In Reply: Dr Pontiroli raises 2 important questions. First, what are the effects of different bariatric surgical methods on cardiovascular events? With 2010 surgery patients and 2037 controls at baseline, we did not have the statistical power to analyze the treatment effects in surgical subgroups and so did not include the information in the article. Compared with controls, the hazard ratios (HRs) for fatal cardiovascular events (myocardial infarction and stroke combined) were 0.30 (95% CI, 0.10-0.87; P = .03) for gastric banding; 0.45 (95% CI, 0.25-0.81; P = .007) for vertical banded gastroplasty; and 1.07 (95% CI, 0.45-2.58; P = .87) for gastric bypass. There were 49 fatal cardiovascular events in 2037 controls, 4 events in 376 patients with gastric banding, 18 events in 1369 patients with vertical banded gastroplasty, and 6 events in 265 patients with gastric bypass. For total cardiovascular events (fatal and nonfatal), the HRs were 0.54 (95% CI, 0.36-0.81; P = .003) for gastric banding; 0.91 (95% CI, 0.73-1.14; P = .44) for vertical banded gastroplasty; and 0.94 (95% CI, 0.61-1.45; P = .77) for gastric bypass. There were 234 total cardiovascular events in 2037 controls, 29 events in 376 patients with gastric banding, 147 events in 1369 patients with vertical banded gastroplasty, and 23 events in 265 patients with gastric bypass. With these findings, we were not prepared to draw any firm conclusions about superiority of a given surgical method. It may be possible to answer Pontiroli's question when more events are observed during extended follow-up of the patients.