Purpose: Bipolar disorder is a common, severe mental health condition and a major financial burden for healthcare systems across the globe. There is some evidence that unrecognized bipolar disorder is prevalent amongst patients with depression in primary care which can lead to non-optimal treatment. However, a systematic synthesis of this literature is lacking. We aimed to determine the percentage of primary care patients who are diagnosed with depression that have unrecognized bipolar disorder. Methods: Medline, Embase, Cochrane and PsycINFO were searched to January 2019. We included quantitative observational studies. Risk of bias was assessed using the Newcastle Ottawa cohort scale. Analyses were performed using random-effects models, heterogeneity was quantified using I
and formal tests of publication bias were undertaken. Results: Ten studies with 3803 participants with depression in primary care were included. The pooled prevalence of bipolar disorder in those with depression was 17% (95% CI = 12 to 22). The prevalence of unrecognized bipolar depression was higher in studies which used questionnaires as assessment tools for bipolar disorder compared to studies which used clinical interviews but this difference was not significant (14%, 95% CI = 8 to 20 versus 22%, 95% CI = 16 to 28, Q = 1.27, p = 0.12). The prevalence of unrecognized bipolar disorder was not significantly affected by study-level variations in the risk of bias and we found no evidence for publication bias. Conclusion: Over 3 in 20 patients with depression have unrecognized bipolar disorder in primary care which can lead to harmful patient outcomes. Increased awareness of unrecognition of bipolar disorder in primary care patients with depression and efficient assessment strategies in primary care are warranted.