Bone Marrow and Peripheral Blood Findings in Patients Infected by SARS-CoV-2

Cynthia K. Harris, MD; Yin P. Hung, MD, PhD; G. Petur Nielsen, MD; James R. Stone, MD, PhD; Judith A. Ferry, MD


Am J Clin Pathol. 2021;155(5):627-637. 

In This Article

Materials and Methods

This study was approved by our hospital's institutional review board. This is a single-center retrospective study of 22 patients who tested positive for COVID-19 at our institution from March to July 2020 who had bone marrow available for histopathologic examination. The bone marrows from 2 surgical specimens and the first 20 consecutive autopsies of patients positive for COVID-19 from March to April 2020 were evaluated. Nasopharyngeal swabs were taken premortem during hospitalization or postmortem in the case of deaths that occurred outside our institution and were analyzed using real-time reverse transcription polymerase chain reaction. Normal values used for hematologic parameters were as established at our institution: WBC count, 4.5 to 11 × 109/L; hemoglobin, 13.5 to 17.5 g/dL; hematocrit, 41 to 53 L/L; platelets, 150 to 400 × 109/L; absolute neutrophil count (ANC), 1.8 to 7.7 × 109/L; and absolute lymphocyte count (ALC), 1 to 4.8 × 109/L. Postmortem examination was performed on 20 patients; the bone marrow was obtained from rib squeeze.[10,11]

For purposes of comparison, we collected the bone marrow at the time of autopsy from 10 patients who tested negative for COVID-19; the bone marrow was obtained from rib squeeze. The autopsy cases were performed by the same prosecutor during a 2-week period. All 10 patients tested negative for COVID-19 premortem at least once within the 2 weeks before death, and 7 of the 10 patients tested negative for COVID-19 twice within the 2 weeks before death. The patients were selected solely based on their negative COVID-19 status with no additional clinical information provided to the reviewer at the time of histologic review.

Additional bone marrow samples from 2 living patients who tested positive for COVID-19, obtained during clinical care, were also examined, one from a below-the-knee amputation and one from the sternum during treatment of a pericardial effusion. The formalin-fixed paraffin-embedded tissues were sectioned in 5 μm onto glass slides and stained using H&E. The bone marrow was evaluated for overall cellularity; myeloid-to-erythroid ratio; overall appearance of myeloid cells, erythroid cells, and megakaryocytes; presence and extent of hemophagocytosis; and any additional findings such as the presence of lymphoid aggregates. Immunohistochemical staining for CD68 was performed to highlight the macrophages. Selected cases were evaluated with antibodies to E-cadherin, CD71, glycophorin, myeloperoxidase, CD61, CD20, CD3, CD5, LEF1, and cyclin D1 to highlight hematologic elements Table 1. The patients' electronic medical records were retrospectively reviewed, and demographic information, laboratory data, clinical presentation, hospital course, and comorbidities were documented. We assessed the patients' risk for secondary HLH by calculating their HScore, which provides numerical point values for each of the following 9 variables: (1) known immunosuppression, (2) high temperature, and (3) organomegaly; laboratory values for (4) triglycerides, (5) ferritin, (6) serum glutamic oxaloacetic transaminase, and (7) fibrinogen; (8) cytopenia(s); and (9) the histologic presence of hemophagocytosis in the bone marrow.[12] We used an HScore of greater than or equal to 169 as our cutoff value, which corresponds to sensitivity of 93% and specificity of 86% for the diagnosis of HLH.[12] Statistical analysis including 2-tailed paired t tests and the Fisher exact test was performed using GraphPad Prism version 5.02 (GraphPad Software), with significance defined as P < .05.