Background: The efficacy of influenza vaccination in older nursing home residents is frequently overestimated due to frailty selection bias. Limited data exist to examine this issue.
Methods: We conducted a prospective cohort study from December 2009 to November 2010 to evaluate the efficacy of influenza vaccination in old nursing home residents with respect to their functional status. Participants were stratified according to the Barthel Index (BI) into good functioning (GF; BI > 60), intermediate functioning (IF; BI = 5-60), and poor functioning (PF; BI = 0). Participants were vaccinated by monovalent H1N1 2009 and trivalent seasonal influenza vaccinations (H1N1-TIV), TIV alone, or remained unvaccinated by choice. The associations between all-cause mortality, vaccination efficacy, and functional status were examined.
Results: A total of 711 older nursing home residents were enrolled (GF group: N = 230; IF group: N = 246; PF group: N = 235). At 12 months, H1N1-TIV recipients had the lowest all-cause mortality, whereas unvaccinated residents had the highest all-cause mortality in all three functional status groups. In the comparison between H1N1-TIV recipients and TIV alone recipients, the hazard ratios (HRs) of all-cause mortality were lower in the GF group and higher in the PF group (GF group: HR 0.30 [0.07-0.95], p < .05; IF group: HR 0.40 [0.18-0.86], p < .05; PF group: HR 0.53 [0.28-0.99], p < .05). The same observation was found in comparison between other vaccination statuses (H1N1-TIV vs unvaccinated and TIV alone vs unvaccinated).
Conclusions: Influenza vaccination was associated with reduced all-cause mortality in older nursing home residents with different functional statuses. Vaccine efficacy in reducing mortality declined with increasingly impaired functional status.