T. Vesikari Vaccine Research Center, University of Tampere Medical School, Tampere, Finland
Two live oral rotavirus (RV) vaccines, a human attenuated G1 RV strain and a bovine-human reassortant five component combination vaccine, were licensed in 2006, and in the next several years universal RV vaccination of all healthy infants was introduced in most Latin American countries, the United States, Australia, South Africa, and several European countries. More recently RV vaccination with GAVI support has been integrated into EPI programs of many African and other developing countries. In addition, locally produced live oral lamb RV vaccine is being used in China, and human RV strain vaccines are being launched in India and Vietnam. Altogether, universal RV vaccination has been introduced in about 80 countries, and the global experience has shown that RV gastroenteritis is a vaccine preventable disease. Live oral RV vaccines prevent effectively severe RV disease but have little effect on RV infection as such. In Finland, representing optimal conditions, the real life effectiveness against RV associated hospitalizations has been 95%, in Latin America around 80% and in Africa between 50% and 70%. The mechanism of protective immunity is not fully understood, but it is likely that much of the protection against severe RV disease is not serotype-specific but mediated by immunity against the inner core major protein and group antigen VP6. The lower effectiveness in developing countries is due to several reasons not easily remedied. All live oral RV vaccines appear to be associated with a (low) risk of intussusception (IS). Risk of IS and less than optimal effectiveness in developing countries are the main drivers towards development of nonlive candidate RV vaccines.
RotaShield, RotaTeq, Rotavac, Rotavin-M1
The discovery of RVs in 1973 by Bishop et al. (1973) and Flewett et al. (1973) was a major breakthrough because up to that point the etiology of acute gastroenteritis (GE) in young children was largely unknown. After this discovery it soon became apparent that RVs caused approximately one half of the cases of acute GE in industrialized countries and were also the most common causative agents of childhood diarrhea in developing countries (Steinhoff, 1980; Maki, 1981).
A recent estimate of the mortality attributable to RV disease puts the number at 453,000 deaths in 2008 in children less than 5 years of age (Tate et al., 2012). This may be an overestimate since total diarrheal disease mortality in children was estimated to be less than 0.6 million in 2012 (Liu et al., 2015), and RV disease-associated mortality of young children was at around 200,000 deaths in 2011 (Walker et al., 2013). Still, mortality from RVGE has declined less than the total mortality from diarrheal disease and other measures of diarrheal disease control, such as oral rehydration therapy, improvement of hygiene and sanitation, and promotion of breastfeeding have had proportionally less effect on RVGE associated deaths than mortality from other diarrhea. This emphasizes the global need for RV vaccination, and RVGE ranks high among vaccine-preventable diseases.
The incidence of RVGE in developed countries is not much different from that in developing countries, but the outcome is. In Europe, deaths from RVGE are rare and the main argument for RV vaccination is prevention of hospitalization (Soriano-Gabarro et al., 2006). Even severe RVGE can be successfully managed in appropriate facilities that are available in developed countries. Nevertheless, the episodes of severe RVGE cause distress to the infants and anxiety for the families, not to mention the costs of hospitalization. Thus, there are compelling medical and financial arguments for universal RV vaccination to be established also in developed countries.
2. Rotavirus vaccines
All past and current RV vaccines are live viruses given orally to multiply in the intestines and induce immunity mimicking that after natural RV infection. The vaccines can be divided into three categories: heterologous animal RVs, animal-human reassortant RV strains, and attenuated human RV strains (Table 2.11.1).
Live Oral Rotavirus Vaccines Licensed or at Late Stage Development
Status in development
Animal RV strains
Lamb strain G10P
Lanzhou Institute of Biologica products (China)
Licensed and in use in China
Animal–human reassortant strains
Tetravalent human–rhesus reassortant
G1–G4 + P7
Wyeth (United States)
Withdrawn 1999, Phase II trials of neonatal administration in Ghana
Murdoch Childrens Research Institute (Australia) and Biofarma (Indonesia)
Phase III trial
2.1. Heterologous (Animal) Rotavirus Vaccines
2.1.1. RIT4237 (NCDV) Bovine RV Vaccine
Bovine RV strain RIT4237 (G6P) was the first RV vaccine tested in humans. The vaccine is highly attenuated with a history of 154 cell culture passages, much more than any other RV vaccine strain. The high level of attenuation was meant to minimize risk of RV disease in calves in case of transmission (and for possible use as veterinary vaccine) and was not for attenuation in humans as bovine RV would not cause any symptomatic disease in humans anyway. However, the adaptation to cell culture resulted in growth to a high titer over 108 TCID50 /mL (Delem et al., 1984), which may explain the good efficacy of this vaccine in the early trials in Finland.
In its first efficacy trials in 1983, a single dose of RIT4237 vaccine induced 50% protection against any and 88% protection against severe (clinically significant) RV disease (Vesikari et al., 1984a), a result, which remains a benchmark for other vaccines and trials. Two doses of the same did not improve efficacy (Vesikari et al., 1985a). In Peru the efficacy was 40% against any and 75% against severe RVGE (Lanata et al., 1989) and in The Gambia the efficacy was 33% but severity of RVGE was not measured (Hanlon et al., 1987). Thus, a gradient of efficacy between developed and developing countries was observed, which has later been confirmed to be true for other RV vaccines as well.
The higher efficacy of RV vaccine against severe RV disease was also documented using a 20-point score, later often referred to as “Vesikari-score” (Ruuska and Vesikari, 1990). Figure 2.11.1 shows the efficacy of neonatal vaccination with bovine RV vaccine on RV disease of varying severity (Ruuska, 1991). The same score has later been applied to measure efficacy of the other RV vaccines.
The series of studies of RIT4237 bovine RV vaccine also established the following general principles that were largely confirmed later for other RV vaccines: (1) a higher titer of oral inoculum resulted in greater uptake (Vesikari et al., 1985b), (2) RV vaccine virus was sensitive to gastric acidity, and buffering such as milk feeding was needed for successful vaccine uptake (Vesikari et al., 1984b), (3) breast milk or breast feeding did not suppress the uptake to any significant degree, especially if the titer of vaccine was high (Vesikari et al., 1985b, 1986a). Earlier studies also showed that concurrent administration of live oral polio vaccine (OPV) suppressed the uptake of bovine RV vaccine, whereas RV vaccine had only minimal inhibitory effect (interference) on OPV (Vodopija et al., 1986; Giammanco et al., 1988).
The vaccine was withdrawn from further development for a combination of reasons: (1) the level of protection in developing countries was perceived as insufficient. (2) The concept of protection against severe disease only was new, and this end point was regarded as inadequate. (3) There was concern by the WHO that oral RV vaccine might compromise the success of the poliovirus eradication program, which was considered to be of the highest priority. The same issues remain relevant today for the current licensed vaccines, but are no longer regarded as obstacles for introduction of RV vaccination.
2.1.2. WC-3 Bovine RV Vaccine
WC-3 stands for Wistar Calf according to Wistar Institute where the vaccine was developed (Clark et al., 1986). This bovine G6P RV strain is less adapted to cell culture (about 20 passages) and therefore grows only to a titer of approximately 107 PFU/mL. In the main efficacy trial in the United States the WC3 vaccine yielded a mere 20% protection against RVGE, and was not developed further (Bernstein et al., 1990). However, the strain is important because it is the backbone for bovine–human reassortants that are included in the current “pentavalent” (RV5) bovine–human reassortant RV vaccine (see later).
2.1.3. Rhesus RV (RRV) Vaccine
Rhesus (monkey) RV strain (G3P) was grown 9 times in monkey kidney and 7 times in fetal rhesus lung cells and adopted as human vaccine (Kapikian et al., 1986). Unlike bovine RV vaccine, RRV is clearly virulent in humans by causing fever, though not diarrhea (Vesikari et al., 1986b). At a high titer level of 105 PFU/dose the RRV vaccine was efficacious but reactogenic (Flores et al., 1987; Gothefors et al., 1989). Importantly, however, RRV served as a backbone for the development of the rhesus–human reassortant vaccines (Midthun et al., 1985), of which the tetravalent composition (RRV-TV) became the first licensed RV vaccine in 1998 (see later).
2.1.4. Lamb RV Vaccine
Lamb rotavirus strain LLR-37 (G10P) was developed in Lanzhou Institute for Vaccines, China, and is a licensed RV vaccine in China (Li et al., 2015). No formal efficacy trial nor any head-to-head comparison of LLR-37 with other RV vaccines has been done, but there is post marketing evidence for efficacy (Fu et al., 2012). The vaccine is recommended for two doses at ages 2 months–2 years, and widely available in the private market in China. LLR has also served as a backbone for the development of a trivalent lamb-human reassortant vaccine, which is undergoing clinical trials in China.
2.2. Animal–Human Rotavirus Reassortant Vaccines
2.2.1. Rhesus–Human Reassortant Vaccine
Rhesus–human reassortant tetravalent (RRV-TV) vaccine (RotaShield, Wyeth) is the prototype of reassortant vaccines and contains four viruses expressing human G1, G2 and G4 VP7 antigens on the rhesus G3P RV genetic backbone plus the RRV itself. RRV-TV retained the high reactogenicity (for fever) of RRV and accordingly, a relatively low dose of the vaccine (4 × 104 PFU) was used in the licensed formulation. RotaShield vaccine was in use in the United States in 1998–99, but was withdrawn because of association with intussusception (IS) (Murphy et al., 2001; Simonsen et al., 2005).
RRV-TV is a more efficacious vaccine than the parent rhesus RV vaccine (Rennels et al., 1996). In the pivotal efficacy trials before licensure of RotaShield the RRV-TV vaccine showed an efficacy of 100% against any hospitalization for RVGE and against any RVGE 68% in Finland (Joensuu et al., 1997) and 75 and 48%, respectively, in Venezuela (Pérez-Schael et al., 1997). These results were instrumental for wider acceptance of protection against severe RVGE as the primary end point of RV vaccine efficacy and for acknowledgment of the somewhat lower vaccine efficacy in developing countries as being satisfactory, thus considering the introduction of RV vaccination as worthwhile.
RotaShield vaccine was withdrawn in 1999 in the United States because of association with IS (see earlier), before it was launched in Europe or tested in developing countries. Recently, RRV-TV, given to neonates (to minimize risk of IS (Vesikari et al. 2006d)), was tested in Ghana with promising results, and the possibility of reintroduction in developing countries remains viable. The vaccine efficacy was 63% for two doses (Armah et al., 2013). Whether transplacentally transmitted maternal RV-specific antibodies might have reduced the efficacy can only be speculated.
2.2.2. Bovine–Human Reassortant Vaccine
The “pentavalent” bovine–human reassortant RV vaccine (RotaTeq®, Merck, also termed RV5) is a combination of four G-type reassortants (for G1–G4) and one P-type (P) reassortant on the WC-3 bovine RV genetic backbone (Clark et al., 1996). Since the WC-3 is a G6P virus, these bovine G and P types are also present in the vaccine. The terms pentavalent or RV5 refer to the five mono-reassortants present in the vaccine. However, the term is not accurate as there are more RV types represented in the vaccine and more importantly because it is now well established that the protection induced by the vaccine is not limited to the G or P types contained in the product (see later).
The dose of the RotaTeq vaccine is approximately 106 PFU and was determined in a small scale dose-finding efficacy trial, which showed that a dose one log higher did not much improve efficacy (and would be more expensive to produce) whereas a dose one log lower was clearly less efficacious (Vesikari et al., 2006a). The vaccine is given in three doses. This was determined early on to accommodate the US childhood immunization program, but was also based on the demonstration of incremental immunogenicity up to the third dose (Clark et al., 2003).
The efficacy and safety of RotaTeq vaccine were established in a large (70,000 infants) Rotavirus Efficacy and Safety Trial (REST). The overall efficacy against health care utilization (combined end point of hospital admission and outpatient clinic treatment) was about 95% with a narrow confidence interval (Vesikari et al., 2006b). An extension study of REST in Finland of 21,000 children confirmed that RotaTeq was not only efficacious against G1, G3, and G4, all P, but also against G9P and the fully heterologous G2 P, which are not among the G-types in the vaccine. The level of efficacy against various RV G-types was not significantly different (Vesikari et al., 2010). RotaTeq was licensed in 2006 and is now one of the two major RV vaccines used globally (the other one being Rotarix®, see later)
The G1 and P reassortants included in RotaTeq vaccine may re-reassort with each other and form vaccine-derived (vd) double reassortants on the bovine RV VP6 core (Donato et al., 2012; Hemming and Vesikari, 2012). vdG1P viruses may be more virulent than the original single reassortant vaccine viruses and they may be responsible for the low rate (about 1%) of diarrhea seen after vaccination, and may also be capable of transmission (Payne et al., 2010). If transmitted to immunocompromised subjects, vaccine viruses may cause prolonged infection (Patel et al., 2010). Shedding of live infectious virus after the first dose of vaccination was found in about 9% in the REST study, but using RT-PCR shedding may be found in about 50% of the recipients, with G1 reassortants being the most common ones (Markkula et al., 2015).