A random effects meta-analysis investigating the prevalence of bipolar disorder in people with fibromyalgia: An updated analysis

Abstract

I read with great interest the innovative review paper recently published by Kudlow et al. (2015). The authors address an important topic, investigating the prevalence and odds of bipolar disorder (BD) among people with fibromyalgia (FM), which has received minimal attention in the literature to date. Whilst this paper was helpful and clearly advanced the field, the authors determined the 'pooled prevalence rate' based upon a calculation using the sum of the total number of people with BD across all studies divided by the total sample size and did not perform a meta-analysis. Therefore, utilising the author's data across the eight studies, I recalculated the prevalence of BD in FM with a random effects meta-analysis together with the 95% CI. Across all 8 studies, the pooled prevalence of BD was 15.2% (95% CI 5.3-36.3%) with high heterogeneity (I(2)=95%). Although the Egger test indicated no significant publication bias (=-6.70, p=0.12) the Duval and Tweedie trim and fill adjusted prevalence of BD in fibromyalgia was 17.9% (6.68-40.2%) with one study adjusted for publication bias. Sub group meta-analysis determined that the prevalence of BD was similar in case control (15.2%, 95% CI 2.9-54.0%, 4 studies) and cohort studies (14.2%, 95% CI 2.7-49.7%, P=0.94, 4 studies). Thus, the actual prevalence of BD in FM may be slightly lower than that the reported by the authors. Nonetheless, more should be done to accurately identify and manager people with FM and comorbid BD. Future research might also consider common neurobiological underpinnings.