Recurrent events in meta-analysis of multiple clinical trials.

AIM: To study the efficacy and safety of drug or therapy with recurrent events in meta-analysis of multiple clinical trials. METHODS: A nonparametric approach is proposed to estimate the rates of recurrent events for meta-analysis of trials. The method was used in meta-analysis of angiotensin-converting enzyme (ACE) inhibitor clinical trials to analyze the relative rates and the excess rates between ACE inhibitor and placebo treatment groups for endpoints of hospitalizations for CHF, hospitalizations for CHF or cardiac death, and hospitalizations for CHF or any death, respectively. RESULTS: The estimates of those three endpoints were 69%, 74%, and 76% (P < 0.01). Compared with placebo, ACE inhibitor reduced 30 cases of hospitalizations for CHF per 1000 person-years, or 40 cardiac deaths or hospitalizations for CHF per 1000 person-years (P < 0.01). CONCLUSION: The method was a simple and efficient approach to conduct meta-analysis of clinical trials with recurrent events.