To evaluate the applicability of Nutritional Risk Screening 2002 (NRS2002), Mini Nutritional Assessment (MNA), and Malnutrition Universal Screening Tool (MUST) in hospitalized patients with multidrug-resistant pulmonary tuberculosis (MDR-PTB), using the Global Leadership Initiative on Malnutrition (GLIM) criteria as the reference standard.

A descriptive study was conducted, consecutively enrolling 350 patients diagnosed with MDR-PTB at Xi′an Chest Hospital between June 2019 and December 2021. Nutritional risk screening was performed using the 3 toolsSensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and agreement were calculated for each tool. Receiver operating characteristic (ROC) curves were constructed, and the area under the curve (AUC) was computed. Univariate and multivariate logistic regression analyses were conducted to assess the association between nutritional status and clinical outcomes, with sputum culture conversion at 2 months and the occurrence of adverse events as primary endpoints.

Among the 350 MDR-PTB patients, 204 cases (58.29%) were diagnosed with malnutrition according to the GLIM criteria, and 146 cases (41.71%) were well-nourished. The nutritional risk detection rates of various screening tools were as follows: MNA in 294 cases (84.00%), NRS2002 in 228 cases (65.14%), and MUST in 240 cases (68.57%). The screening time required for MNA [(3.40 ± 2.07) minutes] was significantly longer than that for NRS2002 [(2.23±0.76)minutes] and MUST [(1.54±0.33) minutes], with a statistically significant difference (F=0.578, P<0.001). According to GLIM criteria, 58.29% of patients were diagnosed with malnutrition. Performance analysis revealed that MNA had the highest sensitivity (99.51%) but the lowest specificity (37.67%). NRS2002 and MUST demonstrated higher specificity (66.44% vs. 63.70%) and greater AUC values (0.771 vs. 0.777). Multivariate binary logistic regression indicated that, compared with well-nourished patients, those identified as malnourished by GLIM criteria (OR=20.511, 95%CI: 4.509~93.308) and those at high risk by NRS2002 (OR=3.492, 95%CI: 1.147~10.627) had a significantly higher risk of non-conversion in sputum culture at 2 months. Similarly, the GLIM malnutrition group (OR=2.951, 95%CI: 1.023~8.509) and the MNA high-risk group (OR=17.213, 95%CI: 1.794~165.179) were associated with a higher risk of adverse treatment outcomes.

Among the three nutritional screening tools, MNA showed the highest sensitivity, while NRS2002 exhibited the highest specificity, MUST demonstrated better overall screening performance. For focusing on short-term clinical outcomes, NRS2002 and MUST have higher screening efficacy, while MNA is superior for early identification of nutritional risk and assessment of long-term clinical outcomes, and may be more suitable for the clinical nutritional management strategy of MDR-PTB patients.