Methods: We performed a retrospective whole genome sequencing and epidemiologic analysis of institutional influenza outbreaks occurring during 2014-2015 influenza season in Toronto, Canada. Only outbreaks with at least 2 PCR positive samples for influenza H3N2 were considered. We sequenced the two earliest submitted pairs of influenza positive samples from 39 reported institutional outbreaks in long-term care facilities.
Results: 108 of 147 H3N2 influenza outbreaks occurring during the 2014-2015 season had samples submitted to the provincial public health laboratory. 87 outbreaks had two or more PCR positive samples, and of these 39 outbreaks with satisfactory samples for genome sequencing were evaluated. Whole genome sequencing revealed that the majority of sample pairs were highly related. Inclusion of community samples demonstrated that outbreak sources were likely community introductions. Pairwise distance analysis using whole genome and HA specific genes allowed identification of thresholds for discrimination of within and between outbreak pairs, with the AUC’s ranging from 92-93%. The majority of outbreak pairs falling outside determined thresholds could be identified when within outbreak pairwise distance was greater than pairwise distance to a contemporaneous but separate outbreak sample.
Conclusion: Routine whole genome sequencing for defining linked influenza outbreaks in long term care facilities is unlikely to add significant benefit to the current clinical definition. Sequencing may prove useful for investigating specific sources of outbreak introductions.
T. Athey, None
S. Perusini, None
R. Olsha, None
A. Li, None
J. Gubbay, None
B. Hanage, None