Poliovirus immunity among adults in the Democratic Republic of the Congo: a cross-sectional serosurvey

While no cases of wild poliovirus (WPV) have been detected in the Democratic Republic of the Congo (DRC) since 2011 [1], endemic transmission continues in Afghanistan, and Pakistan [2]. The elimination of WPV in the DRC was a gradual process: the last reported cases occurred in 1997, 2009 and 2011 for WPV serotypes 2, 3 and 1, respectively [3]. Furthermore, WPV type 2 has been declared globally eradicated, with the last case of paralytic polio from WPV2 observed in Uttar Pradesh, India in 1999 [4]. However, circulating vaccine-derived poliovirus (cVDPV) outbreaks remain a concern and, at the time of writing, there were 9 independent cVDPV outbreaks occurring in the DRC [5]. Until polio is eradicated and oral polio vaccine (OPV) withdrawn from use, the risk of outbreaks from WPV importation and emergence of cVDPV cases may persist due to sub-optimal vaccination coverage [6].

Prior to the initiation of routine and supplemental immunization programs, high rates of endemic infection among children resulted in infrequent paralytic poliomyelitis cases among adults in developing countries [7]. However, the introduction of routine immunization reduced poliovirus circulation and, as unvaccinated or inadequately vaccinated subgroups of children, outbreak-associated cases may occur among older children, adolescents and adults [7]. In the DRC, past outbreaks have affected adult populations and underscore the role and importance of adults in transmission, outbreaks, and efforts to eradicate the disease. In 2010 and 2011, respectively, there were WPV1 outbreaks of 100 cases (in 5 provinces) and 93 cases (in 6 provinces) in the DRC [3]. Among all cases with known ages, 5% in 2010 and 29% in 2011 were identified among individuals aged ≥ 15 years (age range 16–31 years). Prior to these outbreaks, nearly all documented WPV cases in the DRC had occurred in children < 15 years of age [8]. At the same time, the Republic of Congo experienced an outbreak of 445 polio cases, with nearly 75% occurring among adults [6]. During that outbreak, the median age of laboratory-confirmed and clinical polio cases was 20 years (range, < 1–63 years) and individuals 15 to 29 years of age had the highest incidence of all age groups. In response to the outbreaks, national and sub-regional supplemental immunization activities (SIAs) targeted the entire population, adapting strategies to reach adults. Furthermore, cVDPV2 cases have also been documented in adults: while the vast majority of the 315 cVDPV2 cases identified in Nigeria between 2005 and 2010 occurred in children between 18 and 36 months of age, cases were identified in adults up to 22 years of age [9].

Serological protection among adults is also related to transmissibility of wild and vaccine viruses [10, 11]. While the DRC Expanded Program on Immunization (EPI) was introduced in 1978, many adults in the current study were born when routine immunization coverage was low, and wild polioviruses were endemic. Historically, SIAs typically targeted children less than 5 years of age; however, during the 2010–2011 outbreak, all-age campaigns were conducted in certain provinces throughout the DRC, making some adults eligible for supplemental vaccination. This variation makes precise estimation of adult participation in transmission, or generalization of any such estimate, difficult. Since acute flaccid paralysis (AFP) monitoring is the cornerstone of polio surveillance, this estimation is further complicated by the fact that most infections present sub-clinically. Approximately 1 in 150 primary infections cause paralytic poliomyelitis; thus, reports of paralytic cases provide a vast underestimate of total viral spread [4].

Among adults who were not vaccinated, serological protection must be due to exposure to WPV, cVDPV, or through exposure to the Sabin live, attenuated oral poliovirus shed by a vaccine recipient. Individuals with serum immunity against polio are protected from developing paralysis; however, as intestinal immunity to poliovirus wanes over time, exposure to the virus may result in infection of the gut, resulting in shedding and exposure to others [12, 13]. This mechanism may be an underlying reason for the finding of a case–control study in Angola that identified travel of adult household members outside the province of residence as a risk factor for poliomyelitis in children [14]. In addition, oral polio vaccine (OPV) strains can protect contacts of vaccine recipients through environmental secondary immunization, increasing the reach and impact of the vaccine [15]. Since transmission and reversion of OPV can also facilitate emergence of cVDPV in low-immunity settings, it is important to better understand secondary spread of OPV strains among adult household and community contacts.

Given the important role adults may play in the transmission of poliovirus and the global effort to eradicate polio, understanding adult immunity against polioviruses could be useful in identifying high risk populations in the DRC [16]. In this paper, we describe results from a serosurvey conducted in collaboration with the 2013–2014 DRC Demographic and Health Survey (DHS) on serological protection from poliovirus types 1, 2, and 3 among adults (persons 15 years of age and older). Identifying potential high-risk areas with large numbers of susceptible adults is essential for planning, management and eradication efforts due to the risk of continued cVDPV emergence or importation, as well as importation of WPV.

Among adults participating in the 2013–2014 DRC-DHS, 74%, 72%, and 57% were seropositive for neutralizing antibodies for type 1, type 2 and type 3, respectively (Table 1). While little difference was observed by sex for type 1, women had higher seroprevalence rates compared with men for types 2 and 3. Immunity was positively correlated between serotypes: overall, 40% of adults were seropositive for all three types of polio (46% for women and 33% for men), while only 9% were negative across all serotypes.

The relationship between seropositivity and sociodemographic characteristics varied by serotype (Table 2). We generally saw higher rates of seroprevalence among adults in households with more children. Calculated from the number of children in the household, we observed 73%, 68%, and 52% seroprevalence for type 1, 2 and 3, respectively, for those without children and 74%, 74%, and 58% seroprevalence for type 1, 2 and 3, respectively, for those with five or more children. As both children and adults took part in the DHS serosurvey, we conducted exploratory analysis of seropositivity among children [19] and adults within households. A significant intrahousehold relationship in neutralizing antibodies was identified for poliovirus type 2 only (see Additional file 1: Table S2). Seroprevalence was generally higher in older populations, while we did not see much variation in seroprevalence by wealth index, education, or residence (rural vs urban).

While little difference was observed by the type of residence (urban or rural), seroprevalence varied widely across provinces (Fig. 1). Bas-Uele had the lowest seroprevalence in adults (28% for type 3), while the highest estimates of seropositivity were observed in Kinshasa for type 1 (96%), Equateur for type 2 (87%) and Haut-Lomami for type 3 (67%).

As seroprevalence was higher among women overall, and this trend generally persisted within age groups as well (Fig. 2). Among adults older than 20 years of age, a higher proportion of women trended seropositive than men across all serotypes. The largest differences in seroprevalence estimates across sex were observed in the 30 to 34 year-olds, while those differences appeared smaller at the extremes of age. The largest gap in seropositivity between men and women occurred in those aged 25–29 for poliovirus type 1 (13%), in those 30–34 years old for type 2 (26%), and in those 30–34 years old for type 3 (25%) (Additional file 2, Additional file 3).

This study provides the first nationally representative estimates of neutralizing antibodies against poliovirus type 1, 2, and 3 in adults of the DRC. Among adults represented in the survey (men ages 15 to 59 years and women ages 15 to 49 years), an estimated 26%, 28% and 43% remain susceptible to poliovirus type 1, 2, and 3, respectively. As only 9% were seronegative for all three, the vast majority of adults exhibit some level of protection against one or more serotypes (approximately 40% of adults were positive for all three). Seropositivity was generally higher among women, those in households with more children, and among older age groups, and also varied by province.

Overall, we observed relatively low seroprevalence among adults across the country. In some areas, the low seroprevalence, particularly to type 3 poliovirus (57% overall), may indicate a limited viral circulation and/or possibly weak poliovirus transmissibility in the population, which permits individuals to escape infection through adulthood. In addition, this result may be influenced by the lower overall immunogenicity of poliovirus type 3 administered by trivalent OPV vaccine as compared to that from the type 1 and 2 components [22]. The buildup and geographic spread of susceptible populations may, in part, explain previous outbreaks that affected adult populations, a pattern more commonly observed in developed countries with high levels of sanitation [7]. In the context of the DRC, low seroprevalence could be instead explained by relatively low population movement and restricted geographic spread due to the difficulty of transportation between communities and historically rural character, rather than good sanitation and limited ability to spread to contacts of infected individuals. This is observed in some of the recent cVDPV2 outbreaks which are often regionally localized rather than country-wide. Future studies that include modeling of potential household secondary transmission of poliovirus from vaccinated children to adults could refine estimates of natural immunity and better clarify the extent of historical virus circulation.

In the present study, type 1 had the highest seroprevalence across all age and sex groups followed by types 2 and 3. Although the assay does not distinguish between neutralizing antibodies induced by exposure to WPV, OPV, or VDPV, this finding may reflect the atypical outbreak of wild poliovirus type 1 that occurred from 2010 to 2011 [3]. This outbreak affected a larger proportion of persons ≥ 15 years of age than expected in 4 of the 6 outbreak provinces, and triggered OPV campaigns covering all ages. We found the high seroprevalence for type 1 in Kinshasa (95%) and Kwilu province (80%) which had the highest proportion of cases that occurred among individuals 15 years of age and older during the outbreak (24% and 52%, respectively). Additionally, male sex was associated with case status among the adults from the 2010–2011 WPV1 outbreak, corresponding with our findings that there were larger immunity gaps among men. As the OPV virus is shed through fecal matter, the difference in seropositivity by sex may be a result of females taking on primary caregiver roles for infants and young children receiving OPV.

Similar to the serosurvey findings of 2013–2014 DRC DHS child participants aged 6–59 months [19], we also identified the high prevalence of seropositive individuals for type 1 and 2 in Kinshasa and type 3 in both Nord Kivu and Kinshasa. However, in contrast to the children’s results, adult DHS respondents showed a positive relationship between poliovirus immunity and the number of children residing in their households. This finding may again be suggestive of immunity resulting from infant care and fecal matter contact [10]. Behavioral transmission dynamics such as this are of particular importance for polio prevention and control through improved hygiene practices [23]. The correlation between serotypes, household size, and sex seen here all suggest that indirect vaccination by exposure to OPV shed from children may have contributed to immunity among adults in this survey.

While this study provides the first nationally representative estimates of poliovirus immunity among adults in the DRC, there are several limitations. As in the 2013–2014 child polio serosurvey, the adult serosurvey was carried out in conjunction with the household-based DHS, thus mobile or remote, difficult-to-reach populations may not be well represented in the sampling frame; indeed, there were no selected clusters near the previous cVDPV2 outbreak areas [19]. Despite efforts to split the adult specimens for even distribution across province, sex and age, and adjust survey weights for differential missingness, if other factors affecting serostatus were not evenly distributed across groups this may have affected the representativeness of the results. Another limitation is the lack of information on history of vaccination among adults, which was not collected by the DHS, other than tetanus for women who gave birth to one or more children. Given past SIAs that have included older children and adults, it would be of interest to assess seroprevalence in relation to this information. Finally, as noted above, we are unable to distinguish between immune responses induced by natural infection versus exposure to vaccine-related poliovirus in the laboratory analysis, a limitation of the neutralization assay [24].

Our finding of high susceptibility to poliovirus among adults in the DRC is of concern, given the potential for importation of WPV1 or cVDPV. These results may have been influenced by several factors. While we found evidence for increased seropositivity with household size, and number of children in the household, secondary spread of vaccine viruses was not sufficient to achieve high immunity for all serotypes. Second, low seropositivity suggests more limited historical circulation of virus, potentially due to population distribution and rurality. Lastly, while it is generally believed that polio immunity is lifelong once acquired, waning immunity below the limit of detection in adulthood may be possible. Given the global effort to eradicate polio, this study provides a first step in understanding adult immunity against polioviruses in a setting with continued clusters of outbreaks; this may influence future vaccination policies countrywide.