Background: Highly transmissible viruses such as influenza are a potential source of nosocomial infections and thereby cause increased patient morbidity and mortality.
Aim: To assess whether influenza virus sequence data can be used to link nosocomial influenza transmission between individuals.
Methods: Dutch A(H1N1)pdm09-positive specimens from one hospital (N = 107) were compared with samples from community cases (N = 685). Gene fragments of haemagglutinin, neuraminidase and PB2 were sequenced and subsequently clustered to detect patients infected with identical influenza viruses. The probability of detecting a second patient was calculated for each hospital cluster against the background diversity observed in hospital and community strains. All clusters were further analysed for possible links between patients.
Findings: Seventeen A(H1N1)pdm09 hospital clusters were detected of which eight had a low probability of occurrence compared with background diversity (P < 0.01). Epidemiological analysis confirmed a total of eight nosocomial infections in four of these eight clusters, and a mother-child combination in a fifth cluster. The nine clusters with a high probability of occurrence involved community cases of influenza without a known epidemiological link.
Conclusion: If a background sequence dataset is available, the detection of hospital sequence clusters that differ from dominant community strains can be used to select clusters requiring further investigation by hospital hygienists before a nosocomial influenza outbreak is epidemiologically suspected.
Copyright © 2012 The Healthcare Infection Society. Published by Elsevier Ltd. All rights reserved.