Letters Section Editor: Robert M. Golub, MD, Senior Editor.
To the Editor: The study of long-term mortality associated with aprotinin by Dr Mangano and colleagues1 did not enroll all patients with coronary artery bypass graft (CABG) surgery in the 69 designated medical centers. There is a potential for site-level bias related to unequal enrollment between European and North American centers, corresponding differences in medication use, and possible associated variations in the quality of care. Based on information from the Web site of the Ischemia Research and Education Foundation describing 70 sites,2 40 were located in North America, 24 in Europe, and 6 elsewhere. Combining this information with the data given in Table 1,1 it appears that each North American site enrolled a mean of 45 patients during the study period, whereas each European site contributed a mean of 87 patients. Each institution would have been expected to contribute a total of about 180 patients during the study period of 3.6 years; however, the North American sites enrolled only 25% of the expected patients, whereas in Europe the corresponding rate was 48%. This may have introduced a selection bias.
Scharnetzky E, Schill W, Garbe E. Long-term Mortality Associated With Aprotinin Following Coronary Artery Bypass Graft Surgery. JAMA. 2007;297(22):2475-2477. doi:10.1001/jama.297.22.2475-a