Abstract
In this study, risk factors for coronary slow flow (CSF) patients were examined, and a clinical prediction model was created. This study involved 573 patients who underwent coronary angiography at our hospital because of chest pain from January 2020 to April 2022. They were divided into CSF group (249 cases) and noncoronary slow flow (NCF) group (324 cases) according to the coronary blood flow results. According to a 7:3 ratio, the patients were categorized into a training group consisting of 402 cases and a validation group consisting of 171 cases. The outcome was assessed by employing multiple logistic regression analysis to examine the factors that influenced it. The model's recognizability was assessed by calculating the consistency index and plotting the receiver operating characteristic curve. Its consistency was assessed by calibration curve, decision curve, and Hosmer-Lemeshow testing goodness-of-fit. The multivariate model included factors such as male, BMI, smoking, diabetes, ursolic acid, and high-density lipoprotein cholesterol. The model validation showed that the consistency index was 0.714, and the external validation set had a consistency index of 0.741. The areas under the curve for the training and external validation sets were respectively 0.730 (95% CI: 0.681-0.779) and 0.770 (95%CI: 0.699-0.841). Nomogram calibration curves indicated intense calibration, and the results of the Hosmer-Lemeshow goodness-of-fit test indicated that χ² = 1.118, P = .572. The nomogram combining various risk factors can be used for individualized predictions of CSF patients and then facilitate prompt and specific treatment.