Childhood maltreatment and unfavourable clinical outcomes in bipolar disorder: a systematic review and meta-analysis

Background Bipolar disorder affects up to one in 25 individuals and identification of early risk indicators of negative outcomes could facilitate early detection of patients with greatest clinical needs and risk. We aimed to investigate the association between childhood maltreatment and key negative outcomes in patients with bipolar disorder. Methods For this systematic review and meta-analysis we searched MEDLINE, PsycINFO, and Embase to identify articles published before Jan 1, 2015, examining the association of maltreatment (physical, sexual, or emotional abuse, neglect, or family conflict) before age 18 years with clinical features and course of illness in bipolar disorder. Data were extracted from published reports and any missing information was requested from investigators. We did 12 independent random-effects meta-analyses to quantify the associations between childhood maltreatment and course of illness or clinical features. Findings We initially identified 527 records and after unsuitable studies were removed, our search yielded 148 publications of which 30 were used in the meta-analysis. Patients with bipolar disorder and history of childhood maltreatment had greater mania severity (six studies, 780 participants; odds ratio [OR] 2.02, 95% CI 1.21-3.39, p=0.008), greater depression severity (eight studies, 1007 participants; 1.57, 1.25-1.99, p=0.0001), greater psychosis severity (seven studies, 1494 participants; 1.49, 1.10-2.04, p=0.011), higher risk of comorbidity with post-traumatic stress disorder (eight studies, 2494 participants; 3.60, 2.45-5.30, p<0.0001), anxiety disorders (seven studies, 5091 participants; 1.90, 1.39-2.61, p<0.0001), substance misuse disorders (11 studies, 5469 participants; 1.84, 1.41-2.39, p<0.0001), alcohol misuse disorder (eight studies, 5040 participants; 1.44, 1.13-1.83, p=0.003), earlier age of bipolar disorder onset (14 studies, 5733 participants; 1.85, 1.43-2.40, p<0.0001), higher risk of rapid cycling (eight studies, 3010 participants; 1.89, 1.45-2.48, p<0.0001), greater number of manic episodes (seven studies, 3909 participants; 1.26, 1.09-1.47, p=0.003), greater number of depressive episodes (eight studies, 4025 participants; 1.38, 1.07-1.79, p=0.013), and higher risk of suicide attempt (13 studies, 3422 participants; 2.25, 1.88-2.70, p<0.0001) compared with those with bipolar disorder without childhood maltreatment. Overall, these associations were not explained by publication bias, undue effects of individual studies, or variation in study quality. Interpretation Childhood maltreatment predicts unfavourable clinical features and course of illness in patients with bipolar disorder.