Factors determining anti-poliovirus type 3 antibodies among orally immunised Indian infants

Highlights • 88.1% of 8454 children screened had protective antibodies to poliovirus serotype 3.• The number of tOPV doses received was the main determinant of seroprevalence.• Age, gender, residence and number of tOPV doses are associated with seroprevalence.


Introduction
The global incidence of polio cases has declined with only two countries now considered polio endemic [1]. This remarkable reduction was achieved by the effective use of vaccines, with oral poliovirus vaccines (OPV) playing the greatest role in decreasing disease and interruption of transmission in developing countries. Although OPV has many practical advantages for mass immunisation in field settings [2], like other oral vaccines the immunogenicity and effectiveness of OPV is impaired in lower-income countries [3][4][5]. Potential contributing factors for low immunogenicity in these settings include a high prevalence of diarrhoea, infection of the gut with other pathogens, malnutrition, micronutrient deficiencies, and tropical enteropathy [3,6]. http  Diarrhoea is independently associated with a failure to seroconvert following administration of OPV after adjusting for potential confounders like season, breast feeding, mass campaigns and maternal antibodies [7]. In northern India, reduced take of OPV was significantly associated with season [8]. Concurrent enteric infections with lower OPV response in low-income settings have been described [9,10]. Tropical enteropathy, resulting from high environmental exposure to enteric pathogens, is common among children living in poverty and may be associated with poor response to oral vaccines, both in terms of primary antibody response and its longevity [11].
Serological data are informative about vaccination coverage, immunogenicity, secondary spread of vaccine poliovirus and exposure to wild-type infections. However, there are limited published data available on antibodies to polio in Indian children in the recent past, particularly from southern India. Table 1 presents a comparison of recent data on seroprevalence from developing countries.
After vaccination with tOPV, antibody responses are greatest to poliovirus type 2 and usually lowest to serotype 3 (PV3) [3]. In 2009, the baseline seroprevalence of antibodies to PV3 among infants aged 6-9 months was just 48% in a community-based randomised clinical trial conducted in a high risk area, Moradabad in northern India [12].
We now report a community-based seroprevalence study of anti-poliovirus type 3 antibodies among infants of age 5-11 months who had not previously received inactivated poliovirus vaccination (IPV) residing in rural and urban areas of Vellore district of Tamil Nadu, southern India. This study was done to screen for a clinical trial on the effect of azithromycin on the immunogenicity of serotype-3 monovalent oral poliovirus vaccine (mOPV3) given to healthy infants without antibodies to serotype-3 poliovirus [10], which found that removal of bacterial pathogens by azithromycin treatment did not increase the proportion of children who responded to mOPV3.

Study design and setting
The cross sectional survey was carried out in 210 health subcentres (HSC, each serves a population of 5000) of 42 primary health centres (PHC, each serving approximately 20,000-30,000) in 14 health blocks (serving 80,000-120,000 and as referral facility for 3-4 PHCs) of the rural and urban parts of Vellore district of Tamil Nadu between July 2014 and January 2015.
Infants in the study area receive routine immunisation either from government or private health care facilities, tOPV is given with BCG at birth and at 6, 10 and 14 weeks along with DPT in the study area. IPV was not available in the government sector during the study period. The last supplemental immunisation activity (SIA) was in February 2014 and no SIAs were carried out during the study period.
The Christian Medical College Institutional Review Board and the Imperial College Research Ethics Committee approved the study and appropriate central and state governmental permissions were obtained prior to conducting the screening. Investigators and study coordinators met with local community leaders, private and government health providers and informed them about the study and requested their cooperation.

Study population
We did a door to door survey to identify infants and written informed consent was obtained from all willing parents of eligible healthy infants aged between 5 and 11 months. The Village Health Nurses (VHNs) of the concerned Health sub-centres discussed the study with potential participants and motivated the families to participate. A screening camp was organised in each village, and parents brought the child to the camp. Each infant was assigned a unique screening identification number and basic demographic details were collected. A study physician examined the infant for eligibility for screening and recorded the infant's age and polio vaccination history from the immunisation cards. Additional doses received during National Immunization Days were obtained from verbal history as these doses are not recorded on immunisation cards and if immunisation cards were not available, the mother's statement was recorded. Exclusion criteria included children who had received IPV-through private healthcare providers, had any congenital or chronic illness or had high grade fever or any other illness that prevented participation as decided by the study physician. Infants temporarily excluded because of minor illnesses were asked to visit the camp held in a neighbouring village.

Laboratory methods
Blood specimens were collected by trained phlebotomists and study nurses. Samples were stored on ice and delivered to the laboratory on the same day. Assessment of poliovirus-specific neutralising antibodies to serotype 3 was done using a microneutralisation assay according to the World Health Organization (WHO) protocol with modifications [13]. Briefly, a 2-fold dilution of each serum sample (50 ll) ranging from 1/4 to 1/8 was mixed with 50 ll of approximately 100 median tissue culture infective dose (TCID 50 ) of Sabin 3 poliovirus in replicate wells at each dilution and the mixture was incubated at 37°C (5% CO 2 ) for 1 h. 100 ll of Vero cell suspension (5000 cells/well) was then added to all the wells and the plates were incubated for 3 days at 37°C (5% CO 2 ). As part of quality control, standard polio antisera from the U.S. Centers for Disease Control and Prevention were included in each run. For each assay, a back-titration titre of 30-300 TCID 50 was considered acceptable. Cell controls were included in each assay. A reciprocal titre of <8 was considered non-protective.

. Data management and analysis
Single data entry of the questionnaire was done using Epi-Info 3.5.1 (CDC, GA, USA) and was verified by a statistician. SPSS 18 (SPSS Inc., IL, USA) and STATA 13 (StataCorp, TX, USA) software were used for analysis. All statistical tests were carried out at 5% significance level (p < 0.05) and confidence interval (CI) set at 95%. Student's t-test was used to compare means. Univariable and multivariable logistic regression was performed on sex, age, number of OPV doses received and place of residence and odds ratio (with 95% CI) calculated to ascertain the strength of association between the exposure and outcome variable. In addition, multilevel logistic regression (MLR) was performed to examine the effect of contextual (location, period, administrative) and compositional (individual variation) effects on model parameter estimates [14]. We estimated the probability of seroconversion for each dose of OPV using the binomial likelihood function, allowing for an arbitrary baseline level of seroprevalence among infants who were reported to have received OPV.

Baseline characteristics
Of 12609 children in the target age group invited to participate; 429 (3%) refused participation, 1224 (10%) had received IPV, 2150 (17%) did not visit camp sites because they were unlikely to be available and 352 (3%) were screen failures based on medical grounds (congenital heart disease, fever, etc.). Samples were collected from 8454 children from 10 rural health blocks and 4 urban health centres with an average of 604 children per block (range 89-1184), with a mean age of 8.3 (SD-1.8) months. Baseline characteristics of the study population are shown in Table 2.

Univariable and multivariable regression
The results of the univariable, multivariable and multivariable mixed effects logistic regression analysis are presented in Table 4. In the multivariable regression, male children were more likely to be seropositive (OR 1.27, 95% CI: 1.11-1.46). In this same model, children had higher odds for PV3 seropositivity for each month increase in age (OR 1.17, 95% CI: 1.12-1.23) and one dose increase of OPV (OR 1.74, 95% CI: 1.61-1.89). Similarly children residing in urban areas had higher odds of being seropositive compared to rural children (OR 1.24, 95% CI: 1.05-1.45) in the standard multivariable regression analysis but this was not significant when analysed in the multivariable mixed effects logistic regression at the block level (OR 1.18, 95% CI: 0.98-1.41). The mixed effects regression performed to investigate the cluster effect at various levels show a significant random effects at the block level (likelihoodratio v 2 = 11.3 with 1 degree of freedom; p = 0.0004) and not at the PHC and HSC levels. However, we did not find any appreciable changes in the parameter estimates.

Discussion
The seroprevalence among infants of age 5-11 months in rural and urban areas of Vellore district of Tamil Nadu who had not previously received IPV was 88.1% (95% CI: 87.4-88.8). This is similar to rates in Pakistan (Table 1), but slightly lower than in Sri Lanka [15].
The seroprevalence increased with number of OPV doses received and is comparable to the seroconversion rates reported by John [16] and in concurrence with Indian [17][18][19] and other studies [20][21][22][23]. Similar to other studies, seroprevalence rates increased with age [19,21,24,25]. This could be due to a more mature immune system or to receipt of OPV doses that are not recorded in the vaccination history taken for the child.
Interestingly, there was limited variability in seroprevalence rates across the various rural and urban blocks (Table 3), demonstrating that the Tamil Nadu's state immunisation system is able to substantially deliver vaccines.
Our study demonstrated that urban infants have higher seroprevalence which may be due to increased coverage, accessibility and higher literacy rates. This is similar to a Mexican study where urban children had higher antibody prevalence rates (88.2%) than    rural children (82.9%) [23] but differs from other studies that report no difference by place of residence [26,27]. Seroprevalence was marginally higher (89.2%) in males compared to females (87.0%) consistent with another study [22], but differing from Nigeria where females had higher antibody titres for all the three serotypes in one study [28] or were not different [29] and a Chinese study that also found no difference [30]. Sex differential non-specific effects of vaccines are common in developing countries with negative non-specific effects (NSE) of inactivated vaccines more common in girls than boys [31,32] in some countries, but with no differences in high income countries [33]. Though the general pattern is both negative and positive NSE are stronger in females [34], a randomised controlled trial in Guinea-Bissau negated the hypothesis that mortality rates in boys would be lower if they had not received OPV0 [35]. The reasons why OPV uptake is slightly higher for males in our study area are unknown.
Though factors such as age, gender and urban setting have no remedial solution from a public health perspective, this study indicates that response may be affected by factors that are not amenable to modification. This is important for the understanding of oral vaccine performance in low-income countries.

Limitation
One of the limitations of this study was that we were unable to collect more detailed demographic information, morbidity and nutritional status, other than the physician's assessment because of logistic constraints.

Conclusion
To our knowledge this is the largest seroprevalence study in a low income setting in a narrow age range. Seroprevalence was associated with age, gender, number of OPV doses received and place of residence. The high rate of seropositivity to the weakest antigen in the trivalent OPV in recent studies in India differs markedly from older studies where low rates of seropositivity to this antigen were reported. Perhaps the recent switch from tOPV to bOPV will further increase immunogenicity to type 3 by elimination of the interference by type 2.
India has seen improvements in immunisation coverage such that seroprevalence is now reasonable. However, immunogenicity remains lower than in high-income countries, consistent with results for other oral vaccines. For future oral vaccines, understanding the reasons for poor performance in low-income settings remains important.