Influenza vaccination and risk of stroke: self-controlled case-series study

Background Previous studies suggest that stroke may be triggered by respiratory infections including influenza. Influenza vaccination could therefore reduce risk of stroke. Studies showing an association between influenza vaccination and reduction in stroke incidence may have been affected by residual confounding. We aimed to investigate the association between influenza vaccination and risk of stroke using the self-controlled case series method which eliminates the effect of fixed confounders. Methods We used the self-controlled case series design. The Clinical Practice Research Datalink (CPRD) was used to extract records of patients aged 18 years and over recorded with a first stroke (fatal or non-fatal) occurring between September 2001 and August 2009. Statistical modelling with conditional Poisson regression was used to compute incidence rate ratios (IRR). The incidence rates of stroke during fixed time periods following vaccination were compared with the incidence rate during a baseline period. Results We included 21256 cases with stroke in the analysis. The incidence of stroke was significantly reduced in the first 180 days following influenza vaccination compared with the baseline period. We found a reduction of 41% (IRR 0.59; 95% CI 0.53-0.65) in the first 14 days after vaccination, and 20% (0.80; 0.72-0.88) 23% (0.77, 0.72-0.83), 15% (0.85, 0.79-0.91), 13% (0.87, 0.81-0.94) and 9% (0.91, 0.86-0.95) at 15-28 days, 29-59 days, 60-90 days, 91-120 days and 121-180 days after vaccination respectively. There were no differences in the association by age group (above or below 65 year) or sex. Conclusions Influenza vaccination is associated with a reduction in incidence of stroke. This study suggests a beneficial association between influenza vaccination and stroke which should be tested in a future randomised controlled study