A Prospective Study Evaluating Cumulative Incidence and a Specific Prediction Rule in Pulmonary Embolism in COVID-19

RATIONALE: Abnormal values of hypercoagulability biomarkers, such as D-dimer, have been described in Coronavirus Disease 2019 (COVID-19), which has also been associated with disease severity and in-hospital mortality. COVID-19 patients with pneumonia are at greater risk of pulmonary embolism (PE). However, the real incidence of PE is not yet clear, since studies have been limited in size, mostly retrospective, and PE diagnostic procedures were only performed when PE was clinically suspected. OBJECTIVES: (1) To determine the incidence, clinical, radiological, and biological characteristics, and clinical outcomes of PE among patients hospitalized for COVID-19 pneumonia with D-dimer > 1,000 ng/mL. (2) To develop a prognostic model to predict PE in these patients. METHODS: Single-center prospective cohort study. Consecutive confirmed cases of COVID-19 pneumonia with D-dimer > 1,000 ng/mL underwent computed tomography pulmonary angiography (CTPA). Demographic and laboratory data, comorbidities, CTPA scores, treatments administered, and clinical outcomes were analyzed and compared between patients with and without PE. A risk score was constructed from all these variables. RESULTS: Between 6 April 2020 and 2 February 2021, 179 consecutive patients were included. The overall incidence of PE was 39.7% (71 patients) (CI 95%, 32–47%). In patients with PE, emboli were located mainly in segmental/subsegmental arteries (67%). Patients with PE did not differ from the non-PE group in sex, age, or risk factors for thromboembolic disease. Higher urea, D-Dimer, D-dimer-to-ferritin and D-dimer-to-lactate dehydrogenase (LDH) ratios, platelet distribution width (PDW), and neutrophil-to-lymphocyte ratio (NLR) values were found in patients with PE when compared to patients with non-PE. Besides, lymphocyte counts turned out to be lower in patients with PE. A score for PE prediction was constructed with excellent overall performance [area under the ROC curve-receiver operating characteristic (AUC-ROC) 0.81 (95% CI: 0.73–0.89)]. The PATCOM score stands for Pulmonary Artery Thrombosis in COVID-19 Mallorca and includes platelet count, PDW, urea concentration, and D-dimer-to-ferritin ratio. CONCLUSION: COVID-19 patients with pneumonia and D-dimer values > 1,000 ng/mL were presented with a very high incidence of PE, regardless of clinical suspicion. Significant differences in urea, D-dimer, PDW, NLR, and lymphocyte count were found between patients with PE and non-PE. The PATCOM score is presented in this study as a promising PE prediction rule, although validation in further studies is required.