Your article reports 86 heart attacks among 15,560 rosiglitazone (Avandia) users, versus 72 others in a control group of 12,283. A study coauthor then says that “after statistical adjustment, that yields a 43 percent higher risk of heart attacks among rosiglitazone users.” Simple arithmetic would seem to indicate it was the control group that had a slightly higher risk.

Vincent Fecher
San Antonio, Texas

The study reported was based on data from diverse trials. Simply adding all the numbers to reach a risk calculation would require that the studies be identical in design. Since they were not, scientists had to account for various differences. But it’s fair to say that the results appear counterintuitive. —N. Seppa