Key result
Incorporating the STOP-BANG score into the GRACE risk model improved F1-score, accuracy, and precision for predicting in-hospital cardiovascular events after myocardial infarction, though the increase in ROC-AUC from 0.70 to 0.76 was not statistically significant (p=0.38).
Why the study?
Obstructive sleep apnea may impact outcomes in ACS patients, and it was unknown whether incorporating the STOP-BANG score enhances the predictive ability of the GRACE score.
Does incorporating the STOP-BANG score into the GRACE score improve prediction of in-hospital cardiovascular events in myocardial infarction patients?
Observational (n=227)
No
Does incorporating the STOP-BANG score into the GRACE score improve prediction of in-hospital cardiovascular events in myocardial infarction patients?
Incorporating the STOP-BANG questionnaire into the GRACE score may improve risk stratification for in-hospital cardiovascular events following myocardial infarction.
May support adding STOP-BANG to GRACE for MI risk stratification; leaves open whether it improves discrimination in larger cohorts.
Obstructive sleep apnea (OSA) may impact outcomes in acute coronary syndrome (ACS) patients. The Global Registry of Acute Coronary Events (GRACE) score assesses cardiovascular risk post-ACS. This study evaluated whether incorporating the STOP-BANG score (a surrogate for OSA) enhances GRACE's predictive ability. A total of 227 myocardial infarction (MI) patients were included, with 66 (29.07%) experiencing in-hospital cardiovascular events. Patients with events were older, predominantly male, and had worse clinical markers, including lower hemoglobin and ejection fraction and higher RDW, creatinine, CRP, and GRACE scores (p < 0.001). While STOP-BANG was higher in event patients, risk group classification was non-significant (p = 0.3). Three models were trained: (1) all selected features, (2) GRACE alone, and (3) GRACE + STOP-BANG. The Extra Trees Classifier performed best (ROC-AUC = 0.82). Adding STOP-BANG improved the F1-score, accuracy, and precision but had a non-significant effect on ROC-AUC. The decision curve analysis showed an increased net benefit when STOP-BANG was incorporated. Feature importance analysis ranked STOP-BANG highest in models, reinforcing its relevance. While this study showed that STOP-BANG improved risk stratification, further multicenter validation is needed to confirm its clinical utility in ACS risk models.
No takes yet. Share an insight, caveat, or question.
Shahri et al. (2025) conducted an observational in Myocardial Infarction (n=227). GRACE + STOP-BANG score vs. GRACE score alone was evaluated on ROC-AUC for predicting in-hospital cardiovascular events (95% CI 0.61-0.88, p=0.38). Incorporating the STOP-BANG score into the GRACE risk model improved F1-score, accuracy, and precision for predicting in-hospital cardiovascular events after myocardial infarction, though the increase in ROC-AUC from 0.70 to 0.76 was not statistically significant (p=0.38).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: