Skip to main content

Our take —

This study provides credible estimates that by May 4, 2020, the 11 European countries in this study had achieved control of the COVID-19 epidemic by bringing the reproduction number below 1, after 3.2% to 4.0% of the population had become infected (far more than has been reported). The authors further estimate that country-wide non-pharmaceutical interventions (e.g., lockdowns and school closures) prevented over 3 million deaths. Results suggest the need for suppression measures to continue for a long time, since susceptible individuals are still estimated to represent a large fraction of the European population. However, this implication relies on model assumptions that preclude the ability of other factors, including individual behavior changes, to reduce transmission.

Preprint Review

This expert summary is for the peer-reviewed article linked above. We also summarized this paper before it underwent peer-review.  You can find the original review of the preprint by clicking here.

Study design

Ecological; Modeling/Simulation

Study population and setting

A semi-mechanistic Bayesian hierarchical model was fit to observed COVID-19 deaths in 11 European countries, and was used to estimate the number of cases and the reproduction number (Rt) of the infection. The model used partially pooled data across countries, along with the timing of country-specific non-pharmaceutical interventions, to estimate the impact of these interventions on Rt (with particular attention to whether Rt has been driven below 1) and on the number of infections and deaths in these countries up to May 4, 2020.

Summary of Main Findings

The reproduction number of the virus is estimated to have been reduced from 3.8 (95% credible interval: 2.4 to 5.6) to below 1 in all countries. Many times more people are estimated to have been infected by SARS-CoV-2 than have been confirmed: across all 11 countries, estimated cases totaled 12 to 15 million for an overall attack rate of 3.2% to 4.0%. Estimates of country-specific attack rates ranged from 0.46% (0.34% to 0.61%) in Norway to 8.0% (6.1% to 11.0%) in Belgium. 3.1 million deaths (2.8 million to 3.5 million) are estimated to have been averted by all non-pharmaceutical interventions by May 4. Of the categories of intervention, only lockdown had an identifiable impact.

Study Strengths

The model is fit to observed deaths, which are likely to be more reliable than case counts or hospitalizations. Estimates of case counts are supported by evidence from serologic surveys. The model reproduces observed data up to May 4th, 2020 very well. Uncertainty from various sources is appropriately handled by the model explicitly and by the discussion implicitly. Prior distributions and parameter values are chosen based on current, best available data.

Limitations

The model did not allow for other factors, such as changes in individual risk-avoiding behavior, to affect Rt. The infection fatality ratio, which is key to estimation of the number of infections, is treated as a fixed value by the model; there remains considerable uncertainty about the true value of this parameter. The timing of country-specific interventions makes it difficult to distinguish between the effects of specific interventions. Interventions are assumed to have the same impact across countries and time. Effect estimates in the model are heavily influenced by countries with a high number of deaths that implemented interventions earlier.

Value added

This study is the most thorough model-based estimate of the impact of European country-wide non-pharmaceutical interventions so far.

Our take —

This study, available as a preprint and thus not yet peer reviewed, confirms that cats are highly susceptible to SARS-CoV-2 infection, and that dogs are less susceptible than cats, with limited viral shedding, and that both animals develop antibodies to the virus following infection which provide protection against re-infection in cats. More research will be needed to confirm if cats can transmit the virus back to humans or to other animals.

Study design

Prospective cohort; Ecological

Study population and setting

The study population included seven adult cats (1 male, 6 female, 5-8 years old) and three adult dogs (all female, 5-6 years old). Cats and dogs were housed in groups. Weight, body temperature, clinical status, and oral swabs were obtained prior to viral exposure, and all cats tested negative for feline enteric coronavirus antibody prior to starting the study. Cats were inoculated with 3×10^5 plaque forming units of SARS-CoV-2 and dogs received 1.4×10^5 plaque forming units intranasally. Three inoculated cats (cohort 1) were monitored for viral shedding via orapharyngeal swabs, nasal flushes, and blood samples between for 14 days post-inoculation. At 28 days post-inoculation, cats were re-challenged with 3×10^5 plaque forming units of virus, then oronasal samples were taken between for 14 days post re-challenge (up to 42 days after initial infection). Cats were euthanized at day 42 and tissues were collected for histopathology. Two cats (cohort 2) were exposed to the virus like cohort 1, then two naïve cats were introduced into the room with the infected cats 48 hours following initial infection. Inoculated cats were euthanized at day 5 post-inoculation, and tissues were collected for virus isolation and histopathology. Contact cats were euthanized at day 30 post-inoculation and necropsied. Dogs were sampled at the same frequency as cat cohort 1 for 42 days post-inoculation, but were not re-challenged with the virus.

Summary of Main Findings

None of the cats or dogs showed any clinical signs of infection, including fever or changes in body weight; the clearest pathological sign observed in cats was moderate rhinitis and minor interstitial pneumonia. All three cats in cohort 1 shed virus orally and nasally for up to five days post-inoculation. Contact cats from cohort 2 shed virus orally as early as 24 hours post-exposure. Cats in both cohort 1 and direct contact cats from cohort 2 developed antibodies to SARS-CoV-2 as early as 7 days post-inoculation that remained high until the end of the study. Re-challenged cats showed no signs of viral shedding during the 7 days following re-exposure, but did show a moderate increase in antibody titer. Viral shedding was not detected in any of the dogs post-infection, but dogs developed antibodies by day 14 post-exposure, although with lower titers than cats.

Study Strengths

Compared to previous experimental studies of SARS-CoV-2 infection in cats, this study examines shedding kinetics over time, assesses virus neutralization, seroconversion, and transmission in the same experiment. This is also the first study to report protective immunity against SARS-CoV-2 following repeated exposure.

Limitations

All animals were adults, so it is unclear whether the lack of observed clinical signs of infection changes with animal health status, age, and comorbidities. The study also does not investigate the potential for inter-species transmission, such as between cats and dogs, or between cats and humans.

Value added

The study confirms that cats are highly susceptible to SARS-CoV-2 infection in experimental conditions, and that they develop neutralizing antibodies that protect against subsequent challenge with the virus.

Our take —

In this study available as a preprint and thus not yet peer reviewed, farmed mink at two locations in the Netherlands became infected with SARS-CoV-2, and developed respiratory symptoms following suspected exposure to farm workers with COVID-19 at each farm. The presence of viral RNA in inhalable dust from the mink housing areas suggests airborne transmission between minks, and may pose a risk for mink-to-human transmission. The Dutch government has now reported two cases of COVID-19 in farm workers that are likely due to mink-to-human transmission, and will cull animals on infected mink farms in the country.

Study design

Ecological

Study population and setting

Signs of respiratory disease and mortality were reported in minks on two farms (NB1 and NB2) in Noord Brabant province, Netherlands on April 19 to 20, 2020. Suspecting infection with SARS-CoV-2, researchers collected samples from lungs of three deceased animals from each farm on April 21 and 25, 2020. Samples from animals on each farm each farm (one index animal and four additional animals from NB1; one index animal and five additional animals from NB2) were collected to sequence genetic material of the virus. In the following week, 36 deceased animals (18 per farm) were collected and necropsied, and individual throat and rectal swabs were collected for detection of the viral genetic material. Inhalable dust inside of mink houses on farms was also sampled between April 28 and May 2, 2020 via air pumps for 5-6 hours to test for viral genetic material.

Summary of Main Findings

At the time of reporting, the sick minks showed signs of watery nasal discharge, with severe respiratory distress in some animals; necropsied animals showed signs of pneumonia in lungs tissues (16/18 from NB1 and 12/18 from NB2). Mortality in animals on the farms was 2-4 times higher in the period from April 19 to 30, 2020 than background mortality from past data. SARS-CoV-2 RNA was found in all seven mink from which organs were collected, in the nasal conchae, lung, throat swabs, and rectal swabs, with a total of 36/36 positive throat swabs and 34/36 positive rectal swabs.

One farm worker on farm NB1 had shown symptoms consistent with COVID-19 at the beginning of April, 2020 but was not tested for SARS-CoV-2. A farm worker at NB2 was diagnosed with COVID-19 and hospitalized on March 31, 2020 but could not produce a sufficient viral load for sequencing. The viral genomes from positive minks on each farm were similar to other sequences from human cases from within the Netherlands, but the viral sequences from animals from the two farms clustered into separate genetic groups by farm, suggesting independent exposure events on each farm from infected workers. Additionally, viral RNA was detected in 3/6 inhalable dust samples from NB1 and 1/3 from NB2.

Study Strengths

The sequencing of viral RNA from animals on each farm clarifies the separate origins of infection on the farms. Sequencing of viral RNA in inhalable dust provides some evidence that transmission of the virus between animals is through an airborne route.

Limitations

The authors can only suspect that the mink infections on each farm were seeded by the sick farm workers due to the lack of genetic sequences from the human cases. It is also unclear from the limited sampling what the overall incidence of infection was on each farm. Additional testing of animals would be needed to assess the number of animals that were infected, but showed no clinical signs or mortality, and if mink develop antibodies to SARS-CoV-2 following infection.

Value added

This study broadens the number of domesticated and farmed animals that are susceptible to SARS-CoV-2 following exposure to human COVID-19 cases.

Our take —

This study reports 86% (32 of 37) of reporting jurisdictions report at least COVID-19 case among US correctional and detention facilities. There were 4,893 detected cases among incarcerated or detained people, and 2,778 cases among facility staff. This is an important case count, however only 69% of jurisdictions responded to the CDC request for data, and facilities did not use a universal testing strategy, therefore the numbers are likely an undercount. More granular data from facilities are needed to help understand the situation within these vulnerable populations.

Study design

Ecological; Other

Study population and setting

The study consisted of aggregate data of COVID-19 cases in correctional and detention facilities from 37 state and territorial health department jurisdictions (out of 54 jurisdictions requested) from April 22 to 28, 2020, and laboratory-confirmed cases identified and reported from January 21 to April 21, 2020.

Summary of Main Findings

Of the 37 jurisdictions reporting, COVID-19 cases were reported from 32 (86%) jurisdictions and 420 facilities within these jurisdictions. Of these, 221 (53%) reported COVID-19 cases only among staff members. There were 4,893 total cases among incarcerated or detained persons, 491 (10%) hospitalizations, and 88 (2%) deaths due to COVID-19. Of 2,788 staff member cases, 79 (3%) were hospitalized, and 15 (1%) died due to COVID-19.

Study Strengths

This study is the first report of laboratory-confirmed COVID-19 cases among correctional and detention facilities. The study reported the number of cases due to detained and incarcerated people, as well as the number due to staff members; this disaggregation is important to describe risks among two very different populations that are both present in these facilities.

Limitations

The study only had data from 69% of the total 54 jurisdictions requested, therefore there may be selection bias and these results may not be generalizable across US facilities, and some jurisdictions only had state facility data, rather than local jails, federal, or private facilities. Also, most facilities do not provide universal testing to either incarcerated/detained people or to staff members, and therefore these numbers are likely undercounts. With aggregate numbers, none were disaggregated by state or region, or with any individual information, therefore it is unclear if these cases cluster in any particular area or among any particular population.

Value added

This is one of the first reports of the aggregate number of cases throughout the US in correctional and detention facilities.

Our take —

To date, incomplete and inconsistent reporting of data on race/ethnicity has made it difficult to assess the impact of the U.S. COVID-19 epidemic on racial and ethnic minorities. This study documents the early and disproportionate burden of COVID-19 infection and death shared by black communities, particularly among those in non-urban settings. Although the ecological design limits its ability to accurately measure disparities and their causes, the results are consistent with our current understanding of how segregation and structural disadvantage, particularly in housing and access to care, produce inequities in health.

Study design

Ecological

Study population and setting

U.S. county-level data (n=3,412 counties) on COVID-19 diagnosis and deaths through mid-April 2020 were linked to county-level data on demographics (age group, race, unemployment, urbanicity), environment (urbanicity, air quality, social distancing), and healthcare access (insurance) from various sources from 2014-2020, including the U.S. Census Bureau, the Centers for Disease Control, and Unacast. Bayesian hierarchical models adjusting for time since first case were used to estimate rate ratios for counties with a higher vs. lower proportion of Black residents (as compared to the national average), adjusting for county-level characteristics. Population attributable fractions were estimated for all, disproportionately black, and non-disproportionately Black counties.

Summary of Main Findings

By April 13, 2020, 52% of diagnoses (n= 283,750) and 58% of deaths (n=12,748) were reported in communities where 13% or more of residents were Black (higher proportion). After adjustment for county-level characteristics, counties with a higher proportion of Blacks remained associated with higher rates of diagnosis (ratio = 1.24; 95%CI 1.17-1.33) and death (ratio = 1.18; 95%CI 1.00 – 1.40). For diagnosis, significantly higher rates were observed across the urbanicity spectrum; associations with death were only significant among small metropolitan and non-core areas. Population attributable fractions estimated excess deaths associated with occupancy crowding (280,112) and lack of health insurance (126,985).

Study Strengths

The study analyzed data from nearly all US counties, included a wide range of predictors associated with infection and severe COVID19 prognosis, and used appropriate regression methods for small area estimation.

Limitations

As an ecological study, it cannot link data on race, diagnosis, death, or confounders at the individual level. As such, measures of disparate COVID-19 diagnosis or death, and their causes, were not estimated by this study. The reported community-level associations may be underestimated due to barriers to diagnosis and treatment of COVID19 complications.

Value added

This study confirms that, nationally, higher rates of COVID-19 infection and death were found among communities of color even with adjustment for additional community factors.

Our take —

A study of dogs from households with COVID-19 cases in Hong Kong demonstrates that occasional human-to-dog transmission of SARS-CoV-2 can occur, producing infection with limited to no clinical symptoms, viral shedding in the nasal passages, and the development of antibodies following resolution of infection. While there is still no evidence that dogs can infect other dogs or humans, dogs in households with COVID-19 cases should be isolated to prevent infection.

Study design

Prospective cohort; Ecological

Study population and setting

The study focuses on fifteen dogs and seven cats from households with known COVID-19 cases that were quarantined and tested by the Hong Kong Agriculture, Fisheries, and Conservation Department in Hong Kong, China, as of March 27, 2020.

Summary of Main Findings

Two dogs out of the fifteen tested positive for SARS-CoV-2: a 17-year-old male Pomeranian with pre-existing diseases and a 2.5-year-old male German Shepherd in good health. The owner of the Pomeranian developed symptoms of COVID-19 on February 12, 2020 and was diagnosed on February 24, 2020; a female domestic helper developed fever on February 16, 2020. Five consecutive nasal swabs from the Pomeranian tested positive for SARS-CoV-2 genetic material between February 25 and March 9, 2020; all rectal and fecal samples tested negative. The owner of the German Shepherd developed symptoms on March 10, 2020 and was diagnosed with COVID-19 on March 17, 2020. Oral and nasal swabs from the dog tested positive for SARS-CoV-2 genetic material on March 18 and 19, 2020; rectal swabs from March 18, 2020 also tested positive, although with lower viral load than oral and nasal swabs. A second dog from the same household tested negative on five occasions between March 18 and 30, 2020. Both positive dogs showed no clinical signs of infection but produced measurable antibody titers, on March 3, 2020 for the Pomeranian and March 23,2020 for the German Shepherd. Sequencing of viral genetic material from both dogs showed that the viruses in both dogs were identical to those from their respective owners, but the viral clusters in the two households were distinct from one another.

Study Strengths

Compared to other studies focused on testing community dogs, the study focused only on dogs from COVID-19 patients to ascertain whether human-to-dog transmission can occur. The sequencing of the virus in the human cases and in the infected dogs was important for confirming human-to-dog transmission (rather than independent infection from a community source).

Limitations

With the very small sample size, it is difficult to determine the frequency of human-to-dog transmission of SARS-CoV-2, the possibility of dog-to-dog transmission, and the absence of clinical signs in infected dogs. The focus in this study on dogs from COVID-19 patients certainly overestimates the frequency of human-to-dog transmission relative to the larger population of dogs.

Value added

The evidence presented in this study demonstrating occasional human-to-dog transmission of SARS-CoV-2 confirms anecdotal reports of dogs becoming infected with SARS-CoV-2, and experimental transmission studies confirming that dogs are susceptible to SARS-CoV-2 infection, although to a lesser degree than cats.

Our take —

In this study available as a preprint and thus not yet peer reviewed, social distancing in the 25 U.S. counties with highest COVID-19 incidence was correlated with a decrease in the growth rate of COVID-19 cases, with an estimated 9 to 12 day lag period. Social distancing was measured via aggregated mobility data, and does not necessarily equate to a direct reduction in physical or social contact.

Study design

Ecological

Study population and setting

Data from mobile phone records (via Teralytics) were aggregated from January 1 to April 20, 2020, and used to quantify social distancing within the 25 counties in the United States with the highest number of reported cases of COVID-19 as of April 16, 2020. Social distancing was measured using a ratio comparing the number of individual trips made per day in each county (incoming, outgoing, or within) after January 24, 2020 to a baseline period over the last three weeks of January 2020. Reported changes in COVID-19 cases were compared to the social distancing ratio in each county using a linear regression model.

Summary of Main Findings

There were varying levels of social distancing behavior across states and counties, with many southern states displaying travel patterns closer to the baseline. Social distancing was observed to occur prior to state directives, suggesting either an early implementation of local directives or other incentives driving this behavioral change. A “lag” period of 11 (9-12) days was identified as the difference in time between the beginning of social distancing and the resulting reduction in cases. Social distancing was significantly correlated with a decrease in COVID-19 incidence in 23 of the 25 counties.

Study Strengths

Use of mobility data and number of trips, as opposed to inferred movement based on travel distances or transmission rates, captured real-time movement of people from the selected counties. Authors used averages over a few days to estimate baseline mobility and the growth rate of new COVID-cases, which is helpful when dealing with volatility in reported data. The 11-day lag window is consistent with the 4-5 day median incubation period of COVID-19. Results from a sensitivity analysis, using aggregated case reports, were consistent with the main findings.

Limitations

Mobile data is a coarse measure of movement, does not necessarily distinguish individual behavior, and does not capture the movement of non-mobile users. Links between reduced movement and changes in social or physical contacts remain unclear. Other mitigating factors in reducing case growth are ignored. The reported number of cases may have underestimated the actual prevalence within the counties due to testing limitations and reporting biases. Selected counties were those with the most cases in mid-April 2020, but may not be representative of later hotspots.

Value added

This study evaluates the impact of social distancing measures in a setting other than China, which has been the focus of most early literature surrounding the relationship between mobility and COVID-19.

Our take —

This study presents important data about COVID-19 in occupational settings at increased risk for infection. It found 3.0% (n=4,913) of 130,578 workers at meat and poultry plants that had reported infection, were confirmed COVID-19 cases. However, because this was an aggregate assessment, it is not possible to further assess whether certain plants are were particular risk. The qualitative analysis found infection challenges across structural, occupational, socio-cultural, and economic domains, and proposed a range of solutions.

Study design

Ecological; Other

Study population and setting

Across the US, 115 meat and poultry processing facilities in 19 states were reported to have COVID-19 infections from April 9 to 27, 2020. These states included Colorado, Delaware, Georgia, Illinois, Iowa, Kansas, Kentucky, Mississippi, Missouri, Nebraska, North Carolina, Ohio, Pennsylvania, South Dakota, Tennessee, Texas, Virginia, Washington, and Wisconsin. Pennsylvania had the most plants affected (n = 22), while Nebraska reported the highest number of total people working at affected plants (n=19,911). Poultry plants were the most commonly reported type across all states (n= 12), followed by beef plants (n = 10).

Summary of Main Findings

Among 130,578 total workers who worked at a plant with at least one SARS-CoV-2 infection, 3.0% (n=4,913) were confirmed COVID-19 cases. Of these cases, 0.4% (n=20) died of COVID-19-related causes to date. Qualitative data from facility risk assessments found a number of challenges to reducing exposure risk, including structural challenges (such as difficulty maintaining distance during breaks and at entrances/exits), operational challenges (such as difficulty adhering to face covering recommendations), socio-cultural challenges (such as difficulty communicating through language and cultural barriers), and economic challenges (such as incentivizing employees to work while ill).

Study Strengths

This study was strengthened by its scope across a number of states with reported infections. The study not only reported challenges, but proposed solutions for the qualitative challenges it found, which helps give guidance to potential policymakers and stakeholders.

Limitations

Data were limited across all of the states: for instance, Pennsylvania reported the most number of plants with infections, but did not report how many workers were affected, or the type of facilities. These aggregate counts do not allow for more granular assessments, such as whether a particular type of plant or geographic region was more likely to experience cases among workers. There also may be testing differences between facilities, or lag time in reporting to health departments that can affect counts.

Value added

This study is one of the first to report the number of workers at meat or poultry plants who were confirmed COVID-19 cases, and the rate of COVID-19-related deaths in this population.

Our take —

In this study, available as a preprint and thus not yet peer reviewed, transmission experiments of SARS-CoV-2 in fruit bats, ferrets, pigs, and chickens confirm that pigs and chickens are not susceptible to infection; fruit bats are susceptible but show transient viral shedding and limited transmission to contact animals; and ferrets are susceptible and capable of efficient transmission to contact animals. These results could guide the development of effective infection models in bats and ferrets to study viral shedding and possible vaccines. Human-to-bat transmission is a potential risk that could complicate the management of the pandemic, so contact with wild bats by researchers and wildlife managers should be limited during this time.

Study design

Ecological; Other

Study population and setting

The study population consisted of nine fruit bats (Rousettus aegyptiacus), nine ferrets, nine pigs, and seventeen chickens inoculated intranasally with SARS-CoV-2 (oculo-oronasally in chickens). Three direct contact animals per species were included 24 hours post-inoculation. The authors also tested the susceptibility of embryonated chicken eggs and three different cell lines from pigs typically used for virus isolation. All animals tested negative for SARS-CoV-2 genetic material and antibodies prior to the experiment. Viral shedding was tested from nasal washes and rectal swabs (ferrets), oral swabs and pooled feces (bats), nasal and rectal swabs (pigs) or oropharyngeal and cloacal swabs (chickens) between day 2 and day 21 post-infection. Two or three animals of each species were sacrificed on days 4, 8, and 12 days post-infection; all remaining animals were euthanized on day 21.

Summary of Main Findings

Pigs, pig cell lines, chickens, and chicken eggs were not susceptible to SARS-CoV-2 infection; fruit bats and ferrets were susceptible to infection, with limited pathological signs (predominantly rhinitis) in both animals. All inoculated bats became positive and oral swabs tested positive for viral genetic material between days 2 and 12 post-inoculation. Oral swabs from 2/3 contact animals were positive 8 days post-inoculation. Pooled feces from all bat cages also tested positive days 2 and 4 post-inoculation. All inoculated bats produced detectable SARS-CoV-2 antibodies starting at 8 days post-inoculation, and one contact bat at day 21, but titers were low. Eight of nine inoculated ferrets tested positive for SARS-CoV-2 between days 2 and 8 post-inoculation, and all three contact ferrets tested positive starting at day 8; rectal swabs also tested positive, although with lower amounts of virus than in nasal washes. Antibodies to SARS-CoV-2 were detected in all inoculated ferrets by 8 days post-infection, and by 21 days in all contact ferrets.

Study Strengths

The study corroborates results from Shi et al. finding that pigs and chickens are not susceptible to infection, and that ferrets are susceptible. This work improves on this study by including contact animals to demonstrate transmission among ferrets. This is the only study to date that investigates replication and transmission of the virus in bats.

Limitations

Based on previous work on SARS-related coronaviruses, fruit bats are probably not the reservoir hosts of SARS-CoV-2, but rather insectivorous Rhinolophus spp. However, the ease of maintaining fruit bats in captivity compared to insectivorous bats makes them a useful model for studying virus-host interactions and transmission, but may have limited relevance in the field and to our understanding of the ecological maintenance of the virus. The study only tested an intranasal route of infection in animals; the possibility of alternative routes including fecal-oral transmission will need to be assessed. Inter-species transmission was also not investigated.

Value added

The study confirms that fruit bats and ferrets are susceptible to SARS-CoV-2 infection, and can transmit the virus between animals of the same species. It provided experimental verification that bats can be infected. Fruit bats are not the primary hosts of SARS-related coronaviruses, so this could have implications for understanding the ecology of the virus and the risk of transmission from humans into atypical host species (therefore further complicating management). Ferrets are useful animal models for studying the efficacy of antivirals and vaccines.

Our take —

This non-peer-reviewed preprint describes a set of novel tools to rapidly identify mutations and possible recombination in SARS-CoV-2 and highlight mutations that may be worth examining more deeply. The authors show the utility of the tool by identifying the D614G mutation in the viral spike protein. However, they use unproven methods that do not properly control for other epidemiological factors to claim that the G614 variant is outcompeting other variants due to hypothetical biological advantages.

Study design

Cross-sectional; Ecological

Study population and setting

This study used SARS-CoV-2 viral sequences taken over time from around the world that are deposited into the GISAID database (gisaid.org). The authors used a sequence analysis pipeline to rapidly identify mutations in the virus over time, and then attempted to determine if there is evidence that viral variants possess any biologically relevant advantage, e.g., transmission potential, that might explain changes in their frequency in infected populations over time. Additional sequencing data from COVID-19 patients in Sheffield, England was used to compare changes in the frequency of viral variants over time with other patterns observed in other countries, and to test whether viral variants influence patient clinical outcomes.

Summary of Main Findings

The authors identified a mutation in SARS-CoV-2, D614G in the viral spike protein, that has now circulated widely in infected populations since its initial emergence in January 2020. The increase in frequency of the G614 variant over the D614 variant is shown in multiple countries and states in Europe, Asia, and North America, and is paralleled in the data from Sheffield, England. Within the Sheffield population, there was no significant association between D614G variants and clinical outcomes (hospitalized versus not hospitalized), but there was a significant association with the cycle threshold for viral detection in cases, suggesting potentially higher viral loads in patients carrying the G614 variant. The authors also identify potential recombination events in SARS-CoV-2 that could indicate the presence of mixed infections of viral variants that exchange genetic material.

Study Strengths

The dataset used for the study was extensive, and the researchers have significant experience working with large-scale viral sequence data. The pipeline to rapidly identify emerging mutations was robust and appears useful. In addition, pairing this mutational identification with preliminary structural analysis can also help to highlight mutations that may be important to examine further. The data from Sheffield was also important for supporting a biological explanation for the rapid increase in frequency of the G614 variant virus.

Limitations

The initial limitation of the technique described here is that it does not take into account other factors that might play a critical role in the spread of different viral variants globally. In particular, the authors claim that increasing frequency of the G614 variant is due to higher transmissibility. While this is supported by potentially higher viral loads in patients infected with the G614 variant, there are other equally plausible explanations for the observed patterns. One of these is a pervasive founder effect, wherein the G614 variant simply increased in frequency due to random chance after introduction into Europe from China, and was then reinforced by repeated introductions from Europe to other countries around the world. Without controlling for these other neutral factors through more complex modeling, many of the claims made in the paper on the emergence of “dominant” strains are unfounded. Although structural modeling is an interesting tool added to this analysis pipeline, it should only be used for generating questions to be explored functionality. The authors speculate that the mutations they identify may have critical biological advantages based on these structural models and some clinical data. However, the researchers would need much denser sequence data from patients over time or infection experiments in cell cultures or live animals to establish that the G614 variant is more transmissible. In addition, the results indicating lower cycle threshold values for detecting the G614 variant do not account for the timing of sample collection relative to symptom onset. Additional sampling or experiments will be needed to verify these results and establish whether the potentially higher viral loads in patients with the G614 variant translate into more transmission.

Value added

This study provides an interesting and novel series of tools for researchers to rapidly examine the large amount of SARS-CoV-2 sequence data that is being generated daily, and provides preliminary identification of possibly important mutations as they arise.