Abstract Background Risk stratification of out-of-hospital sudden cardiac death (SCD) patients after acute myocardial infarction (AMI) is crucial for timely therapeutic interventions. However, left ventricular ejection fraction (LVEF) as a traditional clinical tool has major limitations, necessitating improved risk stratification strategies. High-throughput proteomic offers the potential to detect novel biomarkers to improve risk assessment. Purpose To identifiy optimal circulating protein combination predicting out-of-hospital SCD post-AMI. Methods We conducted a two-center retrospective nested case-control study of cases with out-of-hospital SCD and controls with AMI who discharged alive. In a discovery cohort, we carried out plasma proteomics profiling of 6,592 peptides and 522 proteins using mass spectrometry. The proteins associated to SCD were further assessed using targeted quantitative proteomics, and the optimal protein combination for predicting out-of-hospital SCD post-AMI was identified by predictor selection classifier. Finally, ELISA validated its predictive value by comparing with a clinical stratification tool (LVEF ≤35%) and two reported risk stratification models within discovery and two independent validation cohorts. Results In the discovery cohort (Figure 1, 210 participants including 105 SCD cases [median age: 68.0 years, 59.05% men]), mass spectrometry discovered 44 differential proteins not previously associated to SCD. Targeted proteomics quantified these proteins and identified optimal protein combination, SCD-warning three protein combination (SCD-W3P), including coronin-1A, haptoglobin, and complement factor D (CFD). A model based on expression levels of SCD-W3P outperformed LVEF alone (Figure 2, C-statistic: 0.752 vs. 0.548, P<0.001), improving its risk prediction (ΔC-statistic = 0.281, P<0.001) and net reclassification improvement (9.5%, P<0.001). Similar incremental discrimination metrics were observed in two reported stratification models combined with SCD-W3P, particularly within the LVEF-preserved population. These findings were repeatably validated in two independent cohorts (Figure 1 and 2, validation 1: 160 participants including 40 SCD cases [median age: 72.5 years, 67.50% men]; validation 2: 96 participants including 24 SCD cases [median age: 74.5 years, 79.17% men]). Finally, CFD inhibition protection for mortality and pro-malignant arrhythmias in AMI mice supported biological plausibility of the critical protein in SCD-W3P. Conclusions This two-center retrospective nested case-control study with two independent validation serves as the largest-scale investigation with the most comprehensive proteomics profiling in out-of-hospital SCD after AMI, identifying an optimal protein combination which may contribute to enhance early identification for timely intensive management. These findings suggest pivotal proteins for improving understanding SCD pathophysiology.Methods overview.Comparison of model performances for pre