History, Physical, and D-Dimer: Another Clinical Decision Rule for Diagnosing DVT

*Patients with a score ≤3 did not receive a referral for ultra-sonography; patients with a score ≥4 did.

†Calf circumference was measured 10 cm below the tibial tubercle.

§Inverness Medical, Bedford, United Kingdom

From Buller HR, ten Cate-Hoek AJ, Hoes AW, et al. Safely ruling out deep venous thrombosis in primary care. Ann Intern Med. 2009;150:229-35. Reprinted with permission.

A prospective management study utilized more than 300 primary-care practices in the Netherlands to assess 1,028 consecutive patients with clinically suspected DVT. A clinical decision rule was devised that included history and physical findings as well as a point-of-care D-dimer assay. The latter used fingerprick capillary blood to measure elevated D-dimer by immunochromatography read as abnormal if, next to the control, there appears a second band within 10 minutes. This test has proven to be highly sensitive and to have a high negative predictive value for DVT.^2-3  Clinicians calculated a score using the clinical decision rule (see Table). Patients with a score of 4 or more were referred for ultrasonography. DVT was considered present when one proximal vein in the leg was non-compressible. The primary outcome was the incidence of symptomatic venous thromboembolic disease (VTE) — DVT and/or pulmonary embolus.

Seven of the 500 patients (1.4%) with a score of 3 or less who did not have ultrasonography developed VTE. In contrast, 125 of the 499 (25%) with a score of 4 or more who had ultrasonography showed DVT (overall prevalence 125/1,002 or 13%). Of the 374 patients with normal ultrasounds, only four developed VTE within the subsequent three months. If a clinician had used only the clinical characteristics without the D-dimer and had not referred patients for ultrasound with scores of 3 or less, DVT would have been missed in about 10 percent. If the clinician had just used D-dimer negativity, DVT would have been missed in about 3 percent. In contrast, patients with negative D-dimer and scores greater than 4 had a 23.5 percent probability of having DVT.

The authors concluded that primary-care physicians can use a simple clinical decision rule that includes point-of-care D-dimer testing to safely rule out 50 percent of patients presenting with signs and symptoms suggesting DVT and thereby reduce referral for secondary testing and care. The D-dimer test is most useful when the clinical score is low and less useful when the clinical score is high.

An accompanying editorial by Reilly and Evans⁴  suggests caution before adopting these prediction rules in clinical practice. These results have not been compared with the better-known Wells rule^5,6  in a head-to-head comparison. Büller, et al. were concerned that 1.8 percent to 2.3 percent of low-risk patients had DVT in validation studies of the Wells rule. However, the Büller predictor variables may be less subjective. Furthermore, extrapolation to other patient settings requires caution since Büller’s study included many patients at very low risk for DVT (leg trauma) and had a lower DVT prevalence (13%) than most other studies. Finally, the Büller prediction rule has not undergone an impact analysis examining clinicians’ actual decisions when using the rule or not.

From Washington, the recent economic stimulus bill has provided monies for comparative effectiveness research. Clearly, studies as described above help primary-care physicians on the front lines effectively care for patients with suspected DVT. This study echoes what we learned in the second year of medical school: A good history, physical examination, and targeted, confirming laboratory tests reflect the qualities of the most effective hematologists and primary-care physicians.