Clinical Epidemiology and Global Health (Oct 2021)
The factors predicting mortality in pulmonary tuberculosis with acute respiratory failure
Abstract
Background: Acute respiratory failure might complicate pulmonary tuberculosis (TB), leading to a high mortality rate. The inadequate number of respiratory intensive-care unit (RICU) for TB patients might contribute to the mortality rate. However, other factors should be considered as predictors of mortality. This study aimed to analyse the predictors of mortality in hospitalized pulmonary TB patients with acute respiratory failure. Methods: This prospective cohort study was implemented in a tertiary hospital in Surabaya, East Java, Indonesia. Demographic data, acid-fast bacilli (AFB) sputum smear, radiographic conclusion, biochemical analysis of blood, and clinical outcomes from active pulmonary TB patients who had acute respiratory failure were collected. A chi-square test was run to analyse the relationship between variables. The statistically significant variables were then evaluated with multivariate logistic regression analysis. Odds ratio (OR) and their 95% confidence intervals (CI) indicated factors that predicted mortality. Results: 35 patients had acute respiratory failure among the 234 pulmonary TB patients who were being hospitalized in the pulmonary ward from July 2019 to September 2019. The average age of patients was 50.2 ± 15.1 years-old, 45.7% had Diabetes Mellitus, albumin level were 3.07 ± 0.56 g/dL, median procalcitonin 0.73 ng/mL, median Body Mass Index (BMI) 20.7 kg/m2, 54.3% with PaO2/FiO2 ratio 200–300 mmHg, mortality rate 42.9%. Multivariate logistic regression analysis illustrated lower albumin level (<3 g/dL)(OR 12.254 (CI 95% [1.924–78.062]; p = 0.008) and Diabetes Mellitus (OR 8.448, CI 95% [1.350–52.872]; p = 0.023) were the independent predictor factors correlated to mortality. Conclusions: Pulmonary TB cases that were aggravated with acute respiratory failure showed a high mortality rate. Lower albumin levels and Diabetes Mellitus were significantly related to in-hospital mortality.
