Letters Section Editor: Robert M. Golub, MD, Senior Editor.
In Reply: Dr Scharnetzky and colleagues raise questions about site-level bias, suggesting differential enrollment between European and North American centers. However, patient selection criteria and sampling (every nth patient) were strictly prespecified and identical at each participating center. Neither site nor investigator had control over patient selection or sampling.
The use of blood-sparing therapy differed among the 69 centers in 16 countries. Post-hoc analyses of long-term all-cause mortality rates for patients treated with aprotinin for the European centers showed a 5-year covariate-adjusted HR of 1.48 (95% confidence interval [CI], 1.10-2.00); for the North American centers, the adjusted HR was 1.39 (95% CI, 0.96-2.00). However, these post-hoc subset analyses must be interpreted with caution because of limited power.1 We estimate that for the North American subset, there is only 42% power to detect aprotinin-associated mortality differences, compared with 94% power for the entire sample. Survival analysis demonstrated significant differences in both subsets, consistent with results for the total sample (Kaplan-Meier log-rank P = .01 for European centers, and P <.001 for North American centers).
Mangano DT. Long-term Mortality Associated With Aprotinin Following Coronary Artery Bypass Graft Surgery—Reply. JAMA. 2007;297(22):2475-2477. doi:10.1001/jama.297.22.2477-a