Predictive Mortality of the Prognostic Nutritional Index Combined with APACHE II Score for Critically Ill Tuberculosis Patients.

Abstract

High mortality rates are commonly found in critically ill patients with tuberculosis (TB), which is due partially to limitations in the existing prognostic evaluation methods. Therefore, we aimed to find more effective prognostic evaluation tools to reduce the mortality rate. Data from critically ill patients with TB admitted to the intensive care unit of The Second Hospital of Nanjing, Nanjing, China, between January 2020 and December 2022 were analyzed retrospectively. A total of 115 patients were enrolled and divided into a survival group (n = 62) and a death group (n = 53) according to 30-day survival. Univariate and least absolute shrinkage and selection operator (LASSO) regression analyses were used to investigate the risk factors for 30-day death in critically ill patients with TB. A prediction model for risk of 30-day mortality was developed for critically ill patients with TB in the intensive care unit. The LASSO regression model showed that the prognostic nutritional index (PNI) and Acute Physiology and Chronic Health Status (APACHE II) scores on the third day after admission to the intensive care unit were independent risk factors for 30-day mortality in critically ill patients with TB (P <0.05). The area under the curve value and that PA3 represents the combination of the PNI and APACHE II score on the third day, which was 0.952 (95% CI: 0.913-0.991, P <0.001), was significantly higher than that of the PNI or the APACHE II score on the third day. The new model is as follows: PA3 = APACHE II score (on the third day) × 0.421 - PNI × 0.204. The PNI combined with the APACHE II score on the third day could well predict the 30-day mortality risk of critically ill patients with TB.