Temporal association of Streptococcus, Staphylococcus, and parainfluenza pediatric hospitalizations and hospitalized cases of Henoch-Schönlein purpura.

Abstract

To determine if hospitalizations for specific infectious exposures are associated with hospital admissions for Henoch-Schönlein purpura (HSP). We conducted a retrospective cohort study using administrative data of children admitted to 40 children's hospitals between January 1, 2002, and December 31, 2008. We examined the association of standardized rates of group A ß-hemolytic Streptococcus (GABS), Staphylococcus aureus, parainfluenza, influenza, adenovirus, and respiratory syncytial virus (RSV)-associated hospital admissions with standardized rates of HSP hospital admissions on a month by month basis using autoregressive moving average process models to account for temporal autocorrelation and clustering by hospital. Among the 3,132 admissions for HSP observed over the 7-year study period, hospital admissions were most frequent September through April, but with substantial variability between hospitals for each month. Accounting for these month by month differences within each hospital, the rate of HSP admissions in a given month increased significantly as the standardized rates of GABS (p = 0.01), S. aureus (p < 0.01), and parainfluenza (p = 0.03) admissions increased. Our results demonstrate a local month by month temporal association between hospitalization for GABS, S. aureus, and parainfluenza and hospitalization for HSP. Future investigations will be required to determine causality.