Maternal depressive symptoms during and after pregnancy are associated with attention-deficit/hyperactivity disorder symptoms in their 3- to 6-year-old children

RESEARCH ARTICLE Ma e nal dep essi e symp oms du ing and a e p egnancy a e associa ed wi h a en ion- de ici /hype ac i i y diso de symp oms in hei 3- o 6-yea -old child en Elina Wol o d 1 *, Ma ius Lah i 1,2 , Soile Tuo inen 1 , Ja i Lah i 1,3,4 , Ja i Lipsanen 1 , Ka i Sa olainen 1 , Ka i Heinonen 1 , Esa Ha ¨ma ¨la ¨inen 5 , Ee o Kajan ie 6,7,8 , Anu- Ka iina Pesonen 1 , Pia M. Villa 9 , Hannele Lai uo i 10,11,12,13 , Rebecca M. Reynolds 2 , Ka i Ra ¨ikko ¨nen 1 1Depa men o Psychology and Logopedics, Facul y o Medicine, Uni e si y o Helsinki, Helsinki, Finland, 2Uni e si y/B i ish Hea Founda ion Cen e o Ca dio ascula Science, Queen’s Medical Resea ch Ins i u e, Uni e si y o Edinbu gh, Edinbu gh, Uni ed Kingdom, 3Helsinki Collegium o Ad anced S udies, Uni e si y o Helsinki, Helsinki, Finland, 4Folkha ¨lsan Resea ch Cen e, Helsinki, Finland, 5Depa men o Clinical Chemis y, Uni e si y o Helsinki, Helsinki, Finland, 6Na ional Ins i u e o Heal h and Wel a e, Helsinki, Finland, 7Child en’s Hospi al, Helsinki Uni e si y Hospi al and Uni e si y o Helsinki, Helsinki, Finland, 8PEDEGO Resea ch Uni , MRC Oulu, Oulu Uni e si y Hospi al and Uni e si y o Oulu, Oulu, Finland, 9Obs e ics and Gynaecology, Uni e si y o Helsinki and Helsinki Uni e si y Hospi al, Helsinki, Finland, 10 Facul y o Medicine and Li e Sciences, Uni e si y o Tampe e, Tampe e, Finland, 11 Depa men o Obs e ics and Gynecology, Tampe e Uni e si y Hospi al, Tampe e, Finland, 12 Medical and Clinical Gene ics, Uni e si y o Helsinki and Helsinki Uni e si y Hospi al, Helsinki, Finland, 13 Ins i u e o Molecula Medicine Finland, Helsinki Ins i u e o Li e Science, Uni e si y o Helsinki, Helsinki, Finland *elina.wol o d@helsinki. i Abs ac Ma e nal dep essi e symp oms du ing p egnancy ha e been associa ed wi h child beha- iou al symp oms o a en ion-de ici /hype ac i i y diso de (ADHD) in ea ly childhood. How- e e , i emains unclea i dep essi e symp oms h oughou p egnancy a e mo e ha m ul o he child han dep essi e symp oms only du ing ce ain imes, and i ma e nal dep essi e symp oms a e p egnancy add o o media e any p ena al e ec s. 1,779 mo he -child dyads pa icipa ed in he P edic ion and P e en ion o P e-eclampsia and In au e ine G ow h Res ic ion (PREDO) s udy. Mo he s illed in he Cen e o Epidemiological S udies Dep es- sion Scale biweekly om 12+0–13+6 o 38+0–39+6 weeks+days o ges a ion o deli e y, and he Beck Dep ession In en o y-II and he Conne s’ Hype ac i i y Index a he child’s age o 3 o 6 yea s (mean 3.8 yea s, s anda d de ia ion [SD] 0.5). Ma e nal dep essi e symp oms we e highly s able h oughou p egnancy, and child en o mo he s wi h consis- en ly high dep essi e symp oms showed highe a e age le els (mean di e ence = 0.46 SD uni s, 95% Con idence In e al [CI] 0.36, 0.56, p<0.001 compa ed o he low g oup), and p opo ion (32.1% s. 14.7%) and odds (odds a io = 2.80, 95% CI 2.20, 3.57, p<0.001) o clinically signi ican ADHD symp oms. These associa ions we e no explained by he e ec s o ma e nal dep essi e symp oms a e p egnancy, which bo h added o and pa ially media ed he p ena al e ec s. Ma e nal dep essi e symp oms h oughou p egnancy a e associa ed wi h inc eased ADHD symp oma ology in young child en. Ma e nal dep essi e PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0190248 Decembe 21, 2017 1 / 13 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Ci a ion: Wol o d E, Lah i M, Tuo inen S, Lah i J, Lipsanen J, Sa olainen K, e al. (2017) Ma e nal dep essi e symp oms du ing and a e p egnancy a e associa ed wi h a en ion-de ici /hype ac i i y diso de symp oms in hei 3- o 6-yea -old child en. PLoS ONE 12(12): e0190248. h ps://doi. o g/10.1371/jou nal.pone.0190248 Edi o : Ma ianna Mazza, Uni e si a Ca olica del Sac o Cuo e Sede di Roma, ITALY Recei ed: Augus 23, 2017 Accep ed: Decembe 11, 2017 Published: Decembe 21, 2017 Copy igh : ©2017 Wol o d e al. This is an open access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal au ho and sou ce a e c edi ed. Da a A ailabili y S a emen : The PREDO S udy da a con ain sensi i e and po en ially iden i ying pa ien in o ma ion, e en a e social secu i y numbe s and names ha e been emo ed om he da a se . The use and sha ing o such da a a e s ic ly con olled by se e al clauses o he Finnish law, designed o ensu e he p o ec ion o p i acy. The esea che s in e es ed in using he da a mus ob ain app o al om he PREDO S udy Boa d. The S udy Boa d is accoun able o he na ional egis e au ho i y ( he Finnish Na ional Ins i u e o Heal h symp oms a e p egnancy add o, bu only pa ially media e, he p ena al e ec s. P e en- i e in e en ions sui ed o he p egnancy pe iod may bene i bo h ma e nal and o sp ing men al heal h. In oduc ion A en ion-de ici /hype ac i i y diso de (ADHD) is cha ac e ized by a pe sis en pa e n o ina en ion, impulsi i y, and hype ac i i y. I is one o he mos p e alen neu ode elopmen al diso de s in child en wi h p e alence a es a ying om 5.9 o 7.1% [1,2]. These symp oms may no only be associa ed wi h impai men s in academic, socioeconomic and social domains [3], bu also p edic p ema u e mo ali y [4]. In he ecen decade, he p e alence a es o ADHD ha e shown a nea ly 30% inc ease [5]. As ou gene ic makeup has no changed, his inc ease canno be a ibu ed o gene ic ac o s. E en hough inc easing a ailabili y o se ices, ecogni ion o , and sc eening o ADHD may pa ly unde lie his inc ease [6], ano he con ibu ing ac o may lie in he exposu e o en i- onmen al ad e si ies in he p ena al s age o li e. Among hese ad e si ies a e exposu e o ma e nal p e-p egnancy obesi y and hype ensi e and diabe ic p egnancy diso de s. They complica e an inc easing numbe o p egnancies and ha e been linked wi h an inc eased isk o ADHD symp oma ology in he child [7–10]. They a e also among he majo unde lying causes o p e e m bi h and low bi hweigh , which p e ious s udies ha e also linked wi h an inc eased isk o ADHD symp oma ology [11–13]. Ye , al hough ex ensi e esea ch on he e ec s o ma e nal dep ession on o sp ing ou - comes has s a ed o eme ge [14,15], ela i ely li le a en ion has been de o ed o he associa- ion wi h o sp ing ADHD symp oma ology. I has been es ima ed ha 7 o 20% o women expe ience clinically signi ican le els o dep essi e symp oms a di e en s ages o p egnancy [16,17]. We a e awa e o only one la ge-scale e ospec i e s udy which has shown ha he mo he s o 2-11-yea -old child en wi h an ADHD diagnosis we e mo e likely o be diagnosed wi h dep ession he yea be o e he bi h o he child [18]. Two addi ional p ospec i e s udies ha e shown ha ma e nal dep essi e symp oms epo ed a ges a ional week 20 we e associ- a ed wi h a highe isk o child a en ion p oblems a 3 yea s o age [19], and when epo ed a ges a ional weeks 18 and 32, hey we e associa ed wi h a highe isk o child a en ion and hype ac i i y p oblems a 4 and 11 yea s o age [19,20]. These s udies a e, howe e , limi ed by a numbe o easons. Fi s , hey measu ed dep essi e symp oms “du ing he pas se en days o las wo weeks” a one o wo ime poin s du ing p egnancy no co e ing he en i e p egnancy [19,20]. Second, while wo o hese s udies accoun ed o ma e nal dep ession a e p egnancy [19,20], none o he s udies es ed i ma e - nal dep ession a e p egnancy added o o media ed ei he ully o pa ially he p ena al e ec s. Finally, he s udies ailed o accoun o ma e nal p e-p egnancy obesi y and common p egnancy diso de s, which in addi ion o inc easing he child’s ADHD isk [8–10], may o en also accompany ma e nal dep ession [21]. Hence, we es ed, in a la ge sample o p egnan Finnish women, i dep essi e symp oms, measu ed biweekly om ges a ional week 12 onwa ds un il e m o deli e y, we e associa ed wi h ADHD symp oms in hei 3- o 6-yea -old child en. The biweekly assessmen s allowed us o add ess ges a ion-week and imes e -speci ic e ec s, and ma e nal e- a ings o dep essi e symp oms a he ime o a ing he 3- o 6-yea -old child allowed us o add ess i any e ec s we e speci ic o he p ena al s age. Ou s udy also es ed i ma e nal dep essi e symp oms a e Ma e nal dep essi e symp oms du ing p egnancy and child a en ion-de ici /hype ac i i y diso de symp oms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0190248 Decembe 21, 2017 2 / 13 and Wel a e) and o he e hics commi ees which ha e app o ed he s udy p o ocol ( he E hics Commi ees o Obs e ics and Gynaecology, Child en’s Diseases and Psychia y, and Women, Child en and Psychia y o he Hospi al Dis ic o Helsinki and Uusimaa and by he pa icipa ing hospi als in Finland). Collabo a ion in da a analysis is possible h ough speci ic esea ch p oposals sen o he PREDO S udy Boa d [p edo. [email p o ec ed]] o p ima y in es iga o s Ka i Ra¨ikko¨nen [ka i. aikkonen@helsinki. i] o Hannele Lai uo i [[email p o ec ed]]. Funding: The PREDO s udy is unded by he Academy o Finland, E aNe , EVO (a special s a e subsidy o heal h science esea ch), Uni e si y o Helsinki Resea ch Funds, he Signe and Ane Gyllenbe g Founda ion, he Emil Aal onen Founda ion, he Finnish Medical Founda ion, he Jane and Aa os E kko Founda ion, he No o No disk Founda ion, he Pa¨i ikki and Saka i Sohlbe g Founda ion, he Sig id Juselius Founda ion, and he Si Jules Tho n Cha i able T us . The unde s had no ole in s udy design, da a collec ion and analysis, decision o publish, o p epa a ion o he manusc ip . Compe ing in e es s: The au ho s ha e decla ed ha no compe ing in e es s exis wi h No o No disk Founda ion (an independen ounda ion wi h co po a e in e es s) o any o he unde s. This does no al e ou adhe ence o PLOS ONE policies on sha ing da a and ma e ials. p egnancy added o o media ed any o he p ena al e ec s. Finally, we es ed i ma e nal p e- p egnancy obesi y, hype ensi e p egnancy diso de s, and ges a ional diabe es, o ma e nal ADHD symp oms accoun ed o any obse ed e ec s. We ha e p e iously demons a ed in his coho associa ions be ween ma e nal dep essi e symp oms and child in e nalizing, ex e - nalizing, and o al p oblems, including DSM IV-o ien ed ADHD p oblems [16] as measu ed by he Child Beha iou Checklis (CBCL) [22]. This s udy adds o hese indings by es ing associa ions wi h he Conne s’ Hype ac i i y Index (CHI) [23], which has been shown o be alid in iden i ying child en 3 yea s and olde wi h ADHD [23,24]. CHI is a comp ehensi e measu e o ADHD symp oms in child en 3 yea s and olde , including aspec s o ina en ion, hype ac i i y, and emo ional labili y [23–25], while he ADHD p oblems scale o he CBCL is sui able o child en 1.5 yea s and olde and ocuses mo e on hype ac i e-impulsi e and ina - en i e symp oms [22]. Ma e ials and me hods Pa icipan s The P edic ion and P e en ion o P e-eclampsia and In au e ine G ow h Res ic ion (PREDO) s udy comp ises al oge he 4,777 mo he s and hei single on o sp ing bo n ali e in Finland be ween 2006 and 2010 [26]. The women we e ec ui ed when hey a ended he i s ul asound sc eening be ween 12+0–13+6 weeks+days o ges a ion in an ena al clinics a one o he en s udy hospi als in Sou he n and Eas e n Finland. O hem, 3,402 (71.2%) assessed hei dep essi e symp oms biweekly du ing p egnancy. In 2011–2012 we in i ed 4,586 mo he -child dyads ( h ee child en had died be o e he ol- low-up, 33 had no da a in he Finnish Medical Bi h Regis e (MBR), 55 women declined pa - icipa ion in a ollow-up, and o 100 women, add esses we e no aceable), and 2,667 (58.2%) pa icipa ed. O hem 2,312 (68.0% o hose wi h da a on dep essi e symp oms du ing p eg- nancy) had p egnancy as well as ollow-up da a a ailable a he child’s age o 1.9 o 6.3 yea s (50.6% boys). Since he CHI is alida ed o child en who a e 3 yea s and olde [23], we excluded 533 child en om he analy ic sample. Hence, he cu en s udy comp ised o 1,779 mo he -child pai s who had bo h p egnancy as well as ollow-up da a a ailable a he child’s mean age o 3.8 yea s (S anda d De ia ion (SD) = 0.5 yea s, ange 3.0 o 6.3 yea s; 51.5% boys). Compa ed o he women who we e in i ed bu did no pa icipa e in he ollow-up (n= 2,274), he women who pa icipa ed and whose child en in he ollow-up we e 3 yea s and olde (n= 1,779) we e olde a deli e y (31.9 s. 31.1 yea s, p<0.001), had mo e o en a e ia y educa ion (61.6% s. 54.1%, p<0.001), we e less o en single (1.5% s. 3.6%, p<0.001), we e less o en mul ipa ous (57.7% s. 64.4%, p<0.001), smoked less o en h oughou p egnancy (2.6% s. 7.5%, p<0.001), and epo ed less o en a his o y o a dep es- sion diagnosis (9.0% s. 12.6%, p= 0.002). The PREDO s udy p o ocol was app o ed by he E hics Commi ees o Obs e ics and Gynaecology, Child en’s Diseases and Psychia y, and Women, Child en and Psychia y o he Hospi al Dis ic o Helsinki and Uusimaa and by he pa icipa ing hospi als. All pa icipa ing women signed in o med consen o ms and all p ocedu es con ibu ing o his wo k comply wi h he Helsinki Decla a ion o 1975, as e ised by he 59 h WMA Gene al Assembly, Seoul, Republic o Ko ea, Oc obe 2008. Child ADHD symp oms A he child’s age o 3 o 6 yea s, hei mo he s a ed he en CHI ques ions on he child’s beha- iou al symp oms o ADHD on a scale o “no a all” (0) o “ e y much” (3) [23]. A sum-sco e o 10 o abo e indica es clinically signi ican ADHD symp oms [27]. The scale has good Ma e nal dep essi e symp oms du ing p egnancy and child a en ion-de ici /hype ac i i y diso de symp oms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0190248 Decembe 21, 2017 3 / 13 in e nal consis ency [24,25] and disc iminan alidi y [24,25]. In ou sample, i showed high in e nal consis ency (α= .91). Ma e nal dep essi e symp oms du ing and a e p egnancy Dep essi e symp oms we e epo ed biweekly up o 14 imes h oughou p egnancy s a ing om 12+0–13+6 o 38+0–39+6 weeks+days ges a ion o deli e y using he Cen e o Epide- miological S udies Dep ession Scale (CES-D) [28]. The CES-D has 20 ques ions a ed on a scale o none o he ime (0) o all he ime (3) du ing he pas week, wi h highe sco es indica - ing mo e equen symp oms o dep ession, and a sum-sco e o 16 indica ing a isk o clini- cal dep ession [28]. In he ollow-up, dep essi e symp oms we e epo ed using he Beck Dep ession In en- o y-II (BDI-II) [29]. This scale comp ises 21 ou -s a emen se s (sco ed om 0 o 3) wi h each s a emen e lec ing inc easing se e i y o dep essi e symp oms du ing he pas wo weeks [29]. A sum-sco e o 14 indica es a leas mild dep essi e symp oms [29]. Bo h dep ession scales ha e good psychome ic p ope ies [28–30], and he CES-D has been used ex ensi ely and alida ed also in p egnan popula ions [30]. We ha e p e iously shown ha in ou sample he CES-D (C onbach‘s α= .88 o .92 in he 14 biweekly measu e- men poin s) and he BDI-II (α= .91) showed high in e nal consis ency [16]. Ma e nal p e-p egnancy obesi y and p egnancy diso de s Ma e nal p e-p egnancy obesi y (BMI, 30 kg/m 2 ), ges a ional diabe es (yes s. no) and hype ensi e p egnancy diso de s (p e-eclampsia, ges a ional hype ension; yes s. no) we e ex ac ed om he MBR and/o om medical eco ds independen ly e i ied by a clinical ju y. Co a ia es These included ma e nal his o y o physician-diagnosed dep ession which was epo ed in a ques ionnai e a 12+0–13+6 weeks+days o ges a ion. Ques ions on ma e nal ADHD p oblems we e embedded in he Adul Sel -Repo (ASR) [31], which he mo he s comple ed in he ol- low-up (T-sco e 65 poin s indica es bo de line signi ican p oblems). Ma e nal age a deli - e y (yea s), an idep essan use (yes s. no), psycho opic medica ion use (yes s. no), smoking du ing p egnancy (did no smoke/ qui du ing he i s imes e / smoked h oughou p eg- nancy), pa i y (p imipa ous s. mul ipa ous), p e-p egnancy/ch onic hype ension (yes s. no), ype 1 diabe es (yes s. no), child’s sex, ges a ional leng h (weeks), bi hweigh (g) adjus ed o sex and ges a ion leng h, and amily s uc u e (cohabi a ing/ma ied s. single) a child- bi h we e de i ed om he MBR, and ma e nal alcohol use (yes s. no) and educa ion (sec- onda y o less, uppe seconda y, lowe e ia y, uppe e ia y) we e mo he - epo ed du ing p egnancy, and child’s age a he ollow-up. S a is ical analyses We i s examined ma e nal dep essi e symp oms p o iles du ing p egnancy wi h a la en p o- ile analysis. We compa ed solu ions wi h wo o eigh clus e s, and iden i ied he mos op imal one by using Akaike In o ma ion C i e ion, sample size-adjus ed Bayesian In o ma ion C i e- ion, and Vuong-Lo-Mendell-Rubin Likelihood Ra io Tes and Lo-Mendell-Rubin Adjus ed Likelihood Ra io Tes s. We hen es ed i he child ADHD symp om sco es, ea ed as a con in- uous ou come a iable, and he p opo ion o child en wi h clinically signi ican ADHD symp- oms, ea ed as a dicho omous a iable using he ADHD symp om sco e 10 o abo e as a clinical cu o [27], di e ed be ween he g oups o mo he s wi h di e en dep essi e symp om Ma e nal dep essi e symp oms du ing p egnancy and child a en ion-de ici /hype ac i i y diso de symp oms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0190248 Decembe 21, 2017 4 / 13 p o iles du ing p egnancy. These g oup di e ences a e p esen ed as mean di e ences (MD) and odds a ios (OR) and hei 95% Con idence In e als (CI) om linea (con inuous ADHD symp om sco es) and logis ic eg ession analyses (ADHD symp om sco es dicho omized a clinical cu o ), espec i ely. We also examined i he associa ions be ween ma e nal dep essi e symp oms du ing p eg- nancy and child ADHD symp oms we e ges a ion-week- o imes e -speci ic. In hese es s, we used linea eg ession analysis when we ea ed child ADHD symp oms as con inuous and logis ic eg ession analysis when we dicho omized child ADHD symp oms sco es a he clini- cal cu o . Fu he , in hese analyses ma e nal dep essi e symp om sco es (biweekly alues; i s imes e alue, mean alues o he second and hi d imes e alues; imes e -weigh ed mean alue) we e squa e oo ans o med o imp o e linea model i ing. In all o he abo e analyses we i s made adjus men s o child’s sex and age a ollow-up (model 1). The ea e , we addi ionally adjus ed o ma e nal age a childbi h, pa i y, amily s uc u e, educa ion le el, ype 1 diabe es, ch onic hype ension, his o y o physician-diag- nosed dep ession, an idep essan and o he psycho opic medica ion use, alcohol use and smoking du ing p egnancy, and ges a ion leng h and weigh a bi h adjus ed o sex and ges- a ion leng h (model 2); o ma e nal p e-p egnancy obesi y, ges a ional diabe es, ges a ional hype ension, and p e-eclampsia (model 3); o ma e nal ADHD p oblems (model4); and inally, o all o he abo e and ma e nal dep essi e symp oms a ollow-up pa allel o a ing he child (model 5). We also es ed i ma e nal dep essi e symp oms a e p egnancy added o he p ena al e ec s wi h an in e ac ion e m o ma e nal imes e -weigh ed mean dep essi e symp oms du ing p egnancyma e nal dep essi e symp oms a e p egnancy ha was added o he linea (con inuous ADHD symp om sco es) and logis ic eg ession models (ADHD symp oms sco es dicho omized a he clinical cu o ). In addi ion, we es ed i ma e nal dep essi e symp- oms a e p egnancy media ed he e ec s o ma e nal imes e -weigh ed mean dep essi e symp oms du ing p egnancy using he PROCESS mac o o media ion in SPSS 24 wi h 5000 boo s apping e-samples wi h bias-co ec ed CIs [32,33]. Finally, we conduc ed sensi i i y analyses by unning he linea eg ession analyses sepa a ely in g oups acco ding o ma e nal p e-p egnancy obesi y and p egnancy diso de s, child’s sex, ma e nal his o y o physician- diagnosed dep ession, and ma e nal ADHD p oblems. We used MPLUS and SPSS 24 da a packages o he analyses. Resul s Cha ac e is ics o he s udy pa icipan s a e in S1 Table. Co ela ions be ween he co a ia es wi h child ADHD symp oms a e in S2 Table. Ma e nal biweekly, imes e -speci ic and imes- e -weighed mean alues o dep essi e symp oms du ing p egnancy we e signi ican ly co e- la ed (Pea son ’s anged om .51 o .92, all p- alues <.001) and we e also signi ican ly co ela ed wi h dep essi e symp oms a e p egnancy (Pea son ’s anged om 0.36 o 0.46, all p- alues <0.001) [16]. The median numbe o consecu i e dep essi e symp om measu e- men s du ing p egnancy in he en i e sample was 13 and he in e qua ile ange was 12 o 14. The e we e al oge he 879 (49.4%) women wi h da a on all 14 measu emen poin s du ing p egnancy and only 334 (18.8%) women had mo e han wo missing alues du ing p egnancy. Ma e nal dep essi e symp oms du ing p egnancy and child ADHD symp oms Fig 1 (Panel A) shows ha he mos op imal la en p o ile solu ion (in compa ison o solu ions wi h h ee o eigh g oups) iden i ied wo g oups o women wi h consis en ly low and high Ma e nal dep essi e symp oms du ing p egnancy and child a en ion-de ici /hype ac i i y diso de symp oms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0190248 Decembe 21, 2017 5 / 13 le els o dep essi e symp oms h oughou p egnancy (Akaike In o ma ion C i e ion = 147821.10, sample-size-adjus ed Bayesian In o ma ion C i e ion = 147920.30, Vuong-Lo- Mendell-Rubin LRT and Lo-Mendell-Rubin-adjus ed likelihood a io es p- alues <0.001). Fo he wo la en p o ile g oups, he pe cen age o women wi h da a on all 14 measu emen poin s compa ed o he ones wi h a leas one missing alue was no signi ican ly di e en (50.9% in he low and 45.9% in he high dep essi e symp om le el g oup had all 14 measu e- men poin s, p= 0.051 o g oup di e ence). Fig 1 (Panels B and C) also shows ha child Fig 1. La en p o ile analysis showing he mos op imal, wo g oup, solu ion o mo he s wi h consis en ly low and high sco es on he Cen e o Epidemiological S udies Dep ession Scale (CES-D) h oughou p egnancy (Panel A), and he child’s beha iou al symp oms o a en ion-de ici / hype ac i i y diso de on he Conne s’ Hype ac i i y Index (CHI) es ima ed ma ginal mean sco es (Panel B), and p opo ion o child en wi h sco es abo e he clinical cu o (10) in he CHI (Panel C). E o ba s e e o he 95% Con idence In e als (95% CI), and numbe s o mean di e ence (MD) (Panel B) and odds a io (OR) (Panel C) and hei 95% CIs in model 1, and p- alues o models 1–5. Fo di e en adjus men models, please see oo no e in Table 1. h ps://doi.o g/10.1371/jou nal.pone.0190248.g001 Ma e nal dep essi e symp oms du ing p egnancy and child a en ion-de ici /hype ac i i y diso de symp oms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0190248 Decembe 21, 2017 6 / 13 ADHD symp oms and he p opo ion o and odds o child en wi h clinically signi ican ADHD symp oms we e highe in he g oup o women who had consis en ly high dep essi e symp oms h oughou p egnancy. Fu he , highe ma e nal biweekly, imes e -speci ic (S3 Table) and imes e -weigh ed mean dep essi e symp om sco es (Table 1) we e signi ican ly associa ed wi h highe ADHD symp om sco es and highe odds o clinically signi ican ADHD symp oms (Table 1) in chil- d en. These associa ions we e signi ican ac oss all adjus men models wi h all co a ia es, including ma e nal ADHD symp oms. Addi i e e ec s o ma e nal dep essi e symp oms du ing and a e p egnancy on child ADHD symp oms Fig 2 (Panels A and B) shows ha ma e nal dep essi e symp oms a e p egnancy added o he e ec o ma e nal dep essi e symp oms du ing p egnancy (bo h p- alues o in e ac ions = 0.03 o dep essi e symp oms du ing p egnancydep essi e symp oms a e p egnancy in e ac ion on child con inuous and clinically signi ican ADHD symp oms sco es). Ac oss all adjus men models, child ADHD symp om sco es and p opo ion o child en wi h clinically signi ican symp oms we e he highes i he mo he epo ed dep essi e symp oms abo e he clinical cu - o bo h du ing and a e p egnancy (Fig 2). Media ion o ma e nal dep essi e symp oms du ing p egnancy on child ADHD symp oms ia ma e nal dep essi e symp oms a e p egnancy Fig 3 shows, ha he e ec o ma e nal dep essi e symp oms du ing p egnancy on child ADHD symp oms was pa ially media ed ia ma e nal dep essi e symp oms a e p egnancy. The model also shows ha ma e nal dep essi e symp oms du ing p egnancy s ill had a di ec and signi ican e ec on child ADHD symp oms a e adjus ing o dep essi e symp oms a e p egnancy. Toge he , dep essi e symp oms du ing and a e p egnancy accoun ed o 11.4% o he a ia ion o he child’s ADHD symp oms. Table 1. Associa ion be ween ma e nal dep essi e symp oms du ing p egnancy and child beha iou al symp oms o a en ion-de ici /hype ac i - i y diso de on he Conne s’ Hype ac i i y Index a age 3.5 yea s. Ma e nal Cen e o Epidemiological S udies Dep ession Scale imes e -weigh ed mean sco e du ing p egnancy Child’s Conne s’ Hype ac i i y Index Sum sco e Child’s Conne s’ Hype ac i i y Index Sum sco e10 B (95%CI) pOR (95%CI) p Model 1 0.26 (0.22, 0.31) <0.001 1.83 (1.62, 2.08) <0.001 Model 2 0.25 (0.21, 0.30) <0.001 1.85 (1.62, 2.11) <0.001 Model 3 0.26 (0.21, 0.30) <0.001 1.86 (1.63, 2.12) <0.001 Model 4 0.24 (0.20, 0.29) <0.001 1.77 (1.55, 2.03) <0.001 Model 5 0.15 (0.10, 0.20) <0.001 1.49 (1.29, 1.73) <0.001 B indica es he S anda d De ia ion (SD) inc ease in child ADHD symp oms when ma e nal dep essi e symp oms inc ease by 1 SD. OR (Odds Ra io) indica es he isk o clinically signi ican ADHD symp oms pe 1 SD uni inc ease in ma e nal dep essi e symp oms du ing p egnancy. Model 1: adjus ed o child sex and age a ollow-up Model 2: adjus ed o model 1 + ma e nal age a childbi h, pa i y, amily s uc u e, educa ion le el, ype 1 diabe es, p e-p egnancy/ch onic hype ension, his o y o physician-diagnosed dep ession, an idep essan and o he psycho opic medica ion use, alcohol use and smoking du ing p egnancy, and ges a ion leng h and in an ’s bi hweigh adjus ed o sex and ges a ion leng h Model 3: adjus ed o model 2 + ma e nal p e-p egnancy obesi y, ges a ional diabe es, ges a ional hype ension and p e-eclampsia Model 4: adjus ed o model 3 + ma e nal ADHD p oblems Model 5: adjus ed o model 4 + ma e nal dep essi e symp oms a e p egnancy h ps://doi.o g/10.1371/jou nal.pone.0190248. 001 Ma e nal dep essi e symp oms du ing p egnancy and child a en ion-de ici /hype ac i i y diso de symp oms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0190248 Decembe 21, 2017 7 / 13 Fig 2. Es ima ed ma ginal means (Panel A) o he child’s beha iou al symp oms o a en ion-de ici /hype ac i i y diso de on he Conne s’ Hype ac i i y Index (CHI) and p opo ion o child en wi h sco es abo e he clinical cu o (10) in he CHI (Panel B) acco ding o he ma e nal Cen e o Epidemiological S udies Dep ession Scale (CES-D) imes e -weigh ed mean sco e (16) du ing p egnancy and Beck Dep ession In en o y-II (BDI-II) sum sco e (14) a e p egnancy abo e and below he clinical cu o s. E o ba s e e o 95% Con idence In e als (95% CI), and numbe s o mean di e ences (MD) (Panel A) and odds a ios (OR) (Panel B) and hei 95% CIs in model 1, and p- alues o models 1–4. Women who sco ed below he clinical cu o in bo h he CES-D du ing p egnancy and in he BDI-II a e p egnancy we e used as he compa ison g oup. Fo di e en adjus men models, please see oo no e in Table 1. h ps://doi.o g/10.1371/jou nal.pone.0190248.g002 Fig 3. Media ion model o he e ec s o ma e nal dep essi e symp oms du ing p egnancy on child’s beha iou al symp oms o a en ion-de ici /hype ac i i y diso de on he Conne s’ Hype ac i i y Index ia ma e nal dep essi e symp oms a e p egnancy. Numbe s e e o uns anda dized eg ession coe icien s (B) and hei 95% Con idence In e als om models adjus ed o child’s age and sex, and o he p opo ion (R 2 ) ma e nal dep essi e symp oms du ing and a e p egnancy explain o he child’s ADHD symp oms. h ps://doi.o g/10.1371/jou nal.pone.0190248.g003 Ma e nal dep essi e symp oms du ing p egnancy and child a en ion-de ici /hype ac i i y diso de symp oms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0190248 Decembe 21, 2017 8 / 13 Sensi i i y analyses S4 Table shows ha ma e nal p e-p egnancy obesi y and p egnancy diso de s, child’s sex, ma e nal his o y o physician-diagnosed dep ession, o ma e nal ADHD p oblems had no e ec s on he indings: ac oss all g oups, he associa ions be ween ma e nal dep essi e symp- oms du ing p egnancy and child ADHD symp oms we e signi ican . Discussion Ou s udy showed, i s , ha ma e nal dep essi e symp oms h oughou p egnancy a e associ- a ed wi h child ADHD symp oms. We also showed ha ma e nal dep essi e symp oms du ing p egnancy we e highly s able, and child en o mo he s wi h consis en ly high dep essi e symp- oms du ing p egnancy showed highe le els o ADHD symp oms a he age o 3 o 6 yea s. These child en also showed a highe p opo ion (o e 32%) and 2.8- imes highe odds o clin- ically signi ican ADHD symp oms. I was he e o e no su p ising ha we ound no ges a- ion-week o imes e -speci ic associa ions be ween ma e nal dep essi e symp oms du ing p egnancy and child ADHD symp oms. None o hese associa ions we e accoun ed o by a numbe o pe ina al, ma e nal and neona al cha ac e is ics, and a se ies o sensi i i y analyses demons a ed ha he associa ions did no ei he a y by ma e nal p e-p egnancy obesi y, hype ensi e p egnancy diso de s, o ges a ional diabe es, child’s sex, ma e nal his o y o phy- sician-diagnosed dep ession, o ma e nal ADHD p oblems. Ou s udy also showed ha highe le els o ma e nal dep essi e symp oms a e p egnancy we e associa ed wi h highe child ADHD symp om sco es. These highe le els o dep essi e symp oms a e p egnancy only pa ially accoun ed o he p ena al e ec s as ma e nal dep es- si e symp oms du ing p egnancy also had a signi ican di ec e ec on he child’s ADHD symp oms when adjus ing o he symp oms a e p egnancy. They did, howe e , add o he p ena al e ec s, such ha child ADHD symp om sco es and he p opo ion and odds o chil- d en wi h clinically signi ican ADHD symp oms we e he highes among hose women wi h clinically signi ican dep essi e symp oms bo h du ing and a e p egnancy. Toge he ma e nal dep essi e symp oms du ing and a e p egnancy accoun ed o 11% o he a ia ion in he child’s ADHD symp oms. Ou indings co espond wi h he De elopmen al O igins o Heal h and Disease (DOHaD) amewo k sugges ing ha p ena al exposu e o en i onmen al ad e si y may ca y endu ing e ec s on b ain de elopmen al sequelae, including isk o ADHD symp oma ology [12,13,34]. Ou indings a e also in alignmen wi h ou own ecen s udy on psychia ic beha - iou p oblems [16], and he o he wo p e ious p ospec i e s udies based on he ALSPAC and he Gene a ion-R coho s on child a en ion and hype ac i i y p oblems [19,20] showing ha ma e nal dep essi e symp oms du ing p egnancy a e associa ed wi h child ADHD symp oms. Also in alignmen wi h he ALSPAC s udy, ou s udy showed ha ma e nal dep essi e symp- oms a e p egnancy did no en i ely accoun o he e ec s o ma e nal dep essi e symp oms du ing p egnancy on child’s ADHD symp oms. In con as , in he Gene a ion-R s udy ma e - nal dep essi e symp oms a e p egnancy ende ed he p ena al e ec s on child a en ion p ob- lems non-signi ican [19]. While ou s udy is o ou knowledge he i s o o mally es o media ion, bo h he ALSPAC and Gene a ion-R indings poin o media ion: he ALSPAC indings poin o pa ial and he Gene a ion-R indings poin o ull media ion o he p ena al dep ession e ec s ia he dep ession e ec s a e p egnancy. An ob ious s udy limi a ion is ha we a e no able o speci y he b ain s uc u al o unc- ional no biological o beha iou al unde lying mechanisms. Exis ing li e a u e sugges s ha highe ma e nal dep essi e symp oms and/o sali a y co isol le els du ing p egnancy a e linked wi h al e ed o sp ing b ain s uc u e and unc ional connec i i y [35], and wi h co ical Ma e nal dep essi e symp oms du ing p egnancy and child a en ion-de ici /hype ac i i y diso de symp oms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0190248 Decembe 21, 2017 9 / 13