Estimating age-specific vaccine effectiveness using data from a large measles outbreak in Berlin, Germany, 2014/15: evidence for waning immunity

BACKGROUND: Measles elimination is based on 95% coverage with two doses of a measles-containing vaccine (MCV2), high vaccine effectiveness (VE) and life-long vaccine-induced immunity. Longitudinal analysis of antibody titres suggests existence of waning immunity, but the relevance at the population-level is unknown. AIM: We sought to assess presence of waning immunity by estimating MCV2 VE in different age groups (2–5, 6–15, 16–23, 24–30 and 31–42 years) in Berlin. METHODS: We conducted a systematic literature review on vaccination coverage and applied the screening-method using data from a large measles outbreak (2014/15) in Berlin. Uncertainty in input variables was incorporated by Monte Carlo simulation. In a scenario analysis, we estimated the proportion vaccinated with MCV2 in those 31-42 years using VE of the youngest age group, where natural immunity was deemed negligible. RESULTS: Of 773 measles cases (median age: 20 years), 40 had received MCV2. Average vaccine coverage per age group varied (32%–88%). Estimated median VE was > 99% (95% credible interval (CrI): 98.6–100) in the three youngest age groups, but lower (90.9%, 95% CrI: 74.1–97.6) in the oldest age group. In the scenario analysis, the estimated proportion vaccinated was 98.8% (95% CrI: 96.5–99.8). CONCLUSION: VE for MCV2 was generally high, but lower in those aged 31-42 years old. The estimated proportion with MCV2 should have led to sufficient herd immunity in those aged 31-42 years old. Thus, lower VE cannot be fully explained by natural immunity, suggesting presence of waning immunity.