Skip to main content

Our take —

This study described an outbreak of the SARS-CoV-2 Delta variant that was associated with an indoor gymnastics facility in Oklahoma in the late spring of 2021. This cluster included 47 cases among 194 exposed individuals; two patients were hospitalized, and there were no fatalities. Vaccine coverage among those who were exposed was low (in part, due to vaccine eligibility guidelines), and a small number of vaccine breakthrough infections were observed. This study demonstrates that the Delta variant is highly transmissible in indoor sports settings and within households. Current evidence indicates that vaccines remain effective against the Delta variant. Multicomponent prevention strategies (i.e., masking, social distancing, quarantine) remain important for limiting transmission, especially in high-risk settings with low rates of vaccination and high prevalence of the Delta variant.

Study design

Other

Study population and setting

This study described an outbreak of SARS-CoV-2 Delta variant cases associated with an Oklahoma gymnastics facility between April 15 and May 3, 2021. This cluster of cases was initially identified by the Oklahoma Department of Health, Acute Disease Service through routine surveillance and sequencing. Contact tracing was performed to identify potential exposures and linked cases. Team and staff rosters from the Oklahoma facility and participant lists from two regional meets that took place during this time period were compared to Oklahoma’s COVID-19 surveillance data to identify additional linked cases. State immunization records were used to confirm vaccination status of all cases and contacts. Available specimens were sequenced to confirm variant status.

Summary of Main Findings

This cluster of SARS-CoV-2 Delta variant infections included 47 cases (median patient age, 14 years; range, 5-58 years) among 194 identified exposed persons between April 15 and May 3, 2021 (overall attack rate: 24%). Cases included 23 gymnasts (from 10 of 16 cohorts), 3 staff members, and 21 household contacts (from 7 of 26 interviewed households). Two adults were hospitalized, one of which required treatment in the ICU; neither were vaccinated. At the time, 74 (38%) of the exposed persons were eligible for vaccination based on local guidelines, but only 17 (9%) were fully vaccinated. Four vaccine breakthrough cases were identified in this cluster (Moderna and Pfizer); three additional cases were identified in individuals who had received only one vaccine dose. Despite the participation of 4 infected gymnasts in two regional meets during the contagious period, there was no evidence of secondary spread at either event. Of the confirmed cases, 21 were attributed to the Delta variant based on sequencing results, while 26 were attributed based on direct epidemiologic linkage to another variant cases. Environmental risk factors identified during household interviews included quarantine nonadherence, low rates of testing, delayed infection recognition due to mild symptoms, lack of mask wearing, poor ventilation, overlapping cohorts, and inadequate cleaning.

Study Strengths

This study describes a SARS-CoV-2 Delta variant outbreak associated with an indoor gymnastics facility. Cases were identified through robust contact tracing efforts, and results were contextualized based on known vaccination status.

Limitations

Only cases reported to the state surveillance system were included in counts; additional exposed persons with mild or asymptomatic infection may have not been tested for SARS-CoV-2, leading to underestimates of case counts. Voluntary patient interviews could have led to underreporting of contacts and infection related details. Delays in updates to the state vaccine database could have resulted in underestimation of vaccination rates. Less than half of outbreak cases had samples available for sequencing to confirm variant status. Vaccine effectiveness could not be determined because not all exposed individuals could be interviewed and because state vaccination data was incomplete.

Value added

This is a thorough description of a cluster of SARS-CoV-2 Delta variant infections that was associated with a gymnastics facility in Oklahoma in the late spring of 2021.

Our take —

This study, available as a preprint and thus not yet peer-reviewed, investigated six symptomatic COVID-19 cases in vaccinated individuals after an outdoor wedding in Houston, Texas in April 2021. All patients were infected with the Delta variant, and phylogenetic analysis confirms likely transmission during the wedding event. However, limited data on time since vaccination (or antibody titers) for these patients makes it difficult to confirm how many were true vaccine breakthroughs, and the total number of asymptomatic infections acquired at the wedding is unknown. Breakthrough infections should continue to be monitored, especially as Delta or other new variants increase in frequency. However, it is important to remember that vaccine breakthroughs are to be expected and do not suggest vaccine ineffectiveness; in fact, vaccination may have protected some of these patients (and other wedding guests) from more severe outcomes.

Study design

Other

Study population and setting

This study describes SARS-CoV-2 infections acquired at a wedding in Houston, Texas with 92 attendees in April 2021. All wedding activities occurred outside in a large open-air tent, and attendees were required to be fully vaccinated. Two guests (Patients 0a and 0b) travelled from India to attend the wedding and tested positive four days after the wedding. Patient 0b complained of fatigue the first night of the wedding but associated it with jet lab. Patient 0a developed a cough two days after the wedding, and both developed a fever three days after the wedding. They had both received their second dose of the Covaxin BBV152 vaccine 10 days prior to traveling to Texas. Five other guests reported close contact with patients 0a and 0b, and four tested positive for SARS-CoV-2. Two had been vaccinated with the Pfizer vaccine, and two had been vaccinated with the Moderna vaccine, though vaccination dates for these patients were not reported. All six positive patients reported symptoms, and at least two required hospitalizations. Patient 0a was one of those hospitalized, and he died approximately one month after the wedding from complications of COVID-19.

Summary of Main Findings

Whole genome sequencing revealed that all six cases were caused by the SARS-CoV-2 Delta variant. Phylogenetic analysis revealed that the cases were closely related (viruses formed a distinct cluster within the Delta clade of the phylogenetic tree). The authors characterize all six infections as vaccine breakthroughs due to reported vaccination status of all patients.

Study Strengths

In this study, the phylogenetic and epidemiological data both suggest the same conclusion: that transmission of SARS-CoV-2 occurred at a wedding attended by all six patients. The study also provided detailed information on patient demographics and comorbidities, and includes travel history of two patients from India, indicating the likely source of initial infection.

Limitations

The manuscript does not provide any details on testing of other wedding attendees, including those that may have been asymptomatic, limiting our understanding of the true number of attendees that may have been infected. Additionally, the manuscript does not discuss the severity of symptoms in patients other than 0a and 1. Finally, the paper states that Patients 0a and 0b received their second vaccine dose only 10 days prior to travel, which may have not been long enough for them to mount a full immune response. Therefore, only up to four of these cases may have been true vaccine breakthrough infections (though time since vaccination was not reported for the other patients).

Value added

This manuscript presents additional evidence of symptomatic infections — particularly of the Delta variant — in vaccinated individuals.

Our take —

The Virus Evolution Working Group of the World Health Organization established a new naming system specifically for SARS-CoV-2 variants of interest and concern based on the Greek alphabet. New variant labels will be designated by a centralized global authority (the World Health Organization), will avoid names associated with people or places, and will avoid complex alphanumeric designations that are difficult to remember and prone to error. The WHO names for the current VOCs are: Alpha (UK variant, B.1.1.7), Beta (South Africa variant, B.1.351), Gamma (Brazil variant, P.1), and Delta (India variant, B.1.617.2).

Study design

Other

Study population and setting

This report announces a new naming scheme for SARS-CoV-2 variants of interest and concern. It was published by a group convened and led by the Virus Evolution Working Group of the World Health Organization. This working group specifically tracks variants that may provide a selective advantage to the virus (variants of interest and concern; VOI and VOC) and identified four challenges in naming these variants for the purposes of global tracking: (1) there are three existing nomenclature systems for naming SARS-CoV-2 lineages, none of which was designed specifically to track VOIs and VOCs. (2) There is no central authority responsible for naming these variants, resulting in confusion regarding which nomenclature system to use. (3) The existing alphanumeric naming schemes are confusing and liable to misreporting and misunderstanding. (4) Using the name of the place where a VOI or VOC was first identified as its name is both misleading and potentially stigmatizing, and should be avoided.

Summary of Main Findings

The group convened by the Virus Evolution Working Group established a naming system for officially designated VOIs and VOCs based on Greek alphabet letters. These labels will be linked to the existing three genetic-based nomenclature systems and updates will be published regularly by the World Health Organization (WHO). The WHO names for the current VOCs are: Alpha (UK variant, B.1.1.7), Beta (South Africa variant, B.1.351), Gamma (Brazil variant, P.1), and Delta (India variant, B.1.617.2).

Study Strengths

NA

Limitations

NA

Value added

This report announces a new nomenclature system for SARS-CoV-2 variants of interest and concern.

Our take —

While bats are the reservoir hosts of SARS-related coronaviruses and the most likely source of the evolutionary ancestor of SARS-CoV-2, the absence of live bats from the market shops implies that other intermediate hosts may have needed to be involved in transmission to humans. The current study provides evidence that wild animal species that are susceptible to SARS-related coronaviruses were sold at markets in Wuhan prior to the emergence of SARS-CoV-2 in November 2019. However, without evidence of viral infection in these animals, their involvement in SARS-CoV-2 emergence remains unknown. Additional studies that investigate viral exposure in potential intermediate host animals being sold or transported to markets or on the farms sourcing markets and the humans in close contact with these animals during transportation and trade are needed to provide additional support for this hypothesis of SARS-CoV-2 emergence.

Study design

Other

Study population and setting

The study reports on observations of live animals sold at markets around Wuhan between May 2017 and November 2019. The surveys were performed as part of a study to identify the animal source of tick borne disease in Hubei Province. A total of 17 shops selling live animals were surveyed within four large markets in Wuhan: Huanan Seafood market, Qiyimen live animal market, Baishazhou market, and Dijiao outdoor pet market. Researchers interviewed vendors to ask about what animal species were sold, in what quantities (either individuals or by weight for some species), the price per unit sold, and the origin of animals (whether wild-caught or captive bred/farmed). The investigators also checked to see if vendors had licenses to sell livestock and wild animal products, which are required by Chinese law, and whether the permits were visible to customers.

Summary of Main Findings

Including animals sold by weight, a total of 47,381 animals were reported from the 17 shops over the time period of the surveys. Most animals were alive when sold and were kept in stacked cages. There were 38 non-domesticated animal species reported from the shops, 31 of which are protected species under Chinese law. No bat species or pangolin species were found for sale in the shops. Of the 38 species, 21 had individuals sourced directly from the wild as inferred from wounds (gunshots or leg-hold traps) or based on interviews with vendors; the remaining 17 species were sourced from farms or captive breeding operations. While 13/17 shops did have the necessary permits for selling wild animal species for food, species names were given in Chinese only and none of the shops posted the origin or quarantine certificates, so were not compliant with legal requirements. Furthermore, vendors displayed 20 species that were caught in the wild that are protected under Chinese law restricting the sale of wild-caught individuals for food, making all sales illegal.

Study Strengths

The primary strength of the study is the uniqueness of the data. The vendors were more transparent with the authors about trading practices than if the investigators had been connected to law enforcement.

Limitations

The authors did not collect samples to test for the presence of viruses in animals being sold at shops. Therefore, we have no information on whether the animals were infected or exposed to SARS-CoV-2 or related viruses. The authors also provided no data on how the numbers of animals sold per species changed over time at each of the markets.

Value added

This study provides unique data on the live animal sales occurring at markets in Wuhan in the time leading up to the first reported cases of COVID-19. Bats (Rhinolophus species) and pangolins are known to carry viruses related to SARS-CoV-2, but none of these were present at the markets. However, four species with known susceptibility to SARS-related coronaviruses were found among the species sold at the markets: raccoon dogs, hog badgers, masked palm civets, and mink. Hog badgers and masked palm civets were important intermediate hosts for SARS-CoV emergence in China in 2001. Experimental infection trials have shown that raccoon dogs are susceptible to SARS-CoV-2. Finally, farmed mink populations experienced massive outbreaks of SARS-CoV-2 in Europe and North America in 2020. While individuals of all four carnivore species sold at markets were sourced from farms, some individual raccoon dogs and hog badgers were wild-caught. The presence of these animals at the markets increases the plausibility of the hypothesis that one or more intermediate hosts may have been the source of SARS-CoV-2 in Wuhan. In such a scenario, animals carrying the virus, having been infected from a reservoir host in the wild or on a farm, were brought to the market and spread the virus to humans involved in the transportation or sale of animals, as occurred for SARS-CoV.

Our take —

This study, available as a preprint and thus not yet peer-reviewed, analyzed SARS-CoV-2 sequences belonging to the B.1.620 lineage. The authors find that this is a particularly diverse lineage with several mutations associated with antibody-induced immunity, and find evidence for local transmission of this lineage in Lithuania and other parts of Europe (including Germany, France, and Belgium). When available, the authors also used travel history metadata to conclude that this lineage likely originated in Cameroon or a nearby country in central Africa, despite limited sequencing there. This work highlights the need for global genomic surveillance to limit SARS-CoV-2 spread everywhere, as well as the importance of keeping and sharing travel history information as part of routine surveillance efforts. Evaluation of the effectiveness of existing COVID-19 vaccines against the B.1.620 lineage will be critical to understand the potential impact of these mutations on public health outcomes.

Study design

Other

Study population and setting

This study was an analysis of all publicly available whole genome sequences of SARS-CoV-2 belonging to lineage B.1.620 as of April 29, 2021 (85 sequences, after quality filtering). The authors were interested in this lineage because of the large number of mutations found in sequences belonging to it, including several mutations that have been found individually in other variants of concern, but never together. Because this lineage was found in a large cluster in Lithuania, the authors analyzed the 85 B.1.620 sequences alongside the other lineages observed there. And, because several B.1.620 cases turned out to be in travelers returning from Cameroon, the analysis also included other sequences from Cameroon showing high similarity to the B.1.620 lineage. The 85 B.1.620 sequences and 141 Lithuanian and Cameroonian sequences were subsequently combined with 150 randomly selected global sequences to understand B.1.620 in the context of global diversity.

Summary of Main Findings

The authors found that B.1.620 lineage carries several mutations and deletions that have been previously observed in other variants of concern individually, but never together. The majority of these mutations have been shown in other studies to likely escape antibody-mediated immunity. The authors also found compelling evidence for local transmission of B.1.620 in Lithuania and other parts of Europe, including Germany and France. However, a large number of cases were among travelers returning from Cameroon (7 out of 85 B.1.620 sequences) or Central African Republic (6 out of 85), which borders Cameroon. The authors concluded that these findings suggested a central African origin of the B.1.620 lineage, and detection of the lineage in neighboring countries suggests that the lineage may be circulating widely in Central Africa, despite limited genomic surveillance and detection.

Study Strengths

The study uses sophisticated phylogeographic techniques to estimate the likely origin of B.1.620 in Central Africa. This was enabled by collection of detailed travel histories of several sequenced European cases.

Limitations

As the authors themselves point out, the study is limited by the lack of metadata (specifically travel history) for the majority of B.1.620 sequences. Furthermore, limited genomic surveillance in parts of Central Africa means that the authors were not able to define the extent and geographic coverage of B.1.620 circulation in the region.

Value added

This study highlights the importance of global genomic surveillance of SARS-CoV-2. In particular, the authors describe how lack of surveillance makes it difficult to reconstruct the order of mutations that occurred during the evolution of the B.1.620 lineage, which limits our ability to understand how viral mutations affect fitness. The authors also highlight a new lineage (B.1.620) with mutations seen in other variants of concern and ensured that these sequences are categorized correctly in public databases.

Our take —

In a large study of students and staff associated with New York City Public Schools, it was found that from October 2020 to December 2020 — when both hybrid learning (3 days in person/week) and fully remote learning options were being implemented — the school-based prevalence of COVID-19 was 0.4%. The cumulative incidence in schools was 341.1 cases per 100,000 from October 9, 2020 to November 19, 2020 (some hybrid learning) and 464.4 per 100,000 from December 7, 2020 to December 18, 2020 (primarily remote learning). Cumulative incidence associated with schools was consistently lower than that in the population overall. The secondary attack rate among school-based close contacts was low. With adequate COVID-19 mitigation strategies in place, the burden of COVID-19 was no higher in schools than in the general community. Lower attendance over this period and a study period which preceded the emergence and gradual takeover of SARS-CoV-2 variants in the population both suggest that results may change under different conditions, and consistent implementation of risk mitigation strategies remains important.

Study design

Retrospective Cohort

Study population and setting

From October 12 to November 20, 2020, students within New York City public schools were given the option of either hybrid (i.e., one to three days in-person per week) or fully remote learning. Schools then went fully remote, with only elementary and special education schools reopening in December. The goal of this study was to examine SARS-CoV-2 infections in students and staff within New York City public schools between October and December 2020. Three primary analyses were conducted: estimations of 1) the prevalence of SARS-CoV-2 from testing in schools, 2) the incidence of COVID-19 from testing in both schools and the community, and 3) the secondary attack rate based on monitoring of school-based close contacts. Cases were ascertained from three different sources: 1) mandatory reporting to the Department of Health and Mental Hygiene and subsequent interviews regarding school association; 2) school-affiliated individuals voluntarily notifying the school if they tested positive; 3) routine PCR testing of asymptomatic persons at schools (at least monthly at each school). School period-prevalence was defined as the number testing positive for SARS-CoV-2 divided by the total number of tests performed for a given school week. School incidence was defined as all cases reported across the whole period associated with in-person learning divided by the number of people estimated to be present in-person during that period. School cumulative incidence was calculated for two periods during the study period in order to account for the gap created by the Thanksgiving holidays. Prevalence and incidence estimates were compared to estimates derived from the larger community.

Summary of Main Findings

Among 234,132 individuals tested between October 9 and December 18, 2021, 0.4% tested positive (n=986), representing the period prevalence from in-school testing. Among students, test positivity was highest at elementary schools (n=355, 0.54%). Test positivity among those in secondary schools was 0.23%, n=5. From October 9 to November 19 (hybrid school was an option), the cumulative incidence in schools was 341.1 cases per 100,000, compared with 528.9 per 100,000 in the general population. From December 7 to December 18 (only elementary schools open), the cumulative incidence in schools was 464.4 cases per 100,000, compared with 509.6 per 100,000 in the general population. Among 36,423 school-based close contacts of confirmed cases, 191 (0.5%) tested positive.

Study Strengths

A major strength of this study was case ascertainment from three different sources: mandatory reporting, voluntary notification, and routine testing. Routine testing was done on a sample of asymptomatic individuals, mitigating selection biases caused by only the sickest individuals going in for testing.

Limitations

Findings from this study are based on routinely collected data, and there are some data missing due to lack of data availability, lack of consent for data access, etc. These results represent a short time period, in which there was either hybrid learning or fully remote learning; they also precede the emergence of an increasing prevalence of variant B.1.1.7 or the so-called New York variants in the population.

Value added

With mitigation approaches in place, including hybrid education and many students engaged in fully remote learning, there was no increased prevalence or incidence of SARS-CoV-2 in schools compared with the general community.

Our take —

This study used daily case rates and genomic sequences sampled throughout South Africa from May to December 2020 to describe how the SARS-CoV-2 501Y.V2 variant arose in South Africa during the second wave of the nation’s epidemic. The 501Y.V2 variant contains new mutations that appear to lead to increased transmissibility of the virus and may allow for immune escape, although recent data released by Johnson & Johnson indicate that their COVID-19 vaccine still provides protection against 501Y.V2. Continuing to track viral genetic changes will allow public health entities to effectively respond to variants that increase transmission or immune escape. This work clearly illustrates how variants can emerge in areas with ongoing community transmission, reinforcing the importance of public health measures aimed at minimizing the spread of SARS-CoV-2.

Study design

Other

Study population and setting

This study described the emergence and spread of the SARS-CoV-2 501Y.V2 variant during the second wave of the South African epidemic. Daily COVID-19 case rates were obtained from publicly available databases maintained by the National Department of Health and the National Institute for Communicable Diseases. Phylogenetic analysis included SARS-CoV-2 genomic sequences (n=2,759) randomly sampled across South Africa between March and December 2020, and a global background sequence data set (n=2,753). 501Y.V2 variant sequences (n=341) collected between October and December 2020 were used to establish a mutation profile for the variant and to create a structural model of the variant spike protein.

Summary of Main Findings

The second wave of the South African COVID-19 epidemic started in October 2020, beginning with a resurgence of cases in the Nelson Mandela Bay region of the Eastern Cape (EC) and rapidly expanding to include both the Western Cape (WC) and KwaZulu-Natal (KZN). In response, genomic surveillance efforts were intensified in these provinces, leading to the identification of a new variant of concern (501Y.V2). By mid-November, 501Y.V2 became the predominant lineage in EC, WC, and KZN, and preliminary models indicate that its rate of transmission may be increased by as much as 50%. The 501Y.V2 variant contains eight lineage-defining non-synonymous mutations in the spike protein, three of which (K417N, E484K, and N501Y) correspond to the receptor binding domain and have previously been associated with either increased transmissibility or immune escape. Structural modeling demonstrated that two of these mutations (K417N, E484K) are found in key locations targeted by host neutralizing antibodies.

Study Strengths

This study included a large dataset of randomly sampled SARS-CoV-2 whole genome sequences from South Africa, including 341 501Y.V2 sequences. Mutations specific to the 501Y.V2 variant were assessed using structural modeling, estimates of selection pressure, and prior studies.

Limitations

Sequence metadata did not include location information, so health facility location was used to approximate the geographic origin for each sample. Additionally, functional relevance of lineage-defining mutations was assumed based on prior work describing the impact of individual mutations on variants from other areas. Although authors indicate that they have preliminary data supporting the 501Y.V2 variant’s ability to escape neutralization, these data were not presented here.

Value added

This work describes how high levels of SARS-CoV-2 transmission in South Africa led to the emergence and spread of a new variant of concern (501Y.52). This study also describes variant-defining mutations and explores their potential functional relevance in terms of both transmissibility and the ability of 501Y.V2 to evade host immune responses.

Our take —

Several variants with mutations in the S protein, the target of most vaccines including Pfizer/BioNTech, have been identified, and concerns have been raised about the efficacy of the vaccines against these variants. While the data from this study were derived from in-vitro experiments, these findings suggest that antibodies induced by the Pfizer/BioNTech vaccine robustly neutralize the three new variants from the UK, South Africa and Brazil. However, further clinical data will be needed to fully estimate the level of protection that the vaccine provides against these variants.

Study design

Other

Study population and setting

This study analyzed the ability of sera from 15 individuals who received two doses of the Pfizer/BioNTech mRNA vaccine to neutralize several of the newly identified SARS-CoV-2 variants. The authors generated mutated S-proteins from one early viral isolate to contain all the mutations in the variants of interest that were identified in the UK, South Africa and Brazil. To determine the S protein region that is more significant in lowering immune response against viral variants, 2 other viruses were made that have all the deletions seen in the South African plus two other sets of mutations in receptor binding region.

Summary of Main Findings

The study found that the sera from the fully vaccinated participants efficiently neutralized the early US isolate of the virus, as well as all tested variants. Neutralization of the South African variant was lower than other variants, but still robust relative to the US isolate and other variants, with neutralization dilution more than 1:40. Neutralization of viruses with the South African variant deletion, but with different receptor binding region mutations were higher than that of the viruses with all mutations seen in the South African variant, suggesting mutations in the receptor binding region are more significant in antibody neutralization than the deletion.

Study Strengths

The study included analysis of the mutations found in all three of the current variants of interest.

Limitations

This is an in-vitro study that focuses only on the ability of the sera from vaccinated individuals to neutralize the variants of interest. In addition to antibodies, The Pfize/BioNTech vaccine was shown to induce a cell mediated immune response as well. Given that the actual correlate of protection is unknown, further clinical studies are needed to fully determine the actual efficacy of Pfize/BioNTech vaccine against the new variants.

Value added

This study showed that Pfizer/BioNTech (BNT162b2) vaccine-induced neutralizing antibodies are effective against the three main SARS-CoV-2 variants of interests.

Our take —

This systematic review synthesized data from 404 serosurveys globally from December 2019-December 2020, and included seroprevalence, antibody testing data from 5,168,360 individuals. Authors found that close contacts and high risk healthcare workers had a generally higher seroprevalence than low risk healthcare workers and the general population, though there was significant variation across geographic regions. Many of the reviewed studies were determined to be of poor quality, making it difficult to make meaningful comparisons across studies and regions. Within the higher quality studies, combined seroprevalence estimates across studies ranged from 4.2% among low-risk healthcare workers to 18.0% among contacts of known COVID-19 cases.

Study design

Other

Study population and setting

The goal of this study was to collate and synthesize all data from serological surveys for antibodies for SARS-CoV-2 that were published between December 1, 2019 and Dec 22, 2020. The authors systematically searched three databases that house peer-reviewed literature (PubMed, Embase and Web of Science), along with five preprint servers (medRxiv, bioRxiv, SSRN, Wellcome Open Research, and Europe PMC). In order to be included in this review, studies had to include a measure of seroprevalence in non-COVID-19 clinical cases, be conducted after the first reported case in the study area, and report the specific assays used. Only studies in English were included. Seroprevalence was the main outcome, and was defined as the prevalence of seropositive results in each original study.

Summary of Main Findings

A total of 404 of the 10,538 original records identified were included in the study: 8 from the African region, 120 from the region of the Americas, 19 from the Eastern Mediterranean region, 194 from the European region, 19 from the South-East Asia region, and 44 from the Western Pacific region. These 404 studies represented tests done on 5,168,360 unique individuals. 64% (259/404) of studies came from convenience samples (no random sampling). The overall quality of the included studies was low, with only 82 of 404 (20%) classified as of higher quality. Among studies of higher quality, seroprevalence was 18% (95% CI: 15.7-20.3) among close contacts, 17.1% (95% CI: 9.9-24.4) among high-risk health care workers, 4.2% (95% CI: 1.5-6.9) among low-risk health care workers, and 8.0% (95% CI: 6.8-9.2) in the general population. Pooled seroprevalence estimates were highest for the South-East Asia region (19.6% 95%CI 5.5-33.6, all in India). There was a significant variation across the included studies.

Study Strengths

This study is a comprehensive, systematic review of the available evidence at the time of publication. This study incorporated data from a very large body of literature from over 5 million people. The study followed PRISMA guidelines (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) and included an assessment of the quality of included papers by adapting existing quality assessment tools for their purposes. These quality assessments highlight the variation in the test performance, immunoglobulin isotypes, and thresholds used.

Limitations

The main limitation of this study is that it attempts to conduct a meta-analysis on heterogenous estimates that are not standardized and utilize different methods of measurement. Serosurveys reviewed in this systematic were conducted at different points in each country’s epidemic trajectory.

Value added

This study synthesizes the available serological evidence on SARS-CoV-2, globally, and sheds light on the generally low quality of the existing seroprevalence studies and the need to standardize methods.

Our take —

In this preprint that has not yet been peer-reviewed, the authors report new evidence of alphacoronaviruses and betacoronaviruses in bats sampled in Yunnan, China, in 2019 and 2020. The results contribute to evidence that Rhinolophus species bats are a rich source of viruses related to SARS-CoV and SARS-CoV-2. However, a virus with sufficient similarity to SARS-CoV-2 to represent a direct ancestor in bats was not found in the study.

Study design

Other

Study population and setting

To examine the diversity of coronaviruses circulating in bats in southern China, authors collected samples from 23 bat species in the Xishuangbanna Tropical Botanical Garden and nearby areas in Mengla County, Yunnan province, China between May 2019 and November 2020. In total, 411 samples from 342 bats were collected, including 283 fecal samples, 109 oral swabs, and 19 urine samples. Samples were merged into 100 pooled libraries based on sample type, species, and collection date; 18 pools were discussed in a previous study (https://doi.org/10.1016/j.cub.2020.05.023). Metatranscriptomic sequencing was performed on pools to detect viral RNA and pools that were positive for coronaviruses related to SARS-CoV-2 were further tested using PCR and sequencing to identify individual samples that were positive prior to pooling. In addition to viral detection, the authors used occurrence data and numerous environmental variables to predict the geographic distributions of 49 Rhinolophus bat species in Southeast Asia and China.

Summary of Main Findings

Coronavirus sequences were detected in 40/100 libraries, producing 26 long sequences from 20 libraries. Of the 26 sequences, 9 were identified as the genus Betacoronavirus and 17 were genus Alphacoronavirus. All 9 betacoronaviruses were related to SARS-CoV or SARS-CoV-2. PCR tests confirmed that 9 individual samples from Rhinolophus malayanus, R. pusillus, and R. stheno sampled between May and July 2020 were positive for betacoronaviruses. Two viruses (RmYN01 and RmYN02) closely related to SARS-CoV-2 from one pool of fecal samples from R. malayanus were described previously (https://doi.org/10.1016/j.cub.2020.05.023). Three viruses shared between 91.6 and 93.3% sequence identity across the genome with SARS-CoV and other related viruses from bats sampled previously in Yunnan. Four viruses (RpYN06, RsYN04, RmYN05, and RmYN08) were more similar to SARS-CoV-2. RpYN06 from R. pusillus shared 94.5% sequence identity with SARS-CoV-2 across the genome while the latter three were more distantly related to SARS-CoV-2 and were more similar to pangolin coronaviruses. Models using environmental variables showed high accuracy in their prediction of Rhinolophus bat species distributions, showing that up to 23 Rhinolophus species can coexist in Southeast Asia, particularly in Thailand, Laos, Vietnam, and southern China.

Study Strengths

This is the first study reporting results of new sampling and detection of coronaviruses in bats in China after the beginning of the COVID-19 pandemic. Specific methodological strengths include correcting for the possibility of false positives due to sequencing errors (e.g., index hopping) by removing short sequences with high similarity to others within the same chip and same lane in the sequencer. They also confirmed via RT-PCR which bat samples were positive within the pools with amplified coronavirus sequences. Finally, the identities of bat species that were positive for coronaviruses were confirmed by sequencing mitochondrial DNA.

Limitations

The supplementary material is not currently available with the preprint, so it is unclear how many pools or individual samples from each bat species were positive for coronaviruses. Additionally, the authors only validated the presence of coronaviruses in individual bat samples with PCR for pools positive for betacoronaviruses. Thus, it is not clear which individual bat samples were positive for the alphacoronaviruses described.

Value added

The paper provides valuable new information on the circulation of coronaviruses in bat species in Yunnan sampled in 2019 and throughout 2020. The results indicate multiple betacoronaviruses related to SARS-CoV and SARS-CoV-2 can be found in Rhinolophus species bats in this region. Paired with the ecological modeling of Rhinolophus species ranges, this manuscript contributed to existing evidence that Yunnan and other parts of Southeast Asia are regions with high diversity of Rhinolophus species and are likely important hotspots for betacoronavirus diversity. Nevertheless, the viruses amplified do not provide evidence that bats harbor the direct ancestor of SARS-CoV-2. Even though the described RpYN06 virus was 94.5% similar to SARS-CoV-2 across the whole genome, this virus is still too genetically divergent to be the progenitor of SARS-CoV-2.