Effectiveness of neuraminidase inhibitors in preventing hospitalization during the H1N1 influenza pandemic in British Columbia, Canada

OBJECTIVES In British Columbia (BC), Canada, neuraminidase inhibitors (NIs) were publicly funded during the 2009 A(H1N1)pdm09 pandemic for treatment of high-risk patients and/or anyone with moderate-to-severe illness. We assessed antiviral effectiveness (AVE) against hospitalization in that context. METHODS A population-based cohort study was conducted using linked administrative data. The cohort included all individuals living in BC during the study period (1 September to 31 December 2009) with a diagnostic code consistent with influenza or pandemic H1N1. The main study period pertained to the second-wave A(H1N1)pdm09 circulation (1 October to 31 December 2009), with sensitivity analyses around the more specific pandemic peak (18 October to 7 November). Exposure was defined by same-day NI prescription. The main outcome was all-cause hospitalization within 14 days of the outpatient influenza diagnosis. Cox proportional hazards models assessed AVE with 1 : 1 propensity-score matching and covariate adjustment. RESULTS After matching, there were 304/58,061 NI-exposed and 345/58,061 unexposed patients hospitalized during the main study period. The very young [<6 months (35.0; 95% CI 16.7-73.4)], the old [65-79 years (13.7; 95% CI 10.1-18.6)] and the very old [≥80 years (38.7; 95% CI 26.6-56.5)] had the highest hospitalization rate per 1000 patients overall. Fully adjusted AVE against all-cause hospitalization during the main study period was 16% (95% CI 2%-28%), similar to the pandemic peak (15%; 95% CI -4%-30%). CONCLUSIONS The use of NIs was associated with modest protection against hospitalization during the 2009 pandemic, but appeared underutilized in affected age groups with the highest hospitalization risk.