Evaluation for Pulmonary Embolism in Patients with COVID-19
By Joshua Easter, MD, MSc
University of Virginia Health System emergency physician and UVA School of Medicine associate professor
Case: A 40-year-old female with no significant past medical history presents to your ED with chest pain. Ten days prior she developed cough, fever, and shortness of breath. She saw her PCP and had a positive covid-19 PCR test. Her fever and shortness of breath improved over the subsequent few days. Then yesterday she developed renewed shortness of breath and pleuritic chest pain. On exam she is afebrile with a heart rate of 110, blood pressure of 110/76, respiratory rate of 20, and oxygen saturation of 94% on room air. Her lungs are clear to auscultation with no respiratory distress, and she has strong peripheral pulses without lower extremity edema. A chest radiograph is normal and electrocardiogram shows only sinus tachycardia. Your differential diagnosis includes pulmonary embolism. What is your next step to evaluate for a pulmonary embolism in your patient with covid-19?
Pulmonary embolism is a known complication of covid-19 infection, occurring in 3-9% of hospitalized patients (Sakr, 2020). For non-covid patients, clinicians commonly employ d-dimer assays as a screening test to rule out pulmonary embolism; d-dimer has high negative predictive value, albeit with low positive predictive value. It is unclear if these test characteristics are maintained in patients with covid-19 infection, who appear to have a different pre-test probabilities for pulmonary embolism than patients without covid-19 (Roncon, 2020; Middeldorp, 2020; Suh, 2021).
Elberts et al. conducted an observational retrospective review of all patients undergoing computed tomography pulmonary angiography (CTPA) and d-dimer testing at five EDs in Philadelphia from 2019-20 (Elberts, 2021). Patients with covid-19 routinely underwent d-dimer testing upon admission to the hospital to assess disease severity. The investigators excluded any patients that did not undergo both d-dimer testing and CTPA. Initially only symptomatic patients underwent covid-19 testing. Later in the study due to changes in institutional policy, all patients underwent covid-19 testing. Reviewers blinded to clinical data determined the presence of pulmonary emboli on CTPA; subsegmental pulmonary emboli were included in the outcome measure.
Table. Test characteristics of d-dimer for pulmonary embolism.
During the study period, 1,158 patients underwent d-dimer testing and CTPA, and over one quarter (28%) of these patients had covid-19. Pulmonary emboli were present in 110(10%) patients, including 12% of covid-19 patients and 9% of patients without covid-19. D-dimer demonstrated excellent sensitivity in both patients with and without covid-19 (Table). It missed subsegmental pulmonary emboli in two patients without covid-19. It had poor specificity in both groups. Increasing the threshold value of what constituted a positive d-dimer improved specificity without significant reductions in sensitivity.
The largest limitation to these results comes from the observational nature of the study. Only patients undergoing CTPA were included, and this could impact the results in multiple ways. This selection bias could result in inclusion of patients with a greater risk of thromboembolism (as low risk patients would have negative d-dimers and not undergo CTPA). This increased prevalence of thromboembolism would be expected to reduce the negative predictive value of d-dimer. It is also possible that patients with covid-19 and pulmonary emboli had negative d-dimers (false negatives) and therefore did not undergo CTPA. These patients would have been excluded from the study, which might have made the negative predictive value appear higher than the true value. It also important to consider that the lower bound of the 95% confidence interval for sensitivity was only 88% in patients with covid-19. A larger study is required to determine the true sensitivity of d-dimer for pulmonary embolism in covid-19 patients.
Despite the limitations inherent with a retrospective review conducted during a time when the standard of care was evolving for covid-19, this study suggests d-dimer remains highly sensitive for ruling out pulmonary embolism. The negative predictive value of d-dimer is similar in patients with and without covid-19. Also similar to non-covid patients, the specificity of d-dimer is quite low; covid-19 infection itself is associated with higher d-dimer levels. Nevertheless, d-dimer appears to be a helpful screening tool for pulmonary embolism in patients with covid-19.
Outcome: You are concerned about your patient having a pulmonary embolism. Your patient is low to moderate risk for venothromboembolism by clinical gestalt and Wells’ criteria. Therefore, you decide to obtain a d-dimer. It is negative, and you do not pursue further testing for pulmonary embolism. You discharge the patient from the ED, and her symptoms resolve over the subsequent few days.
About the VACEP EM Lit Review: Every month, VACEP members will share their readings of the latest medical literature. Submit yours to us by emailing Executive Director Sarah Marshall.
