Background
Diarrheal diseases (DD) are one of the leading causes of death in children ≤5 years old, accounting for almost 10% of deaths in this age group [
1‐
4]. Globally, rotavirus A (RVA), norovirus (NoV) genogroup II, astroviruses (HAstV),
Campylobacter sp.,
Cryptosporidium sp
., enterotoxigenic
Escherichia coli, and
Shigella sp. are the most prevalent agents of DD [
5‐
9]. The Global Rotavirus Surveillance Network has shown that, although approximately 40.3% of DD cases can be attributed to RVA globally; in countries of the Americas with universal vaccination this proportion is 12.2% [
10]. In China, a country where the two licensed RVA vaccines were not included in the routine vaccination schedule, the overall rate of RVA positivity in children with DD is 30% [
11].
RVA possesses an RNA genome with 11 gene segments and is commonly classified using a binary system based on two outer and most immunogenic capsid proteins [
12]. Among the most prevalent RVA genotypes, G1P[8], G3P[8], G4P[8], G9P[8] and G12P[8] belong to the Wa-like genomic constellation, while G2P[4] belongs to the DS-1-like constellation [
13]. These two major genomic assemblages display nucleotide sequence identities varying from 75 to 90% [
13,
14].
In 2017, Brazil completed a decade of vaccine implementation (the attenuated monovalent G1P[8] vaccine (Rotarix®, RV1) in the National Immunization Program (NIP), which has expanded substantially in the last years. Vaccination with RV1 consists of two doses. Infants aged 6 weeks to 8 months are vaccinated. The first dose should be given until the age of 3 months and 15 days, and the last dose up to 7 months and 29 days.
According to the World Health Organization and the Pan American Health Organization (WHO/PAHO), RVA vaccines are strategic to reduce DD burden, along with oral rehydration, breastfeeding, zinc administration and improvement of sanitation [
15]. So far, more than 81 countries introduced RVA vaccination since October 2016. From these, 63 countries introduced RV1 and 18 countries implemented vaccination with the pentavalent vaccine RotaTeq® (RV5) (four countries introduced both RV1 and RV5). Currently, the Global Alliance for Vaccines and Immunization (GAVI), supports RVA vaccination in 45 developing countries (
https://www.gavi.org/results/countries-approved-for-support/).
A meta-analysis of RVA surveillance studies – including data from countries that participated in the WHO RVA surveillance network from 2008 to 2013 – estimated a reduction on a global scale from 528,000 to 215,000 RVA-associated deaths in children ≤5 years old from 2000 to 2013. In the same period, the RVA detection rates in children with DD declined from 42.5 to 37.3% [
16]. The positive impact of RVA vaccination on DD-associated hospitalizations and deaths has been well demonstrated in Brazil and several other Latin American countries [
17,
18]. In Brazil, effectiveness is higher among infants aged up to 12 months, decreasing in older children [
19‐
21].
In this study, we accessed the impact of Rotarix after ten years of its implementation in the NIP in Brazil. For this propose, we explored RVA detection rates and genotype distribution in DD samples collected from children in the pre- and post-vaccination periods.
Discussion
The current study demonstrates, by laboratory-based surveillance, a decrease in the frequency of infection with RVA in children presenting with DD after RV1 implementation in Brazil. As recently reviewed, some studies have demonstrated the impact of universal anti-RVA vaccination in Brazil; significant declines of diarrhea-associated hospitalization rates among children ≤5 years-old and infants have been described [
23‐
28]. The present study demonstrated that the reduction in the frequency of RVA infection occurred mainly among children aged 4–11 months-old and 12–24 months-old. The main goal of RV1 vaccination is to prevent severe RVA infections during the first two years of life, and it is well known that DD is more severe in age groups aged less than 24 months-old, the group in which hospitalization occurs due to severe dehydration leading to more frequent deaths. Therefore, it was expected that the main impact of vaccine introduction was likely to occur in age groups less than two-year old. Several studies have shown that RV1 induced immunity protects children from RVA infection in the first two years of life [
21,
23,
24,
29].
Interestingly, after RV1 introduction, RVA-positivity showed an increasing trend in children aged 25–48 months-old. Our data are consistent with data reported in the USA, that also demonstrated a shift in the age group distribution of RVA infections, following the introduction of the anti-RVA vaccination [
30]. The changes in the age at which children are more likely to become infected with RVA should be considered a beneficial effect of the vaccine.
A somewhat cyclical pattern of genotype circulation was observed, with a 10-year interval between two G2P[4] detection peaks. We observed a long cycle, where DS-1 like and Wa-like genotypes alternated in a 10-year interval and short cycles, where Wa-like genotypes, including G1, G9, G3 and G12 alternated at 2–3 year intervals. The peak of genotype G9 observed in 2005 was mostly attributed to a large outbreak of DD that affected more than 12,000 patients in the state of Acre, Amazonian region of Brazil [
31]. At that time, the epidemic of DD was associated with RVA, mainly with genotype G9P[8]. In the same year, other studies described the high circulation of the genotype G9 worldwide. Important changes in RVA genotype distribution have been reported in many countries in the last decades. Among these changes, it is worth pointing out the emergence of G9P[8] in the late 1990s, becoming a very frequent genotype together with G1P[8]. Nonetheless, the most striking global shift in RVA genotype distribution was the reemergence of genotype G2P[4] twelve years ago, shortly preceding and just after implementation of RV1 introduction in Brazil, as well as in countries which did not implement universal RVA vaccination. The fact that RV1 efficacy and effectiveness against genotype G2P[4] is somewhat lower than that observed for Wa-like strains has led to the hypothesis that the long period of G2P[4] predominance could be related to vaccination with RV1 [
18‐
21].
The putative influence of vaccination on the temporal cycling of RVA genotypes was analyzed, and demonstrated an alternation between P[8] and G2P[4]; in turn, G1P[8] and G9P[8] also alternated with each other [
32]. However, it should be observed that distinct genetic variants of G2P[4] circulated between 2005 and 2011 in Brazil, and no evidence of selective pressure imposed by the RV1 massive vaccination was observed [
33,
34]. In addition, the comparison of G2[4]/G[NT] detection rates in vaccinated and unvaccinated children does not suggest that breakthrough infections have occurred more frequently by this genotype. Interestingly, this was observed with G12P[8]/P[NT], which was detected more often among vaccinates than among non-vaccinates, suggesting some level of vaccine escape for this genotype. RV1 vaccine coverage in Brazil increased from 87 to 95% between 2011 and 2015, having decreased to 88 and 77% in 2016 and 2017, respectively.
The second period of G2P[4] predominance in Brazil lasted 5 years. A recent study performed in Brazil, revealed that new variants of G2P[4] started circulating in Southeastern, Northeastern and Southern regions in 2008, and Northeastern and Southeastern regions in 2010 [
34]. It was observed that the re-emergence of G2P[4] was a global phenomenon, and was reported even in countries that had not introduced anti-RVA vaccine [
35,
36]. It should be pointed out that in Argentina, a neighbor country where universal vaccination against RVA was implemented in 2015, long lasting predominance of G2P[4] strains started in 2004, and extended until 2011 [
37,
38].
Although the re-emergence of G2P[4] appeared not to be associated with the onset of heterotypic vaccination with RV1, we cannot exclude that the massive predominance of G2P[4] in Brazil from 2006 to 2010 may have been influenced by vaccination with RV1. Nevertheless, G2P[4] could not stay for more than 5 years in the environment of vaccinated children, possibly due to the natural induction of homotypic immunity and depletion of the susceptible population. Another noteworthy finding of our study is the re-emergence of G3 from 2011 onwards, replacing G2 predominance after its exhaustion. The G3P[8] genotype has been detected in a higher frequency in the USA, Australia and other countries in the years that followed massive vaccination with RV5 [
39]. Reemergence of G3P[6] and G3P[8] was also reported between 2011 and 2012 in Northern Brazil [
40]. In the last year of our observation period (2017), another significant increase in the detection rate of G3P[8] was observed. Also observed was a peak of G12P[8] in 2014 and 2015. Luchs et al. (2016) reported a countrywide spread of genotype G12P[8] in the years of 2014 and 2015 in Brazil [
41]. Moreover, a global G12 emergence has been observed in the last five years [
42].
Our data demonstrates that after RV1 introduction, RVA has been detected in significantly higher frequencies among non-vaccinated children compared to vaccinated ones. These differences were greater in children aged 4 to 11 months, followed by children aged 12 to 24 months. Even in children older than 24 months the RVA detection rates was significantly lower in vaccinated than in non-vaccinated children.
Our study design has limitations, since the health services spontaneously send fecal samples, and consequently there was no systematic sampling in space and time, making data susceptible to bias. However, the results, because they are comprehensive and have been generated by the official surveillance system, shed light in the RV1 vaccination impact in Brazil, and its putative influence in the burden of RVA in the country.
Monitoring other DD viral agents, especially norovirus – detected in high frequency in children with DD – is a current challenge in this new scenario. Continuous viral surveillance must be carried out in Brazil to monitor the circulation of distinct RVA genotypes and other enteric viruses.