Biological basis of vaccinology
Protection against rotavirus infection has been associated with the presence of antirotavirus IgA antibodies in the gastrointestinal mucosal surface.54
Although the IgA response is often used to measure vaccine immune response, levels of serum rotavirus IgA antibody do not always correlate with levels of IgA antibody in the gut.13
Presence of virus-specific IgA in feces or serum was not predictive of protection against disease in studies of the simian – human reassortant vaccine. Therefore, it remains difficult to identify the best immune correlate of protection from a rotavirus vaccine.
It is also unclear how many serotypes a rotavirus vaccine should contain. Although initial infection does not confer complete immunity to rotavirus, clinical studies show that primary infection does seem to protect against severe disease upon reinfection. For example, one study in West Africa showed that primary infection conferred 70% (95% CI: 29–87) protection against subsequent rotavirus diarrhea.55
Similarly, although neonatal infection does not confer complete protection against disease in future, children who had rotavirus infection as neonates appear to have less-severe disease later in childhood. This protection does not appear to be strain specific, that is, children who were infected by one rotavirus strain as neonates had less-severe diarrhea later in childhood even when infected with a different strain.56
However, protection is stronger when a child is exposed to a G-type with which they have been previously infected.57
When measuring immunologic response to natural rotavirus infection, it appears that the antibody response is higher against the infecting G-type than against other G-types.58
The benefits of a multiserotype vaccine vs a single-serotype approach to vaccine development have not been established.
Several studies have evaluated animal strain vaccine candidates, with inconsistent results. The RIT4237 bovine strain was isolated from a calf and attenuated in cell culture, and found to be immunogenic by serum immunoglobulin response in human infants.59
Although initial efficacy trials in Finland appeared positive, subsequent studies showed little or no protection against rotavirus disease.60
The bovine strain WC3 was isolated from a calf in Pennsylvania in 1981. This strain alone as a vaccine was not consistently protective in all efficacy trials, with protection ranging from <10% to 76.1%.63
A third strain was the rhesus rotavirus vaccine (RRV), which was isolated from a monkey and also attenuated in cell culture. Similar to RIT4237, while initial efficacy studies showed a modest level of protection, subsequent studies showed little to no effect.62
The poor results seen when using single-animal rotavirus strains as human vaccine candidates lead to the development of two major categories of rotavirus candidate vaccines: attenuated human rotavirus strains and recombinant (reassortant) rotaviruses containing human and animal rotavirus components. Attenuated human rotavirus strains are produced by serial passage of rotavirus strains isolated from humans in cell culture to reduce their pathogenicity. Reassortant rotavirus vaccines take advantage of the segmented rotavirus genome to create viruses that combine the RNA-encoding VP7 proteins from a human rotavirus with the remaining RNA segments of an animal rotavirus. The goal was to invoke the immune response to a G-type antigen from a human virus. Reassortants were initially created by coinfecting a monkey with bovine and human rotavirus and allowing the gene reassortant to occur by chance. Subsequent reassortants were created in laboratories and propagated in Vero cells.13
A quadrivalent human – animal reassortant vaccine (RRV-TV) containing serotypes G1, G2, G3, and G4 was developed by the National Institutes of Health and Wyeth Laboratories (Wyeth Laboratories, Monmouth Junction, NJ). The vaccine contained the RRV strain and three simian – human reassortant strains. Several different doses were tested for safety, immunogenicity, and efficacy. In the two trials done in the United States, the dose containing 105
plaque-forming units per strain had an efficacy against all rotavirus disease of 49% (95% CI: 31–63) and 57% (95% CI: 26–67) and against severe rotavirus disease of 80% (95% CI: 56–91) and 82% (95% CI: 29–88).62
There was some concern about the efficacy of the vaccine in developing countries; the higher dose vaccine when tested in Venezuela had an efficacy of 48% (95% CI: 33–61) against all rotavirus disease and 88% (95% CI: 61–96) against severe rotavirus disease, which was similar to the results found in the United States.66
The vaccine was licensed in the United States in 1998 under the trade name Rotashield®
(Wyeth Lederle, Philadelphia, PA).
After several cases of intussusception were reported via the US Vaccine Adverse Event Reporting System (VAERS), a case – control investigation was conducted in 429 infants with intussusception and 1,763 matched controls. An increased risk of intussusception 3–14 days after the first dose was found (adjusted odds ratio [OR] = 21.7, 95% CI: 9.6–48.9).67
Intussusception is not a known consequence of natural rotavirus infection. The possible explanations of why RRV-TV might cause intussusception include the following: (1) one strain in the vaccine may have been pathogenically unique from wild-type rotavirus, (2) the vaccine virus may be absorbed in the intestine in a different manner than wild type, and 3) the immune response induced by the vaccine strain might be different.13
Although the actual mechanism of the relationship between RRV-TV and intussusception is not clear, the simian strain RRV is considered the most likely causative vaccine strain. Following the results of this analysis, the use of the vaccine was discontinued. This vaccine was subsequently withdrawn from the market, and assessing the risk of intussusception has remained a key component of all subsequent rotavirus trials.
The oral human-bovine pentavalent reassortant rotavirus vaccine (Rotateq®; Merck and Co., Whitehouse Station, NJ) consists of 5 human-bovine reassortants suspended in a fully liquid buffer-stabilized formulation (RV5). The vaccine is based on the creation of a new rotavirus strain that contains a single human virus coat protein on the viral surface, with the rest of the structural proteins from a bovine strain (WC3).
Candidate reassortants were created in vitro
by coinfecting cells with WC3 and a human rotavirus strain. The progeny containing the appropriate strain mixture is selected using molecular and immunologic selection. The human-bovine reassortants for this vaccine were initially cultured using the monkey cell line MA104, and then propagated in the commonly used monkey cell line Vero using standard cell-culture technique for production. The five human serotypes contained in the vaccine are G1, G2, G3, G4, and P1A. The reassortant viruses with G1–G4 express the attachment protein P7 from the bovine strain. The reassortant virus expresses the attachment protein P1A from the human strain and the outer capsid protein G6 from the bovine strain.68
A comparison of three different potencies of the RV5 vaccine for safety, immunogenicity, and efficacy was conducted in Finnish infants aged 2–8 months from 1998 to 2001.14
Efficacy estimates for gastroenteritis of any severity in the first year following vaccination were 68.0% (95% CI: 31.1–86.4), 74.3% (95% CI: 37.9–91.0), and 57.6% (95% CI: 11.8–80.9) in the high-potency, middle-potency, and low-potency groups, respectively; the middle-potency vaccine was chosen for the subsequent phase III efficacy trials.
Subsequently, clinical trials and postintroduction studies have demonstrated the efficacy of three doses of vaccine given with routine infant immunizations (). From 2001 to 2004, the Rotavirus Efficacy and Safety Trial (REST), a double-blind, placebo-controlled, randomized trial, was conducted in over 68,000 infants in 11 countries.69
In an immunogenicity study in a small subset of the children, seroconversion rates for serum antirotavirus IgA were 95.2% (95% CI: 91.2–97.8) in 189 vaccine recipients and 14.3% (95% CI: 9.3–20.7) in 161 placebo recipients. In the per-protocol efficacy analysis, among 4,512 subjects, vaccine efficacy against G1–G4 gastroenteritis of any severity was 74% (95% CI: 66.8–79.9) and 98% (95% CI: 88.3–100) for severe disease in the first rotavirus season. Vaccine efficacy in the second season of RV gastroenteritis was 62.6% (95% CI: 44.3–75.4) against any disease and 88% (95% CI: 49.4–98.7) against severe disease. Efficacy has been shown across several regions. In fully vaccinated infants in the REST trial, reductions in RV-associated hospitalizations and emergency department visits up to 2 years after vaccination were 94.7% (95% CI: 90.9–96.9) in Europe, 94.9% (95% CI: 84.0–98.9) in the United States, and 90.0% (95% CI: 29.4–99.8) in Latin America/Caribbean.70
Comparison of vaccine efficacy and effectiveness estimations from clinical trials of RV1 and RV5 against any serotype severe rotavirus gastroenteritis, stratified by country income status.
More recently, efficacy trials were implemented in Kenya, Ghana, Mali, Bangladesh, and Vietnam to determine the protection in lower income settings, with results now available for the first year of follow-up (). Preliminary results from the efficacy trials in Africa (Kenya, Ghana, Mali) show a three-dose efficacy of 64% (95% CI, 40–79) against severe rotavirus gastroenteritis. In Asia (Bangladesh and Vietnam), a three-dose efficacy of 51% (95% CI, 13–73) has been shown against severe rotavirus gastroenteritis.71
Safety and immunogenicity have also been demonstrated in Taiwan.72
Since its introduction in the United States in 2006, RV5 has had a dramatic impact on rotaviral disease. Postintroduction surveillance data from the Centers for Disease Control and Prevention (CDC) in the United States for the 2007–2008 and 2008–2009 seasons compared with the prevaccination period showed that the rotavirus seasons were reduced from 26 weeks to 14–17 weeks, and the peak percentage of rotavirus positive tests reduced from 43% to 17%–25%.73
An evaluation of children in the United States vaccinated in the first two seasons after licensure compared with an unvaccinated cohort showed a vaccine efficacy of 100% (95% CI: 76–100) against rotavirus gastroenteritis hospitalizations and emergency department visits.74
A case – control study in Texas showed a three-dose vaccine efficacy of 100% (95% CI: 71–100) against severe rotavirus gastroenteritis requiring hospitalization. Vaccine effectiveness of one and two doses against hospitalization and emergency department visits was 69% (95% CI: 13–89) and 81% (95% CI: 13–96), respectively.75
In addition, RV5 has also shown promise in low-income settings. A postintroduction evaluation in Nicaragua showed a three-dose efficacy of 58% (95% CI: 30–74) against severe rotavirus diarrhoea.76
In the 2007 rotavirus season, despite a vaccination coverage of approximately 26%, hospitalizations and outpatient visits for diarrheal illness declined by 11% and 23%, respectively, compared with prevaccination years.77
Due to the safety concerns associated with rotavirus vaccines, enhanced monitoring for possible vaccine-associated episodes of intussusceptions and other adverse effects is ongoing in many countries. The CDC published reports of postmarketing surveillance of intussusception after RV5 vaccination through September 25th, 2007. Under the assumptions that 100% of distributed doses were given and 100% of cases of intussusception were reported, the number of cases of intussusception reported through VAERS were lower than what would have been expected for the age-adjusted baseline rates at 1–7 days after vaccination (relative risk [RR] = 0.51; 95% CI: 0.32–0.81) or at 1–21 days after vaccination (RR = 0.30; 95% CI: 0.20–0.44). The Vaccine Safety Datalink (VSD), encompassing persons enrolled in 8 large health maintenance organizations, is also being used to monitor intussusception risk post vaccination. As of May 31st, 2008, 207, 621 doses of RV5 had been administered to infants in VSD-associated health management organizations; a total of five cases of intussusception were observed in children who had received all doses compared with an expected number of 6.75 episodes (RR =0.74). 78
The rotavirus vaccine RV1 (Rotarix®
; GlaxoSmithKline, Genval, Belgium) is a monovalent vaccine composed of an attenuated human rotavirus strain G1P. This parent strain was isolated during a clinical trial of an early rotavirus vaccine in 1988.2
Although the vaccine tested in this study was eventually found to be ineffective, observation over subsequent years showed that children infected with the naturally circulating strain during the clinical trial were protected against 81% of subsequent rotavirus infections and 100% protected against severe rotavirus disease. In addition, serum antibodies produced during infection were able to neutralize G types 1–4.79
An isolate of this strain, denoted 89–12, was serially passed in tissue culture, resulting in an attenuated viral strain that was subsequently used in clinical trials.2
Phase I and II clinical trials in Europe, the United States, and Latin America demonstrated excellent safety, immunogenicity, and efficacy.80
Early clinical trials showed an efficacy of 89% (95% CI: 65.4–94.5) after two doses of vaccine resulting in adoption of the two-dose schedule.81
Subsequent large randomized, double-blind, placebo-controlled phase three clinical trials have demonstrated the efficacy of the vaccine given with the first and second dose of routine infant vaccination (). A study in Finland and 11 Latin American countries following 20,169 vaccine and placebo recipients for 9–10 months, following completion of the two-dose series, found an 85% reduction (95% CI: 71.7–92.4) in severe rotavirus gastroenteritis and a 42% (95% CI: 29–53) reduction in hospitalization for all-cause gastroenteritis.83
Subsequent follow-up in Finland showed no reduced protection during the second rotavirus season.84
A similar but smaller trial of 3,994 infants in six European countries demonstrated efficacy across two subsequent rotavirus seasons.85
During the first rotavirus season, the vaccine recipients has 87.1% (95% CI: 79.6–92.1) fewer rotavirus episodes and 95.8% (95% CI: 89.6–98.7) fewer severe rotavirus episodes. During the second season, the efficacy dropped slightly to 71.9% (95% CI: 61.2–79.8) against any rotavirus gastroenteritis and 85.6% (95% CI: 75.8–91.9) against severe rotavirus disease. A study in Hong Kong, Singapore, and Taiwan following 10,708 children until 24 months of age found higher efficacy, noting a 96.1% (95% CI: 85.1–99.5) reduction in severe rotavirus gastroenteritis and 30.3% (95% CI: 13.1–44.2) against reporting one or more episodes of severe all-cause gastroenteritis.86
To assess RV1 efficacy in low-income countries, a randomized trial was conducted in 3,166 South African and 1,773 Malawian infants.87
In South Africa, the overall efficacy was similar to what was seen in high-income country settings, with the vaccine recipients having 76.9% (95% CI: 56.0–88.4) fewer cases of severe rotavirus gastroenteritis and 44.1% (95% CI: 19.8–61.0) lower incidence of all-cause severe gastroenteritis. However, in Malawian children, the efficacy of the vaccine was reduced, preventing only 49.4% (95% CI: 19.2–68.3) of severe rotavirus gastroenteritis. This difference could not be explained by serotype differences, as both populations showed similar efficacy against G1 and non-G1 serotypes. Interestingly, despite the reduced efficacy, more disease was prevented per recipient in Malawi than in South Africa due to the higher burden of disease. In Malawi, vaccine prevented 3.9 episodes of severe rotavirus gastroenteritis per 100 vaccinated children compared with preventing 2.5 episodes per 100 vaccinated children in South Africa.88
In addition, the study compared two- and three-dose schedules of RV1, with vaccine given at 6, 10, and 14 weeks or 10 and 14 weeks. No statistical difference was found in the efficacy of the vaccine in children who received two versus three doses.
RV1 has demonstrated efficacy against vaccine serotype and nonvaccine serotype diseases. In Finland and Latin America, RV1 showed serotype-specific efficacy for G1P of 91.8% (95% CI: 74.1–98.4) and a combined efficacy against G3P, G4P, and G9P of 87.3% (95% CI: 64.1–96.7).83
Too few cases of non-G1 or P serotypes were isolated to determine efficacy against these groups. In Europe, the highest efficacy was seen against G1 serotypes, with the vaccine preventing 89.8% (95% CI: 82.9–94.2) of any severe gastroenteritis.85
In Asia, RV1 prevented 100% (95% CI: 80.8–100) of G1 serotypes and 93.6% (95% CI: 74.7–99.3) of non-G1 serotypes of severe gastroenteritis.86
Several Brazilian studies have shown efficacy against non-vaccine serotypes.89
Among these, a case control study comparing children hospitalized with G2P rotavirus gastroenteritis to those hospitalized with acute respiratory infections demonstrated an efficacy of 77% (95% CI: 43–90) in infants aged 6–11 months.90
However, the efficacy dropped to 15% (95% CI: −101 to 64) in children older than 12 months. Further studies are needed to establish the efficacy of RV1 against serotypes that are neither G1 nor P.
Postintroduction impact of RV1 has been demonstrated in several countries.91
A retrospective case – control study in indigenous Australians during an outbreak of G9P rotavirus found RV1 to be 84.5% (95% CI: 23.4–96.9) effective at preventing hospitalized rotavirus infections.92
RV1 was introduced to the national immunization system in Brazil in 2006, with a vaccination coverage of 46.5% in 2006 and 78.3% in 2007.91
In 2006, there were 26% fewer hospitalizations in children aged younger than 1 year due to gastroenteritis compared with the average number of yearly hospitalizations from 1998 to 2005. In 2007, the hospitalization reduction was 48% compared with prevaccination levels. In 2007, a decrease in all-cause diarrheal disease was seen in children aged 1–5 years, reversing an increasing trend in gastroenteritis in this age group that had been seen in the past 10 years. Mexico introduced RV1 in 2007. In 2008, the incidence of diarrhea-related deaths decreased to 11.8 per 100,000 children younger than 5 years compared with 2003–2006 with an average of 18.1 deaths per 100,000 children; a rate reduction of 35% (95% CI: 29–39). This reduction continued though the 2009 rotavirus season. Such a dramatic reduction in diarrhea mortality is very promising for reducing mortality due to rotavirus infections.
Due to the previous association of rotavirus vaccination with intussusception, an extended intensive follow-up phase to assess severe side effects was built into most clinical trials.83
Over 75,000 children were followed in these studies for 31–100 days but no increased risk of intussusception was found. In fact, across all studies, there were significantly fewer serious adverse events noted in the vaccine recipients compared with placebo recipients.83
In addition, a study in East Asia reported no difference in nongastrointestinal serious adverse events in vaccine vs placebo recipients.86
A live, attenuated oral rotavirus vaccine (LLV) was developed by the Lanzhou Institute of Biological Products. The vaccine was developed by passing a wild-type group A serotype G10P lamb rotavirus through primary calf kidney cells. After 37 passages, the virus was subsequently tested in several clinical studies. In 2000, LLR vaccine was licensed by China Drug Inspection and Management Bureau using a one-dose oral schedule. It is currently licensed in China to be given to children aged 2 to 36 months, followed by yearly boosters.94
However, the vaccine is relatively expensive in China, costing $18.4 per dose; as a result, few children received more than one dose.95
Between 2001 and 2008, approximately 10,000,000 doses were administered in China.94
The vaccine is not routinely being used under China’s national immunization program.
The strongest evidence for efficacy of LLV comes from a case control study in Guangzhou province comparing 838 children aged 2 months to 5 years hospitalized with rotavirus infections to 838 matched community controls.95
This study demonstrated a 73.3% (95% CI: 61.2–81.6) efficacy of one dose of LLV against hospitalized rotavirus gastroenteritis. The efficacy was found to be higher in older children, with efficacy in 12–23 month olds found to be 80.9% (95% CI: 65.4–89.4) compared with 60.0% (95% CI: 28.6–77.6) in 2–11 month olds. This may be due in-part to older children receiving vaccine at a later date, resulting in booster effect on an already present wild-type rotavirus response. A larger efficacy trial done in 4,000 infants aged from 6 to 24 months showed an efficacy against all-cause rotavirus gastroenteritis of 78%.94
However, this study was not placebo controlled, and the results are not available in a peer-reviewed journal.
A nonrandomized cohort study in Guangzhou province in southern China following 102 vaccinated children and 145 unvaccinated children aged 6 months to 3 years for a total of 6 months found 53% fewer cases of rotavirus gastroenteritis in vaccinated children.96
In addition, disease was less severe; the average duration of disease was reduced by 24% (P
< 0.001). A study comparing 225 children with rotavirus disease, 34 previously vaccinated and 191 unvaccinated, found a reduced duration (P
< 0.001), severity (P
= 0.041), and risk of hospitalization (P
= 0.022) in the rotavirus vaccinated group.97
Immunologic studies among children aged 6–24 months before and after vaccination showed that LLR induced neutralizing antibody against rotavirus of all the four G types ranging from 40% to 70%.94
However, these studies did not include non-vaccinated controls.
Available data suggests that there are no major safety concerns with LLR. Trials carried out in Beijing, Guangxi province, and Zhejiang province reported mild side effects, with 5.6%–8.2% of vaccinated children experiencing low-grade fever and 0%–2.1% reporting high-grade fever.98
There has been one case report of intussusception following vaccination with LLR.101
Appropriately powered studies to assess differences in intussusceptions rates in vaccinated vs unvaccinated children have yet to be reported.
LLR has never been tested in a randomized, placebo-controlled phase III clinical trial,95
so the true efficacy of the vaccine is unknown. In all of the reported trials, the majority of children were vaccinated during or after the peak age for rotavirus disease. In China, 50% of rotavirus infections occur between the ages of 6 and 11 months.102
It is unknown whether vaccinated children had prior natural rotavirus infection, in which case LLV could be boosting an already present immune response. The efficacy of LLR in rotavirus-naive populations is unknown. Controlled studies in which children are vaccinated prior to peak age of rotavirus incidence are necessary to demonstrate its true efficacy and potential impact on rotavirus gastroenteritis.
Vaccines under development
In India, two live, cell line-adapted human viruses obtained from asymptomatic neonatal strains (116E and I321) recently underwent early clinical evaluation.103
Candidates using both vaccines were shown to be safe, with adverse event rates not statistically different in the vaccinated vs placebo groups. 116E, a predominately human vaccine strain of serotype G9P, with the VP4 appearing to be a natural reassortant from a bovine strain of rotavirus, seems to be more promising. Immunologic evaluation showed a 73% serum IgA conversion rate compared with 20% in the placebo group. I321 was less immunogenic, resulting in 39% seroconversion. A subsequent immunogenicity study of 116E in 93 and 91 children receiving 104
focus-forming units of vaccine virus, respectively, were compared with 184 placebo recipients.104
There was a ≥4-fold increase in antirotavirus IgA titers in 66.7% of children receiving the lower dose and 62.1% of children receiving the higher dose after the first dose compared with 18% of placébo recipients. After 3 doses, 64.5% of infants receiving the lower dose and 89.7% of those receiving the higher dose had a ≥4-fold increase in IgA compared with 25% of placebo recipients. No increase in adverse events was found between vaccine and placebo recipients. This vaccine is scheduled to begin phase III efficacy trials in late 2010.
Early rotavirus vaccine development studies are underway in Vietnam in an effort to produce inexpensive rotavirus vaccine locally. Three human rotavirus strains (genotypes G1P, G1P, and G4P) have been characterized for potential use in a live attenuated vaccine.105
The wild-type viruses for these strains have been passed more than 30 times, each through cell culture with the goal of developing an effective indigenously produced vaccine. The results of clinical trials are not yet available.