Results:

For the derivation cohort, average age was 67.2 ± 11.1 years, and 22.8% were female. In the derivation cohort, there were 10 deaths and 93 patients with major morbidity, as characterized by the STS outcomes. The STS score predicted 194 patients who were low risk (0%-10%), 308 intermediate risk (10%-25%), and 164 high risk (>25%). Multivariable predictors of in-hospital mortality or major morbidity were a restrictive mitral filling pattern (odds ratio [OR], 2.96; p = 0.001), RV fractional area change (FAC) <35% (OR, 3.03; p = 0.01), RV myocardial performance index (MPI) >0.4 (OR, 1.89; p = 0.02), and mitral regurgitation ≥ moderate (OR, 1.96; p = 0.07). For all-cause mortality at 3.2 years, only RV FAC (OR, 2.02; p = 0.07) and PASP ≥50 mm Hg (OR, 3.54; p < 0.001) were predictive. The area under the curve for the STS score was 0.68, which increased to 0.73 with the addition of the echocardiographic parameters. LVEF was a univariate predictor of events (OR, 5.2; p < 0.001), but was no longer significant when diastolic dysfunction and RV function were entered into the variable model. RV FAC and MPI provided complementary information in that 5% of patients with RV hypokinesis and abnormal RV FAC had 50% mortality or major morbidity, and RV MPI identified 24% with subclinical RV impairment with a 26% event rate. In the validation cohort, there were 33 (17.6%) with postoperative mortality or major morbidity. Restricted LV filling (OR, 4.73), abnormal RV MPI (OR, 2.54), and mitral regurgitation (OR, 1.65) were the optimal predictors of events.