Data synthesis and analysis
Meta-analysis and meta-regression were performed using Comprehensive
Meta-Analysis software. Random-effect pooled odds ratios (ORs) were
calculated with the corresponding 95% confidence intervals (CIs) to
summarize the results overall and within subgroups. Heterogeneity was
assessed by using the I2 statistic. We performed
subgroup analyses by study type (RCT, cohort). Random-effects
meta-analysis was used to pool analyses. Meta-regression analyses were
performed to evaluate whether differences in pravastatin dosage and
timing of treatment initiation modified the association between exposure
to pravastatin and outcomes, and to explain the heterogeneity in the
estimated effect size. The data were analyzed using the Comprehensive
Meta-Analysis software.