Evidence for off-target effects of live Shingles vaccination against all-cause death and infection-associated hospitalisation in older adults in England: a population-based cohort study

BackgroundNon-specific live vaccine effects have been described in paediatric populations. We investigated whether non-specific effects are observed in older adults with live shingles (Zostavax(R)) immunisation. MethodsA population-based cohort study was performed using primary- and secondary-care data from the Clinical Practice Research Datalink in England. Shingles vaccine eligible individuals aged 70 years and over who had received pneumococcal immunisation were included. All-cause death, all-cause hospitalisation, and infection-associated hospitalisation rates were compared between live shingles vaccinated and unvaccinated person-time using time-to-first-event Cox regression and recurrent events modelling. Live shingles vaccine exposure was included as a time-varying exposure. The models were run on an overlap-weighted pseudo-population to minimise confounding. FindingsBetween September 2013 and June 2019 314,618 participants aged 70 years and over who had received pneumococcal immunisation prior to study entry were identified, of which 55% had live shingles vaccine exposure. The overlap weighted pseudo-population consisted of 60021 lives shingles vaccine unexposed and 60245{middle dot}6 exposed participants. Live shingles vaccine exposure in the overlap pseudo-population was associated with a reduction in all-cause death (hazard ratio 0{middle dot}64; 95% confidence interval 0{middle dot}62, 0{middle dot}66), a reduction in recurrent all-cause hospitalisation (hazard ratio 0{middle dot}83 (0{middle dot}79, 0{middle dot}87)), and a reduction in first and recurrent infection-associated hospitalisation (hazard ratio for first event 0{middle dot}81 (0{middle dot}79, 0{middle dot}83), and for recurrent events 0.75 (0.73, 0.77)). Protective vaccine effects were observed for at least five years post-immunisation. InterpretationsReceipt of live shingles vaccine associates with lower mortality and morbidity in older adults in England. The potential for causal linkage should be validated in robust prospective studies, with major implications for national immunisation policies. FundingNo specific funding for this project. KD was supported by an NIHR academic clinical fellowship. Research in contextO_ST_ABSEvidence before this studyC_ST_ABSA body of research exists demonstrating an association between live vaccines and a reduction in mortality and illness unrelated to the vaccine target. This has been termed non-specific vaccine effects for which heterologous immune programming has been proposed as a possible mechanism. To date, non-specific vaccine effects have largely been demonstrated in paediatric populations in relation to live measles, oral polio, and Bacillus Calmette-Guerin (BCG) immunisation. Added value of this studyThis is the first study to evaluate mortality and hospitalisation effects in an older adult population with live shingles immunisation. Older adults are a growing population in many settings worldwide and require cost-effective interventions to reduce morbidity and mortality. This study finds a reduction in all-cause death, all-cause hospitalisation, and infection-associated hospitalisation in over 70-year-olds in England associated with live shingles vaccine exposure. These findings were not explained by any measurable confounding effect of age, socio-economic status, comorbidity, nor health-seeking behaviour. Implications of all the available evidenceThis work and other emerging reports advocate for studies with an experimental design to provide definitive evidence of non-specific vaccine effects to guide policy makers and to evaluate the possible immune mechanisms in an older immunosenescent population. There is potential for very significant benefits relating to both use of shingles vaccine and type of vaccine that needs clarity.