[Hospital mortality at a cardiosurgical unit in Torino: international comparisons and time trend]

To compare hospital mortality in a cardiac surgery unit with external data and to assess changes in time (patients undergoing surgery in two different periods).Data on risk factors for hospital mortality were collected from clinical records (retrospectively for the first period and prospectively for the second) for all patients undergoing open heart surgery at the Heart Surgery Unit of the University of Turin (Italy) during 1991 and 1995 (n = 1794) and 1999 (n = 892). Comparisons of in-hospital mortality, expressed as Standardized Mortality Ratios (SMR), were adjusted for risk factors defined according to EuroSCORE (European System for Cardiac Operative Risk Evaluation).In the first and second period, complete information on all the 17 EuroSCORE items was available for 58.3% and 89.6% patients respectively. After exclusion of patients with one or more missing data, observed and expected numbers of death were found to be very similar, with SMRs ranging between 0.82 (isolated bypass in the second period) and 1.06 ("other" surgery in the first period). Mortality was higher among patients with missing data, but at least in 1999 the latter had a limited impact on the overall estimates. Compared to the first period, mortality was reduced during 1999 (from 5.9% to 5.4%), in particular for isolated bypass (from 4.4% to 3.4%).In the unit under investigation, hospital mortality following heart surgery was similar to that predicted from EuroSCORE and seemed to be lower in 1999 than in 1991-95, particularly for isolated bypass. Incompleteness of data on individual risk factors may have been a source of bias, especially when data were collected retrospectively.