Secular variation in United States rotavirus disease rates and serotypes: implications for assessing the rotavirus vaccination program.

BACKGROUND: Since 2006, we have conducted population-based surveillance for rotavirus disease in children seen in hospitals and emergency departments (EDs) in Monroe County, NY (Rochester), Hamilton County, OH (Cincinnati), and Davidson County, TN (Nashville).
METHODS: During the 2006 and 2007 rotavirus seasons, clinical information and stool specimens were obtained from county children who were <3 years presenting with diarrhea and/or vomiting to the hospital or ED of the only children's hospital in each county. Specimens were tested for rotavirus and genotyped, and rates of hospitalization and ED visits were calculated.
RESULTS: While aggregate rotavirus hospitalization rates for the 3 sites were similar in 2006 and 2007 (22.5/10,000 and 26.8/10,000, respectively), individual rates for the 3 counties differed considerably. The rotavirus hospitalization rate in Rochester between 2006 and 2007 increased 3-fold, but decreased by 33% in Cincinnati and 41% in Nashville over the 2 study years. G1 strains accounted for >80% of strains at all 3 sites in 2006. However, in 2007, the uncommon P[8], G12 strain was detected in 69% of Rochester specimens, while the P[8], G1 strain remained predominant in the other 2 sites. No subjects received rotavirus vaccine in 2006 and coverage with 2 to 3 vaccine doses reached 15% in all 3 communities by June 2007.
CONCLUSIONS: During the 2006 and 2007 rotavirus seasons, with only limited vaccine use, remarkable variability was observed in the population-based rates of severe rotavirus and in the rotavirus serotypes across the 3 sites. This natural secular variability in rotavirus disease must be considered in the assessment of the impact of vaccine on disease rates and rotavirus serotypes.