Age-structured modelling of Ebola convalescence and re-emergence risk: a Bayesian approach

Background. Increasing evidence indicates that Ebola virus disease (EVD) survivors can remain a source of infection long after clinical recovery. Confirmed survivor-associated transmission events and genomic evidence linking the 2021 Guinea outbreak to viral lineages from the 2013-2016 West African epidemic have demonstrated that persistent infection in survivors can contribute to post-epidemic re-emergence. However, the population-level conditions under which survivor reservoirs may sustain recrudescence remain poorly understood. Methods. We developed an age-structured Bayesian transmission model to quantify survivor-driven recrudescence risk using historical Ebola outbreak data (1976-2022) and empirical viral persistence data from male survivors. Age-specific viral clearance probabilities were estimated for three age groups (less or equal to 25, 26-35, and >35 years). The recrudescence reproduction number (Rc) was derived using the next-generation matrix approach. Sensitivity analyses examined alternative assumptions regarding viral clearance and the potential contribution of female survivors. Results. The posterior mean recrudescence reproduction number remained below the persistence threshold (Rc=1) across all viral-clearance scenarios under the assumption of no female survivor contribution. Only by assuming the slowest rate of viral clearance and maximal female survivor contribution did the posterior mean for Rc exceed one (1.052; 95% CrI: 0.428-2.229), suggesting that survivor-driven transmission alone is unlikely to sustain Ebola re-emergence given our current understanding of recrudescence dynamics. Simulations showed that survivor-driven outbreak pressure (rate of survivor-initiated outbreaks) was driven primarily by outbreak size and clustering. Outbreaks involving less or equal to 5,000 EVD cases generally produced outbreak pressure below the estimated natural spillover rate, whereas outbreaks comparable in size to the 2013-2016 West African epidemic generated transient survivor-driven outbreak rates up to 7.8-fold higher than the natural spillover rate before declining to comparable levels within 3-7 years. Moreover, across all viral-clearance scenarios, older (>35 years) male survivors consistently exhibited the longest effective persistence durations and made the largest contribution to the recrudescence reproduction number. Conclusions. The human survivor reservoir represents a plausible complementary pathway for Ebola re-emergence, particularly following large epidemics and should be considered alongside zoonotic spillover as an important source of future outbreaks. Age-dependent viral clearance strongly shapes recrudescence dynamics, with older survivors contributing disproportionately to transmission potential. These findings support age-stratified survivor monitoring, extended persistence surveillance, and improved characterization of viral persistence in both male and female survivors to strengthen post-epidemic preparedness.