Abstract
BACKGROUND: Studies have demonstrated reduced rotavirus vaccine effectiveness (VE) in resource-limited settings. Enteropathogen coinfections in rotavirus cases have been hypothesized to contribute to the lower VE in such settings. We sought to determine if coinfections affect rotavirus VE in Botswana. METHODS: Between June 2013 and April 2015, children <60 months old, presenting with severe gastroenteritis at 4 hospitals as part of a national rotavirus surveillance were enrolled. Rotavirus enzyme immunoassay (EIA)-positive samples were tested with an in-house real-time polymerase chain reaction (PCR) panel that detected 9 pathogens and a commercial 15 multiplex PCR gastrointestinal pathogen panel. Coinfection was defined as detection of rotavirus plus 1 of the 5 pathogens with the highest attributable fractions for diarrhea. Vaccine status was compared between rotavirus case patients and non-rotavirus "test-negative" controls. VE was also calculated restricting cases to those with rotavirus as the only pathogen detected. RESULTS: Two hundred and forty-two children tested rotavirus EIA positive, and 368 children were negative. Of the 182 rotavirus EIA-positive samples tested with the gastrointestinal pathogen panel assay, coinfections were detected in 60 (33%). The overall adjusted 2-dose VE was 59% (95% confidence interval [CI]: 27-77) in the rotavirus coinfection group and 51% (95% CI: -14 to 79) in the rotavirus monoinfection subgroup. Using in-house multiplex PCR panel, of 213 rotavirus EIA-positive subjects, coinfections were detected in 98 samples (46%). The overall adjusted VEs for 2 doses were 48% (95% CI: -2 to 74) and 62% (95% CI: 25-80) in rotavirus monoinfection subgroup. CONCLUSIONS: We could not find evidence of an effect of enteric coinfections on the effectiveness of rotavirus vaccine.