Modeling undetected live poliovirus circulation after apparent interruption of transmission: implications for surveillance and vaccination

Abstract

BackgroundMost poliovirus infections occur with no symptoms and this leads to the possibility of silent circulation, which complicates the confirmation of global goals to permanently end poliovirus transmission. Previous simple models based on hypothetical populations assumed perfect detection of symptomatic cases and suggested the need to observe no paralytic cases from wild polioviruses (WPVs) for approximately 3-4 years to achieve 95% confidence about eradication, but the complexities in real populations and the imperfect nature of surveillance require consideration.MethodsWe revisit the probability of undetected poliovirus circulation using a more comprehensive model that reflects the conditions in a number of places with different characteristics related to WPV transmission, and we model the actual environmental WPV detection that occurred in Israel in 2013. We consider the analogous potential for undetected transmission of circulating vaccine-derived polioviruses. The model explicitly accounts for the impact of different vaccination activities before and after the last detected case of paralytic polio, different levels of surveillance, variability in transmissibility and neurovirulence among serotypes, and the possibility of asymptomatic participation in transmission by previously-vaccinated or infected individuals.ResultsWe find that prolonged circulation in the absence of cases and thus undetectable by case-based surveillance may occur if vaccination keeps population immunity close to but not over the threshold required for the interruption of transmission, as may occur in northwestern Nigeria for serotype 2 circulating vaccine-derived poliovirus in the event of insufficient tOPV use. Participation of IPV-vaccinated individuals in asymptomatic fecal-oral transmission may also contribute to extended transmission undetectable by case-based surveillance, as occurred in Israel. We also find that gaps or quality issues in surveillance could significantly reduce confidence about actual disruption. Maintaining high population immunity and high-quality surveillance for several years after the last detected polio cases will remain critical elements of the polio end game.ConclusionsCountries will need to maintain vigilance in their surveillance for polioviruses and recognize that their risks of undetected circulation may differ as a function of their efforts to manage population immunity and to identify cases or circulating live polioviruses.Electronic supplementary materialThe online version of this article (doi:10.1186/s12879-015-0791-5) contains supplementary material, which is available to authorized users.

Environmental Surveillance System Characteristics and Impacts on Confidence About No Undetected Serotype 1 Wild Poliovirus Circulation

“…Prior mathematical modeling studies explored the probability of undetected circulation of WPVs in the absence of reported cases or poliovirus detections by an ES system. Our 2015 study (Kalkowska et al., ) used a stochastic adaptation of a poliovirus dynamic transmission model (Duintjer Tebbens et al., ) and included an analysis of the 2013 serotype 1 WPV introduction into Israel followed by more than a year of transmission detected by the extensive ES system in Israel (Kalkowska et al., ). The study considered the impact of acute flaccid paralysis (AFP) surveillance and ES on the confidence about no undetected poliovirus circulation as a function of time after the last detection through either system (Kalkowska et al., ).…”

“…Our 2015 study (Kalkowska et al., ) used a stochastic adaptation of a poliovirus dynamic transmission model (Duintjer Tebbens et al., ) and included an analysis of the 2013 serotype 1 WPV introduction into Israel followed by more than a year of transmission detected by the extensive ES system in Israel (Kalkowska et al., ). The study considered the impact of acute flaccid paralysis (AFP) surveillance and ES on the confidence about no undetected poliovirus circulation as a function of time after the last detection through either system (Kalkowska et al., ). The results suggested that it takes up to three years without paralytic cases to reach 95% confidence about no circulation in the context of perfect (i.e., detecting every case) AFP surveillance (without ES).…”

“…The results suggested that it takes up to three years without paralytic cases to reach 95% confidence about no circulation in the context of perfect (i.e., detecting every case) AFP surveillance (without ES). Less than perfect AFP surveillance implies longer times, while the addition of high‐quality ES reduces the times to reach 95% confidence (Kalkowska et al., ). The GPEI generally considers three years of high‐quality AFP surveillance with no reported polio AFP cases as providing sufficient confidence to certify the eradication of a WPV serotype.…”

“…The development of ES raises questions about whether the criteria used to certify the world as free of WPV1 and/or WPV3 should include the time since the last detected event (i.e., either a polio AFP case or a positive isolate from an ES sample). We previously demonstrated that poliovirus transmission can continue to occur in the absence of cases, and that die‐out (i.e., clearance of the virus from the last excreting individual) will occur well after the last reported AFP case (Kalkowska et al., ). Evaluating the potential for ES to effectively increase confidence about no undetected circulation depends on understanding the extent to which the AFP and ES surveillance systems will detect transmission.…”

“…The circulation of LPVs remains complicated in Pakistan and Afghanistan as both countries continue to pursue WPV1 eradication and monitor the disappearance of serotype 2 LPVs. Using the same framework as a prior study (Kalkowska et al., ), the subsequent analysis (Kalkowska et al., ) used a more detailed approach to model the ES system, and explored the likely future confidence about the end of transmission in the absence of any epidemiological signals (i.e., after the last reported AFP case and/or detection of poliovirus in ES based on analysis of limited data available). We emphasize that our prospective modeling (Kalkowska et al., ) differs from and complements epidemiological modeling efforts that seek to characterize retrospective events using the currently available incomplete and/or inconsistent data (Mercer et al., ; Molodecky et al., ; O'Reilly et al., ).…”