Combined use of pretest clinical probability score and latex agglutination D-dimer testing for excluding acute deep vein thrombosis.

OBJECTIVE: Currently, the latex agglutination D-dimer assay is widely used for excluding deep vein thrombosis (DVT) but is considered less sensitive than the enzyme-linked immunosorbent assay-based D-dimer test. The purpose of the present study was to determine if a combination of different cutoff points, rather than a single cutoff point of 1.0 microg/mL, on the latex agglutination D-dimer assay and the pretest clinical probability (PTP) score would be able to reduce the use of venous duplex ultrasound (DU) scanning in patients with suspected DVT.

METHODS: The PTP score and D-dimer testing were used to evaluate 989 consecutive patients with suspected DVT before venous DU scanning. After calculating the clinical probability scores, patients were divided into low-risk (< or =0 points), moderate-risk (1-2 points), and high-risk (> or =3 points) pretest clinical probability groups. Receiver operating characteristic (ROC) curve analysis was used to determine the appropriate D-dimer cutoff point for each PTP with a negative predictive value of >98% for a positive DU scan.

RESULTS: There were 886 patients enrolled. The study group included 609 inpatients (68.7%) and 277 outpatients (31.3%). The prevalence of DVT in this series was 28.9%. There were 508 patients (57.3%) classified as low-risk, 237 (26.8%) as moderate-risk, and 141 (14.9%) as high-risk PTP. DVT was identified in 29 patients (5.7%) with low-risk, 118 (49.8%) with moderate-risk, and 109 (77.3%) with high-risk PTP scores. ROC curve analysis was used to select D-dimer cutoff points of 2.6, 1.1, and 1.1 microg/mL for the low-, moderate- and high-risk PTP groups, respectively. In the low-risk PTP group, specificity increased from 48.9% to 78.2% (P < .0001) with use of the different D-dimer cutoff value. In the moderate- and high-risk PTP groups, however, the different D-dimer levels did not achieve substantial improvement. Despite this, the overall use of venous DU scanning could have been reduced by 43.0% (381 of 886) if the different D-dimer cutoff points had been used.

CONCLUSIONS: Combination of a specific D-dimer level with the clinical probability score is most effective in low-risk PTP patients for excluding DVT. In moderate- and high-risk PTP patients, however, the recommended cutoff points of 1.0 microg/mL may be preferable. These results show that different D-dimer levels for patients differing in risk is feasible for excluding DVT using the latex agglutination D-dimer assay.