Combination of a clinical risk assessment score and rapid whole blood D-dimer testing in the diagnosis of deep vein thrombosis in symptomatic patients

AbstractPurpose: The clinical diagnosis of deep vein thrombosis (DVT) is unreliable in more than 50% of cases, and, in recent years, plasma D-dimer assays have been used to predict the presence of DVT with high sensitivity and negative predictive values. This study investigated the use of a rapid whole blood test that could be performed and interpreted by the bedside in addition to a clinically derived risk assessment score (RAS) to determine whether a practical and cost-effective clinical model could be developed for the diagnosis and exclusion of DVT in symptomatic patients. Methods: Two hundred consecutive patients who were referred to the vascular laboratory with clinically suspected DVT underwent the following procedures: (1) clinical assessment and stratification into low, moderate, or high risk for DVT on the basis of an RAS related to history, symptoms, and physical examination findings; (2) rapid (5-minute) whole blood D-dimer testing with a semi-qualitative agglutination technique on a fingerprick blood sample; and (3) color flow duplex ultrasound scanning with standard criteria for the diagnosis of proximal and calf DVT. Results: Forty-six patients (23%) had acute DVT on duplex ultrasound scanning. Of these cases, 28 (61%) had proximal DVT and 18 (39%) were confined to the calf. A total of 88 patients were classified as low risk, 67 were classified as moderate, and 45 were classified as high risk on the basis of the RAS method. Of the patients who were classified at low risk, 4.5% (4/88) had DVT, as compared with 17.9% (12/67) in the moderate clinical risk group and 66.7% (30/45) in the high risk group. Isolated calf DVT was found in 30% (9/30), 50% (6/12), and 75% (3/4) of the high, moderate, and low risk groups, respectively. The sensitivity (8.7%, 26.1%, 65.2%), specificity (45.5%, 64.3%, 90.3%), positive predictive value (4.5%, 17.9%, 66.7%), and overall accuracy rate (37%, 55.5%, 84.5%) of the low, moderate, and high risk groups, respectively, in the diagnosis of DVT increased significantly with increasing risk score stratification. The sensitivity, specificity, positive and negative predictive value, and overall accuracy of the rapid D-dimer test for all DVT were 91.3%, 81.9%, 60%, 96.9%, and 84%, respectively, with a 100% sensitivity and negative predictive value for proximal DVT. A combination of the RAS and D-dimer identified a low risk group with a negative D-dimer as having less than a 1% likelihood of DVT and a high risk group with positive D-dimer as having a likelihood of more than 90%. Conclusion: A combination of clinical assessment and rapid D-dimer testing provides an effective means of excluding proximal DVT in symptomatic patients. The application of a clinical diagnostic model on the basis of these parameters has the potential for saving a large proportion of unnecessary duplex scans with the associated time and costs. (J Vasc Surg 1999;30:794-804.