Background Changing patterns of vaccine breakthrough can clarify vaccine effectiveness. Aim To compare breakthrough infections during a SARS-CoV-2 Delta wave vs unvaccinated inpatients, and an earlier Alpha wave. Methods In an observational multicentre cohort study in Israel, hospitalised COVID-19 patients were divided into three cohorts: breakthrough infections in Comirnaty-vaccinated patients (VD; Jun–Aug 2021) and unvaccinated cases during the Delta wave (ND) and breakthrough infections during an earlier Alpha wave (VA; Jan–Apr 2021). Primary outcome was death or ventilation. Results We included 343 VD, 162 ND and 172 VA patients. VD were more likely older (OR: 1.06; 95% CI: 1.05–1.08), men (OR: 1.6; 95% CI: 1.0–2.5) and immunosuppressed (OR: 2.5; 95% CI: 1.1–5.5) vs ND. Median time between second vaccine dose and admission was 179 days (IQR: 166–187) in VD vs 41 days (IQR: 28–57.5) in VA. VD patients were less likely to be men (OR: 0.6; 95% CI: 0.4–0.9), immunosuppressed (OR: 0.3; 95% CI: 0.2–0.5) or have congestive heart failure (OR: 0.6; 95% CI: 0.3–0.9) vs VA. The outcome was similar between all cohorts and affected by age and immunosuppression and not by vaccination, variant or time from vaccination. Conclusions Vaccination was protective during the Delta variant wave, as suggested by older age and greater immunosuppression in vaccinated breakthrough vs unvaccinated inpatients. Nevertheless, compared with an earlier post-vaccination period, breakthrough infections 6 months post-vaccination occurred in healthier patients. Thus, waning immunity increased vulnerability during the Delta wave, which suggests boosters as a countermeasure.