years) M = 28.6, SD = 5.8Married (%) 65Primiparous (%) 45Education (years) 14.5Annual household income (%)$0 to $30,000 29$30,001 and $60,000 22$60,001 and $100,000 28Over $100,000 21

Race/Ethnicity (%)Non-Hispanic white 51Hispanic 34Asian 11

120 Elysia Poggi Davis et al.

2010 The AuthorsJournal of Child Psychology and Psychiatry 2010 Association for Child and Adolescent Mental Health.

Maternal plasma cortisol

In the afternoon, at least one hour after women hadeaten (mean time of day ranged from 13:19 to 13:30across the five assessment intervals), maternalblood samples were withdrawn by antecubital veni-puncture into EDTA vacutainers. EDTA vacutainerswere chilled on ice immediately, then centrifuged at2000 g (15 minutes), decanted into polypropylenetubes and stored at )70C until assayed. Plasmacortisol levels were determined by a competitive bindingsolid phase enzyme-linked immunosorbent assay (IBLImmuno Biological Laboratories America, Minneapolis,MN). All samples were assayed in duplicate. The inter-and intra-assay coefficients of variance are reported asless than 8% with a minimum detectable level of .25 lg/dL.

Maternal psychological assessments

Maternal psychosocial stress was evaluated using threestandardized measures all of which have been usedextensively with pregnant and non-pregnant subjects(e.g., Glynn, Dunkel Schetter, Hobel, & Sandman,2008; Marcus, Flynn, Blow, & Barry, 2003). General-ized or non-specific stress was evaluated using the12-item version of Cohens Perceived Stress Scale (PSS;Cohen, Kamarck, & Mermelstein, 1983). The PSS eval-uated participants feelings about how they were able tohandle day-to-day problems and hassles, how oftenthey felt nervous and stressed and how often they feltthings were going well during the past week. Responseswere made on a 5-point Likert scale ranging from 0(never) to 4 (almost always) and the final score couldrange from 0 to 48. The short form of the Center forEpidemiological Studies Depression Inventory was usedto evaluate maternal depression (CES-D; Santor &Coyne, 1997). Responses to each of the 9 items in thismeasure were recorded on a four-point Likert scale witha range of 0 to 3. Anchor points, in terms of days perweek during the last week, were rarely or none of thetime (less than 1 day) to most or all of the time(57 days). The final score could range from 0 to 27,with a higher score indicating greater impairment. The10-item State Anxiety subscale of the State-TraitPersonality Inventory (Speilberger, 1983) assessed thedegree to which participants had experienced anxiety-related symptoms or emotions in the last few days.Responses were made using a 4-point Likert scaleranging from 1 (not at all) to 4 (very much) and scorescould range from a minimum of 10 to a maximum of 40.

Prenatal course and birth outcome

An extensive structured medical interview was con-ducted by a research nurse at each prenatal visit toassess maternal health and pregnancy-related compli-cations. Maternal and infant medical records werereviewed to assess pregnancy complications and birthoutcome. A binary score assessing the presence orabsence of prenatal obstetric risk was derived (Hobel,1982). This obstetric risk score includes assessment ofthe presence or absence of infection, pregnancy-induced hypertension, gestational diabetes, oligohy-dramnios, polyhydramnios, preterm labor, vaginal

bleeding, placenta previa and anemia, in the indexpregnancy.

Infant data collection

Infant response to the heel-stick blood draw was eval-uated between 13 and 35 hours after birth (M = 23 hrs47 min, SD = 4 hrs 13 min). Infants were monitoredcontinuously for one hour prior to the assessment toensure that they were not handled or fed. The assess-ment began with baseline measurements. During the10-minute baseline period, behavioral state wasrecorded at 20-second intervals. Baseline salivary cor-tisol was then assessed, followed by a clinically indi-cated heel-stick blood draw performed by a neonatalnurse. For this protocol, the participants nurse cleanedthe heel with an alcohol swab, applied an automatedlancet, and then repeatedly squeezed the heel until asufficient blood sample was obtained. The medianlength of heel-stick was 4 minutes. Behavioral assess-ments continued during the heel-stick procedure andduring the 5 minutes immediately following the heel-stick (recovery). To capture the peak cortisol responseto the heel-stick, salivary cortisol samples were taken at20 minutes and 40 minutes after the start of the heel-stick (Davis, Townsend et al., 2004; Gunnar, 1989;Gunnar, Hertsgaard, Larson, & Rigatuso, 1992). Dur-ing the study period, infants were not fed or handledbeyond the blood draw protocol.

Infant salivary cortisol assessment

Saliva was obtained (without any stimulant) by placinga swab in the infants mouth. Samples can be collectedin this manner without disturbing or waking the infant(e.g., Davis et al., 2004). Salivary cortisol reflects theunbound or active fraction of cortisol and is highlycorrelated with plasma cortisol (Gunnar, 1989; Kirs-chbaum & Hellhammer, 1989). The collection swab wasthen placed in a saliva extraction tube (Roche Diag-nostics). Saliva samples were spun and stored at )20Cuntil assayed. Salivary cortisol levels were determinedby a competitive luminescence immunoassay (LIA; IBL-America, Minneapolis, MN) with detection limits of.005 lg/dl. The intra- and inter-assay coefficients ofvariance are 5.5% and 7.6%, respectively. All sampleswere assayed in duplicate and averaged.

Infant behavioral responses to stress

Behavioral state was assessed using a modified versionof a coding system designed for the assessment ofneonatal behavioral regulation (Als, Lester, Tronick, &Brazelton, 1988). This scheme was used to categorizeeach infants state on a scale of 1 to 6: quiet sleep,active sleep, quiet awakeness/drowsy, awake and alert,awake and fussy, and crying. Behavior was coded inreal time in 20-second epochs. Codes recorded repre-sented the highest state noted during each 20-secondepoch. For 15% of the infants, two observers indepen-dently coded behavioral state. Percent agreement forstate codes was 85.1%. Coders were blinded to allinformation about the mothers (e.g., maternal prenatalcortisol and psychological state).

Prenatal cortisol and infant development 121


Data transformation and analysis plan

Analyses were performed using SPSS 17 and HLM6software. Inspection of maternal cortisol data indicatedthat levels were normally distributed and analyses wereperformed with raw data. Summary scores were createdto characterize infant cortisol and behavioral responsesto the heel-stick. For infant cortisol, delta scores werecalculated by subtracting baseline levels from values at20 minutes (response) and 40 minutes (recovery) afterthe start of the heel-stick. Behavioral state compositescores were calculated for each of the three phases.Baseline was defined as the average behavioral statescore during the 10-minute period prior to the heel-stick procedure. The heel-stick response score wasdefined as the average behavioral state score during thefirst minute of the heel-stick procedure. This intervalwas selected to ensure that all infants had completedata despite variability in length of heel-stick. Recoverywas defined as the average behavioral state score dur-ing the 5-minute period following the end of the heel-stick. Repeated measures ANOVAs, with post hoc asneeded, were implemented to determine if cortisol oraverage behavioral state changed across the threeassessment intervals (baseline, response and recovery).Preliminary analyses were performed to identify

variables that might influence infant cortisol or behav-ioral regulation, including sociodemographic factors(marital status, race/ethnicity, education, and house-hold income), medical risk (prenatal medical risk, par-ity, maternal age, mode of delivery) and infant factors(gestational age at birth, birth weight, sex, Apgar score,and postnatal age in hours). None of these factors wereassociated significantly with infant cortisol or behaviorat baseline, response or recovery and so none wereincluded in the model as covariates (all ps > .10).Additionally, aspects of the heel-stick event, time of day(ps > .14) and length of the heel-stick procedure(ps > .12) were not associated with patterns of infantcortisol or behavior. Even within the restricted range ofsample collection, maternal cortisol was correlated withcollection time (rs ranged from .49 to .33). Thus, timeof maternal sample collection was included as acovariate in all subsequent analyses. Additional corre-lational analyses determined that there were no signif-icant associations between prenatal maternal cortisolor psychosocial stress and baseline infant cortisol andbehavioral state (all ps > .15).Hierarchical linear modeling (HLM) growth curve

analyses (Raudenbush & Bryk, 2002) were used todescribe the trajectory of prenatal maternal cortisol andpsychosocial stress across pregnancy. HLM, when usedwith repeated measures, treats the data in a hierar-chical fashion with observations nested within persons.

This approach allows variance to be modeled at multi-ple levels and provides several advantages over ordinaryleast squares (OLS) regression. First, standard regres-sion models are limited to one component of variability,the deviation of the individual from the group mean. Incontrast, HLM includes the within-person variabilityassessed over time. Second, HLM uses precise mea-sures of timing (i.e., gestational week) of data collectionrather than nominal estimates of assessment intervals.Third, estimates of goodness of fit in modeling eachindividuals data are derived and the most reliable dataare given greater statistical weight. Fourth, HLM pro-duces robust estimates despite missing values for therepeated measure. In HLM cases with complete data areweighted more heavily, but all cases are included in theestimation of effects. Ninety-two mothers had completecortisol data at all prenatal assessments. Twenty-oneparticipants were missing data from one prenatalassessment and three participants were missing datafrom two assessments.Separate two-level models were constructed to

determine the association between the trajectory ofmaternal cortisol and maternal psychosocial stressacross gestation and infant cortisol and behavioralresponses to stress. A linear model was used for per-ceived stress, state anxiety and depression. For mater-nal plasma cortisol a quadratic function was a better fit(p < .05). In all cases predictor and outcome variableswere modeled as continuous variables. Level 1 vari-ables, or those evaluated longitudinally across the fiveprenatal assessment days, included: maternal cortisol,maternal psychosocial stress measures, time of day andgestational week. Infant cortisol levels (response-delta20 min post and recovery-delta 40 min post) or behav-ioral composite scores (response and recovery) were theoutcomes of interest and were modeled continuously atlevel 2. These analyses were used to identify the periodduring gestation during which maternal cortisol andpsychosocial stress were most strongly associated withinfant stress responses. The model was tested for dif-ferences in intercept at each gestational week within therange of actual endocrine and psychosocial assess-ments (13 to 38 weeks of gestation).

Results

Prenatal maternal cortisol

As anticipated, maternal cortisol increased from10.7 lg/dl at 15 gestational weeks to 25.5 lg/dl at36 gestational weeks (see Table 2). Further, asshown in Table 3, maternal plasma cortisol levelswere moderately correlated across gestation.

Table 2 Maternal cortisol and psychosocial stress (means and standard deviations) at the five prenatal assessment intervals

Gestational week

15 19 25 31 37

Maternal cortisol lg/dl 10.7 (3.9) 14.6 (5.4) 19.1 (6.5) 22.0 (6.4) 25.5 (7.5)Perceived stress 16.6 (7.3) 15.0 (7.7) 15.5 (7.9) 14.2 (7.7) 14.6 (7.1)State anxiety 19.6 (5.3) 18.7 (5.8) 18.5 (5.8) 18.1 (5.6) 18.9 (5.7)Depression 6.7 (5.3) 6.0 (4.7) 6.2 (5.0) 6.7 (4.9) 6.7 (4.9)



Maternal psychological state

As shown in Table 2, maternal perceived stress(rs ranged from .44 to .67, ps < .001), depression (rsranged from .44 to .67, ps < .001) and state anxiety(rs ranged from .41 to .58, ps < .001) were relativelystable across gestation. Further, within each assess-ment time point the three measures were highlycorrelated (rs ranged from .47 to .83, ps < .001).

Prenatal maternal cortisol and maternalpsychosocial stress

Maternal anxiety and depression scores were notassociated significantly with maternal cortisol at anyof the five prenatal assessments (all ps > .10).Among all of the psychosocial stress measures, onlyperceived stress at 25 weeks gestation was associ-ated with maternal cortisol after controlling for timeof day [partial r(115) = .21, p < .05]. This associa-tion did not remain significant after correction formultiple comparisons. Perceived stress at the otherfour assessment time points was not associatedsignificantly with maternal cortisol (all ps > .10).

Infant stress responses

Infant cortisol levels changed in response to the heel-stick [F(2,114) = 10.9, p < .01]. Post hoc testsrevealed that infant cortisol levels were higher at 20

[t(115) = 4.3, p < .01] and 40 minutes [t(115) = 3.3,p < .01] as compared to baseline levels. Similarly,infants responded behaviorally to the heel-stick[F(2,114) = 66.2, p < .01]. Infants were in a signifi-cantly more aroused behavioral state during theheel-stick response [t(115) = 10.2, p < .01] and therecovery period [t(115) = 2.3, p < .05], as compared tobaseline. Infant cortisol was not significantly associ-ated with behavioral state scores during the baselineor response period (ps > .35), but was modestlyassociated with behavior during the recovery period.Infants who displayed a greater cortisol increase at20 minutes and 40 minutes after the start of the heel-stick stressor also showed delayed behavioral recov-ery from the heel-stick stressor [r(116) = .21, p < .05and r(116) = .17, p = .06 respectively]. Infant stressresponses did not differ by sex (ps > .20).

Prenatal maternal cortisol and infant stressregulation

The trajectory of maternal cortisol across gestationwas associatedwith the infant cortisol response to theheel-stick stressor. Growth curve analyses, includingmaternal and infant cortisol as continuous variables,revealed significant intercept differences between 21and 35 gestational weeks (bs: .72 to 1.11; ts: 2.0 to2.5, ps < .05), indicating that elevated maternalcortisol was associatedwith a larger increase in infantcortisol in response to the heel-stick (Figure 1A). Asshown in Figure 1B, high maternal cortisol between20 to 27 gestational weeks predicted elevated infantcortisol at the recovery assessment (bs: .52 to .68; ts:2.0 to 2.2, ps < .05). The strongest association (i.e.,largest b) between prenatal maternal cortisol andinfant cortisol response to the heel-stick procedurewas at 25 gestational weeks. This association is pre-sented in Figure 2, illustrating that infants exposedto the highest (top quartile) or lowest (bottom quartile)maternal cortisol have the largest and smallest HPAresponses respectively.

Table 3 Intercorrelations among prenatal maternal plasmacortisol samples

Gestational week

19 25 31 37

15 Wks GA .33* .32** .24* .30**19 Wks GA .52** .32** .25**25 Wks GA .29** .24*31 Wks GA .33**

* p < .05, ** p < .01.

Gestational Week

15 wks 19 wks 25 wks 31 wks 37 wks

Gestational Week

15 wks 19 wks 25 wks 31 wks 37 wks

Mate

rnalC

ortis

olg/d

l

0

10

15

20

25

30Smaller infant cortisol response at 40 minLarger infant cortisol response at 40 min

Mate

rnalC

ortis

olg/d

l

0

10

15

20

25

30

Smaller infant cortisol response at 20 minLarger infant cortisol response at 20 min

(A) (B)

Figure 1 Figures A and B illustrate the trajectory of maternal cortisol across gestation that is associated with theinfant cortisol response to the heel-stick. Data were analyzed continuously using growth curve modeling. For illus-trative purposes, prenatal maternal cortisol trajectories (with standard error bars) are displayed for mothers ofinfants with the highest cortisol response (top quartile) and the lowest cortisol response (bottom quartile) at20 minutes (Figure 1A) and 40 minutes (Figure 1B) after the heel-stick



The pattern of maternal cortisol across gestationalso predicted infant behavioral state during therecovery period. Mothers with high cortisol early inpregnancy had infants who displayed slower behav-ioral recovery from the painful stress of the heel-stick procedure. Specifically, as shown in Figure 3,elevated maternal cortisol levels at the earliest timepoints measured (13 to 14 gestational weeks) wereassociated with higher infant behavioral arousalduring recovery (bs: .54 to .65; ts: 2.1 to 2.2,ps < .05). In Figure 4 behavioral state scores arepresented for infants born to mothers with high (topquartile) and low (bottom quartile) cortisol levels at13 weeks gestation. Infant behavioral state duringthe heel-stick response period (ps > .3) was not sig-nificantly associated with maternal prenatal cortisol.

Prenatal maternal psychosocial stress and infantstress regulation

The pattern of maternal psychosocial stress acrossgestation also was associated with infant behavioralstate during the recovery period. Elevated levels ofperceived stress, anxiety and depression were asso-ciated with slower infant behavioral recovery. Infantsborn to mothers with elevated levels of maternalperceived stress throughout gestation maintained ahigher level of behavioral arousal during recovery(from 13 to 35 weeks; bs: .09 to .10; ts: 1.9 to 2.4,ps < .05, between 36 and 38 weeks; bs: .086 to.087; ts: 1.8 to 1.9, ps < .10). Elevated behavioralstate scores during recovery were additionally asso-ciated with elevated maternal state anxiety (from 26to 38 gestational weeks; bs: .06 to .09; ts: 1.9 to 2.1,

ps < .05) and elevated maternal depression (from 24to 38 gestational weeks; bs: .07 to .13; ts: 1.9 to 2.4,ps < .05). Figure 5 illustrates the associationbetween maternal perceived stress scores acrossgestation and infant behavioral state during recov-ery. Figure 6 depicts behavioral state scores forinfants born to mothers with high (top quartile)and low (bottom quartile) perceived stress scores at13 weeks gestation. Maternal psychosocial stress

Sampling TimeBaseline 20 min post 40 min post

Infa

nt C

ortis

ol

g/dl

0.0

0.5

1.0

1.5

2.0

Low maternal cortisol at 25 wks GA High maternal cortisol at 25 wks GA

Figure 2 This figure illustrates the infant cortisol pro-file that is associated with high (top quartile) and low(bottom quartile) prenatal maternal cortisol. Data arepresented based on maternal cortisol at 25 weeksgestation (the point with the strongest associationbetween prenatal maternal cortisol and infant cortisolresponse to the heel-stick procedure based on growthcurve models). Note: Both maternal cortisol and infantcortisol were analyzed as continuous variables. Forgraphing purposes we have presented data based onthe top and bottom quartile

Gestational Week15 wks 19 wks 25 wks 31 wks 37 wks

Mat

erna

l Cor

tisol

g/

dl

0

10

15

20

25

30Lower infant behavioral arousal during recoveryHigher infant behavioral arousal during recovery

Figure 3 The trajectory of maternal cortisol acrossgestation that is associated with infant behavioralrecovery scores is presented. Although data were ana-lyzed continuously using growth curve modeling, forgraphing purposes the trajectory of cortisol acrossgestation (with standard error bars) is shown for themothers of infants with the highest behavioral recoveryscores (top quartile) and the lowest behavioral recoveryscores (bottom quartile)

Infa

nt B

ehav

iora

l Sta

te

2.0

2.5

3.0

3.5

4.0

4.5

5.0

Low maternal cortisol at 13 wks GAHigh maternal cortisol at 13 wks GA

0 1 2 3 5 64

Heel-stick Recovery

Time (min)

Figure 4 This figure illustrates the pattern of behav-ioral responses that is associated with maternal corti-sol. Behavioral state scores are presented for infantsborn to mothers with high (top quartile) or low (bottomquartile) cortisol levels at 13 weeks gestation (the pointwith the strongest association between prenatalmaternal cortisol and infant behavioral responses to theheel-stick procedure based on HLM analyses)



measures were not associated with behavioral stateduring the response period or infant cortisolresponse or recovery (all ps > .12).

Discussion

In a prospective study with multiple prenatalassessments we document that prenatal maternalcortisol and psychosocial stress program infant

stress regulation. Confidence in these data isincreased because associations are observed amonghealthy full term neonates and thus cannot beexplained by postnatal influences such as parentingstyle. Prenatal maternal psychosocial stress andcortisol each exert influences on infant stress regu-lation, independent of baseline functioning, andthese influences are dependent upon the gestationalperiod during which the fetus is exposed.

Prenatal maternal cortisol and infant stressregulation

The results of this study indicate that the trajectoryof maternal cortisol levels across gestation predictsinfant behavioral and physiological responses tostress. This likely occurs through the actions ofcortisol on the developing central nervous system(CNS). Cortisol is important for normal maturation inmost regions of the developing CNS initiating termi-nal maturation, remodeling axons and dendrites,myelination and cell survival (Kapoor et al., 2008;Raschke, Schmidt, Schwab, & Jirikowski, 2008;Setiawan, Jackson, MacDonald, & Matthews, 2007).Moreover, receptors for cortisol are highly expressedin the developing brain (Jacobson & Sapolsky, 1991;Sanchez et al., 2000). Exposure to cortisol duringgestation has widespread effects upon neuronalstructure and synapse formation and the specificconsequences will be dependent on the gestationalinterval of exposure (Seckl & Meaney, 2006).The prenatal period of susceptibility to the pro-

gramming influence of maternal cortisol differs forphysiological and behavioral stress regulation. Asreported here, elevations in maternal cortisol early inpregnancy affect infant behavioral stress responses,but later elevations affect infant cortisol responses.The distinct consequences of maternal cortisol basedon the timing of exposure may result both fromchanges in placental physiology regulating fetalexposure tomaternal cortisol (Brown et al., 1996) andvariations in the sensitive periods of development ofdifferent fetal physiological systems (Seckl &Meaney,2006). Between the gestational ages of 8 and16 weeks migrating neurons form the subplate zone,awaiting connections from afferent neurons originat-ing in the thalamus, basal forebrain, and brainstem.Once neurons reach their final destination, theyarborize and branch in an attempt to establishappropriate connections (Sidman&Rakic, 1973). Theconnectivity between brainstem, limbic and corticalbrain regions is responsible for behavioral regulation(Geva & Feldman, 2008). It is possible that earlyexposure tomaternal cortisol affects infantbehavioralregulation because of the influence of cortisol on thedevelopment of connectivity between these regions.Exposure to maternal cortisol later in gestation

may influence the development of the fetal HPA axisby modifying glucocorticoid receptor development inregions such as the paraventricular nucleus of the

Gestational Week15 wks 19 wks 25 wks 31 wks 37 wks

Mat

erna

l Per

ceiv

ed S

tress

10

12

14

16

18

20 Lower infant behavioral arousal during recoveryHigher infant behavioral arousal during recovery

Figure 5 The trajectory of perceived stress across ges-tation associated with infant behavioral recovery scoresis presented. Although data were analyzed continuouslyusing growth curve modeling, for graphing purposes thetrajectory of perceived stress across gestation (withstandard error bars) is presented for the mothers ofinfants with the highest behavioral recovery scores (topquartile) and the lowest behavioral recovery scores(bottom quartile)

Infa

nt B

ehav

iora

l Sta

te

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

0 1 2 3 5 64

Heel-stick Recovery

Time (min)

Low maternal perceived stress at 13 wks GAHigh maternal perceived stress at 13 wks GA

Figure 6 The pattern of behavioral responses that isassociated with maternal perceived stress is shown.Behavioral state scores are presented for infants born tomothers with high (top quartile) and low (bottom quar-tile) perceived stress at 13 gestational weeks (the pointwith the strongest association between prenatal mater-nal perceived stress and infant behavioral responses tothe heel-stick procedure based on HLM analyses)



hypothalamus, hippocampus, and amygdala (Her-man & Cullinan, 1997; Levitt, Lindsay, Holmes, &Seckl, 1996). These regions are particularly sensitiveto excessive levels of glucocorticoids. Findings fromanimal models suggest that exposure to elevatedlevels of GCs alters the density of both types of cor-tisol receptors (mineralocorticoid and glucocorticoid)in the hippocampus and the amygdala with conse-quences for HPA axis feedback sensitivity (Kapoor,Dunn, Kostaki, Andrews, & Matthews, 2006). Lim-ited information exists regarding the time course ofprenatal development of cortisol receptors inhumans. There is evidence that both types of cortisolreceptors are present in the human hippocampus by24 gestational weeks (Noorlander, De Graan, Mid-deldorp, Van Beers, & Visser, 2006), indicating thatin the latter half of gestation this region is suscepti-ble to the consequences of excess cortisol. Altera-tions to neurological systems at different timesduring fetal development resulting from prenatalexposure to GCs may determine the neonates abilityto respond behaviorally and physiologically tostressors in the postnatal environment.Data suggesting that elevations in maternal corti-

sol are associated with increased behavioral stressreactivity are consistent with animal studies (Abeet al., 2007) and with studies in humans that linkmaternal cortisol with increased fearful or reactivebehavior during infancy (Davis et al., 2007; C. deWeerth et al., 2003). The observed associationbetween maternal cortisol between 21 and 35 weeksand infant cortisol reactivity is partially consistentwith studies which showed associations betweenmaternal cortisol measured between 15 and 17 ges-tational weeks and child cortisol levels (Guttelinget al., 2004, 2005). The current investigation is theonly one to evaluate timing of exposure with multipleassessments across gestation. It is plausible thatdifferences in the type of stressor and age at assess-ment account for the variability across studies.

Maternal psychosocial stress and infant stressregulation

Maternal psychosocial stress is additionally associ-ated with infant behavioral stress regulation. Highermaternal perceived stress throughout gestation andhigher anxiety and depression during the end of thesecond and the third trimester predict a slower rate ofrecovery of behavioral stress responses in the new-born. The positive association between prenatalmaternal stress and neonatal behavioral reactivity isconsistent with both experimental animal studies(Kapoor et al., 2008; Maccari & Morley-Fletcher,2007) and human work (Bergman et al., 2007; Davis,Snidman et al., 2004) linking prenatal maternalstress with increased behavioral reactivity in theoffspring in response to challenge. Prenatal psycho-social stress is not significantly correlated with neo-natal cortisol regulation, suggesting that HPA

functioning immediately after birth is not influencedby prenatal maternal psychosocial stress. It remainspossible, however, that associations between prena-tal psychosocial stress andHPA axis regulation in theoffspring will emerge later in development (OConnoret al., 2005; Van den Bergh et al., 2007) and thatassociations may be present in high-risk populationsin which exposures are more dramatic (Grant et al.,2009; Keenan, Gunthorpe, & Grace, 2007).Consistent with prior work, it seems unlikely that

maternal cortisol mediates the effect of maternalreport of psychosocial stress on infant behavioralreactivity (Bergman, Glover, Sarkar, Abbott, &OConnor, 2010; Davis et al., 2007; Davis & Sand-man, 2010; C. de Weerth & Buitelaar, 2005). This isso because early in pregnancy (the time whenmaternal cortisol is associated with infant behavior)maternal cortisol and psychosocial stress are notcorrelated and because maternal cortisol later ingestation does not predict infant behavior. Maternalpsychosocial stress exerts widespread influences ona number of stress sensitive systems other than theHPA axis, including the immune and vascularsystems (Dunkel Schetter & Glynn, 2010). Thesesystems are candidates for future studies examiningpossible pathways that might mediate effects ofpsychosocial stress on the developing fetus. Further,although smoking and drug and alcohol use wereexclusion criteria for this study, elevated psychoso-cial stress is likely associated with other health-related behaviors such as nutrition and physicalactivity (see Dunkel Schetter & Glynn, 2010, forreview). These behavioral profiles could havecontributed to associations between maternal psy-chosocial stress and infant behavioral regulation.

Limitations

First, because this study relied on naturally occur-ring variations in maternal cortisol, rather thanexperimental manipulations, it is difficult to sepa-rate the effects of cortisol from the consequences ofother factors that might contribute to this associa-tion such as shared genes. The current findings are,however, consistent with animal models where ran-dom assignment is possible (Kapoor et al., 2008 forreview) as well as human studies evaluating theconsequences of randomly occurring traumaticevents, such as natural disasters (LaPlante et al.,2004; Yehuda et al., 2005). Further, recent humanstudies have documented developmental conse-quences of prenatal stress among children conceivedby in vitro fertilization that were not geneticallyrelated to their mother (Rice et al., 2010). Second,maternal cortisol is an indirect measure of fetalcortisol exposure. There is evidence with humanfetuses, however, that maternal and fetal cortisollevels are significantly correlated, suggesting thatlevel of maternal cortisol is a valid index of fetalexposure (Gitau et al., 1998). Third, because of the



invasive nature of a blood draw, maternal cortisollevels were assessed only on one day during each ofthe five assessment intervals. The growth curvemodeling techniques used in this study considercortisol trajectories, enabling a more stable charac-terization of maternal cortisol across gestation.

Implications

The large and prolonged cortisol response and pro-tracted behavioral responses to the stress of painfulstimulation reported here may be indicative of a dys-functional stress system. This is particularly likelygiven the cost associated with an inability to turn offthe stress response after the stressor has passed(McEwen, 1998). Neonateswho aremore reactivemaycarry a greater risk for the development of behavioralinhibition and anxiety. There is evidence that infantsand toddlers who display greater physiological arou-sal to challenge tend to be more fearful or anxious inresponse to novel situations (Buss, Davidson, Kalin,& Goldsmith, 2004; Kagan, Snidman, & Arcus, 1998;Stansbury & Gunnar, 1994). Further, poor behav-ioral regulation during the neonatal period is predic-tive of increased stress reactivity and poor attentionand emotional regulation during infancy and child-hood (Geva & Feldman, 2008; Lewis, Worobey, &Thomas, 1989). It is plausible that maternal cortisolinfluenced the development of the fetal HPA axis,including alterations to receptor development. Aconsequence of these changes may be the inability torespond appropriately to stressors experienced in thepostnatal environment and is a potential mechanismthat may underlie fetal programming of later healthanddevelopment (Drake et al., 2007; Seckl&Meaney,2006). Longitudinal follow-up of this cohort willdetermine the developmental implications of altera-tions to neonatal stress responses.

Conclusion

Prenatal maternal signals facilitate fetal adaptationto the postnatal environment. Child and adult

functioning, including vulnerability to disease,appears to be determined, in part, by exposures thatoccur during the fetal period (Barker, 1998). Grow-ing evidence indicates that prenatal influences playa role in the development of psychiatric disorderssuch as anxiety, depression and externalizingbehavior problems (Bohnert & Breslau, 2008; Cos-tello et al., 2007; Hellemans, Sliwowska, Verma, &Weinberg, 2010). It has been proposed that disrup-tions to HPA axis functioning may be responsible forthis effect (Kapoor et al., 2008; Seckl & Meaney,2006). The data observed here, in a large prospectivestudy, are among the first with humans to demon-strate that prenatal maternal stress and stress hor-mones alter the functioning of stress regulatorysystems in the offspring, independent of postpartuminfluences, and may be a potential mechanism forfetal programming of later psychiatric disorders.Although the observed associations are likely notexplained by postnatal influences, it is highly likelythat the prenatal and postnatal environment willjointly determine developmental trajectories (e.g.,Bergman, Sarkar, Glover, & OConnor, 2008). It isour intention to continue to follow this cohort todetermine the effects of prenatal and early postnatalexperiences on the development of stress regulatorysystems.

Acknowledgements

This research was supported by grants from the NIH(HD-50662 to EPD and NS-41298 to CAS). Theauthors wish to thank the families who participatedin this project. The assistance of Cheryl Crippen,Jocelle Marucut, and Carol Holliday is gratefullyacknowledged.

Correspondence to

Elysia Poggi Davis, 333 City Blvd W., Suite 1200,Orange, CA 92868, USA; Tel: 714-940-1924; Fax:714-940-1939; Email: [email protected]

Key points

Retrospective studies have demonstrated that adult health outcomes are determined, in part, by prenatalexperiences.

Elevated maternal cortisol is associated with a larger infant cortisol response and a delayed behavioralrecovery to the painful stress of a heel-stick blood draw.

Elevated maternal psychosocial stress is associated with a prolonged behavioral response to the painfulstress of a heel-stick blood draw.

Increased physiological and behavioral reactivity is a risk factor for the development of behavioral inhi-bition and anxiety.

Prenatal maternal stress and stress hormones are a potential mechanism for fetal programming of laterpsychiatric disorders.



References

Abe, H., Hidaka, N., Kawagoe, C., Odagiri, K., Watanabe,Y., Ikeda, T., et al. (2007). Prenatal psychological stresscauses higher emotionality, depression-like behavior,and elevated activity in the hypothalamo-pituitary-adre-nal axis. Neuroscience Research, 59, 145151.

Als, H., Lester, B.M., Tronick, E., & Brazelton, T.B. (1988).Manual for the assessment of preterm infants behavior(APIB). In H. Fitzgerald, B.M. Lester, & M. Yogman (Eds.),Theory and research in behavioral pediatrics (Vol. 1, pp.65132). New York: Plenum.

Ashwood, P.J., Crowther, C.A., Willson, K.J., Haslam, R.R.,Kennaway, D.J., Hiller, J.E., et al. (2006). Neonataladrenal function after repeat dose prenatal corticoster-oids: A randomized controlled trial. American Journal ofObstetrics and Gynecology, 194, 861867.

Barker, D.J. (1998). In utero programming of chronicdisease. Clinical Science, 95, 115128.

Barker, D.J. (2002). Fetal programming of coronary heartdisease. Trends in Endocrinology and Metabolism, 13,364368.

Bergman, K., Glover, V., Sarkar, P., Abbott, D.H., &OConnor, T.G. (2010). In utero cortisol and testosteroneexposure and fear reactivity in infancy. Hormones andBehavior, 57, 306312.

Bergman, K., Sarkar, P., Glover, V., & OConnor, T.G.(2008). Quality of childparent attachment moderatesthe impact of antenatal stress on child fearfulness.Journal of Child Psychology and Psychiatry, 49, 10891098.

Bergman, K., Sarkar, P., OConnor, T.G., Modi, N., &Glover, V. (2007). Maternal stress during pregnancypredicts cognitive ability and fearfulness in infancy.Journal of the American Academy of Child and Adoles-cent Psychiatry, 46, 14541463.

Bohnert, K.M., & Breslau, N. (2008). Stability of psychiat-ric outcomes of low birth weight: A longitudinal investi-gation. Archives of General Psychiatry, 65, 10801086.

Brown, R.W., Diaz, R., Robson, A.C., Kotelevtsev, Y.,Mullins, J.J., Kaufman, M.H., et al. (1996). The ontogenyof 11b-hydroxysteroid dehydrogenase type 2 and miner-alicorticoid receptor gene expression reveal intricatecontrol of glucocorticoid action in development. Endocri-nology, 137, 794797.

Buss, K., Davidson, R.J., Kalin, N.H., & Goldsmith, H.H.(2004). Context specific freezing. Developmental Psychol-ogy, 40, 583594.

Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A globalmeasure of perceived stress. Journal of Health and SocialBehavior, 24, 385396.

Costello, E.J., Worthman, C., Erkanli, A., & Angold, A.(2007). Prediction from low birth weight to femaleadolescent depression: A test of competing hypotheses.Archives of General Psychiatry, 64, 338344.

Davis, E.P., Glynn, L.M., Dunkel Schetter, C., Hobel, C.,Chicz-DeMet, A., & Sandman, C.A. (2007). Prenatalexposure to maternal depression and cortisol influencesinfant temperament. Journal of the American Academy ofChild and Adolescent Psychiatry, 46, 737746.

Davis, E.P., & Sandman, C.A. (2010). The timing ofprenatal exposure to maternal cortisol and psychosocialstress is associated with human infant cognitive devel-opment. Child Development, 81, 131148.

Davis, E.P., Snidman, N., Wadhwa, P.D., Dunkel Schetter,C., Glynn, L., & Sandman, C.A. (2004a). Prenatalmaternal anxiety and depression predict negative behav-ioral reactivity in infancy. Infancy, 6, 319331.

Davis, E.P., Townsend, E.L., Gunnar, M.R., Georgieff,M.K., Guiang, S.F., Cifuentes, R.F., et al. (2004b). Effectsof prenatal corticosteroid exposure on regulation ofstress physiology in healthy premature infants. Psycho-neuroendocrinology, 29, 10281036.

Davis, E.P., Townsend, E.L., Gunnar, M.R., Guiang, S.F.,Lussky, R.C., Cifuentes, R.F., et al. (2006). Antenatalbetamethasone treatment has a persisting influence oninfant HPA axis regulation. Journal of Perinatology, 26,147153.

de Weerth, C., & Buitelaar, J.K. (2005). Physiologicalstress reactivity in human pregnancy a review. Neuro-science and Biobehavioral Reviews, 29, 295312.

de Weerth, C., van Hees, Y., & Buitelaar, J. (2003).Prenatal maternal cortisol levels and infant behaviorduring the first 5 months. Early Human Development,74, 139151.

Drake, A.J., Tang, J.I., & Nyirenda, M.J. (2007). Mecha-nisms underlying the role of glucocorticoids in the earlylife programming of adult disease. Clinical Science, 113,219232.

Dunkel Schetter, C., & Glynn, L. (2010). Stress in preg-nancy: Empirical evidence and theoretical issues toguide interdisciplinary research. In R. Contrada, & A.Baum (Eds.), Handbook of stress science (Ch. 24). NewYork: Springer.

Geva, R., & Feldman, R. (2008). A neurobiological model forthe effects of early brainstem functioning on the devel-opment of behavior and emotion regulation in infants:Implications for prenatal and perinatal risk. Journal ofChild Psychology and Psychiatry, 49, 10311041.

Gitau, R., Cameron, A., Fisk, N.M., &Glover, V. (1998). Fetalexposure to maternal cortisol. Lancet, 352, 707708.

Gitau, R., Fisk, N.M., Teixeira, J.M., Cameron, A., &Glover, V. (2001). Fetal hypothalamicpituitaryadrenalstress responses to invasive procedures are independentof maternal responses. Journal of Clinical Endocrinologyand Metabolism, 86, 104109.

Glynn, L.M., Dunkel Schetter, C., Hobel, C., & Sandman,C.A. (2008). Pattern of perceived stress and anxiety inpregnancy predict preterm birth. Health Psychology, 27,4251.

Grant, K., McMahon, C., Austin, M., Reilly, N., Leader, L.,& Ali, S. (2009). Maternal prenatal anxiety, postnatalcaregiving and infants cortisol responses to the still-faceprocedure. Developmental Psychobiology, 51, 625637.

Gunnar, M.R. (1989). Studies of the human infantsadrenocortical response to potentially stressful events.New Directions for Child Development, 45, 318.

Gunnar, M.R., Hertsgaard, L., Larson, M., & Rigatuso, J.(1992). Cortisol and behavioral responses to repeatedstressors in the human newborn. Developmental Psy-chobiology, 24, 487505.

Gutteling, B.M., De Weerth, C., & Buitelaar, J.K. (2004).Maternal prenatal stress and 46 year old childrenssalivary cortisol concentrations pre- and post-vaccina-tion. Stress, 7, 257260.

Gutteling, B.M., de Weerth, C., & Buitelaar, J.K. (2005).Prenatal stress and childrens cortisol reaction to the firstday of school. Psychoneuroendcrinology, 30, 541549.

Hellemans, K.G., Sliwowska, J., Verma, P., & Weinberg, J.(2010). Prenatal alcohol exposure: Fetal programmingand later life vulnerability to stress, depression andanxiety disorders. Neuroscience and Biobehavioral Re-views, 34, 791807.

Herman, J.P., & Cullinan, W.E. (1997). Neurocircuitry ofstress: Central control of the hypothalamio pituitary adrenocortical axis. Trends in Neuroscience, 20, 7884.



Hobel, C.J. (1982). Identifying the patient at risk. In R.J.Bolognese, R.H. Schwartz, & J. Schneider, (Eds), Peri-natal medicine: Management of the high risk fetus andneonate (pp. 328). Baltimore, MA: Williams & Wilkins.

Jacobson, L., & Sapolsky, R. (1991). The role of thehippocampus in feedback regulation of the hypothalamicpituitary adrenocortical axis. The Endocrine Reviews, 12,118134.

Kagan, J., Snidman, N., & Arcus, D. (1998). Childhoodderivatives of high and low reactivity in infancy. ChildDevelopment, 69, 14831493.

Kapoor, A., Dunn, E., Kostaki, A., Andrews, M.H., &Matthews, S.G. (2006). Fetal programming of hypo-thalamo-pituitary-adrenal function: Prenatal stress andglucocorticoids. Journal of Physiology, 572(Pt 1), 3144.

Kapoor, A., Petropoulos, S., & Matthews, S.G. (2008). Fetalprogramming of hypothalamicpituitaryadrenal (HPA)axis function and behavior by synthetic glucocorticoids.Brain Research Reviews, 57, 586595.

Keenan, K., Gunthorpe, D., & Grace, D. (2007). Parsing therelations between SES and stress reactivity: Examiningindividual differences in neonatal stress response. InfantBehavior and Development, 30, 134145.

Kirschbaum, C., & Hellhammer, D.H. (1989). Salivarycortisol in psychobiological research: An overview. Neu-ropsychobiology, 22, 150169.

LaPlante, D.P., Barr, R.G., Brunet, A., Du Fort, G.G.,Meaney, M.J., Saucier, J.F., et al. (2004). Stress duringpregnancy affects general intellectual and languagefunctioning in human toddlers. Pediatric Research, 56,400410.

Levitt, N.S., Lindsay, R.S., Holmes, M.C., & Seckl, J.R.(1996). Dexamethasone in the last week of pregnancyattenuates hippocampal glucocorticoid receptor geneexpression and elevates blood pressure in the adultoffspring in the rat. Neuroendocrinology, 64, 412418.

Lewis, M., Worobey, J., & Thomas, D. (1989). Behavioralfeatures of early reactivity: Antecedents and conse-quences. New Directions for Child Development, 45, 3346.

Maccari, S., & Morley-Fletcher, S. (2007). Effects ofprenatal restraint stress on the hypothalamuspitui-taryadrenal axis and related behavioural and neurobi-ological alterations. Psychoneuroendocrinology,32(Suppl. 1), S1015.

Marcus, S.M., Flynn, H.A., Blow, F.C., & Barry, K.L.(2003). Depressive symptoms among pregnant womenscreened in obstetrics settings. Journal of WomensHealth, 12, 373380.

McEwen, B.S. (1998). Stress, adaptation, and disease.Allostasis and allostatic load. Annals of the New YorkAcademy of Sciences, 840, 3344.

Murphy, V.E., & Clifton, V.L. (2003). Alterations in humanplacental 11beta-hydroxysteroid dehydrogenase type 1and 2 with gestational age and labour. Placenta, 24,739744.

Nathanielsz, P.W. (1999). Life in the womb: The origin ofhealth and disease. Ithaca, NY: Promethean Press.

Noorlander, C.W., De Graan, P.N., Middeldorp, J., VanBeers, J.J., & Visser, G.H. (2006). Ontogeny of hippo-campal corticosteroid receptors: Effects of antenatalglucocorticoids in human and mouse. Journal of Com-parative Neurology, 499, 924932.

OConnor, T.G., Ben-Shlomo, Y., Heron, J., Golding, J.,Adams, D., & Glover, V. (2005). Prenatal anxiety predictsindividual differences in cortisol in pre-adolescent chil-dren. Biological Psychiatry, 58, 211217.

Raschke, C., Schmidt, S., Schwab, M., & Jirikowski, G.(2008). Effects of betamethasone treatment on centralmyelination in fetal sheep: An electron microscopicalstudy. Anatomy, Histology and Embryology, 37, 95100.

Raudenbush, S.W., & Bryk, A.S. (2002). Hierarchical linearmodels: Application and data analysis methods (Vol. 1).Thousand Oaks, CA: Sage.

Rice, F., Harold, G.T., Boivin, J., van den Bree, M., Hay,D.F., & Thapar, A. (2010). The links between prenatalstress and offspring development and psychopathology:Disentangling environmental and inherited influences.Psychological Medicine, 40, 335345.

Sanchez, M.M., Young, L.J., Plotsky, P.M., & Insel, T.R.(2000). Distribution of corticosteroid receptors in therhesus brain: Relative absence of glucocorticoid recep-tors in the hippocampal formation. The Journal ofNeuroscience, 20, 46574668.

Sandman, C.A., & Davis, E.P. (2010). Gestational stressinfluences cognition and behavior. Future Neurology,5(5), 675690.

Sandman, C.A., Glynn, L.M., Dunkel Schetter, C., Wadwha,P.D., Garite, T., & Hobel, C. (2006). Elevated maternalcortisol early in pregnancy predicts third trimester levelsof placental corticotropin releasing hormone (CRH):Priming the placental clock. Peptides, 27, 14571463.

Santor, D.A., & Coyne, J.C. (1997). Shortening the CES-Dto improve its ability to detect cases of depression.Psychological Assessment, 9, 233243.

Schneider, M.L. (1992). Prenatal stress exposure alterspostnatal behavioral expression under conditions ofnovelty challenge in rhesus monkey infants. Develop-mental Psychobiology, 25, 529540.

Seckl, J.R. (2008). Glucocorticoids, developmental pro-gramming and the risk of affective dysfunction. Progressin Brain Research, 167, 1734.

Seckl, J.R., & Meaney, M.J. (2006). Glucocorticoid pro-gramming and PTSD risk. Annals of the New YorkAcademy of Sciences, 1071, 351378.

Setiawan, E., Jackson, M.F., MacDonald, J.F., & Mat-thews, S.G. (2007). Effects of repeated prenatal gluco-corticoid exposure on long-term potentiation in thejuvenile guinea-pig hippocampus. Journal of Physiology,581, 10331042.

Sidman, R.L., & Rakic, P. (1973). Neuronal migration, withspecial reference to developing human brain: A review.Brain Research, 62, 135.

Speilberger, C. (1983). State-Trait Anxiety Inventory. Red-wood City, CA: Mind Garden.

Stansbury, K., & Gunnar, M.R. (1994). Adrenocorticalactivity and emotion regulation. Monographs of theSociety for Research on Child Development, 59, 108134.

Van den Bergh, B. (1990). The influence of maternalemotion during pregnancy on fetal and neonatal behav-ior. Pre- and Peri-natal Psychology, 5, 119130.

Van den Bergh, B.R.H., Van Calster, B., Smits, T., VanHuffel, S., & Lagae, L. (2007). Antenatal maternalanxiety is related to HPA-axis dysregulation and self-reported depressive symptoms in adolescence: A pro-spective study on the fetal origins of depressed mood.Neuropsychopharmacology, 33, 536545.

Yehuda, R., Engel, S.M., Brand, S.R., Seckle, J., Marcus,S.M., & Berkowitz, G.S. (2005). Transgenerational effectsof posttraumatic stress disorder in babies of mothersexposed to the World Trade Center attacks duringpregnancy. Journal of Clinical Endocrinology and Metab-olism, 90, 41154118.

Manuscript accepted 14 July 2010



j25 davis 2011 prenatal stress & hpa.pdf

Documents