Further evidence for robust familiality of pediatric bipolar I disorder: results from a very large controlled family study of pediatric bipolar I disorder and a meta-analysis.

OBJECTIVE To determine the risk for bipolar I disorder in first-degree relatives of children with DSM-IV bipolar I disorder via meta-analysis and expanded controlled study. DATA SOURCES AND EXTRACTION For the meta-analysis, PubMed was searched for scientific articles published in the world literature in English through 2011. The keywords searched were bipolar disorder, first-degree relatives, family study, and control. All online abstracts were reviewed, and relevant full manuscripts were collected and reviewed. Citations were also examined for other potentially relevant articles. The analysis included only controlled family studies that examined rates of bipolar I disorder in all first-degree relatives (parents and siblings) of pediatric bipolar I probands and that had age- and sex-matched controls. Family history studies were excluded, as were studies that were not in English, did not report bipolar I rates for all first-degree relatives, or reported only bipolar spectrum rates. Also excluded were family studies that included only adult probands. A meta-analysis was conducted of the 5 controlled family studies of pediatric bipolar I probands that met the search criteria using the random-effects model of DerSimonian and Laird. METHOD For the family study, our previous sample of DSM-IV bipolar I probands was greatly expanded using structured diagnostic interviews. The new study included 239 children aged 6-17 years who satisfied full DSM-IV diagnostic criteria for bipolar I disorder (n = 726 first-degree relatives), 162 attention-deficit/hyperactivity disorder (ADHD) probands (without bipolar I disorder; n = 511 first-degree relatives), and 136 healthy control probands (without ADHD or bipolar I disorder; n = 411 first-degree relatives). The Kaplan-Meier cumulative failure function was used to calculate survival curves and cumulative lifetime risk in relatives. Cox proportional hazard models were used to calculate the risk of bipolar I disorder in relatives. RESULTS The pooled odds ratio for bipolar I disorder in relatives was estimated to be 6.96 (95% confidence interval [CI], 4.8-10.1). First-degree relatives of bipolar I probands were also significantly more likely than first-degree relatives of both ADHD probands (hazard ratio [HR] = 3.02; 95% CI, 1.85-4.93; P < .001) and control probands (HR = 2.83; 95% CI, 1.65-4.84; P < .001) to have bipolar I disorder. CONCLUSIONS Our results document an increased familial risk for bipolar I disorder in relatives of pediatric probands with DSM-IV bipolar I disorder.