The report of an increased risk of AF (atrial fibrillation) with Zoledronic acid by Black et al. (NEJM 356:1809-1822, 2007) and the observations of a nominally, but not statistically significant, increased risk of serious events of AF in the original alendronate FIT trial by Cummings et al. (in a letter to the editor in the same issue of NEJM, pp. 1896-1896) had prompted us to explore the incidence of AF and other related cardiovascular (CV) endpoints in clinical trials conducted by Merck which compared alendronate to placebo. While the effort of locating all relevant studies and corresponding data was ongoing, we also received a detailed request from the FDA to submit AF related data from all of our trials on alendronate. The FDA had a plan to conduct their own meta-analysis on all biophosphonates (including alendronate) for AF. Some of the trials were run by Merck while others were run for Merck by contract research organizations over a period of several years. We had to deal with the problem of extracting data from different sources and different platforms with varying definitions of some terms. Some of the adverse events were rare and we had several trials with zero events in one or both arms. These trials had different durations and included different doses of alendronate. Thus we faced both logistical and statistical problems. The statistical approach we used was somewhat standard for estimation of relative risk of alendronate versus placebo. The primary method of analysis for all endpoints was an analysis of rates using exact Poisson regression with time as an offset parameter. The stability (sensitivity) of the estimates was evaluated by conducting exact Poisson regression meta-analysis with each study eliminated one at a time. Based on the analysis, alendronate use was not found to be associated with an increased risk of AF.