Shedding of Oral Poliovirus Vaccine (OPV) by HIV-Infected and -Uninfected Mothers of OPV-Vaccinated Zimbabwean Infants

Abstract

Community circulation of oral poliovirus vaccine (OPV) likely begins with household transmission. We analyzed stool collected from Zimbabwean mothers who were infected with human immunodeficiency virus (HIV) and those who were uninfected with HIV 1 to 24 weeks after infant oral poliovirus vaccination. Overall, only 5% of the mothers had detectable OPV (16 of 304) despite high infant shedding rates. OPV shedding was similar between HIV-infected mothers and those who were uninfected (11 [6.4%] of 171 vs 5 [3.8%] of 133, respectively) and between mothers of HIV-infected infants and those of uninfected infants (2 [3.5%] of 57 vs 9 [6.3%] of 144, respectively). Mothers of vaccinated infants are unlikely to shed OPV, even when they are infected with HIV.

INTRODUCTION

The oral poliovirus vaccine (OPV), the backbone of polio-eradication efforts, may jeopardize eradication because it can cause vaccine-associated paralytic poliomyelitis and can mutate into vaccine-derived polioviruses (VDPVs), which also can cause poliomyelitis. VDPVs develop through prolonged community transmission (circulating VDPVs) or prolonged intestinal replication in those with humoral immunodeficiencies [1]. In 2012–2013, 10 countries reported VDPV outbreaks, and 7 found VDPVs in sewage [2]. Successful polio-eradication efforts must limit the emergence of VDPVs.

OPV recipients shed vaccine virus in their stool, which confers herd immunity, one of the greatest attributes of the OPV. However, vaccine shedding also enables community transmission. Up to 90% of infants shed OPV after vaccination [3] and can spread OPV to their household contacts. A few published studies have assessed shedding by healthy adult contacts of OPV recipients, but most of these studies were limited to 8 weeks after vaccination, possibly missing transmissions that occurred later [1, 3–5].

These contact studies were also conducted when wild-type poliovirus was endemic, which resulted in robust mucosal immunity in healthy adults [3, 4]. However, evidence suggests that younger adults represent an underimmunized population overlooked by poliovirus surveillance and supplementary campaigns, which primarily target children younger than 5 years. A study of pregnant women in the Democratic Republic of Congo found that seroprevalence to poliovirus serotypes 1 and 3 was lowest in 15- to 29-year-olds in districts with subsequent outbreaks disproportionately affected young adults [6]. Thus, young adults may play a significant role in community poliovirus persistence.

Contact studies have also focused primarily on healthy adult contacts. However, adults with humoral immunodeficiencies who chronically shed OPV are potential sources of circulating VDPVs, especially in settings with low polio vaccine coverage [1]. In addition to cell-mediated immunodeficiency, human immunodeficiency virus (HIV) infection confers humoral and mucosal immune dysfunction [7], which may disrupt an HIV-infected individual's response to OPV after re-exposure. Three studies evaluated OPV shedding in HIV-infected adults [8–10], but they did not evaluate those who are at the greatest risk of poliovirus exposure as a result of contact with contaminated stool—primary caretakers of recently and newly vaccinated infants. We studied the incidence and determinants of OPV shedding by young HIV-infected mothers and HIV-uninfected mothers of OPV-vaccinated Zimbabwean infants up to 5 months after their primary vaccination series.

METHODS

Study Design

This prospective cohort study, conducted from 2008 to 2011, followed families with infants receiving primary OPV at 3, 4, 5, and 18 months of age in Chitungwiza and Harare, Zimbabwe [11]. Each infant's mother's age, HIV status, current medication(s), polio immunization status, and social history were recorded, and stool samples were collected at each study visit. Pertinent samples were collected from each mother when her vaccinated child was 4, 6, and 9 months old.

In this analysis, we included mothers whose infant's polio immunizations were confirmed and who submitted at least 1 stool sample within 1 to 6 weeks of the infant's first dose of OPV or within 1 to 24 weeks of the infant's third dose. Samples submitted within 7 days of OPV receipt were not included, because we assumed that these early samples might not represent spread from the vaccinated infant to the mother. All maternal samples collected during the inclusion window were tested for OPV. However, to investigate the association between maternal and infant OPV shedding, we chose 1 maternal sample within 6 weeks of an OPV dose: either the sample with detectable OPV or the latest collected sample. We included each of the infants' samples collected before or on the same day as their mother's stool.

The Medical Research Council of Zimbabwe, the Research Council of Zimbabwe, and the institutional review board of Stanford University School of Medicine approved this study, and each mother gave informed consent.

Fixed-Sample-Size Power Calculation

Each HIV-uninfected (125) and HIV-infected (153) mother had samples collected after their infant received the third OPV dose. Assuming that 1% of HIV-uninfected mothers shed OPV [1, 3], we had 87% power to detect a 10% difference in the rates of shedding (α = .05).

Stool Assays

Stool samples were shipped on dry ice to Stanford University and stored at –80°C. Each sample underwent a real-time polymerase chain reaction assay to detect OPV serotypes 1, 2, and 3 as previously described [12].

Statistical Analysis

Analysis was conducted by using R 3.0.3 and SAS 9.3. Categorical data were compared by using Fisher's exact test (2-tailed), and continuous data were compared with the t test (2-tailed). A P value of ≤.05 was considered statistically significant.

RESULTS

A total of 544 stool samples from 304 mothers met our inclusion criteria and were included in this analysis (110 samples collected between 1 and 6 weeks and 434 collected 1–24 weeks after the infant's first and third dose).

Overall, 5% (16 of 304) of the mothers had detectable OPV, and no mother had more than 1 sample that tested positive. Figure 1 shows the proportion of maternal OPV-positive stool samples collected after each infant's OPV dose. The peaks in this graph reflect the timing of study visits for infants at 4, 6, and 9 months of age. Shedding within 6 weeks of vaccination was not associated with OPV dose (4 [3.8%] of 104 and 8 [3.7%] of 217 after the first and third OPV doses, respectively). However, less shedding was detected in the samples collected >6 weeks after the third OPV dose (4 [2%] of 204). In contrast, a higher proportion of infant stools that tested positive was detected during the same periods (41.9% [26 of 62] and 22.4% [46 of 205] within 6 weeks of the first and third OPV dose, respectively, and 6.8% [28 of 409] more than 6 weeks after the third dose) [11]. No seasonal variation in OPV shedding was detected (data not shown).

Figure 1.

Frequency of maternal oral poliovirus vaccine (OPV) shedding plotted according to week since each mother's vaccinated child's most recent OPV dose (first or third). Stools that tested positive for OPV are indicated by black. The proportions of shedding after the first and third OPV were similar (4 [3.7%] of 107 vs 12 [2.8%] of 429, respectively). The peaks in the number of stools depicted here represent the timing of study visits according to study protocol.

Of the 16 OPV-shedding mothers, 3 shed serotype 1, 6 shed serotype 2, and 7 shed serotype 3; no mother shed more than 1 serotype. Nine mothers shed the same serotype as their infant.

A total of 279 mothers were linked to at least 1 infant sample collected before or at the same time as their samples, including 11 shedding mothers. Mothers who shed were more likely to have an infant who had shed OPV previously, although the differences were not significant (3 of 3 shedding vs 57 of 99 nonshedding mothers' infants shed after their first OPV dose [P = .27], and 7 of 8 shedding vs 157 of 212 nonshedding mothers' infants shed after their third OPV dose [P = .68]).

Demographic differences between OPV-shedding mothers and nonshedding mothers were not statistically significant (Supplementary Table 1). Most mothers in both groups were the OPV-vaccinated infant's primary caregiver (300 [99%] of 304), and most of them reported a remote history of OPV receipt (260 [86%] of 304). The remainder of the mothers did not recall their vaccination history (24 [8%] of 304) or had missing data (20 [6%] of 304).

OPV shedding rates were similar between HIV-infected mothers and uninfected mothers (11 [6.4%] of 171 HIV-infected mothers vs 5 [3.8%] of 133 HIV-uninfected mothers [P = .44]). Among HIV-infected mothers, there was no difference in shedding on the basis of the infants' HIV status (9 [8%] of 114 HIV-uninfected children vs 2 [3.5%] of 57 HIV-infected children [P = .34]). At study enrollment, only 25% [42 of 171] of the HIV-infected mothers reported taking antiretroviral therapy (ART), including 4 of 11 HIV-infected mothers who shed OPV. However, ART was not associated with OPV shedding (4 [9.5%] of 42 mothers on ART vs 6 [4.8%] of 126 mothers not on ART shed OPV [P = .27]). No medication history was available for 3 HIV-infected mothers, one of whom shed OPV.

DISCUSSION

Maternal OPV shedding was uncommon in this prospective cohort despite high rates of OPV shedding by the vaccinated infants. These mothers likely maintained protective mucosal immunity from previous OPV receipt or exposure. OPV vaccination in Zimbabwe began in the late 1960s, and the last clinical case of poliomyelitis (without virologic confirmation) was reported in 1999 [13]. Mothers enrolled in our study were born between 1967 and 1992. Thus, although 10% of the enrolled mothers did not recall their polio vaccination history, all of them likely had mucosal immunity as a result of previous exposure to wild-type poliovirus or OPV.

We found no correlates of maternal OPV shedding; analysis was limited by the small number of mothers who shed OPV. Given our fixed sample size, we were powered to detect ≥10% shedding difference between the mothers on the basis of their HIV status. Although not significant, we found a slightly higher proportion of shedding in HIV-infected mothers than in HIV-uninfected mothers, which contrasts with results of other studies that found no OPV shedding by HIV-infected adults [8–10]. However, only 1 of these studies enrolled adults who were the primary caretaker of a recently OPV-vaccinated child, but it was limited by a small sample size (n = 28) and included children who were up to 5 years old and likely received multiple doses of OPV [10]. In contrast, our study followed mothers whose infant was receiving the primary vaccination series.

Most maternal shedding occurred within 20 to 37 days of vaccination, which is consistent with previous reports of household contact shedding, primarily in siblings [1, 3]. Four mothers in our study shed OPV between 89 and 142 days after vaccination. However, routine administration of OPV to infants occurred year-round, and OPV was also administered during supplementary campaigns. Thus, it is possible that some of these mothers acquired OPV from their community instead of from their vaccinated child.

Our study had some limitations. First, we likely underestimated the true incidence of maternal OPV shedding. Stool samples were collected at 4- to 12-week intervals and may have missed shedding between collections. We excluded samples collected within 7 days after OPV administration, which may have resulted in us missing early shedding. However, the main aim of this study was to determine persistent poliovirus circulation assessed by long-term stool collection. Second, although the HIV status of these mothers was available, their degree of immunosuppression was unknown. However, no differences in shedding according to HIV status were detected.

In conclusion, we found a low rate of maternal OPV shedding after routine infant vaccination despite high rates of infant shedding. Thus, mothers and other OPV-immunized primary caretakers are unlikely to shed OPV, even when they are infected with HIV. This result is important given the global HIV burden (especially among countries that use OPV), and it provides reassurance that HIV is not a risk factor for persistent OPV shedding and transmission that could jeopardize the polio end game. However, it remains likely that community circulation of OPV begins with household transmission. As the world transitions from OPV to routine inactivated poliovirus vaccine (IPV) use, which confers less mucosal immunity, understanding OPV circulation in undervaccinated and primarily IPV-vaccinated communities is critically important.