Background: Mortality from cardiac surgery is an essential indicator of quality and forms the basis of treatment strategy decisions in eligible patients. No contemporary complete data on unselected adult cardiac surgery patients are available in Germany.
Methods and results: A registry was started in June 1997 of all patients referred to surgery from 85 cardiology centres in Germany. The registry was intended to include 10 000 patients and this number was reached in March 1998. Follow-up of the patients was by simple questionnaire, reporting the date of surgery, major complications, and symptomatic improvement. If the questionnaire was not returned, a reminder letter was sent and, if necessary, further telephone investigations were performed. This resulted in 99.9% complete data. Of 10 525 patients operated on, 3.91% had died by 30 days after surgery. The overall operative mortality was 4.57%, which included 69 patients who died after more than 30 days from complications related to surgery. By multivariate analysis, the following predictors of mortality were identified: previous surgery, emergency or complex operation; age >75 years, female gender, cardiac failure, angina CCS class IV, and three-vessel coronary disease. An integral part of the registry was a pre-operative prediction of surgical risk in five categories. This risk estimate revealed a surprisingly correct prediction of the mortality observed.
Conclusions: In a representative unselected group of cardiac surgery patients, operative mortality was 4.57%. Several procedural and clinical parameters were significantly correlated with mortality, but the risk increment by each of these factors was small. Unstructured clinical judgement reliably predicted the operative risk.