Introduction

There is a lack of scientific evidence regarding the association between the ACEF score and the prediction of coronary slow flow (CSF).

Aim

To investigate the relationship between the ACEF score and CSF.

Material and methods

This retrospective, single-center study enrolled 257 patients who underwent elective coronary angiography between January 2024 and August 2025. Patients were stratified into two groups: the CSF group (n = 121) and the normal coronary flow (NCF) group (n = 136). Receiver operating characteristic (ROC) curve analysis evaluated the accuracy of the ACEF score, glycated hemoglobin (HbA1c), and low-density lipoprotein cholesterol (LDL-C) in predicting CSF. The incremental value of the ACEF score relative to its components was assessed using C-statistics, IDI, and NRI. Logistic regression was conducted to evaluate independent predictors of CSF presence. The model included sex, hypertension, peripheral artery disease, atrial fibrillation, hemoglobin, white blood cell count, platelet count, HbA1c, triglycerides, LDL-C, high-density lipoprotein cholesterol (HDL-C), and the ACEF score.

Results

In ROC analysis, the cut-off value of the ACEF risk score was 0.94, with 81% sensitivity and 53% specificity for predicting CSF (AUC = 0.763, 95% CI: 0.706–0.821). For HbA1c, the optimal cut-off value was 5.44%, with a sensitivity of 70% and a specificity of 52% to predict CSF (AUC = 0.610, 95% CI: 0.538–0.682). The risk of CSF increased by 28% for every 0.1-unit increase in the ACEF score (OR per 0.1-unit increase: 1.278, 95% CI: 1.168–1.396). The ACEF score demonstrated superior discrimination (AUC: 0.763 vs. 0.729, p = 0.018) and significant incremental predictive value (IDI = 0.052, p < 0.001) compared to its individual components.

Conclusions

This study suggests that the ACEF score is closely associated with the presence of CSF, independent of its individual components. The discriminatory ability of the ACEF score was superior to that of HbA1c in patients with CSF.