(263) Assessing The Effect of Statin Use on Influenza Vaccine Effectiveness: A Prior Event Rate Ratio Analysis in Primary Care Data

Publication Number: 1308

Background: Influenza (flu) vaccination is routinely advised for many at-risk groups including adults aged≥65 years. Many of these patients are also likely to be prescribed statins for control of cardiovascular risk. Previous studies have suggested that statin use may reduce the effectiveness of flu vaccination. However, these studies are susceptible to bias from unmeasured confounding.

Objectives: The Statin Treatment OPtions during vaccination against inFLUenza (STOPFlu) study is the first investigation of the potential modulating effect of taking statins at the time of receiving the flu vaccine in the UK healthcare system. A new-user quasi-experimental design was used to address sources of measured and unmeasured confounding.

Methods: We recruited cohorts of adults aged≥65y vaccinated against influenza from general practices registered to the Clinical Practice Research Datalink for each year from 2010 to 2018. New statin users and nonusers were identified in the six-months preceding vaccination. The composite influenza outcome was defined as hospital admission for influenza or influenza-like illness diagnosed in primary care. We estimated associations between statin-use at the time of vaccination and risk of influenza using Cox regression. We adjusted for observed confounding bias by inverse probability treatment weighting (IPTW) on the propensity score for receipt of statins. We also used influenza events in statin users and non-users in the previous influenza season to help assess the degree of confounding bias from unobserved sources.

Results: Preliminary results from pilot work using IPTW to account for observed confounders did not suggest an increased risk of influenza among new statin-users following vaccination (HR 0.83, 95% confidence interval: 0.67, 1.02) compared to statin-free patients. Further analysis is being conducted using the prior event rate ratio method to account for unmeasured confounding bias, with careful consideration to testing the plausibility of the assumptions.

Conclusions: Adjustment for observed covariates of flu risk, alone, in health registry data is insufficient to account for all confounding bias. New results from the STOPFlu study on the moderating effect of statins on flu vaccine effectiveness that account for unmeasured confounding will be presented at the ICPE2022 conference. Results will also be available on the possible prophylactic effect of statins themselves against influenza.