Reclassification of CADILLAC risk score with lung ultrasound at admission in patients with STEMI in a one-year follow-up

Abstract

Abstract Funding Acknowledgements Type of funding sources: Foundation. Main funding source(s): JC-M received Original Research Projects Grant from the Catalan Society of Cardiology and LR-S received Scholarship for Training and Research from the Association of Ischemic Heart Disease and Cardiological Critical Care of the Spanish Society of Cardiology. Introduction Lung ultrasound (LUS) has emerged as a valuable diagnostic and prognostic tool in heart failure (HF) and shows promise in risk stratification for ST-segment elevation myocardial infarction (STEMI) during the acute phase. Current risk stratification scales for STEMI primarily rely on the Killip scale to assess the severity of heart failure (HF) upon admission. However, these scales were developed before the widespread use of LUS, and as a result, LUS findings have not been incorporated into any existing risk stratification scales for STEMI. Purpose To assess the potential improvement in prognostic predictive value achieved by incorporating LUS findings into a current risk score for STEMI. Methods In a prospective multicenter cohort study, an independent operator conducted LUS, and offline B-line counting was performed upon admission. Patients were categorized as having wet lung (at least one site with ≥3 B-lines) or dry lung based on LUS results. The CADILLAC score was calculated, and patients were classified as either low risk (≤2 points) or medium/high risk (>2 points). We evaluated the Net Reclassification Index (NRI) and conducted a Receiver Operating Characteristics (ROC) comparison based on the prediction of a composite endpoint which included death, hospitalization for HF, acute coronary syndrome, or stroke within one year following revascularization. Results The CADILLAC score was computed for 353 patients after excluding those admitted with Killip IV. Among these, 216 (61.2%) had a low-risk score (≤2), and 137 (38.8%) had a medium-high score. LUS significantly reclassified CADILLAC results, resulting in an absolute NRI of 12.4%. ROC comparison of models comprising CADILLAC and CADILLAC + LUS demonstrated that the addition of LUS significantly improved the accuracy of predicted risk (0.722 vs. 0.770, p=0.05). Conclusion The inclusion of LUS findings in the CADILLAC score enhances its prognostic capacity.Figure 1