Age-at-onset in bipolar-I disorder: Mixture analysis of 1369 cases identifies three distinct clinical sub-groups

Biostatistics and Bioinformatics Unit, Wales School of Medicine, Cardiff University, Heath Park, Cardiff, CF14 4XN, United Kingdom.
Journal of Affective Disorders (Impact Factor: 3.38). 07/2009; 116(1-2):23-9. DOI: 10.1016/j.jad.2008.10.021
Source: PubMed


To assess whether bipolar disorder type I segregates into three clinically distinct sub-groups defined by age-at-onset.
Clinical data were available on 1369 individuals with DSM-IV bipolar I disorder. Mixture analysis was performed on the age-at-onset (AAO) data to determine whether they were composed of more than one normal distribution. Individuals were allocated to groups according to the results of the mixture analysis. Categorical logistic regression was then used to investigate relationships between AAO and nine clinical characteristics.
The distribution of AAOs in our sample comprised a mixture of three normal distributions with means of 18.7 (SD=3.7), 28.3 (SD=5.5) and 43.3 (SD=9.1) years, with relative proportions of 0.47, 0.39 and 0.14 respectively. Individuals were allocated into three groups dependent on their AAO: < or = 22; 25-37; and > or = 40 years, producing a sample of 1225 individuals (144 with borderline values were excluded). Eight out of the nine clinical characteristics showed evidence for a statistical association with AAO group.
Systematic and non-systematic recruitment of participants. Some data relied on retrospective recall.
Our results provide further robust evidence to suggest that the AAO distribution of individuals affected with bipolar disorder is composed of three normal distributions. Substantial clinical heterogeneity between the three AAO groups may reflect genetic heterogeneity within bipolar I disorder. Future genetic studies should consider AAO grouping as potential sub-phenotypes.

26 Reads
  • Source
    • "maximum = +0.001), thus ensuring that each individual had a different AAO (Hamshere et al., 2009). To compare our theoretical distribution replicate with the theoretical distribution replicate generated using the parameters of the other admixture analysis studies we used the online version of the 2-sample KS test ( "
    [Show abstract] [Hide abstract]
    ABSTRACT: A major obstacle to the identification of the neurobiological correlates of schizophrenia is the substantial diagnostic heterogeneity of this disorder. Dividing schizophrenia into "early" and "late" subtypes may reduce heterogeneity and facilitate identification of biomarkers related to this disease. Our objective was to assess the presence of different sub-groups in schizophrenia by age at onset analysis. The participants in this study were 612 unrelated patients with schizophrenia. Admixture analysis was applied in order to identify a model of separate normal distributions of age at onset characterized by different means, variances and population proportions to evaluate the effect of winter birth and ethnicity on early onset schizophrenia. The best-fitting model suggested three subgroups with means and standard deviations of 17.11±2.09, 21.96±3.43 and 30.02±7.1years, comprising 34.6%, 42.6% and 22.8% of the sample respectively. We considered as predictors of early onset schizophrenia: male gender, winter birth, white ethnicity and positive family history for psychiatric disorders. Earlier onset was significantly associated with male gender. We also compared our age at onset distribution with those published in other studies and we found significant differences with several studies suggesting heterogeneity in age at onset that is likely influenced by diagnostic heterogeneity in applying the DSM-IV criteria. Overall, our study showed that a typical early onset schizophrenia patient is more likely to be a white male with cannabis abuse and positive family history of psychiatric disorders. The heterogeneity in reporting age at onset across different studies suggests the application of more stringent criteria in diagnosing schizophrenia. Copyright © 2015 Elsevier B.V. All rights reserved.
    Schizophrenia Research 03/2015; 164(1-3). DOI:10.1016/j.schres.2015.03.004 · 3.92 Impact Factor
  • Source
    • "Nineteen studies involving 15,924 patients published since 2000 had at least one finding that was replicated in at least one other such report (Table 1). Among such studies, most (14/19, involving 9583 patients) were based outside of the United States, and used aggregated childhood/adolescent/ 7early adulthood as earlyonset groups that comprised 41.8% of these 9583 patients, with a mean onset age of 17.2 years (Azorin et al., 2013; Bellivier et al., 2001, 2003; Benazzi, 2009; Carter et al., 2003; Coryell et al., 2013; Ernst and Goldberg, 2004; Etain et al., 2012; Hamshere et al., 2009; Javaid et al., 2011; Lin et al., 2006; Manchia et al., 2008; Ortiz et al., 2011; Tozzi et al., 2011). In contrast, fewer reports (5/19, involving 6341 patients), primarily based in the United States, used separate childhood-and adolescent-onset groups that comprised 20.7% and 48.9%, respectively, of these 6341 patients, with mean onset ages of 8.6 and 13.2 years, respectively (Baldessarini et al., 2012; Leverich et al., 2007; Moor et al., 2012; Perlis et al., 2009; Rende et al., 2007). "
    [Show abstract] [Hide abstract]
    ABSTRACT: The strengths and limitations of considering childhood-and adolescent-onset bipolar disorder (BD) separately versus together remain to be established. We assessed this issue. BD patients referred to the Stanford Bipolar Disorder Clinic during 2000-2011 were assessed with the Systematic Treatment Enhancement Program for BD Affective Disorders Evaluation. Patients with childhood- and adolescent-onset were compared to those with adult-onset for 7 unfavorable bipolar illness characteristics with replicated associations with early-onset patients. Among 502 BD outpatients, those with childhood- (<13 years, N=110) and adolescent- (13-18 years, N=218) onset had significantly higher rates for 4/7 unfavorable illness characteristics, including lifetime comorbid anxiety disorder, at least ten lifetime mood episodes, lifetime alcohol use disorder, and prior suicide attempt, than those with adult-onset (>18 years, N=174). Childhood- but not adolescent-onset BD patients also had significantly higher rates of first-degree relative with mood disorder, lifetime substance use disorder, and rapid cycling in the prior year. Patients with pooled childhood/adolescent - compared to adult-onset had significantly higher rates for 5/7 of these unfavorable illness characteristics, while patients with childhood- compared to adolescent-onset had significantly higher rates for 4/7 of these unfavorable illness characteristics. Caucasian, insured, suburban, low substance abuse, American specialty clinic-referred sample limits generalizability. Onset age is based on retrospective recall. Childhood- compared to adolescent-onset BD was more robustly related to unfavorable bipolar illness characteristics, so pooling these groups attenuated such relationships. Further study is warranted to determine the extent to which adolescent-onset BD represents an intermediate phenotype between childhood- and adult-onset BD. Copyright © 2015 Elsevier B.V. All rights reserved.
    Journal of Affective Disorders 03/2015; 179:114-120. DOI:10.1016/j.jad.2015.03.019 · 3.38 Impact Factor
  • Source
    • "The emergence of bipolar disorder involves the interaction of complex genetic mechanisms (Burmeister et al., 2008; Craddock and Sklar, 2013; Petronis, 2003) and environmental factors (Tsuchiya et al., 2003). Based on 6 studies of 2509 patients with bipolar I disorder, the weighted mean age of onset falls into 3 groups, having peaks at ages 18.1, 26.9 and 42.7 years, with 55% of patients in the middle or late onset groups (Bellivier et al., 2001, 2003; González Pinto et al., 2009; Hamshere et al., 2009; Lin et al., 2006; Manchia et al., 2008). This broad range of onset and the polygenic basis of bipolar disorder suggest that environmental factors have an influential role (Burmeister et al., 2008; Craddock and Sklar, 2013; Wright et al., 2003). "
Show more