In 2006, a new rotavirus vaccine, RotaTeq (Merck & Co, Inc., Whitehouse Station, NJ), was licensed in the United States and recommended by the Advisory Committee on Immunization Practices and the American Academy of Pediatrics1,2 for the routine immunization of all American children. This decision was based on recognition of the burden of rotavirus disease in this age group and the safety and efficacy of the vaccine as demonstrated in large scale clinical trials. The estimated median coverage of RotaTeq vaccine (one or more doses among 3-month-old infants from sentinel US sites) has increased steadily since June 2006 and reached 58% by the end of 2007.3 This program already has had great success as measured by a greater than 45% reduction in the number of hospitalizations for gastroenteritis among children less than 5 years old during the 2007–2008 rotavirus season4,5 and the number of stool specimens testing positive for rotavirus has dropped markedly in laboratories around the country.3,6 Similar benefits have been noted in other settings as well.4
One unanswered question raised before the vaccine was introduced was whether pressure from the vaccine might alter the distribution of rotavirus strains in circulation. Vaccines could put new evolutionary pressures on circulating rotavirus strains, favoring those that are most resistant to the neutralizing activity of vaccination-induced immunity. Reports from Brazil and Australia have described changes in rotavirus genotype prevalence following vaccine introduction but it has not been established that these observations are a result of vaccine-induced immune pressure or simply reflect natural temporal variations in rotavirus genotype circulation.7–12
Rotavirus strains are commonly characterized by the presence of 2 neutralizing antigens on their outer capsid, a glycosylated outer surface protein (G-protein) encoded by the VP7 gene, and the protease-cleaved protein (P-protein) encoded by the VP4 gene. The standard genetic methods to characterize these strains have been reverse transcription-polymerase chain reaction (RT-PCR) genotyping and sequence analysis of the gene segments that encode each of these proteins.13–15
RotaTeq is a pentavalent vaccine which expresses the 4 most common human rotavirus G-types, G1, G2, G3, and G4 and the most common P-type, P1A and was produced by laboratory reassortment of human rotavirus strains with the bovine rotavirus strain WC3.16 If the vaccine failed to protect against any single strain, or against any novel strains, surveillance of strains before and after vaccine introduction might provide clues that could affect the outcome of this vaccine strategy.
Since 1996, the Rotavirus Laboratory at the Centers for Disease Control and Prevention (CDC) has been monitoring rotavirus strains collected from sentinel laboratories across the United States to understand the distribution of genotypes that a vaccine might have to address if it were to be effective. In this study, we report results of rotavirus strain surveillance for the 3 seasons, the season immediately before the vaccine was introduced (2005–2006) and the 2 subsequent seasons (2006–2007, 2007–2008). Our goal was to determine if vaccine introduction might have begun to alter the distribution of strains in the United States.
MATERIALS AND METHODS
Laboratories recruited for the National Rotavirus Strain Surveillance System represented a subset of those participating in the National Respiratory and Enterovirus Surveillance System (NREVSS; Fig. 1). 17 Aliquots of stool were retained from specimens that tested positive for rotavirus antigen by enzyme immunoassay (EIA) performed as part of routine clinical testing at laboratories across the country. All specimens were collected from November to June (the “rotavirus season”) and were shipped to the CDC along with corresponding stool collection date.
At the CDC, samples were stored at 4°C until confirmatory detection of rotavirus antigen in human stool specimens was performed by EIA using the Premier Rotaclone Rotavirus Detection Kit (Meridian Diagnostics, Inc, Cincinnati, OH).
Viral RNA Extraction and RT-PCR Genotyping
For EIA confirmed positive samples, rotavirus double-stranded RNA or total nucleic acid was extracted by using 1 of 3 methods and corresponding standard protocols: the automated NucliSens Extractor (Biomerieux, Durham, NC); the automated KingFisher extraction system (Thermo Electron Corporation, Vantaa, Finland); or the semi-automated miniMAG magnetic extractor (Biomerieux, Durham, NC).
G-type (VP7) and P-type (VP4) genotyping were carried out following previously described 2-step amplification methods13,14 with modification on a GeneAMP PCR System 9700 thermocycler (Applied Biosystems, Foster City, CA). During 2005–2006 and 2006–2007 seasons, VP7 was first amplified by RT-PCR using primers 9Con1-L and VP7-RDg and VP4 was first reverse-transcribed and amplified using primers con3 and con2. During 2007–2008 season, the RT-PCR step was modified to use the One-Step RT-PCR kit (Qiagen, Inc., Valencia, CA) on a GeneAMP PCR System 9700 thermocycler (Applied Biosystems, Foster City, CA). In this procedure, the dsRNA was denatured at 97°C for 5 minutes and then the reaction conditions were set up as described in the kit instructions. RT was performed at 42°C for 30 minutes, followed by 15 minutes at 95°C, and the PCR was performed for 30 cycles consisting of 30 seconds at 95°C, 30 seconds at 42°C, and 1 minute at 72°C. Reactions were then subjected to a final extension step of 72°C for 7 minutes then held at 4°C. For all 3 seasons, genotyping PCR was performed as described previously.13,14 G-typing used primer 9Con1-L in combination with primers 9T1–1, 9T1-Dg, 9T2, 9T3P, 9T-4, and 9T-9B.13,14,18 P-typing used primer con3 in combination with primers 1T-1, 1T1-V, 2T-1, 3T-1, 4T-1, 5T-1, and JRG23714,18,19 and new primer 1T1-A (5′-CAC GTC GAT CCA GTA GA-3′). Genotyping reactions were analyzed by electrophoresis on a 3% agarose gel using a 2:1 ratio of NuSieve GTG: SeaPlaque (FMC Bioproducts, Rockland, ME). The G- and P-types obtained were classified according to the system described by Estes and Cohen.20
Our genotyping strategy was based on the total number of specimens submitted from each site in an individual season. If 100 or fewer samples were submitted by a laboratory during a given season, all EIA positive samples were both G- and P-genotyped, but if more than 100 specimens were submitted, a subset was genotyped for both genes. Selection of each subset was based on the sample collection month, the EIA absorbance value, and the G genotype observed.
Nucleotide Sequencing and Sequence Analysis
Isolates that did not yield a visible PCR product by RT-PCR genotyping were characterized by nucleotide sequencing of the con3/con2 and 9Con1-L/VP7-RDg RT-PCR amplicons. In cases when a con3/con2 product was not obtained, the VP4 gene was reverse transcribed and amplified with the primer pair primers VP4F and VP4R21 using the one-step RT-PCR kit. Similarly, in cases where a 9Con1-L/VP7-RDg RT-PCR product was not obtained, primer pair Beg9/End9 was employed.15 Analysis of RT-PCR reactions by gel electrophoresis, amplicon purification, and DNA sequencing was carried out as described previously.22 Forward and reverse sequences were assembled using Sequencher software versions 4.7 or 4.8 (Gene Codes Corporation, Inc, Ann Arbor, MI). The consensus sequences obtained were compared with existing rotavirus sequences in the NR/NT database using the BLASTN program at the National Center for Biotechnology Information website (available at: http://www.ncbi.nlm.gov/BLAST/). Specimens that did not yield a visible product by RT-PCR were considered to be nontypeable.17
Nonparametric statistical analyses of genotype and age data were performed by using Prism Version 5.02 Software for Windows (GraphPad Software, Inc., La Jolla, CA). The means of 3 or more groups were compared by Kruskal-Wallis test, with Dunn Multiple Comparison Post-Test used to evaluate the difference in the rank sums between any 2 data columns. The Mann-Whitney U Test was used to compare the distributions of 2 groups. Data from patients over 35 months of age were included in the overall genotyping results but excluded from age analyses.
At the CDC, all 1248 rotavirus-positive samples from the 3 seasons were retested by EIA. In 2005–2006, 438 of 479 (91.4%) were confirmed as positive, as were 493 of 562 (87.7%) in 2006–2007 and 179 of 207 (86.5%) in 2007–2008, yielding a total of 1110 EIA confirmed rotavirus-positive specimens. From this group, 996 samples were G-typed and 623 were tested for both G- and P-types (Tables 1 and 2).
From G-typing results, G1 was the most common strain in the 2005–2006 and 2006–2007 seasons, with overall detection rates of 51.2% and 78.7%, respectively (Table 1). Moreover, G1 was the most prevalent genotype detected in 5 of 6 sites (83%) in 2005–2006 and in 10 of 12 (83%) sites in 2006–2007, with 9 of the sites from the 2006–2007 season demonstrating a G1 prevalence of 70% or greater. In the 2007–2008 season, G1 was the most prevalent genotype in only 3 of 7 (43%) sites and the overall detection rate of the G1 genotype decreased to 30.7%. The decrease in G1 prevalence was particularly apparent in 2 cities when the 2006–2007 and 2007–2008 seasons were compared; the G1 prevalence decreased from 92.0% to 4.3% in Forth Worth, TX and from 92.5% to 6.1% in Omaha, NE. Despite the reduction in overall G1 detection rates, G1 was still the first or second most common genotype at all sites during the 2007–2008 season.
G3 was the most frequently detected G-type in the 2007–2008 season, increasing from an overall detection rate of 1.2% in 2005–2006 and 1.6% in 2006–2007 to 36.3% in 2007–2008 (Table 1). G3 was detected at all sites in 2007–2008 with a prevalence of 10% or greater at 5 of 7 sites. In Forth Worth, TX and Omaha, NE, G3 appeared to almost totally replace G1 (95.7% and 90.9% of G3 prevalence, respectively).
The percentage of G2 genotype specimens remained approximately the same in the 2006–2007 and 2007–2008 seasons (14.2% and 12.8%, respectively) though this was reduced compared with the 28.1% prevalence seen in the 2005–2006 season (Table 1). G9 was detected infrequently except in Seattle, WA where it made up 18.8% of the samples in 2005–2006 and 39.3% in 2007–2008 (Table 1). Genotype G12, which was the third most prevalent G-type during the 2005–2006 season with a prevalence of 6.2%, declined in the 2 later seasons to 1% or less. During the 2007–2008 season, it was detected only at the same 2 sites where it had been detected in 2005–2006. Genotype G4 was rare, appearing infrequently in the 2005–2006 and 2006–2007 seasons and was not detected at all in specimens from the 2007–2008 season. The number of nontypeable samples for G remained low, ranging from 1.8% to 3.9% of samples over the 3 seasons. The frequency of mixed G infections declined with each season from 6.2% to 2.4% to 0.5%.
For these 3 rotavirus surveillance seasons, G- and P-typing was attempted on 623 specimens, including all EIA-confirmed specimens from the 2007–2008 season and a subset from the earlier seasons (Table 2). G1P comprised 85% or more of the G1 strains in all 3 seasons. G2P was the most common G2 strain in all 3 seasons and comprised all G2 strains in the 2007–2008 season. During 2007–2008, G3P was the most common strain (35.8%) among the samples, followed by G1P (29.6%), G9P (14.5%), and G2P (12.8%). Uncommon strains observed during the 3 surveillance seasons were G2P, G1P, G2P, G4P, G1P, G3P, G12P, and G12P (Table 2). The rare strain G2P was observed in 1 city, Boston, MA, during the 2006–2007 season but was not detected in other cities or seasons. The frequency of nontypeable samples for P decreased consistently with each season going from 5.7% to 3.6% to 2.2% of the total number for which P typing was attempted. Mixed P-type infections were identified in 2.8% of samples in 2005–2006 and 3.3% in 2006–2007, but were not detected in 2007–2008. The percentage of samples for which G and P were nontypeable ranged from 0.7% to 2.2% over the 3 seasons.
We examined the age distribution of the patients under 3 years of age with EIA-confirmed rotavirus gastroenteritis (n = 919) to identify changes over time (Table 3). The mean age of rotavirus cases among young children in the 2007–2008 season (17.73 months) was significantly higher than the mean ages in the 2005–2006 and 2006–2007 seasons (13.1 and 13.26 months, respectively, P < 0.05, Dunn Multiple Comparison Post-Test) and the same significant trend was seen in the subset of samples that were both G- and P-typed (Table 3). In addition, the mean age of G2P rotavirus gastroenteritis cases in the 2007–2008 season (23.77 months) was significantly higher than the mean ages in the 2005–2006 (11.17 months, P < 0.001) and 2006–2007 seasons (15.21 months, P < 0.05). The mean age of G1P cases did not change significantly over the 3 seasons. The small number of non-G1P and non-G2P cases in the 2005–2006 and 2006–2007 seasons did not allow for cross-season analyses of other genotypes. When mean case ages associated with specific genotypes within a given season were analyzed, the mean age associated with genotype G2P cases (11.17 months) was significantly lower than that of G1P cases (14.04 months, P < 0.01, Mann Whitney U Test, 2-tailed) within the 2005–2006 season. Within the latter 2 seasons, however, significant differences in mean age of cases associated with specific genotypes were not observed.
The aim of this study was to examine whether or not we could detect any changes in the distribution of rotavirus genotypes following the widespread introduction of a new vaccination program, which has already had a major impact in reducing hospitalizations for rotavirus disease. Our results do not provide a definitive answer but do indicate that continued surveillance is definitely warranted.
For the baseline season, 2005–2006, and the first season following introduction, 2006–2007, no difference in strain distribution was apparent but some fluctuations were seen at individual sites. G1, the universally common strain, remained the predominant strain in most locations. However, during the 2007–2008 season, a less common variant, G3, was the predominant strain at 2 sites and genotypes other than G1 (ie, G2, G3, G9) were most common at 4 of 7 sites. This could well represent normal year-to-year variation so further surveillance will be necessary. However, this emergence of G3 as the predominant strain is of some particular interest for 2 reasons. It has been demonstrated that in infants vaccinated with 3 doses of RotaTeq, seroconversion rates were lowest against the G3 reassortant strain of the vaccine,16 thus providing a possible immunologic mechanism for G3 emergence following RotaTeq introduction. Also in Australia, where 3 states have introduced the RotaTeq vaccine while 3 other states and 2 territories use another live-attenuated rotavirus vaccine, Rotarix (GlaxoSmithKline, Research Triangle Park, NC; discussed later in the text), G3 strains tended to be more prevalent in RotaTeq states during the 2007–2008 season compared with Rotarix states and territories.10,11 In the next rotavirus season (2008–2009), G1P predominated in 1 RotaTeq state with strain data and both G1P and G3P were prevalent in the other, whereas G2P strains tended to dominate in states/territories using Rotarix.12 While the emergence of G3 strains in certain states in Australia and some locations in the United States after RotaTeq introduction is intriguing, it should be noted that this strain shift has occurred in the context of dramatic declines in severe rotavirus disease in both countries.3–5,23 Furthermore, a postlicensure case-control study conducted in a Texas hospital in 2008 when G3 accounted for almost 50% of severe rotavirus cases showed that a full 3-dose RotaTeq series was 85% to 89% protective against severe rotavirus disease.24 Thus, additional seasons of monitoring are needed to help assess the significance of the observed trends in circulating strains.
Analysis of our surveillance data indicted that there was an upward shift in the mean age of cases under 36 months of age compared with the previous 2 seasons and this could be a direct result of increased vaccination coverage with RotaTeq. Starting in 2006, vaccination of infants at 2, 4, and 6 months of age1 would reduce the number of susceptible individuals in younger age groups, resulting in both a decrease in the overall number of rotavirus gastroenteritis cases and an increase in the mean age of case in the 2007–2008 season. Children older than 22 months of age at the start of the 2007–2008 season would not have been eligible to receive RotaTeq vaccine because they were already beyond the maximum age for starting the series when vaccine first became available. The rotavirus transmission model formulated by Pitzer et al25 predicted that higher vaccine coverage would result in an increase in average age of severe diarrhea cases. The mean age shift observation for G2P cases in this study also supports data which indicate that RotaTeq is effective against G2P rotavirus strains even though this vaccine does not contain the P antigen.16,26
A theoretical concern with RotaTeq vaccine is the potential that the vaccine strains themselves may either cause disease or reassort with wild-type rotavirus to produce a virulent strain. Acute gastroenteritis associated with infection by a RotaTeq strain reassortant has been reported.27 The occurrence of such events in this study cannot be specifically ruled out because we did not screen samples for WC3-derived RotaTeq strain components nor did we examine genes other than VP4 and VP7. Development of PCR-based assays for detection of vaccine strains and vaccine strain components is currently underway at CDC.
The other live, attenuated rotavirus vaccine, Rotarix, is currently licensed in the United States, but was not available during this study period. Rotarix, approved for use in April 2008, is a live, attenuated vaccine based on a genotype G1P human strain, which relies on a heterotypic immune response to protect against other human rotavirus strains. As mentioned earlier, Australian states and territories using Rotarix vaccine reported a predominance of genotype G2P, which shares neither a G or P antigen with the vaccine.12 G2P strains have also predominated in Brazil following the introduction of Rotarix.9,28 However, in both Australia and Brazil, postlicensure case-control studies have shown good effectiveness of Rotarix vaccine against severe rotavirus disease caused by G2P strains.29 Genotype G2P was the second most prevalent strain in the United States in the 2005–2006 and the 2006–2007 seasons, accounting for over 50% of the rotavirus positive stool samples per season in some cities. As use of both rotavirus vaccines continues in the United States, surveillance will be critical to monitor for potential selection of strains through vaccine pressure.
While we can now characterize over 94% of all strains, some strains still remain nontypeable and additional laboratory investigations will be needed to improve future genotyping studies. In the 2006–2007 season, the number of variant P strains that required genotyping by sequencing increased. These P strains had multiple mutations in the 1T1 and 1T1-V primer binding sites that precluded their amplification with our standard genotyping primers. Other studies have documented extensive variation in the P genotyping primer binding sites.21,30–33 Though no revisions of the standard multiplexed RT-PCR primers were made during this study, it would be useful for future studies to develop new primers for P-typing these variant P strains.
Since hospitalizations for rotavirus in the United States have decreased markedly in the past 2 years, we will need to increase our collection of stool specimens from a greater number of hospital laboratories to have an adequate number of specimens to understand the national patterns. The number of hospital laboratories our data to not reflect fully the overall prevalence and diversity of circulating US rotavirus strains. In the future, it would be beneficial to increase the number of participating hospital laboratories in the National Rotavirus Strain Surveillance System and retain these study sites from year to year. Expanded surveillance and some investigation of vaccination status among those children hospitalized for rotavirus diarrhea with unusual strains could greatly enhance our understanding of interaction of the vaccine with strain selection and replacement.
In closing, this study lays out some clear directions for the future of US rotavirus surveillance and raises important questions that need to be addressed as rotavirus vaccination programs continue, both in the United States and abroad.
The authors thank Slavica Mijatovic-Rustempasic and Rashi Gautam for their reviews of this manuscript and helpful comments.
1.Cortese MM, Parashar UD. Prevention of rotavirus gastroenteritis among infants and children: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep
2.Anonymous. Prevention of rotavirus disease: guidelines for use of rotavirus vaccine. Pediatrics.
3.Tate JE, Panozzo CA, Payne DC, et al. Decline and change in seasonality of US rotavirus activity after the introduction of rotavirus vaccine. Pediatrics
4.Cortese MM, Tate JE, Simonsen L, et al. Reduction in gastroenteritis in United States children and correlation with early rotavirus vaccine uptake from national medical claims databases. Pediatr Infect Dis J
5.Curns AT, Steiner CA, Barrett M, et al. Reduction in acute gastroenteritis hospitalizations among US children after introduction of rotavirus vaccine: analysis of hospital discharge data from 18 US states. J Infect Dis
6.Anonymous. Delayed onset and diminished magnitude of rotavirus activity—United States, November 2007–May 2008. MMWR Morb Mortal Wkly Rep
7.Leite JP, Carvalho-Costa FA, Linhares AC. Group A rotavirus genotypes and the ongoing Brazilian experience: a review. Mem Inst Oswaldo Cruz
8.Gurgel RQ, Correia JB, Cuevas LE. Effect of rotavirus vaccination on circulating virus strains. Lancet
9.Gurgel RQ, Cuevas LE, Vieira SCF, et al. Predominance of rotavirus PG2 in a vaccinated population, Brazil. Emerg Infect Dis
10.Kirkwood CD, Cannan D, Boniface K, et al. Australian Rotavirus Surveillance Program annual report, 2007/08. Commun Dis Intell
11.Matthijnssens J, Bilcke J, Ciarlet M, et al. Rotavirus disease and vaccination: impact on genotype diversity. Future Microbiol
12.Kirkwood CD, Boniface K, Bishop RF, et al. Australian Rotavirus Surveillance Program annual report, 2008/2009. Commun Dis Intell
13.Das BK, Gentsch JR, Cicirello HG, et al. Characterization of rotavirus strains from newborns in New-Delhi, India. J Clin Microbiol
14.Gentsch JR, Glass RI, Woods P, et al. Identification of group-A rotavirus gene-4 types by polymerase chain-reaction. J Clin Microbiol
15.Gouvea V, Glass RI, Woods P, et al. Polymerase chain-reaction amplification and typing of rotavirus nucleic-acid from stool specimens. J Clin Microbiol
16.Vesikari T, Matson DO, Dennehy P, et al. Safety and efficacy of a pentavalent human-bovine (WC3) reassortant rotavirus vaccine. N Engl J Med
17.Griffin DD, Kirkwood CD, Parashar UD, et al. Surveillance of rotavirus strains in the United States: identification of unusual strains. J Clin Microbiol
18.Freeman MM, Kerin T, Hull J, et al. Enhancement of detection and quantification of rotavirus in stool using a modified real-time RT-PCR assay. J Med Virol
19.Esona MD, Steele D, Kerin TK, et al. Determination of the G and P types of previously non-typeable rotavirus strains from the African Rotavirus Network from 1996–2004: identification of unusual G types. J Infect Dis.
20.Estes MK, Cohen J. Rotavirus gene structure and function. Microbiol Rev
21.Simmonds MK, Armah G, Asmah R, et al. New oligonucleotide primers for P-typing of rotavirus strains: strategies for typing previously untypeable strains. J Clin Virol
22.Esona MD, Geyer A, Page N, et al. Genomic characterization of human rotavirus G8 strains from the African rotavirus network: relationship to animal rotaviruses. J Med Virol
23.Lambert SB, Faux CE, Hall L, et al. Early evidence for direct and indirect effects of the infant rotavirus vaccine program in Queensland. Med J Aust
24.Boom JA, Tate JE, Sahni LC, et al. Effectiveness of pentavalent rotavirus vaccine in a large urban population in the United States. Pediatrics
25.Pitzer VE, Viboud C, Simonsen L, et al. Demographic variability, vaccination, and the spatiotemporal dynamics of rotavirus epidemics. Science
26.Vesikari T. Rotavirus vaccines. Scand J Infect Dis
27.Payne DC, Edwards KM, Bowen MD, et al. Sibling transmission of vaccine-derived rotavirus (RotaTeq) associated with rotavirus gastroenteritis. Pediatrics
28.Nakagomi T, Nakagomi O. A critical review on a globally-licensed, live, orally-administrable, monovalent human rotavirus vaccine: Rotarix. Expert Opin Biol Ther
29.Correia JB, Patel MM, Nakagomi O, et al. Effectiveness of monovalent rotavirus vaccine (Rotarix) against severe diarrhea caused by serotypically unrelated G2P strains in Brazil. J Infect Dis
30.Iturriza-Gomara M, Green J, Brown DW, et al. Diversity within the VP4 gene of rotavirus P strains: implications for reverse transcription-PCR genotyping. J Clin Microbiol
31.Iturriza-Gomara M, Kang G, Gray J. Rotavirus genotyping: keeping up with an evolving population of human rotaviruses. J Clin Virol
32.Lee JI, Song MO, Chung JY, et al. Outbreak of rotavirus variant P in Seoul, South Korea. J Med Virol
33.Fischer TK, Eugen-Olsen J, Pedersen AG, et al. Characterization of rotavirus strains in a Danish population: high frequency of mixed infections and diversity within the VP4 gene of P strains. J Clin Microbiol
Keywords:© 2011 Lippincott Williams & Wilkins, Inc.
RotaTeq; rotavirus; strain; surveillance; US