Bilirubin has potential predictive and prognostic value for myocardial infarction (MI), but the clinical evidence remains controversial. We performed this meta-analysis to systematically quantify the relationships between circulating bilirubin levels and the incidence of MI and post-MI adverse events.We searched the PubMed, Cochrane Library, Embase, and Web of Science databases for ad-hoc studies, published up to October 17, 2020, recording bilirubin before MI (predictive analyses) or adverse events (prognostic analyses). Relative risks (RR) were pooled by a random-effects model. The dose-response analysis was conducted by restricted cubic splines. In patients without previous MI, increased total bilirubin (TB) reduced the risk of long-term (3 year) first MI by 22% (95% confidence interval [CI]: 0.69-0.88, n = 4). The dose-response analysis indicated that the RR for first MI decreased by 2.7% per each 2 μmol/L increment of TB (three studies, 95% CI: 1.3%-4.1%, P 0.001), with a cut-off value of 12.60 μmol/L for RR 1.00. Elevated bilirubin reduced the incidence of first and recurrent MI by 36% (95% CI: 0.42-0.98, n = 7). However, after suffering MI, higher TB concentrations could not decrease the risk of recurrent MI (RR: 1.02, 95% CI: 0.67-1.55, n = 5) and increased the incidence of short-term (1 year) post-MI major adverse cardiovascular events, all-cause mortality, and cardiovascular mortality, but not long-term (≥1 year).Higher TB levels within a physiological range reduced the incidence of long-term first MI, with a cut-off value of 12.60 μmol/L.