Indicators of recurrent hospitalization for pneumonia in the elderly

OBJECTIVES: To identify modifiable risk factors of late unplanned readmissions for elderly with community-acquired pneumonia. DESIGN: A case-control study. SETTING: Three university-affiliated tertiary-care hospitals. PARTICIPANTS: Two hundred four case-control pairs. Case patients referred to all patients readmitted with pneumonia 30 days to 1 year after discharge. Control subjects were matched for age, admission date, and residence before admission. MEASUREMENTS: Baseline sociodemographic information, clinical data, activity of daily living (ADLs) information, and Charlson Comorbidity Index score were obtained. The Pneumonia Severity Index was calculated with swallowing dysfunction and pattern and extent of radiographic abnormalities, antimicrobial coverage, and total duration recorded. RESULTS: Median time to readmission was 123 days (interquartile range=65-238 days). Readmission was not associated with increased severity or length of hospital stay. In a Cox proportional hazards regression model, swallowing dysfunction (hazard ratio (HR)=2.15, 95% confidence interval (CI)=1.46-2.97), current smoking (HR=2.04, 95% CI=1.48-2.82), use of tranquilizers (HR=1.5, 95% CI=1.02-2.22), and lower ADL scores (HR=1.06, 95% CI=1.02-1.10) were independently associated with readmission for pneumonia. The receipt of angiotensin-converting enzyme inhibitors (HR=0.46, 95% CI=0.27-0.78) and prior pneumococcal vaccination (HR=0.59, 95% CI=0.42-0.82) had a protective effect. CONCLUSION: Although there are limited effective measures to improve functional status, preventive strategies that include smoking cessation and pneumococcal vaccination should be actively pursued. Routine evaluation of swallowing dysfunction and use of pharmacological agents to improve the cough reflex deserve further evaluation in multicenter controlled trials.