Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Sep 1.
Published in final edited form as: J Affect Disord. 2020 May 22;274:126–135. doi: 10.1016/j.jad.2020.05.131

The Effect of Traumatic Events on the Longitudinal Course and Outcomes of Youth with Bipolar Disorder.

Maria Andreu Pascual 1, Jessica C Levenson 1, John Merranko 1, Mary Kay Gill 1, Heather Hower 2,3,4, Shirley Yen 2,5, Michael Strober 6, Tina R Goldstein 1, Benjamin I Goldstein 7, Neal D Ryan 1, Lauren M Weinstock 2,8, Martin B Keller 2,8, David Axelson 9, Boris Birmaher 1
PMCID: PMC7365761  NIHMSID: NIHMS1598435  PMID: 32469795

1. Introduction

Bipolar Spectrum Disorder (BD) in youth is an episodic illness with an average prevalence rate of 3.9%, which is associated with significant psychosocial difficulties, financial burden and behavioral health costs, as well as high risk for suicidality and substance abuse (Birmaher, 2018; Peele, 2004; Van Meter et al., 2019). Due to the episodic and potentially lifelong course of early-onset BD, longitudinal studies are needed to understand the factors associated with its trajectory and outcomes, to inform treatment, and to improve outcomes (Birmaher, 2018).

Given the complex causal architecture of BD, the role played by individual risk factors and their likely interactions in determining how an illness trajectory unfolds is not easily determined. Nonetheless, exposure to Traumatic Events (TEs)1 has emerged as one of the factors having substantial impact on the course of BD (Agnew-Blais and Danese, 2016). TEs are 2.6 times more prevalent in adults with BD compared to healthy controls (Palmier-Claus et al., 2016). Compared to BD adults who have not experienced TEs, BD adults with TEs have earlier BD onset, greater depressive and psychotic symptoms, psychiatric comorbidity and suicidality, greater number of recurrences, and more psychosocial stressors (Anand et al., 2015; Cohen et al., 2004a; Etain et al., 2013; Garno et al., 2005; Gershon et al., 2013; Hammersley et al., 2003; Leverich et al., 2002). TEs could impact health outcomes (Leclerc et al., 2018), and they have been associated with biological changes such as inflammation (Danese et al., 2007) and changes in gray and white matter integrity in various brain areas potentially implicated in emotion regulation and threat sensitivity (Poletti et al., 2016; Stevelink et al., 2018), possibly explaining the functional deficits observed in BD.

Among TEs, sexual and/or physical abuses are the most widely studied, and they are the strongest predictors of an adverse BD course in adults (Aas et al., 2016; Daruy-Filho et al., 2011). Abuse affects neurocognitive performance, worsens functioning and mood symptomatology, increases suicidality, and is associated with higher rates of psychiatric comorbidity in both cross-sectional and longitudinal studies in BD (Farias et al., 2019; Larsson et al., 2013; Leverich et al., 2002; Maniglio, 2013; Savitz et al., 2007).

While much of the research on the impact of TEs on BD has been conducted among adults, fewer studies have examined its impact among BD youth. Like adults, cross-sectional/concurrent studies and chart reviews show that BD youth are more likely to report TEs than healthy youth and youth with non-BD-psychopathology (Romero et al., 2009b; Rucklidge, 2006; Tillman et al., 2003). These studies also suggest that history of TEs among BD youth is associated with more hospitalizations, decreased treatment response, delayed diagnosis, greater suicidality, higher rates of psychiatric comorbidity, worse psychosocial functioning, and earlier age of BD onset compared to BD youth without TE and non-BD youth (Cazala et al., 2019; Du rocher Schudlich, 2015; Marchand et al., 2005; Rucklidge, 2006).

Longitudinal studies examining clinical and functional outcomes among BD youth with TEs have been few, especially those that have used frequent assessments of trauma over time. Daglas and colleagues showed that BD youth with past TEs had more symptoms of mania and depression, general psychopathology, and poorer social and functional outcomes at a 12-month evaluation, as compared to those without TEs (Daglas et al., 2014). Neria and colleagues followed a large sample of youth and adults hospitalized for bipolar psychosis (Neria et al., 2005), finding that childhood-TEs were associated with more severe symptomatology and lower remission rates 24-months after discharge. Kim and colleagues, showed that high stress (repeatedly measured over a 12-month-span) was associated with less mood symptoms improvement after one-year among 38 adolescents with BD compared to those with low stress (Kim et al., 2007). Conus and colleagues studied 118 youth experiencing their first-psychotic-mania using a retrospective file-audit, finding that those with abuse experienced poorer adherence to an 18-month early-intervention treatment than those without abuse exposure (Conus et al., 2010). However, other longitudinal studies showed only partial/no effects of TEs over BD course and onset (Strawn et al., 2010; Tijssen et al., 2010).

Nevertheless, the above-noted studies have been limited by multiple methodological factors including: 1) cross-sectional designs; 2) existing longitudinal designs limited to a maximum of 2 years for BD samples; 3) few assessed symptoms and/or TEs longitudinally; 5) infrequent follow-ups; 5) small sample sizes (N<150); 6) focus only on BD-I participants; 7) unadjusted confounding factors (e.g., socioeconomic status (SES), family history); 8) reliance on chart reviews, self-reports, and lack of direct interviews with the youth participant, and 9) retrospective measures that were covering long periods of time, increasing the likelihood of recall bias (Hardt and Rutter, 2004; Widom et al., 2004). Studies that use frequent prospective assessments of trauma and abuse and allow for the longitudinal examination of TEs impact on the course of early-onset BD could be valuable. Such studies may potentially reveal clinical and psychosocial outcomes not yet studied, providing an opportunity to identify risks and to create effective detection and intervention strategies.

The Course and Outcome of BD Youth (COBY) study is a multisite longitudinal research study of youth with childhood-onset BD followed into adulthood. This study provides a unique opportunity to examine the longitudinal effects of trauma in this age group and includes assessments of traumatic events and clinical course at various time-points. Previous cross-sectional findings from COBY have suggested that youth with BD had higher rates of negative life events and lower rates of positive life events compared to healthy controls, and 20% had a history of abuse (Romero et al., 2009a; Romero et al., 2009b).

The goals of this study are to extend the prior findings by prospectively examining: 1) the rate and effects of TEs among BD youth; and, specifically, 2) the effects of abuse on participants’ course and outcomes over an average period of 8.7 years. Based on the existing literature, we hypothesized that: BD participants with lifetime TEs, particularly those with lifetime history of abuse, will spend less time euthymic, more time with threshold mania and depression, have more recurrences and comorbid disorders (e.g., substance use disorders [SUD] and Post-Traumatic Stress Disorders [PTSD]), and worse functioning, compared to those without past history of TEs.

2. Method

2.1. Participants

Youth participating in the COBY study provided data for this analysis. Details of the COBY sample are described elsewhere (Axelson et al., 2006; Birmaher et al., 2014). Briefly, from October 2000 through July 2006, COBY enrolled 446 participants, ages 7 to 17 years 11 months at intake (mean=12.7 ± 3.2 years old, 53% males, and majority White), meeting Diagnostic and Statistical Manual of Mental Disorders-IV (DSM-IV) ((Association, 2004) criteria for BD-I (n = 260), BD-II (n = 32), or BD-Not Otherwise Specified (NOS), based on the COBY operationalized definition (n = 154) (For criteria see (Axelson et al., 2006)). BD onset age was defined by the age of onset of any DSM-IV mood episode or an episode fulfilling the COBY’s BD-NOS criteria. Participants were enrolled regardless of current mood state, and recruited from outpatient clinics (84.4%), inpatient units (4.4%), advertisements (6.7%), and referrals from other physicians (4.4%) at the three study sites (University of Pittsburgh, Brown University, and University of California at Los Angeles [UCLA]). Youth with schizophrenia, an intellectual disability, autism, and mood disorders secondary to medical conditions or substance use were excluded from the study.

Since the instrument for screening severe lifetime TEs (Traumatic Events Screen: TES) (Kaufman et al., 1997) (Supplemental Screen 1) was added in 2007; this paper includes 375 participants (mean age 17, age range: 8.7–25.8) who had at least one prospective follow-up visit during which the TES was completed. Participants were prospectively interviewed every 7 months for an average of 8.7years with a retention rate of 94%. Except for the included participants having significantly higher rates of family history of psychosis (p<0.01), there were no other significant demographic or clinical differences between the included and excluded samples.

2.2. Procedure

Each university’s Institutional Review Board approved the study before enrollment, and informed consent/assent was obtained from the participants and parents at intake. COBY research staff administered semi-structured interview assessments to participants and parents. Participants age 18/older chose whether to include parents’/secondary informants’ in study interviews. Participant’s symptomatic and psychosocial course was reviewed by a research staff along with a study investigator who was ultimately responsible for the clinical ratings. After caregiver/participants interviews, these rating scores were used for the current analyses.

2.2.1. Measures

At intake, participants and their primary caregivers were directly interviewed for presence of current and lifetime psychopathology using the K-SADS-PL (Kaufman et al., 1997). Mood symptoms severity was assessed using the K-SADS Mania Rating Scales (K-MRS) (Axelson et al., 2003) and the depression section of the K-SADS-P (K-DRS), both derived from the mood disorder sections of the K-SADS-Present Episode (K-SADS-P; 4th revision) (Chambers et al., 1985). Diagnostic Interrater agreement was high for the K-SADS-PL (range 93%−100%). The K-SADS-PL has excellent overall diagnostic reliability with kappa coefficients for psychiatric disorders of ≥0.80 (Kaufman et al., 1997). The intraclass correlation coefficients for the K-SADS-MRS and K-SADS-DEP-P were≥0.83. Kendall’s concordance coefficient was 0.85 for major depressive disorder and 0.78 for mania/hypomania. Week-by-week longitudinal changes in psychiatric symptoms were assessed using the Adolescent Longitudinal Interval Follow-Up Evaluation (A-LIFE) (Keller et al., 1987) using this instrument’s Psychiatric Status Rating Scales (PSR) with good psychometric properties (Warshaw et al., 2001). The PSR used numeric values linked to DSM-IV criteria, ranging from 1 to 6 for mood disorders (scores indicate respectively: ≤2, euthymia, 3–4 subsyndromal symptoms, and ≥5 syndromal symptoms). Full recovery was defined as 8 consecutive weeks with a score of≤2 (minimal/no mood symptoms). Recurrence was defined by the presence of a score of ≥5, with duration of 1-week for mania/hypomania and 2-weeks for depression. PSR reliability in COBY was good/very good (Axelson et al., 2011). Intraclass correlation was 0.85 when assessing percent of time meeting full DSM-IV criteria for a mood episode and 0.82 for subthreshold mood symptoms. Reliability for PSR mood disorder ratings over COBY’s course had an average Kendall’s W of 0.8.

The presence of lifetime (past/present) TEs was obtained using the Traumatic Events Screen (TES) (Supplemental Screen 1), a brief interview including the 11 items derived from the PTSD section of the K-SADS-PL (Kaufman et al., 1997), plus an additional item (victim of intimate partner violence). The screen included events such as being in serious accident, witnessing a disaster or violence, being the victim of a crime, or experiencing physical/sexual abuse at any point in their lives (i.e., physical abuse in adulthood such as being hurt or threatened by peers/significant others). If a participant endorsed an event, the total number of occurrences and dates of the first, second and most recent event were recorded. Events were ascertained at each follow-up separately for parents and participants, and a summary score for each item was created. The events in the TES are answered as yes/no. In general, if either parent/participant endorsed an event, then the summary score was counted as positive. If there was discrepancy between informants, further clarification was obtained from informants. If a parent reported an event as traumatic, but participant indicated that he/she did not perceive this as traumatic, the event would not be given a positive score. Spearman correlation was strong between summary scores and participant reports at 0.95 and was slightly weaker between summary scores and parent reports at 0.72, indicating that summary scores depended more heavily on participant reports than parent reports. Spearman correlation between parents’ and participants’ reports was 0.59.

Information about lifetime sexual and physical abuse was obtained at the time of study recruitment with the KSADS-PL, and over follow-up with the TES.

SES was ascertained using the 4-factor Hollingshead Scale (Hollingshead, 1975). Participants’ psychosocial functioning was assessed using the Psychosocial Functioning Scale (PSF), the Children’s Global Assessment Scale (CGAS) (< age 22) and the Global Assessment of Functioning (GAF) (> age 22) (Jones et al., 1995; Keller et al., 1987; Shaffer et al., 1983). The parent/caretaker was interviewed at study intake about his/her own personal psychiatric history using the Structured Clinical Interview for DSM-IV (First, 1996), and first/second degree relatives psychiatric status was ascertained using an enhanced version of the Family History Screen (Weissman et al., 2000).

2.3. Statistical Analyses

Univariate group contrasts were performed via chi-squared, Fisher’s exact, and t-tests as appropriate. Contrasts of suicide attempts and self-injury were made via negative binomial regression.

For the analyses, TEs were divided into six different clusters depending on the nature of the event. TEs were clustered as follows: 1) Witnessing a Traumatic Event, 2) Experiencing an accident, 3) Being victim of violence or abuse, 4) Being confronted with traumatic news, 5) Experiencing other traumatic events, which included an open question about any past TEs such as abortion, separation from family, or bankruptcy, and finally, 6) Any of the above (See Supplemental Screen 1). All analyses involving TEs rates were first run using participants’ overall rates of any TEs, and then re-run separately by TE cluster. To test the association between participants’ rates of TEs and rates of mood symptoms, Poisson regressions were fit covarying for age at end of follow-up as well as demographics, comorbid diagnoses, and family history retained by Least Absolute Shrinkage and Selection Operator (LASSO). Briefly, a LASSO is a modified form of linear or generalized linear regression that penalizes overfit models via a regularization parameter that proportionally shrinks the magnitude of predictor coefficients toward zero, and in the case of less important predictors, coefficients shrink all the way to zero. In doing so, covariate selection is implicitly performed, as less important variables are removed from the model without the potential biases of other variable selection techniques such as multiple comparisons and collinearity between predictor variables. Cross-validation selected the optimal LASSO regularization parameter via the one-standard-error-rule, which conservatively implements the most regularized model whose model error is within one standard error of the minimum (Hastie T, 2009). Poisson regression effect sizes were estimated via standardized incidence rate ratios (RR). To test the association between participants’ rates of TEs during recovery periods and recurrence risk, Andersen and Gill recurrent events Cox proportional hazards regressions (Andersen PK, 1982) were fit covarying for age, BD onset age, and number of previous recurrences, as well as demographics, comorbid diagnoses, and family history retained by backward selection (retention criterion for each model: p<0.1); effect sizes were estimated via standardized hazard ratios (HR).

Since PTSD is a common diagnosis among participants with TEs, logistic regression was used to test whether increased TE rates aggregated over follow-up were associated with increased likelihood to develop PTSD onset over follow-up (both considering rates of any TE as well as separately by cluster). Effect sizes were estimated via odds ratios (OR).

Of all TEs, sexual/physical abuse had more specific and reliable occurrence dates, which enabled more sophisticated longitudinal and survival analyses. Mixed logistic regressions modeled the probability of mood symptoms during each follow-up period as a function of age and time-varying lifetime abuse history, which was expressed as a dichotomous variable indicating whether abuse had ever been reported as of each follow-up period (random intercept accounted for within-participant correlation). These models controlled for demographics, comorbid diagnoses, and family history retained by LASSO. Lastly, Kaplan-Meier and Cox proportional hazards regression analyses (controlling for demographics, comorbid diagnoses, and family history retained by LASSO) tested for associations between lifetime abuse history and the subsequent risk to develop new onset non-affective comorbidities.

LASSO models used R 3.5.1(Team, 2019); all other statistical analyses used SAS 9.4 (SAS, 2019).

3. Results

Of the 375 participants, 46.7% were female (n=175), and 72.0% (n=270), 14.1% (n=53) and 13.9% (n=52) had a diagnosis of BD-I, BD-II and BD-NOS respectively as of their last follow-up assessment. See Table 1 for demographics.

Table 1:

Intake demographic, diagnostic, clinical, family history and psychosocial variables associated with traumatic events

variable Mean ± SD or N (%) Test Stat p-value
No Traumatic Events (n=59) Traumatic Events (n=316)
Demographics Baseline Age 18.3 (3.5) 16.7 (3.8) t=2.98 0.003
Baseline SES 4.5 (0.8) 4.1 (1.1) t=3.06 0.003
Female 32 (54.2) 143 (45.3) χ2=1.61 0.2
White 47 (79.7) 260 (82.3) χ2=0.23 0.6
Both Parents 20 (33.9) 95 (30.1) χ2=0.34 0.6
Follow-up Length 5.5 (3.6) 8.3 (2.0) t=5.94 <0.0001
Lifetime Clinical Variables Bipolar Onset Age 11.3 (4.0) 8.9 (3.8) t=4.42 <0.0001
Bipolar Disorder Subtype
 BD-I 46 (78.0) 224 (70.9) χ2=1.83 0.4
 BD-II 8 (13.6) 45 (14.2)
 BD-NOS 5 (8.5) 47 (14.9)
ADHD 23 (39.0) 223 (70.6) χ2=21.98 <0.0001
Disruptive Behavior Disorders 22 (37.3) 194 (61.4) χ2=11.83 0.0006
Anxiety Disorder 24 (40.7) 234 (74.1) χ2=25.80 <0.0001
PTSD 2 (3.4) 81 (25.6) χ2=14.27 0.0002
SUD 9 (15.3) 153 (48.4) χ2=22.28 <0.0001
Suicidality, Self-Injury Suicidal Ideation 6 (10.2) 93 (29.4) χ2=9.49 0.002
Suicide Attempts (per year) 0.06 (0.2) 0.08 (0.2) Negative Binomial Wald
χ2=0.03		 0.9
Self-Injury (per year) 0.05 (0.2) 0.11 (0.4) Negative Binomial Wald
χ2=1.56		 0.2
Family History Depression 43 (72.9) 286 (90.5) χ2=14.35 0.0002
Mania 22 (37.3) 198 (62.7) χ2=13.20 0.0003
ADHD 18 (30.5) 158 (50.0) χ2=7.58 0.006
CD 18 (30.5) 122 (38.6) χ2=1.39 0.2
Schizophrenia 4 (6.8) 26 (8.2) Fisher’s Exact 1
Psychosis 0 (0) 69 (21.8) Fisher’s Exact <0.0001
Anxiety 35 (59.3) 244 (77.2) χ2=8.36 0.004
SUD 35 (59.3) 229 (72.5) χ2=4.12 0.04
Suicidality 23 (39.0) 173 (54.8) χ2=4.95 0.03
Open in a new tab

SD=Standard Deviation; SES=Socio Economic Status; CGAS=Children’s Global Assessment Scale; GAF=Global Assessment of Functioning; BD=Bipolar Disorder; ADHD=Attention Deficit and Hyperactivity Disorder; CD=Conduct Disorder; SUD=Substance Use Disorder; PTSD= Post-Traumatic Stress Disorder. a=Hollingshead Redlich criteria

3.1. Prevalence of Traumatic Events

Eighty-four percent (316/375) reported at least one TE during follow-up. Participants with at least one TE subsequently reported TEs in 39% of their follow-up visits. As compared to participants without TEs, univariate analyses showed that those with at least one TE were significantly more likely to have younger intake age, younger BD-onset, lower SES, longer follow-up duration, lower psychosocial functioning, and higher rates of suicidal ideation (SI), Attention Deficit and Hyperactivity Disorder (ADHD), Disruptive Behavior Disorders (DBD), anxiety, and SUD. Also, they had higher rates of family history of depression, mania, ADHD, psychosis, anxiety, SUD, and suicidality (Table 1). Differences in the rates of comorbid disorders remained significant after covarying for between-group significant demographic and family history factors.

As shown in Table 2, among those with at least one lifetime TE, the most prevalent TE clusters were: 1) being confronted with traumatic news (80%), among which the most common event was being confronted with death (73%) (i.e., death or serious illness of a loved one); 2) experiencing an accident (58%), among which the most common event was being in a car accident (35%); and 3) experiencing any kind of violence or abuse (51%), among which the most common events were physical abuse (34%) and being a victim of a violent crime (27%) (See Supplemental Screen 1). TEs occurred approximately once every two years and at higher rate during adulthood than before 18 years old (0.9/year vs. 0.6/year, with rate ratio:1.5; p<0.0001).

Table 2:

Traumatic events descriptive statistics (clusters, specific types, and mean rate of events) over follow-up period

Traumatic Events Descriptive Statistics (N=375) Percent of Sample Who Reported Event Mean Rate of Events per Year (SD)
Witness Traumatic Event 31% 0.06 (0.14)
 Witness to Natural Disaster 7% 0.01 (0.02)
 Witness to Violent Crime 20% 0.03 (0.10)
 Witness to Domestic Violence 11% 0.02 (0.08)
Accidents 58% 0.12 (0.20)
 Car Accident 35% 0.05 (0.09)
 Fire 9% 0.01 (0.07)
 Other Accident 36% 0.06 (0.12)
Violence and Abuse 51% 0.17 (0.35)
 Victim of Violent Crime 27% 0.04 (0.13)
 Physical Assault 21% 0.02 (0.05)
 Sexual Assault 4% 0.00 (0.02)
 Victim of Intimate Partner Violence 14% 0.05 (0.21)
 Physical Abuse 34% 0.04 (0.08)
 Sexual Abuse 17% 0.01 (0.05)
Confronted with Traumatic News 80% 0.25 (0.23)
 Death 73% 0.15 (0.16)
 Other Traumatic News 53% 0.10 (0.14)
Other (divorce, separation, bankruptcy…) 62% 0.11 (0.14)
Any of the Above 84% 0.71 (0.64)
Open in a new tab

SD=Standard deviation

3.2. Association between TEs and Mood Symptomatology

After adjusting for potential confounders retained by LASSO (age at end of follow-up, demographics, comorbid diagnoses, and family history), participants who experienced higher rates of any type of TE reported significantly lower rates of euthymia (p=0.001;Table 3) and significantly higher rates of sub/threshold Major Depressive Episodes (MDE) over follow-up (p<0.001). When analyzing this effect differentially by TE cluster, participants experiencing any TEs reported significantly lower rates of euthymia and significantly higher rates of sub/threshold MDE (all p≤0.03). There were no significant associations between witnessing a TE or experiencing accidents and increased rates of mood symptoms. Estimated standardized incident rate ratios indicated that as compared to average participants (those at the mean rate of TEs), those whose rates of those TEs were one standard deviation above the mean had 5–9% lower rates of euthymia, 13–24% higher rates of threshold MDE, and 7–12% higher rates of subthreshold/worse MDE (depending on type of TE). Further, participants whose rates of violence/abuse were one standard deviation above the mean also had an estimated 17% higher rate of threshold hypo/mania (p=0.008).

Table 3:

Poisson regressions of PSR* mood states (% of time) as a function of traumatic events rate

Traumatic Event Cluster Mood State Standardized Rate Ratio ** Wald χ2 p-value
Any Euthymia 0.92 (0.88, 0.97) 10.31 0.001
Threshold MDE *** 1.24 (1.12, 1.37) 18.12 <0.0001
Threshold Hypo/mania 1.12 (0.96, 1.30) 2.03 0.2
Subthreshold MDE 1.11 (1.05, 1.18) 11.35 0.0008
Subthreshold Hypo/mania 1.05 (0.98, 1.14) 2.01 0.2
Witness Traumatic Event Euthymia 0.96 (0.92, 1.01) 2.59 0.1
Threshold MDE 1.09 (0.99, 1.20) 3.13 0.08
Threshold Hypo/mania 1.10 (0.97, 1.24) 1.99 0.2
Subthreshold MDE 1.04 (0.99, 1.11) 2.19 0.1
Subthreshold Hypo/mania 1.06 (0.99, 1.13) 3.11 0.08
Accidents Euthymia 1.00 (0.96, 1.05) 0.03 0.9
Threshold MDE 1.06 (0.95, 1.17) 1.00 0.3
Threshold Hypo/mania 0.87 (0.71, 1.06) 1.94 0.2
Subthreshold MDE 1.01 (0.95, 1.08) 0.12 0.7
Subthreshold Hypo/mania 0.97 (0.90, 1.05) 0.52 0.5
Violence and Abuse Euthymia 0.95 (0.90, 0.99) 5.05 0.02
Threshold MDE 1.13 (1.03, 1.23) 7.42 0.006
Threshold Hypo/mania 1.17 (1.04, 1.31) 7.02 0.008
Subthreshold MDE 1.07 (1.01, 1.13) 5.70 0.02
Subthreshold Hypo/mania 1.04 (0.97, 1.12) 1.35 0.2
Confronted with Traumatic News Euthymia 0.94 (0.90, 0.99) 6.02 0.01
Threshold MDE 1.17 (1.04, 1.31) 7.14 0.008
Threshold Hypo/mania 1.02 (0.87, 1.20) 0.05 0.8
Subthreshold MDE 1.08 (1.01, 1.15) 4.94 0.03
Subthreshold Hypo/mania 1.06 (0.98, 1.14) 2.29 0.1
Other (divorce, separation, bankruptcy…) Euthymia 0.94 (0.90, 0.99) 5.93 0.01
Threshold MDE 1.20 (1.09, 1.32) 13.47 0.0002
Threshold Hypo/mania 0.98 (0.83, 1.15) 0.08 0.8
Subthreshold MDE 1.12 (1.05, 1.19) 12.69 0.0004
Subthreshold Hypo/mania 1.01 (0.94, 1.09) 0.12 0.7
Open in a new tab
*

PSR=Psychiatric Status Rating Scale; numeric values linked to DSM-IV criteria, range 1–6 for mood disorders. Scores ≤2 indicate euthymia, scores of 3–4 indicate subsyndromal symptoms, and scores ≥5 syndromal symptoms.

**

Models adjust for age at end of follow-up + demographics, comorbid diagnoses, and family history retained by LASSO.

***

MDE=Major Depressive Episode

3.3. Associations between TEs and PTSD diagnosis

PTSD diagnosis was evaluated throughout each follow-up interval and diagnostic rates were compared between TE clusters. After adjusting for confounders, participants who reported above average rates of any TE had more than double the estimated odds of developing PTSD (OR=2.19, p<0.0001). Specifically, every TE cluster, except witnessing a traumatic event, was significantly associated with increased odds of developing PTSD (all OR>1.40, ps<0.009)

3.4. Association between TEs and risk for Mood Recurrences

After adjusting for number of prior recurrences and potential confounders retained by LASSO, Cox regressions showed that as compared to participants with average number of TEs, those whose rates of any TE during recovery periods were one standard deviation above the mean had significantly increased recurrence risk of any type of episode (HR=1.42, p=0.001; Table 4). When analyzing differentially by TE cluster, all clusters except accidents and “other TEs” (i.e., abortion, separation from parents) were associated with significantly increased recurrence risk (HR=1.12–1.61, p<0.03). Participants who experienced violence/abuse during recovery periods had 1.6 times the estimated hazard of subsequent recurrence risk.

Table 4.

Cox regressions of recurrence* risk as a function of traumatic events rate during recovery periods

Traumatic Event Cluster Standardized Hazard Ratio ** Wald χ2 p-value
Any 1.42 (1.15, 1.76) 10.48 0.001
Witness Traumatic Event 1.12 (1.04, 1.20) 9.45 0.002
Accidents 1.04 (0.90, 1.19) 0.25 0.6
Violence and Abuse 1.61 (1.07, 2.40) 5.31 0.02
Confronted with Traumatic News 1.11 (1.01, 1.22) 5.02 0.03
Other (divorce, separation, bankruptcy…) 1.11 (0.95, 1.30) 1.67 0.2
Open in a new tab
*

Recurrence was defined by the presence of a score of ≥5 in the Psychiatric Status Rating Scale, with duration of 1-week for mania/hypomania and 2-weeks for depression.

**

Models adjust for age, BD onset age, and number of previous recurrences (0 vs. 1 vs. 2+) + demographics, comorbid diagnoses, and family history retained by backward selection (p-value<0.1).

3.5. Mood Course and history of sexual/physical Abuse

Of the 375 participants, 167 (44.5%) reported lifetime sexual and/or physical abuse. Abuse-exposed participants were mostly females, less likely to live with both parents, and had more comorbidities, suicidal behaviors, family psychopathology, and worse functioning (all p<0.03) than those without abuse. There were no significant age differences between participants with/without abuse history. Among abuse-exposed participants, 34% reported physical abuse and 17% reported sexual abuse. Further, Kaplan-Meier estimation indicated that 20% of the sample reported abuse by age 21, and 30% by age 24 (Supplemental Figure 1).

In contrast with other TEs, abuse dates were more specifically and reliably reported in the sample and were obtained by two different sources, which enabled more sophisticated longitudinal analyses involving abuse. Considering all participants’ longitudinal observations, after adjusting for time-varying confounding factors retained by LASSO, mixed logistic regression found a significant association between lifetime history of abuse (i.e., lifetime presence/absence of abuse history for each distinct follow-up period) and prevalence of mood symptoms, with increased likelihood of threshold and subthreshold/worse MDE (OR=1.82 and 1.38, p<0.03) and subthreshold/worse hypomania (OR=1.40, p=0.03), and a marginally non-significant trend for reduced likelihood of euthymia (OR=0.75, p=0.056; Table 5).

Table 5:

Mixed logistic regressions of mood symptoms as a function of presence vs. absence of lifetime abuse

Mood State Odds Ratio F-stat p-value
Euthymia 0.75 (0.56, 1.01) 3.63 0.06
Threshold MDE 1.82 (1.37, 2.42) 16.96 <0.0001
Threshold Hypo/mania 1.16 (0.81, 1.67) 0.65 0.4
Subthreshold MDE 1.38 (1.04, 1.84) 4.90 0.03
Subthreshold Hypo/mania 1.40 (1.03, 1.89) 4.66 0.03
Open in a new tab

Models adjust for current age + demographics, comorbid diagnoses, and family history retained by LASSO

The more specific and reliable abuse dates enabled analyses of the association between lifetime abuse history and the risk to develop new onset non-affective disorders. Eligible sample sizes for each new onset non-affective disorder excluded participants for whom we could not establish whether abuse preceded new onset of the disorder. Also, disorders with small prevalence were not included. Kaplan-Meier and Cox regression analyses (controlling for demographics and family history retained by LASSO) found no significant association between abuse history and risk of new onset Generalized Anxiety Disorder (GAD) or DBD (Supplemental Table 1). However, a significant effect was established between abuse history and risk of new onset SUD. Kaplan-Meier estimation indicated that mean age of SUD onset was 23.5 years old among participants with abuse and 25.6 years old among those without abuse (Log-Rank χ2stat=23.88, p<0.0001, Figure 1). After adjusting for confounders retained by LASSO, Cox regression estimated that abuse-exposed participants had over twice the hazard of SUD-onset as compared to those without abuse (HR=2.14, p<0.0001).

Figure 1.

Figure 1.

Open in a new tab

Kaplan-Meier survival estimation: substance use disorder risk stratified by history of abuse *

*Model controls for demographics, comorbid diagnoses, and family history retained by Least Absolute Shrinkage and Selection Operator (LASSO).

4. Discussion

To our knowledge, this is the largest sample of youth with BD followed into young adulthood longitudinally, thereby allowing prospective examination of the long-term clinical and social effects of exposure to severe TEs, including sexual/physical abuse.

There were two major findings. First, consistent with our hypotheses, we observed that after adjusting for confounders, participants with a lifetime history of at least one TE (84%), particularly those with exposure to abuse (44.5%) and violence, showed worse mood course, more non-mood psychopathology, lower SES, and poorer psychosocial functioning as compared to those without TEs. Participants with lifetime TEs had earlier BD onset (Cohen’s d=0.63) and greater sub/threshold mood symptomatology. Specifically, participants who had above average rates of TEs had up to 24% more threshold MDE and up to 11% more sub/threshold MDE compared to those with average TE rates, and those who experienced above average rates of violence and abuse had up to 17% more hypo/mania than those with average rates of violence/abuse. These participants had more mood recurrences (almost 1.5 times the subsequent risk of recurrence if TEs occurred during recovery periods), more SI, and comorbid disorders (anxiety, ADHD, DBD, SUD, PTSD) and family psychopathology (MDE, mania, ADHD, psychosis, anxiety, SUD, suicidality). Second, after the first abuse incident, participants had more severe MDE and hypo/mania symptoms compared to before the first abuse incident and also compared to participants who never experienced abuse. Participants who had been exposed to abuse had 82% greater odds of threshold MDE, 38% greater odds of subthreshold MDE and 40% greater odds of subthreshold hypomania compared to participants who had not experienced abuse. Abuse-exposed participants were mostly females, less likely to live with both parents, had worse mood course, and had higher rates of SI, suicide attempts, and self-injury. Furthermore, participants with lifetime abuse developed SUD more than twice as frequently and with earlier onset than participants without abuse history.

In a recent meta-analysis, the lifetime prevalence of TEs among adults with BD ranged from 8–77% (Palmier-Claus et al., 2016), while the prevalence of TEs among BD youth in some longitudinal studies has been reported around 40% (Daglas et al., 2014; Neria et al., 2005). The wide range of trauma prevalence in BD in the literature may be accounted for by methodological variability (e.g., definition of trauma, different instruments). The lifetime prevalence of abuse in BD has been reported at 24% in a systematic review (Maniglio, 2013), while concurrent/retrospective studies have reported the prevalence of abuse among BD youth around 11–24% (Conus et al., 2010; Du rocher Schudlich, 2015). The higher prevalence in our sample (84% with at least one TE and 45% with physical and/or sexual abuse) may be accounted by the fact that the COBY study included a referral sample and participants were followed for an average of 8.7years, increasing the likelihood to experience TEs. Furthermore, in contrast with other studies, COBY participants had frequent assessments (including self-reports, interviews, and collateral information), increasing the likelihood of identifying TEs.

Similar to prior literature, TEs presence among BD youth in our study was associated with poorer clinical and psychosocial outcomes (Daglas et al., 2014; Daruy-Filho et al., 2011; Marchand et al., 2005), and TEs increased as youth aged (Kim et al., 2007). Consistent with BD youth studies, we replicated the finding that participants with TEs have significantly greater psychiatric comorbidities, such as SUD and PTSD (Du rocher Schudlich, 2015). In addition, our findings uniquely showed that participants who had TEs above the average rate had more symptomatology than those with average TEs rates. Although this association between clinical severity and TEs rate has been reported in youth with other psychopathologies (Boe et al., 2018; Taylor and Gotham, 2016), this has not yet been studied among bipolar youth. Further, the risk of mood recurrence associated with TEs has been mostly described among BD adults (Cohen et al., 2004a; Leverich et al., 2002), with very few studies including BD youth (Neria et al., 2005; Strawn et al., 2010). Our findings showed that rates of even one standard deviation above the average of any TEs experienced during recovery periods predicted a 42% increased risk of subsequent mood recurrence compared to periods without TEs. However, a previously published longitudinal study did not show similar findings, reporting instead that trauma symptoms did not predict mood recurrences (Strawn et al., 2010). This difference may be attributable to the use of trauma symptomatology scores, not the presence of TEs.

While experiencing more TEs than average was generally associated with poorer course and more recurrences in our study, this was not the case for experiencing an accident, and witnessing a TE was not related to more severe mood symptoms. These results are similar to reports showing that experiencing an accident or witnessing trauma were not related to increased suicide risk among community adults and depressed adolescents (Jeon et al., 2014; Nrugham et al., 2010). This could be explained by the relatively independent nature of these events, as their occurrence may be more outside the individual’s control. Still, we did not measure whether the TEs were dependent on the participants behaviors, so we cannot assure this explains the difference in findings.

Consistent with the adult BD literature (Carbone et al., 2019), our findings suggest that among various types of trauma, abuse has greater impact on the course of BD youth than other TEs, particularly during recovery periods. Studies among BD youth with abuse exposure are mixed. While some showed that individuals with abuse experience poorer functioning, more severe mood symptoms, more frequent episodes, and more SUD comorbidity compared to those without abuse (Du rocher Schudlich, 2015), others have reported that abuse exposure among BD youth was not associated with poorer symptomatic and functional outcomes (Conus et al., 2010). Unlike our study, prior analyses did not consider temporal associations. In our study, there was a worsening in mood symptoms after abuse occurred. Abuse dates were more reliably reported in COBY, enabling this comparison of mood before/after abuse, representing a unique contribution among BD youth literature highlighting the detrimental effects of abuse on mood after it occurs.

Another important finding in our study was that overall, TEs were associated mainly with depression, as shown in other BD studies (Bart et al., 2019; Farias et al., 2019; Shapero et al., 2017). However, when abuse was analyzed separately in our study, it was also associated with increased risk for hypo/mania. Other studies have also reported that abuse seems to pose a high risk for mania onset/recurrences and is associated with more severe manic symptoms compared to those without abuse (Du rocher Schudlich, 2015; Gilman et al., 2015), but other TEs were not evaluated. Although there are no clear explanations for this finding, abuse-exposed participants might have had more prior manic episodes than participants with other TEs, and abuse could contribute to increase subsequent mania risk (i.e., episode sensitization). Yet, analyses of prior episode polarities were not evaluated separately for abuse-exposed participants as compared to other participants with other TEs.

Although it is not known how TEs could affect the course and outcome of BD, there are several biological and psychosocial mechanisms that may explain this association. The central nervous system (CNS) is still maturing in adolescence; thus, TEs could affect its optimal development by interfering in affective regulation, attachment, and adaptation to environment (Cicchetti and Toth, 2005). Moreover, the early adversity sensitization (EAS) hypothesis suggests that early-TEs could have enduring effects on the catecholamine-stress-response (Otte et al., 2005), hyper-reactivity of corticoid systems (Heim and Nemeroff, 2001), and functioning of prefrontal cortex and hippocampus (Bremner, 2002). This could potentially lower the threshold for future mood episodes, leading to worsened BD course (Shapero et al., 2017). Further, individuals with mood disorders could actively contribute to their stressful environment (dependent events), which could be putatively associated with their behavior and could play a role in the precipitation of future episodes (stress generation model) or place them at risk for future re-victimization and TEs (Hammen, 2006). It is also possible that early-TEs could be associated with prodromal BD symptoms (i.e., irritability, hypersexuality), that could perpetuate this daunting cycle. Finally, stress affects sleep in childhood, potentially inducing biological changes in circadian rhythms, precipitating BD symptoms (Landgraf et al., 2014).

Among all new onset non-affective disorders examined, we found that SUD onset remained the only disorder predicted by abuse. Abuse-exposed participants had almost twice the risk of SUD onset and developed it at a younger age than those without abuse. Other studies have also reported that abuse was associated with SUD among bipolar youth/adults (Brown et al., 2005; Du rocher Schudlich, 2015; Maniglio, 2013). In contrast to our study, these studies were correlational and therefore, the direction of this association was not determined. Others have showed that childhood-onset-BD by itself poses a higher risk of SUD than adult-onset, although the role of abuse was not examined (Goldstein and Levitt, 2006; Wilens et al., 2016). The interplay between genetics (e.g. SUD family history), other stress systems (corticoid release expression), and the effects of early-abuse history in the brain could partially explain the association. Moreover, abuse survivors might use substances as a coping strategy in order to alleviate painful internal states (Roesler and Dafler, 1993). Specific mechanisms that mediate SUD onset among abuse-exposed BD youth should be further studied.

4.1. Limitations

The results presented should be considered in light of the following limitations. First, prospective data was gathered longitudinally in COBY and assessed retrospectively at every follow-up period encompassing an average of 7-months. While all reports of TEs are retrospective by nature, our design limited the interval between assessments to minimize retrospective recall. Second, TEs were ascertained with the TES, which contains a limited number of severe events, excluding minor, but significant events, or qualifying information on the dependence/independence on the participants’ behavior, or if the event was perceived as a threat/traumatic and its severity. Emotional abuse and neglect were not ascertained with the TES. Moreover, comparison of the effects of sexual abuse and physical abuse were not separately analyzed. However, a future paper will address this issue. Third, as a longitudinal phenomenology study, COBY did not recruit a control group. Fourth, COBY participants were recruited from different clinical settings and thus might not be representative of general population. Nonetheless, similar course/morbidity of BD youth has been observed in non-clinical populations (Lewinsohn et al., 2000). Fifth, participants were predominantly White; thus, findings might not be representative of cultural minority groups. Sixth, all participants had a BD diagnosis at the time of study entry. Thus, it was not possible to reliably ascertain whether the TEs occurred before/after the BD diagnosis. Finally, except for abuse, other TEs dates captured by the TES were approximate; thus, we used a conservative approach in limiting our discussion of causality regarding the TEs effects on mood symptoms course. Nevertheless, our results indicate that TEs, particularly abuse, could contribute to worse overall BD course and outcomes.

4.2. Conclusions

Our findings provide further evidence of the impact of severe lifetime TEs on the course and outcome of early-onset BD, showing that TEs, especially sexual/physical abuse and violence were associated with more severe mood symptomatology, increased risk of mood recurrences, SI, and increased rates of comorbid disorders across all follow-up. Further, the prospective nature of our study showed that after abuse occurrence, participants experienced worsening of their mood symptomatology and were more likely to develop new-onset SUD.

The high volume and frequency of COBY follow-ups allowed for TEs assessment in a way that minimized the retrospective nature of reporting, representing an important contribution given prior literature recall biases. History of TEs, and particularly abuse, appear to be a marker that identifies high risk patients indicating that trauma screening and early intervention is recommended to minimize or potentially prevent the associated consequences.

Given the effects of TEs on the course and outcome of BD youth, it is important to periodically assess for TEs following the recommendations described in prior TEs literature (Read, 2007). Moreover, although not specific for BD youth, more specific treatments for TEs such as Trauma-Focused-Cognitive-Behavioral Therapy, which is designed to improve family functioning and decrease interpersonal stress associated with TEs, should be considered (Cohen et al., 2004b). Finally, studies focusing on putative biological and psychosocial mechanisms of the effects of exposure to TEs to develop more specific interventions for BD youth are warranted.

Supplementary Material

1
2
3

Acknowledgments

Dr. Andreu Pascual has received a grant from the Alicia Koplowitz Foundation. Dr. Levenson receives salary support and grant funding from NICHD, grant funding American Academy of Sleep Medicine Foundation and the University of Pittsburgh, and royalties from American Psychological Association Book. Dr. Birmaher reports grants from NIMH, during the conduct of the study; royalties from Random House, Woltas Kluwer (UpToDate) and Lippincott, Williams & Wilkins, outside of the submitted work. Ms. Hower has received research support from NIMH, and honoraria from the Department of Defense (DOD). Dr. Yen has received research support from NIMH, NCCIH, and the American Foundation for Suicide Prevention. Dr. Strober has received research support from NIMH, and support from the Resnick Endowed Chair in Eating Disorders. Dr. T. Goldstein reports grants from NIMH, The American Foundation for Suicide Prevention, University of Pittsburgh Clinical and Translational Science Institute (CTSI) and The Brain and Behavior Foundation and royalties from Guilford Press, outside the submitted work. Dr. B Goldstein reports grants from Brain & Behavior Research Foundation, Brain Canada, Canadian Institutes of Health Research, Heart & Stroke Foundation, and the departments of psychiatry of Sunnybrook Health Sciences Centre and the University of Toronto Department of Psychiatry. Dr. Ryan reports grants from NIH. Dr. Weinstock has received research support from NIMH, NCCIH, the NIH OBSSR, and NIJ. Dr. Keller has received research support from NIMH and the John J. McDonnell and Margaret T. O’Brien Foundation. Dr. Axelson reports grants from NIMH, during the conduct of the study; personal fees from Janssen Research and Development, LLC, and UpToDate, outside the submitted work. Mr. Merranko, Ms. Gill report no financial relationships with commercial interests.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

TEs: Traumatic Events

References:

  1. Aas M, Henry C, Andreassen OA, Bellivier F, Melle I, Etain B, 2016. The role of childhood trauma in bipolar disorders. International journal of bipolar disorders 4, 2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Agnew-Blais J, Danese A, 2016. Childhood maltreatment and unfavourable clinical outcomes in bipolar disorder: a systematic review and meta-analysis. Lancet Psychiatry 3, 342–349. [DOI] [PubMed] [Google Scholar]
  3. Anand A, Koller DL, Lawson WB, Gershon ES, Nurnberger JI, Bi GSC, 2015. Genetic and childhood trauma interaction effect on age of onset in bipolar disorder: An exploratory analysis. Journal of affective disorders 179, 1–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Andersen PK GR, 1982. Cox’s Regression Model for Counting Processes: A Large Sample Study. Annals Statististics 10, 1100–1120. [Google Scholar]
  5. Association AP, 2004. The Diagnostic and Statistical Manual of Mental Disorders-IV (DSM-IV). American Psychiatric Association, Arlington, VA. [Google Scholar]
  6. Axelson D, Birmaher B, Strober M, Gill MK, Valeri S, Chiappetta L, Ryan N, Leonard H, Hunt J, Iyengar S, Bridge J, Keller M, 2006. Phenomenology of children and adolescents with bipolar spectrum disorders. Archives of general psychiatry 63, 1139–1148. [DOI] [PubMed] [Google Scholar]
  7. Axelson D, Birmaher BJ, Brent D, Wassick S, Hoover C, Bridge J, Ryan N, 2003. A preliminary study of the Kiddie Schedule for Affective Disorders and Schizophrenia for School-Age Children mania rating scale for children and adolescents. Journal of child and adolescent psychopharmacology 13, 463–470. [DOI] [PubMed] [Google Scholar]
  8. Axelson DA, Birmaher B, Strober MA, Goldstein BI, Ha W, Gill MK, Goldstein TR, Yen S, Hower H, Hunt JI, Liao F, Iyengar S, Dickstein D, Kim E, Ryan ND, Frankel E, Keller MB, 2011. Course of subthreshold bipolar disorder in youth: diagnostic progression from bipolar disorder not otherwise specified. Journal of the American Academy of Child and Adolescent Psychiatry 50, 1001–1016 e1003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Bart CP, Abramson LY, Alloy LB, 2019. Impulsivity and Behavior-Dependent Life Events Mediate the Relationship of Reward Sensitivity and Depression, but Not Hypomania, Among at-Risk Adolescents. Behav Ther 50, 531–543. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Birmaher B, 2018. Bipolar Disorders In: In: Martin A V, FR, editor. (Ed.), Lewis’s Child and Adolescent Psychiatry: A Comprehensive Textbook, 5th edition ed. Wolters Kluwer, Philadelphia. [Google Scholar]
  11. Birmaher B, Gill MK, Axelson DA, Goldstein BI, Goldstein TR, Yu H, Liao F, Iyengar S, Diler RS, Strober M, Hower H, Yen S, Hunt J, Merranko JA, Ryan ND, Keller MB, 2014. Longitudinal trajectories and associated baseline predictors in youths with bipolar spectrum disorders. The American journal of psychiatry 171, 990–999. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Boe T, Serlachius AS, Sivertsen B, Petrie KJ, Hysing M, 2018. Cumulative effects of negative life events and family stress on children’s mental health: the Bergen Child Study. Soc Psychiatry Psychiatr Epidemiol 53, 1–9. [DOI] [PubMed] [Google Scholar]
  13. Bremner JD, 2002. Neuroimaging of childhood trauma. Semin Clin Neuropsychiatry 7, 104–112. [DOI] [PubMed] [Google Scholar]
  14. Brown GR, McBride L, Bauer MS, Williford WO, Cooperative Studies Program 430 Study, T., 2005. Impact of childhood abuse on the course of bipolar disorder: a replication study in U.S. veterans. Journal of affective disorders 89, 57–67. [DOI] [PubMed] [Google Scholar]
  15. Carbone EA, Pugliese V, Bruni A, Aloi M, Calabro G, Jaen-Moreno MJ, Segura-Garcia C, De Fazio P, 2019. Adverse childhood experiences and clinical severity in bipolar disorder and schizophrenia: A transdiagnostic two-step cluster analysis. Journal of affective disorders 259, 104–111. [DOI] [PubMed] [Google Scholar]
  16. Cazala F, Bauer IE, Meyer TD, Spiker DE, Kazimi IF, Zeni CP, Zunta-Soares GB, Soares JC, 2019. Correlates of childhood trauma in children and adolescents with bipolar disorder spectrum: A preliminary study. Journal of affective disorders 247, 114–119. [DOI] [PubMed] [Google Scholar]
  17. Chambers WJ, Puig-Antich J, Hirsch M, Paez P, Ambrosini PJ, Tabrizi MA, Davies M, 1985. The assessment of affective disorders in children and adolescents by semistructured interview. Test-retest reliability of the schedule for affective disorders and schizophrenia for school-age children, present episode version. Archives of general psychiatry 42, 696–702. [DOI] [PubMed] [Google Scholar]
  18. Cicchetti D, Toth SL, 2005. Child maltreatment. Annual review of clinical psychology 1, 409–438. [DOI] [PubMed] [Google Scholar]
  19. Cohen AN, Hammen C, Henry RM, Daley SE, 2004a. Effects of stress and social support on recurrence in bipolar disorder. Journal of affective disorders 82, 143–147. [DOI] [PubMed] [Google Scholar]
  20. Cohen JA, Deblinger E, Mannarino AP, Steer RA, 2004b. A multisite, randomized controlled trial for children with sexual abuse-related PTSD symptoms. Journal of the American Academy of Child and Adolescent Psychiatry 43, 393–402. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Conus P, Cotton S, Schimmelmann BG, Berk M, Daglas R, McGorry PD, Lambert M, 2010. Pretreatment and outcome correlates of past sexual and physical trauma in 118 bipolar I disorder patients with a first episode of psychotic mania. Bipolar disorders 12, 244–252. [DOI] [PubMed] [Google Scholar]
  22. Daglas R, Conus P, Cotton SM, Macneil CA, Hasty MK, Kader L, Berk M, Hallam KT, 2014. The impact of past direct-personal traumatic events on 12-month outcome in first episode psychotic mania: trauma and early psychotic mania. The Australian and New Zealand journal of psychiatry 48, 1017–1024. [DOI] [PubMed] [Google Scholar]
  23. Danese A, Pariante CM, Caspi A, Taylor A, Poulton R, 2007. Childhood maltreatment predicts adult inflammation in a life-course study. Proceedings of the National Academy of Sciences of the United States of America 104, 1319–1324. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Daruy-Filho L, Brietzke E, Lafer B, Grassi-Oliveira R, 2011. Childhood maltreatment and clinical outcomes of bipolar disorder. Acta psychiatrica Scandinavica 124, 427–434. [DOI] [PubMed] [Google Scholar]
  25. Du rocher Schudlich T, 2015. Physical and Sexual abuse in Early-onset bipolar disorder in youths receiving outpatient services: Frequent, but not specific. Journal of abnormal child psychology 43, 453–463. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Etain B, Aas M, Andreassen OA, Lorentzen S, Dieset I, Gard S, Kahn JP, Bellivier F, Leboyer M, Melle I, Henry C, 2013. Childhood trauma is associated with severe clinical characteristics of bipolar disorders. The Journal of clinical psychiatry 74, 991–998. [DOI] [PubMed] [Google Scholar]
  27. Farias CA, Cardoso TA, Mondin TC, Souza LDM, da Silva RA, Kapczinski F, Magalhaes P, Jansen K, 2019. Clinical outcomes and childhood trauma in bipolar disorder: A community sample of young adults. Psychiatry research 275, 228–232. [DOI] [PubMed] [Google Scholar]
  28. First, M.B.S. RL; Gibbon M; Williams JBW, 1996. Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I). American Psychiatry Publishing, Inc. [Google Scholar]
  29. Garno JL, Goldberg JF, Ramirez PM, Ritzler BA, 2005. Impact of childhood abuse on the clinical course of bipolar disorder. The British journal of psychiatry : the journal of mental science 186, 121–125. [DOI] [PubMed] [Google Scholar]
  30. Gershon A, Johnson SL, Miller I, 2013. Chronic stressors and trauma: prospective influences on the course of bipolar disorder. Psychological medicine 43, 2583–2592. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Gilman SE, Ni MY, Dunn EC, Breslau J, McLaughlin KA, Smoller JW, Perlis RH, 2015. Contributions of the social environment to first-onset and recurrent mania. Molecular psychiatry 20, 329–336. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Goldstein BI, Levitt AJ, 2006. Factors associated with temporal priority in comorbid bipolar I disorder and alcohol use disorders: Results from the national epidemiologic survey on alcohol and related conditions. The Journal of clinical psychiatry 67, 643–649. [DOI] [PubMed] [Google Scholar]
  33. Hammen C, 2006. Stress generation in depression: reflections on origins, research, and future directions. J Clin Psychol 62, 1065–1082. [DOI] [PubMed] [Google Scholar]
  34. Hammersley P, Dias A, Todd G, Bowen-Jones K, Reilly B, Bentall RP, 2003. Childhood trauma and hallucinations in bipolar affective disorder: preliminary investigation. The British journal of psychiatry : the journal of mental science 182, 543–547. [DOI] [PubMed] [Google Scholar]
  35. Hardt J, Rutter M, 2004. Validity of adult retrospective reports of adverse childhood experiences: review of the evidence. Journal of child psychology and psychiatry, and allied disciplines 45, 260–273. [DOI] [PubMed] [Google Scholar]
  36. Hastie T, Tibshirani R, Friedman J, 2009. The elements of statistical learning: prediction, inference and data mining. Springer-Verlag, New York. [Google Scholar]
  37. Heim C, Nemeroff CB, 2001. The role of childhood trauma in the neurobiology of mood and anxiety disorders: preclinical and clinical studies. Biological psychiatry 49, 1023–1039. [DOI] [PubMed] [Google Scholar]
  38. Hollingshead AB, 1975. Four Factor Index of Social Status. New Haven, CT: Yale University. [Google Scholar]
  39. Jeon HJ, Park JI, Fava M, Mischoulon D, Sohn JH, Seong S, Park JE, Yoo I, Cho MJ, 2014. Feelings of worthlessness, traumatic experience, and their comorbidity in relation to lifetime suicide attempt in community adults with major depressive disorder. Journal of affective disorders 166, 206–212. [DOI] [PubMed] [Google Scholar]
  40. Jones SH, Thornicroft G, Coffey M, Dunn G, 1995. A brief mental health outcome scale-reliability and validity of the Global Assessment of Functioning (GAF). The British journal of psychiatry : the journal of mental science 166, 654–659. [DOI] [PubMed] [Google Scholar]
  41. Kaufman J, Birmaher B, Brent D, Rao U, Flynn C, Moreci P, Williamson D, Ryan N, 1997. Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (K-SADS-PL): initial reliability and validity data. Journal of the American Academy of Child and Adolescent Psychiatry 36, 980–988. [DOI] [PubMed] [Google Scholar]
  42. Keller MB, Lavori PW, Friedman B, Nielsen E, Endicott J, McDonald-Scott P, Andreasen NC, 1987. The Longitudinal Interval Follow-up Evaluation. A comprehensive method for assessing outcome in prospective longitudinal studies. Archives of general psychiatry 44, 540–548. [DOI] [PubMed] [Google Scholar]
  43. Kim EY, Miklowitz DJ, Biuckians A, Mullen K, 2007. Life stress and the course of early-onset bipolar disorder. Journal of affective disorders 99, 37–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Landgraf D, McCarthy MJ, Welsh DK, 2014. Circadian clock and stress interactions in the molecular biology of psychiatric disorders. Curr Psychiatry Rep 16, 483. [DOI] [PubMed] [Google Scholar]
  45. Larsson S, Aas M, Klungsoyr O, Agartz I, Mork E, Steen NE, Barrett EA, Lagerberg TV, Rossberg JI, Melle I, Andreassen OA, Lorentzen S, 2013. Patterns of childhood adverse events are associated with clinical characteristics of bipolar disorder. BMC Psychiatry 13, 97. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Leclerc E, Mansur RB, Grassi-Oliveira R, Cordeiro Q, Kapczinski F, McIntyre RS, Brietzke E, 2018. The differential association between history of childhood sexual abuse and body mass index in early and late stages of bipolar disorder. Journal of affective disorders 227, 214–218. [DOI] [PubMed] [Google Scholar]
  47. Leverich GS, McElroy SL, Suppes T, Keck PE Jr., Denicoff KD, Nolen WA, Altshuler LL, Rush AJ, Kupka R, Frye MA, Autio KA, Post RM, 2002. Early physical and sexual abuse associated with an adverse course of bipolar illness. Biological psychiatry 51, 288–297. [DOI] [PubMed] [Google Scholar]
  48. Lewinsohn PM, Klein DN, Seeley JR, 2000. Bipolar disorder during adolescence and young adulthood in a community sample. Bipolar disorders 2, 281–293. [DOI] [PubMed] [Google Scholar]
  49. Maniglio R, 2013. The impact of child sexual abuse on the course of bipolar disorder: a systematic review. Bipolar disorders 15, 341–358. [DOI] [PubMed] [Google Scholar]
  50. Marchand WR, Wirth L, Simon C, 2005. Adverse life events and pediatric bipolar disorder in a community mental health setting. Community mental health journal 41, 67–75. [DOI] [PubMed] [Google Scholar]
  51. Neria Y, Bromet EJ, Carlson GA, Naz B, 2005. Assaultive trauma and illness course in psychotic bipolar disorder: findings from the Suffolk county mental health project. Acta psychiatrica Scandinavica 111, 380–383. [DOI] [PubMed] [Google Scholar]
  52. Nrugham L, Holen A, Sund AM, 2010. Associations between attempted suicide, violent life events, depressive symptoms, and resilience in adolescents and young adults. J Nerv Ment Dis 198, 131–136. [DOI] [PubMed] [Google Scholar]
  53. Otte C, Neylan TC, Pole N, Metzler T, Best S, Henn-Haase C, Yehuda R, Marmar CR, 2005. Association between childhood trauma and catecholamine response to psychological stress in police academy recruits. Biological psychiatry 57, 27–32. [DOI] [PubMed] [Google Scholar]
  54. Palmier-Claus JE, Berry K, Bucci S, Mansell W, Varese F, 2016. Relationship between childhood adversity and bipolar affective disorder: systematic review and meta-analysis. The British journal of psychiatry : the journal of mental science 209, 454–459.27758835 [Google Scholar]
  55. Peele P, 2004. Use of Medical and Behavioral Health Services by Adolescents With Bipolar Disorder. Psychiatric Services 55. [DOI] [PubMed] [Google Scholar]
  56. Poletti S, Vai B, Smeraldi E, Cavallaro R, Colombo C, Benedetti F, 2016. Adverse childhood experiences influence the detrimental effect of bipolar disorder and schizophrenia on cortico-limbic grey matter volumes. Journal of affective disorders 189, 290–297. [DOI] [PubMed] [Google Scholar]
  57. Read J, 2007. To ask, or not to ask, about abuse--New Zealand research. Am Psychol 62, 325–326; discussion 330–322. [DOI] [PubMed] [Google Scholar]
  58. Roesler TA, Dafler CE, 1993. Chemical dissociation in adults sexually victimized as children: alcohol and drug use in adult survivors. J Subst Abuse Treat 10, 537–543. [DOI] [PubMed] [Google Scholar]
  59. Romero S, Birmaher B, Axelson D, Goldstein T, Goldstein BI, Gill MK, Iosif AM, Strober MA, Hunt J, Esposito-Smythers C, Ryan ND, Leonard H, Keller M, 2009a. Prevalence and correlates of physical and sexual abuse in children and adolescents with bipolar disorder. Journal of affective disorders 112, 144–150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Romero S, Birmaher B, Axelson DA, Iosif AM, Williamson DE, Gill MK, Goldstein BI, Strober MA, Hunt J, Goldstein TR, Esposito-Smythers C, Iyengar S, Ryan ND, Keller M, 2009b. Negative life events in children and adolescents with bipolar disorder. The Journal of clinical psychiatry 70, 1452–1460. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Rucklidge JJ, 2006. Psychosocial functioning of adolescents with and without paediatric bipolar disorder. Journal of affective disorders 91, 181–188. [DOI] [PubMed] [Google Scholar]
  62. SAS, I.I, 2019. Inc. SAS Institute Cary, NC, USA. [Google Scholar]
  63. Savitz J, van der Merwe L, Stein DJ, Solms M, Ramesar R, 2007. Genotype and childhood sexual trauma moderate neurocognitive performance: a possible role for brain-derived neurotrophic factor and apolipoprotein E variants. Biological psychiatry 62, 391–399. [DOI] [PubMed] [Google Scholar]
  64. Shaffer D, Gould MS, Brasic J, Ambrosini P, Fisher P, Bird H, Aluwahlia S, 1983. A children’s global assessment scale (CGAS). Archives of general psychiatry 40, 1228–1231. [DOI] [PubMed] [Google Scholar]
  65. Shapero BG, Weiss RB, Burke TA, Boland EM, Abramson LY, Alloy LB, 2017. Kindling of Life Stress in Bipolar Disorder: Effects of Early Adversity. Behav Ther 48, 322–334. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Stevelink R, Abramovic L, Verkooijen S, Begemann MJH, Sommer IEC, Boks MP, Mandl RCW, van Haren NEM, Vinkers CH, 2018. Childhood abuse and white matter integrity in bipolar disorder patients and healthy controls. European neuropsychopharmacology : the journal of the European College of Neuropsychopharmacology 28, 807–817. [DOI] [PubMed] [Google Scholar]
  67. Strawn JR, Adler CM, Fleck DE, Hanseman D, Maue DK, Bitter S, Kraft EM, Geracioti TD, Strakowski SM, DelBello MP, 2010. Post-traumatic stress symptoms and trauma exposure in youth with first episode bipolar disorder. Early intervention in psychiatry 4, 169–173. [DOI] [PubMed] [Google Scholar]
  68. Taylor JL, Gotham KO, 2016. Cumulative life events, traumatic experiences, and psychiatric symptomatology in transition-aged youth with autism spectrum disorder. J Neurodev Disord 8, 28. [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Team, R.C., 2019. R: A language and environment for statistical computing. R Foundation for Statitstical Comuting Vienna, Austria. [Google Scholar]
  70. Tijssen MJ, Van Os J, Wittchen HU, Lieb R, Beesdo K, Wichers M, 2010. Risk factors predicting onset and persistence of subthreshold expression of bipolar psychopathology among youth from the community. Acta psychiatrica Scandinavica 122, 255–266. [DOI] [PubMed] [Google Scholar]
  71. Tillman R, Geller B, Nickelsburg MJ, Bolhofner K, Craney JL, DelBello MP, Wigh W, 2003. Life events in a prepubertal and early adolescent bipolar disorder phenotype compared to attention-deficit hyperactive and normal controls. Journal of child and adolescent psychopharmacology 13, 243–251. [DOI] [PubMed] [Google Scholar]
  72. Van Meter A, Moreira ALR, Youngstrom E, 2019. Updated Meta-Analysis of Epidemiologic Studies of Pediatric Bipolar Disorder. The Journal of clinical psychiatry 80. [DOI] [PubMed] [Google Scholar]
  73. Warshaw MG, Dyck I, Allsworth J, Stout RL, Keller MB, 2001. Maintaining reliability in a long-term psychiatric study: an ongoing inter-rater reliability monitoring program using the longitudinal interval follow-up evaluation. Journal of psychiatric research 35, 297–305. [DOI] [PubMed] [Google Scholar]
  74. Weissman MM, Wickramaratne P, Adams P, Wolk S, Verdeli H, Olfson M, 2000. Brief screening for family psychiatric history: the family history screen. Archives of general psychiatry 57, 675–682. [DOI] [PubMed] [Google Scholar]
  75. Widom CS, Raphael KG, DuMont KA, 2004. The case for prospective longitudinal studies in child maltreatment research: commentary on Dube, Williamson, Thompson, Felitti, and Anda (2004). Child Abuse Negl 28, 715–722. [DOI] [PubMed] [Google Scholar]
  76. Wilens TE, Biederman J, Martelon M, Zulauf C, Anderson JP, Carrellas NW, Yule A, Wozniak J, Fried R, Faraone SV, 2016. Further Evidence for Smoking and Substance Use Disorders in Youth With Bipolar Disorder and Comorbid Conduct Disorder. The Journal of clinical psychiatry 77, 1420–1427. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1598435-supplement-1.docx (28.6KB, docx)
2
NIHMS1598435-supplement-2.docx (19.5KB, docx)
3
NIHMS1598435-supplement-3.docx (16.6KB, docx)

RESOURCES