Comparison of Blood Counts and Markers of Inflammation and Coagulation in Patients With and Without COVID-19 Presenting to the Emergency Department in Seattle, WA

Christopher M. Chandler, MD; Molly C. Reid, MPH; Sindhu Cherian, MD; Daniel E. Sabath, MD, PhD; Kerstin L. Edlefsen, MD


Am J Clin Pathol. 2021;156(2):185-197. 

In This Article

Materials and Methods

This study was approved by the University of Washington Human Subjects Division Institutional Review Board (STUDY00009972) and was limited to adult patients (≥18 years old) who received a SARS-CoV-2 RT-PCR test at one of the University of Washington hospitals in Seattle, before April 13, 2020, the date on which routine screening for COVID-19 was enacted at our institution. Patients were stratified by SARS-CoV-2 status, mortality, and sending location of laboratory results Figure 1. Patients with results originating from an ED location only were assigned to the ED clinical course. Patients with results from an ED location and an inpatient location (but not ICU) were assigned to the admitted clinical course. Finally, patients with results from an ED location and an ICU location at any point during their admission were assigned to the critical clinical course. Results from patients in locations with ICU, intensive, or critical in the location name were classified as ICU. Results from patients in a special care unit at one of the local hospitals were also classified as originating from an ICU. To better capture early laboratory alterations associated with illness severity that could be used to stratify those at increased risk of poor outcome, patients who died within 15 days of their last SARS-CoV-2 test were put in their respective critical group; this added 20 patients to the critical COVID-19–negative group and 18 patients to the critical COVID-19–positive group, respectively. Two patients with COVID-19 and 35 patients without COVID-19 died outside of the 15-day window from their last COVID-19 test and were not included in the analysis.

Figure 1.

Flow diagram showing patient selection process. Patients who had at least 1 positive SARS-CoV-2 RT-PCR test at any point were assigned to the detected group. Patients who never had a positive or inconclusive SARS-CoV-2 test were assigned to the not detected group. Those patients who had only an inconclusive SARS-CoV-2 test were excluded from analysis. Patients were stratified for disease severity according to sending location of laboratory results. Patients with only ED locations were classified as ED COVID-19–negative or ED COVID-19–positive, depending on whether they fell into the SARS-CoV-2 detected or not detected group. In a similar manner, patients with laboratory testing submitted from only an inpatient ward (not ICU locations) were classified as admitted COVID-19–negative or admitted COVID-19–positive. Lastly, patients with laboratory results submitted from an ICU at some point during their clinical course or died within 15 days of their last SARS-CoV-2 test were classified as critical COVID-19–negative or critical COVID-19–positive. aPatients with SARS-CoV-2 detected (n = 2) or not detected (n = 35) were excluded. COVID-19, coronavirus disease 2019; ED, emergency department; ICU, intensive care unit; RT-PCR, reverse transcription polymerase chain reaction; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

Patients with laboratory values originating from locations or wards of the University of Washington's affiliated cancer center were excluded due to the high prevalence of patients with leukemia and lymphoma. Similarly, patients from locations including the word oncology were excluded. Select laboratory results were extracted from the Department of Laboratory Medicine and Pathology's data warehouse for 1 week following each SARS-CoV-2 PCR test. Demographics, International Classification of Diseases-Tenth Revision (ICD-10) diagnostic codes, and mortality data were pulled from the health care system's shared electronic data warehouse. No values outside of an assay's reportable range were included and all values were exact (no greater or less than results). Results were organized and analyzed using RStudio version 1.2.1335.

For most laboratory tests, such as CBC, we defined the presentation value as the first test sent from an ED location for each patient (or as the first value identified within 1 day of the patient's first set of tests). For some markers, we also evaluated the highest documented (maximum) and lowest documented (minimum) values, in addition to the presentation levels. Continuous variables were tested for normality using the D'Agostino-Pearson omnibus normality test. Due to the frequency of nonparametric distributions in the selected laboratory values, the Mann-Whitney U test and Kruskal-Wallis test were used for continuous variables. Fisher exact test was used in comparison of categorical variables. For pairwise comparisons, the Wilcoxon signed rank test was used. We set the significance level (α) at P = .05. Statistical analysis was performed using GraphPad Prism8.