The transfer of maternal antibodies and dynamics of maternal and natural infection-induced antibodies against coxsackievirus A16 in Chinese children 0–13 years of age: a longitudinal cohort study

Coxsackie A16 (CVA16) belongs to the picornavirus family, which can cause a wide spectrum of diseases, such as encephalitis, paralysis or myelitis, meningitis, hand-foot-and-mouth disease (HFMD), upper and lower respiratory tract diseases, pleurodynia, herpangina, and myopericarditis [1]. HFMD is a common infectious disease, with CVA16 as the major pathogen, mainly observed among children under 5 years old in the Asia-Pacific region, especially in China. CVA16 predominated in China in 2012, 2014, 2016 and 2018, accounting for 58.6% of HFMD outbreaks between 2011 and 2018. An increasing number of HFMD cases caused by CVA16 have been observed since 2017 [2]. Although most patients show minor symptoms, CVA16 can cause severe and fatal HFMD, and increasing numbers of reports of such outcomes have been noted in the US, France, Japan, mainland China, and Taiwan of China [3]. In Shenyang, northeastern China, 20.7% (19/92) of HFMD cases with neurological symptoms were caused by CVA16 infection, of which 2 patients had brainstem encephalitis and 1 had acute flaccid paralysis [4].

There is currently no specific treatment for HFMD, and the CVA16 HFMD vaccine is still in the preclinical phase, with promising results in animal models [3]. The only HFMD vaccine licenced in mainland China is the enterovirus A71 (EV-A71) monovalent HFMD vaccine, but this vaccine fails to confer cross-protection against HFMD caused by CVA16 or other enteroviruses [5‐7]. All of these elements highlight the urgent need for the development of vaccines targeting CVA16.

Maternal antibodies have been recognized as an effective source to protect newborns against measles virus and influenza [8, 9]. Given the absence of a CVA16 vaccine, anti-CVA16 maternal antibodies, as the only source of passive immunity, may play an important role against CVA16 during the early months of life. After the waning of maternal antibodies, young children may become susceptible to infection. Serological studies are necessary to quantify maternal antibody transfer and anti-CVA16 antibody dynamics, including maternally transmitted antibodies as well as natural infection-induced antibodies. Such understanding may help select the target population for future CVA16 vaccination campaigns. Few studies have reported a positive correlation of anti-CVA16 antibody levels in mother-neonate pairs [10‐13]. Only one study reported the average transfer ratio of anti-CVA16 maternal antibodies [11], and the half-life of maternal antibodies was not reported. Limited cohort studies describing maternal antibody dynamics showed that seroprevalence in neonates decreased from 72% at birth to 0–26.4% at 6–7 months of age. However, these results may be affected by the small sample size [12] and limited follow-up [14, 15]. In addition, previous studies were cross-sectional studies, which cannot reveal the dynamics of natural infection-induced antibodies over time [16‐20].

Here, we quantify the transfer of maternal antibodies and the dynamics of antibodies from birth to 13 years old using two population-based cohorts: one was a paired mother-neonate cohort (abbr. neonate cohort), and the other was a 1- to 9-year-old paediatric cohort (abbr. paediatric cohort). More specifically, we quantified concentrations of CVA16-specific antibodies in paired maternal and cord serum samples and assessed the transplacental transfer efficiency of maternal antibodies from a large cohort of mothers and neonates. We then analysed the antibody kinetics from birth to age 13 years (including the decline in maternal antibody levels and subsequent increases due to natural infection) using the paediatric cohort and neonate cohort together.

A total of 3013 children, including 538 neonates and 2475 children aged 1–9 years old, were enrolled in our study between Sep 2013 and Sep 2018. The follow-up rate of the neonate cohort was above 60% (Fig. 1A). Over 80% of children in the paediatric cohort were followed up annually. The group with semiannual follow-up had at least an 85% follow-up rate from Sep 2013 to Sep 2016 (Fig. 1B). The local circulation of CVA16 during 2013–2018 in Anhua County, Hunan Province, is depicted in the Appendix (Additional file 1: page 6–8, line 6–42).

The general characteristics of the enrolled neonates and children at baseline are provided in Table 1. The median age of the participants was 3 years old (IQR: 1–6), and 1532 (51%) children were boys. The follow-up time ranged from 1 to 41 months (median: 22 months; IQR: 11–29 months). The median age of mothers at baseline was 24 years old (IQR: 26–29 years old). Ninety percent of the mothers’ sera were collected at delivery (Additional file 1: Table S4 (a)). Comparisons between participants and local nonparticipants are presented in the appendix (Additional file 1: pages 12–16, line 118–125).

As shown in Fig. 2A, the baseline anti-CVA16 antibody titre of neonates at birth was highly correlated with the maternal anti-CVA16 antibody titre (ρ=0.80, 95% CI 0.77–0.83). The GMT and seroprevalence were similar between mothers and neonates at birth (mother vs. neonate: 9.82 (95% CI 9.14–10.55) vs. 9.61 (95% CI 8.97–10.30), p= 0.102; 44.80% (95% CI 40.53–49.10) vs. 41.82% (95% CI 37.61–46.12), p= 0.349, respectively).

The probability of neonatal positivity increased as the maternal antibody titre rose until the maternal antibody titre was >28 for CVA16, with the neonatal probability being close to 100% (Fig. 2B). The overall maternal anti-CVA16 antibody titre transfer ratio was 0.88 (95% CI 0.80–0.96). It decreased with maternal antibody levels (Fig. 2C), with an average value larger than 1 when maternal antibody titres were ≤13. Although the transfer efficacy decayed, neonates can successively obtain positive maternal antibodies via placental transmission as long as the maternal antibody titre reaches the positive level (Fig. 2C, Additional file 1: Fig. S4).

Multivariate analysis (Additional file 1: Table S8) indicated that maternal anti-CVA16 antibody titres were highly correlated with neonates’ titres at birth and the status of antibody titres (positive or not). For every 2-fold increase in maternal titres, neonatal titres increased by 69.1% (95% CI 63.4–75.2). The probability of acquiring positive antibody levels for neonates at birth increased by 17.7 times (95% CI 15.5–20.0) for every 2-fold increase in maternal titres. Having a twin decreased the probability by 15.7% (95% CI 1.4–28.0) compared to being a single neonate. The transfer ratio decreased by 14.8% (95% CI 11.8–17.6%) for every 2-fold increase in maternal antibody titre. The transfer ratio was reduced by 16.4% (95% CI 2.8–28.1%) when the mother had liver system disease.

Sensitive analyses with a positive threshold at 16 are depicted in the appendix (Additional file 1: Fig. S7).

Maternal transmitted antibody levels in neonates diminished rapidly. This drop was followed by an increase in titres due to natural infection, as was apparent in both the GMT and seroprevalence stratified by age groups (Fig. 3A, B). Forty-two per cent of neonates had positive antibody levels, with a GMT of 10 (95% CI 9–10). However, the GMT decayed below the positive threshold after 1.5 months (range: 1.0–2.0 months) and further decreased to a minimum at 6.8 months. The seroprevalence experienced a sharp drop, falling to 20% after 1.3 months, with the lowest point at 5.8 months. Then, due to natural infections, the seroprevalence increased rapidly from 6 months to 5 years old (60%), then slowly increased to 80% at 13 years old. The GMT increased above the positive threshold after 1.4 years (range: 1.2–1.5 years) and then reached 32 by the age of 5 years. It slightly dropped thereafter. For the sensitivity analysis with a higher threshold of 16, the kinetics of seroprevalence and the GMT were consistent with those in the baseline analysis. The GMT increased above 16 by 2.7 (range: 2.4–3.1) years of age (Additional file 1: Fig. S7). The kinetics of antibodies from high-level, low-level and negative maternal antibody groups were significantly different within 6 months. The higher the level of maternal antibody, the longer the duration of neonatal titre decay to under the detective threshold (high-level vs. low-level: 5.3 months vs. 1.2 months) and to the bottom (high-level vs. low-level: 11.0 months vs. 8.1 months). The antibody dynamics in the lower level of maternal antibody groups, especially the negative group, rose earlier due to natural infection, and the corresponding times for the three groups to rise to 8 were 1.5, 2.0 and 2.1 years, respectively. However, the difference was not obvious when age more than 1.5 years old (Additional file 1: Fig. S9).

The maternally transmitted antibody levels decayed faster than natural infection-induced antibody titres (p<0.001), with corresponding antibody half-lives of 2.0 months (95% CI 1.9–2.2 months) and 13.9 months (95% CI 13.3–14.5 months), respectively (Additional file 1: Fig. S10). Half of the neonates who were seropositive at birth were seronegative by 2.4 months of age (95% CI 2.3–2.5 months) (Fig. 4A). The average time to loss of seropositivity was 2.8 months (95% CI 2.5–3.1 months). For every 2-fold increase in maternal antibody, the duration of protection was prolonged by approximately 38 days (95% CI 28–48 days). When the maternal antibody titre was ≥64 and >512, the duration of protection for neonates was extended to 6 and 12 months, respectively (Fig. 4B). The sensitivity analyses are shown in Additional file 1: Table S13.

Our study revealed that the average anti-CVA16 antibody titres and seroprevalence were similar in maternal and cord blood, with a mean transfer ratio of maternal antibodies of 0.88 (95% CI 0.80–0.96). Although the transfer ratio decreased as the maternal antibody titre increased, the positive maternal antibodies could still be successfully transmitted to neonates. However, due to the low maternal seroprevalence, only 42% of neonates acquired positive concentrations of anti-CVA16 antibodies from mothers, which declined rapidly, with a half-life of 2.0 months (95% CI 1.9–2.2 months). Among them, the average duration of protection from maternal antibodies was 2.8 months (95% CI 2.5–3.1 months), and the duration was extended to 6 months on average for mothers with titres of ≥64. After the disappearance of maternal antibodies, antibody titres increased because of rising natural infections.

Transplacental antibody transfer from mother to foetus provides protection from infection in the first weeks of life, such as measles, rubella, diphtheria, and tetanus, with various transfer efficacy. A strong association between maternal anti-CVA16 antibodies and neonates with a correlation coefficient of 0.8 was observed in our study, which was close to 0.56–0.75 in previous studies [10‐12]. The transfer of anti-CVA16 maternal antibody is efficient, with an overall transfer ratio of 0.88, which is similar to Fu’s observation of 0.9 [11] and is lower than that for EV-A71 [21], measles, and poliomyelitis III antibodies [11, 21]. The trend of the transfer ratio decreasing as the maternal antibody titre increases has also been observed for measles [26] and EV-A71 [21], which may be due to the saturation of a finite number of Fc receptors [26, 27]. Maternal antibody level is the major factor associated with the transfer ratio, following the mother’s health. Liver system disease in the mother was associated with a lower transfer ratio. Other factors associated with mothers, such as maternal age, gestational diabetes, gestational hypertension, parity and gravidity, as well as factors associated with neonates, such as gestational age and birth weight, were not found to be related to transfer in our study.

The seroprevalence and GMT in our study were minimal at the ages of 5.8 months and 6.8 months, respectively, in line with previous longitudinal observations [12, 14, 15, 28]. The neonatal seroprevalences at birth and 2-4 months of age were lower than those in the studies by Jiangsu, potentially because a higher prevalence (90%) was observed in mothers in those studies [15, 29]. The time when the GMT reached 8 and when it reached its minimum value were similar to those in Guangzhou studies (at 3 and 6 months old) [12, 28]. No previous longitudinal study investigated the kinetics of the anti-CVA16 antibody level, and only cross-sectional studies reported age-specific seroprevalence and GMT. Our age-specific seroprevalence results between 6 months and 13 years were comparable with those of previous reports in China [30], Singapore [17], Korea [16] and Thailand [18].

Using a neonatal cohort, we showed in another study that the dynamics of anti-EV-A71 antibodies had a similar pattern [21]. With higher maternal seroprevalence, neonatal EV-A71 seroprevalence was significantly higher than that of CVA16 for neonates aged 0–4 months old, so it took slightly longer for the GMT to decay below the positive threshold. After the decay of maternal antibody, the seroprevalence was similar between CVA16 and EV-A71 within 4 months to 3-year-old children (unpublished study), although the EV-A71 GMT was higher than the CVA16 GMT. A similar pattern was also observed in previous studies [31‐33], showing that EV-A71 might elicit stronger neutralizing antibody responses than CVA16. The capsid proteins of the two viruses have approximately 20% sequence differences, as well as antigenic variation caused by the different surface structures of the two viruses, which may explain this phenomenon [34].

After the rapid decay of maternal antibodies, the seroprevalence increased with age as a result of natural infection; however, 40–95% of children from 6 months to 5 years old were still seronegative. In addition, to curb the transmission of CVA16 infection, a threshold of 59–96% for herd immunity is needed [2, 35‐37]. Meanwhile, 0- to 5-year-old children, accounting for 99% and 96.5% of severe and mild cases in Hunan Province, respectively, constitute the majority of HFMD cases associated with CVA16 (Additional file 1: Fig. S11). Among them, the 1-year-old group had the largest number of mild and severe cases, which was in line with a previous report [23]. The above evidence showed that children aged 0–5 years old represent the main susceptible population, which should be targeted for vaccination.

Overall, both avoiding potential maternal antibody interference and preventing the largest incidence, especially severe cases, in the 1-year-old group, we recommend that the optimal vaccination time be between 1.5 months and 1 year old. Complete vaccination during this period could prolong the positive antibody titre until 5 years old, although this hypothesis is based on the half-life of anti-CVA16 antibody induced by infection. Further research is required on antibody dynamics induced by vaccines. Thus, we call for accelerating the development of a human CVA16 vaccine and recommend clinical trials to further assess doses, safety, and effectiveness in susceptible young children aged 1.5 months to 5 years old.

Vaccinating pregnant mothers to provide maternal antibodies to infants who are too young to be vaccinated could be an alternative intervention to protect these infants, which has proven efficacious for influenza, pertussis and COVID-19 [8, 38, 39]. As we estimated, anti-CVA16 maternal antibodies can be effectively transmitted to newborns, and high maternal antibody levels can prolong the duration of protection for infants, with protection lasting up to 6 months when the maternal antibody level is ≥ 64. Meanwhile, the results of antibody dynamics specified by maternal antibody levels also indicated that a higher maternal antibody group was associated with later antibody decay to undetectable levels and later natural infection. It is possible to increase the mother’s antibody titre to extend protection to 1 year of age. Given that, vaccination of mothers during pregnancy can be a solution to protect infants from disease and death in early life, especially those under 1 year old. This conclusion is based on observations of antibodies mounted upon natural infection rather than vaccination, and we therefore still require further research to elucidate the efficacy or effectiveness of maternal CVA16 vaccination, antibody response and duration of immunity produced by maternal CVA16 vaccination.

A study in Vietnam found that adults can have HFMD and contribute to household transmission [40]. In our study, the seroprevalence of mothers (44.8%) was lower than that of mothers (87–93%) and adults (65–83%) in some coastal provinces of China [11, 15], Germany [19] and Vietnam [40]. Given the absence of sex-specific and age-specific differences in CVA16 prevalence [19, 41, 42], males and adults of other ages are also susceptible in theory. Thus, vaccination for adults in close contact with infants could help reduce household transmission and provide indirect protection for vulnerable infants.

This study is a population-based longitudinal study with a large sample size, prolonged follow-up design and high follow-up rates, in contrast with those of previous studies [10, 11, 14‐17]. The quality control of the neutralization test was stringent. The conclusion using children with complete follow-up is consistent with the main results (Additional file 1: Fig. S8). However, our study still had some limitations. First, we were unable to correlate antibody levels with clinical symptoms because only a few cases were reported during the follow-up period. Moreover, there are currently no studies confirming the association between antibody titre and symptom reduction or avoidance of infection. Thus, the positive threshold in this study is a threshold for detecting antibodies in neutralization assays. It is expected that a higher titre is associated with higher protection; thus, a sensitivity analysis with 16 as the positive threshold was conducted, which showed a conclusion consistent with that of the baseline analysis. Second, we were unable to compare the demographic information between the study participants and the local eligible nonenrollees in the 1–9 years old cohort because they refused to provide personal information. Finally, the sample size of mothers with higher antibodies (≥256) in our study was relatively small, causing the relative decreasing trend of directly calculated prevalence when the maternal antibody titre was 512, while the predicted probability was stable due to logit transformation in logistic regression.

In conclusion, positive titres of anti-CVA16 antibodies were efficiently transferred to neonates from mothers through the placenta. However, titres declined rapidly to below the positive threshold by 1.5 months of age. Increasing the maternal antibody titre can effectively prolong the duration of protection. The majority of children aged 0–5 years old are susceptible, especially 1-year-old children with a high risk of developing severe disease. It is important to assess the feasibility of vaccinating children between 1.5 months and 1 year old, pregnant women or adults in close contact with infants to protect 0–5-year-old children.