Predictive value of STEP-HF risk score in patients with ST-segment elevation myocardial infarction (Killip I) treated by primary percutaneous coronary intervention

Abstract

Abstract Background New-onset heart failure (HF) during the first hospitalization was considered as an earlier indicator for poor prognosis in acute ST-segment elevation myocardial infarction (STEMI) without overt HF (Killip I). Purpose This study aims to develop a practical risk score to predict the in-hospital occurrence of new-onset HF and to evaluate its ability for prognosis. Methods The Chinese STEMI pPCI Registry is registered with ClinicalTrials.gov (NCT04996901). Within 24 h of symptom onset, 2940 STEMI patients with Killip class I on admission were treated by pPCI from five hospitals between January 2015 and January 2021; 1869 patients from three of the hospitals served as a derivation cohort, and 1071 from the other two hospitals constituted the external validation cohort. In-hospital HF was defined as Killip class ≥II at the first hospital presentation (8.7% in the derivation cohort vs. 14.3% in the validation cohort). Multivariable logistic regression was used to develop the in-hospital HF score. For 2-year following-up, all-cause mortality rate was estimated using the Cox regression analysis. Results Eight variables were independently associated with the development of in-hospital HF: age >65 years, previous atrial fibrillation, triple vessel lesion, anterior myocardial infarction, symptom onset-to-balloon time > 5 h, TIMI flow grade 0 before pPCI, neutrophil-to-lymphocyte ratio > 6.3 and aspartate aminotransferase > 65U/L on admission. STEP-HF score was created based on these 8 variables, and classified patients into three risk categories: low (0-3), intermediate (4-7), and high (≥8). Observed in-hospital HF rates were 3.47% (3.5%), 12.35% (8.34%), and 30.8% (17.05%) across the three risk categories, respectively in derivation (validation) cohort. The STEP-HF score demonstrated high discrimination (c-statistic of 0.705 in the derivation cohort, 0.656 in the validation cohort) and adequate calibration (Hosmer-Lemeshow goodness-of-fit test P values > 0.05) in both cohorts. After two-year follow up, 202 patients were lost to follow-up and finally a total of 2738 patients were included. 30-day all-cause mortality occurred in 0.13%, 0.66% and 1.71% of patients in low, intermediate and high risk group, respectively (p = 0.041). The association persisted during 1 year and 2 years of follow-up. Two-year all-cause mortality were 8 (1.03%), 32 (2.33%), and 37 (6.32%) for the low, intermediate, and high score groups, respectively (p<0.001). Receiver operating characteristic (ROC) analysis demonstrated the STEP-HF score can effectively predict all-cause mortality as compared with GRACE score for 30d, 1 year, and 2 years, respectively (STEP-HF: AUC 0.76, 0.71,0.68 vs. GRACE: AUC 0.75, 0.67, 0.67, p = 0.923, 0.202, 0.619, respectively). Conclusion The STEP-HF score can be used to predict prognosis in STEMI (Killip I) with pPCI. Higher scores are associated with higher in-hospital HF at first admission and all-cause mortality for long-term following up.