A simple clinical scoring model to predict significant left main/three-vessel disease in non-ST elevation acute coronary syndrome: The MADISON score

Background: Patients with non-ST elevation acute coronary syndrome (NSTE-ACS) who have significant left main and/or three-vessel disease (LM/3VD) are at particularly high risk for poor outcomes. Timely identification and revascularization of these patients is essential. However, growing healthcare resource limitations necessitate patient triage and prioritization to avoid delays in care. Aims: To identify clinical predictors of LM/3VD available at initial NSTE-ACS presentation and validate a prediction model in independent cohorts. Methods: Retrospective data of NSTE-ACS patients from National Cardiovascular Data Registries of three institutional cohorts were analyzed. Logistic regression identified predictors of LM/3VD in an index cohort to develop a predictive model. The model's performance was subsequently evaluated in the remaining two cohorts. Results: Among 1706 NSTE-ACS patients, 24 % had LM/3VD. Seven clinical predictors—age ≥ 65, male sex, diabetes, ECG abnormalities, heart rate > 90 bpm, history of stroke/TIA, and signs of heart failure—were used to derive the MADISON score. The model demonstrated an AUC-ROC of 0.67 across cohorts. At the prespecified clinical threshold (score ≥ 4), the MADISON score showed high specificity (83.5–95 %) but low-to-moderate sensitivity (3.5–34.4 %) for LM/3VD detection. Conclusions: A high MADISON score (≥4) reliably identifies patients at increased risk of LM/3VD with high specificity, supporting its potential use for triage and prioritization of patient care in resource-limited settings.