key: cord-0840879-qtfhoaqx
authors: Gottlieb, Michael; Sansom, Sarah; Frankenberger, Casey; Ward, Edward; Hota, Bala
title: Clinical Course and Factors Associated with Hospitalization and Critical Illness Among COVID‐19 Patients in Chicago, Illinois
date: 2020-08-06
journal: Acad Emerg Med
DOI: 10.1111/acem.14104
sha: 5adedbd2f38ff79ab602dd9b11df357130c2a032
doc_id: 840879
cord_uid: qtfhoaqx

BACKGROUND: SARS‐CoV‐2 is a global pandemic associated with significant morbidity and mortality. However, information from United States cohorts is limited. Understanding predictors of admission and critical illness in these patients is essential to guide prevention and risk stratification strategies. METHODS: This was a retrospective, registry‐based cohort study including all patients presenting to Rush University Medical Center in Chicago, Illinois with COVID‐19 from March 4(th), 2020 to June 21(st), 2020. Demographic, clinical, laboratory, and treatment data were obtained from the registry and compared between hospitalized and non‐hospitalized patients, as well as those with critical illness. We used logistic regression modeling to explore risk factors associated with hospitalization and critical illness. RESULTS: 8,673 COVID‐19 patients were included in the study, of whom 1,483 (17.1%) were admitted to the hospital and 528 (6.1%) were admitted to the intensive care unit. Risk factors for hospital admission included advanced age, male sex (OR 1.69; 95% CI 1.44‐1.98), Hispanic/Latino ethnicity (OR 1.52; 95% CI 1.18‐1.92), hypertension (OR 1.77; 95% CI 1.46‐2.16), diabetes mellitus (OR 1.84; 95% CI 1.53‐2.22), prior CVA (OR 3.20; 95% CI 1.99‐5.14), coronary artery disease (OR 1.45; 95% CI 1.03‐2.06), heart failure (OR 1.79; 95% CI 1.23‐2.61), chronic kidney disease (OR 2.60, 95% CI 1.77‐3.83), end‐stage renal disease (OR 2.22; 95% CI 1.12‐4.41), cirrhosis (OR 2.03; 95% CI 1.42‐2.91), fever (OR 1.43; 95% CI 1.19‐1.71), and dyspnea (OR 4.53; 95% CI 3.75‐5.47). Factors associated with critical illness included male sex (OR 1.45; 95% CI 1.12‐1.88), congestive heart failure (OR 1.45; 95% CI 1.00‐2.12), obstructive sleep apnea (OR 1.58; 95% CI 1.07‐2.33), bloodborne cancer (OR 3.53; 95% CI 1.26‐9.86), leukocytosis (OR 1.53; 95% CI 1.15‐2.17), elevated neutrophil‐to‐lymphocyte ratio (OR 1.61; 95% CI 1.20‐2.17), hypoalbuminemia (OR 1.80; 95% CI 1.39‐2.32), elevated AST (OR 1.66; 95% CI 1.20‐2.29), elevated lactate (OR 1.95; 95% CI 1.40‐2.73), elevated D‐Dimer (OR 1.44; 95% CI 1.05‐1.97), and elevated troponin (OR 3.65; 95% CI 2.03‐6.57). CONCLUSION: There are a number of factors associated with hospitalization and critical illness. Clinicians should consider these factors when evaluating patients with COVID‐19.

In December 2019, a cluster of cases with pneumonia of unknown etiology was identified in Wuhan City, Hubei province, China. In January 2020, a novel coronavirus, now designated as severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), was identified as the cause of coronavirus disease 2019 . 1 Despite containment efforts, SARS-CoV-2 spread rapidly to other countries. The first case in the United States was identified in mid-January of 2020 and it has now been reported in all 50 states. 2, 3 The disease was declared a global pandemic on March 11, 2020 and spread of SARS-CoV-2 has been confirmed in over 200 countries resulting in over 12.1 million cases and 550,000 deaths. 1 Large cities in the United States, including Chicago, Illinois, have been profoundly affected by this pandemic, with over 50,000 COVID-19 confirmed cases and 2,600 deaths among Chicago residents. 4 However, information from large cohorts within the United States remains limited and it is unclear if data from other countries are generalizable to local community dynamics and outcomes.

Observational studies have identified several medical comorbidities potentially associated with adverse clinical outcomes, including older age, [5] [6] [7] [8] [9] [10] [11] male sex, 8, 12, 13 cardiovascular disease, 6, 10, [13] [14] [15] diabetes mellitus, 6, 7, 13, 15 and chronic respiratory disease. 10, [13] [14] [15] The importance of these comorbidities has not yet been fully elucidated due to small observational cohort sizes, inadequate adjustment for confounding factors, and likely under-reporting. 16 COVID-19 is associated with significant morbidity and mortality. However, only a subset of patients become critically ill. Therefore, it is important to identify the clinical and demographic features that predict admission and critical illness among this population using large data sets to enhance our understanding of this rapidly expanding disease. Understanding the morbidity and mortality in patients with COVID-19 is essential to guide prevention and risk stratification strategies for this population.

The primary goal of this study was to present the clinical and demographic features of patients who presented to a major academic institution in Chicago, Illinois with laboratory-confirmed COVID-19 infection as of June 21 st , 2020. As a secondary outcome, we sought to identify risk factors associated with hospitalization and critical illness.

This article is protected by copyright. All rights reserved

We conducted a retrospective case-control study to evaluate the risk factors for severe COVID-19 infection. This study was conducted at Rush University Medical Center, a 664-bed urban tertiary care hospital in Chicago, Illinois with an annual Emergency Department volume of 70,000 patients per year. We collected our first samples for testing on January 23, 2020 and the first patient to test positive with SARS-CoV-2 was seen at Rush University Medical Center on March 4, 2020. During this time period, we tested patients with symptoms concerning for COVID-19. From March 26th to May 21st, we also screened all admissions for COVID-19 regardless of symptoms. All patients who were tested for SARS-CoV-2 were designated as a person under investigation (PUI) in the electronic health record (EHR). Patients screened as PUIs were captured in the Epic EHR (Epic Systems, Verona, WI) as a part of normal clinical workflows. These indicators subsequently were filtered to an enterprise data warehouse (EDW).

All patients with diagnosed COVID-19, regardless of age, were included through June 21 st , 2020. We excluded patients who were transferred from other inpatient hospitals (n = 71), as these reflected a different patient cohort. These patients presented to the Rush University Medical Center intensive care unit (ICU) several days into their illness course and were often already intubated, thereby limiting the ability to use this population to predict admission or critical illness.

For patients considered as PUIs, a data mart was developed using detailed demographics, diagnoses, symptoms, comorbidities, treatments, laboratory results, and outcomes data. These data were filtered from the EDW comprising clinical and administrative data derived from the clinical EHR and associated information technology systems, including registration and intake surveys, laboratory, radiology, and billing information systems. This single source EDW supports all clinical, research, and operational needs and contains administrative data for more than 10 years and clinical data back to 2007. All data were de-identified before sharing. The local institutional review board evaluated this project and approved it with a waiver of informed consent.

This article is protected by copyright. All rights reserved

All patients were confirmed to have SARS-CoV-2 infection using a molecular amplification assay and nasopharyngeal, mid-turbinate, or nasal swab samples that were collected by trained clinical staff members. Testing was done at Illinois Department of Public Health Laboratory using the CDC-developed assay, or in the clinical microbiology laboratory at Rush University Medical Center using a laboratory-modified version of the CDC assay, the Real-Time SARS-CoV-2 assay on the m2000 (Abbott Laboratories, Abbott Park, IL), or the ID NOW™ COVID-19 (Abbott Laboratories, Abbott Park, IL). The assay used depended on the date and location of testing.

Our primary outcome was to provide a descriptive summary of the clinical and demographics features of patients who presented to our institution with COVID-19 over a greater than threemonth period. As a secondary outcome, we sought to identify risk factors associated with inpatient hospitalization and critical illness. Inpatient hospitalization was defined as any patient requiring admission to the hospital. For patients with more than one hospitalization (n = 376), only the most recent hospitalization was utilized. Critical illness was defined as a patient requiring ICU admission.

We obtained the following information from the EDW for analysis: age; sex; race; ethnicity; tobacco use (current and former); marital status; recent international travel; asthma; chronic obstructive pulmonary disease (COPD); hypertension; hyperlipidemia; diabetes mellitus; prior ischemic or hemorrhagic cerebrovascular accident (CVA); coronary artery disease (CAD); congestive heart failure (CHF); chronic kidney disease (not requiring dialysis); end-stage renal disease (ESRD) requiring dialysis; cirrhosis; obstructive sleep apnea (OSA); bloodborne cancer (e.g., leukemia, lymphoma); solid organ cancer; human immunodeficiency virus (HIV) infection; solid organ transplantation; body mass index (BMI); vital signs on arrival (e.g., temperature, heart rate, oxygenation saturation, respiratory rate, systolic and diastolic blood pressure); symptoms (i.e., anosmia, abdominal pain, bruising/bleeding, conjunctivitis, cough, diarrhea, dyspnea, fever, headache, joint pain, myalgias, rash, sore throat, vomiting, weakness); and

This article is protected by copyright. All rights reserved laboratory testing (complete blood count, absolute neutrophil count (ANC), absolute lymphocyte count (ALC), creatinine, total bilirubin, albumin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), lactate dehydrogenase (LDH), lactic acid, D-dimer, C reactive protein, ferritin, creatinine kinase, troponin, brain-type natriuretic peptide (BNP), pregnancy test). We also collected hospitalization outcome data on co-infections (e.g., positive influenza test, positive respiratory syncytial virus (RSV) test), complications (e.g., cardiac dysrhythmias, positive blood culture, positive sputum culture, subsequent bacterial pneumonia, acute respiratory distress syndrome (ARDS), rhabdomyolysis), and interventions performed (e.g., high flow nasal cannula, continuous and bi-level positive airway pressure ventilation, prone positioning, inhaled pulmonary vasodilators, extracorporeal membrane oxygenation, vasopressors, steroids, lopinavir/ritonavir, hydroxychloroquine, remdesivir, tocilizumab). All predictor variables were obtained at the time of their initial presentation. We also collected the following outcome data: hospital length of stay, intensive care unit length of stay, number of patients who were intubated, and days of mechanical ventilation. Length of stay and days on mechanical ventilation were determined only for patients who were discharged or died.

Categorical data were presented as percentage frequencies, with continuous data presented as median and interquartile range (IQR). All data obtained from the EDW were evaluated for anomalies by a clinical data analyst. Categorical variables were analyzed using χ2 or Fisher exact test, as appropriate. Continuous variables were analyzed by Student's t-test or Mann-Whitney U test, as appropriate. Variables found to be significant at the p ≤ 0.10 level were included in a multivariable logistic regression model. Two-tailed tests were used. Two logistic regression models were created. The first used inpatient hospitalization as the primary outcome and various co-factors as independent variables. The second used critical illness (defined as requiring ICU admission) as the primary outcome and various co-factors as independent variables. A plot of the standard deviance results demonstrated no extreme outliers in either model. A plot of leverage found two observations with high leverage in the admission model and no extreme influential points in the critical illness model. Removal of both observations in the admission model did not significantly influence the findings. There was no evidence of multicollinearity with a variance inflation factor < 4. There were no continuous covariates in the model. Goodness of fit was

This article is protected by copyright. All rights reserved assessed for both models using a c-statistic and Hosmer-Lemeshow test. For the admission model, the c-statistic was 0.92 and the Hosmer-Lemeshow test had a p = 0.111. For the critical care model, the c-statistic was 0.75 and the Hosmer-Lemeshow test had a p = 0.204. Findings from the multivariable logistic analyses were presented as odds ratios (ORs) with 95% confidence intervals (CIs). All analyses were performed using SAS version 9.4 (Cary, NC).

During the study period, our health system tested 41,153 patients for COVID-19, of whom 8,673 The demographics, clinical symptoms, and laboratory findings are shown in Table 1 . Among all patients, the most common age group affected was 19-44 years of age (52.2%), followed by 45- Events and interventions during hospitalization are demonstrated in Table 2 . Co-infection with influenza (3.6%) or respiratory syncytial virus (≤1%) was uncommon in our population. The most common complications were acute respiratory distress syndrome (17.9%), cardiac dysrhythmias (9.6%). and rhabdomyolysis (8.7%). Three-quarters of patients received high-flow

This article is protected by copyright. All rights reserved nasal cannula oxygen therapy and 23.3% received hydroxychloroquine. Steroids were given in 45.6% of cases.

Multivariate logistic regression identified several factors associated with an increased risk of hospitalization ( 

This analysis of 8,673 cases at a major academic health system in Chicago is one of the largest collective series of COVID-19 cases to date and one of very few large data sets performed in the United States. In this report, we summarized the most common features of the disease in our cohort and predictors of both hospitalization and critical illness.

While much of the literature has arisen out of areas that had a rapid escalation of cases at the start of the pandemic (e.g., China, Italy, New York, Seattle), there is limited literature evaluating

This article is protected by copyright. All rights reserved secondary sites in the United States which have had more time to prepare for COVID-19. When compared with the largest data set from New York, our cohort had a significantly lower rate of mortality and critical illness. 17 There are many factors which may have contributed to this lower mortality rate, which may include the additional time available to prepare the hospital system, lower overall exposure rates due to social isolation efforts, and differences in resource availability. Of note, our population was also younger, more likely to be female, and had fewer comorbidities than this cohort, which may also have been protective. 17 We may also have tested more liberally than other sites due to increased availability of testing later in the pandemic.

Consistent with prior literature, cough was the most common symptom, present in 72% of cases. 6, 18 However, contrary to the study by Guan et al, we found significantly higher rates of dyspnea in our cohort. 18 Interestingly, while 48% of patients reported a history of subjective fevers, only 12% of patients had a fever on presentation. This is much lower than existing literature, which have found the rates of fever ranging from 31% to 94%. 6, 17 Similar to prior studies, we found that an elevated ferritin level and lymphopenia were common. 6,17,18 However, we identified significantly higher C reactive protein levels and D-dimer levels than prior literature. 17, 18 The rate of concomitant influenza infection in our study was low, though higher than the New York data. 17 We also found that 13% of our population developed bacteremia during their hospitalization, which is higher than expected for hospitalized patients but similar to data among ICU patients without COVID-19. 19 We found that the most common factors associated with hospitalization were increased age, male sex, Hispanic/Latino ethnicity, hypertension, diabetes mellitus, prior CVA, CAD, CHF, chronic kidney disease, and ESRD, which is consistent with prior data demonstrating increased risk of worse outcomes in these patients. 6, 8, 9, 15, 17, 18 Interestingly, we also identified an increase rate of hospitalization among patients with cirrhosis, which has not been previously described. The reason for this is unclear, but it is possible that this may reflect worsened baseline health similar to the aforementioned factors.

Several factors were associated with an increased risk of critical illness. These included male sex, CHF, OSA, bloodborne cancer, leukocytosis, elevated ANC/ALC, hypoalbuminemia,

This article is protected by copyright. All rights reserved elevated AST, elevated lactate, elevated D-Dimer, and elevated troponin. Among these, one the most predictive was an elevated troponin. There has been increasing recognition of the association between COVID-19 and cardiovascular complications. 20 Therefore, it is not surprising that an elevated troponin is associated with an increased risk of critical illness.

Leukocytosis was another predictor, which was also demonstrated in the study by Guan et al. 18 Early literature has also proposed that an elevated ANC/ALC may be predictive of critical illness in this population. 21 Hypoalbuminemia was another risk factor in our population. While there has been limited analysis of this in prior studies, 6, 22, 23 it is possible that this may signify a patient population with less physiologic reserve who may be more likely to deteriorate.

It is important to consider several limitations with regard to the present study. First, this was studied at a single hospital system in Chicago, Illinois and may not reflect other locations.

However, this is the first large study to be performed in this location and we believe it lends unique insights into the disease. Additionally, while a dedicated data registry was used, some data were not reported or available. Where possible, we have accounted for this in the logistic regression, though we acknowledge that it would be preferable to have all data available for each outcome. Due to the retrospective nature of this database, we were unable to control for unmeasured confounders. Moreover, many discharged patients did not receive laboratory testing, so we are unable to comment on the role of laboratory testing on admission decisions. We are also not able to determine the degree to which COVID-19 influenced admission decisions.

Finally, testing was primarily limited to symptomatic patients, so it is likely that we may have missed some asymptomatic patients in this cohort.

This study described the clinical and laboratory features associated with COVID-19 among patients in Chicago, Illinois. Advanced age, male sex, Hispanic/Latino ethnicity, hypertension, diabetes mellitus, prior CVA, CAD, CHF, chronic kidney disease, ESRD, cirrhosis, and symptoms of fever or dyspnea were more commonly associated with admission, while male sex, CHF, OSA, bloodborne cancer, leukocytosis, an elevated ANC/ALC, hypoalbuminemia, an elevated AST, an elevated lactate, and elevated D-Dimer and an elevated troponin were

This article is protected by copyright. All rights reserved associated with critical illness. We believe these findings will be valuable for understanding the epidemiology of COVID-19 in secondary sites within the United States and for assessing risk factors associated with hospitalization and critical illness. Future studies should validate these findings at other locations in the United States and utilize them in conjunction with other cohorts to develop decision tools to risk stratify patients for admission and critical illness. 

Washington State 2019-nCoV Case Investigation Team. First Case of 2019 Novel Coronavirus in the United States

COVID-19) Cases in U.S. 2020

First known person-to-person transmission of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in the USA

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Risk Factors Associated With Acute Respiratory Distress Syndrome and Death in Patients With Coronavirus Disease

Accepted Article This article is protected by copyright. All rights reserved

Baseline Characteristics and Outcomes of 1591

Patients Infected With SARS-CoV-2 Admitted to ICUs of the Lombardy Region

Severe Outcomes Among Patients with Coronavirus Disease 2019 (COVID-19) -United States

Characteristics of and Important Lessons From the Coronavirus Disease 2019 (COVID-19) Outbreak in China: Summary of a Report of 72314 Cases From the Chinese Center for Disease Control and Prevention

Prediction for Progression Risk in Patients with COVID-19 Pneumonia: the CALL Score

Clinical characteristics of 113 deceased patients with coronavirus disease 2019: retrospective study

Prevalence of comorbidities in the novel Wuhan coronavirus (COVID-19) infection: a systematic review and meta-analysis

Preliminary Estimates of the Prevalence of Selected Underlying Health Conditions Among Patients with Coronavirus Disease 2019 -United States

Covid-19: risk factors for severe disease and death

Presenting Characteristics, Comorbidities, and Outcomes Among 5700 Patients Hospitalized With COVID-19 in the

Clinical Characteristics of Coronavirus Disease 2019 in China

EPIC III Investigators. Prevalence and Outcomes of Infection Among Patients in Intensive Care Units in 2017

Accepted Article This article is protected by copyright. All rights reserved 20

The diagnostic and predictive role of NLR, d-NLR and PLR in COVID-19 patients

Clinical characteristics of refractory COVID-19 pneumonia in Wuhan, China

A Tool to Early Predict Severe Corona Virus Disease 2019 (COVID-19): A Multicenter Study using the Risk Nomogram in Wuhan and Guangdong, China

CI, confidence interval; CVA, cerebrovascular accident; ESRD, end-stage renal disease

HIV, human immunodeficiency virus

CI, confidence interval; CVA, cerebrovascular accident; ESRD, end-stage renal disease

HIV, human immunodeficiency virus

BMI, body mass index Accepted Article

This article is protected by copyright. All rights reserved

This article is protected by copyright. All rights reserved