Abstract Objectives Compare reported rates of mood-shifts from major depression to mania/hypomania/mixed-states during antidepressant (AD)-treatment and rates of diagnostic change from major depressive disorder (MDD) to bipolar disorder (BPD). Methods Searching computerized literature databases, followed by summary analyses. Results In 51 reports of patients diagnosed with MDD and treated with an AD, the overall risk of mood-switching was 8.18% (7837/95,786) within 2.39±2.99 years of treatment, or 3.42 (95% CI: 3.34–3.50) %/year. Risk was 2.6 (CI: 2.5–2.8) times greater with/without AD-treatment by meta-analysis of 10 controlled trials. Risk increased with time up to 24 months of treatment, with no secular change (1968–2012). Incidence rates were 4.5 (CI: 4.1–4.8)-times greater among juveniles than adults (5.62/1.26 %/year; p<0.0001). In 12 studies the overall rate of new BPD-diagnoses was 3.29% (1928/56,754) within 5.38 years (0.61 [0.58–0.64] %/year), or 5.6-times lower (3.42/0.61) than annualized rates of mood-switching. Conclusions AD-treatment was associated with new mania-like responses in 8.18% of patients diagnosed with unipolar MDD. Contributions to mood-switching due to unrecognized BPD versus mood-elevating pharmacological effects, as well as quantitative associations between switching and later diagnosis of BPD not associated with AD-treatment remain uncertain. Limitations Rates and definitions of mood-switching with ADs varied greatly, exposure-times rarely were precisely defined, and there was little information on predictive associations between mood-switches and BPD-diagnosis.