J Heart Valve Dis 2001 Sep;10(5):572-7; discussion 577-8
Risk factors for early mortality after valve surgery in
Europe in the 1990s: lessons from the EuroSCORE pilot program.
Roques F, Nashef SA, Michel P; The EuroSCORE study group.
Hopital La Meynard, Fort de France, Martinique, French West Indies.
BACKGROUND AND AIM OF THE STUDY: The characteristics of valve surgery are
evolving. The study aim was to explore its demographics and risk factors in
Europe in the 1990s, using the EuroSCORE database. METHODS: For the EuroSCORE
program, information on 98 variables regarding risk factors, procedures and
outcome were collected for 5,672 patients undergoing valve surgery under
cardiopulmonary bypass in 128 European centers. Bivariate (i.e. Mann-Whitney
test or chi-square when appropriate), then logistic regression analyses were
carried out to identify risk factors for early mortality. The predictive value
of EuroSCORE was analyzed using the Hosmer-Lemershow test and by computing the
area under the receiver operating characteristic (ROC) curve. RESULTS: Aortic
valve stenosis was the most common diagnosis (47.6%), whilst mitral valve
surgery accounted for 42% of procedures. Coronary surgery was performed
concomitantly in 21% of cases. Hospital mortality was 6.1%. Predictive factors
for early mortality were: age (p = 0.0001), preoperative serum creatinine >200
micromol/l (p = 0.014), previous heart surgery (p = 0.0001), poor left
ventricular function (p = 0.008), chronic congestive heart failure (p = 0.0001),
pulmonary hypertension (p = 0.0001), active acute endocarditis (p = 0.0001),
emergency procedure (p = 0.05), critical preoperative status (p = 0.0001),
tricuspid surgery (p = 0.015), aortic and mitral surgery (p = 0.002), combined
thoracic surgery (p = 0.0001), and combined coronary surgery (p = 0.0001). The
predictive value of EuroSCORE for mortality was good (area under the ROC curve =
0.75). CONCLUSION: The 'valve' subset of the EuroSCORE database provides an
instant picture of European valve surgery in the 1990s that can be used either
for individual assessment, or for country- or institution-based epidemiological
studies of risk factors and practices.