Oral polio vaccine response in the MAL-ED birth cohort study: Considerations for polio eradication strategies

Background: Immunization programs have leveraged decades of research to maximize oral polio vaccine (OPV) response. Moving toward global poliovirus eradication, the WHO recommended phased OPV-to-IPV replacement on schedules in 2012. Using the MAL-ED prospective birth cohort data, we evaluated the influence of early life exposures impacting OPV immunization by measuring OPV response for serotypes 1 and 3. Methods: Polio neutralizing antibody assays were conducted at 7 and 15 months of age for serotypes 1 and 3. Analyses were conducted on children receiving ≥3 OPV doses (n = 1449). History of vaccination, feeding patterns, physical growth, home environment, diarrhea, enteropathogen detection, and gut inflammation were examined as risk factors for non-response [Log2(titer) < 3] and Log2(titer) by serotype using multivariate regression. Findings: Serotype 1 seroconversion was significantly higher than serotype 3 (96.6% vs. 89.6%, 15 months). Model results indicate serotypes 1 and 3 failure was minimized following four and six OPV doses, respectively; however, enteropathogen detection and poor socioeconomic conditions attenuated response in both serotypes. At three months of age, bacterial detection in stool reduced serotype 1 and 3 Log2 titers by 0.34 (95% CI 0.14–0.54) and 0.53 (95% CI 0.29–0.77), respectively, and increased odds of serotype 3 failure by 3.0 (95% CI 1.6–5.8). Our socioeconomic index, consisting of Water, Assets, Maternal education, and Income (WAMI), was associated with a 0.79 (95% CI 0.15–1.43) and 1.23 (95% CI 0.34–2.12) higher serotype 1 and 3 Log2 titer, respectively, and a 0.04 (95% CI 0.002–0.40) lower odds of serotype 3 failure. Introduction of solids, transferrin receptor, and underweight were differentially associated with serotype response. Other factors, including diarrheal frequency and breastfeeding practices, were not associated with OPV response. Interpretation: Under real-world conditions, improved vaccination coverage and socio-environmental conditions, and reducing early life bacterial exposures are key to improving OPV response and should inform polio eradication strategies. The MAL-ED study, described elsewhere [11], [34], differs from much of the polio vaccine response literature that primarily describes controlled, clinical trials; in contrast, MAL-ED was an observational study that evaluated vaccine response under real-world conditions, which include supplemental immunization to maximize OPV response. MAL-ED enrolled participants within 17 days of birth and followed them intensively for the first two years of life. Children were included in this analysis if they received at least three doses of OPV before the protocol blood draws; those receiving IPV were excluded. The study was conducted under human use research protocols approved by local and/or national ethical review committees at each site. Signed consent was obtained for participation. Blood collection was scheduled at 7 and 15 months of age ±14 days to accommodate participant availability and illness. Poliovirus serum neutralizing antibody titers were measured using WHO-standardized microneutralization assays [12], [35]. The primary outcomes were serotype-specific non-response, defined as Log2 (titer) < 3 (hereafter called seroconversion failure), and Log2(titer). Exposures of interest are briefly defined below and in Supplemental Table 1. Children were vaccinated at local health facilities and during vaccine campaigns; not by the MAL-ED study. Structured monthly questionnaires were administered to record dates of vaccination, along with a quarterly assessment of confirmed dates and receipt of vaccination [12], [36]. Locally-defined rainy seasons were also identified to classify OPV timing. Twice-weekly household surveillance captured the occurrence of diarrheal symptoms (≥3 loose stools in 24 h) [34]. Diarrheal stools collected during household visits and non-diarrheal stools (separated by ≥2 diarrhea-free days) collected monthly in the first year and quarterly in the second year, were tested for ≥40 enteropathogens [37]. Frequency of diarrhea episodes and enteropathogen detection scores were computed at early ages (4, 8, 12 and 16 weeks) and at the time of blood draw (7 and 15 months). Diarrhea frequency was additionally assessed 1, 3, and 5 days before and after an OPV dose. Enteropathogen scores were computed as the cumulative number of pathogen detections divided by the total stools collected up to a specified age. Scores were computed separately by stool type (diarrhea vs. non) and for all stools combined. We evaluated scores for individual ((Campylobacter, Cryptosporidium, enteroaggregative Escherichia coli (EAEC), Giardia) and pathogen categories (bacteria, viruses, parasites, all combined). Finally, gut inflammation and permeability were measured using fecal α-1 antitrypsin, neopterin, myeloperoxidase, and urinary lactulose:mannitol ratio [38]. Nutritional status was measured using monthly anthropometry and serum biomarkers at 7 and 15 months. Monthly anthropometry (length (cm), weight (kg)) was converted to length-for-age (LAZ), weight-for-age (WAZ), weight-for-length (WFL) Z-scores and categorized (stunted LAZ < −2, wasted WAZ < −2, underweight WFL < −2) based on WHO standards [39]; quality control procedures revealed bias in length measures from Naushero Feroze (Pakistan) thus children from this site were excluded in analyses involving length. Growth velocity during the first three months of life was also computed. Biomarkers of nutrient status (retinol, ferritin, transferrin receptor, hemoglobin, zinc, alpha-1-acid glycoprotein) were measured from the same blood samples as OPV titers [40]. Infant feeding patterns, including, frequency of breastfeeding and age at introduction of non-breastmilk liquids and solids, were recorded during household surveillance visits [40]. Breastfeeding status was characterized as exclusive, partial, or predominant (Supplemental Table 1 for definitions). A socioeconomic status index was developed for the MAL-ED study [41]. The index is a composite of: Water/sanitation, household Assets, Maternal education, and household Income (WAMI, components range from 0 to 8; components are summed and divided by 32; WAMI ranges from 0 to 1). WAMI components were measured at 6, 12 and 18 months; however, little variation existed over time, so mean scores for these time points were used. A modified version of the Home Observation for the Measurement of the Environment (HOME [42], [43]) was administered by the MAL-ED study at 6, 24, and 36 months of age [44]. Two HOME factor scores (range 0–4) were computed: Clean and Safe Environment, which reflects (permanent) environments conducive to the safety and health of the child; and Child Cleanliness, which reflects cleanliness of the child [45]. This analysis uses the 6 month scores and change from 6 to 24 months. Analyses focused on response to serotypes 1 and 3. Univariate analyses were used to compare characteristics across sites and to assess differences in response (seroconversion failure and Log2 titers) across factors using the Cochrane-Armitage Trend test (for continuous factors divided into ordered categories) and t-tests (for factors consisting of two groups). Two multivariate models were fit for serotypes 1 and 3 with random effects to adjust for correlation within site and a fixed effect for age at sample collection (in months): a logistic model for serotype failure and linear mixed model for Log2 titer (See supplementary material). Model selection was guided using AIC statistics. Significance was interpreted at the 0.05 level; however, with five thematic areas evaluated (vaccination history, infant feeding practices, nutritional status, enteric infection, home environment), a Bonferonni corrected alpha-level of 0.01 is also provided. All analyses were conducted using SAS 9.4 (SAS Institute Inc., Cary, NC, USA) [46]. The Bill & Melinda Gates Foundation did not play any role in the writing of the manuscript nor in the study design, data collection, data analysis, or interpretation of results. The corresponding author had full access to all the data in the study and final responsibility for the decision to submit for publication.