OBJECTIVES: To improve the prognostic value of the age, creatinine, and ejection fraction (ACEF) score following percutaneous coronary intervention (PCI) by integrating the residual SYNTAX score (rSS). BACKGROUND: ACEF score was proposed for predicting the operative mortality risk in elective cardiac operations and has been validated in numerous studies. However, it does not incorporate coronary lesion-based variables for risk assessment of patients who undergo PCI. METHODS: Overall, 10,072 patients who underwent PCI at our hospital in 2013 were enrolled. The endpoint was 2-year cardiac death after PCI, defined as death that was not attributed to a non-cardiac cause. ACEF-rSS was constructed with incremental weights attributed to the ACEF score and rSS according to their estimated coefficients. RESULTS: 2-year cardiac death occurred in 63 patients (0.63%). In multivariable analyses, the ACEF score and rSS > 8 were independently associated with the risk of cardiac death. ACEF-rSS was computed as age (years)/ejection fraction (%) + 1 (if creatinine ≥2.0 mg/dl) + 1 (if rSS >8). The discrimination of ACEF-rSS was significantly better than that of the ACEF score based on receiver operating characteristic (ROC) curve analysis and integrated discrimination improvement (IDI) (C-statistics = 0.835 vs. 0.776 for ACEF-rSS and ACEF score, respectively, p = .029; IDI = 0.014, p < .001). Compared with all other SYNTAX-derived risk scores, ACEF-rSS had significantly better discrimination ability based on ROC curve analysis, net reclassification improvement, and IDI. CONCLUSIONS: Combining the ACEF score with rSS to produce the ACEF-rSS enhanced the predictive ability for long-term cardiac mortality.
OBJECTIVES: To improve the prognostic value of the age, creatinine, and ejection fraction (ACEF) score following percutaneous coronary intervention (PCI) by integrating the residual SYNTAX score (rSS). BACKGROUND: ACEF score was proposed for predicting the operative mortality risk in elective cardiac operations and has been validated in numerous studies. However, it does not incorporate coronary lesion-based variables for risk assessment of patients who undergo PCI. METHODS: Overall, 10,072 patients who underwent PCI at our hospital in 2013 were enrolled. The endpoint was 2-year cardiac death after PCI, defined as death that was not attributed to a non-cardiac cause. ACEF-rSS was constructed with incremental weights attributed to the ACEF score and rSS according to their estimated coefficients. RESULTS: 2-year cardiac death occurred in 63 patients (0.63%). In multivariable analyses, the ACEF score and rSS > 8 were independently associated with the risk of cardiac death. ACEF-rSS was computed as age (years)/ejection fraction (%) + 1 (if creatinine ≥2.0 mg/dl) + 1 (if rSS >8). The discrimination of ACEF-rSS was significantly better than that of the ACEF score based on receiver operating characteristic (ROC) curve analysis and integrated discrimination improvement (IDI) (C-statistics = 0.835 vs. 0.776 for ACEF-rSS and ACEF score, respectively, p = .029; IDI = 0.014, p < .001). Compared with all other SYNTAX-derived risk scores, ACEF-rSS had significantly better discrimination ability based on ROC curve analysis, net reclassification improvement, and IDI. CONCLUSIONS: Combining the ACEF score with rSS to produce the ACEF-rSS enhanced the predictive ability for long-term cardiac mortality.