Blunted endocrine and cardiovascular reactivity in young healthy women reporting a history of childhood adversity

Blunted endocrine and cardiovascular reactivity in young healthy women

reporting a history of childhood adversity

Annette Voellmin

, Katja Winzeler

, Evelin Hug

, Frank H. Wilhelm

, Valérie Schaefer

, Jens Gaab

, Roberto La Marca

, Jens C. Pruessner

, Klaus Bader

aPsychiatricClinicsoftheUniversityofBasel,CenterforSpecificPsychotherapy,CBTUnit,Wilhelm Klein-Strasse27,4012Basel,Switzerland

bDepartmentofPsychology,DivisionofClinicalPsychologyandPsychotherapy,UniversityofBasel, Missionsstrasse60/62,4055Basel,Switzerland

cUniversitySalzburg,DepartmentofPsychology,DivisionofClinicalPsychology,Psychotherapy,andHealth Psychology,Hellbrunnerstrasse34,5020Salzburg,Austria

dDepartmentofPsychology,DivisionofClinicalPsychologyandPsychotherapy,UniversityofZurich, Binzmuehlestrasse14/26,8050Zurich,Switzerland

eDepartmentsofPsychiatry,Neurology,andNeurosurgery,McGillUniversity&DouglasHospitalResearch Centre,Montreal,QC,Canada

KEYWORDS Hypothalamus—

pituitary—adrenal axis;

Sympatheticnervous system;

Adversechildhood experiences;

Female;

Stressreactivity;

Resilience;

Trauma

Summary

Background: Chronic or prolonged stress exposure in childhood can alter structural and functionalbraindevelopment,leadingtomentalandphysicalillnessandalterationsofpsy- chobiological stress systems in adulthood.Recently, attenuationin stress reactivityof the hypothalamic—pituitary—adrenal (HPA)axisandcardiovascular systemhavebeenrelated to thenumberofadversechildhoodexperiences(ACEs).Wesetouttoinvestigatetheassociation ofACEdurationandageofACEoccurrenceonstressreactivity.

Methods:104womenintheagerange18—25years(mean=21.7)freeofmentalandphysical illnessunderwentpsychosocialstresstestingwiththeMontrealImagingStressTask(MIST).Free salivacortisolandheartratewereassessedrepeatedlybeforeandaftertheMIST.

Results:Number ofACEswas associated with attenuated cortisol andheart rateresponses tostressinadose-responserelationship.Whereas overalldurationofACEs wassignificantly associatedwithanattenuatedcortisolresponse,thespecificageoffirstACEoccurrencedid notcontributefurthertothedampenedstressresponse.

∗Correspondingauthor.Tel.:+41613255119;fax:+41613257671.

E-mailaddress:Klaus.Bader@upkbs.ch(K.Bader).

Konstanzer Online-Publikations-System (KOPS) URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-2-tvnfu5latka6

https://dx.doi.org/10.1016/j.psyneuen.2014.09.008

Conclusions: ACEsareassociatedwithbluntedendocrineandcardiovascularstressreactivityin youngandhealthywomen.Adverselifeeventsinchildhood,particularlyiftheyoccurrepeat- edlyandchronically,showastrongassociationwithalterationsinstressreactivityinadulthood, potentiallypredisposingforlatermentalorphysicaldisorders.

©2014ElsevierLtd.Allrightsreserved.

1. Introduction

Adverse childhood experiences (ACEs), including physi- cal, emotional and sexual abuse, affect a significant portion of the population and have been shown to be risk factors for the development and persistence of mental disorders such as depression, anxiety disorders, substance usedisorders,or attention-deficit/hyperactivity disorder.Changesinstresssensitivityandfunctioningofthe hypothalamic—pituitary—adrenal(HPA)axishavebeensug- gestedascausalfactors(DeBellis,2002;TarulloandGunnar, 2006;Heim etal.,2008).TheHPAaxis,togetherwith the sympatheticnervous system (SNS), arekey players in the formation of the stress response (Chrousos, 2009). Mod- elsofstressreactivityandhealthoutcomeshavetherefore beenamainfocusinresearchtowardsunderstandingadver- sityandresilienceprocesses.Differentreactivityphenotype patternshaveemergedindifferentstudies,withanumber ofstudiesshowingexaggeratedHPAaxisandSNSresponses in thecontext ofACEsand psychopathology(Heim etal., 2000b;HeimandNemeroff,2001;Bremneretal.,2003;Rao etal.,2008).

In contrast to findings in clinical samples, a growing numberofstudiesonhealthyparticipantsreportedblunted endocrine(Carpenteretal.,2007,2011;Elzingaetal.,2008;

Lovalloetal.,2012)andcardiovascular(Lovalloetal.,2012) responsesinassociationwithACEs.Forexample,Carpenter etal.(2007)reportedbluntedplasmacortisolresponsesto a psychosocial stress test in a healthy sample with a his- toryofACEscomparedtoparticipantswithoutahistoryof childhood maltreatment. Also, a recent study by Lovallo et al.(2012) showed in a large sample of healthy partic- ipants diminished cortisol aswell as heart rateresponses withanincreasingnumberofadverselifeevents,indicating aninversedose-responserelationshipofACEsandreactivity toamentalstresstest.

To the best of our knowledge, the association of age of occurrence and durationof ACEshas not been investi- gatedinhealthysamplessofar.Thereisevidencethatthese factorscouldhave adifferentialimpactonstress reactiv- ityinadulthood(TarulloandGunnar,2006;Schoedletal., 2010; Tottenham and Sheridan, 2010). Sincemost studies usedbriefself-reportquestionnairesorlifeeventschecklists toassessACEs(e.g.Carpenteretal.,2011;Lovalloetal., 2012), questions of durationof the events or the respec- tive agewhen theyhappened widelyremain unanswered.

Froma developmentalperspective, ageat traumatization isbelieved tobeanimportantfactor(Boschetal.,2012).

Brain components involved in stress response show large plasticity during pre- and postnatal periods and during early childhood, and some plasticity during later child- hood and adolescence (Fumagalli et al., 2007; Andersen et al., 2008). Also, the duration of adverse experiences

couldbe associatedwith psychobiologicalconstructs.Par- ticularlythoseenvironmentaleventsthatcauseexceeding or prolongedstimulationofthestresssystemduringthese criticaldevelopmentalperiodscouldbelinkedtoabnormal neurodevelopmentandthereforeberiskfactorsforlasting alterationsin stressreactivityoftheHPAaxisandtheSNS (Schoedletal.,2010).

Therefore, the aim of the present study wasto repli- catethefindingsofattenuatedendocrineandcardiovascular stress reactivity in association with a history of ACEs in a young, healthy, female adult sample. Furthermore, we aimedtoinvestigatetheassociationofdurationaswellas age of occurrenceof adverselife events inchildhoodand adolescencewiththestressreactivityinadulthood.Toelu- cidate theassociation of age ofoccurrence andduration, in thisstudy,theEarly TraumaInventory-Self Report(ETI- SR) served asa more detailedmethod in measuring ACEs (Bremner etal., 2007). The questionnaire retrospectively assessesawiderangeofstressandtraumaexposurebefore the age of 18 andconsiders age of occurrenceaswell as durationoftheevents.

In contrast to Lovallo et al. (2012) who employed standard public speakingand mentalarithmetic stressors, and in contrast to the standardly used Trier social stress test (TSST, (Kirschbaum et al., 1993), we used the Mon- treal imagingstress task(MIST), that has been developed tobecompatiblewithfunctionalmagneticresonancebrain imaging(Dedovic etal.,2005),butcanbeusedinlabora- torystressstudiesaswell(LaMarcaetal.,2011).TheMIST offersapromisingalternativetoconventionalpsychosocial stresstests.AdvantagesoftheMISTarethatparticipantssit stillduringthestressprotocolanddonotspeak,therefore causingfewerartifactsmeasuringpsychobiologicalparame- terslikeheartrateorelectrodermalactivity.Furthermore, sincetheMISTisconsideredtoevokeamoderateendocrine stressresponse,theobservedfindingsrepresentHPAactivity closetonaturalisticsettings(Smythetal.,1998).

2. Methods and materials

Thesampleincluded104youngandhealthyfemalesinthe ageof18to25years(M=21.7;SD=1.5),recruitedatthree schoolsforhealthcareprofessionsandsocialworkinBasel, Switzerland. The sample waspart of an ongoing project, which included only femaleparticipants. The recruitment material referredtoa 14-daysleepassessment describing thenatureandprevalenceofsleepandsleepingdisorderand didnotincludeexplicitstatementsaboutchildhoodtrauma oradversechildhoodexperiences.

Exclusion criteria were current physical or psychiatric illness, pregnancy, regular and heavy tobacco use (>5

cigarettesaday),theconsumptionofillegaldrugs,andthe useofmedicationthatinterfereswiththecentralnervous ortheadrenocorticoidsystem.

Furthermore, participants were requested tominimize physical exercise during the hour preceding the labora- tory examination and not totake large meals, coffee, or cigarettes.Forparticipants taking nooral contraceptives, thelaboratory assessmentwasheldin thelutealphaseof theparticipant’smenstrualcycle(Kirschbaumetal.,1999).

Participants receivedmonetary compensation for their participationandprovidedwritteninformedconsentprior toparticipation.The ethical principles ofthe Declaration of Helsinki werefollowed andthe study wasaccepted by thelocalEthicsCommittee.Allappointmentstookplacein alaboratory ofthe PsychiatricClinics ofthe Universityof Basel,Switzerland.

2.2. Procedure

After a preliminary screening assessment, participants reportedtothelaboratoryforthestressexamination,which took place between 3:30 pm and 6:00 pm to control for circadianvariationandlastedforapproximately2.5h.Par- ticipants were told that the laboratory assessment would includeatestoncognitiveperformance.

Uponarrival,participantswereseatedinacomfortable chairin frontof a tablewith a computerscreen andsev- eralmagazines.Aftertheheartratesensorswereattached, aten minuteresting periodfollowed tocustomizepartic- ipantswith thelaboratory.Then,abaselinemeasurement was conducted for five minutes. Immediately before the baselinemeasurement,participantsprovidedthefirstsaliva sample. Anothersaliva sample wascollectedimmediately beforeparticipantsengagedintheMIST.Followingthestress exposure,arecoveryperiodwasconductedduringwhichfive additionalsalivarycortisolsampleswerecollectedtogether with self-report measures of the participants’ emotional responsetothestresstask.Attheendofthelaboratorytest- ing,participantsweredebriefedandsignedasecondwritten informedconsenttoapprovethefurtheruseoftheirdata.

2.3. Stressinduction

TheMIST(LaMarcaetal.,2011)wasusedtoinduceapsy- chosocialstressresponse.TheMIST(Dedovicetal.,2005)is astandardizedpsychosocialstresstestduringwhichpartici- pantshavetosolvearithmetictasksdisplayedonacomputer screenundertimepressureandsocialevaluation.Thesoft- ware adapts the difficulty of the tasks to the individual performance level of each participant, so that it is not possibletocorrectlyanswermorethan45—50%ofthearith- metictasksintheexperimentalcondition.Participantshad tocompletethreeexperimentalruns,eachlastingfourmin- utes.

Toinduceasocialevaluativethreat,participantsaretold thattheirperformancehastobecloseorequaltotheaver- age performanceof a normative sample that is shown on thetopofthecomputerscreen, andthatotherwise,their datacannotbeusedfortheresearchpurposes.Also,partic- ipantsareinformedthatthestudyleaderiswatchingtheir performancenextdoor.Furthermore,aftereachofthefirst

tworuns,tofurtherenhancesocialevaluative threat,the studyleaderinformsparticipantsthattheirperformanceis poor.Inconcreteterms,afterthefirstexperimentalrun,the experimentercallsthestudyleadertoaskwhattodoinsuch anunusualsituation.Theparticipantthenistoldtorepeat thetaskandtodobetter.Then,afterthesecondunsuccess- fulperformance,theslightlyannoyedanddemandingstudy leader enters the laboratory andinterrogates thepartici- pantaboutherindividualreasonsforherpoorperformance.

Thestudyleaderexplainsthehighcostsoftheexperiment in case ofa possible exclusionifthe participantdoes not achieveabetterperformance.Then,thelastrunoftheMIST startswhilethestudyleaderstaysintheroomandwatches thesubject’sperformancestandingrightbehindher.

Thesebehaviorswerestandardizedandpracticedbefore studyonset.

2.4. Measures

2.4.1. Biologicalmeasures

Salivawascollectedatsevenmeasurementpoints,whereof twotookplacebeforethestresstest(−10and−1min)and five after the stress test (+1, +10, +25, +40 and +55min) usingsalivettes(Sarstedt,Sevelen,Switzerland).Allsaliva samples were first stored at −22^◦C, then thawed and centrifuged at 3000rpm, before cortisol concentration in saliva was determined by enzyme immunoassay (ALPCO Diagnostics,Salem,USA).Becauseofunexpectedhighval- ues, cortisol concentration was reanalyzed with a more established commercially available chemiluminescence immunoassaywith highsensitivity(IBLInternational,Ham- burg,Germany).The intra-andinterassaycoefficientsfor cortisolwerebelow8%.

HeartratewasrecordedusingVitaport3dataacquisition system(TEMECInstrumentsB.V.,Netherlands).Electrocar- diogram(ECG)recordingsweretakenusingLead-IIelectrode placement(RedDot^TM, 2248-50,3F HealthCare,Germany) onthethoraxwiththreedisposableelectrodes.Asampling rateof1024HzwasusedforECGrecordingswithalowpass filter of 512Hz and a high pass filterof 0.5Hz. Anslab, a software for scientific analysis of physiological data, was usedtoanalyzedetectedconsecutiveR-wavesandcalculate R—Rintervals,whichweretransformedtoheartrate(Auto- nomic nervous system laboratory, Wilhelm& Peyk, 2005).

Heart rate wasaveraged for Baseline, Stress 1, Stress 2, Stress3,andRecoveryperiodsinreferencetotimemarkers manuallysetinaccordancewiththevarioussectionsofthe experiment.

2.4.2. Psychologicalmeasures

A diagnostic screening including the German version of the‘‘StructuredClinicalInterviewfor DSM-IV/AxisIDisor- ders’’(SCID-I)wasconductedinordertodetectandexclude participants suffering from a mental disorder (Wittchen et al., 1997). Lifetime history of mental disorders (n=9) wasassessed.Participants reportedtohave sufferedfrom depression(n=7),panicdisorder(n=1),andposttraumatic stress disorder comorbid with depression(n=1) and were fullyremittedatthetimeofthestudy.Relevantdatainclud- ingage, medication,drugconsumption,age ofmenarche,

Table1 ParticipantcharacteristicsandACEscoresofthestudysample(N=104).

Variable M SD Range/%

Age[yr] 21.66 1.54 18—25

Ageofonsetmenarche^a[yr] 12.95 1.28 10—16

Oralcontraceptiveuse,n(%) 59(56.7)

BodyMassIndex[kg/m²] 21.81 2.53 18.37—31.14

Depressivesymptoms(ADS-K) 7.08 4.93 0—24

ACEtotalsumscore 2.76 3.17 0—15

Generaltrauma 1.43 1.63 0—7

Physicalabuse .53 .99 0—5

Emotionalabuse .54 1.25 0—6

Sexualabuse .26 .57 0—3

ACEs≥1year 1.27 2.40 0—12

ACEs<1year 1.44 1.55 0—6

ACEsbeforemenarche 1.96 2.99 0—15

ACEsaftermenarche .76 1.05 0—6

AgeoffirstACEoccurrence^b 7.54 4.83 0—16

a n=103.

b n=80(noexperiencedACE,n=23;noreportedageofoccurrence,n=1).

BMI,dateoflastmenstruation,andintakeofhormonalcon- traceptiveswerealsoassessedduringtheinterview.

ACEs before the age of 18 years were assessed ret- rospectively using a German translation of the ‘‘Early Trauma Inventory-Self Report’’ (ETI-SR) (Bremner et al., 2007), whichincludes generaltrauma (31items),physical (9items),emotional(7items),andsexualabuse(15items).

TheETI-SRhasbeendemonstratedtobeavalidmeasureof earlytrauma,andhasshownhighinternalconsistencyinall traumadomains(Cronbach˛>0.7)(Bremneretal.,2007).

Participants were asked a seriesof questions concern- ing potential trauma and stress exposure, which they answered with yes or no. Next, on positively answered items,ageofoccurrence,frequencyoftraumaorabuse,and emotional impact(0=nonegativeimpact, 1=slightlyneg- ative, 2=moderatelynegative, 3=stronglynegative) were assessed.Intotal,fivedifferentACEscoreswerebuilt.First, a sum scorewascomputed fromall events rated with an emotionalimpactofatleast1(ACEtotalsumscore).Fur- thermore, a sum score for ACEs lasting less than a year (ACEs<one year) and for ACEs lasting more than a year (ACEs>one year) was computed. Age of occurrence was assessedintwoways.Inafirststep,eventswhichoccurred beforeorafteraparticipants’menarcheweresummedup toACEsbeforeandaftermenarche,respectively.Next,age of first ACE occurrence was abstracted for each subject, whiletheemotionalimpactofatleast‘‘slightlynegative’’

wasconsidered.ACEmeanscores ofthestudysample are depictedinTable1.

BecausetheETI-SR does not providecut-off scores for grouping,forillustrationpurposes,evenlydistributedquar- tilegroups(ACEtotalgroups)werebuiltviarankfunctionof SPSSfortheACEtotalsumscore.Thegroupingviarankfunc- tionresultedinthefollowinggroupdistributions:group1=0 ACE,group2=1ACE,group3=2—3ACEs,andgroup4=4or moreACEs.Groupswithregardtodurationwerethenbuilt accordingtothesamegroupdistributionasfortheACEtotal groups.

DepressivesymptomatologywasassessedviatheGerman version of the Center for Epidemiological Studies Depres- sionScale(CES-D;Germanversion:ADS-K;(Hautzingerand Bailer,1993).

Visualanalogscales(VAS)for mood,tension,andstress served as measures of subjective emotional response of participantsduringthepsychosocialstresstest.Thescales rangedfrom‘‘notstressed’’(0)to‘‘verystressed’’(100), experiencing‘‘notension’’(0)to‘‘extremetension’’(100), and ‘‘having a goodmood’’ (0) to‘‘having a bad mood’’

(100),respectively.

2.5. Dataanalysis

Statistical analyses wereperformed usingIBMSPSS Statis- tics,version20(SPSSInc.,Chicago,IL).Descriptivestatistics were conductedforall variables.Skewed datawereloga- rithmicallytransformedwhereappropriate.

First,repeatedmeasuresgenerallinearmodel(GLM)was used toassessif thestress taskledtoa significant stress responseforthedependentvariablessalivarycortisol,heart rate,aswellasforthesubjectiveemotionalresponsestothe MIST. Next,GLMs for repeated measures served todeter- mine theeffectsof ACEsonendocrineandcardiovascular responses.Inthesemodels,theACEtotalsumscoreaswell asthedifferentscoresfordurationandageofoccurrence were used as continuous variables to examine effects of time,ACEscores,andtheinteractionoftimebyACEscores.

Inasecond step,inorder tovisualizetheresults,thedif- ferentACE groupswerethen usedasfixedfactorsfor the GLMs,respectively.

To protect against violation of sphericity, Greenhouse—Geisser corrections were applied where appropriate. Effect sizes were determined by partial eta-square, reflectingsmall (.01), medium (.06),or large (.14)effectsizes(Greenetal.,2000).

Toaccountfortheirpotentialconfoundinginfluenceon cortisol concentration (Kirschbaum etal., 1999), BMI and

use of oral contraceptives were included ascovariates in allstatisticalmodels.Inthissample,depressivesymptoms wereoveralllowandwereneitherrelatedtocortisol,heart ratenorACEsandthereforenotcontrolledforintheanal- yses.Also, lifetimehistoryof mental disorders(n=9) was not related to the outcome measures and therefore not controlledforin theanalyses.Emotional responsestothe MIST(mood,tension,stress)wereenteredascovariatesin post-hocanalyses,usingthetrapezoidformulaforcalcula- tion(areaunderthe curvewith respecttoground, AUCg;

Pruessneretal.,2003).

TechnicaldifficultieswithVitaport3dataacquisitionsys- temledto dataloss inheart rate measurements (missing completelyatrandom).Eventually,heartratemeasuresof 88participantswereavailableandwentintotheanalyses.

Cortisoldataof foursubjects hadtobeexcluded because ofunlikelyhighandfluctuatingvalues,orbecauseofacute illness,andtherefore,cortisolmeasuresof100participants wentintotheanalyses.

3. Results

Demographicandtraumacharacteristicsofthesampleare displayed in Table 1. According to univariate analyses of variance, the ACE groups did not differ significantly in terms of demographic characteristics (e.g. age, age of onsetofmenarche,oralcontraceptiveuse,anddepressive symptoms).However,fortheACEtotalgroups,BMIwassig- nificantlyhigher[p=.02]inwomenreporting4ormoreACEs (M=23.38,SD=3.55)comparedtowomenreportingnoACEs (M=21.16,SD=2.07).

ResultsobtainedbyGLMrepeatedmeasureanalysesindi- catedthatthestresstaskinduced arobustandsignificant increaseincortisollevels[F(1.87,185.52)=27.16,p<.001;

p2=.22]andheartrate[F(1.85,164.95)=216.86,p<.001;

p2=.71]. Subjects experienced significant worsening of mood [F(3.53, 360.36)=31.72, p<.001; p2=.24], as well as increases in tension [F(3.78, 385.39)=32.82, p<.001;

p2=.24], and stress [F(3.90, 398.15)=28.45, p<.001;

p2=.22](Fig.1).Regardingtheiremotionalreactiontothe MIST, participants did notshow differences in their base- lineandpeaklevelsinrelationtothetotalnumberofACEs, asindicated byunivariate analysesof variances(datanot shown).

Furthermore, the associations of the subjective emo- tional responses to the MISTwith cortisol and heart rate responsesweretested.Resultsrevealednosignificantcorre- lations(allp>1).DepressionsymptomscoresandACEtotal sumscorewereuncorrelated(r=.11,p=.27).

3.1. AssociationofACEandcortisolresponsesto stress

Repeatedmeasuresanalysisof cortisolresponsestostress showed a significant interaction of time×ACE total sum score[F(2.33,221.60)=5.89,p<.001;p2=.06]aswell as a significant main effect of ACE total sum score [F(1, 95)=7.52, p<.01; p2=.07]. Results remained significant when the emotional responses to the stress task were enteredadditionallyascovariates.Results aredepictedin Fig.2.

Fig. 1 Subjective emotional responses to the MIST.

(^aTimepoints relative to the MIST: t1=−1min, t2=+1min, t3=+10min,t4=+25min,t5=+40min,t6=+55min.).

Next, it was tested whether the durationof ACEswas associated with the cortisol responses tothe stress task.

Repeatedmeasuresanalysisofcortisolresponsetothestress task resulted in a significant main effect of duration of ACEs>oneyear[F(1,95)=10.10,p<.01;p2=.10]andasig- nificantinteractionoftimeanddurationofACEs>oneyear [F(2.33, 221.16)=5.36, p<.01; p2=.05] (Fig. 3). Results remainedsignificant whentheemotional responses tothe stress task wereentered additionally ascovariates. How- ever, these effects were notobserved for the association between ACEsthat lasted shorter in duration (ACEs<one year) and cortisol responses to the stress task [main

Fig.2 CortisolresponsestotheMISTaredepictedforwomen whoexperienced0,1,2—3or4ormoreACEs.Valuesrepresent averagecortisol±standarderrorofthemeanfortheACEtotal groupscontrolled for oral contraceptiveuseand BMI. Higher ACEsumscoresareassociatedwithbluntedcortisolresponses.

(^a Timepoints relative to the MIST: t1=−1min, t2=+1min, t3=+10min,t4=+25min,t5=+40min,t6=+55min.).

Fig.3 CortisolresponsestotheMISTaredepictedforwomen whoexperienced0,1,2—3orupto4ormoreACEsthatlasted more than ayear. Values represent cortisol±standard error ofthemeanfortheACEs>oneyeargroupcontrolledfororal contraceptive useand BMI. Higher ACE sumscores areasso- ciatedwithbluntedcortisolresponses.(^aTimepointsrelative totheMIST:t1=−1min,t2=+1min,t3=+10min,t4=+25min, t5=+40min,t6=+55min.).

effect, F(1, 95)=.64, p=.43; interaction effect, F(2.24, 213.21)=1.05,p=.36].

Forage of occurrence, a significant interaction effect [F(2.33, 220.99)=6.48, p<.01; p2=.06], and a signifi- cant main effect were observed for the sum of events whichoccurred beforemenarche [F(1,95)=10.26, p<.01;

p2=.09].Foreventswhichoccurredaftermenarche,these effects were not observed [main effect, F(1, 95)=.18, p=.67;interactioneffect,F(2.24,212.34)=.63,p=.55].For thespecificageoffirstACEoccurrence,nosignificantasso- ciationswereobserved[maineffect,F(1,78)=71,p=.40;

interactioneffect,F(2.24,174.75)=.56,p=.58].

3.2. AssociationofACEandheartratereactivity tostress

Repeatedmeasures analysisof heart rateresponse tothe stress task showed a significant main effect of ACE total sum score [F(1, 85)=7.13, p<.01; p2=.08] as well as a significantinteractioneffectoftime×ACEtotalsumscore [F(1.98,168.32)=5.86,p<.01;p2=.07].Resultsremained significant,whentheemotionalresponsestothestresstask wereenteredascovariates(Fig.4).

However, thefurtheranalyses with durationaswell as ageofoccurrenceofACEsrevealednosignificantrelation- shipswithheartrateresponsestothestresstask(datanot shown).

4. Discussion

Wesetouttoassesstheassociation ofACEsandpsychobi- ologicalstressreactivityanditsmodulationbythenumber, duration and age of occurrence of ACEsin healthy young women. Our results are in line with previous reports of

Fig. 4 Heart rate responses to the MIST are depicted for women who experienced0, 1, 2—3, or 4ormore ACEs.Val- uesrepresentaverageheartrate(beatsperminute)±standard error of the mean for the ACE total groups controlled for oral contraceptive use and BMI. Higher ACE sum scores are associatedwithattenuatedheartratereactivity.(^aTimepoints relativetotheMIST:HeartrateaveragedforBaseline(−5min), Stress1(firstrun),Stress2(secondrun),Stress3(thirdrun), andRecoveryperiod(first5minafterlastrun).).

attenuatedendocrine(Carpenteretal.,2007,2011;Elzinga et al.,2008; Lovallo etal.,2012)aswell ascardiovascu- lar(Lovalloetal.,2012)stressresponsestoapsychosocial stresstestinhealthyadultswithahistoryofadversechild- hoodexperiences.Furthermore,ourresultssubstantiatethe importanceofthemeannumberofACEsonendocrineand cardiovascularresponsetopsychosocialstress.Importantly, blunted cortisol and heart rate responses were indepen- dentofemotionalresponses,suggestingthatthediminished endocrine and cardiovascular stress reactivity cannot be explainedbyareducedemotionalreactiontostress(which may beinterpreted asflattened affect) aftera history of childhoodadversity.

To thebestofourknowledge, thepresent studyis the firsttodemonstratethatinhealthyyoungwomen,especially longenduring,chronicACEsshowthestrongestassociation with a blunted cortisol reactivity, adding valuable knowl- edge tothelink betweenchronic childhoodadversityand alterationsoftheHPAaxis.Eventhoughthesumofevents whichoccurredbeforemenarcheshowedanassociationwith abluntedcortisolreactivity,whereaseventsaftermenarche didnot,inthissample, thespecificageofoccurrencedid notcontributetoafurtherunderstandingoftheassociation between timingof ACEsandendocrineandcardiovascular reactivity.

Per se,our results showa deviation froman expected endocrine and cardiovascular stress response in partici- pants free of mental and physical illness in association with a history of ACEs. According to Obradovic (2012), taking together recent findings onstress reactivity in the context of early adversity, it is more accurate to state thatexposuretoearlylifestressmayleadtodysregulated physiological phenotypes rather than toa particular pat- tern of hyper-or hyporesponsivity (Obradovic,2012).The recentlyproposedadaptivecalibrationmodel(DelGiudice etal.,2011)offersanevolutionary-developmentaltheoryof

individualdifferencesinphysiologicalreactivityprocesses.

The authors hypothesize that, at a very general level, a nonlinearrelationbetweenadverselifeeventexposureand stress response exists. However, in the context of high adversity, the model predicts an either vigilant profile, characterized by high biological stress responsiveness, or anunemotional,underresponsiveprofile,characterizedby generallylowHPAaxisandSNSactivity.Thus,theseoppo- site phenotypes might be mediated by other factors and theirinteractions,e.g.thestudysample,thetypeofmal- treatmentortheinteraction ofvariousenvironmentaland geneticfactors.

Intermsofenvironmentalfactors,studieshavedemon- strated that the HPA axis in early human development is understrong social regulation(Tarullo andGunnar, 2006).

Therefore,severalstudiessuggestedparentalcaregivingas a moderator of the HPA reactivity (Gunnar et al., 1992;

Nachmias et al., 1996; Tarullo and Gunnar, 2006). Thus, sensitive parentingappearstobuffercortisol responsesin fearfulsituations,whereasbeingdeprivedofanevolution- arilyexpectablelevelofcare(e.g.institutionalrearing)has been associatedwith blunted cortisolproduction (Carlson and Earls, 1997; Gunnar et al., 2001). However, studies withinstitutionalizedchildrenwerealsoabletoshowthat improved caregiving environments had an effect on nor- malizingdampenedHPAaxisdiurnalrhythms(Dozieretal., 2008;Cicchettietal.,2011;Fisheretal.,2011).

The social buffering of the HPA axis is supported by findingsinanimalmodels.Intheirextensivesummaryoflit- erature,Hostinar etal.(2014)reportfindingsfromanimal studies,whichsuggestthatneuralmechanismsarerespon- sible for behavioral and neuroendocrine changes due to socialbuffering(Hostinaretal.,2014).Findingsfromanimal studiesreportchangesinthedevelopmentoftheHPAaxis anditsfunctioninsocial buffering.Inearly development, astress hyporesponsiveperiod(e.g. Witek-Janusek,1988) duringwhichthemotherstronglycontrolstheinfantscorti- costeronelevelshasbeenreported,whichcanbedisrupted by maternal deprivation and result in a hyperresponsive HPA pattern. In later stages of development, peers can alsobecomeasourceofsocial buffering(Hennessyetal., 2009).Forexample,socialinputhasbeenshowntodampen HPAaxis reactivity in ratpups ifthe motherwas present (e.g.Stanton andLevine,1990; Shionoyaetal., 2007),in maturingrodentswhencohabitatingcompanionsarepresent (Terranovaetal.,1999),orinsquirrelmonkeysinthepres- ence of familiar and unfamiliar conspecifics (Vogt et al., 1981;Hennessy,1984).

The findings in human andanimal studies provide sup- port for an adaptive response of the stress system to its environment,probably in order toenhance survival odds.

It could be speculated that after an initial hypersecre- tionofcortisoldue tochronic stressful environments,the HPA axis could counter-regulate its response and cortisol output might rebound to below normal. A plausible bio- logical explanation could be an increased glucocorticoid negative feedback with a downregulation of CRF recep- tors, or a diminished release of cortisol by the adrenal glands (Heimet al., 2000a; Frieset al.,2005).This view is supported by our finding that only chronic events, not acute events, were associated with a blunted cortisol response.

Recentstudieslinkgeneticandepigeneticalterationsto stressreactivityinassociationwith ACEsaswell.Ahistory ofACEshasbeenassociatedwithanepigeneticregulationof theglucocorticoidreceptorinthehippocampus(McGowan etal.,2009).Moreover,arecentstudyonhealthyadultswho experiencedthelossofaparentduringchildhood,maltreat- ment,orlowparentalcareshowedepigeneticalterationsof aregionofthehumanglucocorticoidreceptorgene,which in turn was associated with a blunted cortisol reactivity afteraneuroendocrinechallengetestintheseparticipants (Tyrkaetal.,2012).Anotherstudylinkedprenatalmaternal depression toincreased methylationof theglucocorticoid receptor gene, and showed exaggerated salivary cortisol outputtostressat3monthsofage(Oberlanderetal.,2008).

Furtherstudiesareneededtoinvestigatethesechangesin central regulation of the glucocorticoid receptor in brain regionsinvolvedinstressresponsesinassociationtoACEs.

Our results raise the challenging question of whether theobservedalterationsinstressresponsivitycanbeinter- pretedasapotentialriskfactorfororasasignofresilience to the development of later mental and physical disor- ders. Even if initiallyadaptive, bluntedcortisol reactivity could compromise future and necessary psychobiological stress reactivity. For example, low cardiovascular and/or endocrinereactivitytoacutepsychologicalstresshasbeen associatedwithdepression,fibromyalgia,obesity,burnout, substanceusedisorders,andchronicpainsyndromes(Griep et al., 1998; Heim et al., 1998; Pruessner et al., 1999;

Lovalloetal.,2000; GoldandChrousos,2002; Guretal., 2004;Phillipsetal.,2011;Jonesetal.,2012).

Consideringthatparticipantsinthepresent studywere recruitedfromaschoolofhighereducationandwerefree of psychopathology in adulthood, they may have been selectedinawaythatthebluntedstressreactivitypattern may stand for resilience to the development of mental illnessin theaftermathof childhoodadverse experiences.

Longitudinal,population-basedresearchisneededtoinves- tigateifparticipantswithoutpresentpsychopathology,who hadshownblunted stressresponsesafterstressinduction, areat higherrisktodevelop psychiatricdisorderslaterin lifeorremainhealthy.

The present study revealed diminished heart rate responses in association with the total number of ACEs, but notwith the subgroupsregardingdurationand ageof occurrence. Recently, Bauer & Boyce have suggested the examinationoftheHPAaxisandtheSNSsimultaneouslyfor a better understanding of their coordination (additive or interactive;or opposingor complementary) (Baueretal., 2002).Onlyfewstudieshaveexaminedtheexactnatureof theircoordinationinadultsamplessofar(AliandPruessner, 2012; Lovalloetal., 2012; AndrewsandPruessner, 2013), andresultsaremixed.Methodologicaldifferencesbetween thereportedstudiescouldaccountforthedifferentfindings.

Moreempiricalresearchisneededtoinvestigatetheexact natureofSNSalterationsafterACEs,aswellasthecoordi- nationandinteractionsbetweenthetwostresssystems.

Ourresultsrevealed noassociationsofthespecificage ofoccurrence,whichcontraststofindingsofotherstudies.

Bosch etal. (2012) reportedthat especially ACEsin pre- and postnatal developmentalstages were associatedwith heightenedcortisolreactivity(Boschetal.,2012).However, their sample included 16-year old adolescents, and also

included concurrent psychopathology. In contrast to the presentstudy,whichusedageoffirstACEoccurrence,and the distinction pre- and postmenarche, the authors used more distinct age groups to examine the association of age and cortisol reactivity. A limitation of our approach, by using the specific age of occurrence, is that type of traumaisneglected.Typeoftraumaisanimportantfactor, which could also explain different HPA axis phenotypes.

Future studies are needed to further analyze timing of ACEs,andmethodologically,thismightbeassessedbestina sampleofparticipantswithonespecificseveretrauma/life event. Furthermore, a limitation of the present study is that no data on pre- and postnatal stress exposure was included. Prenatalmaternal stress(Entringeretal., 2009) and postnatal adversity (Bosch et al., 2012) have been associated with increased HPA responses to stress, which couldresultinasensitizationtostressorsinlong-term.

Another limitation of the present study is that only peripheralreadoutsofstresshormoneactivationweremea- sured. The HPA axis and the cardiovascular system are complexandmultilayersystemsandthereforewe arenot able to identifythe exact mechanismsor locationof the observed dysregulations. Furthermore, that only women wererecruitedandtestedhastobementionedasanother potentially limitating factor. Especially as some of the recentmodelsonstressreactivitychangesafterchronicor traumaticstressmakementionofsexdifferences,itwould havebeeninformativetoassessthestressresponseinmen as well (Bangasser, 2013). Thus, due to the characteris- tics of the study sample, thepresented findings canonly begeneralizedtoyoungwomenfreeof mentalandphysi- calillness.Also,participantswereattendeesofschoolsfor health careprofessionsandsocialwork,which couldhave ledtoaselectionbiasasoutlinedabove.Also,eventhough HardtandRutter(2004)concludedintheirreviewofstudies from1980to2001thatthevalidityofretrospectiverecallof sexual/physicalabuse,physical/emotionalneglectorfamily discord issufficientlyvalid(Hardtand Rutter,2004).Still, the retrospectively assessed adversities could have been underestimated and/orbiased,especiallyreportedevents intoddlerandpre-verbalages.

Because of methodological reasons, the specification between ACEs shorter/longer than one year cannot be assumed to be completely independent from each other.

Sincesome participants reportedboth, acuteandchronic ACEs,withoursamplesize,itwasnotpossibletohavefully distinctandstatisticallyorthogonalgroups.

Despitetheselimitations,thepresentstudystrengthens theassumptionthatadversechildhoodexperiencesgiverise toa bluntedstress reactivityof theHPAaxisand theSNS in young healthy women. In this study population, num- beranddurationofadverseeventsinchildhoodshowedthe strongestassociationwithanattenuatedstressresponsein adulthood.Thesefindingssuggestthatthereactivityofthe humanstresssystemisindeedshapedbytheexperienceof extrinsicchronicstressorsinchildhoodandadolescence.

Contributors

KlausBader(PI)andFrankH.Wilhelmdesignedthestudy andwrotetheprotocol.

JensC.PruessnerdevelopedtheMontrealImagingStress TaskandtogetherwithRobertoLaMarcatrainedusinhow touseitandsupervisedusinadministratingthestresspro- tocoloutsidetheMRT.

Evelin Hugwrotethe firstdraftof thepaper,Annette Voellminwrotethefinaldraftofthemanuscriptandunder- took the statistical analysis. Annette Voellmin managed the recruitment of the participants, together with Katja Winzeler,whowasinvolvedintheliteratureresearchand theproofreadingofthemanuscript.

ValérieSchaeferhelpedintherecruitmentofthepar- ticipantsandwasalsopartofthestudydesignteam.

JensGaabassistedAnnetteVoellmininthemanuscript preparationandinthestatisticalanalyses.

Role of the funding source

ThisstudywassupportedbySwissNationalScienceFounda- tion(SNSF)grant#100014 126635/1(KB,FHW).

Conflict of interest statement

Theauthorshavenobiomedicalfinancialinterestsorpoten- tialconflictsofinteresttoreport.

Acknowledgements

We thank the participants of this study, as well as the graduate students from the University of Basel for their assistance.WealsothankDr.AndreaMeyerforhisstatistical assistanceandDr.ChristianCajochenforhiscontributionsto thestudydesign.

References

Ali,N.,Pruessner,J.C.,2012.Thesalivaryalphaamylaseovercorti- solratioasamarkertoassessdysregulationofthestresssystem.

Physiol.Behav.106,65—72.

Andersen,S.L.,Tomada,A.,Vincow,E.S.,Valente,E.,Polcari,A., Teicher,M.H.,2008.Preliminaryevidenceforsensitiveperiods intheeffectofchildhoodsexualabuseonregionalbraindevel- opment.J.NeuropsychiatryClin.Neurosci.20,292—301.

Andrews, J., Pruessner, J., 2013. The combined propanol/TSST paradigm—–a newmethodforpsychoneuroendocrinology.PLoS One8(2),e57567.

Bangasser,D.A.,2013.Sexdifferencesinstress-relatedreceptors:

‘‘micro’’differenceswith‘‘macro’’implicationsformoodand anxietydisorders.Biol.SexDiffer.4,2.

Bauer,A.M.,Quas,J.A.,Boyce,W.T.,2002.Associationsbetween physiologicalreactivityandchildren’sbehavior:advantagesofa multisystemapproach.J.Dev.Behav.Pediatr.23,102—113.

Bosch, N.M., Riese,H., Reijneveld,S.A., Bakker,M.P., Verhulst, F.C., Ormel, J., Oldehinkel,A.J.,2012. Timingmatters:long termeffectsofadversitiesfromprenatalperioduptoadoles- cenceonadolescents’cortisolstressresponse.TheTRAILSstudy.

Psychoneuroendocrinology37,1439—1447.

Bremner,J.D.,Bolus,R.,Mayer,E.A.,2007.Psychometricproperties oftheearlytraumainventory-selfreport.J.Nerv.Ment.Dis.195, 211—218.

Bremner, J.D., Vythilingam, M., Vermetten, E., Adil, J., Khan, S., Nazeer, A., Afzal, N., McGlashan, T.,Elzinga, B.,Ander- son,G.M.,Heninger,G.,Southwick,S.M.,Charney,D.S.,2003.

Cortisolresponsetoacognitivestresschallengeinposttraumatic stressdisorder(PTSD) relatedtochildhoodabuse.Psychoneu- roendocrinology28,733—750.

Carlson, M., Earls, F., 1997. Psychological and neuroendocrino- logical sequelae of early social deprivation in institution- alized children in Romania. Ann. N.Y. Acad. Sci. 807, 419—428.

Carpenter,L.L.,Carvalho,J.P.,Tyrka,A.R.,Wier,L.M.,Mello,A.F., Mello,M.F.,Anderson,G.M.,Wilkinson,C.W.,Price,L.H.,2007.

Decreasedadrenocorticotropichormoneandcortisolresponses tostressinhealthyadultsreportingsignificantchildhoodmal- treatment.Biol.Psychiatry62,1080—1087.

Carpenter,L.L.,Shattuck,T.T.,Tyrka,A.R.,Geracioti,T.D.,Price, L.H.,2011.Effectofchildhoodphysicalabuseoncortisolstress response.Psychopharmacology(Berl)214,367—375.

Chrousos,G.P.,2009.Stressanddisordersofthestresssystem.Nat.

Rev.Endocrinol.5,374—381.

Cicchetti, D., Rogosch, F.A., Toth, S.L., Sturge-Apple, M.L., 2011. Normalizing the development of cortisol regulation in maltreatedinfantsthroughpreventiveinterventions.Dev.Psy- chopathol.23,789—800.

DeBellis,M.D.,2002.Developmentaltraumatology:acontributory mechanismforalcoholandsubstanceusedisorders.Psychoneu- roendocrinology27,155—170.

Dedovic,K.,Renwick,R.,Mahani,N.K.,Engert, V.,Lupien,S.J., Pruessner,J.C.,2005.TheMontrealimagingstresstask:using functionalimagingtoinvestigatetheeffectsofperceivingand processingpsychosocialstressinthehumanbrain.J.Psychiatry Neurosci.30,319—325.

DelGiudice,M.,Ellis,B.J.,Shirtcliff,E.A.,2011.Theadaptivecal- ibrationmodelofstress responsivity.Neurosci. Biobehav.Rev.

35,1562—1592.

Dozier,M.,Peloso,E.,Lewis,E.,Laurenceau,J.P.,Levine,S.,2008.

Effectsofanattachment-basedinterventiononthecortisolpro- ductionofinfantsandtoddlersinfostercare.Dev.Psychopathol.

20,845—859.

Elzinga, B.M., Roelofs, K., Tollenaar, M.S., Bakvis, P., van Pelt, J.,Spinhoven,P., 2008.Diminishedcortisolresponsestopsy- chosocialstressassociatedwithlifetimeadverseeventsastudy amonghealthyyoung subjects.Psychoneuroendocrinology33, 227—237.

Entringer,S.,Kumsta,R.,Hellhammer,D.H.,Wadhwa,P.D.,Wust, S., 2009. Prenatal exposure to maternal psychosocial stress and HPA axis regulation in young adults. Horm. Behav. 55, 292—298.

Fisher,P.A.,VanRyzin,M.J.,Gunnar,M.R.,2011.MitigatingHPAaxis dysregulationassociatedwithplacementchangesinfostercare.

Psychoneuroendocrinology36,531—539.

Fries,E., Hesse,J., Hellhammer,J., Hellhammer, D.H.,2005. A new view on hypocortisolism. Psychoneuroendocrinology 30, 1010—1016.

Fumagalli,F.,Molteni,R.,Racagni,G.,Riva,M.A.,2007.Stressdur- ingdevelopment: impactonneuroplasticity andrelevance to psychopathology.Prog.Neurobiol.81,197—217.

Gold,P.W.,Chrousos,G.P.,2002.Organizationofthestresssystem and itsdysregulationin melancholic and atypicaldepression:

highvslowCRH/NEstates.Mol.Psychiatry7,254—275.

Green,S.B.,Salkind,N.J.,Akey,T.M.,2000.UsingSPSSforWindows, seconded.PrenticeHall,EnglewoodCliffs,NJ.

Griep, E.N., Boersma, J.W., Lentjes, E.G., Prins, A.P., van der Korst, J.K., de Kloet, E.R., 1998. Function of the hypothalamic—pituitary—adrenal axis in patients with fibromyalgia and low back pain. J. Rheumatol. 25, 1374—1381.

Gunnar,M.R.,Larson,M.C.,Hertsgaard,L.,Harris,M.L.,Brodersen, L.,1992.Thestressfulnessofseparationamongnine-month-old infants:effectsofsocialcontextvariablesandinfanttempera- ment.ChildDev.63,290—303.

Gunnar, M.R., Morison, S.J., Chisholm, K., Schuder, M., 2001.

Salivary cortisol levels in children adopted from Romanian orphanages.Dev.Psychopathol.13,611—628.

Gur,A.,Cevik,R.,Nas,K.,Colpan,L.,Sarac,S.,2004.Cortisoland hypothalamic—pituitary—gonadal axis hormones in follicular- phasewomenwithfibromyalgiaandchronicfatiguesyndrome andeffectofdepressivesymptomsonthesehormones.Arthritis Res.Ther.6,R232—R238.

Hardt,J.,Rutter,M.,2004.Validityofadultretrospectivereportsof adversechildhoodexperiences:reviewoftheevidence.J.Child Psychol.Psychiatry45,260—273.

Hautzinger, M., Bailer, M., 1993. Allgemeine Depressions-Skala (ADS):Manual.BeltzTest,Weinheim.

Heim,C.,Ehlert,U.,Hanker,J.P.,Hellhammer,D.H.,1998.Abuse- related posttraumatic stress disorder and alterations of the hypothalamic—pituitary—adrenal axis in women with chronic pelvicpain.Psychosom.Med.60,309—318.

Heim,C.,Ehlert,U.,Hellhammer,D.H.,2000a.Thepotentialroleof hypocortisolisminthepathophysiologyofstress-relatedbodily disorders.Psychoneuroendocrinology25,1—35.

Heim,C.,Nemeroff,C.B.,2001.Theroleofchildhoodtraumain theneurobiologyofmoodandanxietydisorders:preclinicaland clinicalstudies.Biol.Psychiatry49,1023—1039.

Heim,C.,Newport,D.J.,Heit,S.,Graham,Y.P.,Wilcox,M.,Bon- sall,R.,Miller,A.H.,Nemeroff,C.B.,2000b.Pituitary-adrenal andautonomicresponsestostressinwomenaftersexualand physicalabuseinchildhood.JAMA284,592—597.

Heim, C., Newport, D.J., Mletzko, T., Miller, A.H., Nemeroff, C.B.,2008.Thelinkbetweenchildhoodtraumaanddepression:

insights from HPA axis studies in humans. Psychoneuroen- docrinology33,693—710.

Hennessy,M.B.,1984.Presenceofcompanionmoderatesarousalof monkeyswithrestrictedsocialexperience.Physiol.Behav.33, 693—698.

Hennessy, M.B., Kaiser,S.,Sachser,N., 2009. Socialbuffering of thestressresponse:diversity,mechanisms,andfunctions.Front.

Neuroendocrinol.30,470—482.

Hostinar, C.E., Sullivan, R.M., Gunnar, M.R., 2014. Psychobio- logical mechanisms underlying the social buffering of the hypothalamic—pituitary—adrenocorticalaxis:areviewofanimal modelsandhumanstudiesacrossdevelopment.Psychol.Bull.

140,256—282.

Jones, A., McMillan, M.R., Jones, R.W., Kowalik, G.T., Steeden, J.A.,Deanfield,J.E.,Pruessner,J.C.,Taylor,A.M.,Muthurangu, V., 2012. Adiposityis associated withblunted cardiovascular, neuroendocrineandcognitiveresponsestoacutementalstress.

PLoSOne7,e39143.

Kirschbaum, C., Kudielka, B.M., Gaab, J., Schommer, N.C., Hellhammer, D.H., 1999. Impact of gender, menstrual cycle phase, and oral contraceptives on the activity of the hypothalamus—pituitary—adrenal axis. Psychosom. Med. 61, 154—162.

Kirschbaum,C.,Pirke, K.M.,Hellhammer,D.H.,1993. The‘Trier Social Stress Test’—–a tool for investigating psychobiological stressresponsesinalaboratorysetting.Neuropsychobiology28, 76—81.

La Marca, R.,Waldvogel, P., Thorn, H.,Tripod, M., Wirtz,P.H., Pruessner,J.C.,Ehlert,U.,2011.Associationbetweencoldface test-induced vagal inhibition and cortisol response to acute stress.Psychophysiology48,420—429.

Lovallo,W.R.,Dickensheets,S.L.,Myers,D.A.,Thomas,T.L.,Nixon, S.J.,2000.Blunted stresscortisolresponse inabstinentalco- holicandpolysubstance-abusingmen.Alcohol.Clin.Exp.Res.

24,651—658.

Lovallo, W.R., Farag, N.H., Sorocco, K.H., Cohoon, A.J., Vin- cent,A.S.,2012.Lifetimeadversityleadstobluntedstressaxis reactivity:studiesfromtheOklahomaFamily HealthPatterns Project.Biol.Psychiatry71,344—349.