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


A total of 1,027 patients met criteria for evaluation, comprising 155 patients in the SARS-CoV-2 detected and 872 patients in the SARS-CoV-2 not detected groups Table 1. The median age of patients with COVID-19 was 65 years; 41% were female. The cohort of patients without COVID-19 was younger, with a median age of 56 years (P <.0001); 44% were female. Only patients with COVID-19 in the admitted group were significantly older than corresponding patients without COVID-19 (median age 70 vs 64 years, P = .031). More patients with COVID-19 identified as Hispanic or Latino (28% vs 8%; P <.0001; odds ratio [OR], 4.33; 95% confidence interval [CI], 2.85–6.60), but other demographics were not significantly different between patients with COVID-19 compared to those without. A total of 31 of 155 (20%) patients with COVID-19 and 44 of 872 (5%) patients without COVID-19 died within 15 days of their last SARS-CoV-2 test. The frequency of different ICD-10 diagnostic codes at the time of testing for SARS-CoV-2 were similar between patients with and without COVID-19. Cough, fever, and shortness of breath were within the top 20 diagnostic codes regardless of COVID-19 status concordant with the symptom-driven testing in place at the time (Supplemental Table 1; all supplemental materials can be found at American Journal of Clinical Pathology online).

Several significant differences in basic CBC data were noted between the COVID-19–positive and COVID-19–negative patient groups Table 2, including platelet and WBC counts Figure 2. A number of other parameters showed differences between groups that reached statistical significance, but for many of these (including hemoglobin, RBC count, and RDW) the differences were relatively small. Lymphocyte count was notably lower in patients with COVID-19 overall (median, 1.06 vs 1.43 × 109/L, P <.0001), predominantly attributable to lower median lymphocyte counts in patients with the lowest disease severity (ED group only). While lymphocyte counts were lower in those patients with COVID-19 with a more severe clinical course, this finding did not appear to be specific to COVID-19 (Table 2). Lymphocyte counts were not significantly different among patients admitted to standard inpatient care (admitted course) or critical care (critical course) based on their COVID-19 status (P = .2766 and P = .1954, respectively). Similarly, while 53% and 55% of patients with COVID-19 in the admitted and critical courses were lymphopenic in the ED, similar rates of lymphopenia were observed among admitted and critical COVID-19–negative patients (Supplemental Table 2). While there was a higher likelihood of identifying lymphopenia in patients with COVID-19 overall (45% vs 27%; P <.0001; OR, 2.26), this was largely attributable to differences in the least-severe patient group (ED only, 31% vs 19%; P = .0393; OR, 1.9).

Figure 2.

Overall, patients with COVID-19 had lower levels of leukocytes (P <.0001) (A) and lymphocytes (P <.0001) (B). Patients with COVID-19 had lower levels of neutrophils (P <.0001) (C) and platelets (P <.0001) (D) in comparison to patients without COVID-19. Hemoglobin (P = .0140) (E) was significantly higher and red cell distribution width (RDW) (P = .0354) (F) was significantly lower in patients with COVID-19, but absolute differences in these analytes were small. The neutrophil to lymphocyte ratio (P = .8012) (G) was not significantly different between patients with and without COVID-19 overall. The left panel for each analyte shows a comparison of all patients with and without COVID-19, while the right panel breaks these groups down by clinical course. Dashed lines denote the upper and lower limits of the reference range for each analyte. For analytes with more than 1 reference range (eg, sex or age), the lowest and highest bounds are represented. The center line is the median, the lower and upper hinges are the first and third quartiles (the 25th and 75th percentiles), the shaded region is the interquartile range (IQR), and the upper and lower whiskers extend from the hinge to the largest/smallest value or no further than 1.5 × IQR from the hinge. P ≤ .05 indicates significance. COVID-19, coronavirus disease 2019; ED, emergency department.

Interestingly, we found similar patterns with all leukocyte populations, including significantly lower neutrophil counts among patients with COVID-19 compared to the comparison group across all clinical courses. Median counts of eosinophils and basophils were also significantly lower overall in patients with COVID-19 across the clinical course. Median monocyte counts were lower overall for patients with COVID-19 (0.54 vs 0.64 × 109/L, P <.0001), but without much variation by clinical course (Table 2). Platelet counts were lower overall in patients with COVID-19 compared to those without COVID-19 (Figure 2) and rates of thrombocytopenia were higher (Supplemental Table 2). However, median platelet counts for patients with COVID-19 who followed a critical course were not significantly different from their non-COVID-19 comparison group (P = .11). Additionally, median platelet count did not vary with COVID-19 illness severity (Kruskal-Wallis P = .4266).

The median platelet to lymphocyte ratio (PLR), platelet to neutrophil ratio (PNR), and NLR were not significantly different overall between patients with COVID-19 and those without (P = .10, P = .39, and P = .80, respectively; Figure 2). However, each of these ratios demonstrated a trend of increasing (PLR and NLR) or decreasing (PNR) with illness severity, regardless of COVID-19 status. Considering just patients with COVID-19, the median values for PNR, PLR, and NLR were significantly different by clinical course (Kruskal-Wallis P <.0001 to P = .0049).

Testing for inflammatory markers such as CRP and IL-6 was not as widespread among the ED patients in our dataset (Supplemental Tables 3 and 4). Patients with COVID-19 had a CRP test more often; 61 (39%) COVID-19–positive patients vs 97 (11%) of COVID-19–negative patients. Overall, median maximum CRP levels were significantly higher in patients with COVID-19 (94 vs 28.4 mg/L, P <.0001) driven by substantial elevations in the critical COVID-19–positive group. The data are similarly limited for IL-6; 29 (19%) of patients with COVID-19 had an IL-6 level measured and only 3 patients (less than 1%) in the comparison group had an IL-6 level measured in our dataset. Maximum IL-6 levels increased dramatically with illness severity in COVID-19 (median for admitted COVID-19–positive patients 35 ng/L vs 215 ng/L in the critical COVID-19–positive group, Supplemental Figure 1).

A large majority of patients with and without COVID-19 had a test for serum albumin in our dataset (144 [93%] and 717 [82%], respectively). Minimum albumin levels were lower overall in patients with COVID-19 (33 g/L vs 39 g/L, P <.0001) and decreased with increasing illness severity. Patients with COVID-19 were more likely to have a minimum albumin value below the reference interval (55% vs 24%; P <.0001; OR, 3.79; Supplemental Table 4). The presentation albumin for patients with COVID-19 admitted to a hospital was compared to the minimum value when possible. In both settings, albumin on presentation was significantly higher than minimum values (P <.0001, absolute differences 4 and 5 g/L, for admitted COVID-19–positive and critical COVID-19–positive groups, respectively). Patients without COVID-19 in the corresponding clinical courses had similar but less-pronounced decreases.

PT and PTT were only performed in a subset of patients (Supplemental Table 3 and Supplemental Figure 2). Maximum PT/international normalized ratio [INR] values in patients with COVID-19 were not significantly different from the comparison group and tended to be lower (median 14.2 s/INR 1.1 vs 14.4 s/INR 1.1, P = .86 and P = .83, respectively). Maximum PTT values were significantly higher overall in patients with COVID-19 (median 36 s vs 32 s, P = .0031). Patients with COVID-19 were significantly more likely to have an elevated maximum PTT compared to COVID-19–negative patients (52% vs 33% of those tested; P = .0053; OR, 2.20; 95% CI, 1.31–3.72). However, anticoagulants that could prolong the PTT, such as heparin, were not recorded. The presentation PT and PTT for patients with COVID-19 admitted to a hospital was compared to the maximum value when possible. For both PT/INR and PTT, presentation values were significantly lower than maximum values in admitted patients (P <.0001 for both), although absolute differences were not large (0.15 s and 0.9 s. for PT, 1 s and 5 s for PTT, admitted COVID-19–positive and critical COVID-19–positive groups, respectively). Notably, this finding was also present in patients without COVID-19 in the critical course but not evident for those in the admitted course.

D-dimer testing was concentrated in the critical COVID-19–positive group (66% of all D-dimer tests in patients with COVID-19), while approximately 10% of patients without COVID-19 in each clinical course had a D-dimer value. Overall, maximum D-dimer values in patients with COVID-19 were not different from those without COVID-19 (920 μg/L and 680 μg/L fibrinogen equivalent units, respectively, P = .1972). However, an elevated D-dimer was present in 80% of all patients with COVID-19 tested (28 of 35 patients), which was significantly higher than the rate in patients without COVID-19 (57%, 52 of 91; P = .0225; OR, 3; 95% CI, 1.25–7.53).