Factors predicting length of stay in bronchiolitis

Introduction: Despite advances in medical knowledge, the treatment of viral bronchiolitis is mainly supportive. Antiviral therapies are being investigated in clinical trials. Identifying population-attributable risk factors for RSV hospitalization may help prioritizing targeted treatment. Aim: To utilize MDClone, a data acquisition tool, to examine factors associated with the risk of hospitalization and length of stay (LOS) in bronchiolitis. Methods: A single tertiary medical center retrospective study. Infants discharged with a diagnosis of bronchiolitis between January 2001 and March 2019 were included. Demographic, clinical, laboratory, microbiologic parameters and co-morbidities were collected. Correlations with the risk of hospitalization and LOS were examined. Results: A total of 4793 infants with bronchiolitis, 3851 (80.3%) previously healthy, were seen; 975 visited emergency room only; 3311 were hospitalized in pediatric wards and 507 required pediatric intensive care unit. O-2 saturation, age and fever correlated with the risk of hospitalization (OR = 0.703, p < 0.0001, OR = 0.4, p = 0.024 and OR = 2.388, p < 0.0001, respectively). Saturation, fever, gestational age and birth weight correlated with LOS (r = -0.283, p = 0.000; r = 0.16, p = 0.000; r = -0.12, p = 0.00; and r = -0.117, p = 0.00, respectively). Rates of hospitalization were higher (81.1% vs. 75.6%, p = 0.0008) and LOS was longer (median 2.97 vs. 2.73 days, p < 0.001) in Arabs than in Jews. In a multivariate model, saturation, fever, gestational age and age predicted LOS. Saturation and ethnicity predicted LOS for previously healthy infants. Prematurity and cardiac anomalies increased LOS (p = 0.016 and p < 0.0001, respectively). Conclusions: Population-based data may enable predicting disease severity and LOS in bronchiolitis. Focusing on children at greatest risk may aid targeting new therapies.