Rotavirus vaccines: safety, efficacy and public health impact

Authors


Professor Jim Gray, Norfolk and Norwich University Hospital, Specialist Virology Centre, Microbiology Department, NRP Innovation Centre, Norwich Research Park, Norwich NR4 7GJ, UK.
(fax: 01603 458448; e-mail: james.gray@nnuh.nhs.uk).

Abstract

Abstract.  Gray J (Norfolk and Norwich University Hospital, Norwich, UK). Rotavirus vaccines: safety, efficacy and public health impact (Foresight). J Intern Med 2011; 270: 206–214.

Rotaviruses are the cause of acute gastroenteritis, and disease is widespread amongst infants and young children throughout the world. Also, rotavirus is associated with significant mortality in developing countries with more than 500 000 children dying each year as a result of the severe dehydration associated with rotavirus disease. Efforts have been ongoing for more than 30 years to develop a safe and effective rotavirus vaccine. Currently, two vaccines, RotaRix and RotaTeq, have been licensed for use in many countries throughout the world following comprehensive safety and efficiency trials. Monitoring their effectiveness after licensure has confirmed that their incorporation into early childhood vaccination schedules can significantly prevent severe rotavirus diarrhoea, which would have resulted in hospitalizations, emergency room visits or increased diarrhoea-related mortality. Although the efficacy of both vaccines is lower at approximately 40–59% in developing countries, their use could significantly reduce the mortality associated with rotavirus disease that is concentrated in these countries.

Rotavirus and childhood diarrhoea

Rotavirus is the most common cause of severe gastroenteritis in infants and young children worldwide. Over 500 000 deaths per year are attributed to rotavirus, and rotavirus mortality is heavily weighted to developing countries worldwide. In developed countries, the healthcare burden is associated primarily with the costs of medical visits and hospitalizations.

After an incubation period of 1–3 days, symptoms of watery diarrhoea, vomiting, fever and abdominal pain follow. Rotavirus is excreted at 1011 particles per g or ml of faeces, and with an infectious dose of as little as 10–100 virus particles, rotavirus is easily transmitted from person to person, or following the ingestion of contaminated water or food or through contact with contaminated environmental surfaces. Nosocomial spread is common in hospitals where children are treated for rotavirus infection.

Rotaviruses infect the apical cells of the villi of the small intestine with the resulting necrosis reducing digestion and causing diarrhoea through malabsorption. Watery diarrhoea and the abrupt onset of vomiting might result in severe, life-threatening dehydration. Treatment of rotavirus disease is by oral or intravenous rehydration therapy with appropriate feeds taken during the episode. Breast feeding should be encouraged as breast milk contains maternal anti-rotavirus antibodies and the glycoprotein lactadherin which bind specifically to rotavirus particles and inhibit their replication.

Prevention of rotavirus infection and disease has not been associated with improvements throughout the world in hygiene or water quality. In countries with no universal childhood rotavirus vaccination programme, any reduction in rotavirus-associated mortality is more likely to be associated with improved access to healthcare, rehydration fluids and rehydration therapy.

Rotavirus biology

Rotaviruses are unenveloped viruses with an icosahedral triple-layered capsid of 75 nm in diameter. The rotavirus genome consists of 11 segments of dsRNA which encode six structural and six nonstructural proteins. Rotavirus evolution and diversity result from genome rearrangement, antigenic drift through the accumulation of point mutations which arise during the error prone viral RNA replication and antigenic shift through reassortment of RNA genome segments during dual infection of a single cell [1].

Viruses derived through genome rearrangement are not a significant public health risk as rearrangement usually takes place in genes encoding nonstructural proteins. The accumulation of point mutations might result in antibody escape mutants that can cause outbreaks or epidemics in immunologically naive populations, and reassortment can result in hybrid rotavirus strains with proteins acquired from both human and animal strains again giving rise to epidemics in an immunologically naïve population. Cross-species rotavirus transmission is a frequent event, and although animal strains replicate poorly in the human host, they can acquire – through reassortment with a human strain – the replicative advantage and transmissibility of the human strain and antigens of the animal strain, which are new to the human population (Table 1, Fig. 1).

Table 1. Diversity of rotavirus strains co-circulating throughout Europe between September 2006 and August 2010 and their possible origins
GenotypeaSeasonbTotalPer cent
2006/20072007/20082008/20092009/2010
  1. aRotaviruses have a dual nomenclature reflecting the genetic or antigenic diversity of the surface glycoprotein (G) VP7 and the protease-sensitive spike protein (P) VP4.

  2. bA season is defined as September to the following August.

  3. cEmerged as a common rotavirus strain worldwide in the 1990s following a possible reassortment event between an animal strain containing G9 and a common human strain of the P[8] genotype.

Common human strains
G1P[8]163745353434312912 73549.851
G2P[4]5847497251079313712.280
G3P[8]11332442841212774.999
G4P[8]282130116291201441317.275
G9P[8]c768902615770305511.959
Subtotal338478116831659124 61796.363
Reassortants of common strains
G1P[4]11301727850.333
G2P[8]214624101010.395
G3P[4]11522290.114
G4P[4]16208350.137
G9P[4]4161733700.274
Subtotal3899831003201.253
Possible human/animal hybrid rotaviruses
G1P[6]  4150.020
G2P[6]51736310.121
G2P[9]11  20.008
G2P[10]1   10.004
G3P[3]1   10.004
G3P[6]  185230.090
G3P[9] 23490.035
G4P[6]5 49180.070
G4P[9] 15280.031
G4P[10]  2 20.008
G6P[4] 1  10.004
G6P[8] 21250.020
G8P[4] 363781440.564
G8P[8]236 110.043
G10P[4]12  30.012
G10P[8] 61510310.121
G12P[4]  3 30.012
G12P[8]214652471660.650
Subtotal37841791644641.816
Possible zoonotic introductions
G5P[6]   110.004
G6P[5]   110.004
G6P[6]   110.004
G6P[9]43111190.074
G6P[11]1   10.004
G6P[14]13 6100.039
G8P[14] 12580.031
G8P[6]411170.027
G9P[6]5 5 100.039
G9P[9]231 60.023
G9P[10] 11 20.008
G10P[6] 1 230.012
G10P[10] 1  10.004
G10P[14]51  60.023
G12P[6]144113590.231
G12P[9] 451100.039
Subtotal232357421450.568
Total348280177150689725 546 
Figure 1.

 Reassortment event between an animal rotavirus (G8P[6]) and a human rotavirus (G2P[4]) in which the animal rotavirus acquired the gene encoding VP4 (P[4]) (mauve spike) from the human rotavirus and the human rotavirus acquired P[6] (red spike) from the animal strain.

Rotaviruses infect the young of many species, and zoonotic introductions into the human population are not uncommon. A total of 145/25 546 (0.57%) rotavirus strains, characterized throughout Europe between 2006 and 2010, were of possible animal origin (Table 1). For reassortment to occur, mixed rotavirus infections are required to allow the possibility of two rotavirus strains to infect and replicate within the same cell. In the same European study, 1175/27 459 (4.27%) mixed rotavirus infections were identified and 134 (0.5%) had proteins associated more commonly with animal rotavirus strains (http://www.eurorota.net/).

Rotavirus vaccines

Vaccine development strategies have concentrated on the development of live-attenuated rotavirus vaccines that can be administered by the oral route. Live-attenuated vaccines replicate to produce immunity, and the immune response is similar to that of a natural infection. Oral vaccines produce secretory immunity which can operate at the luminal face of the epithelium and intracellulary in infected cells. Oral vaccination is more efficient at giving rise to secretory (S) IgA antibodies, and live vaccines give rise, not only to SIgA, but also to longstanding serum rotavirus-specific IgG and IgA responses. Natural protection from rotavirus disease is additive with three natural symptomatic or asymptomatic rotavirus infections protecting against subsequent rotavirus diarrhoea in childhood. Rotavirus infection does not result in a sterilizing immunity, and individuals will be re-infected throughout their lifetime, resulting in asymptomatic infections which further boost their immunity or symptomatic infection if rotavirus-specific immunity has waned.

Taking into account the need for mucosal immunity and the genomic and antigenic diversity of rotaviruses co-circulating in the community, efforts to develop rotavirus vaccines have been ongoing for more than 30 years, resulting in the development and licensure of three live-attenuated oral rotavirus vaccines, RotaShield (Wyeth, Marietta, PA, USA) [2], RotaRix (GSK, Rixensart, Belgium) [3] and RotaTeq (Merck, Whitehouse Station, NJ, USA) [4].

RotaShield, a live-attenuated rhesus–human reassortant vaccine, was withdrawn from use because of an association with the development of intussusception in children within 7 days of the fist vaccine dose [5].

The RotaRix live-attenuated oral rotavirus vaccine is monovalent and derived from a G1P[8] human rotavirus strain attenuated through serial passage in cell culture. It has a two-dose schedule with the first dose from 6 weeks of age with an interval of ≥4 weeks till the second dose. The schedule must be completed by 24 weeks of age.

RotaTeq is a pentavalent live-attenuated oral rotavirus vaccine derived through the reassortment of a G6P[5] bovine rotavirus strain with common human rotavirus strains to give human–bovine rotavirus reassortants carrying G1, G2, G3, G4 and P[8] human rotavirus surface proteins. The virus is naturally attenuated being a cattle strain and was further attenuated by passage in cell culture during the production of reassortants. It has a three-dose schedule with the first dose from 6 weeks of age with an interval of ≥4 weeks between doses. The schedule must be completed by 24 weeks of age.

For protection from rotavirus strains of different serotypes, RotaRix vaccine (G1 and P[8]) relies on the protection from rotavirus disease being heterotypic, resulting in protection from antigenically or genetically similar viruses of different serotypes/genotypes. RotaTeq contains antigens from a range of common human rotavirus strains (G1, G2, G3, G4 and P[8]), allowing homotypic protection from viruses of the same genotypes/serotypes.

Rotaviruses express 12 proteins, many of which have been shown to elicit antibody responses. VP7 and VP4 elicit neutralizing antibodies but it is as yet unclear how the immune responses to other structural and nonstructural proteins such as VP6, the most abundant structural protein, and NSP4, a possible viral enterotoxin, provide protection from subsequent disease. Figure 2 shows antigens shared amongst common and possible zoonotic rotavirus strains indicating less genetic and antigenic diversity within these proteins, suggesting cross protection amongst different G and P types might be possible through the immune responses to several structural and nonstructural proteins.

Figure 2.

 Shared rotavirus antigens amongst common rotavirus strains (G1, G2, G3, G4, G9), human/animal hybrid strains (G12P[8]) and possible zoonotic strains (G10P[11], G12P[6]) compared with antigens present in rotavirus vaccines. Antigens shared with the rotavirus vaccines are shown with a tick, blue for RotaRix and red for RotaTeq.

Randomized, double-blind, placebo-controlled rotavirus vaccine trials

Safety

Clinical trials of the new candidate vaccines, RotaRix [3] and RotaTeq [4], were designed specifically to determine the effect of rotavirus vaccination on the risk for intussusception following the withdrawal of the RotaShield vaccine. In two clinical trials, reported in 2006 on the safety and efficacy of RotaRix and Rotateq, 63 225 and 68 038 healthy infants were recruited, respectively. For RotaRix, six cases of intussusception were reported in 31 673 vaccine recipients monitored for 31 days after each dose versus seven cases in 31 552 placebo recipients [relative risk (RR) 0.9, 0.3–2.4]. A RR of 0.3, 0.1–0.8 was determined in a smaller cohort of infants followed up for 1 year after each dose. For RotaTeq, six cases of intussusception were reported amongst 28 038 vaccine recipients followed for 42 days after each dose of vaccine compared with five cases in 27 965 placebo recipients (RR 1.2, 0.3–5.0). A total of 12 cases of intussusception were reported amongst vaccine recipients and 15 in placebo recipients followed up for 1 year after each dose of RotaTeq vaccine (RR 0.8, 0.3–1.8).

For RotaRix, no differences were observed compared to placebo in the incidence of diarrhoea, fever, vomiting and irritability. For RotaTeq, no significant differences were detected in the incidence of fever and irritability between the vaccine and placebo groups but a slight increase in the incidence of mild diarrhoea and vomiting was detected in vaccine recipients. The risk of death or other serious adverse events in vaccinees was not statistically significantly increased when compared with the placebo recipients.

Vaccine efficacy

For RotaRix, the efficacy of the vaccine against severe rotavirus gastroenteritis and rotavirus-associated hospitalization was 85% (Table 2) [3]. For RotaTeq, the efficacy against severe rotavirus gastroenteritis was 98%, and the vaccine reduced emergency department visits and hospitalizations associated with the common human rotavirus genotypes by 95% (Table 2) [4].

Table 2. Efficacy of live-attenuated oral rotavirus vaccines determined in randomized, double-blind, placebo-controlled clinical trials
VaccineYear reportedTrial conducted inEfficacy againstEfficacy (95% confidence intervals)
  1. RVGE, rotavirus gastroenteritis.

  2. aVaccine efficacy was lower in Malawi than in South Africa (49.4% vs. 76.9%).

RotaShield [2]1996USARotavirus episodes49.0%
Very severe episodes82.0%
Dehydrating illness100.0%
RotaShield1997FinlandRVGE of any severity68.0% (57.0–76.0)
Severe RVGE91.0% (82.0–96.0)
RotaTeq [3]2006Finland, USA (including Navajo Nation and White Mountain Apache Tribe)RVGE of any severity74.0% (66.8–79.9)
Severe RVGE98.0% (88.3–100)
RotaRix [4]2006Argentina, Brazil, Chile, Colombia, the Dominican Republic, Honduras, Mexico, Nicaragua, Panama, Peru, Venezuela, FinlandSevere RVGE and rotavirus-associated hospitalizations84.8% (71.1–92.7)
RotaRix [6]2009SingaporeSevere RVGE96.1% (85.1–99.5)
RotaTeq [7]2010Ghana, Kenya, MaliRVGE of any severity30.5% (16.7–42.2)
Severe RVGE39.3% (19.1–54.7)
RotaTeq [8]2010Bangladesh, VietnamRVGE of any severity42.5% (21.1–58.4)
Severe RVGE48.3% (22.1–66.1)
RotaRix [9]2010Malawia, South AfricaSevere RVGE61.2% (44.0–73.2)

Rotavirus vaccine efficacy is significantly lower in clinical trials conducted in developing countries when compared to developed countries (Table 2) [6–9]. This is consistent with the findings in studies of other live-attenuated oral vaccines such as oral poliovirus vaccine and oral typhoid vaccines. It has been hypothesized that this difference might be the result of poor nutritional status, enteric co-infections or through co-administration with other live-attenuated oral vaccines such as oral poliovirus vaccine. Also, the impact of breast feeding with high concentrations of anti-rotavirus antibodies in milk in a population that is boosted regularly or through interference from maternal antibodies might neutralize live virus vaccine.

Health impact in countries with universal childhood rotavirus vaccination

Vaccine effectiveness (VE) and public health impact can be measured in several different ways. Reductions in mortality associated with acute gastroenteritis, hospitalizations for acute gastroenteritis or rotavirus gastroenteritis, nosocomial rotavirus infections, emergency room visits, admission to short stay units, general practice or paediatrician visits and in the number of rotavirus positive laboratory diagnostic tests have all been used as measures of VE.

Postvaccine introduction surveillance data, as well as those obtained through clinical trials (Table 2), demonstrate that RotaTeq and RotaRix are effective vaccines for preventing severe rotavirus disease (VE = 17–100% reduction in disease when compared with the prevaccine era), for reducing acute gastroenteritis-associated mortality (29–35% reduction) and lessening acute gastroenteritis healthcare utilization in general practice, emergency room visits and hospitalizations (25–64% reduction). The number of laboratory diagnosed rotavirus infections has also been reduced by 50–86% (Table 3).

Table 3. Rotavirus vaccine coverage and effectiveness postintroduction as part of the universal childhood vaccination schedule in Europe, Central and South America, the USA and Australia
CountryVaccine coverageVaccine effectiveness (VE) measured through a reduction inComments
Australia [12, 13, 14]89.6% for 1 or more doses, 73.1% for 3 dosesHospitalizations for acute gastroenteritis (AGE)62.3–63.9%Individual states and territories chose either RotaRix (2 doses) or RotaTeq (3 doses). Both vaccines were associated with a marked reduction in gastroenteritis admissions
Hospitalizations for rotavirus gastroenteritis (RGE)89.3–93.9% 
Nosocomial infections87%Study conducted at Childrens Hospital at Westmead, New South Wales
Rotavirus notifications53–57%Laboratory-confirmed rotavirus disease is notifiable in Queensland where this study was performed over 2 rotavirus seasons
Belgium [15]∼30% in 2007 to ∼77% in 2009Hospitalizations for RGE65.0–83.0%Effectiveness improved with increasing vaccine coverage
Laboratory diagnosed rotavirus cases50%Sentinal network of laboratories representing 47–63% of private or hospital microbiology laboratories in Belgium
Brazil [16, 17]74.0% completed 2 coursesDiarrhoea-related mortality29–33%In children 1–4 years of age in 2007 and 2008
84% in <1-year-old childrenHospitalizations for RGE17.00%Three years following vaccine introduction
81% in 1–2-year-old children   
El Salvador [18]∼50–61%Hospitalizations for RGE69.0–81.0%During the 2.5 years following vaccine introduction
Diarrhoea-associated healthcare visits35.0–48.0%Children <5 years of age during 2008 and 2009
Finland [19]50.00%Hospitalizations for RGE93.80%VE for up to 3.1 years following the last vaccine dose
Mexico [20, 21]∼74% in 2008 and ∼89% in 2009Diarrhoea-related mortality35.00%In children <5 years of age in 2008 and 2009
Hospitalizations for AGE40% 
Panama [22]∼63% at end 2006 and ∼94% at end of 2009Hospitalizations for AGE22.0–37.0%In children <5 years of age in 2007 and 2008
USA [23, 24, 25, 26, 27]48–86% as determined at 8 sentinel sites in 2009Hospitalizations for AGE59%In a study using an insurance claims database
Hospitalizations for RGE100%In a study using an insurance claims database
Outpatient visits for AGE28%In a study using an insurance claims database
Outpatient visits for RGE96%In a study using an insurance claims database
Laboratory diagnosed rotavirus cases60%In 2007–2008, National Respiratory and Enteric Virus Surveillance System
Laboratory diagnosed rotavirus cases86%In 2009–2010, National Respiratory and Enteric Virus Surveillance System
All-cause AGE-related activity seen in primary paediatric practice∼25% 
Emergency department visits for RGE89%Case–control study, 2008, Houston, USA

It is clear from the clinical trials data and postvaccine introduction surveillance data that both RotaTeq and RotaRix rotavirus vaccines have a lower efficacy in developing countries but their effectiveness in preventing severe rotavirus disease will result in significant reductions in mortality.

Other measurable effects of universal rotavirus vaccine introduction

Herd immunity  Many researchers involved in the surveillance of rotavirus disease following vaccine introduction have identified a reduction in rotavirus incidence in children born prior to the introduction of rotavirus vaccine as well as in the cohort of children vaccinated. This herd immunity might have arisen through the lack of transmission opportunities amongst families where, prevaccine introduction, infected children would infect their older siblings and parents. Interestingly, the increase in rotavirus disease in older children in the second year after vaccination would suggest that the population of susceptible older children was sufficient for efficient person-to-person transmission.

Temporal and spatiotemporal changes in rotavirus activity

On the European continent, the incidence of rotavirus infection peaks in Spain in December or January, whereas in Hungary and Finland the peak of infection is in May, suggesting its spread across Europe is from the south-west to the north-east [10]. Similarly in the USA, long-term prevaccine surveillance of rotavirus incidence demonstrated a spatiotemporal pattern of annual activity with incidence peaking first in the south-west during the winter months and travelling across the continent to peak finally in the north-east 2–3 months later.

After vaccine introduction in the USA, this pattern no longer occurs, suggesting that the sweep of rotavirus infection across the North American continent is driven by the rate of accumulation of susceptible individuals [11]. In Europe as yet, there is insufficient vaccine coverage to determine whether this phenomenon can be repeated although the peak of infection is later when compared with pre-introduction incidence peaks in those countries with a universal rotavirus vaccination policy.

Current data suggest that both licensed vaccines, RotaRix and RotaTeq, are effective and safe, although efficacy in developing countries is significantly lower than that in the developed world. RotaRix and RotaTeq vaccine introductions throughout the world have delivered significant reductions in severe disease and hospitalizations associated with rotavirus infection and it is expected that in developing countries, significant reductions in rotavirus mortality will follow the introduction of rotavirus vaccine.

The choice of vaccine may be dependant on the ability to integrate vaccination within the current childhood vaccination programmes given that both rotavirus vaccine schedules should be completed by 24 weeks of age, the capacity for storage within the current cold chain depending on size of packaging, and the cost of administration. An at-risk vaccine programme is not sustainable as there are no risk factors to predict progression to severe rotavirus disease, hospitalization and death. Universal childhood vaccination should be adopted as all infants and young children are at risk of rotavirus infection.

Conflict of interest statement

No conflict of interest was declared.

Acknowledgements

I thank the participants of EuroRotaNet (http://www.eurorota.net/) for providing the data on rotavirus strains co-circulating within Europe over the last four rotavirus seasons.

Ancillary