Rotavirus Vaccine

In January 1999 The American Academy of Pediatrics (AAP) published the 1999 recommended childhood immunization schedule for United States children (1). For the first time, it includes three doses of oral rotavirus vaccine administered to infants at 2, 4, and 6 months of age. This marks a significant step forward in prevention of acute infectious diarrhea in children, coming just over 25 years since the discovery of rotavirus (2). The AAP statement on guidelines for use of rotavirus vaccine, published 1 month earlier, supported its use but pointed out the economic issues that have to be reconciled before its routine use can be implemented (3). That statement provides an excellent summary of current understanding of rotavirus disease and prevention. EPIDEMIOLOGY OF ROTAVIRUS INFECTION Rotaviruses are icosahedral triple-shelled viruses containing double-stranded RNA. They infect most mammals, some birds, and other species. Seven groups of rotaviruses (A through G) have been identified (4). These show no cross-reactivity of group-specific antigens on virus protein 6 (VP6). Only group A, B, and C rotaviruses have been shown to infect humans. Most episodes of human disease are caused by group A. Non–group A rotaviruses occur sporadically in humans, with detection rates of 1 in 300 to as high as 4% in one study from Vellore, South India (5). Group A serotypes 1 through 4 are responsible for more than 90% of episodes in humans (6), and vaccine development is targeted at prevention of illness caused by these strains. Atypical strains have been reported in several settings. A study in India found 65% of cases to be associated with atypical serotypes (7). Reassortment between common and uncommon serotypes or animal strains can occur (8,9). If such atypical strains become more prevalent or dominant, that change could have considerable implications for vaccine development, in which the current philosophy is that serotype-specific protection is required. Virtually all children can be expected to be infected by rotavirus before their fifth birthday, and most have repeated infections (10,11). Many of these episodes are asymptomatic, but often the first one causes symptoms of varying severity. Infants aged 6 to 23 months are most vulnerable, and more than 50% of episodes of acute diarrhea requiring hospital admission in this age group are caused by rotavirus (12). A much lower, yet significant, percentage of acute diarrhea in 0-to 5-month-old infants is associated with rotavirus in developed countries, but early infection is more common in this age group in developing countries (13). Rotavirus vaccine, to be effective, must therefore be administered early in life. Rotavirus diarrhea is a winter disease in temperate climates (Fig. 1), with an as yet unexplained spread each year from the southwest to the northeast of North America, not necessarily involving the same strains of rotavirus (14). This geographical pattern is less obvious in Australia where the majority of the population lives in cities, hundreds of miles apart, with very little habitation between them (15).FIG. 1.: Monthly rotavirus isolations from children with acute diarrhea admitted to hospital in Melbourne, Australia, demonstrating midyear winter peaks (Reprinted with permission from J Clin Microbiol 1998;36:133–8).Estimates of the global burden of rotavirus disease provide surprising information. Rotaviruses are responsible for one third of all hospital admissions for diarrhea worldwide and for approximately 600,000 deaths per year. This represents 20% to 25% of all diarrheal deaths, and 6% of all deaths of children less than 5 years of age. Rotavirus caused only 4% of the estimated 1 billion diarrheal episodes per annum worldwide in the early 1990s, but 40% of the deaths (D. Mahalanabis, personal communication, May 1991). In the United States, approximately 55,000 admissions to hospitals occur each year (16). Surveillance of 1 million children on the west coast of the United States showed that 1 in 57 was admitted to a hospital, and 1 in 21 required an emergency room visit during their first 5 years of life (17). In the United Kingdom, a child's risk of hospital admission for rotavirus by age 5 years was estimated to be approximately 1 in 40 (18), and in Australia 1 in 24 (19). Although morbidity is a prevalent problem worldwide, mortality is still a major issue in developing countries. In Bangladesh, 18% of the 3 million children born each year die before their fifth birthdays, 25% as a result of diarrheal disease (20). They include 14,850 to 27,000 childhood deaths due to rotavirus infection. Prevention cannot be achieved by improved hygiene alone. Occurrence rates are similar in developed and developing countries (6). Breast-feeding may have a preventative effect while it continues, but there is evidence of catch-up infection at weaning (21). ROTAVIRUS DISEASE Rotavirus infection may be asymptomatic or associated with fever, vomiting, and watery diarrhea (22). The illness can be as severe as cholera, resulting in severe dehydration and death. Mature epithelial cells on small intestinal villi are the target (23). They are invaded and shed. The resultant diarrhea is due to a combination of malabsorption and fluid secretion by immature crypt cells, and transport defects, possibly exacerbated by toxic effects of one of the viral proteins, NSP4 (24). Recovery occurs when new, healthy epithelial cells repopulate the surface. Treatment includes rehydration and early introduction of feeding. Breast-feeding should continue throughout the illness when possible. IMMUNITY The correlates of protection remain to be determined, and clinical trial results are the ultimate test of vaccine success. It is known that infection by rotaviruses stimulates protective immunity (25). This was demonstrated by a 3-year surveillance of newborn infants who acquired nosocomial rotavirus infection immediately after birth (26). In that small but intensive study, 77% protection against moderate and severe disease and 100% protection against severe disease were observed. This observation was confirmed in young children in developed and developing countries (11,27,28). Primary infection induced natural immunity which protected against subsequent episodes of severe diarrhea during early childhood. However protection is apparently short-lived and incomplete (29). Oral inoculation is not necessary to induce protective, virus-specific response at the mucosal surface. Parenteral inoculation can achieve this, and a combination of oral and parenteral immunization may prove to be superior but will raise considerable practical difficulties for vaccine administration. The challenge is to develop a vaccine that is better at inducing protection against disease than the natural infection, a task achieved by only one other vaccine:Haemophilus influenzae type B. VACCINE EFFICACY Vaccine development has been comprehensively reviewed (30–39). Orenstein et al. (37) pointed out that the rotavirus health burden is comparable to that for measles, pertussis, mumps, and varicella before vaccines for these diseases were routinely given to children. The time is ripe for use of an effective rotavirus vaccine. The logic behind vaccine development is that gastrointestinal mucosal immunity must be stimulated and that a child's first rotavirus infection should be made asymptomatic. Subsequent community rotavirus infections can be expected to be asymptomatic or to cause only mild symptoms and to boost immunity in preparation for the next exposure. A Jennerian approach has been taken, using live, attenuated rotavirus strains from other species to infect the gut while avoiding causing disease (40). The first trial of a bovine-based vaccine caused great excitement (41,42). Protection was encouraging against predominantly serotype 1 community infection. However, it did not do so well against other serotypes or in developing countries. The current licensed (in the United States) vaccine has a rhesus monkey rotavirus base (3). It has been modified by making three reassortants (Fig. 2), each of which incorporates serotype 1–, 2–, or 4–specific human rotavirus protein into the rhesus rotavirus. Rhesus rotavirus itself is serotype 3. The preparation is thus a mixture of four strains, each directed against one of the four common human serotypes, and is known as rhesus rotavirus tetravalent vaccine (RRV-TV).FIG. 2.: Production of reassortant rotavirus vaccine. Three reassortants, each containing human VP7 of serotype 1, 2, or 4 are mixed with the parent rhesus rotavirus which is itself serotype 3, to constitute tetravalent RRV vaccine. (Reprinted with permission from Arch Virol Suppl 1996;12:163–75.)Several significant trial reports have been published (43–49). The United States Rotavirus Vaccine Efficacy Group reported a trial in 1278 healthy infants aged 5 to 25 weeks, each of whom received three oral doses of RRV serotype 1, RRV-TV, or placebo at 2, 4, and 6 months of age (44). Both vaccines significantly reduced the incidence of rotavirus gastroenteritis. The RRV-TV prevented 49% of rotavirus episodes, 80% of very severe episodes, and 100% of dehydrating rotavirus illness. The same group had reported a previous study of 1006 infants using a 10-fold lower dose (45). Efficacy against very severe gastroenteritis was 73% to 82%, but only the tetravalent vaccine provided protection against disease caused by non–serotype 1 rotaviruses during the second year of follow-up. In Finland, 2398 children were enrolled and received at least one dose of RRV-TV or placebo. Efficacy analysis on those receiving 3 doses of vaccine (n = 1128) or placebo (n = 1145) showed RRV-TV to be highly effective against severe disease (46). Because of disappointing results in some trials in developing countries (47,48), there was much interest in a large trial of RRV-TV in Venezuela (49). The vaccine provided 88% protection against severe diarrhea and reduced hospital admissions by 70%. The relationship between serum rotavirus antibody titers and protection is unclear. The outer capsid proteins VP7 and VP4 have been thought to stimulate protective immunity. However immunoglobulin (Ig)A antibody to VP6 and antibody to the nonstructural protein NSP4 are currently under consideration. In one U. S. study no correlation of immunity against rotavirus infection or disease was identifiable, based on seroconversion to any of the six antibodies measured (50). Similarly, no consistent relationship was found between titers of any of the antibodies and protection. A confounder is the decay of maternal IgG antibody over the first months of life, and conventional definitions of IgG seroconversion therefore do not apply. Elucidation of a correlate of protection would greatly speed and economize the development of second-and later generation vaccines. One interesting report indicates that immunogenicity may be improved by administration of RRV-TV with Lactobacillus casei(51). ADVERSE EVENTS Fever has been the main side effect of RRV-TV. In Finland, 387 (33%) of 1182 of vaccine recipients had a temperature of 38°C or higher that developed during days 3 to 5 after the first dose, compared with 27 (2.3%) of 1194 in the placebo group (52). Fever of 39°C or more occurred in 3.4% and 0.2%, respectively. Only minor differences were seen in irritability, decreased appetite, abdominal cramping, and diarrhea, with no difference in vomiting. Two U. S. studies reported fever in 14% (vaccine) versus 7% (placebo) using a dose of 4 × 104 plaque-forming units (pfu) (45), and 7% versus 4%, respectively, using the higher dose of 4 × 105 pfu (43). Four children in the vaccine group were admitted to hospitals within 7 days of receiving the vaccine, for fever (4), vomiting (3), and diarrhea (2); two of them were excreting vaccine virus. The higher incidence of fever in Finnish children raises some questions about reactogenicity. Viral Shedding In a study of excretion of vaccine virus in 248 U. S. infants, shedding was detected in 50.4% of subjects (53). After dose 1, 76% were G3 (serotype 3) strains, but thereafter, they were mainly G2 strains, suggesting uneven development of immunity to different serotypes. A notable observation in the Venezuelan trial was horizontal transmission of vaccine virus in 15% of vaccine recipients and also in 13% of placebo recipients with rotavirus-positive diarrhea (49). In all cases, vaccine virus was shed in conjunction with a second rotavirus strain, which was shown to be wild-type. Shedding of vaccine virus needs further evaluation with respect to herd protection. BREAST-FEEDING Several studies have examined the influence of breast-feeding on seroresponses to RRV-TV. In a trial using 4 × 104 pfu (low-dose) vaccine, similar seroresponses and protection in breast-fed and non–breast-fed children were induced (54). Rennels (55) reviewed 19 reports of four types of rotavirus vaccine. Immune responses were reduced in a few cases. She concluded that it would be difficult, and perhaps prohibitively expensive, in prelicensure studies to prove definitely that rotavirus vaccinations provide no difference in protection in breast-fed versus non–breast-fed children. COADMINISTRATION OF ORAL POLIO AND OTHER VACCINES Serum antibody responses to oral polio vaccine (OPV) and rotavirus vaccines and efficacy of rotavirus vaccines have been compared among control and vaccine groups (56). Neutralizing antibody titers to poliovirus type 1 tend to increase more steeply in placebo than in RRV-TV recipients, but there are no differences in seroprevalence or in geometric mean titers (GMTs) of antibodies to types 1, 2, or 3 among groups. Concurrently administered OPV resulted in lower IgA GMTs to rotavirus in RRV serotype 1 recipients, but not in RRV-TV recipients. Rotavirus gastroenteritis rates did not differ, but sample sizes were small. In an important study, Markwick et al. (57) coadministered RRV-TV with diphtheria-tetanus-pertussis/ H. influenzae type B and OPV. The percentage of infants who attained protective antibody titers, and the distribution of antibody titers against diphtheria toxoid, tetanus toxoid, and H. influenzae type B were not statistically different between RRV-TV and placebo recipients. Antibody titers against Bordetella pertussis, the percentage of infants with neutralizing antibodies against three serotypes of poliovirus, and clinically meaningful postvaccination reactions did not differ. COST EFFECTIVENESS Tucker et al. (58) undertook a cost-effectiveness analysis to estimate the economic impact of a national rotavirus immunization program for the United States. Cost-effectiveness was analyzed from the perspectives of the health-care system alone and of society as a whole. A decision tree was developed with estimates of disease burden, costs, vaccine coverage, efficacy, and price. Vaccine would be administered to infants at ages 2, 4, and 6 months as part of the routine schedule of childhood immunizations. They estimated that this would prevent 1.08 million cases of diarrhea, avoid 34,000 hospital admissions, 95,000 emergency department visits, and 227,000 physician visits in the first 5 years of life. From the health-care perspective, it would cost $103 per case prevented at a vaccine price of $20 per dose. From a societal perspective, net savings of $296 million would occur. Break-even price per dose is $9 for the health-care system alone and $51 for society as a whole. A rotavirus program need not be totally cost neutral or cost saving. Reduction in misery and ill health is worth something to the community. Pediatric gastroenterologists are in a good position to comment knowledgeably in the inevitable cost benefit–cost effectiveness debate. The release price of the first licensed vaccine at $38 per dose puts the cost far beyond the means of developing countries. Ways to reduce the cost must be found. OTHER VACCINE CANDIDATES This discussion has concentrated on RRV-TV, because it is effective and is now licensed, even if there is some question about fever as a side effect and efficacy in developing country settings. There are many other candidates under consideration (59). They include other animal–human reassortants (60), attenuated human strains (61), and neonatal strains. Advances in molecular biology have stimulated a number of innovative approaches that are elegantly discussed by Estes (62). They include such candidates as virus-like particles produced in insect cells (63,64) and a DNA vaccine encapsulated in microparticles (65). The search continues for elusive epitopes that are common to all rotaviruses and that stimulate protective immunity across serotypes and species. Development of second and third generation vaccines is necessary to increase protection against mild disease, decrease the number of vaccines required, and head off possible changes in community strains arising naturally or as a result of vaccine pressure. CONCLUSION Twenty-five years after the discovery of rotavirus as a human pathogen, the first rotavirus vaccine was licensed. This is cause for celebration. A compelling case can be made for universal use of a successful vaccine in developed countries to prevent morbidity, assuming a reasonable vaccine price. In developing countries, the high mortality rate and the enormous morbidity burden demand early introduction of a vaccine, once it has been proved to be effective in that setting. The side effect of fever needs further attention, but it must be remembered that this first-generation vaccine prevents a severe disease. Paediatric gastroenterologists should now actively lobby to ensure implementation of this effective strategy to prevent one of the most common, serious gastrointestinal diseases affecting children worldwide. Postscript: On July 14, 1999, the American Academy of Pediatrics recommended temporary suspension of administration of RRV-TV vaccine. Intussusception occurring after vaccine administration had been reported. The Centers for Disease Control and Prevention was asked to determine whether the association was likely to be one of cause and effect. If so, then other live virus vaccine candidates may also be under suspicion. Acknowledgments: The author thanks Prof. Ruth Bishop, AO, and Dr. Roger Glass for advice and help with this review.