D-Dimer Assessment to Predict Pulmonary Embolism in ICU Patients with COVID-19 Pneumonia

The value of D-dimer assessments in ICU patients with COVID-19 for the prediction of pulmonary embolism (PE) is unclear. The present study had two purposes: 1. To assess the specificity of elevated absolute D-dimer values for PE on admission to the ICU. 2. To assess the specificity of a D-dimer increment for the development of PE during an ICU stay. D-dimer values were paired with the results of a CT pulmonary angiogram (CTPA) and compared in patients with and without PE on admission. In patients without PE on initial imaging and available repeat CTPA during an ICU stay, D-dimer increments between initial and repeat imaging of patients developing PE during an ICU stay were compared with those with persistently no PE. On admission, D-dimers in patients with PE were higher than those in patients without PE (median 850 vs. 6060 mu g/L; p < 0.0001). Using a cut-off of 9000 mu g/L, the specificity for predicting PE was 100% (CI 95.3-100%). Delta D-dimer during an ICU stay was greater in patients with PE (median 7983 vs. 3815 mu g/L; p < 0.005). Using a cut-off of 8000 mu g/L, specificity was 100% (CI 79.4-100%). Strongly elevated D-dimer values on admission and marked increases in D-dimer during ICU stays have a high specificity for predicting pulmonary embolism in critically ill COVID-19 patients.