Cortisol awakening response and diurnal cortisol levels among children at elevated risk for schizophrenia: relationship to psychosocial stress and neurocognitive function

Cortisol

awakening

response

and

diurnal

cortisol

among

children

at

elevated

risk

for

schizophrenia:

Relationship

to

psychosocial

stress

and

cognition

Alexis

E.

Cullen

*

,

Patricia

A.

Zunszain

,

Hannah

Dickson

,

Ruth

E.

Roberts

,

Helen

L.

Fisher

,

Carmine

M.

Pariante

,

Kristin

R.

Laurens

a_{DepartmentofForensicandNeurodevelopmentalSciences(BoxP023),InstituteofPsychiatry,DeCrespigny}

Park,LondonSE58AF,UnitedKingdom

b

SectionofStress,PsychiatryandImmunology,DepartmentofPsychologicalMedicine,InstituteofPsychiatry,

King’sCollegeLondon,125ColdharbourLane,LondonSE59NU,UnitedKingdom

c

MRC Social, Genetic& Developmental Psychiatry Centre, Institute ofPsychiatry, King’s College London,

London,UnitedKingdom

d

ResearchUnit for Schizophrenia Epidemiology,School of Psychiatry,Universityof NewSouth Wales,Sydney,

Australia

e

SchizophreniaResearchInstitute,Sydney,Australia

Received14January2014;receivedinrevisedform19March2014;accepted20March2014

Psychoneuroendocrinology(2014)46,1—13

KEYWORDS

Psychosis; High-risk; HPAaxis;

Negativelifeevents; Hassles;

Memory;

Executivefunction; Development

Summary Abnormalhypothalamic-pituitary-adrenal (HPA)axisfunction, asindexedby

ele-vated diurnalcortisol levels and/or ablunted cortisolawakening response (CAR), has been

observedamongpatientswithfirstepisodepsychosisandassociatedwithneurocognitivedeficits

inthispopulation.However,theextenttowhichthesefeaturesprecedeillnessonsetisunclear.

Thecurrentstudyaimedtodeterminewhetherchildrenwhoareatputativelyelevatedriskfor

psychosis because they present multiple antecedents of schizophrenia (ASz), and high-risk

childrenwith afamily history ofillness (FHx),are characterizedbyabnormal cortisollevels

whencomparedwiththeirtypicallydeveloping(TD)peers.Afurtheraimwastoinvestigatethe

extent to which cortisol levels are associated with psychosocial stress and neurocognitive

function. Thirty-three ASz children, 22 FHx children, and 40 TD children were identified

at age 9—12 years using a novel community-based screening procedure or as relatives of

individuals with schizophrenia. All participants were antipsychotic-naive and not currently

* Correspondingauthorat:DepartmentofForensicandNeurodevelopmentalSciences(BoxP023),InstituteofPsychiatry,DeCrespignyPark, LondonSE58AF,UnitedKingdom.Tel.:+442078480654;fax:+442078480754.

E-mailaddresses:alexis.cullen@kcl.ac.uk(A.E.Cullen),kristin.laurens@kcl.ac.uk(K.R.Laurens). Availableonlineatwww.sciencedirect.com

ScienceDirect

jo u rn al ho m e p age : w w w. el s ev i er. co m / lo c at e /p s y n eu en

http://dx.doi.org/10.1016/j.psyneuen.2014.03.010

1.

Introduction

Theneuraldiathesis-stressmodelof schizophrenia(Walker andDiforio,1997;Walkeretal.,2008)positsthat psychoso-cial stress mayact via the hypothalamic-pituitary-adrenal (HPA)axis(theprimarysysteminvolvedincoordinatingthe physiologicalresponse to stress)to trigger psychosisonset among individualswith an underlyingvulnerability for the disorder.AbnormalHPAaxisfunction,asindexedbyelevated diurnalcortisollevels and/orablunted cortisolawakening response(CAR),hasbeenobservedinindividualswhohave recently experienced their first psychotic episode (Borges etal., 2013). Additionally, elevated cortisol levels among first episode patients have been associated with reduced hippocampalvolume (Mondelli etal., 2010b), andgreater bluntingoftheCARwithmorepronounceddeficitsinverbal memoryandprocessingspeed(Aasetal.,2011).However, theextenttowhichabnormalHPAaxisfunctionprecedesthe onsetofpsychosisiscurrentlyunclear.Whilstthesestudies suggestthatthefirst psychoticepisodeischaracterizedby HPA axisdysfunctionandthatthismaybeassociated with some of the neuroanatomical abnormalities and cognitive deficits observed in psychosis patients, it is possible that theseHPAaxisabnormalitiesmaysimplybeaconsequenceof thestressassociatedwithillnessonset.

Thestudyofindividualsatelevatedriskforschizophrenia providestheopportunitytodirectlytesttheassumptionsof theneuraldiathesis-stressmodel(WalkerandDiforio,1997; Walkeretal.,2008)andthusdeterminewhether vulnerabil-ityforschizophreniaisassociatedwithHPAaxisdysfunction. Traditionally, high-risk approacheshave focused on indivi-dualswithafamilyhistoryofillness.Studiescomparingyoung adultrelativesofpatientswithpsychosistohealthycontrols haveyieldedinconsistentfindings,perhapsdueto methodo-logical differences. A study examining salivary cortisol at random time-pointsthroughout theday (values averaged) reported elevated cortisol among relatives (Collip et al., 2011),whilsttwostudiesmeasuringfastingplasmacortisol ata single time-point didnot(Brunelinetal.,2008; Yang et al., 2012). A more recent strategy focuses on youth consideredatultrahigh-risk(UHR)forpsychosiswhose clin-ical features (typically, attenuated psychotic symptoms) indicatethattheymaybeintheprodromalphaseofillness. Higher salivary cortisol levels (obtained at a single time-pointinthemorning)havebeenreportedamongUHRyouth

relativeto healthycontrols,withmorepronouncedeffects among those who weremedication-free (Sugranyes etal., 2012).Elevatedcortisollevels inUHRyouthweresimilarly observedinastudywhichexaminedsalivarycortisolobtained at three time-pointsthroughoutthemorning (values aver-aged)(Walkeretal.,2013).Usingasimilarsamplingprotocol, work by Walkerandcolleagues also indicatesthat adoles-centswithschizotypalpersonalitydisorder(SPD),whoareat greater risk for psychosis on account of their diagnosis, exhibit higher daytime salivary cortisol levels (Walker et al., 2001; Mittal et al., 2007). Thus, elevated salivary cortisolduringthedayhasbeenconsistentlyobservedamong youth whose clinical features designate them as being at elevatedriskforpsychosis.However,noneofthesestudies obtained the multiple samples distributed throughout the daythatarerequiredfortheexaminationofdiurnalpatterns ofcortisolsecretion(whichmaybeamorereliablemarkerof HPAaxisfunctiongiventhatcortisollevelsvarysubstantially throughtheday).Furthermore,asnostudyofhigh-riskyouth hasyetexaminedtheCAR,itremainsunclearwhetherthe bluntedCARobservedamong individualswith first episode psychosisalsocharacterizestheseyouth.

WhilststudiesofUHRyouthsuggestthatHPAaxis dysfunc-tion precedes the onset of psychosis, these youth are, by definition,alreadysufficientlydistressedastoseektreatment fortheirsymptoms.Thus,elevationsincortisolmightfeasibly beduetodistressrelatingtoemergingillness,asopposedto externalpsychosocialstressors.Indeed,onlyonestudyofUHR youth has examined the association between psychosocial stressexposure andHPA axis dysfunction(Thompson et al., 2007).Inthisstudy,higherplasmacortisollevelsobtainedata single time-point in the morning were correlated with the numberofminorhasslesbutnotwiththenumberofmajorlife eventsexperiencedduringthepastmonth.Studiesofyouth withSPD,althoughnottypicallyconfoundedbyhelp-seeking behaviour,arelimitedbythefactthattheseindividualsalready present with symptoms of sufficient severity as to meet diagnostic criteria for a schizophrenia spectrum disorder.1

seekingtreatmentfortheirsymptoms.Atage11—14years,participantsprovidedsalivarycortisol

samplesandcompletedpsychosocialstressmeasuresandtestsofmemoryandexecutivefunction.

ResultsindicatedthatFHxchildren,butnotASzchildren,werecharacterizedby ablunted CAR

relativetotheirTDpeers(effectsize= 0.73,p=0.01)thatwasnotexplainedbypsychosocialstress

exposureorbydistressrelatingtotheseexperiences.NeitherFHxnorASzchildrenwere

character-izedbyelevated diurnalcortisol.AmongbothFHx andASz children,morepronouncedHPAaxis

functionabnormalities(i.e.,higherdiurnalcortisollevelsand greaterbluntingoftheCAR)were

associatedwith poorer performance ontests of verbal memoryand executivefunction. These

findingssupportthenotionthatatleastsomeHPAaxisabnormalitiesdescribedinpsychosisprecede

illnessonset,ratherthanbeingasubsequentepiphenomenon.WespeculatethatthebluntedCAR

mayconstituteanearly(potentiallygeneticallymediated)markerofpsychosisvulnerability,whilst

elevateddiurnalcortisollevelsmayemergeonlyproximallytodiseaseonset.

#2014TheAuthors.PublishedbyElsevierLtd.ThisisanopenaccessarticleundertheCCBY-NC-ND

license(http://creativecommons.org/licenses/by-nc-nd/3.0/).

1_{Schizotypal(personality)disorderisclassifiedasaschizophrenia}

EstablishingwhetherabnormalHPAaxisfunctionispresentin younger samples of high-risk individuals, who are not yet sufficientlydistressedastoseekhelpfortheirsymptoms,will helptodeterminetheextent towhichHPAaxisdysfunction increasestheriskforpsychosis.Furthermore,giventhat neu-rocognitivedeficitshavebeenconsistentlyidentifiedamong individuals at elevated risk for psychosis(Fusar-Poli etal.,

2012; Agnew-Blais and Seidman, 2013), it is important to

establishwhetherthesedeficitsareassociatedwithHPAaxis dysfunction.

Wehavedevelopedanovelcommunityscreeningmethod for identifying children who may be at elevated risk for psychosisbecausetheypresentwithatriadofantecedents ofschizophrenia(ASz),definedas(i)aspeechand/ormotor delayorabnormality,(ii)asocial,emotional,and/or beha-viouralproblem,and(iii)apsychotic-likeexperience( Lau-rensetal., 2007).Wehave previouslydemonstrated that, relative to their typicallydeveloping (TD)peers, ASz chil-dren,aswellaschildrenwithafamilyhistoryof schizophre-nia (FHx), are more frequently exposed to psychosocial stressorsandmoredistressedbythese experiences(Cullen etal.,2014),andthatbothgroupsshowpoorer neurocogni-tivefunction(Cullenetal.,2010;Dicksonetal.,2014).The current study aimed to determine whether ASz and FHx childrenarealsocharacterizedbyabnormalHPAaxisfunction (elevateddiurnalcortisoland/orabluntedCAR)relativeto TDchildren.Afurtheraimwastoinvestigatewhetheramong ASzandFHx children,HPAaxis functionis associatedwith experiencesofpsychosocialstressandperformanceontasks ofmemoryandexecutivefunction.

2.

Methods

2.1. Recruitmentofchildren presenting antecedentsandtypicallydevelopingchildren

ASzandTDchildrenwererecruitedusingacommunity-based screening procedure(Laurens et al.,2007, 2011).In brief, questionnairesassessingatriadofwell-replicatedantecedents ofschizophreniawerecompletedindependentlybychildren aged9—12yearsatschoolandbytheircaregiversathome.The caregiverquestionnaireincludednineitems(three quantita-tiveandsixqualitative)assessingdelaysorabnormalitiesin speech and/or motor development (Laurens et al.,2007). Social, emotional, and behavioural problems were defined as ascore intheclinicalrange on atleastoneofthefour Strengths and Difficulties Questionnaire [SDQ (Goodman, 2001)]psychopathologyscales: emotionalsymptoms (child-reported), conduct problems, hyperactivity-inattention, or peer relationship problems (caregiver-reported). The child questionnaire also included a measure of psychotic-like experiences(PLEs)(Laurensetal.,2012),assessingnine delu-sion- and hallucination-like experiences, each rated on a three-point scale(0=not true, 1=somewhattrue, 2= cer-tainlytrue;ascoreof2onanyofthenineitemsindicateda positiverating).Childrenpresentingeachofthethree ante-cedentswereeligiblefortheASzgroup.TheTDgroup com-prisedchildrenwhopresentednoneoftheantecedentsand whohad nofirst-,second-,or third-degreerelativeswitha schizophreniaspectrumdisorder,asconfirmedsubsequently with the child’s caregiver via the Family Interview for

GeneticStudies[FIGS(Maxwell,1992)].Screening question-naireswerecompletedby1343childrenandtheircaregivers; ofthese,9.4%presentedtheASztriadand22.9%ofchildren presentednoneoftheantecedents.Ofthe150children eli-giblefortheASzorTDgroupwhoweresubsequentlyinvitedto participateinalongitudinalstudyofchilddevelopment,41%of ASz and42% of TDfamilies declined participation, respec-tively;theseratesareconsistentwithotherstudiesrecruiting childrenviaschoolsinthisregion(Mackieetal.,2011;Stateva etal.,2012).RelativetotheASzandTDchildrenwhodeclined participation,ASz andTDchildren who participated in the studydidnotdifferonage,gender,orethnicity.AmongASz participants,theprevalenceofthetriadcomponentsdidnot differsignificantly,withtheexceptionthatASzparticipants werelesslikelytoscoreinthe‘abnormal’rangeontheSDQ emotional symptoms scale than ASz children who did not participate(p=0.01).

2.2. Recruitmentofchildrenwithafamily historyofschizophrenia

Childrenwith afamilyhistoryofschizophreniaor schizoaf-fective disorder (FHx) were identified via the caregiver screening questionnaire, which included items assessing child and family mental health difficulties. Additionally, medicalrecords of mental healthservice users within the South London andMaudsley National Health Service (NHS) FoundationTrustwerereviewedtoidentifypatientswitha diagnosisof schizophrenia or schizoaffective disorder who hadachildrelativeagedbetween9and12years.Identified familieswereapproachedfollowingliaisonwiththepatient’s careworker.AllFHx children(identifiedeitherviamedical recordsorviacommunityscreening)hadatleastonefirst-or second-degreerelativewithschizophreniaorschizoaffective disorder,asconfirmedbytheFIGS.Ofthe1020childrenfor whomfamilyhistoryinformationonscreeningquestionnaires wasavailable,3.4%werereportedtohaveafamilyhistoryof schizophrenia.Afurther36childrenwhometcriteriaforthe FHxgroup(i.e.,hadafamilyhistoryofschizophrenia, fell withintheappropriateagerange,andforwhomtherewere viablecontactdetails)wereidentifiedviamedicalrecords.In total, 40% of FHx families declined to participate in the longitudinalstudyafterinitialcontact.

2.3. Currentsampleandprocedure

Caregiversandchildrenprovidedwritteninformedconsent andassent, respectively,for participation. Ethical permis-sionforthestudywasgrantedbytheJointSouthLondonand MaudsleyandtheInstituteofPsychiatryNHSResearchEthics Committee.

substanceuseviaquestionnaire(McVieandBradshaw,2005), withregulartobaccousedefinedashavingsmokedatleast onceaweekandregularcannabisusedefinedbyuseonat leastfouroccasions.Pubertalstatuswasassessedusingthe self-report Pubertal Developmental Scale (Carskadon and

Acebo,1993),withhigherscoresindicatingmoreadvanced

pubertaldevelopment.Heightandweightwerecollectedby theresearcherattheassessmentsessionandusedto com-puteBodyMassIndex(BMI;kg/m2).Consistentwithprevious studies conducted in Northern Europe (Vreeburg et al., 2009),samplingmonth wascategorized intolight(March— September)anddarkmonths(October—February).

Ofthe111childrenrecruitedviaquestionnairescreening and/orcaserecordreviewtothelongitudinalstudy,91(82%) providedsalivarycortisoldataatthefollow-upassessmentat age11—14years:29metASzcriteriaonly,18metFHxcriteria only,4metbothASzandFHxcriteria,and40childrenmetTD criteria.InordertomostaccuratelyreflecttheASzandFHx populationsfromwhichtheparticipantsweresampled,the fourASz+FHxcaseswereretainedinbothgroups;ASzand FHxgroupswerethereforeanalyzedrelativetotheTDgroup onlyandnotrelativetoeachother.FHxchildren,butnotASz children,weresignificantly less likely than TDchildren to providesalivary cortisol data at the follow-up assessment (p=0.05). Within-group analyses indicated that children whoprovidedsalivarycortisoldatadidnotdifferfromthose who didnot provide cortisol data on any of thefour SDQ psychopathologyscalesor on totalPLE scoresat screening (p>0.05).Samplecharacteristicsare providedinTable1; therewere no significantbetween-group differences (i.e., FHxvs.TD,orASzvs.TD)inage,pubertalstatus, lapseof time between identification and cortisol assessment, BMI, tobaccouse,awakeningtime,orsamplingmonth(p>0.05). Noneoftheparticipantsreportedregularcannabisuse.ASz childrenweresignificantlymorelikelytobemalecompared toTDchildren(p=0.02).RelativetotheTDgroup,boththe FHx and ASz group were found to differ on ethnicity (p0.05)andsocioeconomicstatus(p0.01).

2.4. Salivary cortisolassessment

Participants wereprovided with a verbal explanation and writteninstructionsforcollectingsalivasamplesathome(at theweekendorduringtheschoolholidays)usingthepassive drool procedure (http://www.salimetrics.com). Further details of our analysis protocol have been described pre-viously (Belvederi Murri et al., 2012). In brief, six saliva sampleswerecollectedthroughoutthedayontwo consecu-tive days: at awakening, andat 15, 30, and 60min after awakening, at 1200h, and 2000h. Participants were instructedtowakebefore1000handcollectthefirstsample immediatelyuponawakening,toavoidfoodconsumptionfor 30minpriortosamplecollection,andtorefrainfrom stren-uous exercise during theday. Samples were storedin the participant’shomefreezeruntilcollectionandsubsequently frozenat 208Catthelaboratory.Afterthawing and cen-trifugation at 3000rpm for 15min, cortisol levels were determined using the Salimetrics High Sensitivity Salivary CortisolELISAKIT(Salimetrics,Suffolk,UK),accordingtothe recommendedprocedure.Theanalyticalsensitivitywasset to 0.33nmol/l. Inter- and intra-assay coefficients ranged from 8% to 11% and 6% to 10%, respectively, which are

consistentwithvaluesreportedinotherstudiesofhigh-risk youth(Collipetal.,2011;Walkeretal.,2013).

CortisoldataweresummarizedusingtwoAreaUnderthe Curve(AUC) computations(Pruessner etal., 2003):(i)AUC withrespecttotheincreaseincortisollevelsfollowing awa-kening(AUCi-CAR:calculatedusingvaluesforawakening,and 15,30,and60minafterawakening),and(ii)AUCwithrespect togroundofcortisollevelsduringtheday(AUCg-DAY: calcu-latedusingvaluesforawakening,1200h,and2000h).Cortisol values were significantly correlated across testing days (p<0.001foreachtime-point);thus,cortisoldataobtained onthefirstsamplingdayonlywereusedforAUCcomputations, exceptwhenday1dataweremissingor whenparticipants demonstratedbettercompliancewiththesamplingprotocol onthesecondday(e.g.,ifparticipantsreportedwakingafter 1000honday1butnotday2).

2.5. Psychosocialstress

Participants completed two psychosocial stress measures within1monthofcollectingsalivarycortisolsamples(mean lapse of time between saliva collection and psychosocial stress measure completion was 3.2 days). As described previously (Cullen etal., 2014), an eight-item self-report measure assessedexposuretoarange ofchild-appropriate negativelifeevents(HeubeckandO’Sullivan,1998),suchas parental separation/divorce,death of someone close, and serious illness. This measure provided information on the totalnumber ofnegativelifeeventseverexperienced and theaveragelevelofdistressexperiencedinrelationtoeach event, both at thetime ofthe event andcurrently (each ratedonafour-pointscale,range:0‘notatall’to3‘alot’). Specifically,distressratingsweresummedanddividedbythe total number of life events experienced to provide two average distress scores (previous and current) reflecting theaveragedistressperevent.Participantsalsocompleted a37-itemquestionnaireadaptedfromHeubeckand O’Sulli-van(1998)toassessexposuretoschool-relateddailyhassles (itemsspannedfourdomains:scholastic,peer,teacher,and home). Participants rated on a four-point scale how fre-quently they hadexperienced each hassle duringthe past 6 months (range:0 ‘never’ to 3‘notat all’)andhow dis-tressedtheywerebythisexperience(range:0‘notatall’to3 ‘alot’).Frequencyratingsweresummedtoprovideatotal dailyhasslefrequencyscore(reflectingthenumberofhassles experiencedandhowoftentheyoccurred)anddistress rat-ings were summed and divided by the number of items endorsed to derive an overall daily hassle distress score (reflecting the average distress per item endorsed). The internal consistency of these two scales was examined in thetotalsampleandfoundtobehigh(Cronbach’sa=0.88 and0.92,respectively).

2.6. Neurocognitivemeasures

Table 1 Sample demographics, psychosocial stress, and neurocognitive performance indices for each participant group.

FHx (n= 22) ASz (n= 33) TD (n= 40) Statistical results

FHx vs. TD ASz vs. TD

Age (years); meanSE 13.30.3 12.80.2 13.10.2 t= 0.50 p= 0.62 t= 1.18 p= 0.24

Pubertal development scale scorea_{; meanSE 2.40.1 2.40.1 2.30.1 t= 0.19 p= 0.85 t= 0.07 p= 0.94}

Sex (male);n(%) 11 (50) 23 (70) 17 (43) x2= 0.32 p= 0.57 x2= 5.40 p= 0.02

Ethnicity;n(%) FE= 17.46 p<0.001 FE= 7.69 p= 0.05

White British 3 (14) 8 (24) 20 (50)

White other 2 (9) 8 (24) 11 (27)

Black Caribbean/African 8 (36) 4 (12) 3 (8)

Other 9 (41) 13 (40) 6 (15)

Socioeconomic status;n(%)a _{FE= 9.74 p= 0.006 FE= 13.44 p= 0.001}

Higher managerial, administrative, and professional 11 (50) 14 (43) 33 (82)

Intermediate 5 (23) 12 (36) 6 (15)

Routine and manual 6 (27) 7 (21) 1 (3)

Tobacco use;n(%) 1 (5) 2 (6) 0 (0) FE p= 0.36 FE p= 0.20

BMI (kg/m2_{); meanSE 19.70.7 20.40.6 19.80.5 t= 0.09 p= 0.93 t= 0.76 p= 0.45}

Waking time; meanSE 8:360:12 8:290:13 8:340:11 t= 0.11 p= 0.91 t= 0.27 p= 0.79

Lighter sampling month (March—September);n(%) 13 (59) 17 (52) 21 (53) x2= 0.01 p= 0.93 x2= 0.25 p= 0.62 Psychosocial stress measure; meanSE (total)

Total number of negative life events 1.80.4 1.80.2 1.10.2 t= 1.83 p= 0.08 t= 2.77 p= 0.007

Distress at the time of negative life event 1.50.3 1.40.2 1.30.2 U= 398.5 p= 0.64 U= 596.5 p= 0.59 Current distress related to negative life event 0.90.2 0.90.2 0.60.1 U= 365.5 p= 0.32 U= 536.0 p= 0.20

Frequency of daily hassles 35.82.9 43.32.2 31.01.8 t= 1.48 p= 0.14 t= 4.33 p<0.001

Distress related to daily hassles 1.10.1 1.30.1 0.90.1 t= 1.45 p= 0.15 t= 3.55 p= 0.001

Neurocognitive measures; meanSE

WRAML2 — Number-letter 12.10.8 11.30.6 13.10.4 t= 1.17 p= 0.21 t= 2.68 p= 0.009

WRAML2 — Verbal memory 21.10.9 22.10.7 23.80.7 t= 2.31 p= 0.02 t= 1.60 p= 0.12

WRAML2 — Visual memory 17.01.1 16.30.8 18.10.6 t= 0.91 p= 0.37 t= 1.79 p= 0.08

WRAML2 — Verbal working memory 9.70.5 9.50.4 10.70.3 t= 1.64 p= 0.11 t= 2.32 p= 0.02

D-KEFS — Letter fluency 10.80.7 10.30.4 11.00.5 t= 0.18 p= 0.86 t= 0.93 p= 0.35

D-KEFS — Category fluency 12.00.6 12.60.6 12.80.5 t= 0.91 p= 0.37 t= 0.23 p= 0.82

D-KEFS — Category switching 10.60.7 9.70.5 12.00.5 t= 1.52 p= 0.13 t= 3.00 p= 0.004

D-KEFS — Inhibition 10.20.5 10.90.4 11.80.3 t= 2.71 p= 0.009 t= 1.52 p= 0.13

D-KEFS — Inhibition/switching 10.50.5 10.70.4 11.10.4 t= 1.07 p= 0.29 t= 0.86 p= 0.39

D-KEFS — Tower test 11.40.3 11.20.3 11.60.3 t= 0.34 p= 0.74 t= 0.78 p= 0.44

Notes. FHx: family history of schizophrenia; ASz: antecedents of schizophrenia; TD: typically developing. Groups are not mutually exclusive (four children meeting both FHx and ASz criteria are retained in both risk groups). SE: standard error; BMI: body mass index; WRAML2: Wide Range Assessment of Memory and Learning 2nd Edition; D-KEFS: Delis—Kaplan Executive Function System; FE: Fisher’s exact test.

a_{Socioeconomic status based on the caregiver with the highest occupational status. Missing data: BMI (n= 2); psychosocial stress (n= 1).}

Cortisol,

stress,

and

cognition

in

children

at-risk

for

schizophrenia

MemoryandLearning2ndEdition(SheslowandAdams,2003) andtheDelis—KaplanExecutiveFunctionSystem(Delisetal., 2001),describedinSupplementaryTable1(meanlapseoftime betweensalivacollectionandneurocognitiveassessmentwas

1.4 months). Standardized scores were derived for each subtestusingmanual-reportedpopulationnorms.

Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/

10.1016/j.psyneuen.2014.03.010.

2.7. Statistical analyses

Independent samples t-tests, Mann—Whitney U tests, chi-squaredtests,andFisher’sexacttestswereusedtoexamine group differences on demographic variables. As AUCi-CAR andAUCg-DAYvalues wereapproximatelynormally distrib-uted,linearregressionanalyseswereconductedtoexamine theeffectofriskstatusonbothAUCmeasures.Standardized mean differences (d) were derived from unstandardized regression coefficients (Lipsey and Wilson, 2001), with; ‘small’,‘medium’,and‘large’effectsizesdenotedbyvalues of0.20,0.50,and0.80,respectively(Cohen,1992). Within-groupcorrelationanalyseswereconducted toexaminethe extent to which cortisol AUC values were associated with psychosocialstress andneurocognitiveperformance (Pear-son’s‘r’correlationswereused forallanalysesexceptfor those examining the negative life event distress variables which were not normally distributed and were therefore examined using Spearman’s rho ‘r’ correlations). Finally, independentsamplest-tests,one-wayANOVA’s,and correla-tionanalyses(Pearson’s‘r’andSpearman’srho‘r’for nor-mallyandnon-normallydistributedvariables, respectively) wereused to examine relationships betweendemographic variablesandcortisolAUCmeasures.Linearregression ana-lyses examiningthe effect of risk statuson AUCi-CAR and AUCg-DAYvalues wererepeatedafteradjustingfor factors foundto beassociated withAUC values. Allanalyseswere conducted using the Statistical Package for the Social Sciences(SPSS)version20.

3.

Results

3.1. Group differencesinthe CARanddiurnal cortisol

MeanAUCi-CARvaluesandcortisolvaluesfortheindividual timepoints(awakening,and15,30,and60minafter awa-kening)arepresentedbygroupinTable2andFig.1, respec-tively. Relative to the TD group, AUCi-CAR values were significantly lower in the FHx group (d= 0.73, p=0.01), reflecting ablunted CARamongFHx children. Incontrast, there was no significant effect of ASz statuson AUCi-CAR values (d= 0.19, p=0.42). This is evident by the visual inspectionofthedatainFig.1,whichshowsthatFHxchildren haveasmallerincreaseincortisollevelsbetweenawakening and15min,andsubsequently,asteeperdeclineincortisol levels between15 and30mincompared to the TD group, whilstASzchildrenshowedasimilarpatternof post-awaken-ingcortisolsecretionasTDchildren.PosthoctestswithinFHx subgroups showed that the magnitude of differences

between FHx and TD children was twice as large among T

FHxchildrenwithafirst-degreerelativewithschizophrenia (d= 1.09, p=0.005)comparedto those withan affected second-degree relative (d= 0.50, p=0.14). AUCg-DAY valuesarealsoshowninTable2.Therewerenosignificant group differences in cortisol levels during the day when eithertheFHxorASzgroupwerecomparedtotheTDgroup (p0.38). Post hoc tests performed on AUCg-DAY values indicated that themagnitude of differences betweenFHx andTDchildrenwassimilaramongFHxchildrenwitha first-degree relative with schizophrenia and those with an

affectedsecond-degreerelative(d=0.10and0.07, respec-tively;p0.05).

3.2. Associations betweenpsychosocialstress andcortisol

Consistentwithourpreviousreportinalarger,overlapping sample(Cullenetal.,2014),relativetotheTDgroup,ASzand FHxchildrenexperienced ahighernumber ofnegativelife eventsthan theTDgroup (p=0.007and p=0.08, respec-tively)and ASzchildrenalso reportedthat they more fre-quentlyexperienceddailyhasslesandweremoredistressed by these experiences (p0.001; Table 1). Within-group correlationanalyses wereconducted to examine the rela-tionshipbetweenpsychosocialstressandcortisol(Table3). AUCi-CARvalues werenotsignificantly associatedwiththe totalnumberofnegativelifeeventsexperiencedinanyofthe three groups (p0.10). However, among FHx children, AUCi-CAR values were positively correlated with distress experiencedinrelationtonegative lifeeventsatthetime oftheevent(r=0.51,p=0.02)andwiththelevelofdistress experienced currently (r=0.52, p=0.02); in contrast, amongTDchildren,AUCi-CARvalueswerenegatively corre-latedwithdistressexperiencedatthetimeofthenegative life event (r= 0.31, p=0.05). These variables werenot significantly correlated in ASz children. AUCi-CAR values werenotassociated withthefrequencyofdailyhassles or withdistressrelatingtotheseexperiences(p0.10).None of the psychosocial stress measures were correlated with AUCg-DAYvalues(p0.10).

0.00 2.00 4.00 6.00 8.00 10.00 12.00

Awakening 15 min 30 min 60 min

Corsol (nmol/l)

TD (n = 40) ASz (n =33) FHx (n = 22)

Figure 1 Mean cortisol levels ( standard error mean) at

awakening and15,30,and 60minpost-awakeningin children

withafamilyhistoryofschizophrenia(FHx),childrenpresenting

antecedents of schizophrenia (ASz), and typically developing

children(TD).

Table3 Correlationsofcortisolwithpsychosocialstressandneurocognitiveperformance.

Cortisolawakeningresponse(AUCi-CAR) Diurnalcortisol(AUCg-DAY)

FHx(n=22) ASz(n=33) TD(n=40) FHx(n=20) ASz(n=33) TD(n=40)

Psychosocialstress

Totalnumberofnegativelifeevents 0.21 0.03 0.03 0.04 0.09 0.27

Negativelifeeventdistress—previous 0.51* 0.16 0.31* 0.06 0.30 0.11

Negativelifeeventdistress—current 0.52* _{0.05 0.27 0.05 0.12 0.20}

Frequencyofdailyhassles 0.18 0.13 0.12 0.00 0.10 0.18

Distressrelatingtodailyhassles 0.17 0.16 0.00 0.16 0.26 0.20

Neurocognitiveperformance

WRAML2—Number-letter 0.29 0.19 0.22 0.62** _{0.01 0.16}

WRAML2—Verbalmemory 0.46* 0.04 0.03 0.47* 0.17 0.12

WRAML2—Visualmemory 0.05 0.05 0.24 0.31 0.22 0.02

WRAML2—Verbalworkingmemory 0.08 0.09 0.09 0.36 0.26 0.06

D-KEFS—Letterfluency 0.23 0.41* 0.14 0.43 0.30 0.27

D-KEFS—Categoryfluency 0.02 0.22 0.18 0.30 0.06 0.15

D-KEFS—Categoryswitching 0.22 0.14 0.06 0.05 0.16 0.11

D-KEFS—Inhibition 0.27 0.16 0.01 0.06 0.09 0.03

D-KEFS—Inhibition/switching 0.27 0.06 0.03 0.02 0.05 0.04

D-KEFS—Towertest 0.18 0.22 0.10 0.55* _{0.27 0.17}

Notes.FHx:familyhistoryofschizophrenia;ASz:antecedentsofschizophrenia;TD:typicallydeveloping.Groupsarenotmutuallyexclusive —fourASz+FHxcasesareincludedinbothgroups.AUCi-CAR:areaunderthecurvewithrespecttoincreaseforthecortisolawakening response;AUCg-DAY:areaunderthecurvewithrespecttogroundforcortisolduringtheday;WRAML2:WideRangeAssessmentofMemory andLearning2ndEdition;D-KEFS:Delis—KaplanExecutiveFunctionSystem.Pearson’s‘r’correlationswereusedforallanalysesexceptfor thoseexaminingthenegativelifeeventdistressvariables,whichwereexaminedusingSpearman’srho‘r’.

* _p0.05, ** _p0.01.

Figure2 Therelationshipbetweenverbalmemoryperformanceandthecortisolawakeningresponse(AUCi-CAR)anddiurnalcortisol

(AUCg-DAY)inchildrenwithafamilyhistoryofschizophrenia(FHx:PanelA),childrenpresentingantecedentsofschizophrenia(ASz:

3.3. Associationsbetweenneurocognitive functionandcortisol

In line with our previous work examining neurocognitive functioninASzandFHxchildrenatage9—12years(Cullen etal.,2010;Dicksonetal.,2014),bothASzandFHxgroups showedpoorerperformance onneurocognitivetestsatage 11—14yearscomparedtotheTDgroup(Table1).Specifically, ASzchildrenobtainedsignificantlylowerscoresonthe num-ber-letter (p=0.009), verbal working memory (p=0.02), andverbalfluency—categoryswitchingsubtests(p=0.004), whilst FHx children obtained lower scores on the verbal memoryindex (p=0.02)andthecolour-wordinterference —inhibitionsubtest(p=0.009).Correlationsbetween cor-tisol AUC measures and neurocognitive performance are presentedinTable3.AUCi-CARvalueswerepositively cor-related with letter fluency among ASz children (r=0.41,

p=0.02) and with verbal memory in the FHx group (r=0.46,p=0.04),butwerenotsignificantlyassociatedwith any neurocognitive measure in the TD group. In contrast, amongFHxchildren,AUCg-DAYvalues werenegatively cor-relatedwithverbalmemory(r= 0.47,p=0.05)andscores on the number-letter subtest (r= 0.62, p=0.008) and towers test(r= 0.55,p=0.02).The contrasting relation-ships betweenverbal memory andcortisol AUC values are furtherillustratedinFig.2.

3.4. Demographiccorrelatesofsalivary cortisol

Neither AUCi-CAR nor AUCg-DAY values were significantly associatedwithage,sex,ethnicity,pubertalstatus,tobacco use,timeofawakening,orsamplingmonth(p0.1). How-ever, therewas asignificantmain effectof socioeconomic statusonAUCg-DAYvalues (F[2,82]=6.31,p=0.003);post hoctestsindicatedthatchildrenclassedinthehighest socio-economiccategory (i.e.,whoseparents wereemployed in highermanagerial,administrative,andprofessional occupa-tions) had significantly lower AUCg-DAY values relative to children in the intermediate socioeconomic category (p=0.001).We thereforerepeatedthe AUCg-DAYanalyses describedabovewhilstadjustingforsocioeconomicstatus, butobservednochangetothepatternofresults(i.e.,there wasstillnosignificanteffectofFHxorofASzstatuson AUCg-DAYvalues).Additionally,theresultswereunchangedafter excludingoneparticipantfromtheASzgroupwhowasbeing treatedwithstimulantmedications(i.e.,ASzchildrendidnot differ significantly to the TDgroup on either AUCi-CARor AUCg-DAYvalues).

4.

Discussion

InthefirststudytoexaminetheCARinchildrenatelevated riskfor psychosis,childrenwitha familyhistoryof schizo-phrenia, butnotthose presentingmultiple antecedents of schizophrenia,werecharacterizedbyabluntedCARrelative totheirtypicallydevelopingpeers.However,neitherFHxnor ASzchildrenwerecharacterizedbyelevateddiurnalcortisol relativetotheTDgroup.TheCARwaspositivelycorrelated with distress relating to negative life events among FHx children,butnegativelycorrelatedintheTDgroup. Further-more,amongFHxandASzchildren,greaterabnormalityof

HPAaxisfunction(thatis,higherdiurnalcortisollevelsanda morebluntedCAR)wasassociatedwithpoorerperformance on tests of verbal memory and executive function. These resultsfurthersupportthenotionthatatleastsomeHPAaxis changes precede illness onset, rather thanbeing a subse-quentepiphenomenon,andmightparticipateinthe devel-opmentofneurocognitiveabnormalities.

We identifiedadissociationintherelationshipbetween psychosocialstressandcortisolamongFHxandTDchildren; specifically, distress relating to negative life events was positively correlated with the CAR in the FHx group but negatively correlated among TD children. Interestingly, a dissociationinstress-cortisolrelationships wasobservedin a study of patients with first episode psychosis which reported a negative correlation between diurnal cortisol and the number of recent stressful events in the patient group but a positive correlation among healthy controls (Mondellietal.,2010a).Fewstudiesofhigh-riskyouthhave measuredpsychosocialstressandcortisolconcurrently.One study of young adult siblings of patients with psychosis observedthatsiblingsshowed anincreasein cortisol after experiencingnegativeeventsduringtheday,butthathealthy controls did not (Collip et al., 2011). Thus, one possible explanationforthedissociationweobservedisthatpsychosis (and psychosis vulnerability) may influence the effect of psychosocial stress on HPA axis function. This may relate tohowthesestressfuleventsareappraised.Thedegreeto which stressors are viewed as potentially controllable or uncontrollablehas been associated with higherand lower morning cortisol levels, respectively (Miller et al., 2007). Thus,itispossiblethatFHxchildrenmayfeelthattheyare personally responsible for the negative life events they experience,whilstTDchildrenmightviewthese eventsas beingoutsideoftheircontrol.Ourfindingthatcortisollevels werenotcorrelatedwithdailyhasslescontrastswithastudy ofUHRyouthinwhichapositiveassociationwas observed betweenmorning cortisollevelsandthenumber ofhassles experiencedduringthepastmonth(Thompsonetal.,2007). Onepossibleexplanation forthelack ofassociationin our studyisthatdailyhassleswerenotalwaysassessedontheday ofcortisolcollection.Forexample,cortisollevelshavebeen foundto be higheron days whenindividualsexperience a greater number of stressors compared to stress-free days (Stawskietal.,2013).

Thecurrentstudyisthefirsttoexaminetherelationship betweenHPAaxisfunctionandneurocognitiveperformance amongindividualsatelevatedriskforpsychosis. Moderate-to-large impairments across a range of neurocognitive domainshave been observedconsistently in patients with firstepisodepsychosisrelative to healthycontrols(

Mesho-lam-Gatelyet al.,2009), andprospective studies indicate

thatsomeoftheseimpairmentsarepresentinchildrenwho goon to developthedisorderin laterlife(Dickson etal., 2012).Thecurrentstudysuggeststhatthedeficitsinmemory andexecutivefunctionthatwehaveobservedinASzandFHx children(Cullenetal.,2010;Dicksonetal.,2014),andwhich arealsoknowntocharacterizeUHRyouth(Fusar-Polietal., 2012),are associatedwithgreaterabnormalityofHPAaxis function. Specifically, a moreblunted CAR was associated withpoorerverbalmemoryandletterfluencyamongFHxand ASzchildren,respectively;moreover,intheFHxgrouponly, higherdiurnal cortisol was correlated with poorer perfor-manceon thenumber-lettersubtest and towers test, and lower verbal memory index scores. In contrast, HPA axis function was not associated with neurocognitive perfor-manceamongTDchildren.Thesefindingsareconsistentwith thoseofarecentstudyofpatientswithfirstepisodepsychosis in which a more blunted CAR was associated with poorer verbal memory and processing speed in patients but not

healthycontrols(Aasetal.,2011).Astudyofpatientswith chronicschizophreniareported thathigherdiurnalcortisol levels werecorrelated with poorer memoryand executive function (Walder et al., 2000). Further, a recent study reportedthatelevatedcortisollevelswereassociatedwith reduced hippocampal volume among individuals with first episodepsychosis(Mondellietal.,2010b).Thus,one expla-nationforourfindingsisthatstress-inducedHPAaxisfunction may have a direct effect on neurocognitive performance (i.e., via the effects of cortisol on the brain structures supporting these functions).Alternatively, common neuro-developmental mechanisms may influence both HPA axis functionandneurocognitiveperformance.

4.1. Limitations

relationship between HPA axis function and more severe formsofpsychosocialstresssuchaschildhoodmaltreatment, whichhasbeenimplicatedinthedevelopmentofpsychosis

(Schafer and Fisher, 2011) and associated with a blunted

cortisolresponsetostress(Ouellet-Morinetal.,2011).Thus, childhoodmaltreatmentmayhavebeenabletoexplainthe bluntedCARobservedamongFHxchildren.Thisisespecially likelyto bethecaseasindividualswithafamilyhistoryof psychosis are at elevated risk for exposure to childhood physicalabuse(Fisheretal.,2014).

4.2. Conclusions

The current study adds to the growing body of evidence suggesting that individuals on the trajectory to psychosis are characterized by abnormal HPA axis function. Albeit longitudinal studiesare needed to further understand the relevanceoftheseabnormalities,thesefindings,andthoseof aforementioned studies, support the targeting of the HPA axisasastrategyforearlyinterventionaimingtoreducethe riskofpsychosis.

Role

of

funding

source

None.

Conflict

of

interest

Nonedeclared.

Acknowledgments

ThisworkwassupportedbyfundingtoKRLfromaNational Institute for Health Research (NIHR) Career Development Fellowship (CDF/08/01/015); a Bial Foundation Research Grant (35/06); a NARSAD Young InvestigatorAward(2005); andtheBritishMedicalAssociationMargaretTempleAwardfor schizophrenia research (2006). AEC was supported by the WaterlooFoundation(164/1719)andHLFwas supportedby aMedicalResearchCouncil(MRC)PopulationHealthScientist award (G1002366).All authors areaffiliated withtheNIHR SpecialistBiomedicalResearchCentre(BRC)forMentalHealth at theSouth LondonandMaudsleyNational Health Service FoundationTrust andInstitute ofPsychiatry, King’s College London,UnitedKingdom.Theauthorsthankthechildrenand caregiverswhoparticipatedinthestudy,andtheresearchers andstudentswhocontributedtodatacollection.

References

Aas,M.,Dazzan,P.,Mondelli,V.,Toulopoulou,T.,Reichenberg,A.,Di Forti,M., Fisher, H.L.,Handley, R.,Hepgul, N., Marques, T., Miorelli, A., Taylor,H., Russo, M., Wiffen, B., Papadopoulos, A.,Aitchison, K.J., Morgan, C., Murray,R.M., Pariante, C.M., 2011.Abnormalcortisolawakeningresponsepredictsworse cog-nitivefunctioninpatientswithfirst-episodepsychosis.Psychol. Med.41,463—476.

Agnew-Blais,J.,Seidman,L.J.,2013.Neurocognitioninyouthand youngadultsunderage30atfamilialriskforschizophrenia:a quantitativeandqualitativereview.Cogn.Neuropsychiatry18, 44—82.

Aiello,G.,Horowitz,M.,Hepgul,N.,Pariante,C.M.,Mondelli,V., 2012.Stress abnormalities inindividuals atriskfor psychosis: areviewofstudiesinsubjectswithfamilialriskorwithatrisk mentalstate.Psychoneuroendocrinology37,1600—1613. AmericanPsychiatric Association,2013. Diagnostic andStatistical

ManualofMentalDisorders, 5thed.AmericanPsychiatric Asso-ciation,Arlington,VA.

BelvederiMurri,M.,Pariante,C.M.,Dazzan,P.,Hepgul,N., Papa-dopoulos,A.S.,Zunszain,P.,DiForti,M.,Murray,R.M.,Mondelli, V.,2012.Hypothalamic-pituitary-adrenalaxisandclinical symp-tomsin first-episode psychosis.Psychoneuroendocrinology 37, 629—644.

Borges,S., Gayer-Anderson, C.,Mondelli, V.,2013. A systematic review of the activity of the hypothalamic-pituitary-adrenal axis in first episode psychosis. Psychoneuroendocrinology 38, 603—611.

Brunelin,J.,d’Amato,T.,vanOs,J.,Cochet,A.,Suaud-Chagny,M.F., Saoud,M.,2008.Effectsofacutemetabolicstressonthe dopa-minergic and pituitary-adrenal axis activity in patients with schizophrenia,theirunaffectedsiblingsandcontrols.Schizophr. Res.100,206—211.

Carskadon,M.A.,Acebo,C.,1993.Aself-administeredratingscale forpubertaldevelopment.J.Adolesc.Health14,190—195. Cohen,J.,1992.Apowerprimer.Psychol.Bull.112,155—159. Collip,D.,Nicolson,N.A.,Lardinois,M.,Lataster,T.,vanOs,J.,

Myin-Germeys,I.,2011.Dailycortisol,stressreactivityandpsychotic experiences in individuals at above average genetic risk for psychosis.Psychol.Med.41,2305—2315.

Corcoran,C.,Perrin,M.,Harlap,S.,Deutsch,L.,Fennig,S.,Manor, O.,Nahon,D.,Kimhy,D.,Malaspina,D.,Susser,E.,2009.Effectof socioeconomicstatusandparents’educationatbirthonriskof schizophreniainoffspring.Soc.PsychiatryPsychiatr.Epidemiol. 44,265—271.

Cullen,A.E.,DeBrito,S.A.,Gregory,S.L.,Murray,R.M.,Williams, S.C.,Hodgins,S., Laurens,K.R., 2013.Temporal lobevolume abnormalities precede the prodrome: a study of children presenting antecedents of schizophrenia. Schizophr. Bull. 39, 1318—1327.

Cullen,A.E., Dickson, H., West, S.A.,Morris, R.G., Mould,G.L., Hodgins,S.,Murray,R.M.,Laurens,K.R.,2010.Neurocognitive performanceinchildrenaged9—12yearswhopresentputative antecedentsofschizophrenia.Schizophr.Res.121,15—23. Cullen,A.E., Fisher,H.L.,Roberts, R.E.,Pariante,C.M.,Laurens,

K.R.,2014.Dailystressorsandnegativelifeeventsinchildrenat elevated risk of developing schizophrenia. Br. J. Psychiatry, http://dx.doi.org/10.1192/bjp.bp.113.127001.

Delis,D.C.,Kaplan,E.,Kramer,J.H.,2001.TheDelis—Kaplan Exec-utive function System: Examiners Manual. The Psychological Corporation,SanAntonio.

Dickson,H.,Cullen,A.E.,Reichenberg,A.,Hodgins,S.,Campbell,D.D., Morris,R.G., Laurens,K.R., 2014. Cognitiveimpairment among childrenat-riskforschizophrenia.J.Psychiatr.Res.50,92—99. Dickson,H.,Laurens,K.R.,Cullen,A.E.,Hodgins,S.,2012.

Meta-analysesofcognitiveandmotorfunctioninyouthaged16years andyoungerwhosubsequentlydevelopschizophrenia.Psychol. Med.42,743—755.

Fisher,H.L.,McGuffin,P.,Boydell,J.,Fearon,P.,Craig,T.K.,Dazzan, P.,Morgan,K.,Doody,G.A.,Jones,P.B.,Leff,J.,Murray,R.M., Morgan,C.,2014.Interplaybetweenchildhoodphysicalabuse andfamilialriskintheonsetofpsychoticdisorders.Schizophr. Bull.,http://dx.doi.org/10.1093/schbul/sbt201.

Fusar-Poli, P., Deste, G., Smieskova,R., Barlati, S., Yung, A.R., Howes,O.,Stieglitz,R.D.,Vita,A.,McGuire,P.,Borgwardt,S., 2012. Cognitive functioning in prodromal psychosis: a meta-analysis.Arch.Gen.Psychiatry69,562—571.

Goodman,R.,2001.Psychometricpropertiesofthestrengthsand difficultiesquestionnaire.J.Am.Acad.ChildAdolesc.Psychiatry 40,1337—1345.

Herman,J.P.,Ostrander,M.M.,Mueller,N.K.,Figueiredo,H.,2005. Limbicsystem mechanisms ofstressregulation: hypothalamo-pituitary-adrenocorticalaxis.Prog.Neuropsychopharmacol.Biol. Psychiatry29,1201—1213.

Heubeck,B.,O’Sullivan,C.,1998.Anexplorationintothenature, frequencyandimpactofschoolhasslesinthemiddleschoolyears. Aust.Psychol.33,130—137.

Kirkbride,J.B.,Errazuriz,A.,Croudace,T.J.,Morgan,C.,Jackson, D.,Boydell, J., Murray, R.M., Jones, P.B.,2012. Incidence of schizophrenia and other psychoses in England, 1950—2009: a systematicreviewandmeta-analyses.PLoSONE7,e31660. Laurens,K.R., Hobbs, M.J., Sutherland,M., Green, M.J., Mould,

G.L.,2012.Psychotic-likeexperiencesinacommunitysampleof 8,000childrenaged9—11years:anitemresponsetheoryanalysis. Psychol.Med.47,1495—1506.

Laurens,K.R.,Hodgins,S.,Maughan,B.,Murray,R.M.,Rutter,M.L., Taylor, E.A., 2007. Community screening for psychotic-like experiences and other putative antecedents of schizophrenia inchildrenaged9—12years.Schizophr.Res.90,130—146. Laurens,K.R.,Hodgins,S.,Mould,G.L.,West,S.A.,Schoenberg,P.L.,

Murray,R.M., Taylor,E.A.,2010. Error-relatedprocessing dys-function in children aged 9 to 12 years presenting putative antecedentsofschizophrenia.Biol.Psychiatry67,238—245. Laurens,K.R.,Hodgins,S.,Taylor,E.,Murray,R.M.,2011.Isearlier

intervention for schizophrenia possible?: Identifying antece-dentsofschizophreniainchildrenaged9—12years. In:David, A.S., McGuffin,P., Kapur,S.(Eds.), Schizophrenia:TheFinal Frontier.PsychologyPress,London.

Lipsey, M.W., Wilson, D.B., 2001. Practical meta-analysis. Sage Publications,ThousandOaks,CA.

Lupien,S.J.,Ouellet-Morin,I.,Trepanier,L.,Juster,R.P.,Marin,M.F., Francois,N.,Sindi,S.,Wan,N.,Findlay,H.,Durand,N.,Cooper, L.,Schramek,T.,Andrews,J.,Corbo,V.,Dedovic,K.,Lai,B., Plusquellec,P.,2013.TheDeStressforsuccessprogram:effectsof a stress education program on cortisol levels and depressive symptomatologyin adolescents making the transition to high school.Neuroscience249,74—87.

Mackie,C.J.,Castellanos-Ryan,N.,Conrod,P.J.,2011. Developmen-taltrajectoriesofpsychotic-likeexperiencesacrossadolescence: impactofvictimizationandsubstanceuse.Psychol.Med.41,47— 58.

Macmanus,D.,Laurens,K.R.,Walker,E.F.,Brasfield,J.L.,Riaz,M., Hodgins,S., 2012. Movementabnormalitiesand psychotic-like experiencesinchildhood:markersofdevelopingschizophrenia? Psychol.Med.42,99—109.

Maxwell,M.E.,1992.FamilyInterviewforGeneticStudies.National InstituteofMentalHealth,St.Louis,MO.

McVie,S.,Bradshaw,P.,2005.AdolescentSmoking,DrinkingandDrug Use,TheEdinburghStudyofYouthTransitionsandCrime.The UniversityofEdinburgh,Edinburgh.

Mesholam-Gately, R.I., Giuliano, A.J., Goff, K.P., Faraone, S.V., Seidman,L.J.,2009.Neurocognitioninfirst-episode schizophre-nia:ameta-analyticreview.Neuropsychology23,315—336. Miller,G.E.,Chen,E.,Zhou,E.S.,2007.Ifitgoesup,mustitcome

down?Chronicstressandthe hypothalamic-pituitary-adrenocor-ticalaxisinhumans. Psychol.Bull.133,25—45.

Mittal,V.A.,Dhruv,S.,Tessner,K.D.,Walder, D.J.,Walker,E.F., 2007.Therelationsamongputativebioriskmarkersin schizo-typal adolescents: minor physical anomalies, movement abnormalities, and salivary cortisol. Biol. Psychiatry 61, 1179—1186.

Mondelli,V.,Dazzan,P.,Hepgul,N.,DiForti,M.,Aas,M.,D’Albenzio, A.,DiNicola,M.,Fisher,H.,Handley,R.,Marques,T.R.,Morgan, C., Navari, S., Taylor, H., Papadopoulos, A., Aitchison, K.J., Murray,R.M., Pariante,C.M.,2010a. Abnormal cortisol levels duringthedayandcortisolawakeningresponseinfirst-episode psychosis: the role of stress and of antipsychotic treatment. Schizophr.Res.116,234—242.

Mondelli,V.,Pariante,C.M.,Navari,S.,Aas,M.,D’Albenzio,A.,Di Forti,M., Handley, R.,Hepgul, N., Marques,T.R., Taylor,H., Papadopoulos, A.S.,Aitchison,K.J., Murray, R.M., Dazzan,P., 2010b. Highercortisol levels are associated with smaller left hippocampal volumeinfirst-episodepsychosis.Schizophr.Res. 119,75—78.

OfficeforNationalStatistics,2010.NeighbourhoodStatisticsbyLocal Authority.NationalStatistics.

Ouellet-Morin,I.,Odgers,C.L.,Danese,A.,Bowes,L.,Shakoor,S., Papadopoulos,A.S.,Caspi,A.,Moffitt,T.E.,Arseneault,L.,2011. Bluntedcortisolresponsestostresssignalsocialandbehavioral problemsamongmaltreated/bullied12-year-oldchildren.Biol. Psychiatry70,1016—1023.

Pruessner,J.C.,Kirschbaum,C.,Meinlschmid,G.,Hellhammer,D.H., 2003.Twoformulasforcomputationoftheareaunderthecurve representmeasuresoftotalhormoneconcentrationversus time-dependentchange.Psychoneuroendocrinology28,916—931. Pruessner,M.,Vracotas,N.,Joober,R.,Pruessner,J.C.,Malla,A.K.,

2013. Blunted cortisol awakening response in men with first episodepsychosis:relationshiptoparentalbonding. Psychoneur-oendocrinology38,229—240.

Schafer, I.,Fisher,H.L.,2011. Childhoodtraumaandpsychosis— whatistheevidence? Dialog.Clin.Neurosci.13,360—365. Sheslow,D.,Adams,W.,2003.WideRangeAssessmentofMemoryand

Learning—SecondEdition(WRAML2).WiderangeInc.,Delaware. Shirtcliff,E.A.,Allison,A.L.,Armstrong,J.M.,Slattery,M.J.,Kalin, N.H.,Essex,M.J.,2012.Longitudinalstabilityanddevelopmental propertiesofsalivarycortisollevelsandcircadianrhythmsfrom childhoodtoadolescence.Dev.Psychobiol.54,493—502. Stateva,M.,Minton,J.,Beckett,C.,Doolan,M.,Ford,T.,

Kallitso-glou, A., Scott, S., 2012. Challenges recruiting familieswith childrenatriskofanti-socialbehaviourintointerventiontrials: lessonsfromtheHelpingChildrenAchieve(HCA)study.J.Child. Serv.7,285—302.

Stawski,R.S.,Cichy,K.E.,Piazza,J.R.,Almeida,D.M.,2013. Asso-ciationsamongdailystressorsandsalivarycortisol:findingsfrom theNationalStudyofDailyExperiences. Psychoneuroendocrinol-ogy38,2654—2665.

Sugranyes, G., Thompson, J.L., Corcoran, C.M., 2012. HPA-axis function,symptoms,andmedicationexposureinyouthsat clini-calhighriskforpsychosis.J.Psychiatr.Res.46,1389—1393. Thompson,K.N., Phillips,L.J.,Komesaroff,P.,Yuen,H.P., Wood,

S.J.,Pantelis,C.,Velakoulis,D.,Yung,A.R.,McGorry,P.D.,2007. StressandHPA-axisfunctioninginyoungpeopleatultrahighrisk forpsychosis.J.Psychiatr.Res.41,561—569.

Vreeburg,S.A.,Kruijtzer,B.P.,vanPelt,J.,vanDyck,R.,DeRijk, R.H.,Hoogendijk,W.J.,Smit,J.H.,Zitman,F.G.,Penninx,B.W., 2009. Associations between sociodemographic, sampling and healthfactorsandvarioussalivarycortisolindicatorsinalarge samplewithoutpsychopathology.Psychoneuroendocrinology34, 1109—1120.

Walder,D.J.,Walker,E.F.,Lewine,R.J.,2000.Cognitivefunctioning, cortisolrelease,andsymptomseverityinpatientswith schizo-phrenia.Biol.Psychiatry48,1121—1132.

Walker,E.,Mittal,V.,Tessner,K.,2008.Stressandthehypothalamic pituitaryadrenalaxisinthedevelopmentalcourseof schizophre-nia.Annu.Rev.Clin.Psychol.4,189—216.

Walker,E.F., Diforio,D., 1997.Schizophrenia:a neural diathesis-stressmodel.Psychol.Rev.104,667—685.

Walker,E.F.,Trotman,H.D.,Pearce,B.D.,Addington,J.,Cadenhead, K.S., Cornblatt, B.A., Heinssen, R., Mathalon, D.H., Perkins, D.O., Seidman, L.J., Tsuang, M.T., Cannon, T.D., McGlashan, T.H.,Woods,S.W.,2013.Cortisollevelsandriskforpsychosis: initialfindingsfromtheNorthAmericanProdromeLongitudinal Study.Biol.Psychiatry74,410—417.

Wicks,S.,Hjern,A.,Gunnell,D.,Lewis,G.,Dalman,C.,2005.Social adversity inchildhoodandtheriskofdevelopingpsychosis: a nationalcohortstudy.Am.J.Psychiatry162,1652—1657. WorldHealthOrganization,1992.TheICD-10ClassificationofMental

andBehaviouralDisorders:ClinicalDescriptionsand Diagnostic Guidelines.WorldHealthOrganisation,Geneva.

Wust,S.,Federenko,I.,Hellhammer,D.H.,Kirschbaum,C.,2000. Genetic factors, perceived chronic stress, and the free

cortisol response to awakening. Psychoneuroendocrinology 25,707—720.

Yang,J.,Liu,X.,Wang,L.,Lv,H.,Yu,J.,Xun,Z.,Yang,G.,2012. Abnormalityofglycometabolismrelatedfactorsinnon-psychotic offspringofschizophrenicpatients.PsychiatryRes.198,183—186. Yehuda,R.,Bierer,L.M.,Schmeidler,J.,Aferiat,D.H.,Breslau,I., Dolan,S.,2000.LowcortisolandriskforPTSDinadultoffspringof holocaustsurvivors.Am.J.Psychiatry157,1252—1259.