Skip to main content

Our take —

This preprint study, which had not been peer reviewed, reports that a 25-year old living in Reno, Nevada, without known immunodeficiencies, was infected with SARS-CoV-2 once in March, recovered in April, and was reinfected in May 2020. Genomic sequencing provides strong evidence that this was a reinfection. Both infections were symptomatic, and the second infection led to severe disease, requiring oxygen support and hospitalization; we have no data on the patient’s immune response following the first infection. This was the first reinfection reported from the USA, but reinfections have now been reported from Hong Kong and Europe. This adds to our understanding that reinfections with SARS-CoV-2 will occur, and that the second infection could be more severe than the first. However, these individual reports do not tell us what the risk for reinfection is, nor the risk of severe disease during the second infection. We understand little about how the virus changes over the course of a single infection and a better understanding of this phenomenon will help us distinguish between continuing and reinfection.

Study design

Case Series

Study population and setting

This is a case report of a 25-year old man residing in Reno, Nevada, USA.

Summary of Main Findings

The patient became ill on March 25, 2020 with symptoms compatible with COVID-19: sore throat, headache, cough, nausea and diarrhea. On April 18, 2020, the patient presented for care and had a nasopharyngeal swab collected which showed evidence of SARS-CoV-2 RNA. The patient reported complete resolution of symptoms on April 27 and was tested again on May 9 and May 27 by RT-PCR, and both tests were negative. On May 31, the patient reported to care with self-report of fever, headache, dizziness, cough, nausea and diarrhea. Five days later, on June 5, the patient presented again to care with hypoxia and was hospitalized; chest x-ray diagnosed atypical pneumonia, and the patient required oxygen support. A respiratory specimen collected on the day of hospitalization confirmed SARS-CoV-2 by RT-PCR, and a blood sample had evidence of IgM and IgG antibodies. The patient had no conditions or treatment that would suppress immune response. Sequence analyses of the viruses in the first and second infections suggests that they are not closely related, though both are consistent with other viruses circulating in Reno. The second infection was concurrent with a confirmed SARS-CoV-2 infection in a household member.

Study Strengths

The authors conducted two independent sequence analysis studies to increase their confidence that the patient had been infected twice. They also conducted an analysis of samples from both infections to confirm that the person who provided them was the same, to rule out mistakes in sample labeling as a reason for the result.

Limitations

There are no data presented on the immune response the patient had following the first infection, so it is impossible to determine if they developed neutralizing antibodies. This study is comprised of only one patient so the implications for this finding on risk of reinfection more broadly are unknown.

Value added

Evidence from other coronaviruses suggests that humans develop some immunity following infection but that reinfection is possible; a number of examples of reinfection with SARS-CoV-2 have now been confirmed using genetic sequencing, from Hong Kong and Europe. This is the first reported reinfection in the USA. This patient’s experience shows that the immune response from one infection with SARS-CoV-2 may not always produce sufficient protection against a second severe episode in a young person without known immune deficiencies.

Our take —

A 33-year old, healthy male in Hong Kong has been infected twice with SARS-CoV-2, ~142 days apart – the first such confirmed reinfection reported. He experienced mild symptoms during his first infection, and no symptoms during this second. Full genome sequence analysis was used to confirm that the viruses causing the two infections were distinct. It is possible that the first infection induced an immune response that prevented disease during the second infection, but no specimens are available to investigate this hypothesis. Despite a growing number of confirmed reinfections, the risk of reinfection with SARS-CoV-2 and the risk of disease during subsequent infections remain unknown.

Study design

Case Series

Study population and setting

This is a case report from a 33-year old male resident of Hong Kong.

Summary of Main Findings

The patient was previously healthy and became ill on March 23, 2020 and presented for care with productive cough, sore throat, fever and headache on March 26 where he had a sample collected that had evidence of SARS-CoV-2 RNA by RT-PCR. His symptoms resolved by March 29, though he was hospitalized – per protocol in Hong Kong – until April 14 when he had two negative RT-PCR tests 24 hours apart. One serum sample was collected from the patient ~16 days post symptom onset and no IgG antibodies were detectable. Following a visit to Spain, traveling via the United Kingdom, the patient was screened upon arrival at the Hong Kong airport on August 15, 2020 and had evidence of SARS-CoV-2 infection by RT-PCR. He was asymptomatic and was hospitalized again, per Hong Kong protocol. The cycle threshold (Ct) values of his repeat RT-PCR tests over the course of the second hospitalization increased, suggesting a decreasing viral load over the course of the second infection. A serum sample collected on the first day of the second hospitalization had no detectable IgG antibodies, but a sample collected 5 days post hospitalization did have IgG. Full genome sequencing confirmed that the two infections were caused by two distinct viruses, the second sequence matched the strain circulating in Europe at the time of his visit.

Study Strengths

The authors used full genome sequence analysis to show that the two infections were related to two distinct viral lineages in order to confirm that patient had two infections.

Limitations

The patient had not developed a detectable IgG response when tested during the first infection, but the sample was collected too early to rule out development of any IgG response during the first infection. It remains unknown whether the first infection conferred any protective antibodies.

Value added

This was the first confirmed reinfection with SARS-CoV-2. Reinfection should be expected, at least in some cases, based on understanding of waning immunity to SARS-CoV-2 over time and reinfection in other coronaviruses. The risk of reinfection remains unclear, despite a growing number of confirmed reinfection reports. In this example, the second infection was asymptomatic, which could be due to some level of immune response following the first infection that protected the patient from disease during the second.

Our take —

Healthcare workers, both those with direct patient contact and non-clinical staff, represented close to 10% of all confirmed COVID-19 cases in Los Angeles County. Half of these confirmed cases were from long-term care facilities, and nurses were the occupational group most affected. Overall, healthcare workers make up a substantial proportion of cases, but reported lower rates of both hospitalization and death compared to all cases. These findings provide evidence about how to direct funding related to personal protective equipment and other resources to reduce transmission.

Study design

Case Series

Study population and setting

From the start of the pandemic to May 31, 2020, Los Angeles County was notified of all confirmed COVID-19 cases and attempted to conduct standardized interviews with positive cases to determine occupational setting, occupational role, date of symptom onset, date last worked, known exposure, and whether or not the patient was hospitalized. Interviews were completed for 60% of all reported cases. Healthcare workers were defined as those working or volunteering in healthcare settings and included those that had both direct patient interaction and non-clinical staff.

Summary of Main Findings

During this period, a total of 57,118 confirmed cases were reported in Los Angeles County of which 5458 (9.6%) were confirmed among healthcare workers; 5118 of these (93.8%) had complete information. Half of all confirmed healthcare worker cases (46.6%) were from long-term care facilities; 27.7% were from hospitals; 6.9% worked at medical offices. Nurses represented about half of all cases (49.4%). Close to half had an unknown exposure to SARS-CoV-2 (45.1%); healthcare related exposures within their facility accounted about 44% of all exposures. Close to 70% reported working during their estimated infectious period. About 5% reported requiring hospitalization due to COVID-19, and 40 deaths (0.7%) were reported.

Study Strengths

This study represents all identified healthcare worker cases of COVID-19 since the start of the pandemic in a large urban setting in the United States.

Limitations

While the study presented the percentage of healthcare workers broken down by occupational setting and role, there is no discussion on the total number or proportion of these disaggregated groups, e.g. burden among hospital workers versus workers from long-term care facilities, or the total number or proportion of doctors versus the total number of nurses.

Value added

This study presents the relative burden of COVID-19 among different groups of healthcare workers in Los Angeles County, and could aid in resource allocation planning.

Our take —

In this study of children hospitalized with COVID-19 from 14 US states, Hispanic and Black children were far more likely to be hospitalized than white children, and they were more likely to have underlying conditions that are associated with disease severity. Although hospitalized children were as likely to be admitted to the ICU as hospitalized adults, a smaller proportion required mechanical ventilation, and only one of 576 children died. We have known for some time that children can have severe COVID-19, and this study highlights the ethnic and racial disparities in the risks faced by US children.

Study design

Case Series

Study population and setting

This study reported on 576 pediatric (<18 years: median age 8 years, 51% male) cases of COVID-19 requiring hospitalization in 14 U.S. states participating in a surveillance network (COVID-NET) from March 1 to July 25, 2020. Cases required a positive test for SARS-CoV-2. Patient data were abstracted from medical records by surveillance personnel. Denominators for hospitalization rates were calculated using post-census population estimates in the surveillance catchment areas from the National Center for Health Statistics.

Summary of Main Findings

Hospitalization rates increased in all age groups during the study period. Among cases with available data on race and ethnicity, 46% were Hispanic, 30% were Black, 14% were white, 5% were Asian/Pacific Islander, and 1% were American Indian/Alaskan Native. Compared to white children, hospitalization rates were 8 times higher among Hispanic children and 5 times higher among Black children. Nineteen percent of cases were infants below the age of 3 months. The highest rate of hospitalization was observed among children aged <2 years (24.8 per 100,000), followed by children aged 5-17 years (6.4) and children aged 2-4 years (4.8); the overall rate among all children <18 years was 8.0 per 100,000. Among the 222 (39%) children with available data, 42% had one or more underlying comorbidity: 38% were obese (defined as a BMI at the 95th percentile or greater for age and sex), 18% had chronic lung disease, and 15% of children <2 years of age had a gestational age <37 weeks at birth. Prevalence of underlying conditions was higher among Hispanic (46%) and Black (30%) children relative to white (15%) children. The most common symptom upon admission was fever and chills (54%); 42% of children had gastrointestinal symptoms. Nine of 83 children (11%) with enough data to support assessment for multisystem inflammatory syndrome in children (MIS-C) received an MIS-C diagnosis. Of 67 children with a chest radiograph, 44 (68%) showed an infiltrate or consolidation. Among the 208 (36%) children with complete chart review, the median duration of hospitalization was 2.5 days. Sixty-nine children were admitted to the ICU (median stay 2 days). Twelve (6%) children required mechanical ventilation, and one child died.

Study Strengths

This study reported data from a population-based surveillance network from 14 US states, and so data are likely to be broadly representative of the pediatric population in these areas.

Limitations

Case counts and rates are likely to be underestimates, since they depend upon laboratory confirmation of SARS-CoV-2 infection. Missing data meant that analysis of outcomes and comorbidities were based on smaller subsets of patients, which may not have been representative of the full population of hospitalized children. In particular, it is possible that the proportion of hospitalizations with ICU admission were overestimated if these records were more likely to be complete, and thus included in the analysis. Hospitalization rates for infants likely reflect increased testing and surveillance of neonates born to mothers with COVID-19, rather than increased disease severity in the <2 age group. Characterization of MIS-C was limited by 1) the requirement for laboratory-confirmed SARS-CoV-2 infection, and 2) the lack of systematic surveillance for MIS-C until late during the study period.

Value added

This study provides a broad, population-based picture of children in the United States hospitalized with COVID-19.

Our take —

Given extensive social distancing measures and rigorous testing and contact tracing, the secondary transmission rate from pediatric COVID-19 cases to secondary cases in the early phase of the South Korean epidemic was low. It is not clear how generalizable these findings are.

Study design

Case Series

Study population and setting

Between January 20 and April 6 2020, all identified pediatric (aged 18 years old or younger) index cases of COVD-19 (n=107) and their household contacts (n=248) in South Korea were identified and reviewed from the National Notifiable Disease Surveillance System. From these data, the secondary attack rate from children to their household contacts was calculated [(secondary case/traced N) x 100%]. Children were confirmed positive through RT-PCR testing, and household contacts were screened through RT-PCR and then placed on a 14-day quarantine regardless of symptoms

Summary of Main Findings

The median age of pediatric index cases was 15 years old (IQR: 10-17). The average number of household contacts per index case was 4.3 (range: 1-67). Of the 41 contacts who were confirmed to have COVID-19, 40 of those had the same exposure as the pediatric index case and were therefore not classified as secondary cases. Only 1 index case-secondary case pair was identified: a 16-year old index case who had returned from the UK to Korea and her 14-year old secondary case sibling, yielding a secondary attack rate of 0.5% (95% CI 0.0% to 2.6%).

Study Strengths

This study represents all identified pediatric index cases of COVID-19 in the early phase of the pandemic and thorough testing and tracing data are available through South Korea’s National Notifiable Disease Surveillance System.

Limitations

These data represent the role of children in household transmission during a period in which schools were closed. These findings should not be extrapolated to settings in which schools have reopened and children have greater mobility and risk of exposure outside the home. No comparison is provided to adults to see how different the results are by age. There is also potential for selection bias in that the cases were only those that were identified and reported. It is also unclear what precautions were being taken within households to prevent transmission.

Value added

This study presents findings on the secondary attack rate from pediatric index cases to household contacts during the early phase of the pandemic within a specific context, that of households in South Korea in the early stages of the epidemic.

Our take —

This is a case series, simply describing COVID-19 infection among passengers of a mid-length (280-minute) commercial flight, and interpretations of the findings should be limited as such. Overall, 7/24 members of a single tourist group tested positive, and 13 of the other 71 passengers that could be reached were tested for antibodies because they were sitting within 2 rows of a case or symptomatic; among those, an additional 2/13 cases were identified. Several major limitations make it impossible to draw any meaningful temporal conclusions between boarding the flight and infection with COVID-19 and other asymptomatic cases may have been missed.

Study design

Case Series

Study population and setting

The study describes COVID-19 infection among passengers of a 280-minute commercial flight that travelled from Tel Aviv, Israel to Frankfurt, Germany on March 9, 2020. 24 of the 102 passengers were members of a tourist group. A week earlier, the tourist group had contact with an individual who later tested positive for COVID-19. However, at the time of the flight, none had been diagnosed with COVID-19 and no protective measures (e.g., masks) were taken before or during the flight by the tourist group. After arrival at Frankfurt airport, a medical evaluation was conducted of the tourist group including collection of a throat swab specimen for COVID-19 testing. About 4-5 weeks later, all passengers were contacted by phone for interviews to assess symptoms, and a COVID-19 antibody test was offered to those passengers who sat within 2 rows of confirmed cases or who reported being symptomatic.

Summary of Main Findings

7 of the 24 members (29.1%) of the tourist group tested positive for COVID-19. However, only 4 were symptomatic during the flight; 2 were pre-symptomatic and 1 remained asymptomatic. 71 of the other 78 passengers who had been on the flight completed the phone interviews, and 13 of them provided serum samples within 6-9 weeks of the flight. COVID-19 antibodies were confirmed in 2 of the 13 individuals who provided serum samples.

Study Strengths

Indeterminate and positive results were verified using a plaque reduction neutralization test.

Limitations

No information was obtained on the airplane crew, and very few passengers outside of the tourist group had an antibody test (13 out of 78). No way to establish causality between boarding the flight and acquiring COVID-19, as all antibody tests were conducted after the flight.

Value added

Describes COVID-19 infection among passengers on an international commercial flight.

Our take —

Despite following the guidelines put forward regarding overnight camps in Georgia, there was an outbreak at an overnight camp that resulted in an estimated attack rate of 44% overall, among which three-quarters had symptoms. These findings are based on identified positive viral cases from the Georgia State Electronic Notifiable Disease Surveillance System. Those whose test results were not found were included and counted as negative, likely underestimating the attack rate. Attack rates were high across age groups, including 6-10 year-olds who had an attack rate of 51%, providing additional evidence of the susceptibility and potential for transmission among youth, including young children.

Study design

Case Series

Study population and setting

Between June 17 and 27, 2020, there was an outbreak of SARS-CoV-2 at an overnight camp in Georgia. Orientation among staff (n=138) and trainees (n=120) was from June 17 to 20, 2020, and the camp opened to campers (n=363) from June 20 to 27, 2020. As per the Georgia state guidelines, the camp required documentation of a negative viral test from all those attending 12 days or fewer days prior to arrival. Masks were required to be worn by staff members, but not required for campers. Campers were cohorted by cabin, attempts were made to limit mixing and windows were shut within cabins. However, singing and cheering that would normally take place was still allowed. A staff member became symptomatic on June 23 and left for testing, confirmed as positive on June 24, and the camp fully closed on June 27, 2020. The Georgia Department of Public Health recommended that all attendees be tested, and conducted an investigation of cases linked to the camp using the State Electronic Notifiable Disease Surveillance System and a specific investigation. Cases linked to the camp were defined as a positive viral RNA test from those attending from when they first arrived until 14 days after they left. Attack rates were calculated as positive test results over the total number of Georgia attendees.

Summary of Main Findings

In total, 597 Georgia residents attended the camp and were included in these analyses (27 out of state campers attended, but were not included as access to results was not available). The median age among staff was 17 years, and among campers was 12 years. Test results were obtained for 344 individuals, and among them 76% (260/344) were positive. The calculated attack rate was 44% overall (260/597). Given that Georgia often does not report back negative results, an assumption was made that those without results were negative. Estimated attack rates by age were: 6-10 years: 51%; 11-16 years: 44%; 18-21 years: 33%; 22-59 years: 29%. Among those testing positive and with symptom data (n=136), 26% (n=36) reported no symptoms.

Study Strengths

This study utilized a surveillance system to determine the attack rate associated with attending this overnight camp.

Limitations

There was an important assumption that any cases captured during this period were acquired as a result of camp attendance, while it is possible that cases were acquired during this period outside of the camp. Those who were not captured in the disease surveillance system were assumed negative in the calculation of the attack rate, likely underestimating the overall attack rate. Those who were not Georgia residents were excluded from these analyses (n=24).

Value added

The results presented here provide evidence for the rapid spread of SARS-CoV-2 from attendance at an overnight camp among children and youth.

Our take —

This report summarizes all publicly available data from state and local health departments on the burden of COVID-19 disease and mortality among children in the US as of July 30, 2020. Despite incomplete reporting, missing data, and divergent methods of reporting among states, it provides an informative description of the extent to which children in the US are diagnosed with, hospitalized, and die from COVID-19. While severe illness due to COVID-19 appears to be relatively rare among children, ongoing monitoring is warranted and the disaggregation of data by age category is important; periodic updates to this summary will be available, providing a useful service for clinicians, the public, and policymakers.

Study design

Case Series, Other

Study population and setting

This non-peer reviewed report provides a summary description of publicly available state-level data on COVID-19 cases, hospitalizations, and deaths among children in the United States. The age range used to define the population of children varied by state/territory; more than half included children aged 0-19 years and additional definitions from most to least commonly used were 0-17, 0-18, 0-14, 0-20, and 0-24 years. Cumulative data obtained from state and local health departments as of July 30, 2020 are presented. All states, except New York and Texas, provided age distributions of COVID-19 cases throughout follow-up, as did New York City, Washington, DC, Puerto Rico, and Guam. Additionally, data on testing, hospitalizations, and death was presented among states providing age distribution for these outcomes.

Summary of Main Findings

As of July 30, 2020, a total of 338,982 cases of COVID-19 were reported among children (8.8% of all documented COVID-19 cases; 447 per 100,000 US children), with 25 states reporting children comprising 10% or more of cases. Between July 16 and July 30, 2020, there was a 40% increase in pediatric cases (going from 241,904 to 338,982), with the majority of new cases occurring in the South and West. Of states reporting (n=8), 3.6-17.8% of children tested were positive. Among states reporting (n=20 and NYC), 0.6-8.9% of all COVID-19 cases among children resulted in hospitalization and 0-0.3% resulted in deaths among states reporting (n=44 and NYC).

Study Strengths

This report uses state- and local-level reports to summarize COVID-19 cases, hospitalizations, and deaths among children in the US.

Limitations

The report includes data that have been reported to local and state health departments, each with inherent limitations and marked differences in the content and format of data across states, yet little detail is given about the methods and limitations. The definition of “child” differed considerably across states, making it difficult to compare metrics between and among states. In addition, the age distribution of cases and hospitalizations was missing for some locations, including New York state (NYC only) and Texas (only reported for 8% of all cases), while Alabama did not disaggregate data below the age of 24 years. The report includes cases, hospitalizations, and death, and therefore does not inform about the rates of SARS-CoV-2 infection among children in the population; further, the proportion of hospitalizations and deaths is reported among cases rather than the population.

Value added

This study is among the first to date to summarize the distribution of COVID-19 cases, testing, hospitalizations, and mortality due to COVID-19 among children in the US using available reports from state and local health departments.

Our take —

In a representative sample of 2,640 individuals from two parishes in Louisiana from May 9 to 15, 2020, 7.8% of individuals tested positive for SARS-CoV-2. There were disparities in seroprevalence by race, with Black participants experiencing a seroprevalence of 9.7% compared with 4.5% among white participants. The overall infection fatality ratio was 1.63%, and there were no significant differences by race. The infection fatality rate was higher than that observed in other settings, and should be interpreted carefully given the limited data; these data warrant further examination to better understand what may be leading to this high infection fatality ratio.

Study design

Case Series

Study population and setting

A representative community sample of 2,640 persons participated in a study designed to estimate the seroprevalence and the infection fatality ratio in Orleans and Jefferson parishes, Louisiana, USA. The study was conducted at 10 sites from May 9 to 15, 2020. The recruitment method ensured representation within the sample across more than 50 characteristics, including social determinants of health and demographics. Both real-time RT-PCR of nasopharyngeal swabs and qualitative IgG antibody blood tests were conducted. Participants were considered to be infected with SARS-CoV-2 if they tested positive on either est. Early-stage infections were excluded from the calculation of the infection fatality ratio.

Summary of Main Findings

Overall, 183 participants tested positive for SARS-CoV-2 (6.9% unadjusted, 7.8% census-weighted). Seroprevalence ranged from 4.5% among white participants to 9.8% among Black participants. There was significant geographic heterogeneity in seroprevalence. The overall infection fatality ratio (IFR) was 1.63%, and statistically similar for white, Black, and multiracial participants.

Study Strengths

The sample is reflective of the demographics and a range of characteristics of the target population (the populations of Orleans and Jefferson Parishes). Additionally, to ensure representativeness, the final estimates were weighted to account for any remaining differences between the sample and target population.

Limitations

While this is a representative sample, the results of this study are primarily relevant for two parishes in Louisiana in May, 2020. In the current analysis, data presented were not disaggregated by symptom status or presented by testing results (e.g. PCR-positive only vs. PCR-positive and IgG-positive vs. IgG-positive only); the case fatality ratio was similarly not reported among those who had been symptomatic.

Value added

This study provides estimates of seroprevalence, and infection fatality ratio for two parishes in Louisiana.

Our take —

Preliminary results from this study suggest that recent and short-term proton pump inhibitor (PPI) use may be associated with more severe outcomes of COVID-19, however, causality may not be inferred as a number of biases could explain the observed association (i.e., residual confounding, selection bias, and reverse causation); concerns around reverse causation are more pronounced as the statistically significant effect of PPI use was only seen among short-term users (<30 days). PPI use was not associated with testing positive for SARS-CoV-2.

Study design

Case Series, Retrospective Cohort

Study population and setting

Data from the Korean national health insurance claims database were used to evaluate whether use of proton pump inhibitors (PPI) (a class of medication to reduce production of stomach acid) influence susceptibility to SARS-CoV-2 infection (primary outcome) or severe clinical complications of COVID-19 (admission to an intensive care unit, invasive mechanical intervention, or death). The study included 132,316 individuals who received COVID-19 testing between January 1 and May 15, 2020. Patients were excluded if they were <18 years, were prescribed H-2 blockers in the previous year, or had a record of NSAID use in the previous month. Given that participants were not randomized to PPI use, statistical techniques were employed that attempted to account for possible differences in factors between those who used PPIs and those who did not, specifically potential factors that might influence the propensity for using PPIs (i.e., propensity score matching). Four rounds of 1:1 propensity score matching were carried out using the greedy nearest-neighbor approach. Among all patients tested for COVID-19, current PPI users (PPI use within past 30 days) were matched to non-users (13,873 pairs), and past PPI users (PPI use 31-365 days before COVID-19 test) to non-users (6,153 pairs); to examine the risk of severe clinical outcomes, this was repeated only among individuals who tested positive (267 pairs for current PPI users vs. non-users, and 148 pairs for past PPI users vs. non-users). To see if duration of PPI use affected any association, the authors repeated the analysis dividing current PPI users into short-term users (<30 days of use) and long-term users (>30 days of use)

Summary of Main Findings

Of the 132,316 individuals who received SARS-CoV-2 testing (mean age 48 years, 51% male, 3.6% test positivity), 14,163 used PPIs in the prior 30 days (current use), 6242 used PPIs in the past, and 111,911 had no history of PPI use. In the final results, current or past PPI users did not experience a greater risk of testing positive for SARS-CoV-2 infection relative to non-users. However, adjusted matched analyses suggested that current PPI use, especially among those with short term use of PPIs (<30 days), was associated with increased likelihood of severe COVID-19 outcomes; there was not, however a statistically significant increase in severe Covid-19 outcomes among long-term PPI users (>30 days).

Study Strengths

The study uses data from a nationwide cohort of individuals tested for SARS-CoV-2 infection in Korea. The authors use sophisticated techniques (i.e., propensity score matching) with multivariable analyses to address some potential sources of confounding.

Limitations

This was an observational study and it is likely that differences between the PPI users and non-users remain even after efforts were taken to account for these potential confounding factors. A key concern is whether individuals who show mild COVID-19 symptoms, which can include gastrointestinal involvement, are prescribed PPIs prior to COVID-19 diagnosis; if so, there may be an issue of reverse causation, where COVID-19 causes prescription of PPIs, rather than PPIs causing severe clinical outcomes. Additionally, participants with missing data and individuals who were unmatched in propensity score matching were excluded from multivariable analysis, which may limit generalizability or introduce a bias if those who participated were different from those who did not participate regarding likelihood of being prescribed PPIs, susceptibility to SARS-CoV-2 infection, or likelihood of experiencing severe disease. Use of PPIs and other medications was determined by electronic health record of prescription and may not reflect actual usage of medications. Finally, relative to the number of variables adjusted for in the multivariable analyses, the number of severe outcomes was low; this may result in overfitting and the observation of spurious association when none actually exist.

Value added

This is one of the first large-scale studies to investigate the association between PPI use and susceptibility to SARS-CoV-2 infection and severe clinical outcomes of COVID-19 disease.