Predictors of mortality in patients of acute exacerbation of chronic obstructive pulmonary disease: A prospective observational study

Background: Acute exacerbations of chronic obstructive pulmonary disease (AECOPD) result in increased morbidity, mortality and tremendous socioeconomic burden. Predicting inhospital mortality may aid prognostication, planning for the site of care, i.e., ward versus intensive care units, and better individualization of treatment. Aim of Study: This study aims in predicting inhospital mortality in patients of AECOPD based on the parameters measured at the time of admission. Methods: Known COPD patients in acute exacerbations admitted in a tertiary care hospital were interrogated for clinical history and examination. All relevant laboratory tests including arterial blood gas analysis, complete blood count, liver and renal function tests, and random blood sugar were done. Based on outcome, patients were grouped into survivors and nonsurvivors. Parameters recorded were then subjected for univariate analysis to get their statistical significance. All significant variables on univariate analysis were then analyzed further with multivariate analysis. Results: Out of the total 140 patients included in the study, 24 (17%) died during their hospital stay. Of the various acute-phase parameters recorded at the time of the admission, only five qualified to be predictors of inhospital mortality based on univariate and multivariate analyses. These were partial pressure of carbon dioxide in arterial blood (PaCO2) (odds ratio [OR], 95% confidence interval [CI] =1.067, 0.993-1.146), pH (OR, 95% CI = 0.001, 0.001-0.584), serum glutamic pyruvate transaminase (SGPT) (OR, 95% CI = 1.032, 1.006-1.059), sodium (OR, 95% CI = 0.779, 0.689-0.881), and random blood sugar (OR, 95% CI = 1.018, 1.007-1.029). With these five factors combined, area under receiver operating characteristic (ROC) curve was 0.9684, sensitivity 79.18%, specificity 96.55%, positive predictive value 82.61%, negative predictive value 95.73%, and correctly classify acute exacerbation in 93.57%. Good survival can be expected if these parameters are within normal limits. Conclusion: pH, PaCO2, SGPT, serum sodium, and random blood sugar at the time of admission are independent predictors of mortality in patients of AECOPD. These can be helpful in developing a prediction tool of inhospital mortality in such patients.