|Home | About | Journals | Submit | Contact Us | Français|
Anxiety disorders such as post-traumatic stress disorder (PTSD), and substance use disorders (SUD) are increasingly recognized as co-morbid disorders in children with bipolar affective disorder (BPD). This study explores the relationship between BPD, PTSD, and SUD in a cohort of BPD and non-BPD adolescents.
We studied 105 adolescents with BPD and 98 non-mood disordered adolescent controls. Psychiatric assessments were made using the Kiddie Schedule for Affective Disorders-Epidemiologic Version (KSADS-E). SUD was assessed by KSADS Substance Use module for subjects under 18 years, or Structured Clinical Interview for DSM-IV (SCID) module for Substance Use Disorders if age 18 or older.
Nine (8%) BPD subjects endorsed PTSD and nine (8%) BPD subjects endorsed subthreshold PTSD compared to one (1%) control subject endorsing full PTSD and two (2%) controls endorsing sub-threshold PTSD. Within BPD subjects endorsing PTSD, seven (39%) met criteria for SUD. Significantly more SUD was reported with full PTSD than with sub-threshold PTSD (χ2=5.58, p=0.02) or no PTSD (χ2=6.45, p=0.01). Within SUD, the order of onset was BPD, PTSD, SUD in three cases; while in two cases the order was PTSD, BD, SUD. The remaining two cases experienced coincident onset of BPD and SUD, which then led to trauma, after which they developed PTSD and worsening SUD.
An increased rate of PTSD was found in adolescents with BPD. Subjects with both PTSD and BPD developed significantly more subsequent SUD, with BPD, PTSD, then SUD being the most common order of onset. Follow-up studies need to be conducted to elucidate the course and causal relationship of BPD, PTSD and SUD.
Post-traumatic stress disorder (PTSD) in children is an increasingly studied disorder associated with substantial distress and morbidity (Giaconia 1995). While Breslau et. al. reported that 40–76% of children had been exposed to a traumatic event by age 17 (Breslau 2006; Giaconia 1995), the estimated lifetime prevalence of the full syndromatic PTSD in the general population is between 1 and 14% (Khantzian 2003), and its prevalence in children is around 6% (Giaconia 1995). Both genders are equally likely to experience trauma, though females are more likely to experience sexual abuse, and females are six times as likely to develop PTSD (Giaconia 1995).
Despite an increased awareness of PTSD in clinical child psychiatry, PTSD in children remains relatively underexamined. For example, children under the age of 15 years were not included in the DSM-IV field trials (Scheeringa 2006). Data also indicate that comorbid psychiatric disorders are common in PTSD (Giaconia 1995). Adolescents with alcohol dependence exhibited significantly higher rates of conduct disorder (CD), oppositional defiant disorder (ODD), attention deficit/hyperactivity disorder (ADHD), major depression, and PTSD than community controls without alcohol dependence (Clark 1997). Rates of alcohol and drug dependence, depression, and anxiety were substantially higher in youths who had a lifetime diagnosis of PTSD before age 18 compared to youths who had never experienced a trauma, and some what higher than youths who had experienced a trauma but did not develop PTSD (Giaconia 1995). Trauma and other adverse life events strongly correlate with alcohol use disorders (Clark 1997). Along the same line, pretrauma psychopathology remains a significant predictor of the development of PTSD in adults (Goldberg 2005). For instance, the existence of anxiety and mood disorders during childhood were associated with developing PTSD in relation to trauma. Among mood disorders in childhood, depression and bipolar disorder (BPD) are increasingly recognized as important clinical entities (Geller 2004).
BPD is often comorbid with other psychiatric disorders, including ADHD, ODD, CD, anxiety disorders, and substance use disorders (SUD), and comorbidity varies with age (Geller 2004; Wozniak 1999). In addition, recent data suggest that juvenile BPD also significantly increases the risk of substance use disorder (SUD, including drug and alcohol abuse or dependence) (Wilens 1997).
Recent work has begun to link PTSD and SUD in BPD adults. In relation to trauma history, Garno et al. (Garno 2005) reported higher risk for SUD, rapid cycling, and suicide attempts in individuals with BPD who had experienced severe childhood abuse. Those who had a history of severe childhood abuse had significantly younger age at onset of BPD, as well as more manic symptoms than those without trauma histories. However, little information is available regarding the interplay of BPD, SUD, and PTSD in adolescents.
Given the increasing recognition of SUD in BPD youth(Wilens 1997) and indications that BPD youth may be at high risk for trauma and PTSD, the relationship of these disorders needs to be examined. To this end, as part of an ongoing controlled longitudinal family study of BPD youth, we examined the relationship of BPD, SUD, and PTSD. Based on the literature we hypothesized that there would be an increased rate of PTSD in children with BPD compared to controls, and, that SUD rates would be higher in the BPD sample with PTSD compared to those without PTSD. In an exploratory manner, we examined the temporal relationship between PTSD, BPD, and SUD. To our knowledge, this is the first prospective controlled study of BPD adolescents to examine this association.
As part of an ongoing family genetic study of adolescent BPD, we ascertained 105 bipolar adolescent probands and 98 non-mood disordered control subjects. Subjects were recruited through newspaper advertisements, clinical referrals to our program (BPD only), internet postings, and internal postings within the Partners/Massachusetts General Hospital (MGH) system (Wilens 1997). Potential subjects were excluded if they had been adopted, or if their nuclear family was not available for study. We excluded youth if they had major sensorimotor handicaps (paralysis, deafness, blindness), autism, inadequate command of the English language, or a Full Scale IQ less than 70. Parents provided written informed consent for their children and children provided written assent to participate. The study was approved by the institutional review board at MGH.
A two-stage ascertainment procedure selected subjects. For BPD probands, the first stage confirmed the diagnosis of BPD by screening all children using a telephone questionnaire with their primary caregiver. The interviewer asked questions about symptoms of BPD and regarding study exclusion criteria. The second stage was the structured psychiatric interview as described below. Only subjects who received a positive diagnosis at both stages were included in the sample.
The diagnosis of BPD was based on a structured psychiatric interview using the Kiddie Schedule for Affective Disorders-Epidemiologic Version (KSADS-E) (Ambrosini 2000). Adolescents diagnosed with BPD type 1 had to have a clinical course characterized by the occurrence of at least one full manic or mixed episode not induced exclusively by substance or attributable to another psychotic illness. Mania was defined by severely impairing, distinct periods lasting more than one week of aberrant mood most of the day, consisting of abnormally elevated, expansive, or severely irritable mood (e.g., sustained rage) accompanied by at least three of the following (four if irritable mood): grandiosity, insomnia, pressured or irrational speech, flight of ideas, change in distractibility from baseline, extreme goal-directed activity, and excessive pleasurable activities/severely poor judgment. Youth diagnosed with BPD type II had a clinical course characterized by one or more major depressive episodes accompanied by at least one hypomanic episode not induced by substances or attributable to another psychotic illness. Hypomania was diagnosed when subjects had active manic mood and associated symptoms most of the day for at least four days, but did not meet duration criteria for mania, or if subjects did not manifest at least marked impairment during an episode (e.g. hospitalization). In all cases, families and/or subjects had to have endorsed moderate to severe impairment attributable to their mood symptoms. Overlap symptoms with ADHD (e.g. distractibility, talkativeness, fidgetiness) referable to BPD had to worsen substantially during a manic, hypomanic, or mixed manic phase to be considered positive.
We also screened potential non-mood-disordered controls in two stages. We eliminated any mood disorder from our ascertainment of controls secondary to concerns of “manic switching” from dysthymia or unipolar depression to BPD. First, control primary caregivers responded to the telephone questionnaire, then eligible controls meeting study entry criteria were recruited for the study and received the diagnostic assessment with a structured interview. Only subjects classified as not having any mood disorder at both stages were included in the control group.
All diagnostic assessments were made by raters with bachelor’s or master’s degree in a mental health field (e.g., psychology, social work, nursing) using DSM-IV based structured psychiatric interviews. Structured Clinical Interview for DSM-IV (SCID) was used to assess mood, anxiety, and substance use disorders in subjects age 18 or over, while the KSADS-E was used to assess all disorders in subjects under 18, as well as all disorders in subjects 18 or older that usually have their onset in childhood, such as ADHD. The interviewers were highly trained according to procedures currently in place for our ongoing studies. These raters make a minimum two-year commitment to work on this project to secure continuity. The interviewers undergo a rigorous training program lasting four months. Training includes mastery of the instruments, learning about DSM-IV criteria, watching training tapes, observing interviews performed by experienced raters, and rating several subjects under the supervision of the project coordinator. Throughout the study, the interviewers are supervised by board-certified child and adolescent psychiatrists and psychologists. This supervision includes weekly meetings, reliability assessment, presentation at diagnostic sign-off meetings, and additional consultations as needed.
Raters were blind to the ascertainment status of the probands. Psychiatric assessments were based on independent interviews with the parents and direct interviews of probands and siblings. For every diagnosis, information was gathered regarding the ages at onset and offset of full syndromatic criteria, and treatment history.
Substance use disorders (SUD) in our analyses included any alcohol or drug (excluding nicotine) abuse or dependence. Alcohol or drug abuse or dependence was diagnosed based on DSM-IV criteria using the KSADS-E Substance Use module, for subjects under 18, or SCID for subjects 18 and over. Recent evidence suggests the utility of structured interview data compared to objective data for ‘lifetime’ SUD determination (Gignac 2005). Rates of disorders reported are lifetime prevalence.
Diagnosis of PTSD was made by DSM-IV criteria, including characteristic symptoms of persistent re-experiencing of a trauma, persistent avoidance of reminders of the event, numbing of general responsiveness, and persistent symptoms of increased arousal. Trauma was defined by experiencing an event that arouses intense fear, helplessness, horror, and that includes real or threatened death, serious injury, or threat to the physical integrity of oneself or others. The diagnostic sign-off committee blindly reviewed all episodes of trauma, as well as all suspected cases of PTSD. Full PTSD was defined as meeting full syndromatic criteria for the disorder by DSM-IV criteria. Subthreshold PTSD was defined as having the cardinal symptoms of PTSD but not meeting full criteria, and broad PTSD was defined as manifesting either full or subthreshold PTSD.
All cases were presented to a committee composed of board certified child psychiatrists and psychologists. A diagnosis presented for review was considered positive only if the diagnosis would be considered clinically meaningful, i.e., a clinical concern due to the nature of the symptoms, the associated impairment, and the coherence of the clinical picture. Since self-report of psychopathology by juvenile probands may be discrepant from their parents’ reports, and recent data suggests the utility of report from both youth and parent for BPD, the diagnostic board assesses the information provided by each respondent based on interviewer notes, the responses to the structured interview, and a review of the audiotaped interview. Using these data, the rule for combining discrepant reports is to use the most severe diagnosis from any source unless the diagnosticians suspect that the source was not supplying reliable information. Diagnosis of CD was made based on DSM-IV symptomatocloy for CD, independent of substance use or SUD. Additionally, all cases of suspected drug or alcohol abuse or dependence were further reviewed with a child and adult psychiatrist with additional addiction credentials.
We computed coefficients of agreement by having 13 experienced, board certified child and adult psychiatrists and licensed clinical psychologists diagnose subjects from blinded audiotaped interviews made by our assessment staff. Based on 500 assessments from interviews by multiple interviewers of children and adults, the median coefficient was .98. Kappa coefficients for individual diagnoses included: ADHD (0.88), CD (1.0), ODD (.90), antisocial personality disorder (.80), major depression (1.0), mania (0.95), separation anxiety (1.0), agoraphobia (1.0), panic (.95), obsessive-compulsive disorder (OCD) (1.0), generalized anxiety disorder (GAD) (0.95), specific phobia (0.95). PTSD (1.0) social phobia (1.0), substance use disorder (1.0), and tics/Tourette’s (0.89). These measures indicated excellent reliability between ratings made by the non-clinician raters and experienced clinicians.
We further estimated the reliability of the diagnostic review process by computing coefficients of agreement between clinician reviewers. For these clinical diagnoses, the median reliability between individual clinicians and the review committee assigned diagnoses was .87. Kappa coefficients for individual diagnoses included: ADHD (1.0), CD (1.0), ODD (.90), ASPD (1.0), major depression (1.0), mania (0.78), separation anxiety (0.89), agoraphobia (.80), panic (.77), OCD (.73), GAD (.90), specific phobia (0.85), PTSD (0.8), social phobia (0.9), substance use disorder (1.0), and tics/Tourette’s (0.68).
We used Student’s t-test for continuous outcomes (e.g., age) and Pearson χ2 tests for dichotomous outcomes. We used logistic regression for dichotomous outcomes when controlling for other clinical characteristics. Independent variables in all regression analyses were dummy variables denoting group membership and any confounding demographic or psychiatric variables (e.g., comorbid disorders). We conducted all statistical analyses using Stata 9.2 (Stata corporation, College Station, TX, USA).
In all we had 105 BPD and 98 non-mood disordered controls. We found that significantly more BPD adolescents manifest PTSD compared to controls. Specifically in BPD youth, 8% had full PTSD and 8% had sub-threshold PTSD compared to controls, while non-mood disordered controls had a 1% rate of PTSD and 2% rate of sub-threshold PTSD [risk of broad PTSD based on BPD status: odds ratio (OR) =6.6, p=0.003, 95% confidence interval (CI): 1.88, 23.15]. Types of trauma included physical and sexual abuse (including rape), as well as witnessing death or intrafamilial violence. As previously reported (Wilens 2008) we found a significantly higher rate of SUD in BPD subjects compared to non-mood disordered controls (OR=7.4, p<0.001, 95% CI: 2.93 18.43).
We examined whether SUD was associated with PTSD in our BPD youth (see Table 1). Since age is significantly positively correlated with SUD in our sample (age by presence of SUD: t=8.8, df=106, p<0.001), we corrected for age in a logistic regression model using level of PTSD symptoms (none, subthreshold, and full symptomatic criteria) to predict SUD. Within our BPD youth, we found significant differences across three levels of PTSD symptoms in prevalence of SUD (omnibus test: χ2=6.7, p=0.03). Significantly more SUD was reported in subjects meeting full criteria for PTSD than in cases of subthreshold PTSD (χ2 =5.83, p=0.02) or no PTSD (χ2 =5.69, p=0.02). We repeated this analysis controlling individually for comorbid ADHD and CD, and then for both disorders. All results supporting this pattern remained significant when controlling for comorbid disruptive behavior disorders. The effect of ADHD was not significant in either case, whereas the effect of CD in predicting SUD was significant in both cases (independent of ADHD: p=0.035; with ADHD: p=0.031).
We further examined the temporal relationship of PTSD, BPD, and SUD. In 50% (9/18) of cases, BPD was diagnosed prior to PTSD, while in the other 50%, PTSD preceded the diagnosis of BPD. Starting with the SUD group, the sequential onset of BPD, PTSD, then SUD occurred in the majority of cases (44%). In 28% of subjects, the order of onset was PTSD, BPD, then SUD, while in another 28%, the simultaneous onset of BPD and SUD was followed by SUD-related trauma leading to PTSD and subsequently more severe SUD.
The most commonly used substance in this sample was alcohol (86 % of those with SUD), followed by marijuana (71%), and tobacco (29%). Two subjects used multiple substances including cough syrup, stimulants, opiates, nitrates, cocaine, and hallucinogens. Of note, within the SUD group, more females demonstrated both PTSD and SUD, while more males demonstrated PTSD without SUD (see Table 2).
In our BPD youth with PTSD, we also examined further psychiatric comorbidity. Increased rates of anxiety disorders were found in those individuals with PTSD (see Table 2). We found no association between CD and ADHD and broad-sense PTSD (either full and subthreshold cases) in our BPD subjects.
The results of these analyses show that as hypothesized, significantly more PTSD was seen in BPD compared to controls. Sixteen percent of youth with BPD had broad PTSD. Moreover, a higher risk of SUD was found in adolescents with BPD who also had PTSD. The rates of SUD were greater in those with full PTSD relative to those with subthreshold PTSD. While exploratory in nature, interesting temporal patterns relating BPD, PTSD, and SUD emerged. These data highlight the high rates of PTSD in BPD adolescents that appear to be related to SUD in these youth.
Our findings demonstrating higher risk for PTSD in BPD are similar to the literature (Garno 2005; Wozniak 1999). We found that 16% of our BPD youth had broad PTSD compared to 3% of non-mood-disordered controls. Our findings in adolescents mirror those in adults in adults; i.e., adults who experienced physical or sexual abuse in childhood or adolescence experienced an earlier age of onset of BPD, an increased number of multiaxial DSM disorders, and more SUD (Leverich 2002). These data add to the literature indicating that BPD is associated with PTSD; and that trauma and PTSD in BPD are associated with more adverse outcomes and a greater number of comorbid disorders.
We also found an increased risk for SUD associated with PTSD in our BPD youth. To our knowledge, this is the first report of the association of PTSD and SUD specifically in adolescents with BPD. High rates of SUD associated with PTSD have been reported in studies of adolescents with other mood disorders. For instance, Clark et al. (Clark 2003) found that major depressive disorder and PTSD were more frequent in alcohol-dependent adolescents than in control adolescents, with females with alcohol dependence having nearly twice as many depressive and PTSD symptoms as males. In a study by Kilpatrick et al. (Kilpartrick 2003) of 4,023 telephone-interviewed adolescents age 12–17, the prevalence of both PTSD and SUD was 0.7% in boys and 0.3% in girls. Variables significantly associated with increased risk were (i) age (OR = 1.4 per year increase); (ii) family history of alcohol use problems (OR=2.5 versus none); and (iii) witnessed violence (OR=9.0 versus none), sexual assault (OR=6.7 versus none), and physical assault (OR= 2.8 versus none) (Kilpatrick 2003). Similarly, Giaconia and colleagues (Giaconia 1995) found high risk of SUD in adolescents with PTSD.
Our findings linking SUD, PTSD, and BPD are similar to those reported in adults. Data suggests that up to 80% of adult women with SUD have experienced a major lifetime trauma, and 30–60% experience PTSD (Najavitis 1997). In the Australian National Comorbidity Study, Mills et al. (Mills 2006) reported that a significant minority of the population (0.5%) experience both PTSD and SUD, and that individuals with both disorders experience significantly greater disability than those with SUD alone. In adult male and female Veterans with BPD, childhood physical or sexual abuse reportedly led to PTSD and alcohol use disorders, with combined physical and sexual abuse leading to earlier age of onset of BPD (Brown 2005).
The temporal relationship of BPD, trauma, PTSD and SUD remains an area of uncertainty. Within our BPD youth without SUD, BPD preceded PTSD in 50% of cases, while PTSD preceded BPD in the remaining 50%. Within our BPD youth with SUD, subjects with preexisting BPD developed PTSD and then SUD. In other studies of mood disorders in adolescents, the onset of physical and sexual abuse was significantly younger than the onset of alcohol use or major depressive disorder (Clark 2003). Alcohol use disorder preceded major depressive disorder in the majority of subjects. Of interest, subjects who had been abused, either with or without alcohol use disorder, experienced much more severe depression than those who had not been abused, similar to work by Mills et al. (2006). One study found that SUD preceded PTSD (Cottler 1992), while another study reported that PTSD preceded SUD (Davidson 1985). Of those with both PTSD and SUD, the trauma occurred before the onset of SUD in 57% of cases, occurred simultaneously in 9%, and in 34% the substance use preceded the PTSD (Davidson 1985).
The precise temporal mechanism linking PTSD and SUD in BPD remains unclear. It is possible that at times BPD and its accompanying impulsivity place children at greater risk for trauma and subsequent PTSD and SUD. SUD in these cases may be related to the self-medication of co-existent anxiety and mood symptoms (Khantzian 2003). In other cases trauma may precipitate the development of symptoms of BPD, consistent with a stress-diathesis model (Mann 2005).
In terms of clinical implications, our findings highlight the heightened risk for PTSD for youth with BPD relative to non-mood disordered controls, and the greater risk of SUD for those BPD youth with PTSD than without PTSD. Given that a sizeable number of youth have the onset of PTSD and SUD after the diagnosis of BPD, treatment of BPD may result in prevention of trauma that may ultimately lead to a reduction in PTSD and SUD risk in these individuals, thus decreasing the negative outcomes associated with the comorbidity of these two disorders (Leverich 2002). Given the high rates of PTSD and SUD in adolescents with BPD, we recommend querying for PTSD as well as continuing to advocate for the aggressive identification of SUD.
There are a number of methodological issues with the analysis. Our non-mood disordered control group had relatively low rates of SUD and PTSD, limiting our comparisons. The rates of PTSD, and PTSD and SUD overall, were low, as were the rates of PTSD and SUD in our combined sample. While our control group was supernormal for any mood disorders, some of our control subjects had an anxiety disorder and thus may have been at risk for a mood disorder later in their course. All diagnoses were made by structured interview (KSADS-E, or SCID for SUD if the subject was 18 years or older) and did not use specific trauma-based instrumentation. While this was the first wave of assessments completed in the longitudinal follow-up study, the diagnosis often relied on retrospective recall of events, symptoms, and diagnosis. Based on our small sample size of affected subjects, we did not examine psychosocial, environmental, or familial factors that may have important contributions to the outcome. However the diagnosis was made through interview with both parent and child, and combined parent/child report has been shown to reveal almost twice the number of symptoms and markedly increase overall diagnosis than parent report alone (Scheeringa 2006). The study also did not differentiate gender or type of trauma.
Despite these limitations, the study’s findings show that BPD places a child at increased risk for the development of PTSD. Full or subthreshold PTSD in adolescents with BPD increases the risk for SUD. In BPD children with PTSD and SUD, BPD either precedes or is coincident with the onset of PTSD, followed by the development of SUD. These findings indicate a need to identify youth with BPD and implement early interventions in order to prevent the development of PTSD and/or SUD.
Funding Source: This study was financially supported by NIH RO1 DA12945 (TW) and K24 DA016264 (TW).
Disclosures: Dr. Steinbuchel, with Children’s Hospital and Research Center of Oakland, has no financial relationships to disclose.
Dr. Wilens, with Massachusetts General Hospital, receives support from Abbott, Alza/Ortho-McNeil, Cephalon, GlaxoSmithKline, Janssen, Eli Lilly, the National Institute on Drug Abuse, National Institute of Mental Health, NICMH, Neurosearch, Novartis, Pfizer, Saegis, Sanofi-Sythelabo, Shire.
Mr. Joel Adamson, statistician at Massachusetts General Hospital, has no financial relationships to disclose.
Ms. Sgambati, with Massachusetts General Hospital, has no financial relationships to disclose.