Prospective Evaluation of a Screening Protocol to Exclude Deep Vein Thrombosis on the Basis of a Combination of Quantitative D-Dimer Testing and Pretest Clinical Probability Score Takashi Yamaki, MD, Motohiro Nozaki, MD, Hiroyuki Sakurai, MD, Masaki Takeuchi, MD, Kazutaka Soejima, MD, Taro Kono, MD BACKGROUND: Clinical signs and symptoms such as swelling, pain, and redness are unreliable markers of deep vein thrombosis (DVT). Because of this venous duplex scanning (VDS) has been heavily used in DVT detection. The purpose of this study was to determine if a combination of D-dimer testing and pretest clinical score could reduce the use of VDS in symptomatic patients with suspected DVT. STUDY DESIGN: One hundred seventy-four consecutive patients with suspected DVT were prospectively eval- uated using pretest clinical probability (PCP) score and D-dimer testing before VDS. After calculating clinical probability scores developed by Wells and associates, patients were divided into low risk ( 0 points), moderate risk (1 to 2 points), and high risk (3 points) PCP. RESULTS: One hundred fifty-eight patients were enrolled. The prevalence of DVT in this study was 37%. Thirty-eight patients (24%) were classified as low risk, 64 (41%) as moderate risk, and 56 (35%) as high risk PCP. DVT was identified in only one patient (2.6%) with low risk PCP. In contrast, DVT was found in 22 (34%) with moderate risk, and 35 (63%) with high risk PCP. In the high and moderate risk PCP groups, positive scan patients had a markedly higher value of D-dimer assay than negative scan patients (p = 0.0001 and p = 0.0057, respectively). In the low risk PCP patients, D-dimer testing provided 100% sensitivity, 46% specificity, 4.8% positive predictive value, and 100% negative predictive value in the diagnosis of DVT. Similarly, in the moderate risk PCP, the D-dimer testing showed 100% sensitivity, 45% specificity, 49% positive predictive value, and 100% negative predictive value. In the high risk group, D-dimer testing achieved 100% sensitivity, 57% specificity, 80% positive predictive value, and 100% negative predictive value in the diagnosis of DVT. These results suggested that 36 of 158 patients who had a non-high PCP (low and moderate PCP) and a normal D-dimer concentra- tion were considered to have no additional investigation, so VDS could have been reduced by 23% (36/158). CONCLUSIONS: A combination of D-dimer testing and clinical probability score may be effective in avoiding unnecessary VDS in suspected symptomatic DVT in the low and moderate PCP patients. The need for VDS could be reduced by 23% despite a relatively high prevalence of DVT. (J Am Coll Surg 2005;201:701–709. © 2005 by the American College of Surgeons) Currently, more than 600,000 cases of pulmonary em- bolism occur with an estimated annual incidence of 23 to 69 per 100,000 population in the US alone. 1-3 Because pulmonary embolism is potentially life threatening, and 80% of cases arise from lower extremity veins, 4 diagnosis and treatment of deep vein thrombosis (DVT) is of pri- mary importance. Although it is well recognized that the risk of DVT increases in patients with specific diseases, 5-7 clinical signs and symptoms such as swelling, pain, and redness are unreliable markers of DVT. So attempts have been made to improve the discriminating power in clin- ical assessment. 8-10 But this method cannot be used alone in clinical decision making. Competing Interests Declared: None. Received May 4, 2004; Revised June 17, 2005; Accepted June 20, 2005. From the Department of Plastic and Reconstructive Surgery, Tokyo Women’s Medical University, Tokyo, Japan. Correspondence address: Takashi Yamaki, MD, Department of Plastic and Reconstructive Surgery,Tokyo Women’s Medical University, 8-1, Kawada- cho, Shinjuku-ku, Tokyo, 162-8666, Japan. 701 © 2005 by the American College of Surgeons ISSN 1072-7515/05/$30.00 Published by Elsevier Inc. doi:10.1016/j.jamcollsurg.2005.06.267