Letters to the Editor

Abstract

4 October 2006 Dear Editor, NOSOCOMIAL ROTAVIRUS INFECTION IN AN AUSTRALIAN CHILDREN’S HOSPITAL There have been some exciting developments in the field of rotavirus vaccination with the successful completion of large phase III trials demonstrating the safety and efficacy of both an attenuated human rotavirus vaccine (Rotarix, GlaxoSmithKline, Boronia, Australia)1 and a pentavalent human-bovine reassortant vaccine (RotaTeq, CSL Limited/Merck & Co Inc, Melbourne, Australia).2 These vaccines were licensed in Australia by the Therapeutic Goods Administration in the first half of 2006 and applications made to the Pharmaceutical Benefits Advisory Committee have resulted in a recommendation for funding under the National Immunisation Program. Rotavirus vaccination is likely to have an impact on community and hospital-treated episodes of rotavirus gastroenteritis in Australia and elsewhere. Vaccination of infants is estimated to prevent around 85–100% of hospitalisations due to rotavirus gastroenteritis in the first year following vaccination.1, 2 Studies from other countries have shown that a significant burden of hospital-treated rotavirus infection is acquired in hospital nosocomially. Nosocomial infection rates have not been studied in Australia since a 1989 report that 14% of children <3 years old admitted to general paediatric wards at the Adelaide Children’s Hospital developed nosocomial rotavirus infection.3 We have collected data on nosocomial rotavirus infections at the Children’s Hospital at Westmead since 1996. Since 1999, all faecal specimens from inpatients with diarrhoea sent for microbiological analysis have been tested for rotavirus using the Rotazyme enzyme-linked immunosorbent assay (ELISA), whether requested or not. Between 1999 and 2004, the incidence density of proven nosocomial rotavirus infection (defined as rotavirus diarrhoea with onset at least 72 h after admission and up to 24 h after discharge) in children <2 years old admitted during the winter rotavirus season, varied between 6.7 and 14.8 cases per 10 000 rotavirus-free days in hospital. When cases of possible nosocomial infection (defined as rotavirus diarrhoea with onset between 24 and 72 h after admission or between 24 and 72 h after discharge) were added to the proven cases, the incidence density for the study period varied between 9.2 and 19.3 cases per 10 000 rotavirus-free days in hospital. Over the study period, 14.5% of all ELISA confirmed rotavirus infections were proven to be hospital-acquired, and 19.4% were either proven or probably hospital-acquired. These ratios of nosocomially acquired compared with all confirmed rotavirus infections are at the lower end of the range reported in a systematic review of 14–51%,4 although these values are likely to be dependent upon the age ranges studied and the institutional referral patterns for hospital and community stool specimens. Because we did not attempt to systematically test stools from all hospitalised children or to follow-up children after discharge, it is likely that the true rates of nosocomial infection may be significantly higher than the rates reported here. In one report, 39.3% of nosocomial infections were asymptomatic and 67.6% occurred after discharge.5 Asymptomatic infection is important in that it probably contributes to the subsequent transmission of the virus. In other institutions, nosocomial rotavirus infection has been found to contribute significantly to patient length of stay and to the costs of care,6 although we have not attempted to directly calculate these parameters. Despite careful attention to infection control practices, nosocomial rotavirus infections remain a significant problem in paediatric hospitals. Universal rotavirus vaccination in infancy is likely to reduce the number of hospitalised children with community-acquired rotavirus and may have a significant impact on reducing nosocomial infection rates in Australian paediatric hospitals.