A Nomogram for Predicting the Possibility of Peripheral Neuropathy in Patients with Type 2 Diabetes Mellitus

Abstract

Background and purpose: Diabetic peripheral neuropathy (DPN) leads to ulceration, noninvasive amputation, and long-term disability. This study aimed to develop and validate a nomogram for forecasting the probability of DPN in type 2 diabetes mellitus patients.

Methods: From February 2017 to May 2021, 778 patients with type 2 diabetes mellitus were included in this study. We confirmed the diagnosis of DPN according to the Toronto Expert Consensus. Patients were randomly divided into a training cohort (n = 519) and a validation cohort (n = 259). In the training cohort, univariate and multivariate logistic regression analyses were performed, and a simple nomogram was built using the stepwise method. The receiver operating characteristic (ROC), calibration curve, and decision curve analysis were computed in order to validate the discrimination and clinical value of the nomogram model.

Results: About 65.7% and 72.2% of patients were diagnosed with DPN in the training and validation cohorts. We developed a novel nomogram to predict the probability of DPN based on the parameters of age, gender, duration of diabetes, body mass index, uric acid, hemoglobin A1c, and free triiodothyronine. The areas under the curves (AUCs) of the nomogram model were 0.763 in the training cohort and 0.755 in the validation cohort. The calibration plots revealed well-fitted accuracy between the predicted and actual probability in the training and validation cohorts. Decision curve analysis confirmed the clinical value of the nomogram. In subgroup analysis, the predictive ability of the nomogram model was strong.

Conclusions: The nomogram of age, gender, duration of diabetes, body mass index, uric acid, hemoglobin A1c, and free triiodothyronine may assist clinicians with the early identification of DPN in patients with type 2 diabetes mellitus.

Keywords: diabetic peripheral neuropathy; nomogram; risk factor; type 2 diabetes.

Figure 1

The nomogram prediction model for DPN risk. Data were collected in patients with type 2 diabetes, and the position of each variable on the corresponding axis was confirmed. A vertical line was drawn from each variable’s position to the top “Points” axis to collect the variable’s score. Then, the users added up the score of each variable to acquire the total score on “Total Points”. They then drew a vertical line from the total points axis to the bottom scale to assess the DPN risk. For example, a 70-year-old (60 points) male (20 points) patient sufferers from a 5-year history of type 2 diabetes (10 points), has 30 kg/m² of BMI (10 points), 500 umol/L of UA (30 points), 7 of HbA1c (10 points), and 8 pmol/L of FT3 (50 points). He receives a total score of 190 points when all of the points are added. The estimated probability of DPN for this patient was calculated. DPN, Diabetic peripheral neuropathy; BMI, body mass index; UA, uric acid; HbA1c, hemoglobin A1c; FT3, free triiodothyronine.

Figure 2

ROC curves of the nomogram in training (A) and validation (B) cohort. The black dot represents the best cut-off value: 1.087 for the training cohort and 1.021 for the validation cohort.

Figure 3

Calibration plot. The x-axis represents the nomogram-predicted probability, and the y-axis represents the actual probability of diabetic peripheral neuropathy. (A) A nomogram calibration plot in the training cohort. A perfect prediction would fall along the 45-degree line (“ideal” line). The “apparent” line represents the training cohort, and the solid black line represents bias corrected by bootstrapping (500 repetitions), indicating observed nomogram performance. (B) A nomogram calibration plot in the validation cohort. A perfect prediction would fall along the 45-degree line (“ideal” line). The “apparent” line represents the validation cohort, and the solid black line represents bias corrected by bootstrapping (500 repetitions), indicating observed nomogram performance.

Figure 4

The decision curve analysis of the nomogram predicts the possibility of DPN in training (A) and validation (B) groups. The horizontal solid black line represents the net benefit when no patient was considered to have DPN. The gray line represents the net benefit when all patients were identified as suffering from DPN. The area along the red line (nomogram model line), the black line, and the gray line representing the net benefit of the nomogram was significantly higher than that of the "no patient" and " all patients" schemes, suggesting that the nomogram has good clinical applicability.