FormalPara Key Points

Since 2006, two live, oral rotavirus vaccines have reduced the substantial burden of rotavirus disease in countries that have introduced the vaccines into their national, routine immunization program.

The effectiveness of rotavirus vaccines is lower in high child-mortality settings, compared with low child-mortality settings; while there are several hypotheses, including differences in co-infections with other enteropathogens at the time of vaccination, gut microbiota composition, and prevalence of Lewis-negative secretor-positive histo-blood group antigens, the exact causes for lower vaccine performance are unknown.

Other rotavirus vaccines are currently under development that make use of alternate rotavirus strains and delivery strategies to ameliorate specific programmatic challenges with the currently available vaccines.

1 Introduction

Rotavirus is a significant cause of diarrheal hospitalizations and deaths among children < 5 years old worldwide [1, 2]. Symptoms of rotavirus infection include watery diarrhea, vomiting, and fever. Rotavirus is primarily spread through the fecal–oral route [3]. Since their licensure in 2006, rotavirus vaccines have dramatically reduced the disease burden. However, vaccines are not yet routinely available in all countries and there are limitations to the currently licensed vaccines. In this article, we present the current state of rotavirus disease burden and vaccines and discuss post-licensure rotavirus vaccine effectiveness, impact, and safety.

2 Disease Burden

Prior to the availability of rotavirus vaccines, rotavirus infection was a common experience for children < 5 years old. Studies of the natural history of rotavirus infections during the pre-vaccine era in Australia, Bangladesh, Canada, Guinea Bissau, India, and Mexico found 64–80% of children had at least one rotavirus infection in infancy and early childhood [4,5,6,7,8,9]. In 2013, an estimated 215,000 deaths due to rotavirus diarrhea among children < 5 years of age occurred worldwide, of which > 95% occurred in countries in Africa and Asia [2]. Estimated rotavirus mortality rates by country are visualized in Fig. 1. Rotavirus infections also represent a significant proportion of diarrhea hospitalizations in young children [10]. The Global Rotavirus Surveillance Network, a sentinel hospital-based active surveillance system for diarrhea coordinated by the World Health Organization (WHO), has found that rotavirus accounted for over one third of diarrhea hospitalizations before the vaccine was introduced [11].

Fig. 1
figure 1

Estimated rotavirus mortality rates by country, 2013

3 Rotavirus Biology and Immunity

Rotavirus consists of a non-enveloped RNA virus with a segmented genome surrounded by a triple-layered protein capsid [12, 13]. While many species of animals can be infected by rotavirus, human infections are primarily caused by group A; two structural proteins in the outer layer—VP7, the G protein, and VP4, the P protein—determine the genotype [12,13,14]. In a synthesis of 30 years of pre-vaccine data, more than two thirds of rotavirus infections globally were attributed to G1P[8] and an additional 20% were due to G2P[4], G3P[8], and G4P[8] [14]. There is some variation in circulating genotypes by year and region; for example, G1P[8] accounted for just 23% of infections in Africa and 34% in Asia. More recently, G9 and G12 viruses have contributed to a larger proportion of global rotavirus infections [15].

Initial natural rotavirus infections, including asymptomatic neonatal infections, have been shown to reduce the frequency and severity of subsequent rotavirus infections during childhood and beyond [4, 5, 7, 16]. First natural infections were 56–77% efficacious against later rotavirus infections and increased with each subsequent infection; protection was higher against diarrheal illness caused by rotavirus [5, 7, 16]. Rotavirus infections also offer some protection against other strains of rotavirus causing diarrheal disease [4, 5]. These findings provided the rationale for the development of an infant vaccine against rotavirus. Despite this body of research, no good correlate of protection against diarrheal disease due to rotavirus has been identified.

4 Licensed Rotavirus Vaccines

Since 2006, two live, oral rotavirus vaccines have been licensed in > 100 countries and pre-qualified by the WHO. RotaTeq (Merck and Co) is a three-dose, pentavalent bovine-human reassortant rotavirus vaccine that is distributed in a single, 2-mL dose vial; and Rotarix (GSK Biologics) is a two-dose, monovalent human G1P[8] rotavirus vaccine that is distributed in a single, 1.5-mL dose vial (Table 1) [17,18,19]. As of September 2017, more than 80 countries have introduced a rotavirus vaccine into their national, routine infant vaccination programs (Fig. 2) [20, 21]. Two newly licensed, live oral rotavirus vaccines—Rotasiil (Serum Institute of India), a three-dose, pentavalent bovine-human reassortant vaccine, and ROTAVAC (Bharat Biotech), a three-dose, monovalent human-bovine G9P[11] vaccine—are seeking WHO pre-qualification as of 2017 [22,23,24]. In addition to these four vaccines, domestically licensed, live oral rotavirus vaccines are available in China (LLR, Lanzhou Institute of Biological Produces) and Vietnam (Rotavin, PolyVac) [24].

Table 1 Licensed rotavirus vaccines
Fig. 2
figure 2

Rotavirus vaccine introduction status by country, 2017

Clinical trials of vaccine efficacy and post-licensure effectiveness evaluations of RotaTeq and Rotarix in high- and upper-middle-income countries demonstrated that the vaccines were > 90% effective in preventing severe rotavirus disease [17, 18, 25,26,27,28]. However, clinical trials in countries with higher child mortality found that rotavirus vaccines were less efficacious in these settings, with published efficacy estimates ranging from 50 to 64% [29,30,31]. Similar to the results from the clinical trials of RotaTeq and Rotarix, ROTAVAC’s vaccine efficacy against hospitalization for rotavirus diarrhea was 54% in a clinical trial in India and Rotasiil’s was 67% in a clinical trial in Niger and 39% in India [22, 23, 32].

ROTAVAC and Rotasiil promised improvements on certain programmatic aspects of Rotarix and RotaTeq. For example, ROTAVAC is expected to cost US$1 and Rotasiil is expected to cost US$2.50 per dose, compared with approximately US$3.20 for RotaTeq and US$2.50 for Rotarix through UNICEF’s supply division [22, 23, 33, 34]. Additionally, ROTAVAC’s presentation as a 0.5-mL vaccine in a five-dose vial reduces the required space in the cold chain and associated costs [19, 24]. Rotasiil, a lyophilized vaccine, is heat stable for 6 months at 40 °C; the clinical trial for Rotasiil in Niger found the vaccine was efficacious after storage at up to 25 °C in the distribution facility and at the ambient temperature once distributed to the vaccination centers [23, 35]. This demonstrated heat stability could reduce the burden on the cold chain, as well as the financial costs and forecasting challenges associated with vaccine wastage due to temperature excursions [36, 37].

5 Vaccine Effectiveness

A recent systematic review of post-licensure vaccine effectiveness (VE) evaluations in countries using RotaTeq and Rotarix supported the clinical trial findings that suggested differential vaccine performance between low and high child-mortality countries (as ROTAVAC and Rotasiil are recently licensed and not yet introduced on a national scale, they were not included in the literature review). Jonesteller et al. reported median Rotarix VE of 84% in low child-mortality countries, 75% in medium child-mortality countries, and 57% in high child-mortality countries (Table 2) [38]. Comparably, median RotaTeq VE was 90% in low child-mortality countries and 45% in high child-mortality countries [38].

Table 2 Post-licensure vaccine effectiveness in countries with low, medium or high child mortality

Because many of the post-licensure evaluations have not been sufficiently powered to detect differences between subpopulations, the systematic review also aimed to clarify questions of VE with a partially completed vaccination series and against a range of disease severities through summary statistics. For Rotarix in low child-mortality countries, VE was nearly the same for children who received the full number of the manufacturer-recommended doses and children who were only partially vaccinated; the median reported difference was 3% points. In medium and high child-mortality countries, VE was higher in children who received the full series compared with children who received a single dose, with a median difference of 10 and 19% points, respectively. For RotaTeq in low child-mortality countries, VE was higher among fully vaccinated children compared with children who received only one or two doses, with a median difference of 16% points. A single evaluation from a high child-mortality country using RotaTeq showed that VE was 65% higher among fully vaccinated children compared with partially vaccinated children [38].

The review found articles defined severity in two ways: by hospitalization status (i.e., emergency department or inpatient) and by Vesikari score. Vesikari score is a grading of seven clinical signs and symptoms; higher numeric values on the 20-point scale indicate more severe and dehydrating diarrheal illness [39]. In low child-mortality countries, the median difference in VE was slightly higher among hospitalized children compared with those treated in the emergency department for both vaccines, with a median difference of 5% points for Rotarix and 9% points for RotaTeq [38]. While there is some variation, overall these results indicate higher effectiveness against more severe rotavirus disease in a variety of settings. In medium child mortality countries, Rotarix VE was a median 25% lower among children with high Vesikari scores compared with lower scores. In high child-mortality countries, the median difference in VE was 17 and 16% points higher among children with high Vesikari scores than among those with low scores using Rotarix and RotaTeq, respectively [38].

6 Explaining and Improving Performance of Oral Rotavirus Vaccines in Developing Countries

The exact reasons for the gradient in VE between low, medium, and high child-mortality countries are unknown; however, several factors are known to reduce the effectiveness of these vaccines. Other oral vaccines, such as cholera and polio, have also demonstrated lower effectiveness in low-income settings. Co-infections with other enteric pathogens at the time of vaccination inhibits immune response to rotavirus vaccine, as well as oral polio vaccine [40, 41]. Transplacental maternal rotavirus antibodies have been shown to reduce vaccine-induced immunity in young infants [42,43,44]. Some improvement to VE may occur naturally as the global polio program transitions from live oral polio vaccine to injectable formulations. Polio and rotavirus vaccines are typically recommended for concomitant administration and oral polio vaccine has been found to interfere with rotavirus vaccine in the gut and immunogenicity of rotavirus vaccine [45,46,47].

Researchers are exploring several additional hypotheses that might explain differential vaccine performance. Matched case-control studies that included Ghanaian, Pakistani, and Dutch infants suggested compelling correlations between gut microbiota composition and rotavirus seroconversion following vaccination. However, the sample sizes in these initial studies were quite small and a corresponding intervention would need to be proposed and tested [48, 49].

There had been some initial concern that immunity induced by RotaTeq and Rotarix vaccines may not protect against non-vaccine strains, which was particularly relevant in Africa and Asia, since the distribution of rotavirus strains varied compared with countries where rotavirus vaccines were initially developed, tested, and introduced [14]. However, a meta-analysis of studies from Australia, Belgium, Brazil, Bolivia, El Salvador, Mexico, Nicaragua, and the US showed that RotaTeq and Rotarix provide protection against homotypic and heterotypic rotavirus strains [50]. Rotavirus recognizes human histo-blood group antigens (HBGA) and individuals with Lewis-negative secretor-positive HBGA phenotypes have an increased susceptibility to P[8] rotavirus strains, compared with non-secretors [51, 52]. As this phenotype is more common in African populations than US or European populations, researchers have hypothesized that Lewis-negative secretor-positive children may be unable to develop protection using the current rotavirus vaccines, as a reason for poorer performance of Rotarix and RotaTeq in sub-Saharan Africa. However, ROTAVAC uses a G9P[11] strain and its efficacy in India was comparable to Rotarix and RotaTeq in medium and high child-mortality settings, indicating HBGA status is not the only reason for the effectiveness gradient. Further research is needed to determine what role HBGAs may play in vaccine-induced immunity against rotavirus and how this might guide future vaccine development.

Several programmatic strategies have also been evaluated to boost rotavirus VE, including improving nutrition status, recommending alternative ages of administration, timing of breastfeeding, and supplementation with zinc [53,54,55,56,57,58,59,60,61,62,63]. Findings from studies that considered alternate ages of administration have been mixed, with one study demonstrating improved immunogenicity with older administration of Rotarix and two others showing no impact [29, 56, 64, 65]. There was concern that lower VE may be because rotavirus vaccine is neutralized by high maternal antibody in breastmilk in settings with high rotavirus burden. However, studies evaluating withholding breastfeeding at the time of vaccination and rotavirus vaccine response have not found any relationship [57, 58, 60, 62]. A recent randomized control trial in India found a modest improvement in immunogenicity following rotavirus vaccination with zinc and probiotic supplementation together; no effect was found when rotavirus vaccine was administered with zinc or the probiotic alone, however the authors report the dosage of zinc may have been too low [63].

While clinical trials in low child-mortality countries did not identify a difference in vaccine efficacy between children < 12 months of age and children ≥ 12 months of age, there was some evidence suggesting waning vaccine efficacy after first year of life in medium and high child-mortality countries [65,66,67]. Similar to the clinical trial findings, in low child-mortality countries differences between age groups were not found by the systematic review of post-licensure VE studies with either vaccine. However, Rotarix VE was almost a third higher among children < 12 months of age in medium child-mortality countries compared with children ≥ 12 months of age; there was no difference in the summary differences in high child-mortality countries. RotaTeq VE was nearly 25% higher among children < 12 months of age in high child-mortality countries compared with children ≥ 12 months of age [38]. The age-group-specific sample sizes in these evaluations were generally too small to detect a difference; nevertheless, the potential waning has important implications for vaccination programs. A clinical trial in Bangladesh demonstrated a booster dose at 9 months of age improved IgG antibodies against rotavirus; a similarly designed analysis from a clinical trial in Mali is forthcoming [68, 69]. These findings have led to optimism that modifying recommendations to include a booster dose of rotavirus vaccine may address the observed waning immunity with the currently available vaccines. However, further research is needed as the published analysis did not include a clinical endpoint to assess vaccine efficacy following a rotavirus booster dose [68].

7 Vaccine Impact

Despite differential VE, rotavirus vaccines have had a sizeable impact on diarrhea morbidity and mortality in low, medium, and high child-mortality countries. Following rotavirus vaccine introduction, rotavirus mortality decreased by 22% in Brazil and 41% in Mexico among children < 1 years old [70, 71]. Countries have experienced a reduction in rotavirus diarrhea hospitalizations by a median of 67% after including rotavirus vaccine in their national immunization programs [72]. In absolute numbers, this translates to thousands of hospitalizations prevented each year. For instance, in the first 7 years of rotavirus vaccine implementation in the US, vaccination prevented an estimated 382,000 hospitalizations and saved over US$1 billion in direct medical costs [73]. In the 29 African countries that had introduced rotavirus vaccine by the end of 2014, nearly 135,000 rotavirus hospitalizations and 21,000 rotavirus deaths were estimated to have been prevented in 2016 [74].

Universal adoption of routine rotavirus vaccination would also reduce the global burden of rotavirus disease. As an example, in all African countries that had not introduced rotavirus vaccine, including Nigeria and the Democratic Republic of Congo, an additional 119,000 hospitalizations and 22,000 deaths were projected to be prevented with national, routine use of rotavirus vaccines at coverage levels comparable to diphtheria-tetanus-pertussis (DTP) vaccines.

8 Safety

An earlier rotavirus vaccine, RotaShield (Wyeth), a rhesus quadravalent rotavirus vaccine licensed in the US in 1998, was removed from the market in 1999 after post-licensure safety assessments identified a 10-fold [adjusted odds ratio (aOR) 10.6; 95% CI 5.7–19.6] elevated risk of intussusception, a rare bowel obstruction caused by telescoping of the small intestine, following rotavirus vaccination, or approximately one excess intussusception case per 10,000 children vaccinated [75,76,77,78].

Large clinical trials of RotaTeq and Rotarix did not detect an increased incidence of intussusception following vaccination [17, 18, 79]. Since the licensure of Rotarix and RotaTeq in 2006, a slightly increased risk of intussusception has been observed in post-licensure safety monitoring, especially in the first 7 days post-vaccination [80,81,82,83,84,85]. A 2015 meta-analysis of intussusception risk following real-world rotavirus vaccination in Australia, Brazil, England, Mexico, Singapore, and the US found an elevated risk of intussusception in the first 21 days following the first dose of Rotarix (OR 2.4; 95% CI 1.5–3.8) and the second dose (OR 1.8; 95% CI 1.3–2.4), or roughly 1.9 and 1.5 excess cases of intussusception per 100,000 children vaccinated, respectively [86]. An analysis of the risk of intussusception following RotaTeq vaccination in Australia and the US also found a similar level of risk for the first and second doses [87]. No association between intussusception and rotavirus vaccination was found in a retrospective evaluation in South Korea, although the number of children included in the analysis was smaller than other post-licensure studies [88]. As more countries with medium and high child mortality introduce rotavirus vaccine, it will be important to evaluate the risk of intussusception specifically in those settings.

Over the decade since licensure of Rotarix and RotaTeq, the benefits of rotavirus vaccination in decreased hospitalizations and deaths have far outweighed the risks. Initially, WHO recommended upper age limits for vaccination to minimize excess cases of intussusception; evidence from RotaShield suggested the highest incidence of intussusception was among older infants, when the natural incidence of intussusception also increases [75,76,77]. These recommendations were changed in 2009 after models demonstrated 44 rotavirus diarrhea deaths would be prevented for every one potential intussusception death if children were eligible to receive rotavirus vaccination up to 1 year of age [75, 89]. WHO recommends administering rotavirus vaccine to children up to 24 months of age concomitantly with DTP vaccine [75].

Newer rotavirus vaccines have evaluated safety in their clinical trials, however the sample size was insufficient to detect a small increase in risk, like that observed with Rotarix and RotaTeq. The phase III clinical trials for both ROTAVAC and Rotasiil reported 0 cases of intussusception in the first month following any dose of vaccine or placebo [22, 23, 32]. Post-licensure safety monitoring will also continue to be important as new vaccines are adopted more widely.

9 Vaccines in Development

Although the existing rotavirus vaccines have dramatically reduced rotavirus diarrhea worldwide, the licensure of additional rotavirus vaccines may address some of the limitations of the current vaccines as well as reinforce the market. Several additional live oral vaccines are being developed. Lanzhou Institute has completed a phase III trial for a trivalent lamb reassortant rotavirus vaccine (Table 3) [24, 90]. Butantan Institute completed phase I clinical trials for a pentavalent human-bovine vaccine [91]. A hexavalent bovine-human rotavirus vaccine by Wuhan Institute of Biological Products (China) is in the process of completing phase I trials [24]. Additionally, Bharat Biotech is developing a liquid stable formulation of ROTAVAC.

Table 3 Rotavirus vaccine in development

Two alternative types of rotavirus vaccines are in development that may eliminate the VE differential between child mortality settings and the small risk of vaccine-associated intussusception. Murdoch Children’s Research Institute (Australia) and PT Biofarma (Indonesia) are developing a live oral human rotavirus vaccine from a G3P[6] strain found to cause asymptomatic infections in neonates [24]. This strain was selected because it is able to replicate in the newborn gut, even in the presence of maternal antibodies, and the strain may also offer protection to individuals with a Lewis-negative and secretor-positive phenotype. A phase IIa clinical trial demonstrated this vaccine was immunogenic with both infant (6, 10, and 14 weeks) and newborn (birth, 6 weeks, and 10 weeks) administration schedules [92]. Administration of a rotavirus vaccine to newborns offers other benefits. A birth dose could offer protection from diarrheal disease and deaths before any children are exposed to wild rotavirus, potentially preventing additional infections that occur between birth and the current recommended age administration of the first dose, usually at 6 weeks or 2 months of age [2, 93]. Other vaccines that have a recommended birth dose, such as polio, Hepatitis B, and bacilli Calmette-Guerin (BCG) vaccines, have high vaccination coverage of children < 1 year old; birth dose administration of rotavirus vaccine could improve coverage compared with the current infant schedule. Additionally, natural intussusception is rare before 3 months of age [94]; a birth dose may eliminate the vaccine-associated risk of intussusception. This vaccine is undergoing a phase IIb clinical trial [24]. RotaShield has also been considered for a neonatal administration schedule; a clinical trial showed 61% efficacy in Ghanaian infants vaccinated during the first and second months of life [95].

Non-replicating vaccines with a parenteral delivery strategy have been proposed as a way to eliminate the risk of intussusception, since immune response would not be dependent on vaccine viruses replicating in the gut. Additionally, these vaccines may not be susceptible to the hypothesized reasons for lower VE in medium and high child-mortality settings such as co-infections, microbiota composition, or interference by oral polio vaccine [24, 96]. Intramuscular administration has additional benefits; for example, it could be included in a combination vaccine with other routine infant vaccinations like DTP and pneumococcal vaccines (PCV). A subunit vaccine developed by PATH showed safety and immunogenicity in a recently completed phase I/II descending dose study; an additional phase I/II study is currently underway [97]. To date, none of the clinical trials of these vaccines have included a clinical endpoint. Intradermal administration of a whole, inactivated particle vaccine, using a microneedle patch of an inactivated vaccine, is also being developed by CDC. Microneedle patches have several programmatic advantages, including thermal stability, reduced packaging, elimination of sharp biohazardous waste, and ease of administration [98, 99]. Additionally, immune response to intradermal administration may allow for a fractional dose to be administered, which could improve the supply chain and reduce costs [24, 100].

10 Conclusions

Over the last 11 years, rotavirus vaccines have been safely and effectively implemented in more than 80 countries worldwide. Large reductions in rotavirus diarrhea deaths and hospitalizations during this period have been sweeping; however, the live, oral rotavirus vaccines have been shown to be less effective in medium and high child-mortality settings than in countries with low child mortality. New vaccines in development are promising to further reduce the burden of rotavirus diarrhea morbidity and mortality among children < 5 years old.