Objective To investigate the influencing factors of malnutrition in inpatients with pulmonary tuberculosis and construct a risk prediction model of malnutrition.

Methods A total of 322 pulmonary tuberculosis patients were selected by convenience sampling method, and divided into the non-risk group (238 cas) and risk group (84 cases) according to the NRS 2002 nutritional risk screening score. The influencing factors of malnutrition in hospitalized patients with pulmonary tuberculosis were analyzed, and a malnutrition risk prediction model was constructed. The prediction effect of model was verified by ROC curve analysis.

Results The incidence of malnutrition was 26.09% (84/322) in 322 patients. There were statistically significant in the age, exercise status, initial treatment, BMI, leg circumference, clinical symptoms (sputum, chest tightness, asthma, fever and poor appetite), number of lung lobes involved in the lesion and laboratory indicators (Hemoglobin, prealbumin and lymphocyte percentage) between the two groups (P ＜ 0.05 to P ＜ 0.01). The Calf circumference, PA, exercise status and BMI were the independent influencing factors of malnutrition in hospitalized patients with pulmonary tuberculosis (P ＜ 0.05 to P ＜ 0.01). The results of ROC curve analysis showed that the AUC of the malnutrition prediction model for inpatients with pulmonary tuberculosis was 0.948, and the sensitivity and specificity were 93.2% and 82.67%, respectively.

Conclusions The incidence of malnutrition risk in hospitalized patients with pulmonary tuberculosis is relatively high. Clinically, patients with BMI ＜ 18.5 kg/m2 and less exercise should be paid attention to, the calf circumference of patients should be monitored, and the .patients with a large decrease in PA in the short termshould also be paid attention to. Based on this, the construction of a risk prediction model has a good predictive value for malnutrition in hospitalized patients with pulmonary tuberculosis.