| BackgroundDecision making for Left main coronary artery disease revascularization (CABG v.s. PCI) is challenging. Based on the4year follow up outcomes of SYNTAX trial, the SYNTAX Score II (SSII) has been developed recently, which was the first individualized risk model aimed to aid in decision making for revascularization. No studies have been to validated SSII in left main stenosis.Objectives1. This study aimes to report the long term outcomes of patients with LMCAD treated by CABG in Fuwai Hosipital from2004to2010.2. To evaluate and validate the prognostic effectiveness of the SSII-CABG risk model in left main coronary artery disease patients, and further assess its utility in clinical practice.3. To assess the predicting ability of different risk models in LMCA disease patients treated by CABG, and modify the established models to achieve better effectiveness of prediction.MethodsAll consecutive patients with angiographic diagnosed left main stenosis>50%treated by CABG were enrolled from2004to2010in Fuwai Hospital. Outcomes of long-term follow up was analysed. SYNTAX Score, EuroSCORE, SinoSCORE, modified ACEF, GRC and SSII(for PCI and CABG) were calculated.Patients were stratified according to tertiles of SSII-CABG, and5-year outcomes were compared among tertiles. The prognostic effectiveness of different risk models was assessed by c-statistics, Hosmer-Lemeshow statistics, IDI and NRI. Outcomes were compared between groups which were divided according to concordance between the actual operation received and SSII recommandation.All patients were divided into3groups according to SYNTAX stratificaiton, and the progonsis of the groups were compared. A modified GRC model with SYNTAX tertiles and EuroSCORE tertiles was constructed, and the new model was further validated by stratification and reclassification analysis.ResultsAt a mean follow up of4.8years, the all-cause mortality, cardiac mortality and MACCE rate were6.4%,2.4%and19.6%respectively.The mortality of low, intermediate and high SSII-CABG tertiles were2.7%,5.9%,10.2% respectively, p<0.001. Multivariable COX regression found SSII-CABG to be an independent predictor of mortality (Hazard ratio:1.390,95%CI[1.104,1.748], p=0.005). The IDI and NRI of SSII-CABG tertile were superior to SYNTAX score but comparable to other risk models. The c-statistics of SSII-CABG was0.680(95%CI [0.624,0.736], p=0.029), the p value of Hosmer-Lemeshow statistics was0.031. The concordance between actual revascularization strategy and SSII recommandation was not associated with5-year outcomes.No significant intergroup differece was observed between SYNTAX groups with respect to5-year outcomes. Multivariable COX regression showed that SYNTAX was not an independent risk factor of all-cause death (HR=0.699,95%CI [0.575,1.448], p=0.913). The best and most balanced model based on clinical variables with respect of discrimination and calibration was EuroSCORE (c-statistics0.647,95%CI [0.620,0.728]; Hosmer-Lemeshow statistics16.88). EuroSCORE was superior to SYNTAX Score in predicting of long term all-cause death in terms of reclassification analysis(IDI=0.017,95%CI [0.004,0.046],p<0.001; NRI=0.235,95%CI [0.059,0.35], p=0.033). GRC was superior to EuroSCORE with respect to all cause death, non-cardiac death and MACCE. Modified GRC exhibited stratification ability in terms of all cause death and non-cardic death(intergroup p value<0.001). Reclassification analysis showed no significant differeces between modified GRC and EuroSCORE.Conclusions1. SSII-CABG has good stratifying ability of long-term prognosis of left main coroanry artery disease treated by CABG.2. SSII model has poor utility in decicion making for left main stenosis revascularization between CABG and PCI.3. SYNTAX has poor performance in prediction of long-term prognosis in patients with LMCA disease treated by CABG.4. EuroSCOREis the most balanced risk model.5. Combination of anatomical model and clinical model does not improve the prediction performance of clinical models. |
PDF Full Text Request