Nomogram Prediction for the Risk of Diabetic Foot in Patients With Type 2 Diabetes Mellitus

• Published in: Frontiers in endocrinology (Lausanne) Vol. 13; p. 890057
• Main Authors: Wang, Jie, Xue, Tong, Li, Haopeng, Guo, Shuai
• Format: Journal Article
• Language: English
• Published: Frontiers Media S.A 13.07.2022
• Subjects: diabetic foot; Endocrinology; individual risk prediction model; nomogram; orthopedics; type 2 diabetes mellitus (T2DM)
• ISSN: 1664-2392; 1664-2392
• DOI: 10.3389/fendo.2022.890057

To develop and validate a nomogram prediction model for the risk of diabetic foot in patients with type 2 diabetes mellitus (T2DM) and evaluate its clinical application value.AimsTo develop and validate a nomogram prediction model for the risk of diabetic foot in patients with type 2 diabetes mellitus (T2DM) and evaluate its clinical application value.We retrospectively collected clinical data from 1,950 patients with T2DM from the Second Affiliated Hospital of Xi'an Jiaotong University between January 2012 and June 2021. The patients were divided into training cohort and validation cohort according to the random number table method at a ratio of 7:3. The independent risk factors for diabetic foot among patients with T2DM were identified by multivariate logistic regression analysis. Then, a nomogram prediction model was developed using the independent risk factors. The model performances were evaluated by the area under the receiver operating characteristic curve (AUC), calibration plot, Hosmer-Lemeshow test, and the decision curve analysis (DCA).MethodsWe retrospectively collected clinical data from 1,950 patients with T2DM from the Second Affiliated Hospital of Xi'an Jiaotong University between January 2012 and June 2021. The patients were divided into training cohort and validation cohort according to the random number table method at a ratio of 7:3. The independent risk factors for diabetic foot among patients with T2DM were identified by multivariate logistic regression analysis. Then, a nomogram prediction model was developed using the independent risk factors. The model performances were evaluated by the area under the receiver operating characteristic curve (AUC), calibration plot, Hosmer-Lemeshow test, and the decision curve analysis (DCA).Multivariate logistic regression analysis indicated that age, hemoglobin A1c (HbA1c), low-density lipoprotein (LDL), total cholesterol (TC), smoke, and drink were independent risk factors for diabetic foot among patients with T2DM (P < 0.05). The AUCs of training cohort and validation cohort were 0.806 (95% CI: 0.775∼0.837) and 0.857 (95% CI: 0.814∼0.899), respectively, suggesting good discrimination of the model. Calibration curves of training cohort and validation cohort showed a favorable consistency between the predicted probability and the actual probability. In addition, the P values of Hosmer-Lemeshow test for training cohort and validation cohort were 0.826 and 0.480, respectively, suggesting a high calibration of the model. When the threshold probability was set as 11.6% in the DCA curve, the clinical net benefits of training cohort and validation cohort were 58% and 65%, respectively, indicating good clinical usefulness of the model.ResultsMultivariate logistic regression analysis indicated that age, hemoglobin A1c (HbA1c), low-density lipoprotein (LDL), total cholesterol (TC), smoke, and drink were independent risk factors for diabetic foot among patients with T2DM (P < 0.05). The AUCs of training cohort and validation cohort were 0.806 (95% CI: 0.775∼0.837) and 0.857 (95% CI: 0.814∼0.899), respectively, suggesting good discrimination of the model. Calibration curves of training cohort and validation cohort showed a favorable consistency between the predicted probability and the actual probability. In addition, the P values of Hosmer-Lemeshow test for training cohort and validation cohort were 0.826 and 0.480, respectively, suggesting a high calibration of the model. When the threshold probability was set as 11.6% in the DCA curve, the clinical net benefits of training cohort and validation cohort were 58% and 65%, respectively, indicating good clinical usefulness of the model.We developed and validated a user-friendly nomogram prediction model for the risk of diabetic foot in patients with T2DM. Nomograms may help clinicians early screen and identify patients at high risk of diabetic foot.ConclusionWe developed and validated a user-friendly nomogram prediction model for the risk of diabetic foot in patients with T2DM. Nomograms may help clinicians early screen and identify patients at high risk of diabetic foot.