Skip to main content

Our take —

This retrospective study found that hospitalization rates, ICU admission rates, and use of mechanical ventilation were similar between individuals (mostly children) diagnosed with COVID-19 and seasonal influenza at a single children’s hospital in Washington DC. However, patients presenting for these two diseases may have differed in a number of ways that influenced outcomes. These findings should be explored in larger studies that adjust for potential confounding factors.

Study design

Retrospective Cohort

Study population and setting

This retrospective cohort study included 315 individuals diagnosed with laboratory-confirmed SARS-CoV-2 infection from March 25 to May 15, 2020, and 1,402 individuals diagnosed with influenza A or B from October 1, 2019 to June 6, 2020 at Children’s National Hospital in Washington, DC. COVID-19 patients were identified from an infectious disease database, and influenza patients were identified from the laboratory informatics system. Testing for COVID-19 and influenza were performed on US FDA-approved PCR assays. Asymptomatic COVID-19 patients who tested positive during pre-admission of pre-procedural screening were excluded.

Summary of Main Findings

Of the 315 COVID-19 patients (52% male, median age: 8.4 years [range: 0.03-35.6 years]), 54 (17%) were hospitalized, 18 (6%) were admitted to the ICU, and 10 (3%) received mechanical ventilation (average duration 10.1 days). These proportions were similar to those seen among the 1,402 influenza patients (53% male, median age: 3.9 years [range: 0.03-40.4 years]): 291 (21%) were hospitalized, 98 (7%) were admitted to the ICU, and 27 (2%) received mechanical ventilation (average duration 7 days). None of these differences in outcomes were statistically significant. Hospitalized COVID-19 patients were older (n=54, median age 9.7 years, 37% ≥15 years) than hospitalized influenza patients (n=291, median age 4.2 years, 6% ≥15 years). None of the patients with COVID-19 died, and only 2 patients with influenza (0.1%), both with influenza A, died. Patients hospitalized with COVID-19 (vs. influenza) were also more likely to have an underlying medical condition (65% vs. 42%), and more often presented with a fever (76% vs. 55%), cough (48% vs. 31%), diarrhea or vomiting (26% vs. 12%), body aches (22% vs. 7%), chest pain (11% vs. 3%), or headaches (11% vs. 3%). The age distribution and symptoms of influenza A and B patients were similar.

Study Strengths

The study compared COVID-19 and influenza patients drawn from the same general population over similar time periods.

Limitations

This was a single-site study, and results may not be generalizable to other settings. Although most patients were children, it is unclear why the study included a small number of adults; results were not disaggregated by age group. None of the statistics presented were adjusted for confounders, and confounding is likely. Due to limited sample size, comparisons between COVID-19 patients and influenza patients who were admitted to the ICU or who received mechanical ventilation were not presented. Testing for influenza was available throughout the study period, but the test positive rate decreased dramatically following school closures on March 15, 2020 (there was only one case after March 22), and it is possible that there are unmeasured differences in the populations who received testing for influenza before and during the COVID-19 pandemic. Comparisons of patients hospitalized for influenza and flu may be influenced by ascertainment bias if testing practices or symptom screening differ based on an individual’s diagnosis. Similarly, the thresholds for a hospital visit may have differed before and during the pandemic, which could bias comparisons of hospitalization rates and outcomes between influenza and COVID-19.

Value added

This is one of the first studies to compare clinical features between COVID-19 and seasonal influenza in children.

Our take —

In this cohort study of 175 patients hospitalized with mild COVID-19 in Shanghai, China, there was wide variation in the level of SARS-CoV-2 specific neutralizing antibodies at discharge, with levels generally reaching their peak 10 to 15 days post symptom onset, and then decreasing following discharge. Interestingly, the same factors that are frequently reported as associated with worse COVID-19 outcomes, including age, male sex, lower lymphocyte counts, and higher CRP levels, were also associated with increased higher neutralizing antibody titers.

Study design

Retrospective Cohort

Study population and setting

This cohort study includes 175 patients with PCR-confirmed SARS-CoV-2 infection who were admitted to Shanghai Public Health clinical Center from January 24 to February 26, 2020, and were classified as having mild symptoms, including fever, respiratory symptoms, and radiologic evidence of pneumonia. Patients were followed up to 2 weeks post discharge or March 16, whichever came earlier. Plasma samples collected at the time of discharge were analyzed to measure SARS-CoV-2 specific neutralizing antibodies (NAb) and levels were compared to 13 healthy controls. Eleven participants additionally had sequential plasma samples collected between hospital admission and discharge in 2 to 4 day intervals.

Summary of Main Findings

Among the 175 patients (median age 50 years, 53% female) with mild COVID-19, none required ICU admission, median length of stay in the hospital was 16 days, and 165 (94%) had developed SARS-CoV-2 specific NAb by the time of discharge; the patients who did not develop NAbs were generally younger (median age 35 years) and 80% were women. Of the 165 who developed SARS-CoV-2 specific NAbs, 30% had low levels of NAb titers (ID50 <500), 17% medium (500-999), 39% medium-high (1000-2500), and 14% high (>2500). NAb levels were higher on average among men (1417 vs. 905), and appeared to increase with age, length of hospital stay, and disease duration. Two of the main factors associated with worse COVID-19 outcomes, low lymphocyte counts and higher CRP counts, were also moderately associated with higher NAb titers at discharge. Among 117 patients with NAb titer from 2 week post discharge, levels had significantly decreased from discharge values (886 vs. 1110), and patients who hadn’t had detected detectable NAb levels at discharge did not go on to develop them. Among 11 patients with serial NAb titers during hospitalization, levels began to increase within 4-6 days of disease onset and peaked 10-15 days post-disease onset.

Study Strengths

This study explores a number of clinical factors associated with SARS-CoV-2 specific NAbs upon discharge from the hospital. Presence of NAb titers was compared to healthy controls, and validated.

Limitations

Despite having NAb titers at discharge for 175 patients, kinetics of NAb during hospitalization were only available for 11 patients, and 2 week follow-up was only available for 117 patients; each of these analyses warrant additional study in larger samples with longer follow-up. Though all patients were classified as mild, they were all hospitalized exhibiting symptoms of fever, respiratory illness, and radiological evidence of pneumonia, and therefore the results may not be generalizable to patients who present with fewer or no symptoms. No adjusted analyses were conducted, and confounding is highly likely. This was a single site study and the degree to which results are generalizable to other populations is unknown.

Value added

This is one of the first studies to characterize clinical characteristics associated with NAb levels among patients with mild COVID-19.

Our take —

In a large cohort of patients from a single health care network in California, obesity had strong associations with mortality, but only among men, and only among those younger than 60 years. These associations were large, and the differences by age and sex are suggestive, but should not be interpreted as causal.

Study design

Retrospective Cohort

Study population and setting

This study assessing obesity as a risk factor for COVID-19 mortality included 6,916 members (55% female, mean age 49 years) of Kaiser Permanente Southern California (KPSC) who were diagnosed with COVID-19 from February 13 to May 2, 2020. Cases could be diagnosed clinically or via laboratory-confirmed SARS-CoV-2 infection. Pregnant women were excluded because of non-comparable BMI. The authors used Poisson models, including age- and sex-stratified versions, adjusted for possible confounding variables to estimate the association between BMI and mortality. BMI was categorized as underweight (<18.5 kg/m2), normal (18.5-24 kg/m2), overweight (25-29 kg/m2), obese class I (30-34 kg/m2), obese class II (35-39 kg/m2), obese class III (40 or greater kg/m2), and those with 45 or greater kg/m2. The outcome was death within 21 days of a COVID-19 diagnosis. Potential confounding covariates were considered at both the individual and neighborhood levels; calendar time was also considered to account for any changes in testing, social distancing, and clinical treatments over the study period. Authors selected covariates based on bivariate associations with mortality; no neighborhood-level covariates were included in final models.

Summary of Main Findings

The mean BMI was 30.6 kg/m2. Fifty-four percent of patients were Hispanic, 78% of patients lived in census tracts with median annual household incomes below $80,000, and 8% received Medicaid. The most common comorbidities were hypertension (24%), high cholesterol (23%), and diabetes (20%). Of the 206 (3%) patients who died within 21 days of diagnosis, 67% were hospitalized and 43% were intubated prior to death. Of survivors, 15% were hospitalized and 3% were intubated. In the primary adjusted model, there was a J-shaped relationship between BMI categories and the risk of mortality. Obese class III (relative risk [RR]: 2.68, 95% CI 1.43-3.02) and BMI ≥ 45 (RR: 4.18, 2.12-8.26) were associated with higher risks of mortality relative to normal weight. These associations, particularly those in the highest BMI categories, were larger in magnitude than those observed for most comorbidities. In separate sex-stratified and age-stratified analyses, associations between BMI categories and mortality were restricted to men, and to those 60 and younger. Race and ethnicity were not associated with mortality in adjusted models.

Study Strengths

The study drew from a very large cohort of patients in one health care system, including both hospitalized and non-hospitalized COVID-19 cases, which allowed the authors to consider a wide range of clinical, demographic, and treatment covariates that were measured with a high degree of standardization.

Limitations

Covariates were chosen for the adjusted models by virtue of their associations with mortality; many of these covariates (such as myocardial infarction and hypertension) are plausibly on the causal pathway between BMI and mortality. Adjusting for these covariates could introduce bias in effect estimates. This possibility is enhanced by the large number of comorbidities included in the models. Similarly, effect estimates for race and ethnicity should not be interpreted as net causal effects, as models were not constructed with the aim of estimating these. Additionally, patients with more severe disease tended to have more complete data.

Value added

This study provides further evidence that obesity is a risk factor for severe COVID-19.

Our take —

This study objective was to describe the prevalence of SARS-CoV-2 among pregnant women presenting to the hospital for delivery. Using universal testing, 375 women were tested for infection across four hospitals, of which 71 tested positive (18.6%), and 64 of those were asymptomatic (70.3%). Although recall bias related to earlier symptoms may lead to some misclassification between symptomatic vs. asymptomatic cases, the majority of women with SARS-CoV-2 infections detected were asymptomatic despite pregnancy-related immunosuppression. This study highlights the need for greater testing in hospital settings serving pregnant people at delivery.

Study design

Retrospective Cohort

Study population and setting

The primary study objective was to examine whether the risk of SARS-CoV-2 infection was associated with prior COVID-19 signs and symptoms among peripartum women. 403 women admitted for delivery across 4 hospitals in New York between April 2 and April 9, 2020 who were potentially eligible for enrollment. Women were excluded for antepartum admissions where no delivery occurred and postpartum readmissions, for a total eligible cohort of 382, of whom 375 were tested under the universal testing protocol. Specimens were obtained within 1 hour of hospital admission, or within 48 hours of admission for scheduled cesarean or induction, and tested for SARS-CoV-2 via RT-PCR tests. Clinical and laboratory data were extracted from electronic medical records (EMRs) and reviewed manually for history of COVID-19 signs and symptoms.

Summary of Main Findings

Of the 375 women tested for SARS-CoV-2, 64 newly tested positive (17.1%), and 7 were previously tested and diagnosed. Thus, the overall prevalence was 18.6% (71/375 women). 64 of those testing positive were asymptomatic (70.3%). The most common symptoms reported among symptomatic patients were cough (57.9%, 11 of 19 symptomatic patients), fever (52.6%, 10 patients) and dyspnea (47.4%, 9 patients). Maternal age and race/ethnicity varied between the four sites, with non-Hispanic white women making up 42.2% of the overall cohort (N=156), followed by Hispanic/Latina women (19.5%, N=72). PCR positive tests also varied between sites, with Southside Hospital reporting the highest prevalence (28.8%, 15 of 52 patients), and North Shore University Hospital representing the lowest (8.8%, 11 of 128 patients).

Study Strengths

The study used a universal testing protocol which was able to identify both symptomatic and asymptomatic patients. It also included all women who delivered at any of the four hospitals, which likely results in a representative sample that may be generalizable to other populations with similar demographics. Using chart review, they were able to capture signs and symptoms based on medical examinations, as opposed to recall alone, which may have resulted in more valid classification.

Limitations

The study used electronic medical records, which may or may not have always taken comprehensive symptom assessments in a structured format, and therefore may be subject to limitations based on what information was previously collected. Participants may have potentially over-reported symptoms out of an abundance of caution, which would lead to an over-estimation of the number of symptomatic cases. The authors also mention women may not have reported symptoms due to fears around COVID-19 diagnosis during this time, which would lead to an underestimation in the number of symptomatic cases. Misclassification from self-report could lead to bias in either direction. Also, these results reflect active COVID-19 infection at delivery, and does not show prior infection among pregnant women.

Value added

This paper estimates the prevalence of infection among pregnant people presenting at hospitals for delivery, providing important information for providers and hospital infection control.

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.

Our take —

There had been early concern that patients with hypertension might be at risk for more severe COVID-19 if they continued taking ACE inhibitors or ARBs. However, several studies have failed to show any increased risk associated with the use of these antihypertensive medications. In this single-center study, hyptertensive patients who continued to use ACE inhibitors or ARBs after COVID-19 hospitalization experienced lower mortality and rates of ICU admission compared to those who discontinued their use. Although these patients may not have been comparable in unmeasured ways, this study provides additional support for keeping patients on these medications while in the hospital.

Study design

Retrospective Cohort

Study population and setting

This study included 614 patients with a history of hypertension who were hospitalized with PCR-confirmed SARS-CoV-2 infection in a single hospital in New York State from February 7 to May 23, 2020. Patients were categorized into three groups: 1) those who had not been taking angiotensin-converting enzyme inhibitors (ACEi) or angiotensin II receptor blockers (ARB) at home (n=279), 2) those who had been taking these medications at home but discontinued in the hospital (n=171), and 3) those who continued taking these medications in the hospital (n=164). The primary outcome of interest was death in hospital, and the secondary outcome was ICU admission. Authors extracted demographic and clinical data from medical records at hospital admission. Multivariable logistic regression was used to compare those who continued vs. those who discontinued ACEi and ARB use.

Summary of Main Findings

Among non-users of ACEi and ARB (group A), 62/279 patients died (22%); among those who discontinued ACEi and ARB in the hospital (group B), 48/171 patients died (28%); and among those who continued using ACEi and ARB in the hospital (group C), 10/164 patients died (6%). Adjusting for age, sex, history of heart disease, COPD, and asthma, those in group C had lower odds of mortality than those in group B (OR: 0.22, 95% CI: 0.10-0.46). In adjusted models, those in group B and C combined had no statistically significant difference in odds of mortality compared to those in group A (OR: 0.81, 95% CI: 0.53-1.24). Because those discontinuing ACEi and ARB use had higher incident hypotension and acute kidney disease (AKI), each of which may have indicated discontinuation, analyses were also stratified by hypotension and AKI. In these analyses, continuing ACEi and ARB use was associated with lower mortality relative to discontinuation among patients without hypotension and among patients without AKI. ICU admission rates were also lower in group C relative to group B in both unadjusted and adjusted analyses; the protective association between continued ACEi/ARB use (vs. discontinued ACEi/ARB use) and ICU admission persisted in the subgroup of patients without AKI and moderately in the group with hypotension.

Study Strengths

This study stratified by two important potential confounders, the development of hypotension and acute kidney disease, both of which are reasons for discontinuation of ACEi and ARB medications.

Limitations

There may have been determinants of ACEi and ARB discontinuation that were not adequately controlled in analyses. The timing of discontinuation among patients was not presented or analyzed, though this may have influenced clinical outcomes. ACEi and ARB medications were not disaggregated. The sample sizes, particularly in stratified analyses, were not large.

Value added

This study provides additional evidence that use of ACE inhibitors and ARBs are not associated with more severe COVID-19 outcomes, even when their use is continued after hospitalization.

Our take —

Contrary to previous studies, blood type A was not associated with testing positive for COVID-19. Blood types AB and B were associated with increased odds of testing positive for COVID-19, while blood type O appeared slightly protective. None of the blood types were associated with intubation or death. The results may not be widely generalizable, as only patients with recorded blood type were included, and residual confounding may exist.

Study design

Retrospective Cohort

Study population and setting

The goal of this study was to examine whether there is a relationship between ABO blood typing and testing positive for COVID-19 or severity of COVID-19 disease. The population included symptomatic adults at five major hospitals in Massachusetts (Partner’s Healthcare System’s Research Patient Data Registry) who were tested for COVID-19 and had blood type recorded. Disease comorbidities were extracted from the medical record based on ICD-9/10 codes and medication history was reviewed over the previous year. The main outcome of interest was COVID-19 severity, defined as intubation or death.

Summary of Main Findings

Among the 7648 symptomatic patients who received a COVID-19 test during the study period, 1289 (16.9%) tested positive for COVID-19, and of those, 162 (12.6%) were intubated or died. In multivariable analyses, adjusting for sex, primary language, age, and rhesus factor with all other blood types as the reference, blood types AB (OR: 1.28, 95% CI: 1.108-1.52) and B (OR 1.37, 95% CI: 1.02-1.83) had higher odds of positive tests, and blood type O (OR=0.84, 95% CI: 0.75-0.95) had a lower odds of testing positive; Rhesus factor was also associated with testing positive (OR=1.22, 95% CI: 1.00-1.50). None of the blood types or Rhesus factor were associated with COVID-19 disease severity (intubation or death) in multivariable analysis.

Study Strengths

This was a relatively large, multi-site study. No data were missing for demographics, comorbidities, or medications.

Limitations

Only participants with recorded blood type were included in the study; how many patients without blood type information were excluded is not reported. Although this missingness is likely random (unrelated to blood type) and would not largely impact the observed association between blood type and COVID-19 test results or disease severity, it may limit the generalizability of the findings, as patients with recorded blood type may be different than patients without recorded blood time in regards to demographics or medical history. The duration of patient follow-up is not reported, and it is possible that some included participants went on to develop severe COVID-19.

Value added

This study adds to a growing body of literature about the relationship between ABO blood type and COVID-19 testing resultings, and is one of the first to examine the relationship between ABO blood type and disease severity.

Our take —

Physiological data derived from wearable technology in 24 US COVID-19 patients revealed latent physiological disturbance patterns—including heightened heart rate, decreased physical activity, and increased sleep duration—prior to symptom onset. A wearable technology detection algorithm correctly identified physiological abnormalities associated with pre-symptomatic COVID-19 infection in two-thirds of COVID-19 patients prior to symptom onset.

Study design

Retrospective Cohort

Study population and setting

Investigators collected survey data and physiological markers, generated from wearable technologies (e.g., Fitbits, Smart Watches), through a smartphone application from a cohort of 5,262 US-based individuals. Investigators integrated participants’ physiological and activity data from wearable technologies with other self-reported metadata (demographics, medical history, daily COVID-19 symptoms, and COVID-19 testing/diagnoses) to: 1) identify physiological changes associated with COVID-19 infection and 2) determine precision with which wearable technologies could detect these physiological changes by, or prior to, symptom onset.

Summary of Main Findings

Using Fitbit data from enrolled participants (n = 24) with self-reported COVID-19 diagnoses and complete physiological markers (from 14 days prior to symptom onset to at least 7 days after), COVID-19 diagnosis was associated with increased heart rate (median: 7 beats/minute increase) 3-7 days before symptom onset. Decreases in daily steps and increased sleep duration were observed primarily in pre-symptomatic periods but following onset of resting heart rate signals associated with COVID-19 illness. There was high variability between individuals’ physiological markers and the progression/severity of COVID-19 illness. Based on these findings, investigators developed an algorithm detecting abnormal resting heart rates associated with pre-symptomatic COVID-19 infection. The algorithm detected 67% of COVID-19 cases prior to symptom onset in 24 participants supplying 28 days of physiological data ahead of symptom onset.

Study Strengths

The authors collated large quantities of physiological data, collected in pre- and post-symptomatic periods among participants with COVID-19 infections, to characterize latent physiological markers of early COVID-19 disease. Additionally, investigators drew from various measurement approaches and statistical modeling techniques to appraise the robustness of their findings.

Limitations

Despite recruiting a large participant cohort, inferences from this study are drawn only from 24 participants with complete Fitbit records and self-reported COVID-19 diagnoses. As information about participants’ activities or behaviors was limited, some observed physiological changes could be attributed to unmeasured events (i.e., stress, other illness), rather than pre-symptomatic COVID-19 infection. In some cases, the time interval between the detection of physiological aberrations and symptoms onset stretched credulity (e.g, 15 days) given the 5 day median incubation period of COVID-19.The high volume of incomplete physiological records indicates participant data may not be missing at random, given individuals with more severe illness symptoms may have temporarily discontinued use of wearable technology. Lastly, Fitbit technology is not a gold standard for measurement of specific physiological markers and could bias the magnitudes of association reported.

Value added

This is the first study to characterize pre-symptomatic physiological changes in individuals with COVID-19 using wearable technology data, and to determine the precision with which wearable technologies, like Fitbits, can detect these physiological changes before symptom onset.

Our take —

Growing evidence suggests that COVID-19 leads to a hypercoagulable and prothrombotic state that may result in a greater risk of neurologic complications such as ischemic stroke. This study is one of the first to indicate that the risk of ischemic stroke is greater with COVID-19 than with influenza, a similar viral respiratory infection that is a known stroke risk factor. Additional investigation of the specific thrombotic mechanisms associated with COVID-19 is warranted to identify effective and timely treatment and prevention strategies.

Study design

Retrospective Cohort

Study population and setting

The objective of this study was to compare the probability of acute ischemic stroke associated with COVID-19 relative to influenza, another viral respiratory infection. This study compared adult patients (>18 years old) with laboratory-confirmed SARS-CoV-2 infection who were hospitalized or visited the emergency department between March 4 and May 2, 2020 with adults who were hospitalized or visited the emergency department with laboratory-confirmed influenza A/B between January 1, 2016 and May 31, 2018 (covering both moderate and severe seasons) at 2 NYC hospitals. Data for the participants with influenza were obtained from the Cornell Acute Stroke Academic Registry (CAESAR). Acute ischemic stroke was confirmed by CT or MRI for both COVID-19 and influenza cohorts. Data on demographics, risk factor, presenting symptoms, illness severity (i.e. ICU admission), treatments, and laboratory and imaging results were obtained by electronic medical records.

Summary of Main Findings

A total of 1916 patients (median age: 64; 57% male) with laboratory-confirmed SARS-CoV-2 infection presented to the two hospitals during the period of study. Of these, 17% required mechanical ventilation and 1.6% (n=31; median age: 69 years) had an acute ischemic stroke. The median duration from symptom onset to stroke was 16 days with an inpatient mortality of 32% (14% among those without stroke). A total of 1486 patients (median age: 62 years; 45% male) with influenza were included; 3% required mechanical ventilation and 3 patients (0.2%) had an acute ischemic stroke. After adjusting for age, sex, and race, patients with COVID-19 were significantly more likely to experience an acute ischemic stroke than the patients with influenza (OR: 7.6; 95% CI: 2.3-25.2). This persisted after adjusting for vascular risk factors and illness severity (OR: 4.6; 95% CI: 1.4-15.7).

Study Strengths

This is one of the first studies to indicate that COVID-19 is associated with a greater likelihood of acute ischemic stroke than other viral respiratory illnesses, specifically influenza, after accounting for other potential risk factors. The study reports that diagnosis of acute ischemic stroke was the same for the COVID-19 and influenza cohorts; both were based on clinical and imaging data. Numerous sensitivity analyses were conducted reinforcing the robustness of the findings.

Limitations

Given the use of historical comparison groups, it is likely there were differences in the criteria for hospital admission, ascertainment of infection status, and diagnosis of ischemic stroke that may have influenced the results; given the prevailing interest in neurologic complications among COVID-19 patients, it is possible they were more often screened for and, thus, diagnosed with ischemic stroke than influenza patients of years past, which could partially explain the observed increased rate of stroke among COVID-19 patients relative to influenza patients. Additionally, the testing criteria for COVID-19 changed considerably over the course of the study period, and hospital burden during the COVID-19 pandemic may have resulted in a population of sicker patients than previous influenza seasons. These measurement and selection issues could result in either an over or under-estimation of the observed differences in stroke outcomes between COVID-19 and influenza. The study population was limited to 2 hospitals in New York City, potentially limiting generalizability. Data are limited to ill patients who visited an ER or were hospitalized and findings may not be generalizable to patients with less severe disease. Only 34 patients combined had the outcome of interest, and adjustment for covariates may overfit the data and overestimate the magnitude of the association. Additionally, the use of logistic regression does not account for censoring due to death or hospital discharge.

Value added

This is one of the first and largest studies (to date) to compare the burden of ischemic stroke associated with COVID-19 relative to influenza by drawing on stroke outcomes among patients with influenza from previous influenza seasons (including both moderate and severe seasons).

Our take —

The substantial limitations with this study prevent drawing any conclusions regarding whether HCQ is harmful or helpful among patients hospitalized with COVID-19. These findings should not be used to guide decisions regarding HCQ use. It does highlight the need for rigorous randomized controlled trials in assessing the effectiveness of putative treatments, such as HCQ.

Study design

Retrospective cohort

Study population and setting

The investigators sought to measure an association between hydroxychloroquine (HCQ) use and in-hospital mortality in a retrospective cohort of consecutive adult patients admitted with COVID-19 to a 6-hospital health system in Michigan. Patients were included if they were admitted for at least 48 hours and diagnosed with SARS-CoV-2 from March 10 to May 2, 2020. COVID-19 treatments were protocol driven. The following independent variables were abstracted: age, gender, race, preexisting medical conditions, mSOFA (modified sequential organ failure assessment) on admission (25% were missing the mSOFA), O2 saturation on admission, receipt of glucocorticoids, receipt of IL-6 pathway agents, ICU admission, and mechanical ventilation. The investigators categorized patients into four treatment groups for analysis: (1) HCQ, (2) HCQ plus azithromycin, (3) azithromycin alone, and (4) neither HCQ nor azithromycin. They used multivariable Cox proportional hazards regression models to calculate hazard ratios adjusted using the abstracted data. The investigators also used the available data to match 1:1 HCQ and no HCQ patients for repeat Cox proportional hazards modeling. During the study period there were 2,948 COVID-19 admissions; 286 (10%) were excluded due to not being discharged, transfer, or leaving against medical advice, leaving 2541 patients for analysis. The median length of hospitalization was 6 days (4-10), and median follow-up time was 28 days. The majority of patients received HCQ (1202 (47%) HCQ alone and 783 (31%) HCQ plus azithromycin). A total of 409 (16%) patients received neither medication. The median time from admission to HCQ initiation was 1 day (interquartile range 1-2).

Summary of Main Findings

In-hospital mortality in this cohort was 18.1%. Among patients receiving HCQ alone, the mortality was 13.5%, for HCQ plus azithromycin it was 20%, for azithromycin alone it was 22%, and for neither drug it was 26%. There were substantial differences between the groups, with the non-HCQ group generally having a greater proportion of characteristics associated with higher mortality. These included older age (median 68 years compared to 63 years among HCQ only), a lower proportion of Black race (Black race was associated with lower mortality in this cohort), and a substantial imbalance in glucocorticoid use: 78% among patients receiving HCQ compared to 35% among those who received neither medication received a glucocorticoid. In addition, only 8% of those non-HCQ received mechanical ventilation compared with 13% among HCQ alone and 29% among HCQ plus azithromycin. Based on the Kaplan-Meier curves, although not explicitly described in the manuscript, it appears that approximately 30% of the deaths occurred in the first 3 days after admission for those not receiving HCQ while few deaths occurred during this time period among those who received HCQ.

Study Strengths

This study includes a large sample size, systematically collected data, and a heterogeneous population.

Limitations

This study had substantial limitations. For example, the inclusion of time prior to HCQ use for estimating survival benefit for HCQ rather than using time dependent variables would favor the HCQ arms because people receiving HCQ had to survive long enough to receive it (which was >2 days for 25% of the HCQ patients). Also, although the study team used multivariable regression methods and equate the groups on a set of observed characteristics, there is a strong potential for additional, unobserved confounders. Perhaps most importantly, unmeasured patient factors likely impacted the decision to treat with HCQ. Very few details are provided regarding how treatment decisions were made expect that a standardized protocol was used. The use of a standardized protocol suggests that patient who did not receive HCQ differed in important ways from those who did. In addition, those who received HCQ had to survive in order to receive HCQ. In particular, the high mortality, low number of ICU admissions, and high prevalence of multiple factors associated with higher mortality seen in the non-HCQ group all suggest that HCQ was preferentially prescribed to less sick patients. It is plausible that decisions were made for palliative care for some individuals, leading to neither admission to the ICU nor use of HCQ, and thus further contributing to an apparent beneficial effect of HCQ.

Value added

This study reinforces the association between age and mortality and uses a relatively large dataset of a diverse set of COVID-19 patients.