Magnitude of D-dimer matters for diagnosing pulmonary embolus

Objective The objective of this study is to determine whether the magnitude of the d-dimer correlates with a higher likelihood of pulmonary embolus (PE). Methods We performed an electronic chart review at our academic, tertiary care center, annual emergency department (ED) census greater than 100 000. All patients with a chest computed tomographic (CT) scan with intravenous contrast and an elevated d-dimer level obtained in the ED between January 2001 and July 2008 were identified. Specific, predetermined, predefined data elements including sex, age, d-dimer level, and final ED diagnosis were recorded by a hypothesis-blinded extractor using a preformatted data form. d-dimer level less than 0.58 μg/mL constitutes the normal laboratory reference range for our turbidometric d-dimer assay. Data were analyzed using standard statistical methods, and a linear regression analysis was performed for correlation analysis of d-dimer and diagnosis of PE. Results We identified 544 subjects who had both a chest CT scan performed and an elevated d-dimer level obtained in the ED. Fifty-eight subjects (10.7%; mean d-dimer, 4.9 μg/mL) were diagnosed with PE, and 486 (89.3%; mean d-dimer, 2.0) did not have a PE. The percentages of PE diagnoses for d-dimers in the ranges 0.58 to 1.0, 1.0 to 2.0, 2.0 to 5.0, 5.0 to 20.0, and greater than 20.0 (n = 11) were 3.6%, 8.0%, 16.2%, 35.3%, and 45.5%, respectively. The positive predictive value of PE for d-dimer level cutoffs of greater than 0.58, greater than 1.0, greater than 2.0, greater than 5.0, and greater than 20.0 was 10.7%, 14.6%, 22.2%, 37.8%, and 45.5%, respectively. Increasing d-dimer values were strongly correlated with the presence of PE (odds ratio, 1.1685 per stratum; P <.001). Conclusion Increasing magnitude of d-dimer correlates with increasing likelihood of PE diagnosed by CT angiography.