Clinical Predictors of Intubation and In-Hospital Death in Lung Cancer Patients Admitted for Acute Radiation Pneumonitis: A Healthcare Cost and Utilization Project (HCUP) Analysis

Abstract

Acute radiation pneumonitis is a dose-limiting toxicity of thoracic radiotherapy which can result in mortality. The purpose of this study was to identify predictors of intubation and in-hospital death among patients with lung cancer admitted for acute radiation pneumonitis. The Healthcare Cost and Utilization Project (HCUP) National Inpatient Sample (NIS) database was queried from 2012 through 2015 to capture hospitalized adult patients with a principal diagnosis of acute radiation pneumonitis. Patients were included in this study if they also had a diagnosis of lung cancer. Variables examined included age, gender, race, smoking status, diagnosis of interstitial lung disease (ILD), congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), connective tissue disease/rheumatoid arthritis, median household income by household zip code, and season/weekday of admission. Chi-squared and multivariable logistic regression modeling were used to determine predictors of in-hospital death, intubation, and length of stay (LOS). Of 674 patients who met inclusion criteria, 6 (0.9%) had a diagnosis of ILD, 113 (16.7%) had a diagnosis of CHF, and 296 (43.9%) were female. The mean age was 70.3 years (range, 35-90), and the average LOS was 6.4 days (range, 0-64). During admission, 53 (7.9%) patients expired, and 72 (10.7%) patients required intubation. Predictors of death on univariable analysis were ILD and higher income quartile by household zip code. On multivariable logistic regression, only ILD (HR 12.5, 95% CI 2.5-63.6, p = 0.002) remained a significant predictor of death. Significant predictors of intubation on univariable analysis were CHF and higher income quartile by household zip code. CHF (HR 2.3, 95% CI 1.3-4.0, p = 0.005) and higher income quartile by household zip code (p = 0.027) remained significant on multivariable logistic regression. Predictors of extended LOS on multivariable analysis were CHF (HR 2.1, 95% CI 1.4 – 3.1, p = 0.001) and COPD (HR 1.6, 95% CI 1.1 – 2.2, p = 0.009). In a large cohort of lung cancer patients hospitalized with a principal diagnosis of acute radiation pneumonitis, ILD was a significant predictor of in-hospital death, while co-morbid CHF was an important predictor of intubation. These findings confirm the importance of comorbid diagnoses in predicting outcomes during hospitalization for lung cancer patients treated with radiotherapy.