Rotavirus Vaccines: a story of success with challenges ahead

Rotavirus disease before the vaccine era

Rotavirus was first visualized in 1973 when electron microscopy was used to examine stools from ill children; subsequently, several decades of epidemiological research concluded that rotavirus was the main cause of acute diarrheal disease in children younger than 5 years of age worldwide^1–3. Community- and hospital-based studies performed before 2006, the year current vaccines were licensed, proved that rotavirus was a key player in childhood gastroenteritis, accounting annually for 25 million outpatient medical visits, 2 million hospitalizations, and 400,000 deaths⁴. At the time, best estimates indicated that every child would be infected by 5 years of age, one in 5 would require a medical visit, one in 65 would be hospitalized, and one in 293 would die, mostly children living in poor regions of the world.

Rotavirus vaccines

Rotavirus burden 10 years after rotavirus vaccine licensing

It is estimated that in 2015 there were nearly 2.4 billion episodes of acute diarrhea, of which nearly 950 million occurred in children younger than 5 years of age. In the same year, diarrheal diseases were responsible for nearly 1.31 million deaths, of which nearly 500,000 occurred in children under 5 years of age. Rotavirus was estimated to cause nearly 147,000 deaths. Between 2005 and 2015, the number of diarrhea cases in children under 5 years of age decreased by about 10%, and deaths due to diarrhea decreased by around 34%, while rotavirus deaths decreased by 44% (95% CI: 33–52%)⁴¹. Attribution of this reduction in rotavirus cases and deaths to vaccine use is difficult, especially because vaccines are not widely used in the countries with the highest disease burden; nevertheless, rotavirus vaccines have most likely played an important role. As of January 2017, 92 countries have introduced rotavirus vaccines. This includes 85 national introductions, two ongoing phased introductions, and five subnational introductions (http://rotacouncil.org/vaccine-introduction/global-introduction-status/). A dramatic decrease (>80%) in the incidence of severe rotavirus diarrhea has been reported in high-income countries, and a decrease of about 50% has been reported in low-income settings³³. Increasingly, evidence shows reductions in diarrhea-associated deaths of 31% in infants younger than 1 year old and 42% in children younger than 5 years old in countries with low child mortality⁴². Specific data from low-income countries, where childhood mortality and specifically diarrhea-associated mortality is highest, are scarce. In South African children receiving two vaccine doses, Groome and colleagues showed 57% (95% CI: 40–68) effectiveness for rotavirus diarrhea requiring at least overnight hospital admission in children younger than 2 years of age⁴³. In Malawi children receiving two doses in an accelerated schedule, Naor Bar-Zeev and colleagues showed 64% (95% CI: 24–83) effectiveness for reduction of rotavirus-positive emergency room visits (compared with rotavirus test-negative controls) in children younger than 5 years old (94% of samples tested from children younger than 2 years of age)^44,45. Furthermore, and most importantly, rotavirus vaccines have not been implemented in countries with the highest rotavirus-associated disease burden. Nearly 40% of sub-Saharan Africa and almost all South-East Asia (with the exception of India and Pakistan, where vaccines are being introduced in a phased format) have not yet introduced rotavirus vaccines.

Challenges ahead

The most imperative current and future challenge is to fulfill the WHO’s recommendation to introduce rotavirus vaccines in all countries, with no exceptions. Funding and support priorities should and are being placed in countries with the highest mortality rates⁴⁶. Unfortunately, in many instances, a lack of funding is not the only limitation to implementation in these countries; there is also a lack of political will and/or recognition of the potential benefits of vaccination. In too many middle-income countries, where diarrhea-associated deaths are uncommon but where rotavirus-associated medical and emergency room visits and hospitalizations are significant, health authorities frequently fail to recognize the need for rotavirus vaccines. Increasing the importance of technical advisory groups, including the use of efficient decision-making tools (such as the Grading of Recommendations, Assessment, Development and Evaluations [GRADE]^47,48), would help balance the viewpoints of health and financial authorities.

Live attenuated rotavirus vaccines have worked well, but we cannot consider current protection levels to be optimal. Unfortunately, efficacy and effectiveness decrease inversely with poverty. The reasons behind this phenomenon are not clear, but there is indirect evidence supporting several possibilities. Firstly, rotavirus infections occurred at younger ages and repeated symptomatic infections were more common in a birth cohort from a very poor area in India compared to a less-deprived area in Mexico^5,49. This observation strongly suggests that viral exposure is significantly higher in poorer regions, most likely due to increased exposure to human feces. Vaccination at earlier ages, more doses, and/or higher dose concentrations may benefit such populations, but a significant increase in protection, from 50–60% to 80–90%, seems unlikely unless there is a concurrent improvement in environmental sanitation. Secondly, the co-administration of oral polio, more commonly used in developing countries, and rotavirus vaccines reduces the immune response to the latter, which most likely has some impact on reduced vaccine efficacy/effectiveness⁵⁰. A third factor may be increased incidence of breastfeeding in low-income areas. While breastfeeding did not significantly reduce the immunogenicity of rotavirus vaccine in Finish children, it was associated with a mild decrease in protection, especially during the second year of life⁵¹. Mexican researchers demonstrated that breastfeeding can interfere with vaccine shedding and immune response⁵². Thus, in poorer regions where breastfeeding tends to be more common, concomitant breastfeeding could play a partial role in reducing the efficacy/effectiveness of the vaccine by neutralizing vaccine replication to some extent. Three field studies addressing this issue provide good evidence that the role of breastmilk in reducing vaccine immune responses is non-existent or minimal. In a well-designed trial in Pakistan, rotavirus IgA seroconversion rates and geometric mean titers were evaluated in vaccinated children randomly assigned to one of two groups, a group where breastfeeding was withheld for at least 1 hour before and after vaccination and a group where infants received at least 20 minutes of breastmilk at most 10 minutes before vaccination. Rather surprisingly, IgA seroconversion rates were roughly 10% higher in the group receiving breastmilk after both the first and the second dose. Hints of possible interference were observed in the antibody titers achieved, which were lower among seroresponders receiving breastmilk after the first dose compared to those not receiving milk; a subset of infants with low maternally derived antibodies receiving high-antibody-containing breastmilk seemed to have a reduced immune response⁵³. A similar study in Indian children in whom breastmilk was withheld for 30 minutes before and after vaccination, or encouraged, showed no differences in seroconversion rates; in this study, seroconversion rates where quite low in both groups (26–27%)⁵⁴. In a South African study with seroconversion rates near 60%, withholding breastmilk for 1 hour showed no difference compared to a breastmilk-encouraged group in anti-rotavirus IgA seroresponse rates or antibody titers achieved⁵⁵. Other factors are possible, such as the increased prevalence of severe malnutrition leading to reduced vaccine immune responses; however, this hypothesis is supported by only one relatively underpowered study⁵⁶.

Environmental enteropathy is a subclinical condition characterized by small intestine inflammation with shortened villi, intestinal barrier dysfunction, and reduced nutrient absorption. This condition seems to be common in children living in poor, unsanitary conditions and is thought to be caused by repeated or chronic exposure to enteropathogens and by malnutrition⁵⁷. Studies using biomarkers associated with this condition show that it is present in over 80% of 12-week-old infants in Bangladesh⁵⁸. Using a complex model, this condition was reported to be associated with decreased seroresponse and failure of the vaccine Rotarix®. In children from El Salvador, this condition was also associated with lower seroresponse rates to RotaTeq®⁵⁹.

Lastly, differences in the gut microbiota/microbiome have been proposed as a factor affecting vaccine effectiveness; while regional differences exist⁶⁰, the potential that the gut microbiome plays any role in differential protection rates will require future studies, which are currently underway⁶¹. A recent publication is enlightening, as it shows that children from Ghana responding to rotavirus vaccination as determined by an anti-rotavirus IgA titer >20 IU/mL have a different microbiome profile compared to non-responders. Interestingly, responders had a microbiome profile more similar to a Dutch infant group compared to non-responders⁶².

Serotype replacement leading to an increase in non-vaccine serotypes over time, due to possible vaccine selective pressure, has been repeatedly postulated during the past decade, but it lacks robust supporting evidence. Increases in the predominance of specific rotavirus serotypes, heterotypic to the vaccines in use, have occurred, but mostly in a temporal manner similar to the known unpredictable regional and temporal variability of rotavirus serotype circulation observed before the vaccine era^7,63. The biological plausibility of a selective pressure phenomenon is low, as rotavirus vaccines do not abolish the circulation of any particular serotype (protection against infection does not surpass 60%) and “competition” between different serotypes within the intestine (as occurs for pneumococcus in the nasopharynx) is unlikely. Low-level emergence of “uncommon serotypes” such as G5, G8, G12⁶⁴, and the more significant G9 serotype, which emerged as a new, frequently predominant serotype during the past few decades, are ongoing phenomena which will most likely continue to occur with or without widespread vaccination. Because different vaccines are most likely not equally protective against all potential human serotypes, and because some novel serotypes may be more fit for the human intestine, it is possible that one or more of these less common types could prevail over others (relative predominance) for a given time period. Importantly, 10 years after vaccine introduction, the emergence of uncommon strains has been mild. A well-performed systematic literature review and meta-analysis including publications with content on rotavirus vaccine effectiveness and strain characterization published from January 2006–2014 concluded that vaccines have similar effectiveness against partly or fully heterotypic strains compared with homotypic strains. It also concluded that the emergence of particular serotypes has not occurred after vaccination⁶⁵. Serotype surveillance will continue to be important, especially in low-income countries, where vaccines are less efficacious and data on rotavirus serotype distribution are scarce; importantly, the interpretation of serotype variations should be carried out with caution.

Further considerations of parenteral protein-based rotavirus vaccine candidates, if proven safe and efficacious, may be of benefit for several reasons. First, we could move away from any risk of intussusception, as it is unlikely that a parenteral vaccine would be a trigger. Second, theoretically any external interference with vaccine take (maternal antibodies, breastmilk, and live poliovirus) would be unlikely, with the potential for increased efficacy, although this is highly speculative at the moment. Third, a combination vaccine with another major cause of diarrhea, such as norovirus, could be considered; one current norovirus VLP and rotavirus VP6 nanostructure-based vaccine demonstrated an interesting adjuvant effect of the rotavirus component on norovirus immune response⁶⁶, although it is unclear if this would provide protection against rotavirus. Combination with rotavirus outer capsid proteins³⁹ may be a future avenue to explore.

Forty-five years after rotavirus discovery, extensive research efforts have led to safe and effective vaccines, which are reducing childhood deaths and suffering. It has been a success story, which is not over. Several pending issues have been discussed here, and the next decade should bring new insights, advances, and answers and, most importantly, significantly more children receiving rotavirus vaccines.