Human Growth Assignment Help

Is developmental timing of trauma exposure associated with depressive and post-traumatic stress disorder symptoms in adulthood?

Erin C. Dunn a,b,c,*, Kristen Nishimi a,d, Abigail Powers e, Bekh Bradley e,f aPsychiatric and Neurodevelopmental Genetics Unit, Massachusetts General Hospital, 185 Cambridge St, Simches Research Building, Boston, MA 02114, USAbDepartment of Psychiatry, Harvard Medical School, 401 Park Drive, 2 West, Room 305, Boston, MA 02215, USAcStanley Center for Psychiatric Research, The Broad Institute of Harvard and MIT, The Ted and Vada Stanley Building, 75 Ames Street, Cambridge, MA 02142, USA dHarvard T.H. Chan School of Public Health, 677 Huntington Avenue, Boston, MA 02115, USAeDepartment of Psychiatry and Behavioral Sciences, Emory University School of Medicine, 201 Dowman Drive, Atlanta, GA 30322, USAfAtlanta VA Medical Center, 1670 Clairmont Road, Decatur, GA 30033, USA article info Article history:

Received 27 March 2016 Received in revised form 18 August 2016 Accepted 1 September 2016 Keywords:

Sensitive periods Depression Post-traumatic stress disorder Child maltreatment Trauma abstract Background:Trauma exposure is a known risk factor for psychopathology. However, the impact of the developmental timing of exposure remains unclear. This study examined the effect of age atﬁrst trauma exposure on levels of adult depressive and posttraumatic stress disorder (PTSD) symptoms.

Methods:Lifetime trauma exposure (including age atﬁrst exposure and frequency), current psychiatric symptoms, and sociodemographic information were collected during interviews with adults partici- pating in a study at a public urban hospital in Atlanta, GA. Multiple linear regression models assessed the association between timing ofﬁrst trauma exposure, classiﬁed as early childhood (ages 0e5), middle childhood (ages 6e10), adolescence (ages 11e18), and adulthood (ages 19þ), on adult psychopathology in 2892 individuals.

Results:Participants exposed to trauma (i.e., child maltreatment, other interpersonal violence, non- interpersonal violence, and other events) at any age had higher depressive and PTSD symptoms compared to their unexposed peers. However, participantsﬁrst exposed to child maltreatment during early childhood had depression and PTSD symptoms that were about twice as high as those exposed during later developmental stages. This association was detected even after controlling for sociodemo- graphic characteristics, exposure to other trauma types, and frequency of exposure. Participantsﬁrst exposed during middle childhood to other interpersonal violence also had depressive symptoms scores that were about twice as high as thoseﬁrst exposed during adulthood.

Conclusions:Trauma exposure at different ages may differentially impact depressive and PTSD symp- toms in adulthood. More detailed examination of timing of trauma exposure is warranted to aid in identifying sensitive periods in development.

©2016 Published by Elsevier Ltd. 1. Introduction Exposure to traumatic life events, including child maltreatment or natural disasters, is increasingly recognized as one of the major social determinants of psychiatric disorders. Trauma exposure hasbeen shown to about double the risk for major depressive disorder (McLaughlin et al., 2010a; Norman et al., 2012; Widom et al., 2007), which is currently estimated to affect 11.7% of adolescents (Merikangas et al., 2011) and 16.6% of adults (Kessler et al., 2005).

Trauma exposure (or more precisely, exposure to events involving *Corresponding author. Psychiatric and Neurodevelopmental Genetics Unit, Center for Human Genetic Research, Massachusetts General Hospital, 185Cambridge Street, Simches Research Building 6th Floor (room 6.252), Boston, MA 02114, USA.

E-mail addresses:[email protected](E.C. Dunn),[email protected](K. Nishimi),[email protected](A. Powers),[email protected] (B. Bradley).

URL:http://www.thedunnlab.com Contents lists available atScienceDirect Journal of Psychiatric Research journal homepage:www.elsevier.com/locate/psychires http://dx.doi.org/10.1016/j.jpsychires.2016.09.004 0022-3956/©2016 Published by Elsevier Ltd. Journal of Psychiatric Research 84 (2017) 119e127 perceived or threatened loss of life, serious injury, or loss of physical integrity), is also a requirement to meet DSM-IV criteria for post- traumatic stress disorder (PTSD) (American Psychiatric Association, 2013), which has a lifetime prevalence of 4.7% among adolescents (McLaughlin et al., 2013) and 7.8% among adults (Kessler et al., 1995). As trauma exposure is common in the popu- lation, with six out of every 10 children (McLaughlin et al., 2013) and one out of two adults in the United States reporting a lifetime trauma exposure (Kessler et al., 1995), greater insight into the role of trauma exposure on both depression and PTSD etiology is needed.

One understudied facet of trauma exposure that may be linked to future psychopathology risk is the developmental period of trauma occurrence. Researchers studying child abuse and neglect, in particular, sometimes consider developmental timing as an important dimension of deﬁning maltreatment,ﬁnding that age at onset to maltreatment may inﬂuence the etiology of mental health problems (Barnett et al., 1993; English et al., 2005). However, few attempts have been made in the broader trauma literature to examine ways in which age at onset to trauma exposure correlates with risk for psychopathology. As a result, we currently lack knowledge about the existence of“sensitive periods”(Knudsen, 2004; Bailey et al., 2001; Bornstein, 1989) for psychopathology risk, meaning windows of time in the course of development when the brain is especially“plastic”and therefore when trauma expo- sure substantially elevates risk for the onset of depression or PTSD.

Determining whether, and when, sensitive periods exist across the lifespan will be important for understanding developmentally- relevant biological pathways implicated in the etiology of psycho- pathology and guiding the investment of limited public health re- sources to the“high-risk”stages when deleterious exposures are most harmful and the“high-reward”stages when enriching ex- posures and interventions could offer their greatest beneﬁt.

Thus far, only a small number of studies have examined possible sensitive periods corresponding to risk for either depression or PTSD. Among these studies, no consensus has emerged regarding whether earlier or later exposure is predictive of elevated risk for either outcome. With respect to depression, three prospective studies observed that individuals with maltreatment prior to age 5 had higher levels of teacher-reported internalizing symptoms in early childhood (Keiley et al., 2001) and self-reported depressive symptoms in early (Thornberry et al., 2010) and early to mid- adulthood (Kaplow and Widom, 2007) compared to those who were either never exposed or exposed during later stages. Retro- spective studies have also found earlier abuse - before age 5 (Dunn et al., 2013), before age 12 (Schoedl et al., 2010; Maercker et al., 2004; McCutcheon et al., 2009), or before age 17 (Chu et al., 2013) or trauma - between 4 and 6 (Rusby and Tasker, 2009) and before age 12 (McCutcheon et al., 2009; Maercker et al., 2004 ) particularly elevates risk for depressive symptoms and major depressive disorder. In two of these studies (McCutcheon et al., 2009; Maercker et al., 2004), early exposure to interpersonal trauma, such as witnessing trauma, physical attacks, and sexual molestation, conferred the largest harm relative to other traumas.

However, prospective studies have also found exposure to maltreatment during adolescence (between 10 and 12 (Harpur et al., 2015)or12e17 (Thornberry et al., 2001)) was more strongly associated with adolescent depressive symptoms than earlier maltreatment. A recent retrospective study also found emotional abuse speciﬁcally at age 14 was most predictive of depression during young adulthood (Khan et al., 2015). Three prospective studies (English et al., 2005; Jaffee and Maikovich- Fong, 2011; Oldehinkel et al., 2014) and two retrospective studies (Manly et al., 2001; Pietrek et al., 2013) found no effect of devel- opmental timing of maltreatment in relation to internalizingsymptoms and adolescent or adult depression.

For PTSD symptoms or PTSD diagnoses, similarly mixedﬁndings have been observed. For instance with respect to child abuse, retrospective studies have found that children with PTSD tended to report a lower age atﬁrst exposure (between 3 and 5 (Glod and Teicher, 1996)), and that sexual assault or physical abuse before age 11, but not childhood neglect, conferred the highest risk for PTSD (McCutcheon et al., 2010). Retrospective studies have also found older children (ages 13e18 (Schoedl et al., 2010)) had higher risk for PTSD relative to their peers exposed at other ages. One prospective study of childhood sexual abuse found no association between age at abuse onset and PTSD symptoms (Kaplow and Widom, 2007). Evidence regarding the effects of age at onset to natural disasters appears more consistent, with both retrospective (Demir et al., 2010) and prospective studies (Green et al., 1991; Copeland et al., 2007) observing higher levels of PTSD symptoms or PTSD diagnoses among older children (around age 7 and above) compared to younger children (those younger than about age 7).

However, results for other traumas is less conclusive, with retro- spective studies suggesting there are no differences (Mueller- Pfeiffer et al., 2013; Maercker et al., 2004) in risk for PTSD based on age at onset of trauma, that early-life trauma is more harmful (Ogle et al., 2013, 2015), or that middle childhood (ages 6e11 ) i s more strongly associated with PTSD (Schrag et al., 2008).

Although these studies suggest the developmental timing of trauma exposure may be associated with subsequent risk for depression or PTSD, these studies are limited by a focus on a small subset of adversities, reliance on small clinical or convenience samples, and failure to account for the correlated nature of adver- sities (McLaughlin et al., 2012a). Moreover, relatively few have examined the time-dependent effects of speciﬁc trauma types.

Instead, most prior studies have focused generally on“early life adversity,”meaning adversities occurring over a broad span of ages (typically birth to age 14). In addition, even fewer studies have accounted for the frequency of exposure to adversity, leaving open the possibility earlier trauma exposure may be confounded by the number of times exposed.

The current study aimed to address these limitations by inves- tigating whether developmentally-sensitive measures of trauma exposure were associated with depressive and PTSD symptoms in a sample of highly-trauma exposed adults. We conducted these an- alyses using a low income, urban-sample of African American adults with both high rates of trauma exposure (>90% exposed to at least one traumatic event) and high rates of depressive and PTSD symptoms (Gillespie et al., 2009), which provided an opportunity to examine the differential effects of age at onset to trauma that would not be possible in samples where the prevalence of trauma was lower. More speciﬁcally, we examined the effect of timing of ﬁrst exposure to trauma, coded as:early childhood(age 0e5 years), middle childhood(6e10 years),adolescence(11e18 years), and adulthood(19þyears), on self-reported depressive and PTSD symptoms in adulthood. Traumatic event types were separated into child maltreatment, other interpersonal trauma, non-interpersonal trauma, and other events in order to determine if the effects of developmental timing of exposure differed based on the type of trauma exposure.

2. Methods 2.1. Sample and procedures Data came from the Grady Trauma Project (GTP), an ongoing National Institute of Mental Health (NIMH)-funded study exam- ining genetic and environmental risk and resilience factors for the development of PTSD and other psychiatric disorders (Gillespie E.C. Dunn et al. / Journal of Psychiatric Research 84 (2017) 119e127 120 et al., 2009; Binder et al., 2008; Bradley et al., 2008). The GTP sample consists of 8886 adults (ages 18e90) who were recruited from general medical clinics and obstetric/gynecological clinics at Grady Hospital in Atlanta, Georgia. The clinics are part of a publi- cally funded, non-proﬁt healthcare center serving a primarily Af- rican American, urban population from low socioeconomic backgrounds. The beneﬁts of this particular sample include iden- tiﬁcation of an understudied population, a group with a high rate of trauma, and a relatively homogeneous socioeconomic status dis- tribution. Participants were recruited from clinic waiting areas.

Eligible participants were at least 18 years old, not actively psy- chotic, and able to give formal written and verbal consent. Con- senting individuals participated in verbal interviews administered by trained research assistants lasting approximately 45e75 min, depending on participant's trauma history and symptoms. Partici- pants received $15 for participation. Emory University's Institu- tional Review Board and the Grady Health Care System Research Oversight Committee approved all study procedures.

In the current analysis, we analyzed data from 2892 African American adults whose data were collected between 2005 and 2013 (74.6% female; mean age¼41.0, SD¼13.8). These adults had complete data on all measures relevant to the current analyses. We restricted analyses to African Americans, as signiﬁcant differences were observed in the distribution of trauma exposure, covariates, and both outcomes by race/ethnicity; restriction to one racial/ ethnic group eliminated the variability associated with race, allowing us to more effectively control for confounding. Adjust- ment or stratiﬁcation by race was not possible as individuals from other racial/ethnic groups comprised only 7.7% of the sample (3.6% were White and 4.1% identiﬁed as other), resulting in low power to detect associations due to small cell counts. Of note, our analytic sample was smaller than the larger study sample due to the fact that participants completed study questionnaires in clinic waiting rooms until the clinic was ready to see them/their family member, thus the majority of participants did not complete all measures.

2.2. Measures 2.2.1. Exposure to trauma Presence or absence of trauma exposure, age atﬁrst exposure to trauma, and trauma frequency were ascertained using the Trau- matic Events Inventory (TEI), a 14-item screening measure assess- ing lifetime history of traumatic events (Gillespie et al., 2009; Schwartz et al., 2005, 2006). We focused on 11 different trau- matic events, which had information about age atﬁrst onset and could plausibly occur in multiple developmental stages. These events were grouped into four types, consistent with prior research (Breslau et al., 1998; McLaughlin et al., 2012b): (1)child maltreat- ment(i.e., witnessing violence between parents or caregivers; being beaten; experiencing emotional abuse; or experiencing sexual abuse); (2)other interpersonal violence(i.e., witnessing or being confronted with a friend or family member being murdered; wit- nessing a family member or friend being attacked with or without a weapon; witnessing a non-family member or friend attacked with or without a weapon); (3)non-interpersonal violence(i.e., experi- encing a natural disaster; witnessing or experiencing a serious accident or injury; experiencing a sudden life threatening illness); and (4)other trauma(i.e., any other event or experience not covered by the previously stated categories that participants self-identiﬁed as a traumatic experience, including witnessing a death or suicide, bereavement, divorce or familial disruption, extended caregiving, job loss, etc).

For each traumatic event, participants reported their age (in year) of theirﬁrst experience. We used this data to develop age categories for age atﬁrst exposure:early childhood(age 0e5 years),middle childhood(6e10 years),adolescence(11e18 years), and adulthood(19þyears). These age categories were chosen to match previous research and minimize recall bias (relative to studying speciﬁc years of age) (Kaplow and Widom, 2007; Dunn et al., 2013).

Participants also reported the frequency of each trauma event occurrence on a scale ranging from 0 (unexposed) to 8 (greater than 20 times). Using this data, we generated a set of indicator variables (one for each traumatic event) to denote low versus high frequency of trauma exposure, with high being at or above the top quartile of frequency for a speciﬁc event. Details regarding cut-points used for each trauma event are denoted in the footnote ofFig. 1. These fre- quency indicators were used in models examining timing of exposure to account for the possibility that people exposed at younger ages would be more likely have a higher number of oc- currences of a given trauma exposure.

2.2.2. Depressive symptoms Depressive symptoms were measured using the Beck Depres- sion Inventory, Second Edition (BDI-II), a 21-item psychometrically validated and widely-used inventory of current depressive symp- toms (Beck et al., 1996; Wang and Gorenstein, 2013; Arnau et al., 2001; Beck et al., 1988). The BDI contains items assessing the presence and severity of depressive symptoms over the past two weeks rated on a scale of 0 (not at all/never) to 3 (extremely/every day). Total BDI score were calculated by averaging all individual items (where at least 19 items were completed) and multiplying that mean by 21; this approach to deriving a summary score enabled us to incorporate data from individuals with small amounts of missing data. In this sample, the BDI demonstrated excellent internal consistency reliability ( a¼0.93).

2.2.3. Post-traumatic stress disorder symptoms Post-traumatic stress disorder symptoms were captured using the modiﬁed Posttraumatic Symptom Scale (MPSS), a psychomet- rically validated self-report measure of the frequency and severity of PTSD symptoms (Coffey et al., 1998). The 17 items on the MPSS correspond to symptom criteria to diagnose PTSD as deﬁned by the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edi- tion (DSM-IV-TR) (Association. AP, 2000). These symptoms encompassed re-experiencing, avoidance, emotional numbing, as well as hyperarousal and reactivity categories. Participants were asked about these symptoms generally, and therefore the symp- toms do not reference a speciﬁc index trauma but can reﬂect symptoms related to any of the traumas reported on the TEI. All items were scored on a scale of frequency from 0 (not at all) to 3 (5 or more times a week) as experienced by the individual in the past two weeks from their point of assessment. MPSS total scores were calculated by averaging all individual items (where at least 15 values were completed) and multiplying the average by 17. In this sample, the MPSS also showed excellent internal consistency reli- ability ( a¼0.92). Of note, the PTSD scale was not administered to individuals that did not report exposure to at least one type of trauma.

2.2.4. Covariates All linear regression models adjusted for the following cova- riates: sex; age (continuous); highest level of education (less than 12th grade; high school graduate or GED; greater than high school graduate or GED/college graduate); household monthly income ($0e499; $500e999; $1000þ), and employment status (unem- ployed; unemployed receiving disability support; and employed, with or without disability support). E.C. Dunn et al. / Journal of Psychiatric Research 84 (2017) 119e127121 2.3. Data analyses To facilitate interpretation and comparison, both outcomes were standardized (mean¼0; SD¼1) prior to analyses. After stan- dardization, we conducted basic univariate and bivariate analyses to compare outcome values by each covariate. We also examined the distribution of exposure to each traumatic event in the total sample and by age atﬁrst exposure among those who were exposed. To determine whether respondents exposed at earlier ages also had more frequent trauma exposure, we examined the percent of respondents exposed to frequent trauma by reported age atﬁrst exposure. Following these analyses, weﬁt a series of linear regression models that examined, separately for each traumatic event and type, the association between trauma exposure and each outcome, after adjusting for covariates. Model 1 examined the ef- fect of being exposed (vs. non-exposed) at any age to trauma. Model 2 (partial adjustment) examined the effect of timing of exposure (compared to the referent group of non-exposed during any period), after adjusting for covariates as well as exposure to all other traumatic events, given that some exposures, particularly child maltreatment events, were highly correlated (tetrachoric correlation values ranged from r¼0.07 to r¼0.63). Model 3 (full adjustment) expanded upon Model 2 by additionally including the indicator for frequency of each trauma event occurrence (0¼low frequency; 1¼high reported number of occurrences of that trauma exposure). For Models 2 and 3, we conducted a test of homogeneity to evaluate whether the beta coefﬁcients (indicating the effect of timing of exposure relative to never exposed) were signiﬁcantlydifferent from each other. In cases where the null hypothesis was rejected for the test for homogeneity (p<0.05), indicating that the effects of developmental timing differed across groups, we then conducted post-hoc Tukey tests to evaluate, after adjustment for multiple testing, which speciﬁc age at exposure groups differed from one another. All analyses were conducted using SAS Version 9.4 (SAS Institute, Inc, Cary, North Carolina).

3. Results The analytic sample comprised mostly women (74.6%) and middle-age adults (mean age¼41.0; SD¼13.8). Depressive and PTSD symptoms signiﬁcantly differed by age, education level, household monthly income, and employment, but not sex (Table 1).

Speciﬁcally, both depressive symptoms and PTSD symptoms were higher among middle-age adults, those with lower education and income, and those who were unemployed or on disability.

3.1. Distribution of trauma exposure and trauma timing Nearly three-quarters of the sample reported being exposed to some type of interpersonal or non-interpersonal violence event (Table 2). Slightly more than half of the sample had been exposed to some type of child maltreatment, with violence between caregivers and sexual abuse being the two most common sub-types.

The mean age atﬁrst exposure to any trauma was 11.4 (SD¼8.8), but ranged between before one years of age through age 62. Overall, middle childhood (ages 6e10) was the most often 0.0 20.0 40.0 60.0 80.0 100.0 Early ChildhoodMiddle ChildhoodAdolescence % of those exposed with high exposure frequency Timing of First Exposure Child Maltreatment Violence between Caregivers Beaten Emotional Abuse Sexual Abuse Any of the Above 0.0 20.0 40.0 60.0 80.0 100.0 Early ChildhoodMiddle ChildhoodAdolescence Adulthood % of those exposed with high exposure frequency Timing of First Exposure Other Interpersonal Violence Family/friend Murdered Witness Attack (Family or friend) Witness Attack (non- family/friend ) Any of the Above 0.0 20.0 40.0 60.0 80.0 100.0 Early ChildhoodMiddle ChildhoodAdolescence Adulthood % of those exposed with high exposure frequency Timing of First Exposure Non-Interpersonal Violence Natural Disaster Serious Accident or Injury Life Threatening Illness NS Any of the Above 0.0 20.0 40.0 60.0 80.0 100.0 Early ChildhoodMiddle ChildhoodAdolescence Adulthood % of those exposed with high exposure frequency Timing of First Exposure Any Other Trauma Any other trauma not yet covered NS Fig. 1.Percent of respondents exposed to frequent trauma by age atﬁrst onset to trauma. Frequency of each trauma event occurrence was ascertained on a scale ranging from 0 (unexposed) to 8 (greater than 20 times). Using this data, we generated indicator variables for each traumatic event denoting low versus high frequency of trauma exposure, with high being at or above the top quartile of frequency for a speciﬁc event. High frequency was deﬁned as greater than or equal to the following values for each individual trauma event: experiencing a natural disaster 2; witnessing or experiencing a serious accident or injury 2; and experiencing a sudden life threatening illness 2; witnessing or being confronted with a friend or family member being murdered 2; witnessing a family member or friend being attacked with or without a weapon 4; witnessing a non-family member or friend attacked with or without a weapon 5; witnessing violence between parents or caregivers 8; being beaten 8; experiencing emotional abuse 8; or expe- riencing sexual abuse 5. For the trauma categories, the 75th percentile for reported frequency of exposure was taken for all events included in the category and high frequency was deﬁned as greater than or equal to the following values for trauma categories: child maltreatment 7, interpersonal trauma 3, non-interpersonal trauma 2; and any other trauma not reported 2. Theﬁgure presents the % of those exposed to frequent trauma within each age category among those exposed. All Chi Square Goodness of Fit models for each trauma event, which evaluated whether there were signiﬁcant differences between the frequencies of exposure to each trauma by age atﬁrst exposure, are signiﬁcant (p<0.0001) unless indicated with NS (p>0.05).E.C. Dunn et al. / Journal of Psychiatric Research 84 (2017) 119e127 122 reported time period forﬁrst exposure to child maltreatment (Table 2). In contrast, adolescence (ages 11e18) and adulthood (ages 19 and above) were the most often reported time periods forﬁrst exposure to other interpersonal violence and non-interpersonal violence, respectively.

For most traumas, there was also an age-frequency gradient, suggesting that thoseﬁrst exposed earlier in the lifespan also tended to report experiencing more frequent exposure (Fig. 1).

3.2. Trauma exposure and depressive and PTSD symptoms As shown inTable 3, respondents reporting exposure to trauma at any age had higher depressive and PTSD symptoms, on average, relative to their non-exposed peers. For example, participants exposed to child maltreatment had, on average, depressive and PTSD symptoms scores that were half of a standard-deviation unit greater than their unexposed peers.3.3. Timing of trauma exposure and depressive symptoms Depressive symptoms varied as a function age atﬁrst exposure to child maltreatment and other interpersonal violence, but not non-interpersonal violence or any other trauma (Table 4). Specif- ically, participantsﬁrst exposed to child maltreatment during early childhood ( b¼0.739) had depressive symptoms scores that were about 1.5 times as high as thoseﬁrst exposed during middle childhood ( b¼0.519) and almost twice as high as thoseﬁrst exposed during adolescence ( b¼0.397), even after adjusting for sociodemographic covariates and exposure to other trauma types (Tukey p-value<0.05 for both comparisons). In other words, the predicted depressive symptom score for thoseﬁrst exposed during early childhood would be 0.470 (SD¼0.27), whereas the predicted depressive symptom score for thoseﬁrst exposed during middle childhood would be 0.245 (SD¼0.27). These differences persisted after additionally adjusting for frequency of trauma exposure Table 1 Distribution of depressive and posttraumatic stress disorder symptoms in the Grady Trauma Project (GTP) analytic sample (N¼2892).

Covariate Total sample Depressive symptoms PTSD symptoms N (%) Mean (SD) F Valuep valueMean (SD) F Valuep value Age 18e25 579 (20.0) 11.78 (9.7) 21.17<0.0001 10.23 (10.9) 23.45<0.0001 26e35 527 (18.2) 14.59 (12.0) 12.39 (11.9) 36e45 499 (17.3) 16.26 (13.3) 14.28 (13.6) 46e55 817 (28.3) 16.24 (12.8) 14.60 (13.0) 56þ470 (16.2) 11.63 (11.1) 9.00 (10.5) Sex Male 733 (25.4) 14.03 (12.4) 0.50 0.500 12.14 (12.7) 0.30 0.585 Female 2159 (74.6) 14.39 (12.0) 12.43 (12.2) Education Less than 12th grade 658 (22.7) 16.89 (13.4) 20.35<0.0001 14.39 (13.2) 11.69<0.0001 High school graduate or GED 1202 (41.6) 13.79 (11.8) 11.68 (12.0) Greater than GED/College graduate 1023 (35.7) 13.24 (11.3) 11.84 (12.1) Income $0e499 879 (30.4) 16.88 (13.5) 36.18<0.0001 13.83 (13.2) 14.29<0.0001 $500e999 779 (26.9) 14.40 (11.8) 12.83 (12.5) $1000 or more 1234 (42.7) 12.40 (10.8) 11.01 (11.5) Employment Status Unemployed 1468 (50.8) 15.34 (12.7) 31.06<0.0001 12.99 (12.7) 17.57<0.0001 Unemployed (disability) 549 (19.0) 15.75 (12.5) 13.84 (13.0) Employed (with or without disability) 875 (30.2) 11.65 (10.2) 10.36 (11.0) Descriptive statistics are presented for the analytic sample. Linear regressions were performed for each outcome by covariate, with F-statistics and corresponding p-values listed. Analyses were performed on the non-standardized outcomes to facilitate interpretation.

Table 2 Distribution of exposure to each traumatic event in total sample and by age atﬁrst exposure to trauma among those who were exposed.

Trauma exposureExposed Age atﬁrst exposure (year)Age atﬁrst exposure (category) Ages 0-5 Ages 6-10 Ages 11-18 Ages 19þ %N MeanSD%N%N%N%N Child Maltreatment Violence between Caregivers 31.0 897 8.06 3.3 8.1 235 16.2 469 6.7 193ee Physical Abuse 19.2 555 8.10 3.3 4.6 133 10.8 311 3.8 111ee Emotional Abuse 25.9 750 9.91 3.8 3.9 112 11.2 325 10.8 313ee Sexual Abuse 30.4 878 9.70 3.7 4.4 128 13.4 387 12.6 363ee Any of the Above 56.9 1624 8.07 3.5 15.5 449 28.0 809 12.7 366ee Other Interpersonal Violence Family/friend Murdered 53.9 1560 22.19 12.0 0.9 26 6.4 186 19.5 564 27.1 784 Witness Attack (Family or friend) 42.2 1220 17.37 10.3 2.7 77 8.7 252 16.5 476 14.4 415 Witness Attack (non-family/friend) 44.0 1271 18.99 10.6 1.4 40 7.0 201 18.9 547 16.7 483 Any of the Above 77.4 2238 17.52 10.5 4.3 123 16.5 476 31.1 900 25.6 739 Non-Interpersonal Violence Natural Disaster 25.7 744 19.68 13.4 1.2 35 6.9 200 7.3 210 10.3 299 Serious Accident or Injury 64.5 1865 20.37 11.9 2.8 81 11.6 334 20.0 572 30.4 878 Life Threatening Illness 23.1 669 34.59 15.5 1.0 30 0.8 24 2.3 67 19.0 548 Any of the Above74.6 2157 18.91 12.2 4.8 140 17.1 495 22.5 651 30.1 871 Any other trauma 31.5 911 27.73 14.2 0.9 26 2.0 59 6.5 188 22.1 638 E.C. Dunn et al. / Journal of Psychiatric Research 84 (2017) 119e127123 (Model 3 results) and appeared driven primarily by exposure to sexual abuse (Supplemental Table 1).

Similarly, after adjusting for sociodemographic characteristics and frequency of trauma exposure, participantsﬁrst exposed to other interpersonal violence during middle childhood ( b¼0.334) had depressive symptoms scores that were about twice as large as thoseﬁrst exposed during adulthood ( b ¼0.169; Tukey p- value<0.05). These results did not appear driven by exposure to a speciﬁc trauma type (Supplemental Table 1).

3.4. Timing of trauma exposure and PTSD symptoms PTSD symptoms also varied as a function age atﬁrst exposure to child maltreatment and to some extent other interpersonal violence, but not non-interpersonal violence or any other trauma (Table 5). After adjusting for all covariates, those exposed to childmaltreatment during early childhood ( b¼0.681) had signiﬁcantly higher levels of PTSD symptoms relative to thoseﬁrst exposed during middle childhood ( b¼0.468) or adolescence ( b¼0.342; Tukey p-value<0.05 for both comparisons). These differences were explained by exposure to child sexual abuse (Supplemental Table 2).

Although exposure to other interpersonal violence appeared initially most damaging in increasing levels of PTSD symptoms (Table 5), no signiﬁcant differences were found across those exposed at different ages (relative to those never exposed) after adjusting for frequency of exposure (Tukey p-value<0.05).

4. Discussion This study examined the association between developmental timing of exposure to trauma and levels of depression and PTSD Table 3 Results of linear regression analyses examining trauma exposure (exposed vs. unexposed) on depressive and PTSD symptoms, adjusting for covariates.

N (%) Depressive symptoms PTSD symptoms Beta 95% CI Beta 95% CI Child Maltreatment 1624 (56.9) 0.549 (0.48, 0.62) 0.575 (0.51, 0.64) Other Interpersonal Violence 2238 (77.4) 0.211 (0.13, 0.29) 0.297 (0.21, 0.38) Non-Interpersonal Violence 2157 (74.6) 0.146 (0.06, 0.23) 0.256 (0.18, 0.34) Any trauma not yet covered 911 (31.5) 0.152 (0.08, 0.22) 0.260 (0.19, 0.33) Cell entries are sample size (n and %), beta coefﬁcients, and 95% conﬁdence intervals (CI). The table presents results from Model 1 examining exposure to trauma (coded as 0¼never exposed; 1¼exposed) on depressive and PTSD symptoms. Both outcomes were standardized to z-scores. Each trauma type was examined separately, thus the table includes results of eight separate multiple regression models (two for each exposure). All models controlled for age, sex, education, income, employment status, and exposure to any other traumatic event type.

Table 4 Results of linear regression analyses examining the effect of timing ofﬁrst exposure to trauma on depressive symptoms, adjusting for covariates and frequency of exposure.

N (%) Model 2 (partial adjustment) Model 3 (full adjustment) Beta 95% CI F-statistic (p-value) Beta 95% CI F-statistic (p-value) Child Maltreatment15.59 (<0.0001) 10.44 (<0.0001) Early Childhood 449 (15.5) 0.739 a,b (0.64, 0.84) 0.633 a,b (0.52, 0.74) Middle Childhood 809 (28.0) 0.519 (0.44, 0.60) 0.444 (0.36, 0.53) Adolescence 366 (12.7) 0.397 (0.29, 0.51) 0.356 (0.25, 0.46) Other Interpersonal Violence4.28 (0.005) 3.84 (0.009) Early Childhood 123 (4.3) 0.304 (0.12, 0.49) 0.291 (0.10, 0.48) Middle Childhood 476 (16.5) 0.344 c,d (0.23, 0.46) 0.334 d (0.21, 0.46) Adolescence 900 (31.1) 0.177 (0.08, 0.27) 0.170 (0.07, 0.27) Adulthood 739 (25.6) 0.172 (0.07, 0.27) 0.169 (0.07, 0.27) Non-Interpersonal Violence0.95 (0.414) 0.50 (0.684) Early Childhood 140 (4.8) 0.099 ( 0.07, 0.27) 0.009 ( 0.18, 0.17) Middle Childhood 495 (17.1) 0.179 (0.07, 0.29) 0.065 ( 0.05, 0.18) Adolescence 651 (22.5) 0.172 (0.07, 0.27) 0.077 ( 0.03, 0.18) Adulthood 871 (30.1) 0.113 (0.02, 0.21) 0.041 ( 0.06, 0.14) Any trauma not yet covered1.26 (0.287) 1.22 (0.302) Early Childhood 26 (0.9) 0.195 ( 0.16, 0.55) 0.210 ( 0.15, 0.57) Middle Childhood 59 (2.04) 0.153 ( 0.09, 0.39) 0.167 ( 0.08, 0.41) Adolescence 188 (6.5) 0.056 ( 0.08, 0.19) 0.070 ( 0.08, 0.22) Adulthood 638 (22.1) 0.210 (0.12, 0.29) 0.221 (0.13, 0.31) The table presents results from Models 2 and 3, which examined age atﬁrst exposure (early childhood¼age 0e5, middle childhood¼age 6e10, adolescence¼age 11e18, adulthood¼age 19þ) on depressive symptoms (z-score standardized). Model 2 (partial adjustment) controlled for age, sex, education, income, employment status, and exposure to any other traumatic event. Model 3 (full adjustment) controlled for all covariates included in Model 2 plus frequency of each trauma eventoccurrence (0¼low frequency; 1¼high reported number of occurrences of that trauma exposure). Each trauma type was examined separately, thus the table includes results of eight separate multiple regression models (two for each exposure). Beta coefﬁcients and 95% conﬁdence intervals (CI) indicate effects for developmental timing of exposure relative to the reference of never exposed in any developmental period. Effects are signiﬁcant at p<0.05 when CI does not include 0. Omnibus tests (F-tests and p-values) indicate the overall main effect of age atﬁrst exposure categories; these test the null hypothesis that the beta coefﬁcients were equivalent across all groups (i.e., early childhood b¼middle childhood b¼adolescence b). aRefers to a signiﬁcant difference (p<0.05) between early childhood versus middle childhood.bRefers to a signiﬁcant difference (p<0.05) between early childhood versus adolescence.cRefers to a signiﬁcant difference (p<0.05) between middle childhood versus adolescence.dRefers to a signiﬁcant difference (p<0.05) between middle childhood versus adulthood. E.C. Dunn et al. / Journal of Psychiatric Research 84 (2017) 119e127 124 symptoms within a sample of highly-trauma exposed adults. By studying a racially and socioeconomically homogeneous sample, where trauma exposure was common, we were able to examine the effects of age atﬁrst exposure to trauma. Such analyses may not have been possible in a heterogeneous sample where trauma exposure was more rare.

In line with previous research (McLaughlin et al., 2010a, 2012a; Norman et al., 2012; Widom et al., 2007), we found that those who were trauma exposed (regardless of age atﬁrst exposure) had signiﬁcantly higher levels of depressive and PTSD symptoms rela- tive to those who were unexposed to trauma. Theseﬁndings add further support to the well-documentedﬁnding that trauma exposure elevates risk for psychopathology across the lifecourse.

However, a more novel observation from this study was that there appeared to be two developmental stages when trauma exposure was associated with elevated levels of subsequent psy- chopathology. First, participants exposed to child maltreatment during early childhood (ages 0e5) had both depression and PTSD symptoms that were up to twice as high as those exposed during later developmental stages. These effects were detected even after controlling for sociodemographic characteristics, exposure to other types of traumas, and the number of occurrences of child maltreatment (i.e., the frequency of exposure). Suchﬁndings are consistent with several prospective (Keiley et al., 2001; Thornberry et al., 2010; Kaplow and Widom, 2007) and retrospective studies (Dunn et al., 2013; Glod and Teicher, 1996) also showing an elevated risk for subsequent depression, in particular, among thoseﬁrst exposed in theﬁrstﬁve years of life. Although the mechanisms linking early trauma exposure to subsequent psychopathology riskare not well known (McLaughlin, 2016), early trauma exposure may be more damaging than later trauma exposure because it com- promises a child's ability to successfully master stage-salient developmental tasks (e.g., self-regulation, secure attachments) (Cicchetti and Toth, 1995) and damages the foundation of brain architecture and neurobiological systems involved in regulating arousal, emotion, stress responses, and reward processing (McLaughlin et al., 2010b, 2011; Fox et al., 2010; Sheridan et al., 2012), which are all implicated in the onset and persistence of stress-related disorders like depression and PTSD.

Second, we also found that participantsﬁrst exposed during middle childhood (ages 6e10) to other types of interpersonal violence, including having a friend or family member murdered or witnessing a friend or family member being attached with or without a weapon, had depressive symptoms scores that were about twice as high as thoseﬁrst exposed during adulthood. Similar results were also detected for PTSD, but were not statistically sig- niﬁcant after adjusting the frequency or number of occurrences of child exposure. These results are consistent with at least some prior studies suggesting that trauma during middle childhood, including severe illness (Schrag et al., 2008), or natural disasters (Demir et al., 2010; Green et al., 1991; Copeland et al., 2007) is associated with an elevated risk for PTSD relative to exposure in other periods.

Although the mechanisms driving this association are unclear, exposure to interpersonal violence events during middle childhood may be more harmful than exposure in adulthood for a number of reasons: school-age children may be in a unique developmental stages where they do not beneﬁ t as much from parental buffering (Garbarino, 2001), when demands from the social environment Table 5 Results of linear regression analyses examining the effect of timing ofﬁrst exposure to trauma on PTSD symptoms, adjusting for covariates and frequency of exposure.

N (%) Model 2 (partial adjustment) Model 3 (full adjustment) Beta 95% CI F-statistic (p-value) Beta 95% CI F-statistic (p-value) Child Maltreatment21.94 (p<0.0001) 15.16 (<0.0001) Early Childhood 449 (15.5) 0.792 a,b (0.69, 0.89) 0.681 a,b (0.57, 0.79) Middle Childhood 809 (28.0) 0.547 c (0.47, 0.63) 0.468 (0.38, 0.55) Adolescence 366 (12.7) 0.386 (0.28, 0.49) 0.342 (0.24, 0.45) Other Interpersonal Violence2.94 (0.032) 1.67 (0.171) Early Childhood 123 (4.3) 0.315 (0.14, 0.49) 0.229 (0.04, 0.42) Middle Childhood 476 (16.5) 0.410 d (0.30, 0.52) 0.345 (0.23, 0.46) Adolescence 900 (31.1) 0.288 (0.19, 0.38) 0.241 (0.14, 0.34) Adulthood 739 (25.6) 0.250 (0.15, 0.35) 0.231 (0.13, 0.33) Non-Interpersonal Violence2.02 (0.109) 1.22 (0.302) Early Childhood 140 (4.8) 0.224 (0.06, 0.39) 0.135 (-0.04, 0.31) Middle Childhood 495 (17.1) 0.305 (0.20, 0.41) 0.211 (0.09, 0.33) Adolescence 651 (22.5) 0.296 (0.20, 0.39) 0.218 (0.11, 0.32) Adulthood 871 (30.1) 0.204 (0.11, 0.30) 0.144 (0.05, 0.24) Any trauma not yet covered1.08 (0.356) 1.12 (0.340) Early Childhood 26 (0.9) 0.363 (0.01, 0.71) 0.341 (-0.01, 0.69) Middle Childhood 59 (2.04) 0.170 (-0.07, 0.41) 0.152 (-0.09, 0.39) Adolescence 188 (6.5) 0.206 (0.07, 0.34) 0.187 (0.04, 0.33) Adulthood 638 (22.1) 0.319 (0.24, 0.40) 0.303 (0.21, 0.39) The table presents results from Models 2 and 3, which examined age atﬁrst exposure (early childhood¼age 0e5, middle childhood¼age 6e10, adolescence¼age 11e18, adulthood¼age 19þ) on PTSD symptoms (z-score standardized). Model 2 (partial adjustment) controlled for age, sex, education, income, employment status, and exposureto any other traumatic event. Model 3 (full adjustment) controlled for all covariates included in Model 2 plus frequency of each trauma event occurrence(0¼low frequency; 1¼high reported number of occurrences of that trauma exposure). Each trauma type was examined separately, thus the table includes results of eight separate multiple regression models (two for each exposure). Beta coefﬁcients and 95% conﬁdence intervals (CI) indicate effects for developmental timing of exposure relative to the reference of never exposed in any developmental period. Effects are signiﬁcant at p<0.05 when the CI does not include 0. Omnibus tests (F-tests and p-values) indicate the overall main effect of age atﬁrst exposure categories; these test the null hypothesis that the beta coefﬁcients were equivalent across all groups (i.e., early childhood b¼middle childhood b¼adolescence b). aRefers to a signiﬁcant difference (p<0.05) between early childhood versus middle childhood.bRefers to a signiﬁcant difference (p<0.05) between early childhood versus adolescence.cRefers to a signiﬁcant difference (p<0.05) between middle childhood versus adolescence.dRefers to a signiﬁcant difference (p<0.05) between middle childhood versus adulthood. E.C. Dunn et al. / Journal of Psychiatric Research 84 (2017) 119e127125 increase (i.e., relating to peers, participating in school activities) (Margolin and Gordis, 2000), and when adaptive coping capacities are still developing (Zimmer-Gembeck and Skinner, 2011).

It is important to emphasize that these developmental timing differences would have been missed had we not considered the effect of timing of trauma exposure. Ourﬁndings therefore un- derscore that a basic comparison of those“exposed”to those who are“unexposed”may potentially mask important within-group differences that are only revealed when examining develop- mental timing of exposure to trauma. Theseﬁndings also empha- size the need to adjust, where possible, for frequency of exposure to trauma, as some developmental timing differences may be atten- uated after considering this information.

Several limitations of the current study must be noted. First, the measure of trauma exposure included in this study did not capture other characteristics of the trauma, including its severity, chro- nicity, or duration. Our frequency indicator variable may have captured some of these domains, though this remains unclear, as the frequency measure was not speciﬁc to a given year. In the case of abuse, the relationship of the perpetrator to the victim was also not examined and could have impacted the speciﬁc results for abuse types. Indeed, there is evidence suggesting more negative psychological outcomes among those experiencing abuse perpe- trated by a family member rather than non-family member (Canton-Cortes and Canton, 2010; Ullman, 2007). Moving forward, larger scales studies are needed to examine the extent to which chronicity, duration, and perpetrator of the trauma varies as a function of developmental timing. Second, trauma exposure and age atﬁrst exposure to trauma were assessed retrospectively in adulthood. Retrospective reports of child maltreatment, in partic- ular, have been shown in some cases to be less reliable and valid than prospective reports, because of memory inaccuracies, a reluctance to disclose personal matters, or current mood states (Hardt and Rutter, 2004). However, recent studies have found retrospective and prospective measures produce similar estimates of effect for mental disorders (Scott et al., 2012), suggesting that trauma exposure is harmful regardless of ascertainment strategy and that even if recall bias is present, effect estimates are unlikely altered. Some studies have documented differential recall bias across the lifespan, showing increasing problems in recall and disclosure of early trauma events as age increases (see for example (Yoshihama and Gillespie, 2002)) as well as an association between earlier age at onset of child abuse with greater amnesia in adult memory recovery of those events (Chu et al., 1999). However, ac- curate recall of memories as early as age 2e3 years old has been documented (Chu et al., 1999; Usher and Neisser, 1993), older in- dividuals show no autobiographical memory recall difference for memories from any point in their lives (Howes and Katz, 1992), and underreporting is more likely than falsely positively reporting early abuse (Hardt and Rutter, 2004). Moreover, by using developmental periods, rather than speciﬁc ages, we were able to maintain con- sistency with prior studies and reduce recall bias as compared to reports focused on single ages. Participants unsure of their age at ﬁrst trauma exposure were also excluded. Prospective research is needed to replicate theseﬁndings; this work would ideally incor- porate repeated measures of trauma exposure and mental health in order to differentiate short-versus long-term effects of trauma timing on psychopathology. Finally, our sample was largely comprised of participants who were low income, female, and Af- rican Americans. Our analyses were also restricted to African Americans, due to the small number of respondents in other racial/ ethnic groups. However, as noted above, a number of theﬁndings from this study are consistent with those found in studies using samples with different demographic proﬁles. Further, this study focuses on a largely under-studied population with high levels ofmental health problems; studying this unique population is key to help determine that factors that might impact the development of psychiatric conditions and the developmental stages when in- terventions or supports may be most beneﬁcial.

In conclusion, our study underscores the need to consider the developmental period of trauma exposure, as the effect of some traumas varied as a function of when in the course of development the traumaﬁrst occurred. Identiﬁcation of these developmental stages of heightened vulnerability will aid in determining sensitive periods and guiding the investment of limited public health dollars towards the life stages when age-tailored interventions can be delivered and possibly yield their greatest returns.

Conﬂict of interest The authors do not report any conﬂicts of interest.

Contributions All authors contributed substantially to the work presented in this paper. ECD and KN wrote the main paper and analyzed data. AP and BB designed and ran the study and contributed to writing the paper. All authors contributed to conceptualizing the analysis, interpreting the results and editing the paper.

Acknowledgements Research reported in this publication was supported by the National Institute of Mental Health within the National Institutes of Health under Award Numbers K01MH102403 (Dunn) and MH102890 (Powers) and the National Institute of Child Health and Human Development under Award Number HD071982 (Bradley).

The content is solely the responsibility of the authors and does not necessarily represent the ofﬁcial views of the National Institutes of Health. Additionally, the contents of this report do not represent the views of the Department of Veterans Affairs or the United States Government.

Appendix A. Supplementary data Supplementary data related to this article can be found athttp:// dx.doi.org/10.1016/j.jpsychires.2016.09.004.

References American Psychiatric Association, 2013. Diagnostic and Statistical Manual of Mental Disorders,ﬁfth ed. American Psychiatric Publishing, Arlington, VA.

Arnau, R.C., Meagher, M.W., Norris, M.P., Bramson, R., 2001. Psychometric evaluation of the Beck Depression Inventory-II with primary care medical patients. Health psychology : ofﬁcial journal of the Division of Health Psychology. Am. Psychol.

Assoc. 20, 112e11 9.

Diagnostic and Statistical Manual of Mental Disorders (4th Edition, Text Revision), fourth ed., 2000. Association. AP, Washington, DC.

Bailey, D.B., Bruer, J.T., Symons, F.J., Lichtman, J.W. (Eds.), 2001. Critical Thinking about Critical Periods. Paul H. Brookes Publishing Company, Baltimore, MD.

Barnett, D., Manly, J.T., Cicchetti, D., 1993. Deﬁning child maltreatment: the interface between policy and research. In: Cicchetti, D., Toth, S.L. (Eds.), Child Abuse, Child Development, and Social Policy. Ablex, Norwood, NJ, pp. 7e73.

Beck, A.T., Steer, R.A., Carbin, M.G., 1988. Psychometric properties of the Beck depression inventory: twenty-ﬁve years of evaluation. Clin. Psychol. Rev. 8, 77e10 0.

Beck, A.T., Steer, R.A., Brown, G.K., 1996. Manual for the Beck Depression Inventory- II. Psychological Corporation, San Antonio, Texas.

Binder, E.B., Bradley, R.G., Liu, W., et al., 2008. Association of FKBP5 polymorphisms and childhood abuse with risk of posttraumatic stress disorder symptoms in adults. JAMA 299, 1291e1305.

Bornstein, M.H., 1989. Sensitive periods in development: structural characteristics and causal interpretations. Psychol. Bull. 105, 179e197.

Bradley, R.G., Binder, E.B., Epstein, M.P., et al., 2008. Inﬂuence of child abuse on adult depression: moderation by the corticotropin-releasing hormone receptor gene.

Arch. Gen. Psychiatry 65, 190e200. E.C. Dunn et al. / Journal of Psychiatric Research 84 (2017) 119e127 126 Breslau, N., Kessler, R.C., Chilcoat, H.D., Schultz, L.R., Davis, G.C., Andreski, P., 1998.

Trauma and posttraumatic stress disorder in the community: the 1996 detroit area survey of trauma. Arch. Gen. Psychiatry 55, 626e632.

Canton-Cortes, D., Canton, J., 2010. Coping with child sexual abuse among college students and post-traumatic stress disorder: the role of continuity of abuse and relationship with the perpetrator. Child. Abuse Negl. 34, 496e506.

Chu, J.A., Frey, L.M., Ganzel, B.L., Matthews, J.A., 1999. Memories of childhood abuse:

dissociation, amnesia, and corroboration. Am. J. Psychiatry 156, 749e755.

Chu, D.A., Williams, L.M., Harris, A.W., Bryant, R.A., Gatt, J.M., 2013. Early life trauma predicts self-reported levels of depressive and anxiety symptoms in nonclinical community adults: relative contributions of early life stressor types and adult trauma exposure. J. Psychiatr. Res. 47, 23e32.

Cicchetti, D., Toth, S.L., 1995. A developmental psychopathology perspective on child abuse and neglect. J. Am. Acad. Child Adolesc. Psychiatry 34, 541e565.

Coffey, S.F., Dansky, B.S., Falsetti, S.A., Saladin, M.E., Brady, K.T., 1998. Screening for PTSD in a substance abuse sample: psychometric properties of a modiﬁed version of the PTSD Symptom Scale Self-Report. Posttraumatic stress disorder.

J. Trauma Stress 11, 393e399.

Copeland, W.E., Keeler, G., Angold, A., Costello, E.J., 2007. Traumatic events and posttraumatic stress in childhood. Arch. Gen. Psychiatry 64, 577e584.

Demir, T., Demir, D.E., Alkas, L., Copur, M., Dogangun, B., Kayaalp, L., 2010. Some clinical characteristics of children who survived the Marmara earthquakes. Eur.

Child. Adolesc. Psychiatry 19, 125e13 3.

Dunn, E.C., McLaughlin, K.A., Slopen, N., Rosand, J., Smoller, J.W., 2013. Develop- mental timing of child maltreatment and symptoms of depression and suici- dality in young adulthood: results from the National Longitudinal Study of Adolescent Health. Depress. Anxiety 30, 955e964.

English, D.J., Graham, J.C., Litrownik, A.J., Everson, M., Bangdiwala, S.I., 2005.

Deﬁning maltreatment chronicity: are there differences in child outcomes.

Child Abuse Negl. 29, 575e595.

Fox, S.E., Levitt, P., Nelson, C.A., 2010. How the timing and quality of early experi- ences inﬂuence the development of brain architecture. Child. Dev. 81, 28e40.

Garbarino, J., 2001. An ecological perspective on the effects of violence on children.

J. Community Psychol. 29, 361e378.

Gillespie, C.F., Bradley, B., Mercer, K., et al., 2009. Trauma exposure and stress- related disorders in inner city primary care patients. Gen. Hosp. Psychiatry 31, 505e514.

Glod, C.A., Teicher, M.H., 1996. Relationship between early abuse, posttraumatic stress disorder, and activity levels in prepubertal children. J. Am. Acad. Child.

Adolesc. Psychiatry 35, 1384e1393.

Green, B.L., Korol, M., Grace, M.C., et al., 1991. Children and disaster: age, gender, and parental effects on PTSD symptoms. J. Am. Acad. Child. Adolesc. Psychiatry 30, 945e951.

Hardt, J., Rutter, M., 2004. Validity of adult retrospective reports of adverse child- hood experiences: review of the evidence. J. Child Psychol. Psychiatry 45, 260e273.

Harpur, L.J., Polek, E., van Harmelen, A.L., 2015. The role of timing of maltreatment and child intelligence in pathways to low symptoms of depression and anxiety in adolescence. Child. Abuse Negl. 47, 24e37.

Howes, J.L., Katz, A.N., 1992. Remote memory: recalling autobiographical and public events from across the lifespan. Can. J. Psychol. 46, 92e116.

Jaffee, S.R., Maikovich-Fong, A.K., 2011. Effects of chronic maltreatment and maltreatment timing on children's behavior and cognitive abilities. J. Child Psychol. Psychiatry 52, 184e19 4.

Kaplow, J.B., Widom, C.S., 2007. Age of onset of child maltreatment predicts long- term mental health outcomes. J. Abnorm. Psychol. 116, 176e18 7.

Keiley, M.K., Howe, T.R., Dodge, K.A., Bates, J.E., Pettit, G.S., 2001. The timing of child physical maltreatment: a cross-domain growth analysis of impact on adoles- cent externalizing and internalizing problems. Dev. Psychopathol. 13, 891e912.

Kessler, R.C., Sonnega, A., Bromet, E., Hughes, M., Nelson, C.B., 1995. Posttraumatic stress disorder in the national comorbidity survey. Arch. Gen. Psychiatry 52, 10 4 8e1060.

Kessler, R.C., Berglund, P., Demler, O., Jin, R., Merikangas, K.R., Walters, E.E., 2005.

Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the national comorbidity survey replication. Archives General Psychiatry 62, 593e602.

Khan, A., McCormack, H.C., Bolger, E.A., et al., 2015. Childhood maltreatment, depression, and suicidal ideation: critical importance of parental and peer emotional abuse during developmental sensitive periods in males and females.

Front. Psychiatry 6, 42.

Knudsen, E., 2004. Sensitive periods in the development of the brain and behavior.

J. Cogn. Neurosci. 16, 1412e1425.

Maercker, A., Michael, T., Fehm, L., Becker, E.S., Margraf, J., 2004. Age of traumati- sation as a predictor of post-traumatic stress disorder or major depression in young women. Br. J. psychiatry J. Ment. Sci. 184, 482e487.

Manly, J.T., Kim, J.E., Rogosch, F.A., Cicchetti, D., 2001. Dimensions of child maltreatment and children's adjustment: contributions of developmental timing and subtype. Dev. Psychopathol. 13, 759e782.

Margolin, G., Gordis, E.B., 2000. The effects of family and community violence on children. Annu. Rev. Psychol. 51, 445e479.

McCutcheon, V.V., Heath, A.C., Nelson, E.C., Bucholz, K.K., Madden, P.A., Martin, N.G., 2009. Accumulation of trauma over time and risk for depression in a twin sample. Psychol. Med. 39, 431e441.

McCutcheon, V.V., Sartor, C.E., Pommer, N.E., et al., 2010. Age at trauma exposure and PTSD risk in young adult women. J. Trauma Stress 23, 811e814.McLaughlin, K.A., 2016. Future directions in childhood adversity and youth psy- chopathology. J. Clin. Child. Adolesc. Psychol. 1e22.

McLaughlin, K.A., Green, J.G., Gruber, M.J., Sampson, N.A., Zaslavsky, A.M., Kessler, R.C., 2010. Childhood adversities and adult psychiatric disorders in the National Comorbidity Survey Replication II: associations with persistence of DSM-IV disorders. Arch. Gen. Psychiatry 67, 124e13 2.

McLaughlin, K.A., Fox, N.A., Zeanah, C.H., Sheridan, M.A., Marshall, P., Nelson, C.A., 2010. Delayed maturation in brain electrical activity partially explains the as- sociation between early environmental deprivation and symptoms of attention- deﬁcit/hyperactivity disorder. Biol. Psychiatry 68, 329e336.

McLaughlin, K.A., Fox, N.A., Zeanah, C.H., Nelson, C.A., 2011. Adverse rearing envi- ronments and neural development in children: the development of frontal electroencephalogram asymmetry. Biol. Psychiatry 70, 1008 e1015.

McLaughlin, K.A., Green, J.G., Gruber, M.J., Sampson, N.A., Zaslavsky, A.M., Kessler, R.C., 2012. Childhood adversities andﬁrst onset of psychiatric disorders in a national sample of US adolescents. JAMA Psychiatry 69, 1151e11 6 0.

McLaughlin, K.A., Green, J.G., Alegria, M., et al., 2012. Food insecurity and mental disorders in a national sample of U.S. adolescents. J. Am. Acad. Child. Adolesc.

Psychiatry 51, 1293e1303.

McLaughlin, K.A., Koenen, K.C., Hill, E.D., et al., 2013. Trauma exposure and post- traumatic stress disorder in a national sample of adolescents. J. Am. Acad. Child Adolesc. Psychiatry 52, 815e830.

Merikangas, K.R., He, S., Burstein, M., et al., 2011. Lifetime prevalence of mental disorders in U.S. Adolescents: results from the national comorbidity survey replication - adolescent supplement (NCS-A). J. Am. Acad. Child Adolesc. Psy- chiatry 49, 980e989.

Mueller-Pfeiffer, C., Moergeli, H., Schumacher, S., et al., 2013. Characteristics of child maltreatment and their relation to dissociation, posttraumatic stress symp- toms, and depression in adult psychiatric patients. J. Nerv. Ment. Dis. 201, 471e477.

Norman, R.E., Byambaa, M., De, R., Butchart, A., Scott, J., Vos, T., 2012. The long-term health consequences of child physical abuse, emotional abuse, and neglect: a systematic review and meta-analysis. PLoS Med. 9, e1001349.

Ogle, C.M., Rubin, D.C., Siegler, I.C., 2013. The impact of the developmental timing of trauma exposure on PTSD symptoms and psychosocial functioning among older adults. Dev. Psychol. 49, 2191e2200.

Ogle, C.M., Rubin, D.C., Siegler, I.C., 2015. The relation between insecure attachment and posttraumatic stress: early life versus adulthood traumas. Psychol. Trauma Theory Res. Pract. Policy 7 (4), 324.

Oldehinkel, A.J., Ormel, J., Verhulst, F.C., Nederhof, E., 2014. Childhood adversities and adolescent depression: a matter of both risk and resilience. Dev. Psycho- pathol. 26, 1067e1075.

Pietrek, C., Elbert, T., Weierstall, R., Muller, O., Rockstroh, B., 2013. Childhood ad- versities in relation to psychiatric disorders. Psychiatry Res. 206, 103e110.

Rusby, J.S., Tasker, F., 2009. Long-term effects of the British evacuation of children during World War 2 on their adult mental health. Aging Ment. Health 13, 391e404.

Schoedl, A.F., Costa, M.C., Mari, J.J., et al., 2010. The clinical correlates of reported childhood sexual abuse: an association between age at trauma onset and severity of depression and PTSD in adults. J. child Sex. Abuse 19, 156e17 0.

Schrag, N.M., McKeown, R.E., Jackson, K.L., Cuffe, S.P., Neuberg, R.W., 2008. Stress- related mental disorders in childhood cancer survivors. Pediatr. Blood Cancer 50, 98e10 3.

Schwartz, A.C., Bradley, R.L., Sexton, M., Sherry, A., Ressler, K.J., 2005. Posttraumatic stress disorder among African Americans in an inner city mental health clinic.

Psychiatr. Serv. 56, 212e215.

Schwartz, A.C., Bradley, R., Penza, K.M., et al., 2006. Pain medication use among patients with posttraumatic stress disorder. Psychosomatics 47, 136e14 2.

Scott, K.M., McLaughlin, K.A., Smith, D.A.R., Ellis, P.M., 2012. Childhood maltreat- ment and DSM-IV adult mental disorders: comparison of prospective and retrospectiveﬁndings. Br. J. Psychiatry 200, 469e475.

Sheridan, M.A., Fox, N.A., Zeanah, C.H., McLaughlin, K.A., Nelson, C.A., 2012. Varia- tion in neural development as a result of exposure to institutionalization early in childhood. Proc. Natl. Acad. Sci. 109, 12927e 12932.

Thornberry, T.P., Ireland, T.O., Smith, C.A., 2001. The importance of timing: the varying impact of childhood and adolescent maltreatment on multiple problem outcomes. Dev. Psychopathol. 13, 957e979.

Thornberry, T.P., Henry, K.L., Ireland, T.O., Smith, C.A., 2010. The causal impact of childhood-limited maltreatment and adolescent maltreatment on early adult adjustment. J. Adolesc. Health 46, 359e365.

Ullman, S.E., 2007. Relationship to perpetrator, disclosure, social reactions, and PTSD symptoms in child sexual abuse survivors. J. Child. Sex. Abus. 16, 19e36.

Usher, J.A., Neisser, U., 1993. Childhood amnesia and the beginnings of memory for four early life events. J. Exp. Psychol. Gen. 122, 155e16 5.

Wang, Y.P., Gorenstein, C., 2013. Psychometric properties of the Beck depression inventory-II: a comprehensive review. Rev. Bras. Psiquiatr. 35, 416e431.

Widom, C.S., DuMont, K., Czaja, S.J., 2007. A prospective investigation of major depressive disorder and comorbidity in abused and neglected children grown up. Arch. Gen. Psychiatry 64, 49e56.

Yoshihama, M., Gillespie, B.W., 2002. Age adjustment and recall bias in the analysis of domestic violence data: methodological improvements through the appli- cation of survival analysis methods. J. Fam. Violence 17, 199e221.

Zimmer-Gembeck, M.J., Skinner, E.A., 2011. The development of coping across childhood and adolescence: an integrative review and critique of research. Int. J.

Behav. Dev. 35, 1e17. E.C. Dunn et al. / Journal of Psychiatric Research 84 (2017) 119e127127