Mortality association between obesity and pneumonia using a dual restricted cohort model

Abstract Background An obesity survival paradox has been reported among obese patients with pneumonia. Aims To determine the impact of obesity on pneumonia outcomes and analyze the correlation between in-hospital all-cause mortality and obesity among patients with pneumonia. Methods The United States Nationwide Readmissions Database (NRD) was retrospectively analyzed for patients with pneumonia from 2013 to 2014. We used a step-wise restricted and propensity score matching cohort model (dual model) to compare mortality rates and other outcomes among pneumonia patients based on BMI. Mortality was calculated by a Cox proportional hazard model, adjusted for potential confounders with propensity score matched analysis. Results A total of 70,886,775 patients were registered in NRD during the study period. Of these, 7,786,913 patients (11.0%) were considered obese and 1,652,456 patients (2.3%) were admitted to the hospital with pneumonia. Based on the step-wise restricted cohort model, the hazard ratio comparing the mortality rates among obese pneumonia patients to mortality rates among normal BMI pneumonia patients was 0.75 (95% CI 0.60–0.94). The propensity score matched analysis estimated a hazard rate of 0.84 (95% CI 0.79–0.90) and the hazard ratio estimated from the dual model was 0.82 (95% CI 0.63–1.07). Conclusions With the application of a dual model, there appears to be no significant difference in mortality of obese patients with pneumonia compared to normal BMI patients with pneumonia.