Skip to main content

Our take —

This study, conducted in long-standing cohorts of HIV-negative and HIV-positive individuals, suggests that there is little difference between those with and without HIV in terms of their symptoms when testing positive for SARS-CoV-2. There was some evidence that people living with HIV may be at increased risk for infection, but some potential confoundin factors remain unaddressed.

Study design

Cross-Sectional

Study population and setting

The study was conducted in the Multicenter AIDS Cohort Study (MACS) and Women’s Interagency HIV Study (WIHS), two cohorts which include HIV positive patients in the United States. The authors conducted a telephone survey among participants in the MACS and WIHS cohorts from April to June 30, 2020 to identify participants who had been infected by SARS-CoV-2 by asking about 14 individual COVID-19 symptoms, their duration and severity, and SARS-CoV-2 testing and results (with medical record verification when possible). They used data from the existing cohort to identify factors associated with diagnosed SARS-CoV-2 infection, including having HIV.

Summary of Main Findings

The authors were able to survey 83% of the cohort participants they contacted (3411/ 4123). Of those, 2078 had HIV and 1333 did not. Among participants surveyed, only 13% (441/4123) had been tested for SARS-CoV-2 with similar proportions among those HIV positive and HIV negative. Nearly all participants reported practicing social distancing (98%) and staying home (97%) and these were similar among those with and without HIV. Those with HIV were more likely to test positive for SARS-CoV-2 (11.2%) than those who were HIV negative (6.1%), but not significantly so. Of all subjects, 53% reported at least one symptom. Reported symptoms were also similar between the two groups in terms of type and severity. Overall, 441 (12.9%) reported testing for SARS-CoV-2, with no differences between those with and without HIV. Among those who were tested, for the odds of testing positive for SARS-CoV-2 were higher among those who were HIV positive compared to negative (adjusted odds ratio (aOR) 2.22 95%CI: 1.01, 4.85) and those living with others versus living alone (aOR 2.95 95%CI 1.18, 7.40).

Study Strengths

The study was conducted in two well documented and described cohorts and as such, the underlying population is easy to identify. They were able to compare those who were HIV positive to those who are HIV negative with confidence due to the underlying cohort design, something few other cohorts are able to do. SARS-CoV-2 diagnoses were confirmed with medical records, where possible.

Limitations

Fourteen percent of cohort members were unable to be reached for the survey, and it’s possible that some participants may have been unreachable due to severe illness. In addition, participants likely had some control over whether or not they were ever tested for SARS-CoV-2, and it’s possible that some infections may have been missed. There is also likely some uncontrolled confounding even after adjustment.

Value added

There is currently little information about the risk for SARS-CoV-2 infection and severe disease among those living with HIV. This study offers unique insights from existing, well-documented cohorts to show that severity of COVID-19 appears similar between those with and without HIV.

Our take —

This study, which has not been peer reviewed, measures components of immune memory at various time points from COVID-19 symptom onset. Despite a small number of repeated measurements, they provide preliminary data that SARS-CoV-2 infection — assuming that it behaves similarly to other pathogens that induce analogous immune responses — can induce sustained immunity to either prevent or lessen the severity of subsequent infection.

Study design

Cross-sectional; Prospective Cohort

Study population and setting

This preprint used cross-sectional data from 185 individuals and longitudinal data from 38 of those individuals to measure SARS-CoV-2-induced antibodies, memory B cells, memory CD8+ T cells, and memory CD4+ T cells from six to 240 days after COVID-19 symptom onset (median 90.5 days). Participants had a documented history of SARS-CoV-2 infection via PCR, serodiagnostics, or both, and lived mostly in California or New York in the United States. Only 7% of participants were hospitalized with COVID-19, 2% of patients were asymptomatic, their ages ranged from 19-81, and participants were more likely to be female than male (57% vs 43%). The authors measured circulating antibody (IgG and IgA), memory B-cells (including surface immunoglobulin isotype), CD8+ T cells, and CD4+ T cells specific to various components of the SARS-CoV-2 virus via assays that included negative and positive controls for each immune component of interest. They created lines of best fit via linear regression or kinetic models for both cross-sectional and longitudinal data and assessed differential immune memory by gender with ANCOVA to assess trends in immune memory after SARS-CoV-2 infection.

Summary of Main Findings

They found that IgG antibody titers tended to decrease slightly from symptom onset with 90% being seropositive at 6-8 months vs 98% at 1 month, regardless of antigen specificity type, although there was high variability in antibody titers across individuals. Memory B cells increased steadily the further individuals got from COVID-19 symptom onset regardless of the antigen type, plateauing around 150 days, with immunoglobulin isotype switching to predominantly IgG+ memory B cells as time progressed. Memory CD8+ and CD4+ T cells, on the other hand, were detectable in a higher percentage of people at earlier measurements (61% and 94% at one month respectively), after which they decreased marginally to 50% and 89% respectively among the 18 participants with data available after six months from COVID-19 symptom onset. These trends seem to hold in individuals who provided data at more than one time point across all immune components. They further homed in on CD4+ T follicular helper cells — cells essential for activating memory B cells that then produce antibodies in response to an antigen — which were stable among tested individuals for at least six months from symptom onset. Finally, they found no significant differences in immune memory by gender or hospitalization status.

Study Strengths

This study provides an in-depth assessment of both circulating antibodies and memory B- and T-cells following symptom onset of natural COVID-19 infection. A subset of their participants provided data at more than one time point, which allowed for comparisons across time in individuals.

Limitations

It appears that in calculating lines of best fit for various antibodies and immune cells, the authors combined data from participants who only provided data at one time point with those who provided data at more than one time point. This introduces data points that are not independent from one another, which biases the lines of best fit in unforeseen ways. Furthermore, it is difficult to compare antibody or immune cell titers across time in different individuals, since, as the authors note, each person’s immune system reacts to pathogens (such as SARS-CoV-2) differently. The small sample size used also makes it difficult to generalize these data to all individuals infected with SARS-CoV-2. Lastly, it is also unknown how immunological responses to SARS-CoV-2 will differ between those who are naturally infected and individuals who are vaccinated.

Value added

The authors provide evidence, through a rather thorough examination of various immune cell subtypes and responses, for persistent SARS-CoV-2 immunity or reduced disease severity following SARS-CoV-2 infection.

Our take —

Of the 766,044 inbound travelers screened for COVID-19 at 15 designated US airports between January 17 and September 13, 2020, 35 symptomatic individuals were referred for diagnostic testing, and nine COVID-19 cases were identified. Low case yield (one identified case per 85,000 screened travelers), resource and personnel requirements, and incomplete data capture render port of entry screening for infections like COVID-19 highly ineffective.

Study design

Cross-sectional

Study population and setting

On January 17, 2020, the United States Centers for Disease Control (CDC) implemented passenger screening at designated airports for travelers arriving by air from Wuhan, China. Enhanced traveler screening for COVID-19, which involved elicitation of travelers’ potential exposures and illness symptoms, was expanded to inbound travelers from areas with sustained, widespread COVID-19 transmission, including mainland China (February 2); Iran (March 2); Schengen areas of Europe, the United Kingdom, and Ireland (March 14 to 17); and Brazil (May 27). After screening, passengers’ information was forwarded to local health departments at their respective travel destinations. Passengers presenting with signs and symptoms of COVID-19 were isolated and referred to local health facilities for further assessment, including COVID-19 testing when available. The authors assessed performance of the screening program by describing the volume of travelers screened and the proportion of COVID-19 cases identified through enhanced screening at U.S. ports of entry.

Summary of Main Findings

Of the 766,044 travelers screened at 15 designated airports between January 1 and September 13, 2020 (the date traveler monitoring was suspended), 298 travelers (0.04%) were referred for follow-up assessment due to self-reported exposures or symptoms of COVID-19 infection. Forty symptomatic travelers were assessed at local health facilities, 35 of whom received COVID-19 testing by RT-PCR. Nine COVID-19 cases were identified of the 35 travelers tested, yielding a case identification ratio of 1 per 85,000 travelers screened. Fourteen additional COVID-19 cases were identified through other non-screening mechanisms, including diagnosis after travel to the United States or retroactive notification of positive test results prior to travel. Passenger information was forwarded to health departments for roughly 68% of screened travelers.

Study Strengths

The authors calculate the number of COVID-19 cases identified as a fraction of total travelers screened at designed U.S. airports from January 17 to September 13, 2020.

Limitations

Given the 14-day incubation period for COVID-19, passenger screening was likely to miss pre-symptomatic or asymptomatic cases; the true number of COVID-19 infections in screened passengers was, therefore, likely underestimated. As the authors aim to evaluate the airport-based screening program’s performance, inclusion of other performance metrics, like health workforce capacity (i.e., ratio of screening personnel to screened travelers), could have helped contextualize the reported findings.

Value added

This study is the first to evaluate the performance of the CDC’s traveler monitoring program for COVID-19 at designated U.S. ports of entry.

Our take —

Compared to the number of laboratory-confirmed COVID-19 illnesses and deaths in other parts of the world, fewer than expected cases and deaths have been reported from sub-Saharan African countries. Using samples from blood donors and Bayesian models, this study provides evidence from Kenya to suggest that transmission in this population has been similarly high to other countries, particularly in urban areas (population-weighted seroprevalence overall was 4.3% and varied from 5.5 – 8.0% in urban areas). Given this finding, the most likely reasons for the lower than expected COVID-19 case reports from Kenya are under-ascertainment from surveillance, milder disease due to younger average age, or hereto unmeasured t-cell mediated immunity from other coronaviruses.

Study design

Cross-Sectional

Study population and setting

3,098 blood transfusion samples from donors aged 15 – 64 years from four Kenya National Blood Transfusion Service regional transfusion centers collected from April 30 – June 16, 2020 were tested for IgG antibodies against SARS-CoV-2. The study used an ELISA test validated by the authors, with a reported sensitivity of 92.7% and specificity of 99.0%. They used Bayesian models to take the crude seroprevalence estimates from this select group of blood donors and adjust to estimate the population-based seroprevalence of Kenya, also taking into account the performance characteristics of the ELISA test used.

Summary of Main Findings

The crude seroprevalence for all participants was 5.6% (174/3098) but was higher among participants from urban areas. The adjusted national seroprevalence estimate was 4.3% (95% CI 2.9 – 5.8%), which remained higher in urban counties of Mombasa (8.0%), Nairobi (7.3%), and Kisumu (5.5%). These seroprevalence estimates are similar to other countries with much higher reported incidence of laboratory-confirmed cases and deaths, and suggest that lower than expected case and death counts in Kenya are unlikely due to differences in number of people infected, unless blood donors included in this study have much higher seroprevalence than others in the population. Other possible contributors to the lower than expected reports of COVID-19 illnesses and deaths in Kenya could be under-ascertainment of these events during surveillance, reduced risk for severe illness and death due to the younger average age in this population, or cell-mediated immunity which is not reflected in seroprevalence estimates.

Study Strengths

Rather than rely on the crude seroprevalence estimates, the authors adjusted for the age, sex and geographic distribution of the underlying population, as well as the performance characteristics of the ELISA test, to estimate the true seroprevalence in this population.

Limitations

The blood samples tested in this study came from blood donors, which were not reflective of the general population; they were more likely to be young adult males. It is possible that their risk for past infection with SARS-CoV-2 may not reflect broader risks in the underlying population.

Value added

The number of laboratory-confirmed COVID-19 cases and deaths has been lower than expected in sub-Saharan Africa, and this study provides evidence to show that this may not be due to a lack of transmission since the proportion of adults tested with antibodies against SARS-CoV-2 is similar to other studies from countries on in North America and Europe.

Our take —

This study, available as a preprint and thus not yet peer reviewed, found that there was a shift in the burden of COVID-19 cases between the first (February-mid-July 2020) and second wave (mid-July-October 2020) of the pandemic based on electronic medical record data from all individuals aged 20-70 years old in Norway. During the first wave, healthcare professionals had the highest odds of infection, however individuals working in the food service industry (bartenders, waiters, etc.) and travel stewards were disproportionately affected during the second wave. Transportation workers had higher odds of infection compared to others of working age during both waves. These data provide comprehensive descriptive information about the shifting demographics of the pandemic and can be used to inform current and future responses to the pandemic.

Study design

Cross-Sectional

Study population and setting

Using data from an emergency preparedness register (BEREDT C19 register) that captured electronic patient records, surveillance data, and employment data for all Norwegian residents (both native and immigrants), this study examined differences in COVID-19 diagnoses and hospitalizations by profession. COVID-19 diagnosis was defined as a positive PCR test or a positive ICD-10 diagnostic code of COVID-19 (U07.1). To be included in the study, individuals had to be between 20 and 70 years old on January 1, 2020, and have a unique personal identification number. Tourists, temporary workers, and asylum seekers were excluded. Occupation was recorded based on the Norwegian Standard Classification of Occupation (STYRK-98) and was aligned to international designations based on the Standard Classification of Occupations. The focus of this study was on occupations having direct contact with others, e.g. health, teaching, sales, cleaners, food service, travel and transportation, recreation and beauty. The first wave of COVID-19 infection was defined as being from February 26 to July 17, 2020, and the second wave from July 18 to October 20, 2020. Crude and adjusted multivariable logistic regression were used to assess the association between occupation and: 1) COVID-19 diagnosis, and 2) hospitalization. Age, sex, and country of birth were adjusted for in the adjusted model.

Summary of Main Findings

A total of 3.5 million individuals were included in this study, of which 0.3% were diagnosed with COVID-19 (n=12,736). During the 1st wave, health care professionals (nurses, physicians, dentists) and transportation works (bus/tram and taxi drivers) had a significantly greater odds of infection (1.5-3.5 times the odds of infection) when compared to everyone else of working age. During the 2nd wave, food service industry workers and transportation workers, again, were disproportionately affected (bartenders, waiters, taxi drivers, travel stewards) with 1.5-4 times the odds of infection. In both waves, there was limited association between occupation type and odds of hospitalization.

Study Strengths

The use of a national emergency preparedness register, linked to employment records, allows for the examination of the entire working age population of Norway for the duration of the pandemic.

Limitations

There is limited information provided on how testing guidelines and protocols evolved over time, which would be of particular importance here given that the study aims to compare those who tested positive during two different time periods. Overall, 24.4% of the sample could not be categorized using the available registry data (e.g. unemployed or not registered with any occupation). There is the possibility that the distribution of occupational status may be different among unregistered individuals which may influence the interpretation of findings. Additionally, the low number of hospitalizations in some professions created very wide confidence intervals, and the limited association between occupation and severe COVID-19 (hospitalization) should be interpreted with caution.

Value added

This study provides valuable descriptive data on occupational groups most impacted by COVID-19 over the course of the pandemic.

Our take —

Although in-person voting for the U.S. presidential election concluded on November 3, 2020, there will be a handful of runoff elections around the U.S. through January 2021, including a high-profile election in Georgia. This study surveyed 522 election poll workers to describe measures to mitigate SARS-CoV-2 transmission in the Delaware statewide primary election, highlighting those with reportedly high uptake (e.g., physical spacing of voting booths) and those with room for improvement (separate portals for entry and exit, incorrect mask use by poll workers and voters, large number of close contacts experienced by poll workers). Differences between survey respondents and those who did not respond, as well as social desirability bias, may have affected the findings.

Study design

Cross-Sectional

Study population and setting

This study reported on a survey of 522 poll workers (median age 59 years, 57% male) who served in the US state of Delaware’s primary election on September 15, 2020. The survey was focused on SARS-CoV-2 transmission mitigation measures undertaken at polling locations, and included questions about supply availability, knowledge, attitudes, training, and directly observed behavior among respondents, other poll workers, and voters.

Summary of Main Findings

Of the 2,498 poll workers who participated in the primary election, 1,595 (64%) were invited to participate via email, and 522 (21%) responded and were deemed eligible. Among the 128 respondents who identified a polling location, 99 distinct polling locations were represented. Thirty-two percent of respondents had at least one underlying condition associated with higher risk of severe COVID-19. Most respondents reported spacing of voting booths at least 6 feet apart (88%), unidirectional movement of people through the location (80%), and visual reminders to remain at least 6 feet apart (84%). Only 45% of respondents reported separate doors for entry and exit. Few respondents reported physical barriers (e.g., plexiglass) at registration desks (5%) and between voting booths (7%). Most respondents reported that hand sanitizer was available for poll workers (94%) and voters (82%), as were masks (88% and 70%, respectively). Eighty percent of respondents reported receiving COVID-19 training, and 92% answered all three COVID-19 survey questions correctly. The majority of respondents (72%) reported contact (within 6 feet) with more than 100 voters, while 27% reported close contact (within 6 feet for 15 minutes or more) with more than 100 voters. Nearly all respondents reported mask use by most (80-100%) poll workers and voters; 73% reported never or rarely witnessing incorrect mask use by poll workers, while 54% reported the same for voters.

Study Strengths

Survey questions addressed not just knowledge and attitudes, but also directly observed behaviors among voters and poll workers, along with physical characteristics of polling locations.

Limitations

Respondents may differ from all Delaware poll workers in meaningful ways; for example, if respondents worked at a subset of polling locations with more extensive SARS-CoV-2 transmission control protocols, results could overestimate the statewide adoption of these protocols. Similarly, only poll workers with valid email addresses were included, who may be different from those without email addresses. Poll workers may have had investment in the success of transmission control measures and their answers may have been subject to social desirability bias, which would result in an overestimate of the uptake of distancing, mask-wearing, etc. Alternatively, poll workers who were upset with non-adherence to control measures may have been more likely to respond.

Value added

This is the first study to focus on measures to reduce SARS-CoV-2 transmission in voting locations during a U.S. election, and it may be relevant to planning for upcoming elections.

Our take —

A survey of U.S. adults conducted over three waves (April, May, and June 2020) found increasing self-reported mask use but declining prevalence of other behaviors that have been recommended by the CDC to reduce SARS-CoV-2 transmission, such as avoiding crowded places and keeping six feet of distance from others. Age was consistently associated with these self-reported behaviors: older adults reported that they engaged in each behavior more often than younger adults. These results support other evidence that younger adults are less likely to adhere to transmission mitigation behaviors; this age group represents a target for public information campaigns. Potential biases such as non response, social desirability and recall bias should be considered in the interpretation of these findings.

Study design

Cross-Sectional

Study population and setting

This study reported on an online and telephone survey of 6,475 U.S. adults at least 18 years old, conducted over three waves from April 20 to June 8, 2020, that focused on personal measures taken to mitigate SARS-CoV-2 transmission. The study population was drawn from a national probability sample, and results were weighted to adjust for nonresponse and over- and under-representation. The survey asked “Which of the following measures, if any, are you taking in response to the coronavirus?” and provided 19 possible responses. These possible responses included six that were aligned with CDC recommendations and White House guidelines (mask use, hand washing/sanitizing, six feet of distance from those outside the household, avoiding public or crowded places, canceling/postponing social/recreational activities, avoiding some or all restaurants).

Summary of Main Findings

Across survey waves, 50% of respondents from the weighted sample identified as female, and the majority of respondents (62-65%) identified as non-Hispanic/Latino white. More than 40% of respondents reported all six of the government-recommended behaviors in each survey wave. Self-reported mask use increased from 78% in April to 83% in May and 89% in June. Each of the other five recommended measures declined (except avoiding restaurants, which did not change) from April to June. In each survey wave, respondents 60 years and older reported the highest prevalence of each of the six government-recommended mitigation behaviors, and those aged 18-29 reported the lowest prevalence. Older adults also reported more cumulative mitigation behaviors than younger adults at all time points. Those who reported not using masks also reported low prevalence of the other five mitigation measures, and their overall engagement in mitigation measures decreased over time.

Study Strengths

The survey was conducted over three waves, allowing for assessment of temporal trends in self-reported behaviors.

Limitations

Survey response rates were low (19.7-26.1%) and biases in participation cannot be ruled out. Respondents might have overreported mitigation behaviors for reasons of social acceptability. Answers are also subject to recall error, which could bias results in either direction. Survey questions did not ask about frequency or duration of mitigation behaviors, and were not specific with regard to the circumstances in which the behaviors applied: for example, “wore a face mask” did not specify the type of mask worn, how often it was worn, or where it was worn. This may have resulted in an overestimate of mitigation behavior prevalence.

Value added

This paper adds to evidence from other U.S. surveys that younger adults are less likely to adhere to behaviors recommended to reduce the risk of SARS-CoV-2 transmission.

Our take —

Previous studies have raised concerns about increased coagulopathy and stroke risk among patients with COVID-19. In this observational study, patients with COVID-19 had considerably lower odds of stroke than other hospitalized patients (OR 0.25). However, selecting only hospitalized patients may have induced bias into the results that was not completely addressed by the statistical methods; moreover, strokes may have been under-diagnosed in patients with COVID-19. Prospective studies using a wider study population are necessary to untangle the relationship between COVID-19 and stroke risk.

Study design

Cross-Sectional

Study population and setting

This study considered 24,808 patients (53% female) who were discharged from six hospitals within a large health care system in New York State from January to April 2020. The authors compared the occurrence of new-onset acute ischemic stroke between discharged patients with COVID-19 (with laboratory confirmed SARS-CoV-2 infection) and discharged patients without COVID-19, using a logistic regression model adjusted for age, sex, race, comorbidities, insurance status, and hospital. The outcome of stroke was restricted to those with symptoms and MRI findings consistent with acute ischemic stroke. Sensitivity analyses used propensity scores to address confounding variables.

Summary of Main Findings

During the study period, 2,513 patients (10.1%) tested positive for SARS-CoV-2 infection. Of these patients, 22 (0.9%) presented with acute ischemic stroke, and among the patients without SARS-CoV-2 infection, 544 (2.4%) presented with stroke. The mean age of those with stroke was 73 years. Those with COVID-19 had lower odds of stroke than other hospitalized patients; the adjusted odds ratio for stroke associated with COVID-19 was 0.25 (95% CI: 0.16 to 0.40). These results were similar in sensitivity analyses that used propensity scores to adjust for confounding. Those with stroke and COVID-19 had 10.5 times the odds of death compared to those with stroke but without COVID-19 (95% CI: 3.5 to 31.2).

Study Strengths

The authors attempted to address possible bias by using propensity scores to weight observations, determined by a model including a broad range of comorbidities.

Limitations

In observational studies like this one, there is a danger of inducing bias by selecting only hospitalized patients, which can create a non-causal association between variables. In this case, because patients could be hospitalized for COVID-19 or for other reasons, those without COVID-19 may have a higher risk of stroke as a consequence of selection, making it seem as if COVID-19 is protective for stroke (this phenomenon has been seen in previous studies of smoking and COVID-19). While propensity score weighting may be helpful in reducing this bias, it can only eliminate the bias if the mechanism for selection into the study is properly characterized, which is not assured in this case. Patients with severe respiratory distress on mechanical ventilation may not have been assessed for stroke, and coagulopathy was not necessarily recognized as an important sign of severe COVID-19 early in the pandemic, which could have led to an undercount of strokes or transient ischemic attacks in patients with COVID-19. This undermines the conclusion that stroke risk is lower among patients with COVID-19. The timing of stroke diagnoses was also not clear (e.g., whether stroke was only diagnosed at presentation to hospital or during admission). Finally, the biological plausibility of lower stroke risk in COVID-19 is questionable.

Value added

This study provides evidence that contradicts the prevailing impression of increased stroke risk among patients with COVID-19.

Our take —

This study compared the age distribution among individuals <18 years old across different categories pertaining to SARS-Cov-2 infection using a large, nationwide registry of COVID-19 in Colombia. Children with mild or asymptomatic SARS-CoV-2 infection were older in age compared to those with severe symptomatic disease. Also, children who recovered from SARS-CoV-2 infection and did not require hospitalization were older in age compared to those who were deceased or hospitalized. Reported cases were likely biased towards more severe disease due to limited testing. Inadequate identification of asymptomatic SARS-CoV-2 infection at younger ages could have biased the results, resulting in an underestimation of SARS-CoV-2 infection in the Colombian NIH database; thus, the findings may not represent the true distribution of SARS-CoV-2 infection across the general population of Colombia.

Study design

Cross-Sectional

Study population and setting

The PEDIACOVID study was a cross sectional study that described the clinical characteristics of COVID-19 infection among those <18 years-old in Colombia. The study used a National Institute of Health (INS) database to identify cases with laboratory-confirmed COVID-19 infection from March 6, when the 1st case was identified, through June 16, 2020. Out of 54,971 confirmed COVID-19 participants, 5062 (9.2%) were <18 years old. Confirmed cases were categorized as asymptomatic or symptomatic. The latter was further categorized into four ordinal categories; mild, moderate, severe or deceased. Confirmed asymptomatic cases in the database were defined as positive PCR tests within 14 days of unprotected exposure to confirmed cases or positive serological antibodies at 11 days or more from exposure to a confirmed case. Confirmed symptomatic cases were defined as positive PCR tests within 14 days of having symptoms or positive serological tests at 11 days or more from having symptoms. Mild symptomatic cases were defined as having mild upper respiratory tract symptoms. Moderate symptomatic cases had dyspnea or tachypnea, but did not require any oxygen supplementation; severe symptomatic cases had pulmonary radiographic findings, elevated serum ferritin, LDH, D-dimer, lymphopenia or thrombocytopenia. All COVID-19 cases with comorbid conditions were categorized as severe. The study compared the age distribution across categories of disease severity and participants’ place of care (home versus general hospital ward versus intensive care unit (ICU)). Comparisons were made using an ANOVA test and a post-hoc Tukey’s test to adjust for multiple comparisons.

Summary of Main Findings

Of the 5,062 cases <18 years of age (49% male), 4022 (80%) were classified as mildly symptomatic, 854 (17%) asymptomatic, and only 8 cases were deceased (0.16%). The majority of the cases were treated at home (2886 [57%]), while 146 (2.8%) were treated in general hospital wards and only 26 (0.5%) in an ICU. The study found a statistically significantly higher mean age in the asymptomatic and mildly symptomatic cases (9.4 years) compared to those who were moderately, severely symptomatic and deceased (6.3, 4.9 and 2.9 years respectively).

There was a statistically significant difference in mean age between those who were treated at home compared to those who were treated in hospitals or ICU (9.3 vs. 6.1 and 4.9, respectively). Participants who recovered were significantly older than those who died (9.3 vs. 2.9 years of age). These results were similar when the data were stratified by the three most-affected regions in Colombia.

Study Strengths

This study used a large national database of confirmed COVID-19 to examine infection and disease severity among individuals less than 18 years old in Colombia. The study adds critical information in regards to the age distribution of severe COVID-19 infection among children.

Limitations

The study categorized any children with comorbid conditions as severe COVID-19 cases regardless of their actual clinical severity of SARS-CoV-2 infection, which may have resulted in a spurious association between age as a risk factor and severe SARS-CoV-2 infection if these comorbidities were more common in younger ages (the authors did not mention the distribution of these comorbidities across different ages in children).

The study used a large national database in which participants had to have medical attention to be tested for SARS-CoV-2 virus, either due to exposure to other COVID-19 subjects or due to having symptoms. Therefore, the sample was not randomly selected from the population. It is plausible that children across different ages with asymptomatic disease were not tested, and thus, were not included in this database. If the distribution of these asymptomatic, untested subjects differed across age groups, the age comparison between asymptomatic and severely symptomatic disease would not be valid. It is unlikely, however, that this would bias the comparison between the deceased and the recovered subjects, unless the database did not capture COVID-19 related deaths in older ages of childhood. Given the descriptive nature of the study, no formal hypothesis testing or adjustment was conducted.

Value added

The study provides information about the age distribution across different categories pertaining to the severity of COVID-19 infection among children (<18 years) using a large, nationwide COVID-19 registry in Colombia. The study results could be potentially extrapolated to other developing countries in South America.

Our take —

This cross-sectional study, available as a preprint and thus not yet peer reviewed, described symptoms and evidence of organ impairment up to four and a half months from acute SARS-CoV-2 infection or a clinical diagnosis of COVID-19 among relatively low-risk participants at two healthcare centers in England. They documented multi-organ MRI changes (lung, heart, pancreas, kidney, liver, and spleen) in up to ⅔ of patients, adding to the growing literature that symptoms and organ abnormalities can linger several months after initial COVID-19 diagnosis. While these findings may have implications for patient management, the lack of pre-diagnosis data means they do not establish a causal relationship between SARS-CoV-2 infection and long-term organ impairment.

Study design

Cross-Sectional

Study population and setting

This cross-sectional pre-print reports the residual impact of COVID-19 in 201 adults (mean age 44 years, 70% female) in southern England at a median follow up of 140 days from their initial symptoms at the baseline visit of an ongoing prospective cohort study. Participants were recruited from clinics in Oxford and London, England between April and August 2020 if they had a history of a positive SARS-CoV-2 PCR (n=62), a positive antibody test (n=63), or a clinical diagnosis of COVID-19 from two independent physicians (n=73) and were excluded if they had current COVID-19 symptoms, a COVID-19 hospitalization in the last 7 days, and/or contraindications to magnetic resonance imaging (MRI) at the time of enrollment. The authors assessed lung, heart, kidney, liver, pancreas and spleen function at follow up using validated symptom assessment scales, fasting laboratory values, and MRI. They compared participants by hospitalization status while symptomatic using Wilcoxon tests, Fisher exact tests, or Spearman correlation as appropriate and created a multivariable model to assess risk factors of a previous hospitalization among participants.

Summary of Main Findings

Of the 201 participants, 20% and 18% of whom reported pre-existing obesity and asthma respectively, 99% were experiencing more than three and 42% were experiencing more than nine COVID-19 symptoms at study enrollment, which occurred a median of 140 days (interquartile range (IQR) 105-160) from participants’ initial COVID-19 symptoms. The most common reported symptoms included fatigue (98%), muscle ache (88%), shortness of breath (87%), and headache (83%), and 52% of participants reported persistent problems resuming usual activities. Participants who were hospitalized with COVID-19 were more likely to have abnormal triglycerides, cholesterol, LDL-cholesterol, and transferrin saturation than those who were not. Organ dysfunction on MRI was also more common among participants who were hospitalized, with evidence of lung (33% of all participants), heart (32%), pancreas (17%), kidney (12%), liver (10%), and spleen (6%) dysfunction on MRI in 66% of participants. Multivariable logistic regression suggested that increasing age (OR=1.06, 95% Confidence Interval (95% CI) 1.02-1.10), liver volume (OR=1.18, 95% CI 1.06-1.30), and multiorgan impairment on MRI (OR=2.75, 95% CI 1.22-6.22) were associated with prior hospitalization adjusted for sex and BMI.

Study Strengths

This study includes a moderate number of participants with few comorbidities and describes well-measured symptom, laboratory, and MRI evidence of the persistent impacts of COVID-19 a median of four and a half months after initial symptoms.

Limitations

It is unclear how participants were approached for inclusion in this study, which could select for individuals still experiencing COVID-19 symptoms who are likely different from those who have recovered without lingering symptoms. This selection bias would likely artificially amplify the prevalence of the reported findings. Furthermore, we cannot conclude that the virus caused the laboratory and/or imaging findings without data from before participants were infected with SARS-CoV-2. Additionally, it is impossible to contextualize the abnormal laboratory and imaging findings without a control group of similar adults who were not exposed to SARS-CoV-2. Finally, it is difficult if not impossible to interpret a model that predicts an outcome (previous hospitalization) that occured before the recorded covariates.

Value added

This is one of the first studies to document the presence of symptoms and organ impairment about four and a half months after initial confirmation of SARS-CoV-2 infection or clinical diagnosis of COVID-19.