Objective: Bipolar disorder (BD) is common in clinical psychiatric practice, and several studies have estimated its prevalence to range from 0.5 to 5% in community-based samples. However, no systematic review and meta-analysis of the prevalence of BD type 1 and type 2 has been published in the literature. We carried out a systematic review and meta-analysis of the lifetime and 1-year prevalence of BD type 1 and type 2 and assessed whether the prevalence of BD changed according to the diagnostic criteria adopted (DSM-III, DSM-III-R vs. DSM-IV). Methods: We searched MEDLINE, Scopus, Web of Science, PsycINFO, and the reference lists of identified studies. The analyses included 25 population-or community-based studies and 276,221 participants. Results: The pooled lifetime prevalence of BD type 1 was 1.06% (95% confidence interval [95%CI] 0.81-1.31) and that of BD type 2 was 1.57% (95%CI 1.15-1.99). The pooled 1-year prevalence was 0.71% (95%CI 0.56-0.86) for BD type 1 and 0.50% (95%CI 0.35-0.64) for BD type 2. Subgroup analysis showed a significantly higher lifetime prevalence of BD type 1 according to the DSM-IV criteria compared to the DSM-III and DSM-IIIR criteria (p < 0.001). Conclusion: This meta-analysis confirms that estimates of BD type 1 and type 2 prevalence are low in the general population. The increase in prevalence from DSM-III and DSM-III-R to DSM-IV may reflect different factors, such as minor changes in diagnostic operationalization, use of different assessment instruments, or even a genuine increase in the prevalence of BD.