Authors' reply

摘要

Sally Doherty and colleagues report the findings of a retrospective study of surgical mortality in Ireland during the same period studied in our prospective European study (EuSOS).1 Using a different method, the authors collected data describing a larger cohort of patients and identified fewer deaths, resulting in a different mortality estimate. While the authors use the original EuSOS data as a reference for their findings, they also suggest these data are inaccurate. The overall mortality for any large population of surgical patients is crucially dependent on the representation of high-risk surgical patients within it. The lower numbers of critical-care admissions and deaths suggest the high-risk group was not so strongly represented in this repeat study population. This difference might also represent a stronger tendency for investigators to include patients undergoing complex surgery in the prospective study. Nonetheless, the hospital mortality of 2-5% is higherthan previous estimates, which range from 1 to 2%,2 and remains a cause for concern.