Supported by an unrestricted educational grant from Wyeth-Ayerst Laboratories
Volume II, Issue 4 - September 1997
Supported by an unrestricted educational grant from Wyeth-Ayerst Laboratories
Rotaviruses are the most common cause of severe vomiting and diarrhea in children worldwide and infect virtually all children by 3 to 4 years of age. Each year among children younger than 5 years of age, about 165 deaths in the United States, and >750,000 deaths worldwide are attributable to rotavirus. Table 1 shows outcomes and costs in the US. This burden is an underestimate because rotavirus also affects adults and is more severe in the elderly. For example, one-third to one-half of parents will be ill during the same week their child is hospitalized for rotavirus diarrhea. Outbreaks of rotavirus infection also occur in hospitals, child care centers, nursing homes, and summer camps.
In general, rotaviruses are detected more often as the severity of illness increases. They account for about 10% of diarrhea episodes that cause a child to receive medical treatment but are associated with about 40% of diarrhea episodes in young children who require hospitalization. In the US and countries with similar climate up to 90% of diarrhea episodes requiring hospitalization during the peak of the winter diarrhea season can be attributed to rotavirus. Many children are infected more than once. First infections with rotavirus usually are the most severe; however, some severe second infections occur. The protective immunity acquired from infection is incomplete and the efficacy of that immunity varies according to the severity of symptoms (Table 2). One rotavirus infection induces immunity sufficient to produce 38% efficacy against subsequent mild infection but 87% efficacy against subsequent moderate-to-severe illness. Two natural rotavirus infections are required to induce immunity strong enough to yield complete protection against subsequent moderate-to-severe infection.
The principal symptoms of illness are diarrhea, vomiting, and fever. Hospitalization is usually needed when dehydration and metabolic acidosis are profound. The illness places a significant stress on the individual's ability to handle fluid and electrolytes. Five to 10 diarrhea episodes and 3 to 5 vomiting episodes per day are common in the first few days of severe illness. Because an infant's fluid needs are around 100 mL/kg/day and each diarrhea or vomiting episode results in a 5 to 10 mL/kg loss of fluid, the infant may need 50% to 125% more fluid each day than usual to maintain metabolic balance.
Rotaviruses are RNA viruses that replicate in the intestine. Rotaviruses share a number of features with influenza viruses, making influenza a useful model for thinking about rotaviruses (Table 3). Rotaviruses have a segmented genome and carry 2 surface proteins that are neutralization antigens and have separate antigenic specificities, called P types (for the protease-sensitive VP4 protein) and G types (for the glycoprotein VP7). Rotaviruses have a wide spectrum of types, groups, and subtypes. Multiple types and subtypes circulate in communities at the same time. Rotaviruses change their surface proteins over time and, in addition, circulate in animals. The animal reservoir is probably important for development of strains new to humans. Rotavirus illness has an annual winter seasonality in temperate regions similar to that for influenza. All of these similarities to influenza have made vaccine development complex.
Rotavirus gene segments can reassort during mixed infections with different strains and lead to infectious viruses with a new genetic composition. For example, when strain A is co-infected with strain B, progeny contain 1 or more segments from strain A and the remaining segments of strain B, to yield 11 segments representing all the protein functions required for an infectious particle. The ability of rotaviruses to reassort has been utilized to produce vaccine candidates with the antigenic properties of human rotaviruses and growth characteristics of animal rotaviruses.
Vaccine development began in the 1970s, shortly after the virus was discovered. Initial vaccine candidates were monovalent animal strains shown to infect children and anticipated to be naturally attenuated. Several monovalent vaccine candidates reached phase 3 trials (Table 4). These vaccines had demonstrable efficacy but at a level considered unacceptable for licensure. The monovalent vaccine candidates were tested under different administration schedules, including different numbers of doses, different ages of first vaccination, and different virus quantities per dose. The results of these trials were considered to be unsatisfactory, especially when it was recognized that rotavirus strains that infect humans include many different serotypes and protection against first infection tends to be serotype-specific. The cumulative results of these vaccine trials led to the conclusion that several doses of vaccine would be required and that evaluation of monovalent vaccine candidates should be discontinued.
Current vaccine candidates are multivalent. The most widely tested of these candidates is a rhesus-rotavirus based tetravalent vaccine (Wyeth-Lederle Vaccines and Pediatrics) that includes types G1 to G4, which account for >90% of infections in the US. This vaccine has been tested in over 10,000 children in the US, Finland, and Venezuela, with efficacy rates of about 80% against severe illness and about 50% against any rotavirus-associated diarrhea. Levels of protection are comparable to those observed after a first rotavirus infection and considerably higher than observed in the monovalent vaccine candidate trials (Table 5).
Whatever the vaccine candidate, the greatest protection provided by the vaccine has been against the most severe illness. This is fortunate, because such severe episodes are those most likely to result in death, longer hospital stays, and higher cost of illness.
The safety profile of these vaccines is excellent. The only reported adverse effect of vaccination for the widely tested rhesus-rotavirus-based tetravalent vaccine is an approximately 4% increase in the rate of low-grade fever (<39°C) 3 to 4 days after vaccination, compared with a placebo. The candidate rotavirus vaccines have been developed as oral preparations, to be given at 2, 4, and 6 months of age to infants. The tetravalent rhesus-rotavirus vaccine can be given safely at the same time as the oral poliovirus vaccine.
Cost and benefit studies indicate that if the vaccine has an overall efficacy of 50% and a cost of $30 per dose based upon a 3-dose series, the immunization program would cost $243 million per year but yield savings of $79 million in healthcare costs (direct costs) and $387 million in society costs (indirect costs) (Table 6). The cost effectiveness was a societal savings of $459 per case and $78 per case in healthcare costs. Three doses of the candidate vaccines would prevent more than one million cases of rotavirus disease, 58,000 hospitalizations, and 82 deaths per year. These are probably underestimates because many of the assumptions in the analyses were for worst cases, the estimates do not reflect the likelihood that mass immunization will decrease overall virus circulation and exposure, and disease burden in the elderly is not included. On the other hand, programs to reduce hospitalization, including greater use of oral rehydration, appear to be reducing numbers of hospitalizations for rotavirus disease.
Informal surveys suggest there are a wide variety of perceptions among practitioners regarding the need for a rotavirus vaccine. Some clinicians report that rotavirus diarrhea "is not a problem in my practice" while others quickly recognize when rotavirus season begins because 10 to 20 children per day with diarrhea seek care in their office. These differences in perception may be related in part to the fact that the virus has been recognized only since 1973 and studies of national disease impact in the US are <10 years old. Systematic studies of physician perception of rotavirus disease importance, perceived need for and acceptability of a rotavirus vaccine, and physician and parent perceptions about the novel formulation of a genetic reassorting virus as a vaccine have not been published. Since the vaccine was developed for oral administration during the same visits when other vaccines are given should aid in vaccine acceptance.
Oral rehydration is an effective treatment that, when given promptly, can reduce rotavirus disease morbidity and need for hospitalization. However, studies of reasons for hospitalization and death indicate delayed treatment and underlying host susceptibility, particularly prematurity, are the principal risk factors for poor outcome of infection. Therefore, although oral rehydration is inexpensive and easy to administer, those most likely to experience severe morbidity are those least likely to receive early treatment. The vaccine offers an alternative strategy by preventing disease.
A 3-dose vaccination series will be complete at about 6 months of age. In some regions, virus exposure occurs at an early age. In Venezuela more than half of children experience their first rotavirus infection by 6 months of age. Studies indicate that like natural infection, vaccine-induced immunity is cumulative with each dose. Accordingly, some children will be ill before full vaccine-induced immunity can be achieved. In addition, because vaccine-induced immunity is not complete, some children will experience severe rotavirus infections despite full immunization. One behavior that reduces the risk of early rotavirus infection and appears to have no effect on vaccine-induced immunity is breast feeding. The efficacy of breast feeding against rotavirus infection is comparable to that induced by a full series of the rhesus-based tetravalent vaccine. Promotion of breast feeding would be an especially suitable complement to a national vaccine program with the current vaccine candidates.
A bovine strain of reassorting, multivalent vaccine in development will likely compete with the product referred to at the beginning of this article. Alternative vaccine approaches that may achieve higher rates of protection are intramuscular injection of inactivated rotavirus virions or subunits at birth; incorporation of rotavirus proteins into an expression vector, such as salmonella; or incorporation of a human P type into a reassorting vaccine.
David O. Matson, M.D., Ph.D.
Center for Pediatric Research
Children's Hospital of The King's Daughters
and Eastern Virginia Medical School
Norfolk, Virginia.
Dr. David O. Matson has grants from Wyeth-Lederle Vaccines and Pediatrics and Merck Research.
Supported by US Public Health Service grant HD13021.
