BackgroundTo conduct a meta-analysis to estimate the incidence of major depressive disorder (MDD) and bipolar disorder (BD) in first-degree relatives (FDRs) of probands affected by MDD or BD. The risk for MDD in FDR of BD probands and vice versa is also investigated. MethodsA systematic review of case-control and cohort studies, which were published between 1977 and 2012; reported relative risks (RR) or odd ratios (OR) or equivalent raw data; made an explicit distinction between MDD and BD; used operational diagnostic criteria; and reported systematic proband recruitment and ascertainment of relatives. Studies were obtained by electronic MEDLINE and EMBASE searches and hand-searching. Estimates were derived from pooled data using random effects methods. ResultsOf an initial sample of 241 articles, 22 were eligible for inclusion. For FDRs of one proband with MDD compared to healthy control probands, estimates for MDD were OR=2.14 (95% CI 1.72–2.67), increasing to OR=3.23 (95% CI 2.11–4.94) for two MDD probands. For FDRs of one BD proband compared to healthy control probands, estimates for BD were OR=7.92 (95% CI 2.45–25.61), and OR=6.58 (95% CI 2.64–16.43) for FDRs of two BD probands. ConclusionsThese findings support previously published data indicating strong familiality for both MDD and BD. Data will be useful in providing individuals with a family history of MDD or BPD with tailored risk estimates.