Safety profile of herpes zoster vaccines in the United States: a descriptive and disproportionality analysis of adverse events reported to the vaccine adverse event reporting system (VAERS), 2017–2024

Background Herpes zoster causes substantial morbidity in older adults, and US herpes zoster vaccination shifted from the live-attenuated Zostavax to the recombinant Shingrix after 2017 We describe the post-marketing adverse-event profile of these vaccines reported to the US Vaccine Adverse Event Reporting System (VAERS) and identify events reported disproportionately for herpes zoster vaccines. Methods In a descriptive cross-sectional design, we analyzed 69,822 US VAERS reports listing a herpes zoster vaccine, received 1 January 2017–25 April 2024. Reports were characterized by demographics, brand, MedDRA preferred terms and system organ classes, and US FDA serious-event indicators; brand subgroups were summarized descriptively. Reporting odds ratios (RORs) with 95% confidence intervals compared the combined herpes zoster vaccine cohort with all other US VAERS reports. Reporting followed the STROBE guideline. Results Shingrix accounted for 58,813 (84.2%) reports and Zostavax for 10,507 (15.0%). Reports were predominantly in women (63.9%) and adults aged 50–69 years (median age 64.0 years). Reactogenic and injection-site terms predominated—pyrexia (18.3%), pain (17.8%), chills (15.9%), headache (15.4%), and injection-site pain (14.2%). Against all other US reports, injection-site and influenza-like reactions were modestly enriched (e.g., injection-site pain ROR 2.92), whereas vesicular rash was strongly over-reported (ROR 43.70); the latter was concentrated in live-vaccine reports and reflects reported vaccine-strain disease rather than an unexpected adverse reaction. Serious reports comprised 6.6% of the cohort (2.9% Shingrix, 27.6% Zostavax). Conclusion Eight years of VAERS data are consistent with the recognized reactogenicity of the recombinant vaccine and the live-vaccine profile of Zostavax; most reported events were transient and non-serious. These hypothesis-generating disproportionality signals require confirmation in active surveillance with denominator data.