Six RCTs (n=13,778) were included in the review.
No statistically significant between-group differences were found for mortality (OR 0.91, 95% CI: 0.77, 1.09; 6 RCTs, n=13,778), nonfatal MI (OR 0.95, 95% CI: 0.80, 1.13; 4 RCTs, n=13,500), hospitalisation (OR 0.97, 95% CI: 0.80, 1.17; 5 RCTs, n=13,718), or a composite of cardiovascular events (OR 0.93, 95% CI: 0.84, 1.03). No evidence of statistical heterogeneity was found. The removal of trials with a treatment course of less than 7 days or trials that did not include participants seropositive for C. pneumoniae did not significantly change the results.
Plots and statistical tests showed a low probability of publication bias.