Routine laboratory testing data for surveillance of rotavirus hospitalizations to evaluate the impact of vaccination

Abstract

Objective: The recent implementation of a rotavirus vaccination program in the United States makes it imperative to assess the impact of immunization on the incidence of severe rotavirus disease leading to hospitalization. Active surveillance for laboratory-confirmed rotavirus hospitalizations is the ideal approach for surveillance, but requires substantial resources to implement. We examined laboratory and hospital discharge data for 2 tertiary care pediatric hospitals to assess the utility of routine laboratory testing data for surveillance of rotavirus gastroenteritis and to estimate rotavirus disease burden.

Design: We obtained all discharge records of hospitalizations for acute gastroenteritis among children <5 years of age at Children's Mercy Hospital (CMH), Kansas City, from July 2000 to June 2005 and at Children's Hospital of Philadelphia (CHOP) from July 2004 to June 2006. We linked these discharge records to laboratory results of rotavirus testing to evaluate epidemiologic differences in children who were tested and not tested for rotavirus and to estimate overall rotavirus burden by extrapolating clinical testing results to the untested group.

Results: At CMH, of the 3702 children with acute gastroenteritis, 69% (n = 2552) were discharged during the winter (January through May) months, when rotavirus is most common. Similarly, at CHOP, 62% (n = 779) of the 1261 gastroenteritis discharges occurred during the winter months. During these months, 47% (n = 1197 of 2552) of the discharges at CMH and 56% (n = 438 of 779) of the discharges at CHOP were tested for rotavirus and of those tested, 71% (n = 853 of 1197) and 55% (n = 242 of 438) were positive, respectively. At both hospitals, children with and without rotavirus testing had similar gender and race/ethnicity, but the rate of testing differed by age at CHOP and by month of admission at CMH. After adjusting for these differences, we estimate that 56%-70% of winter and 34%-48% of year-round gastroenteritis in children <5 years can be attributable to rotavirus. Overall, 3%-5% of all hospitalizations in children <5 years of age were caused by rotavirus.

Conclusions: Sentinel hospitals where a large proportion of children hospitalized for gastroenteritis are routinely tested for rotavirus could provide a useful and cost-efficient platform to complement ongoing active surveillance efforts to evaluate the impact of rotavirus vaccination. The data reaffirm the substantial burden of rotavirus hospitalizations in US children and the potential health benefits of vaccination.