Which score best predicts new-onset AF after primary PCI in STEMI? Head-to-head assessment of four algorithms

Background New-onset atrial fibrillation (NOAF) is a frequent and clinically significant comorbidity following primary percutaneous coronary intervention (PCI) for ST-segment elevation myocardial infarction (STEMI), associated with increased mortality, stroke risk, and healthcare costs. Predicting NOAF in this setting remains challenging. This study aimed to compare the predictive performance of four risk scores—CHA 2 DS 2 -VASc-60, mC2HEST, Mayo-AF, and HATCH—in identifying patients at high risk of NOAF after primary PCI in STEMI. Methods We conducted a single-center, retrospective cohort study at Shaoxing Central Hospital, including consecutive adults with STEMI who underwent primary PCI within 12 h of symptom onset between February 2020 and June 2025. Patients with a history of atrial fibrillation or other exclusions were excluded. The endpoint was NOAF within 1 year after PCI, documented by continuous inpatient telemetry, 12-lead ECG at discharge, and routine follow-up ECGs at 1, 3, 6, 9 and 12 months. We compared the discriminative ability of the four scores using receiver operating characteristic (ROC) curves and assessed their clinical utility through net reclassification improvement (NRI) and integrated discrimination improvement (IDI). Results A total of 1,175 patients were included in the final analysis. The incidence of NOAF within 1 year was 8.2% ( n = 96). The mC2HEST score demonstrated the highest area under the ROC curve (AUC) for predicting NOAF (AUC = 0.786), followed by HATCH (AUC = 0.746) and CHA 2 DS 2 -VASc-60 (AUC = 0.739), while Mayo-AF had the lowest AUC (AUC = 0.736). Pairwise comparisons showed that mC2HEST significantly outperformed CHA 2 DS 2 -VASc-60 ( Δ AUC = 0.046, p = 0.002). Conclusions Among the four risk scores evaluated, mC2HEST provided the best predictive performance for NOAF after primary PCI in STEMI patients. Its superior discriminative ability and significant integrated discrimination improvement suggest that mC2HEST could be a valuable tool for identifying high-risk patients who may benefit from intensified monitoring and early interventions. Future prospective, multicenter studies are warranted to validate these findings.