Enrichment of the Postdischarge GRACE Score With Deceleration Capacity Enhances the Prediction Accuracy of the Long-Term Prognosis After Acute Coronary Syndrome

BACKGROUND: Cardiac autonomic nerve imbalance has been well documented to provide a critical foundation for the development of acute coronary syndrome (ACS) but is not included in the postdischarge GRACE score. We investigated whether capturing cardiac autonomic nervous system (ANS)-related modulations by 24-h deceleration capacity (DC) could improve the capability of existing prognostic models, including the postdischarge Global Registry of Acute Coronary Events (GRACE) score, to predict prognosis after ACS. METHOD: Patients with ACS were assessed with 24-h Holter monitoring in our department from June 2017 through June 2019. The GRACE score was calculated for postdischarge 6-month mortality. The patients were followed longitudinally for the incidence of major adverse cardiac events (MACEs), set as a composite of non-fatal myocardial infarction and death. To evaluate the improvement in its discriminative and reclassification capabilities, the GRACE score with DC model was compared with a model using the GRACE score only, using area under the receiver-operator characteristic curve (AUC), Akaike's information criteria, the likelihood ratio test, category-free integrated discrimination index (IDI) and continuous net reclassification improvement (NRI). RESULTS: Overall, 323 patients were enrolled consecutively. After the follow-up period (mean, 43.78 months), 41 patients were found to have developed MACEs, which were more frequent among patients with DC <2.5 ms. DC adjusted for the GRACE score independently predicted the occurrence of MACEs with an adjusted hazard ratio (HR) of 0.885 and 95% CI of 0.831–0.943 (p < 0.001). Moreover, adding DC to the GRACE score only model increased the discriminatory ability for MACEs, as indicated by the likelihood ratio test (χ(2) = 9.277, 1 df; p < 0.001). The model including the GRACE score combined with DC yielded a lower corrected Akaike's information criterion compared to that with the GRACE score alone. Incorporation of the DC into the existing model that uses the GRACE score enriched the net reclassification indices (NRIe(>0) 7.3%, NRIne(>0) 12.8%, NRI(>0) 0.200; p = 0.003). Entering the DC into the GRACE score model enhanced discrimination (IDI of 1.04%, p < 0.001). CONCLUSION: DC serves as an independent and effective predictor of long-term adverse outcomes after ACS. Integration of DC and the postdischarge GRACE score significantly enhanced the discriminatory capability and precision in the prediction of poor long-term follow-up prognosis.