|Home | About | Journals | Submit | Contact Us | Français|
To compare the most severe lifetime (current or past) mood symptoms, duration of illness, and rates of lifetime comorbid disorders among youth with Bipolar (BP) spectrum disorders.
173 children with BP (<12 y.o), 101 adolescents with childhood-onset BP, and 90 adolescents with adolescent-onset BP were evaluated with standardized instruments.
Depression was the most common initial and frequent episode for both adolescent groups followed by mania/hypomania. Adolescents with childhood-onset BP had the longest illness, followed by children and then adolescents with adolescent-onset BP. Adjusting for sex, socio-economic status, and duration of illness, while manic both adolescent groups showed more “typical” and severe manic symptoms. Mood lability was more frequent in both childhood-onset groups. While depressed, both adolescent groups showed more severe depressive symptoms, higher rates of melancholic and atypical symptoms, and suicide attempts than children. Depressed children had more severe irritability than depressed adolescents. Early BP-onset was associated with ADHD whereas later BP-onset was associated with panic, conduct, and substance use disorders. Above-noted results were similar when each BP subtype was analyzed separately.
Older age was associated with more severe and typical mood symptomatology. However, there were differences and similarities in type, intensity, and frequency of BP symptoms and comorbid disorders related to age-of-onset and duration of BP and level of psychosocial development. These factors and the normal difficulties youth have expressing and modulating their emotions may explain existing complexities in diagnosing and treating BP in youth, particular in young children, and suggest the need for developmentally sensitive treatments.
Although it is accepted that children and adolescents may develop bipolar disorder (BP), there is considerable variability in the frequency of presentation of the manic symptoms across the extant studies, particularly for elation and irritability.1–7 These diverse clinical presentations between children and adolescents may be attributable, at least in part, to the psychosocial and maturational changes that occur after puberty8, 9. Furthermore, the age at which the child or adolescent first experiences symptoms of illness may also have important implications for clinical presentation with early deficits promoting later impairment as the child arrives at each progressive stage of development with inadequate resources available to meet the challenges unique to the ensuing period 10. Additionally, children and adolescents whose illness started during childhood may be more likely to be exposed to stressful life events and diverse types of treatment which may modify the phenomenology and course of their illness.
To date, few studies have compared the symptoms of mania and comorbid disorders between children and adolescents with BP, and no study has compared the BP depressive symptomatology across age groups. Faraone and colleagues 11 compared manic symptoms among a group of children with BP-I (n=68) with adolescents with early-onset (n=25) and adolescents with late-onset mania (n=17). Euphoria was more common in the adolescents with early-onset mania and irritability was least common in the adolescents with late-onset mania. Increased energy was twice as common in both groups with early-onset mania as compared with the late onset manic group. Except for higher rates of ADHD in children with BP, Geller and colleagues 12 did not find any differences in the rates of manic symptoms in a sample of children (n=53) and young adolescents (n= 40) with BP-I. Except for higher rates of substance abuse in adolescents, Findling and colleagues 3 did not find any differences in manic symptomatology or psychosis between children (N= 56) and adolescents (n=34) with BP-I. Finally, Masi and colleagues 13 reported more comorbid ADHD and ODD and a less episodic course in children (N=80) as compared to adolescents (n=56) with BP spectrum disorders. There were no differences between groups in rates of irritability, elation, or severity of illness.
The above-noted studies seem to indicate that, with few exceptions, symptoms of mania are equally manifested in children and adolescents. However, generalizability of these findings is limited because these studies have one or more of the following limitations: small samples, selection of only youth with BP-I, lack of direct interviews with children, absence of analysis of depressive symptomatology, and lack of stratification of the adolescent group by age at illness onset. In addition, none of these studies adjusted for potentially important confounding factors such as demographic factors (e.g., sex, socio-economic status), duration of illness, and presence of comorbid disorders.
The goal of this study was to replicate and extend the above-noted studies in a large sample of children and adolescents with BP spectrum disorders recruited into the Course and Outcome of Bipolar Youth (COBY) multicenter study. In prior publications we reported the clinical picture for the entire age group focusing on the clinical differences between BP-I, BP-II and BP-Not Otherwise Specified (NOS) 5 and psychosis 14 Also, a prior COBY report showed that after adjusting for significant demographic and clinical factors, BP children and adolescents with childhood-onset BP had higher rates of psychopathology in their 1st and 2nd degree biological relatives as compared to adolescents with adolescent-onset BP (≥ 12 y.o) 15. The present study compares the most severe lifetime (current or past) manic and depressive symptoms, other clinical variables, and rates of comorbid disorders among BP youth.
Briefly, after Institutional Review Board approval and after obtaining appropriate consent or assent, children and adolescents fulfilling DSM-IV criteria for BP-I, BP-II, or COBY’s criteria for bipolar disorder Not-Otherwise-Specified (NOS)5 were enrolled at Brown University, University of California Los Angeles, and the University of Pittsburgh Medical Center.
About 70% of the sample was recruited from outpatient clinics and the rest through advertisement and inpatient units. Subjects with current or lifetime DSM-IV diagnoses of schizophrenia, mental retardation, severe autistic spectrum disorders or mood disorders due to substance abuse, a medical condition, or secondary to use of medications were excluded.
In order to evaluate the differential clinical picture between children and adolescents with BP, the sample was divided in 3 subgroups according to onset of significant mood symptomatology and the age at the time of the initial assessment as follows: 1) Children with BP (<12 y.o) (n=173); 2) adolescents with childhood-onset BP (age ≥ 12 and BP onset <12) (n=101); and 3) adolescents with adolescent-onset BP (age ≥ 12 and BP onset ≥ 12) (n= 90). Age of BP-onset was defined as the age of onset of clinically significant manic, hypomanic, and/or depressive symptoms that affected the child’s functioning.5 Illness duration was defined as the subject’s current age minus the age of BP onset. In about 6% (22/364) of cases, parents reported that their children’s mood symptoms (mania, hypomania, and/or depression) appeared ≤ 4 years old. However, given the controversies diagnosing BP in very young children, the minimum age of onset for BP-spectrum illness was arbitrarily set at age 4.5
Since the DSM definition of BP-NOS is not operationalized, BP-NOS was defined as elated mood plus two associated DSM symptoms or irritable mood plus three associated DSM symptoms and a change in functioning. In addition, the above-noted symptoms were required to last a minimum of 4 hours within a 24-hour period and to have occurred at least 4 cumulative lifetime days. Prior papers using the COBY sample showed that children with BP-NOS have similar but less severe clinical picture, comorbid disorders, family history, and longitudinal outcome than the BP-I subjects. Moreover, about 25% of the youth diagnosed as BP-NOS converted into BP-I or BP-II. 5, 16
Children and adolescents and parents (about their children) were directly interviewed for the presence of current and lifetime non-mood psychiatric disorders using the Schedule for Affective Disorders and Schizophrenia for School Age Children, Present and Lifetime Version (K-SADS-PL). 17, 18. As the KSADS-PL only ascertains the presence or absence of symptoms, the K-SADS Mania Rating Scale (K-MRS) 19 and the depression section of the KSADS-P 20 were used to assess the severity of each mood symptom in detail (see Table 2 and and33 for the symptoms included in these scales). Mood symptoms that were in common with other psychiatric disorders (e.g., distractibility) were not rated as present unless they intensified with the onset of abnormal mood. Lifetime comorbid diagnoses were not assigned if they occurred exclusively during a mood episode.
A current mood episode was defined as the worst week of the month prior to intake. Past mood episodes were identified using a mood time line with the child and parents and asking them which of these episodes (mania and/or depression) was the most severe. Duration of the episode, level of impairment, and/or hospitalization were some of the factors considered when determining which past mood episode was the worst. Manic and depressive symptoms with the higher scores of either the current or past mood episode as rated in the MRS and depression section of the KSADS-P were selected for the analyses in this paper. We chose to evaluate the most severe lifetime manic and depressive symptoms in order to capture the illness when manifested at its full expression and not when only mild to moderate symptoms were present. Moreover, this strategy increased the likelihood that parents and children would be better able to remember specific mood symptomatology.
All assessments were completed by research staff trained to reliably administer the above-noted interviews and presented to child psychiatrists/psychologists who confirmed the diagnoses. The overall KSADS kappas for psychiatric disorders were ≥ 0.8. The intraclass correlation coefficients (ICC) for the K-MRS and the K-DEP total scores were ≥ 0.95.
Differences in demographics and clinical symptomatology among the three BP groups were analyzed using standard parametric (e.g., ANOVA) and non-parametric univariate tests (e.g., Kruskal-Wallis) when appropriate.
Generalized Linear Modeling (GLM) was used to evaluate the effects of group, and any significant between group differences in demographics, duration of illness, and other clinical variables.
Age and pubertal status were highly correlated in the three age-of-onset BP groups (rs ≥ .73, ps<0.001), therefore analyses included age only.
Analyses examining continuous and categorical data from the K-MRS and K-SADS depression items yielded similar results. Also, results were similar when each BP subtype was analyzed separately. Thus, for simplicity, only the results using continuous data and the combined BP data are included in this paper.
To inform new hypotheses and guide future studies the data is presented with (indicated in the tables with a symbol (†) and without Bonferroni corrections. Furthermore, to have an idea of the statistical strength of the comparisons, size effects (d) were calculated as described by Cohen23 and included in all tables.
All values are reported as means ± standard deviations. All p-values are based on two-tailed tests with α = 0.05.
As expected, the three groups significantly differed in age and pubertal status. Children with BP included significantly more boys than the other two groups and the SES of both childhood-onset groups was significantly lower than the adolescents with adolescent-onset BP. Adolescents with childhood-onset BP had the longest duration of illness, followed by children and then adolescents with adolescent-onset BP. The onset of significant mood symptomatology preceded the onset of episodes an average of 1.0 ± 1.7 years. Children experienced their first mood episode at an earlier age followed by adolescents with childhood-onset and adolescents with adolescent-onset BP. Adolescents with childhood-onset BP had significantly more at least one depressive episode than the adolescents with adolescents-onset BP followed by the children. However, differences between the two adolescent groups disappeared when adjusting for duration of illness. Both adolescents groups were more likely to have a major depressive episode as their first BP episode and both childhood-onset groups more commonly presented with initial symptoms meeting COBY’s criteria for BP-NOS than adolescents with adolescent-onset BP.
Given the above results, all following analyses were adjusted for sex, SES, duration of illness, and rates of comorbid disorders (see results below).
Adjusting for confounding factors, both adolescent groups showed significantly higher K-MRS-13-item total manic scores and more severe grandiosity, racing thoughts, poor judgment, and increased productivity as compared to children. Adolescents with adolescent-onset BP had more severe elation, goal directed activity, racing thoughts, and high energy than both children and adolescents with childhood-onset BP and more decreased need for sleep, accelerated speech, delusions and sharpened thinking compared to children. Finally, both childhood-onset groups had significantly more severe mood lability as compared to adolescents with adolescent-onset BP.
Adjusting for confounding variables both adolescent groups had significantly higher K-Depression-12-item total scores and more severe hopelessness, anhedonia, fatigue, difficult concentration, psychomotor retardation, circadian reversal, non-restorative sleep, daytime sleep, hypersomnia, weight lost, leaden paralysis, and medical lethality of suicidal attempts as compared to children. In addition, adolescents with adolescent-onset BP showed more severe depressed mood, excessive guilt, social withdrawal, decreased appetite, rejection sensitivity, suicidal ideation, and seriousness of suicidal attempts as compared with children and more terminal insomnia than both childhood-onset groups. Adolescents with childhood-onset BP also exhibited significantly more negative self-image and recurrent thoughts of death than the adolescents with adolescent-onset BP and children. Finally, children had more severe irritability/anger than the adolescents with adolescent-onset BP.
Adjusting for any confounding factors, both adolescent groups had higher prevalence of lifetime conduct and substance use disorders than children. Also, adolescents with adolescent-onset BP had more substance abuse than adolescents with childhood-onset BP and more panic disorder than children. Both childhood-onset groups had more lifetime ADHD than did the adolescents with adolescent-onset BP.
In this large sample of youth with BP spectrum disorders, both adolescent groups had their first episodes meeting criteria for major depression, and were more likely to have a lifetime history of a major depressive episode. In contrast, younger children were more likely to have had subsyndromal manic/hypomanic symptoms as their first episode. After adjusting for sex, SES, and duration of illness, more “typical” and severe symptomatology was observed in both adolescent groups while manic, especially those with adolescent-onset BP. In contrast, mood lability was more frequent in children and adolescents with childhood-onset BP than adolescents with adolescent-onset BP. While depressed, also both adolescent groups showed more severe depressive symptoms and higher rates of melancholic, atypical depressive symptoms, and suicide attempts than children with BP. Irritability was the only symptom that was moderately higher in depressed children when compared with adolescents with adolescent-onset BP.
As expected, earlier-onset of BP was most commonly associated with lifetime comorbid ADHD. In contrast, older age was associated with higher rates of lifetime panic, conduct, and substance use disorders.
Before discussing the above-noted results it is important to note the limitations of this study. First, recollection bias may have influenced the reported rate of symptoms. However, to minimize this problem this study focused on the worst lifetime (present or past) manic and depressive symptoms. Second, the frequency and quality of the symptoms may change from one episode to the next and may vary with the reporter (child vs. parent) 24. Future papers using prospective longitudinal data will address the two above-noted issues. Third, the base rates of some symptoms (e.g., lifetime delusions) were low. Fourth, the data analyzed for this report included only the worst lifetime (current or past) symptoms. These symptoms did not necessarily represent the mood symptomatology experienced by the child during her/his first mood episodes. Fifth, as depicted in tables 2 and and3,3, some symptoms, although significantly different, showed low effect sizes. Nevertheless, even if these symptoms are not taken into account or after adjusting for multiple comparisons, the main story of this paper holds: older age was associated with more typical and severe manic and depressive symptoms. Sixth, lifetime comorbid disorders were not necessarily present during the same time period as the worst mood symptoms. Nevertheless, after adjusting for duration of illness, the types and rates of comorbid disorders follow the expected developmental trajectory (e.g., more ADHD in childhood-onset BP and more SUD in adolescents). Finally, since most subjects included in this study were recruited from outpatient clinics and were Caucasian, the results of this study may not apply to other populations.
It appears that mood lability, as defined in this study by the K-MRS19 (“rapid mood variation with several mood states within a brief period of time which appears internally driven without regard to the circumstances”) and to a lesser degree irritability/anger, are more characteristic of childhood-onset rather than adolescent-onset mania even when adjusting for potential confounding factors such as comorbid disorders. Mood “swings” or lability is often described by parents of children with BP and has been reported in other cross-sectional and longitudinal prospective pediatric BP studies 16,2–5, 13. However, it is important to note that these mood changes are not equivalent to the DSM “rapid cycling” classification and that not every child with severe mood lability, especially if the symptoms are not episodic, has BP. 4, 6, 7, 9, 25
In contrast with prior reports 11, 26, 27, this study found that DSM manic symptoms were more severe in adolescents, particularly in those with adolescent-onset BP, than in children with BP. The reasons for this discrepancy are unclear, but it may be accounted for by methodological differences such as sample size, subject age, instruments used to ascertain symptoms, diagnostic criteria and thresholds used to classify a symptom as positive and counting symptoms that overlap between disorders as positive for two or more disorders without taking into account the overall clinical picture of the child. It is also possible that early-onset BP has a different trajectory than adolescent-onset BP which appears to be more similar to adult BP. Ongoing prospective follow-up of this early-onset group potentially will answer this question.
Speculatively, the greater prominence of cognitive elements of mania such as grandiosity may simply reflect the more advanced cognitive operations that are normatively associated with adolescent development. Also, differentiating abnormal grandiosity from normal fantasy is less difficult in adolescents that in younger children 7. The presence of these symptoms and the increased ability of an adolescent to express his or her affective state as compared to a younger child may explain, at least in part, why it is easier and more acceptable to diagnose BP in adolescents than in children. However, it remains for future more refined neuropsychological research on prospectively observed samples to determine if very early onset of illness compromises the acquisition of more complex cognitive structures which in turn impact on the phenomenology of BP.
Both the duration of illness and the transition into adolescence affected the rates and severity of depressive symptoms with both adolescent groups having their first episode as major depression, higher rates of lifetime major depressive episodes and more melancholic and atypical depressive symptomatology and suicidality when compared to children. In contrast, children showed less severe depressive symptomatology and more irritability. Thus, similar to the unipolar depression literature, the expression of major depression seems to be related to developmental pubertal changes and early age of onset 28–32. There is only one other study that carried out a post-hoc analysis of the symptoms of BP depression in a sample of youth with BP depression and ADHD compared to youth with unipolar depression and ADHD. 33 This study reported higher familial loading for mood disorders and more severe symptomatology, suicide behaviors, anxiety, and conduct disorders in the depressed BP/ADHD youth.
The rates of lifetime comorbid disorders reported in this study are consistent with developmental correlates of psychopathological disorder with childhood-onset BP having higher rates of ADHD and adolescents showing higher rates of substance abuse, and conduct disorders than children. Also, comparable to other pediatric and adult studies of BP anxiety disorders were common across groups, with panic disorder being found most frequently in adolescents 34–36. Interestingly, the rate of comorbid disorders in this study tended to be lower than those reported in some studies.4 This may be due to fact that in COBY, a comorbid diagnosis was not assigned if overlapping symptoms presented only during a mood disorder episode.
This and other studies have reported that early-onset BP has a somewhat different phenotype, higher familial load for mood disorders, and a different picture of comorbid disorders. 7, 9, 26, 37–39 These findings have suggested that early age of onset may carry a different genotype for BP that differentially influences disease phenotype, course, patterns of comorbidity, and functional outcome. However, we cannot determine from this analysis whether the differences found between child and adolescent age of onset subgroups are an independent effect of age when illness begins, or affected by susceptibility genes that confer early age of onset. However, interestingly, studies of adult BP have noted heterogeneity of linkage associated with psychotic features and anxiety 40, as well as familiality of a range of phenotypic characteristics in families with BP 41. Whatever the reason for these differences, it will be crucial for ongoing molecular, genetic, neuroimaging, and treatment studies to take into account the age of BP onset, maturational effects, as well as the presence of comorbid disorders and differential exposure to stress to fully explicate research findings.
In summary, older age was associated with more severe and typical mood symptomatology. However, there are differences and similarities in the type, intensity, and frequency of BP symptoms as well as comorbid disorders in youth depending, at least in part, on the age of onset and duration of their BP and their level of psychosocial development. These factors together with the normal difficulties children have expressing and modulating their emotions may explain the existing complexities in diagnosing and treating BP in this population and suggest the need for developmentally sensitive pharmacological and psychosocial treatments.
The presence of comorbid disorders and sometimes their respective treatments may further complicate the identification and treatment of BP and be associated with worse prognosis 42, 43. Thus, it is critical that these disorders are identified and differentiated from the symptoms of BP, and if appropriate, promptly treated. For example, clinicians should be aware of the possibility that an adolescent with childhood-onset BP may also have ADHD since some symptoms of ADHD tend to improve with age (e.g., hyperactivity). Furthermore, the selective serotonin reuptake inhibitors are efficacious for the treatment of anxiety disorders in youth, but they may induce mood destabilization in youth with BP.
Clinicians should be aware that BP, particularly in adolescents, often initially presents with an episode of major depression posing an important clinical dilemma since the treatment of these youth with antidepressants may worsen their clinical outcome. In these cases the presence of psychosis, pharmacologically induced mania/hypomania, and high familial loading for BP may indicate increased risk to develop BP 44–46. Finally, after puberty the frequency and severity of depression as well as potentially lethal suicide behaviors and substance abuse become more common, indicating the need for early identification and treatment of this disorder to avoid its serious consequences.
The project described was supported by Grant Numbers MH59929 (Dr. Birmaher), MH59977 (Dr. Strober) and MH59691 (Dr. Keller) from the National Institute of Mental Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Mental Health or the National Institutes of Health.
The authors would like to thank Carol Kostek and Editha Nottelmann Ph.D. for their support and Mathew Arruda BA; Amy Broz AS; Jennifer Fretwell BA; Colleen Grimm BA; Risha Henry PhD; Heather Hower MEd; Norman Kim PhD; Marguerite Lee BS; Stephanie Pincince, BA; Nadia Olszanski BA; Nicole Ryan BA; Jeff Ryan BA; Heather Schwickrath MA; Andrea Wolford, BA.
This paper is dedicated to Henrietta Leonard MD
The following authors have no conflicts of interest or financial disclosures: Boris Birmaher MD, David Axelson, MD; Michael Strober PhD, Mary Kay Gill MSN, Mei Yang MS, Neal Ryan MD, Benjamin Goldstein MD, Jeffrey Hunt MD, Christianne Esposito-Smythers PhD, Satish Iyengar PhD, Tina Goldstein PhD, Laurel Chiappetta BS, and Henrietta Leonard MD
The following disclosures are made by Martin Keller MD: Consultant/Honoraria: CENEREX, Cephalon, Cypress Bioscience, Cyberonics, Forest Laboratories, Janssen, JDS, Organon, Novartis, Pfizer, Wyeth. Advisory Boards: Abbott Laboratories, Bristol-Myers Squibb, CENEREX, Cyberonics, Cypress Bioscience, Forest Laboratories, Janssen, Novartis, Oranon, Pfizer. Grants/Research: Pfizer