Skip to main content

Our take —

In this preprint that has not yet been peer-reviewed, the authors show that free-living deer populations in Ohio are becoming infected with SARS-CoV-2. This marks the first evidence of SARS-CoV-2 spread within wild animal populations; another paper showing similar results from deer in Iowa was published contemporaneously. The high prevalence (>30%) of viral RNA detection in both studies, plus evidence of distinct viral lineages present in multiple deer within study sites, suggest that multiple human-to-deer spillover events have occurred and that the virus is likely spreading among deer. It is not clear from these studies how animals are becoming exposed to the virus, how transmission between deer occurs, and how widespread and efficient deer-to-deer transmission is in wild populations and more longitudinal studies will be needed to answer these questions.

Study design

Other

Study population and setting

In light of previous studies finding that white-tailed deer (Odocoileus virginianus) are susceptible to infection with SARS-CoV-2 and capable of indirect transmission in an experimental setting (https://doi.org/10.1128/JVI.00083-21 [https://doi.org/10.1128/JVI.00083-21]) and that free-living deer populations in four US states show serological evidence of SARS-CoV-2 (https://doi.org/10.1073/pnas.2114828118 [https://doi.org/10.1073/pnas.2114828118]), the authors of this study wanted to test whether deer are actively infected with SARS-CoV-2 in Ohio, USA. Between January and March 2021, 360 wild white-tailed deer were culled at nine study sites in northeast Ohio as part of a deer population management program. Nasal swabs were collected from each carcass in the field. Viral RNA was detected in nasal swab samples by real-time polymerase chain reaction (RT-PCR) targeting the SARS-CoV-2 envelope gene; additional RT-PCR assays were performed targeting other genes confirmed the results of the first assay. Presumptive positive samples following RT-PCR screening were sent to the National Veterinary Services Laboratories for whole genome sequencing. After assembly and cleaning, the genomes were assigned to SARS-CoV-2 genetic lineages according to the Pangolin program. Phylogenetic analysis was then performed to compare the genomes from deer to a background dataset compiled from GISAID that included all SARS-Cov-2 sequences from human cases in Ohio during the study period.

Summary of Main Findings

From the 360 nasal swabs collected from deer, 129 (35.8%) were positive for SARS-CoV-2 RNA by RT-PCR over the whole study period. Prevalence estimates varied from 9% to 75% depending on the site and date of sampling. Male deer were more likely to test positive than female deer (chi-squared = 25.45, p-value < 0.0005), and the highest prevalence estimates were observed at four sites nearby urban areas with higher human population densities. Whole genome sequences were obtained from 14 samples across six sites, with lineages including B.1.2, B.1.596, and B.1.582; no Alpha (B.1.1.7) or Delta (B.1.617.2) variants identified. Several sites had multiple deer infected with the same lineage, suggesting deer-to-deer viral transmission. The largest cluster had seven deer sequenced that fell together into clade B.1.596 and had unique amino acid substitutions and deletions that distinguished the genomes from related SARS-CoV-2 genotypes from human cases in the same lineage. Additionally, three distinct clusters of B.1.2 genotypes in deer were detected at different sites, suggesting independent spillover events. Considering all the distinct phylogenetic clusters identified in deer, the authors estimated that there had been at least six independent spillover events from humans to deer that occurred prior to the study period, likely during the winter surge of human SARS-CoV-2 cases in Ohio.

Study Strengths

The sample collection was fortuitous, occurring several weeks after the peak in the winter surge of SARS-CoV-2 in Ohio, when human-to-deer transmission would presumably have been the highest. The collection of samples from multiple sites and the full genome sequencing was important to determining that multiple human-to-deer transmission events had occurred and that unique variants of some lineages (e.g., B.1.596) were likely spreading between deer at some sites.

Limitations

A similar preprint about SARS-CoV-2 RNA detection in white-tailed deer in Iowa was posted only days before this article (https://doi.org/10.1101/2021.10.31.466677), reporting multiple SARS-CoV-2 lineages in the sampled deer, including B.1.2 and B.1.596. However, because the two studies were posted so close together in time, it is unknown whether genotypes in these lineages are shared between deer populations in the two states, which might suggest specific adaptation of the virus to deer, as happened in outbreaks of SARS-CoV-2 in farmed mink. For Ohio sites where only one deer sample was sequenced, it unknown whether deer-to-deer transmission was occurring or not. Moreover, the dynamics of deer-to-deer transmission and whether infection is causing disease and mortality in deer within the Ohio sites cannot be inferred from the limited genetic data. More longitudinal surveillance will be needed. Finally, it is unclear from this study alone how deer are become exposed (e.g., direct exposure to humans in yards or during hunting, or indirectly through contaminated water or trash) and transmitting the virus to each other (e.g., direct contact, airborne, or environmental).

Value added

This study provides direct evidence that free-living white-tailed deer are becoming infected with SARS-CoV-2 via direct or indirect contact with infected humans or contaminated materials. The results corroborate similar findings published contemporaneously (https://doi.org/10.1101/2021.10.31.466677 [https://doi.org/10.1101/2021.10.31.466677]), wherein 94/283 (33.2%) sampled deer (151 free-living and 132 captive) in Iowa between April 2020 and January 2021 were positive for SARS-CoV-2 RNA by RT-PCR. Because white-tailed deer are widespread in the United States and live in areas spanning the rural-urban spectrum, sometimes at high densities, there are likely many opportunities for exposure to SARS-CoV-2 circulating in human populations. If similar human-to-deer transmission events are occurring in many other states, and if deer-to-deer transmission is efficient, then this presents a risk that SARS-CoV-2 may become established in deer populations and potentially lead to deer-to-human transmission events.

Our take —

In this study, available as a preprint and thus not yet peer reviewed, the authors reported a massive effort to describe the diversity of coronaviruses in bats across 14 provinces of China sampled between 2016 and 2021. In 13,064 samples from bats, 146 sarbecoviruses were detected predominantly in Rhinolophus species. While no relatives of SARS-CoV-2 were detected, the closest relative of SARS-CoV in bats was discovered in R. sinicus sampled in 2020, with 95.8% similarity to SARS-CoV at the genome level. Swab samples taken from the Huanan Seafood Market in Wuhan in February 2020 detected animal coronaviruses reflective of animals previously reported being sold in the market, but none of the viruses were related to SARS-CoV-2. This study highlights that China possesses a diverse assemblage of SARS-related coronaviruses in bats, with southern China being a notable hotspot. Additional sampling in this region and in neighboring areas of Southeast Asia could shed more light on the evolutionary origin of SARS coronaviruses.

Study design

Other

Study population and setting

In an effort to describe the diversity of coronaviruses in bats across China, the authors collected 13,064 samples from 56 bat species in 703 locations across 14 Chinese provinces between 2016 and January 2021. Pharyngeal and anal swabs were collected from live bats and then pooled by collection date, species, and site. The field team also had brief access to the Huanan Seafood Market in Wuhan in February 2020, a location where some of the earliest COVID-19 cases visited, and may have been a site where spillover of SARS-CoV-2 from animals occurred. Environmental swab samples (n = 22) were collected from cold storage areas that contained animal products, and 80 swab samples were taken from the environment around stalls selling animal products (ground, walls, sewers, door handles, chopping blocks, knives, and scissors). All samples from bats and the market were tested for the presence of coronavirus RNA using PCR targeting the RdRp gene and with next-generation sequencing. Phylogenetic analysis was then performed to identify different clades of sarbecoviruses in the samples, identity evidence of recombination, and infer whether identified viruses could use human ACE2 based on similarity of their spike protein with other viruses with known ability to enter human cells.

Summary of Main Findings

In the samples collected from bats, 199 of 372 pools were positive for coronavirus RNA: 113 with alphacoronaviruses, 64 with betacoronaviruses, and 22 with both genera. Samples within 44 pools containing sarbecovirus RNA (n = 1,068) were rescreened individually, yielding 146 positive samples mainly from seven Rhinolophus species; 69 of the positive samples produced full genomes from next-generation sequencing. Phylogenetic analysis showed that none of the sarbecoviruses from bats were related to SARS-CoV-2, and instead fell into multiple clades more closely related to SARS-CoV. Six identical genomes (YN2020B-G) from R. sinicus collected in Yunnan Province in 2020 had the highest sequence identity shared with SARS-CoV detected in a bat to date (95.8% across genome, 93.3% within spike); these viruses and another, YN2020H from the same species and year, were predicted to be capable of using human ACE2 based on the phylogenetic clustering with SARS-CoV. In the samples collected from the Huanan Market, three of 11 pools were positive for coronavirus RNA and four coronaviruses were detected, but none were related to SARS-CoV-2 or other sarbecoviruses. Viruses included hedgehog HKU31-related coronavirus in the subgenus Merbecovirus, rabbit HKU14-related coronavirus in the subgenus Embecovirus, canine coronavirus in the subgenus Tegacovirus, and rat coronavirus in the subgenus Embecovirus; these findings were consistent with animal species reported being sold (https://doi.org/10.1038/s41598-021-91470-2 [https://doi.org/10.1038/s41598-021-91470-2]) in the market up to 2019 (e.g., hedgehog, rabbit, bamboo rat).

Study Strengths

This study reports an enormous sampling effort to describe coronaviruses circulating in bats in China, including during 2020 and 2021, and the team had very privileged access to the Huanan Market shortly after it was closed to the public.

Limitations

Due to the pooling strategy, the authors only know the individual-level prevalence of samples containing sarbecovirus RNA that were screened individually; for most other bat species that were not screened individually, these data are unavailable. Such information would have been useful to know if prevalence changed over time, especially in commonly sampled species over the long time period of the study. The inference by the authors about whether viruses could enter human cells was based solely on phylogenetic clustering with other spike sequences that were evaluated in previous studies. In vitro experiments would be needed to evaluate whether other factors beyond ACE2 binding, such as the presence of key proteases, influence infectivity of human cells. Finally, a very small area of the Huanan Market was sampled in the study, so it is unclear what proportion of the animal-selling stalls were sampled in February 2020, and whether or not they were representative of the animal species sold prior to the start of the pandemic.

Value added

This study provides additional data that highlights southern China as a hotspot for sarbecovirus diversity. The absence of relatives of SARS-CoV-2 in the sampled bats, despite sampling in the same areas where previous studies have detected such relatives, suggests that this lineage may not be common in most parts of China and may be more restricted to the southern provinces of China and in neighboring countries in Southeast Asia.

Our take —

In this preprint that has not yet been peer reviewed, researchers surveyed coronavirus diversity in 46 species of bats in northern Laos sampled between July 2020 and January 2021. Among the diverse coronaviruses found, three viruses closely related to SARS-CoV-2 were identified in Rhinolophus species, with BANAL-52 from R. malayanus representing the closest relative of SARS-CoV-2 identified in bats to date, with 96.8% sequence identity at the genome level. Additionally, researchers were able to isolate a coronavirus related to SARS-CoV-2 from a bat in cell culture, representing a first for this lineage and only the fifth time a bat-borne coronaviruses has been isolated. The study indicates that Laos and other countries in Southeast Asia may harbor additional related viruses in Rhinolophus spp. bats that shed light on the evolutionary origin of SARS-CoV-2.

Study design

Other

Study population and setting

To investigate the presence of coronaviruses in bats in Laos, investigators captured and took samples (blood, saliva, feces, anal swabs, and urine swabs) from 645 bats of 46 species sharing caves at four sites in Vientiane and Oudomxay Provinces in northern Laos between July 2020 and January 2021. The authors used polymerase chain reaction (PCR) to identify presence of coronavirus RNA in 539 fecal samples from a subset of bats (methods did not specify number of species). Samples that were positive for betacoronavirus RNA were subjected to full genome sequencing. Betacoronavirus genomes were assessed for evidence of genetic recombination. To test the ability of viruses to enter human cells, the researchers performed molecular simulations of binding between virus spike and the human ACE2 receptor. Potential infectivity of viruses was further analyzed by measuring the entry of pseudotyped lentivirus particles expressing bat coronavirus spike into human kidney cells expressing ACE2. Finally, neutralizing sera from confirmed COVID-19 patients and non-neutralizing control sera from before the pandemic were used to evaluate restriction of pseudotyped lentivirus entry.

Summary of Main Findings

Coronavirus RNA was amplified from 24 bats of 10 species, representing multiple coronavirus subgenera. Sequences of the Sarbecovirus subgenus were only found in Rhinolophus species ­­– R. malayanus, R. marshalli, and R. pusillus – sampled from Feung District in Vientiane Province; alphacoronavirus RNA was found in a single R. affinis and eight other Rhinolophus species were all negative. Five full genomes of sarbecoviruses were sequenced from bats and all five clustered in the novel clade of bat sarbecoviruses related to SARS-CoV-2; three viruses were very closely related to SARS-CoV-2. The virus BANAL-52 from Rhinolophus malayanus is 96.8% similar to SARS-CoV-2 across its genome; two other viruses, BANAL-103 and BANAL-236 were more distant from SARS-CoV-2 but still within the same cluster with BANAL-52. BANAL-236 was successfully isolated in cell culture. Within the spike protein, 16/17 of the amino acid residues that interact with human ACE2 are the same in BANAL-52 or -103 as in SARS-CoV-2 (compared to 11/17 for RaTG13, previously the closest relative of SARS-CoV-2 sequenced from R. affinis collected in Yunnan Province in 2013). Phylogenetic analysis found extensive evidence of recombination in the bat coronaviruses and within SARS-CoV-2, suggesting that transmission of sarbecoviruses among different Rhinolophus species living in the same caves contributes to a complex evolutionary history. The spike of BANAL-52/103/236 binds to human ACE2 with high affinity similar to SARS-CoV-2 and lentivirus particles pseudotyped with BANAL-236 spike readily entered human cells expressing ACE2; entry of particles was successfully blocked by human by neutralizing sera from COVID-19 patients.

Study Strengths

The study used a moderately large sample of bats from a diverse assemblage of species in a hotspot of bat diversity, especially of Rhinolophus spp. Historically, few surveys of coronaviruses in bats have been performed in Southeast Asia relative to China, so this study is an important contribution to the understanding of the biogeography of coronaviruses. The detailed characterization of the detected viruses, including full genome sequencing, spike protein modeling, evaluation of human cell entry, and virus isolation in cell culture, are highly valuable.

Limitations

As with other studies describing relatives of SARS-CoV-2 in bats, the viruses described in this study are still too genetically divergent to be the direct progenitor of SARS-CoV-2, representing decades of evolutionary time.

Value added

The successful isolation of BANAL-236 in cell culture is important because in over 15 years of research on coronaviruses in bats, only five coronaviruses have successfully been isolated in culture; most detections of coronaviruses in bats are only genetic sequences and do not represent cultured viruses, including RaTG13. Thus, this study is the first to obtain a physical specimen of a bat coronavirus related to SARS-CoV-2. The similarity of BANAL-52 to SARS-CoV-2 exceeds the previous record of 96.1% similarity between SARS-CoV-2 and virus RaTG13 from Rhinolophus affinis, making BANAL-52 the closest known ancestor to SARS-CoV-2 in an animal reservoir. The analysis of recombination indicates that the origin of SARS-CoV-2 may not point to a single progenitor virus existing in bats, but rather a complex of viruses that circulate and recombine in multiple species.

Our take —

More than a year and a half into the pandemic, evidence regarding mask effectiveness is scarce, though the balance of evidence suggests a protective effect. In this study of university students in St. Louis, Missouri with COVID-19 and their close contacts, unmasked exposures (in which either the infected person or the contact was unmasked) were about 5 times more likely to result in SARS-CoV-2 transmission than exposures in which both parties were masked. Unfortunately, the analysis did not distinguish between mask use by the infected person and the contact. Other caveats apply: there were only a small number of fully masked exposures, vaccination status was not accounted for (although only 18 contacts were fully vaccinated), the nature of exposure may have varied considerably, and self-reported mask use may have been subject to error. However, these data provide more evidence that masking lowers the probability of SARS-CoV-2 transmission.

Study design

Other

Study population and setting

This study, conducted in St. Louis, MO from January to May 2021, compared rates of SARS-CoV-2 infection by masking status among 378 close contacts during exposure to 265 St. Louis University (SLU) students with confirmed COVID-19. Cases were identified through SLU symptomatic and surveillance testing, and contacts were identified via the SLU contact tracing program. A close contact was defined as any single encounter during a 24-hour period during which individuals spent at least 15 minutes within 6 feet of one another, and mask use was recorded for both case and contact. The exposure was classified as unmasked if either the infected case or the contact was not wearing a mask. The outcome of SARS-CoV-2 infection was assessed via saliva-based RT-PCR testing 5-7 days after exposure. Demographic characteristics and vaccination status (unvaccinated, partial, or full) of contacts were collected. Logistic regression was used to estimate odds ratios of infection following unmasked vs. masked exposures, adjusting for the number of exposures experienced by the contact.

Summary of Main Findings

Of the 378 close contacts identified, 116 (31%) tested positive for SARS-CoV-2 infection. Among all contacts, 26 (7%) reported only masked exposure to the index case, while the remainder reported unmasked exposure (i.e., either the case or the contact was unmasked). Among contacts with only masked exposure, there were 2/26 (7.7%) positive tests; among contacts with any unmasked exposure, there were 114/352 (32.4%) positive tests. The odds ratio for SARS-CoV-2 infection, comparing unmasked to masked exposure, was 4.9 (95% CI: 1.5 to 36.5) after adjustment for the number of exposures. The odds ratio for infection with each additional exposure was 1.4 (1.2 to 1.6). There were 5 positive tests among partially vaccinated contacts (n=24) and no positive tests among fully vaccinated contacts (n=18). No information was provided on vaccination status by masking group.

Study Strengths

Close contacts were identified in an existing contact tracing program, The testing window for close contacts was narrow (5-7 days after exposure).

Limitations

The authors did not distinguish between mask use by the infected case and mask use by the close contact, preventing any inference about source control vs. wearer protection. There was no adjustment for vaccination status in the analysis, nor were fully vaccinated contacts excluded in any sensitivity analysis. If fully or partially vaccinated individuals felt less inclined to wear masks, any protective effect of mask use may have been underestimated. Alternatively, if mask-wearing behavior and vaccine-seeking behavior were positively correlated, the reverse could be true. Mask use was self-reported and thus subject to recall bias (e.g., contacts may have misremembered) and social desirability bias (e.g., contacts may have misrepresented their masking behavior to be viewed more favorably). It is conceivable that some infected contacts acquired their infections from a case other than the one under investigation. There were no data reported on the severity of COVID-19 symptoms among infected contacts, and there were no data on type of masks used. Finally, despite the use of a standard definition of close contact, the nature and duration of exposures may have been highly heterogeneous, which would have an unpredictable effect on estimates.

Value added

Despite the prominence of masks as an object of debate in COVID-19 policy, this is one of the very few studies to have estimated differences in risks of exposure by masking status.

Our take —

Reporting on two SARS-CoV-2 outbreaks in farmed minks in Greece, this paper provides additional evidence that the high density of animals and close contact between farmers and animals during routine care can lead to rapid spread and infection of a majority of animals, resulting in high mortality. Due to the competence of minks as hosts of SARS-CoV-2, the repeated observation of host-specific genetic mutations in mink population, and evidence of transmission from minks to humans on fur farms, improved biosecurity and surveillance procedures must be developed for this industry to prevent further outbreaks, which risk the establishment of mink populations as a permanent reservoir for SARS-CoV-2.

Study design

Other

Study population and setting

The study reports on two investigations of SARS-CoV-2 infection outbreaks on mink farms in the Western Macedonia region of Greece. Farm A, located in the regional unit of Kozani, had approximately 6,500 minks on the property at the beginning of the outbreak in November 2020 when the animals were in the fattening period of their growth prior to pelting. Animals were housed in seven open-sided sheds and animal care was solely managed by two farmers. Farm B was in the regional unit of Kastoria and had 738 animals going through a slimming down phase prior to the outbreak on the property in early 2021.

Summary of Main Findings

The outbreak of SARS-CoV-2 infection on farm A started following a procedure where both farmers took blood samples from all the animals on the farm over the course of three days to test for Aleutian mink virus disease, a common infection in farmed mustelids. Neither farmer wore a mask during blood sampling and one farmer began experiencing COVID-19 symptoms two days after the start of blood sampling; both farmers were positive for SARS-CoV-2 by rapid antigen test and PCR when sampled 12 days later. Minks on farm A began exhibiting signs of illness (reduced food intake) one day after the completion of blood sampling, followed by additional severe signs. Between day 5 and day 26 after observation of reduced food intake, 548 total deaths (8.4% of the population of 6,500) were recorded. A similar introduction event could not be identified for farm B: farmers did not have prolonged contact with animals during the time prior to the outbreak and none of the farmers or farm workers were positive for SARS-CoV-2 RNA or antibodies when tested over the course of the outbreak (days 1–22). Deaths on farm B were first recorded in a single shed one day after the observation of clinical signs and then spread to neighboring sheds. Unlike the more unimodally distributed pattern of deaths on farm A, deaths on farm B were more variable over time. At day 28 when the last death was recorded, 74 minks had died (10% of the population of 738). Ten visibly ill animals were tested from each farm and tested for SARS-CoV-2 via PCR and all oropharyngeal swabs were positive; four samples from each farm were subjected to full-genome sequencing, revealing amino acid substitutions in the spike protein commonly observed in other mink outbreaks (e.g., Y453F) and that the outbreaks resulted from separate introductions of genetically distinct viruses. Serological testing after the end of the outbreaks of 172 animals on farm A revealed 160 (93%) seropositive and 84 of 90 (93.3%) were seropositive on farm B. Taking into account the number of animals that were infected and died, the authors estimated that 93.6% of animals on farm A were infected, 9.01% of infected animals died, and the reproductive number (R­0) at the beginning of the outbreak was 2.9; similar numbers were observed for farm B: epidemic size of 94% and infection fatality rate of 10.7%. Additionally, SARS-CoV-2 RNA was detected on surfaces above mink enclosures and from the air on both farms when sampled in February 2021.

Study Strengths

Because the investigations on the farms were initiated rapidly after the first onset of clinical signs in animals, and the outbreak on farm A was started when hundreds of animals were exposed to an infected farmer, the researchers were able to estimate more epidemiological parameters than other studies of mink outbreaks in Europe. This included the number of infections and deaths in mink attributed to the introduction event, the final outbreak size, and the infection fatality rate. In contrast to other studies in Denmark and the Netherlands, the outbreaks on the two farms in Greece were not ended via culling of animals, allowing the researchers to test whether the outbreaks would resolve after mink populations developed herd immunity.

Limitations

Without an identifiable introduction event for farm B, it was not possible to estimate some epidemiological parameters, including the number of infections and deaths due to the introduction event, R­0, the epidemic growth rate, the doubling time, or the generation time. It was also unclear why the distribution of deaths over time was so different between farms A and B. While this could be attributed to less intensive exposure of animals on farm B to infected human index cases as compared to farm A or limitation of exposure to a smaller number of animal sheds, greater scrutiny of the locations of infections in sheds over time might help to explain the stochasticity in daily deaths. While the two outbreaks were very similar in the number of mink deaths and the proportion of the population infected despite the different growth stages of the animals on each farm (fattening vs. slimming down), potential differences in other epidemiological parameters due to the different growth stages and their potential effects on mink health and immunity were not investigated.

Value added

This article adds to the existing evidence that SARS-CoV-2 is readily transmissible between humans and minks on farms where animals are raised intensively for their fur. The observation that infections and deaths in minks eventually ended on both farms without culling indicates that a herd immunity threshold may exist in mink populations, but at the cost of high mortality (~10%).

Our take —

When comparing observed to expected cases among US National Football League players during the 2020 regular season, the study found that the observed cases were 55% lower than expected. During the season, the NFL put in place a number of mitigation strategies, including daily testing and masking, that may explain this observed result. Comparing NFL players to the general community in terms of COVID-19 cases may be problematic, however, because of differences in socioeconomic status, underlying health, and race. Further, results precede the emergence of the Delta variant and widespread access to the vaccine, both which may affect risks, though in opposite directions. Overall, these data suggest that with intensive mitigation strategies, at least pre-Delta, return to safe play of sports is possible. High rates of vaccination coverage among NFL players (93%) may further increase safety in the upcoming 2021-2022 season.

Study design

Other

Study population and setting

To quantify the impact of holding a regular National Football League (NFL) season on infections among players, observed versus simulated expected cases of COVID-19 among players were compared during the 2020 football season. There were 32 teams who played 256 regular season games, and additionally met for meetings and practices. Between August 1, 2020, and January 2, 2021, the expected number of infections was estimated in a population with similar age and sex using county-level COVID-19 test data for each team (New York Times positive test counts for each day). A binomial distribution was used to simulate expected cases and compare these to observed COVID-19 infections in players. Given that daily testing was being done among the players but not in the general population, an inflation factor (based on CDC estimate for the beginning of the period) of 6.5 (95% CI: 3.5-12.4) cases per positive test was applied to the expected case counts.

Though a “bubble” was not utilized by the NFL (restricting contact between players and the outside community), a number of mitigation strategies were put into place to prevent the transmission of COVID-19, including: daily testing (RT-PCR), mandatory masking, quarantine and isolation, and others.

Summary of Main Findings

Overall, when comparing observed (n=256) to the expected (n=578) simulated cases, it was found that observed cases were 55.7% (95% CI: 53.2%-58.1%) lower. 30 of the 32 teams fell at or below the expectations, while 2 teams had a greater number of cases than expected. Data from 2020 precede the emergence of the Delta variant, which may reduce effectiveness of mitigation strategies and alter the risks faced.

Study Strengths

A strength of this study is that there was daily testing and complete follow-up data on all the players during this period.

Limitations

One major limitation is that NFL players are different from the general population in a number of ways (socioeconomic status, underlying health/fitness, and/or diet, race/ethnicity, etc.) that this study does not take into account. Some of these differences may partially explain the difference in observed vs. expected cases.

Value added

These results support the understanding that testing and behavioral mitigation strategies can facilitate a return to normal activities while limiting the spread of COVID-19.

Our take —

There is growing evidence that myocarditis, or inflammation of the heart muscle, is an infrequent adverse event occurring predominantly in adolescent boys who have received the second dose of the Pfizer-BioNTech vaccine. A recent CDC study on this topic estimated the harms and benefits associated with vaccination and concluded that the benefits exceed the harms across all age groups for both boys and girls. Similarly, other studies have estimated vaccine-associated myocarditis risks using an active surveillance system called the Vaccine Safety Datalink. This study used data from the US Vaccine Adverse Events Reporting System (VAERS) to conduct a similar analysis, and found higher rates of myocarditis than in previous studies. However, VAERS is an open, passive surveillance system that is not suitable for this kind of analysis and is prone to overestimating vaccine side effects because cases are not subject to clinical adjudication. The authors went on to compare these estimated myocarditis risks with pediatric COVID-19 hospitalization rates, which is highly misleading. Preliminary data suggest that the clinical course of vaccine-associated myocarditis is favorable, while COVID-19 hospitalization is more likely to result in ICU admission, and rarely, death. Any implication that vaccination is more dangerous to boys than COVID-19 is entirely unsupported by this study.

Study design

Other

Study population and setting

This study used data from the US Vaccine Adverse Events Reporting System (VAERS), a passive surveillance system not subject to clinical adjudication, from January to June 2021 to estimate rates of reported myocarditis and related symptoms among children aged 12-17 years who received mRNA vaccination against SARS-CoV-2 infection. The authors defined a “cardiac adverse event” (CAE) by the presence of one of the following in the symptom notes: myocarditis, pericarditis, myopericarditis, acute myocardial infarction, elevated troponin, abnormal EKG, abnormal echocardiogram, or cardiac MRI results consistent with myocarditis. Rates of CAE were calculated by age group (12-15 years and 16-17 years) and sex, and by vaccine dose (first or second). The authors then attempted a harm-benefit analysis by comparing their estimates of vaccine-associated CAEs with 120-day pediatric COVID-19 hospitalization rates, stratified by presence/absence of comorbidities. Hospitalization rates were derived from COVID-NET during three periods: January 2021 (high incidence), June 2021 (low incidence), and August 2021 (medium incidence).

Summary of Main Findings

There were 257 cardiac adverse events (90% among males) reported in the VAERS during the study period that met the authors’ search criteria. Among these reports, all children but one (Moderna) had received the Pfizer-BioNTech vaccine, as it was the only one approved for use in children younger than 18 years during the study period. The majority (85%) of CAE reports occurred after the second dose. The estimated incidence of CAEs among boys aged 12-15 years following the second dose was 162 per million; the incidence among boys aged 16-17 years was 94 per million. The estimated incidence of CAEs among girls was 13 per million in both age groups. The incidence of CAEs was considerably lower after the first dose across all age and sex groups. Median peak troponin was 5.2 ng/mL among boys aged 12-15 years, 11.6 ng/mL among boys aged 16-17 years, 0.8 ng/mL among girls aged 12-15 years, and 7.3 ng/mL among girls aged 16-17 years. The authors compared these estimates of CAE incidence against 120-day COVID-19 hospitalization, concluding that for boys without comorbidities, a vaccine-associated CAE was several times more likely than a COVID-19 hospitalization (the ratios varied by sex, age, and whether hospitalization rates were calculated during times of low or high incidence).

Study Strengths

The authors made their data publicly available and easily accessible.

Limitations

A primary problem with this study is its reliance on VAERS data for estimating incidence of myocarditis without any clinical adjudication, secondary data, or control data. The VAERS is an open, passive surveillance system in which any member of the public can submit a report; it is an early-warning system designed to flag symptoms and issues worthy of more rigorous study, and is not designed to be used for the purposes attempted here. The criteria for defining a cardiac adverse event were overly broad, likely overestimating the incidence of vaccine-induced myocarditis to an unknown degree. Additionally, there was no control group used to adjust for baseline myocarditis incidence in the study population. Importantly, it is misleading to compare the incidence of CAE reports with 120-day COVID-19 hospitalization rates. Preliminary evidence suggests that the clinical course of vaccine-associated myocarditis is mild; the clinical course of COVID-related hospitalization is quite serious for a minority of children (with nearly one quarter requiring ICU admission). Additionally, the use of a 120-day window understates ongoing risks posed by SARS-CoV-2 infection; vaccine protection against infection continues beyond the window, and pediatric COVID-19 hospitalization rates have continued to increase.

Value added

A recent MMWR report (Gargano et. al 2021) contained a similar analysis but with outcome data that was subject to physician review and adjudication for alternative myocarditis etiologies. It is not clear what this study adds to the previous report other than further age stratification.

Our take —

The frequency and severity of long-term symptoms following recovery from COVID-19 are important to quantify. However, quantitative data for long-term health consequences following hospitalization remains limited. This ambidirectional cohort study characterized one-year health outcomes among 1,276 COVID-19 survivors who were hospitalized at a single center in Wuhan, China, between January and May of 2020. While hospitalized, 68% of participants required supplemental oxygen, 7% required treatment with high-flow nasal cannula and/or non-invasive ventilation, 1% required mechanical ventilation and/or ECMO, and 4% were admitted to the ICU. Outcomes at six- and twelve-months following hospital discharge were quantified based on physical exam, self-report measures, and laboratory results. While 49% of participants reported persistent long-term symptoms (i.e., fatigue, dyspnea, anxiety/depression) after one year, 88% had returned to normal employment. Furthermore, COVID-19 survivors reported more general health problems (i.e., mobility issues, pain, mental health concerns) than a matched population of healthy non-hospitalized controls (66% vs. 33%), indicating that additional follow-up is required to fully characterize the long-term health impact of SARS-CoV-2 infection.

Study design

Other

Study population and setting

This ambidirectional cohort study characterized one-year health outcomes among COVID-19 survivors who were discharged from a single center in Wuhan, China between January and May of 2020. Participants attended study visits six and twelve months after hospital release. Each visit included a physical examination, detailed interviews, health questionnaires, and laboratory tests. A subset of participants also underwent pulmonary function testing and a chest CT. Data were compared between visits and to healthy controls recruited from the local community, matched 1:1 for age, sex, and relevant comorbidities.

Summary of Main Findings

Of 2,469 COVID-19 survivors eligible for study inclusion, 1,276 (58%) voluntarily participated in both follow-up visits and were included in this analysis. During hospitalization, 68% (864/1,276) participants required supplemental oxygen, 7% (86/1,276) required treatment with high-flow nasal cannula and/or non-invasive ventilation (i.e., CPAP, BiPAP), 1% (8/1,276) required mechanical ventilation and/or ECMO, and 4% (54/1,276) were admitted to the ICU. At the six-month visit, at least one persistent symptom was identified in 68% (831/1227) of participants; this proportion dropped to 49% (620/1272) by the twelve-month visit. Fatigue/muscle weakness (52% at six months; 20% at twelve months) was the most reported symptom. The prevalence of both dyspnea and anxiety/depression increased slightly between the six- and twelve-month visits (from 26% to 30%, and 23% to 26%, respectively). By twelve months, most participants had normal pulmonary function testing, as well as improvement in any prior abnormalities visualized on CT (i.e., ground glass opacities). However, some participants with a more severe hospital course had persistent lung diffusion defects and radiographic abnormalities. Of participants who had been employed prior to their illness, 88% (422/479) had returned to work at twelve months. General health problems (i.e., mobility issues, pain, mental health concerns) were more frequent at twelve months among COVID-19 survivors compared to matched controls (66% vs. 33%).

Study Strengths

This is the largest longitudinal cohort study of COVID-19 survivors following hospitalization. Outcomes were quantified based on multiple data sources, including physical exam, self-report measures, and laboratory results. Outcomes were compared to healthy controls from the local community, matched 1:1 for relevant comorbidities and demographic data.

Limitations

Moderate rate of study participation (58%) among eligible participants may have introduced sample bias, though there were no significant differences in measured baseline characteristics between eligible and participating groups. All participants were enrolled from a single center, and results may not be representative of the general population. Additionally, patients were excluded from study participation if they died following hospital release, were discharged to a nursing home, or had significant physical limitations that prevented normal mobility; as such, this study may not adequately quantify the risk of long-term health consequences following COVID-19 hospitalization. Researchers did not have access to data on the health status of participants prior to infection, which may have confounded results, as participants may have had an increased risk of severe COVID-19 due to factors not considered in control group matching. Furthermore, cases were hospitalized with COVID-19 and controls were from the local community. Despite matching on important variables, hospitalization, especially intensive care unit admission, is likely associated with persistent symptoms. The control group therefore does not provide a baseline level of persistent symptoms after hospitalization for non-COVID-19 illness. Importantly, participants were infected in the early phases of the pandemic and before the rise of the Delta variant. At the time, providers were still identifying best practices for the management of COVID-19 patients, which may have negatively impacted both disease course and long-term health complications. Finally, the early public health response to COVID-19 in China mandated hospitalization for all infected individuals, including those with no or only minor symptoms. As such, this cohort may not be adequately representative of hospitalized patients with moderate to severe COVID-19 disease.

Value added

This is the largest longitudinal study of hospitalized COVID-19 survivors that measures long- term health consequences up to a year after infection.

Our take —

Vaccines are highly efficacious at preventing severe disease and reducing likelihood of infection, but breakthrough infections do occur. This study from the Netherlands was available as a preprint and thus has not yet been peer reviewed. When comparing primary infections occurring between April and December 2020 to breakthrough infections (infections after vaccination) that occurred between April and July 2021, it was found that Ct values were similar but that the probability of detection of infectious virus was lower among breakthrough infections. Comparing these two groups should be done cautiously as there are a number of reasons that could explain these differences in virological characteristics (changes in attitudes and behaviors at the hospital over time, different dominant variants, etc.). This study does not provide sufficient evidence of the comparability of these two groups. Further, given the dominance of the Delta variant which is known to be more virulent, in the latter time period when the breakthrough infections occurred, the effectiveness of the vaccine at reducing viral replication and infectiousness may have been greater had the Alpha variant remained the dominant variant in circulation.  

Study design

Other

Study population and setting

Virological data (RT-PCR and Ct values) were compared between unvaccinated healthcare workers (HCWs) with primary infection (April-December 2020) and vaccinated HCWs with breakthrough infections (April-July 2021, n=161) at two tertiary care centers in the Netherlands. Virological characteristics were compared using the first RT-PCR positive sample. Positive samples underwent sequencing to determine the variant (Alpha, Beta, Gamma, or Delta) and were cultured to assess viability of the virus. Testing algorithms remained constant across the two periods, with symptomatic HCWs seeking testing and contact tracing performed in the event of a positive case.

Summary of Main Findings

Among the 161 breakthrough infections detected between April and July 2021, the majority (70.8%, n=114) were identified as the Delta variant (B.1.617.2). None of the infections required hospitalization and were described as mild. The mean age of HCWs experiencing a breakthrough infection was 25.5 years old. Ct values were significantly lower in symptomatic vs. asymptomatic breakthrough infections (23.2 vs. 26.7), indicating a higher viral load among symptomatic breakthrough infections vs. asymptomatic breakthrough infections. When comparing those with breakthrough infections and those with primary infections, Ct values were similar but the probability of infectious virus detection based on culture was lower among those with breakthrough infections.   

Study Strengths

The sequencing data for all cases is a strength, as was the availability of viral culture data.

Limitations

The authors indicate that the two groups did not differ with respect to demographic characteristics, but a Table indicating symptom status, age, variant, and Ct-value are only provided for the breakthrough infections (April-July 2021), making it difficult for the reader to evaluate similarities and differences between the two groups. Comparison between the two groups is challenged because these two groups were infected in two different time periods (could have been changes in attitudes, behaviors, mandated NPIs) and predominantly by different variants, which could mean different virological characteristics regardless of whether a breakthrough or primary infection. Primary infection sequencing data appear to have been collected, but were not reported in the paper.

Value added

This study helps us better understand differences in transmission likelihood between those who are vaccinated with breakthrough infections and those who are unvaccinated with primary SARS-CoV-2 infections.

Our take —

In this preprint that has not yet been peer-reviewed, researchers at the US Department of Agriculture report that 40% of white-tailed deer sampled in 2021 from four eastern states had antibodies against SARS-CoV-2. Deer are common throughout the US and are frequently in proximity to humans, so exposure may have occurred via many possible routes. It will be important to determine whether deer are transmitting the virus in the wild, the sources of exposure, and the geographic extent of infections in order to prevent the establishment of a permanent reservoir in deer populations.

Study design

Other

Study population and setting

This paper presents preliminary results of a serological surveillance program focusing on SARS-CoV-2 in white-tailed deer (Odocoileus virginianus) in the US. Between January and March 2021, 385 serum samples were collected from wild deer in Michigan, Pennsylvania, Illinois, and New York as part of routine wildlife management and disease surveillance efforts. An additional 239 serum samples collected between 2011 and 2020 from the same four states, plus New Jersey, were used as pre- and early pandemic controls to identify potential cross-reactivity of endemic coronaviruses. All samples were tested using a surrogate virus neutralization test (sVNT) that measures inhibition of binding between SARS-CoV-2 spike protein and the host ACE2 receptor. A selection of 24 samples from 2021 that were positive by sVNT (>30% inhibition) and 24 that were negative were confirmed with a virus neutralization test using live SARS-CoV-2 virus.

Summary of Main Findings

154/385 (40%) of samples from 2021 were positive by sVNT, most having ≥80% inhibition. Antibodies were detected in three samples from 2020 and one sample from 2019, while none were detected in samples from 2011-2018; percent inhibition scores for 2019 and 2020 samples were low (30.03-43.72), just above the threshold of positivity. Seropositive deer were detected in all four states in 2021, with high clustering of detections within particular counties. Samples tested with VNT produced the same number of positives (24) and negatives (24) as in the sVNT.

Study Strengths

The results are strengthened by the inclusion of matched pre-pandemic controls from the same states and counties. Confirmation of sVNT results in a subset of samples from 2021 using VNT increases confidence that the detections were not entirely due to cross-reactivity or other factors leading to false positives.

Limitations

Since the deer in this study were sampled opportunistically as part of other activities (including disease surveillance and urban removals) in only four states, the results are unlikely to be representative of the whole US deer population. It was also unclear which samples from which states were chosen for confirmation by VNT, so the seroprevalence measurements for each state may be overestimates. Without PCR or virus isolation, it is unclear whether animals were actually infectious at any point, and when infection occurred. There was also little information provided on what areas animals were sampled from (e.g., urban versus rural, nearby wastewater sources or farms), which could have provided information on how animals were exposed. Due to the limited number of samples collected later in 2020, the authors could not estimate the rate of increase in detections during the early phase of the pandemic. Finally, it was unclear if some detections by sVNT may be due to cross-reactivity with endemic coronaviruses in deer, although it is unlikely that cross-reactivity would occur in the VNT.

Value added

This study represents some of the most compelling evidence of possible spillback of SARS-CoV-2 into wild animal populations in the US. Previous experiments (https://ncrc.jhsph.edu/research/susceptibility-of-white-tailed-deer-odocoileus-virginianus-to-sars-cov-2/) had indicated that white-tailed deer were susceptible to SARS-CoV-2 infection, showed few clinical signs, and could transmit the virus to naïve contact deer. If transmission is indeed occurring in these animals in the wild, it may form a new reservoir that could continue to seed infections in the US in the future and contribute to the evolution of new SARS-CoV-2 variants.