Clinical risk score to predict poor coronary collateralization in type 2 diabetic patients with chronic total occlusion

Abstract

Background

This study sought to develop and externally validate a score that predicts the probability for poor coronary collateralization (CC) in stable angina patients with type 2 diabetes mellitus (T2DM).

Methods

Clinical and laboratory variables were collected on admission in 1022 T2DM patients with chronic total occlusion (CTO). Coronary collaterals with Rentrop score 0 or 1 were considered as poor CC. Multivariable logistic regression analysis was used to identify independent predictors for poor CC. The external validation cohort comprised 234 T2DM patients with CTO selected randomly from an independent external center.

Results

Eight predictors were independently associated with poor CC and applied to construct the risk model. A score incorporating these factors predicted poor CC, ranging from 7% when all factors were absent to 97% when ≥ 7 factors were present. Internal validation showed an AUC of 0.748 (95%CI, 0.695–0.795) and the external validation had an AUC of 0.754 (95%CI, 0.694–0.808). A cumulative predictive score was developed by summing points assigned to each factor based on its regression coefficient. Smoking and neutrophil > 6.5 × 10⁹/L were assigned 3 points, female gender, hypercholesterolemia, and eGFR < 60 mL/min/1.73 m² were assigned 2 points, age > 65 years, hypertension, and HbA1c > 6.5% were assigned 1 point. The optimal cutoff score was 4 for predicting poor CC with sensitivity 75.4% and specificity 64.1%.

Conclusions

We have demonstrated a risk score based on clinical and laboratory characteristics providing an easy-to-use tool to predict poor CC in T2DM patients with stable coronary artery disease.

Clinical trial number

NCT06054126 Date of registration: September 19th, 2023.