Aim: COVID-19 is a known risk factor for pulmonary embolism (PE). In this retrospective, multicenter study, we aimed to determine an optimal D-dimer cutoff to predict PE in hospitalized patients with COVID-19. Materials & methods: A total of 193 patients underwent computerized tomographic pulmonary angiography imaging and were classified into PE positive and negative groups. Physiological, radiological and biochemical parameters were compared and receiver operator curve analysis was conducted to determine a predictive D-dimer threshold. Results: An optimal D-dimer cutoff of 2494 ng/ml was selected (Youden index: 0.906), giving a sensitivity of 100% (95% CI: 100-100) and specificity of 90.62% (95% CI: 90.5-90.8) for predicting PE. Conclusion: We propose that in the absence of other clinical signs, a D-dimer threshold of 2495 ng/ml could be used with high sensitivity and specificity to predict PE in hospitalized patients with COVID-19.
Keywords: COVID-19; CTPA; D-dimer; computerized tomographic pulmonary angiography; optimal threshold analysis; pulmonary embolism.