Abstract

Background: The immunofiltration D-dimer assay could allow point-of-care testing for pulmonary embolism (PE). A study was undertaken to compare a clinician-performed qualitative D-dimer assay with the automated quantitative D-dimer test.

Methods: A prospective observational study was conducted from January to October 2005 at an urban academic emergency department (ED). 1193 patients of mean (SD) age 47 (16) years (66% female) were enrolled. The study protocol combined pretest probability estimation, D-dimer testing by both a qualitative immunochromatographic assay (Simplify) performed at the point of care by 192 different clinicians and a quantitative D-dimer test performed in a CLIA-certified laboratory. The criterion standard was image-proven PE or deep venous thrombosis within 45 days after enrolment. To test interobserver agreement for the qualitative assay, two blinded observers independently read 841 Simplify cartridges.

Results: Of 1193 patients enrolled, 45 were PE+ (3.8%, 95% CI 2.8% to 5.0%). Qualitative results were available for 1169 (98%) and quantitative results were available for 1136 (95%). Comparison of the qualitative and quantitative D-dimer tests gave the following results: sensitivity 91% (95% CI 78% to 98%) vs 93% (95% CI 80% to 98%); specificity 57% (95% CI 54% to 60%) vs 57% (95% CI 54% to 60%); likelihood ratio negative 0.16 (95% CI 0.06 to 0.37) vs 0.13 (95% CI 0.05 to 0.35). The weighted Cohen’s κ for interpretation of the qualitative assay was 0.69 (95% CI 0.63 to 0.76).

Conclusions: In this very low-risk ED population, a qualitative D-dimer assay performed at the point of care had similar diagnostic accuracy to the quantitative D-dimer test. Interobserver agreement for the qualitative test was good.