Abstract:

Objective: To explore the influencing factors of malnutrition in patients with pulmonary tuberculosis complicated with diabetes mellitus, construct and verify a nomogram prediction model. Methods: The clinical data of 401 patients with tuberculosis combined with diabetes admitted to the Tuberculosis Department of Nanjing Second Hospital from October 2021 to September 2023 were collected with convenience sampling. The patients were divided into modeling group (n=281) and validation group (n=120) according to a ratio of 7∶3. Logistic regression analysis was performed to construct a nomogram prediction model. Area under curve (AUC) and Hosmer-Lemeshow goodness of fit test were used to evaluate the prediction efficiency and calibration degree of the model. Results: Age (OR=3.796, 95%CI: 1.159-12.627), duration of disease >1 month (OR=5.711, 95%CI: 1.879-17.274), glycosylated hemoglobin (OR=5.951, 95%CI: 1.517-23.269), Charlson comorbidity index (OR=8.079, 95%CI: 2.345-27.681) and FRAIL (fatigue, resistance, ambulation, illness and loss) score ≥3 (OR=9.145, 95%CI: 2.404-34.172) were independent risk factors for malnutrition. The above variables were used to construct a nomogram prediction model. Hosmer-Lemeshow test of the model showed that P=0.625, AUC=0.897, the Youden index was 0.584 and the optimal critical value was 0.615. The sensitivity of the verification group was 67.5%, the specificity was 93.8%, and the prediction accuracy was 85.0%. Conclusion: The nomogram model constructed in this study has certain predictive value and clinical applicability, could provide reference for early identification of malnutrition and thereafter taking targeted nutrition intervention measures in clinical work.

Key words: Tuberculosis, pulmonary, Diabetes mellitus, Malnutrition, Forecasting, Nomograms