Global Trends in Norovirus Genotype Distribution among Children with Acute Gastroenteritis

Noroviruses are a leading cause of acute gastroenteritis (AGE) among adults and children worldwide. NoroSurv is a global network for norovirus strain surveillance among children <5 years of age with AGE. Participants in 16 countries across 6 continents used standardized protocols for dual typing (genotype and polymerase type) and uploaded 1,325 dual-typed sequences to the NoroSurv web portal during 2016–2020. More than 50% of submitted sequences were GII.4 Sydney[P16] or GII.4 Sydney[P31] strains. Other common strains included GII.2[P16], GII.3[P12], GII.6[P7], and GI.3[P3] viruses. In total, 22 genotypes and 36 dual types, including GII.3 and GII.20 viruses with rarely reported polymerase types, were detected, reflecting high strain diversity. Surveillance data captured in NoroSurv enables the monitoring of trends in norovirus strains associated childhood AGE throughout the world on a near real-time basis.

Noroviruses are a leading cause of acute gastroenteritis (AGE) among adults and children worldwide. Noro-Surv is a global network for norovirus strain surveillance among children <5 years of age with AGE. Participants in 16 countries across 6 continents used standardized protocols for dual typing (genotype and polymerase type) and uploaded 1,325 dual-typed sequences to the Noro-Surv web portal during 2016-2020. More than 50% of submitted sequences were GII. 4  open reading frame [ORF] 2), which is also the primary neutralization site for antibodies produced after norovirus infection or vaccination (13)(14)(15). Noroviruses are classified into 10 genogroups, GI-GX, and >48 genotypes: 9 genotypes in the GI genogroup, 26 in GII, 3 in GIII, 2 in GIV, 2 in GV, 2 in GVI, 1 in GVII, 1 in GVIII, 1 in GIX, and 1 in GX (16). ORF1 encodes the viral nonstructural proteins including the polymerase, which is classified into >60 polymerase types (P-types) (16). Much about the evolutionary role of recombination among noroviruses, which occurs primarily at the ORF1/ORF2 junction, remains unknown (17)(18)(19). Norovirus classification was recently updated to include typing of the polymerase region (16). This dual typing strategy considers the genotype encoding the major capsid protein and the P-type encoding the polymerase region (16). A short genomic region spanning the 3′ end of the polymerase gene through the 5′ end of the capsid gene is the basis for sequence-based dual typing (20).
Genogroup II genotype 4 (GII.4) viruses have been the most frequently detected noroviruses globally since the mid-1990s, before which GII.3 viruses were dominant (13,21,22). New GII.4 variants regularly emerge and spread across the globe and often contribute to increased illness and death, especially in healthcare settings (23)(24)(25). During 2002-2012, new GII.4 variants with antigenically distinct capsid epitopes, which enable the viruses to escape neutralizing antibodies, emerged and replaced previous variants every 2-3 years (15). These changes indicate that norovirus vaccines might need to be updated regularly. Despite recent recombination events resulting in the global spread of GII.4 Sydney viruses with a novel P16 polymerase, no new variant causing widespread infections has emerged since 2012 (20,26,27). Although GII.4 strains are the most common strains detected among all age groups, non-GII.4 strains, such as GII.2, GII.3, and GII.6 viruses, are common causes of sporadic cases and illness in young children (6,(28)(29)(30)(31)(32). Rare strains (4) and GII.4 variants can circulate, especially among children, for years before spreading globally among all age groups (33,34). Consequently, children might be an important reservoir for emerging norovirus strains against which little or no population immunity exists.
NoroSurv (https://www.norosurv.org), which is maintained by the Centers for Disease Control and Prevention (Atlanta, Georgia, USA), is a global pediatric norovirus strain surveillance network for children <5 years of age with medically attended AGE and can only be accessed by registered NoroSurv laboratories. Surveillance of norovirus strains infecting children is crucial for monitoring the emergence of new or rare strains and for developing vaccines that protect against the most common strains.

NoroSurv
All but 2 participating hospitals and medical centers collected norovirus-positive stool samples from children with AGE; 2 sites in Nicaragua and Australia obtained only samples from symptomatic children in community-based studies. Staff at hospitals, medical centers, universities, and reference laboratories processed and typed the samples. Each laboratory uploaded norovirus sequences; patient demographic data (e.g., deidentified patient age and sex); and information on sample type, collection date, and setting to the password-protected NoroSurv web portal. All laboratories used a standardized protocol for norovirus dual typing that comprised screening by genotype-specific real-time reverse transcription PCR (RT-PCR), conventional RT-PCR, and Sanger sequencing of RT-PCR products (20). Raw DNA chromatogram files or nucleotide sequences were automatically typed by NoroSurv using the most recent reference sequences and classification for noroviruses (16). Ethics approval by the New Zealand component of this study was granted by the Health and Disability Ethics Committee, New Zealand (approval no. 19/CEN/96).

Data Analysis
We analyzed NoroSurv data associated with samples collected during September 1, 2016-August 31, 2020. We excluded samples from children >5 years of age, from asymptomatic patients, or that had missing or low-quality dual typing information. We downloaded sequences and associated data from NoroSurv; we then aggregated, cleaned, analyzed, and visualized the data using R software (The R Project, https://www.r-project.org). After downloading the sequences from NoroSurv as fasta files, we checked the quality of the submitted sequences using Bioconductor (http://bioconductor.org) packages in R. When discrepancies existed between the manually entered and autotyped information, we conducted phylogenetic analysis to confirm the correct type; we updated NoroSurv records accordingly.

Results
A total of 1,325 dual-typed norovirus sequences collected during September 2016-August 2020 from children <5 years of age with AGE were submitted to NoroSurv. Sequences were received from 19 sites in 16 countries in Africa (South Africa, n = 13), Asia (Bangladesh, n = 32; Hong Kong, China, n = 326; India, n = 36; Japan, n = 89; the Philippines, n = 132; and Taiwan, n = 19), Oceania (Australia, n = 71; New Zealand, n = 54), Europe (Germany, n = 111 and Spain, n = 44), North and Central America (Canada, n = 90; Nicaragua, n = 78; and the United States, n = 173), and South America (Brazil, n = 14 and Chile, n = 43) ( Figure 1). Each country submitted a median of 63 sequences (range 13-326); 48% of sequences were from countries in Asia. We excluded 62 of the 1,387 Noro-Surv sequences: 11 that could not be typed because of poor sequence quality or missing fasta files; 7 with sample collection dates before September 1, 2016; 31 from children >5 years of age; and 13 from asymptomatic children.
To compare genotype distribution over time, we defined seasons as September 1-August 31; these periods reflected the seasonality reported for noroviruses, with peak cases often occurring during the cooler months: October-March in the Northern Hemisphere and April-September in the Southern Hemisphere   (Figure 2; Appendix Table  1) and country (Appendix Tables 2-17). The most   Table  1). We detected 149 GII.2 viruses, most (148; 99%) of which were P16. All 64 GII.6 viruses were P7 ( Figure  3; Appendix Table 1).
The 5 most frequently detected dual types were GII. 4

Discussion
We used NoroSurv data to monitor global trends in norovirus genotypes causing sporadic AGE in children <5 years of age. These children would probably benefit most from norovirus vaccines and are a critical group for evaluating future vaccine effectiveness. Although the number of sequences submitted from different countries varied during 2016-2020 on Noro-Surv, the overall surveillance from 16 countries across 6 continents identified common genotypes around the world. Standardized protocols for dual typing across all NoroSurv sites enabled global comparisons, surveillance, and detection of recombinant strains.
During 2016-2020, NoroSurv documented 22 genotypes of norovirus causing illness in young children. GII. 4 Sydney viruses, which globally are the most common among all age groups (13,26), comprised >50% of all NoroSurv sequences. GII.2, GII.3, and GII.6 viruses, which are leading causes of childhood norovirus infections but less common among adults, were also frequently detected (6,28,(30)(31)(32). One study found that among children with sporadic AGE, GII.6 viruses were second most common after GII.4; however, GII.13 viruses were the second most common cause of noroviruses outbreaks in adults (6). Although ≈5% of reported norovirus outbreaks in the United States are caused by GII.3 viruses (17,20), we found they comprised nearly 23% of sporadic cases among children. Thus, GII.2, GII.3, and GII.6 viruses appear to be major causes of AGE in children but might be less transmissible to adults. This lack of transmissibility might be caused by virus-specific properties or long-term immunity in adults after childhood infection. Norovirus vaccines in development focus on the major capsid protein, which is also the genomic region used for genotyping (14,16). Vaccine candidates should protect against a broad diversity of genotypes and be easily adapted to emerging genotypes or GII.4 variants. Noroviruses contribute substantially to the prevalence of diarrheal disease among children (1), causing more severe illness and death in resource-limited countries (30,36). Childhood vaccination might reduce norovirus prevalence among children. If vaccination prevents transmission, then it also might reduce infections among all age groups (6,12).
GII.4 Sydney viruses, primarily associated with P31 and P16 polymerases, were responsible for most norovirus cases during 2012-2019 (33,34). Recombination at the ORF1/ORF2 junction is a common occurrence among noroviruses and contributes to norovirus evolution, although the exact mechanism is poorly understood (17)(18)(19). Acquisition of a novel P16 polymerase did not result in emergence of a novel GII.4 variant or substantial changes to the antigenic region of the capsid (17,37,38). However, changes to the polymerase or other nonstructural proteins might have increased the replicative or transmission fitness of GII.4 viruses (17,18,26).
Overall, 36 dual types were detected in NoroSurv and several genotypes were associated with >1 P-type. GII.3 viruses were primarily associated with P12, but many had P21, P16, and the rare P30 and PNA3 polymerases, indicating a high propensity for recombination among GII.3 strains. Other rarely detected strains included GII.20[P20] and GII.20 [P7]. Several rare and novel norovirus genotypes have been detected only in children (4), suggesting differences in children's and adults' susceptibility to certain strains. We identified a subcluster of GII.4 Sydney (GII.4 untypable) viruses in 6 countries spanning 4 continents during 2017-2019. Complete ORF2 sequences for this strain are needed to analyze possible changes in the antigenic region of the capsid, which could enable viruses to escape antibody neutralization. If such changes exist, or if strains within this subcluster continue to evolve and spread globally, a new GII.4 variant could emerge. A recent study reported that GII.4 variants can begin to circulate, especially among children, for up to 9 years before emerging globally (33). Low-level circulation enables accumulation of mutations and emergence of new strains (18,38) and access to niches in the host environment, thereby promoting spread (33); thus, children might be a reservoir for the recombination and evolution of noroviruses. This concern highlights the necessity of norovirus surveillance among children.
NoroSurv complements NoroNet (34), a wellestablished global network for norovirus surveillance that has illuminated global trends in norovirus strain diversity, recombination, and evolution, including tracking the emergence of novel GII.4 variants. NoroSurv sequences are derived from sporadic cases among children, whereas NoroNet includes sequences from outbreaks and sporadic cases in adults and children. NoroSurv requires standardized protocols for dual typing (20) across all sites to ensure global comparability. However, NoroNet, which was established in 1999, has a much longer history than NoroSurv. Because the importance of dual typing was not well recognized at the time NoroNet was established, many of its sequences are derived from either the polymerase or capsid genes, but not both. In addition, the NoroSurv web portal incorporates a unique automatic typing tool and an internal dashboard of all data by location. In 2021, we plan to make the dashboard publicly available for near real-time data on global trends in sporadic norovirus infections in children.
NoroSurv is a passive surveillance system comprised of voluntary submissions from participating laboratories. As a result, its data do not necessarily correlate with national surveillance records. Furthermore, the number of sequences submitted from each country varies; this number depends in part on the availability of resources such as time and laboratory capacity. Low-income countries are currently underrepresented in NoroSurv, as are countries in Africa and Central America. However, a recent review of norovirus genotypes detected in 8 low-income and 21 low-to-middle income countries showed that GII.4 viruses were the most common genotype, with substantial proportions of GII.3 and GII.6 viruses; in addition, GI.3 viruses were the most commonly detected GI viruses (30). Trends in the genotype distribution of noroviruses in these countries resembled the global trends illuminated in NoroSurv. In future years, NoroSurv aims to expand of the number of countries, sites, and submissions.
The 2019-20 norovirus season coincided with the emergence of the coronavirus disease pandemic, which has limited the capacity and resources for norovirus surveillance. In addition, it is unknown whether the global lockdowns, including school and daycare closures; physical distancing; and heightened hygiene awareness and practices such as handwashing, disinfection, and wearing of face masks (39), will reduce norovirus transmission among children. When settings prone to norovirus outbreaks (e.g., childcare facilities and schools) return to prepandemic capacities, norovirus cases might increase, especially if the use of alcohol-based hand sanitizers, which have limited efficacy against noroviruses (40), are substituted for handwashing in these settings. Although submissions to NoroSurv declined during February-August 2020, users might upload sequences retrospectively. As a result, data for the 2019-2020 season might not fully reflect global trends.
NoroSurv enables the near real-time detection of global norovirus genotype trends and diversity among children <5 years of age with AGE. Our findings support previous research indicating that although some overlap exists between the genotypes detected in children and adults, genotypes such as GII.2, GII.3, and GII.6 are more common among children. Childhood norovirus vaccination will probably reduce the prevalence of norovirus associated AGE among children and interrupt community transmission among all age groups (12). As such, researchers should ensure that candidate vaccines are protective against strains commonly seen in children or produce sufficient cross-protective immunity against those strains. Surveillance of rare genotypes, recombinant strains, and potentially new GII.4 variants can better predict the emergence of new strains, guiding potentially updated vaccine formulation. Sequencing larger regions of the genome, particularly the major capsid gene, can help identify antigenic changes that might enable the virus to escape antibody neutralization, which provides important information for predicting strain emergence and updating vaccine formulations. The continued expansion of the NoroSurv network to include countries with geographic and economic diversity will enhance our understanding of norovirus infections among children worldwide. NoroSurv surveillance will inform efforts to develop and adapt norovirus vaccine candidates; it will also aid in the evaluation of future vaccine efficacy by documenting baseline global strain diversity of noroviruses in children.