Zoster vaccination was introduced in England in 2013, where tackling health inequalities is a statutory requirement. However, specific population groups with higher zoster burden remain largely unidentified. To evaluate health inequalities in zoster disease burden prior to zoster vaccine introduction in England. This population-based cohort study used anonymized U.K. primary care data linked to hospitalization and deprivation data. Individuals aged ≥ 65 years without prior zoster history (N = 862 470) were followed from 1 September 2003 to 31 August 2013. Poisson regression was used to obtain adjusted rate ratios (ARRs) for the association of sociodemographic factors (ethnicity, immigration status, individuals' area-level deprivation, care home residence, living arrangements) with first zoster episode. Possible mediation by comorbidities and immunosuppressive medications was also assessed. There were 37 014 first zoster episodes, with an incidence of 8·79 [95% confidence interval (CI) 8·70-8·88] per 1000 person-years at risk. In multivariable analyses, factors associated with higher zoster rates included care home residence (10% higher vs. those not in care homes), being a woman (16% higher vs. men), nonimmigrants (~30% higher than immigrants) and white ethnicity (for example, twice the rate compared with those of black ethnicity). Zoster incidence decreased slightly with increasing deprivation (ARR most vs. least deprived 0·96 (95% CI 0·92-0·99) and among those living alone (ARR 0·96, 95% CI 0·94-0·98). Mediating variables made little difference to the ARR of social factors but were themselves associated with increased zoster burden (ARR varied from 1·11 to 3·84). The burden of zoster was higher in specific sociodemographic groups. Further study is needed to ascertain whether these individuals are attending for zoster vaccination. © 2018 The Authors British Journal of Dermatology published by John Wiley & Sons Ltd on behalf of British Association of Dermatologists.