Outcomes of high pretest probability patients undergoing d-dimer testing for pulmonary embolism: a pilot study

Abstract

ELISA (enzyme-linked immunosorbent assay) D-dimer testing is commonly used in the evaluation of possible pulmonary embolism (PE) in the emergency department, but is not recommended in high pretest probability patients. Whether a negative ELISA D-dimer can safely rule out PE in these patients is not known, as there have been no large studies comparing ELISA D-dimer results and outcomes in high pretest probability patients. This was a prospective observational pilot study of emergency department patients evaluated for PE. Patients evaluated for PE had pretest probability assessed by the Wells PE Score. High pretest probability was defined as: dichotomized Wells Score>4 points and patients with trichotomized Wells Score>6 points. Patients had an ELISA D-dimer ordered by the treating physician. Pulmonary embolism was defined as: positive computed tomography scan, high probability ventilation/perfusion scan, positive pulmonary angiogram, or PE on 3-month follow-up. We calculated sensitivity, specificity, positive and negative predictive value, and likelihood ratios for the ELISA D-dimer. We prospectively enrolled 541 patients who underwent D-dimer testing for PE, of whom 130 patients had Wells Score>4 and 33 patients had Wells Score>6 (not mutually exclusive). Of subjects with Wells Score>4, 23 (18%) were diagnosed with PE and 40 (31%) had a negative D-dimer. No patient with Wells Score>4 (sensitivity 100%, 95% confidence interval [CI] 82%-100%; specificity 37%, 95% CI 28%-47%) or Wells Score>6 (sensitivity 100%, 95% CI 63%-100%; specificity 56%, 95% CI 35%-76%) who had a negative D-dimer was diagnosed with PE. The likelihood ratio for a negative D-dimer was 0 for both the Wells>4, and Wells>6 groups, however, the upper limits of the confidence interval around the post-test probability for PE were 16% and 33%, respectively, for these high probability groups. In this pilot study, the rapid ELISA D-dimer had high sensitivity and negative predictive value even when applied to patients with high pretest probability for PE. However, with the post-test probability of PE still as high as 16-33% in the negative D-dimer groups, this precludes applying the results to patient care at present. Further testing is warranted to determine whether these findings can be safely incorporated into practice.