Background Heart rate variability (HRV) has been suggested reduced in bipolar disorder (BD) compared with healthy individuals (HC). This meta-analysis investigated: HRV differences in BD compared with HC, major depressive disorder or schizophrenia; HRV differences between affective states; HRV changes from mania/depression to euthymia; and HRV changes following interventions. Methods A systematic review and meta-analysis reported according to the PRISMA guidelines was conducted. MEDLINE, Embase, PsycINFO, The Cochrane Library and Scopus were searched. A total of 15 articles comprising 2534 individuals were included. Results HRV was reduced in BD compared to HC (g = -1.77, 95% CI: −2.46; −1.09, P < 0.001, 10 comparisons, n = 1581). More recent publication year, larger study and higher study quality were associated with a smaller difference in HRV. Large between-study heterogeneity, low study quality, and lack of consideration of confounding factors in individual studies were observed. Conclusions This first meta-analysis of HRV in BD suggests that HRV is reduced in BD compared to HC. Heterogeneity and methodological issues limit the evidence. Future studies employing strict methodology are warranted.