Assessing the Risk of Vaccine-derived Outbreaks After Reintroduction of Oral Poliovirus Vaccine in Postcessation Settings

Abstract

Background: The Polio Eradication and Endgame Strategic Plan 2013-2018 calls for the gradual withdrawal of oral poliovirus vaccine (OPV) from routine immunization. We aimed to quantify the transmission potential of Sabin strains from OPV when it is reintroduced, accidentally or deliberately, in a community vaccinated with inactivated poliovirus vaccine alone.

Methods: We built an individual-based stochastic epidemiological model that allows independent spread of 3 Sabin serotypes and differential transmission rates within versus between households. Model parameters were estimated by fitting to data from a prospective cohort in Mexico. We calculated the effective reproductive number for the Mexico cohort and simulated scenarios of Sabin strain resurgence under postcessation conditions, projecting the risk of prolonged circulation, which could lead to circulating vaccine-derived poliovirus (cVDPV).

Results: The estimated effective reproductive number for naturally infected individuals was about 1 for Sabin 2 and Sabin 3 (OPV2 and OPV3) in a postcessation setting. Most transmission events occurred between households. We estimated the probability of circulation for >9 months to be (1) <<1% for all 3 serotypes when 90% of children <5 years of age were vaccinated in a hypothetical outbreak control campaign; (2) 45% and 24% for Sabin 2 and Sabin 3, respectively, when vaccine coverage dropped to 10%; (3) 37% and 8% for Sabin 2 and Sabin 3, respectively, when a single active shedder appeared in a community.

Conclusions: Critical factors determining the risk of cVDPV emergence are the scale at which OPV is reintroduced and the between-household transmission rate for poliovirus, with intermediate values posing the greatest risk.

Figure 1.

Vaccinations and first shedding events by household. A shedding event was defined as the presence of ≥1 serotype in a stool sample, and only the first shedding event for each individual is displayed. Each row corresponds to a household, and the size of points is scaled to case numbers. Pie chart shows the cumulative number of shedding events in vaccinees, household contacts of vaccinees, and unvaccinated households. (Households that were enrolled in the Mexico study but provided no positive stool samples throughout the study period are not shown.)

Figure 2.

Shedding duration in Mexican study cohort by Sabin serotype, in vaccine recipients and naturally infected individuals.

Figure 3.

Effective reproductive number for nonvaccinees. Overall and within-household effective reproductive numbers (R_e) values estimated at the onset of the Mexico study (A) and 5 years after cessation of oral poliovirus vaccine (B).

Figure 4.

Simulated outbreak response using oral poliovirus vaccine (OPV) in Capoluca 5 years after cessation. A, Predicted cumulative incidence and duration of continued transmission by Sabin serotype, according to vaccination rate among children <5 years of age. B, Sensitivity analysis on transmission rates of Sabin 2 with vaccine coverage fixed at 10%; the same multiplier is applied to within-household and between-household transmission rates. C, Sensitivity analysis on transmission rates of Sabin 2 with vaccine coverage fixed at 90%. D, Sensitivity analysis on shedding duration in children <5 years of age (born after cessation and vaccinated with inactivated poliovirus vaccine alone in routine immunization) and fraction susceptible to infection in persons >5 years of age (born before cessation and exposed to OPV viruses before). Grid color indicates median of the simulation outputs, and annotations represent nonzero probabilities (percentages) of OPV circulating for >9 months in a community. Abbreviation: CI, credible interval.

Figure 5.

Simulated silent circulation triggered by 1 infectious case shedding Sabin virus in Capoluca 5 years after cessation. A, Predicted magnitude and duration of continued transmission by serotype. Scatterplots show results of each simulation; violin plots, marginal distributions of duration and total cases. B, Sensitivity analysis on within- and between-household transmission rates of Sabin 2. Grid color indicates probability of Sabin virus circulating for >9 months in a community. C, Sensitivity analysis on shedding duration in children <5 years of age and fraction susceptible to infection in persons >5 years of age. Abbreviation: cVDPV, circulating vaccine-derived poliovirus.