Practice changer
Use a combination of Wells score and D-dimer test to exclude deep vein thrombosis in low- to intermediate-risk outpatients with suggestive symptoms.1
Strength of recommendation (SOR)
A: Based on one good meta-analysis
Goodacre S, Stevenson M, Wailoo A, Sampson F, Sutton AJ, Thomas S. How should we diagnose suspected deep-vein thrombosis? QJM 2006;99:377–388.
Illustrative case
Your patient is a 47-year-old man who has had pain and swelling in his right leg since yesterday. He has no history of cancer, recent travel, surgery, or blood clots. There has been no known trauma. You observe some tenderness in the calf, and no swelling. A Homan’s sign is negative.
How would you assess for deep venous thrombosis?
Background: What are costs, benefits of different strategies?
Wouldn’t it be nice to be able to rule out deep vein thrombosis (DVT) with a simple history and blood test?
Most patients with suggestive symptoms do not have DVT, but workups for this dangerous condition “range from the accurate but expensive (contrast venography) to the cheap but unreliable (clinical assessment),” noted the researchers who conducted this extensive analysis, seeking a cost-effective strategy to put into practice throughout the United Kingdom’s National Health Service.1 The NHS Health Technology Assessment R&D Program funded the study.
Until now, there has been little clear direction from formal comparisons. Although recent studies2 suggest that combinations of simple diagnostic tests may reduce the need for expensive, definitive tests, none explicitly weigh the costs and benefits of the different strategies; despite a large amount of published data, practice varies considerably.1,3
Clinical Context: Guidelines conflict
While some clinical resources now recommend the Wells criteria and D-dimer as useful tools in the initial workup of suspected DVT,4,5 others still recommend compression ultrasound or impedance plethysmography (IPG) as the initial or confirmatory test in all patients with suspected acute DVT.6 These noninvasive tests do provide reassurance that there is no DVT, but they are costly and less convenient than the Wells score and a D-dimer.
Current guidelines give conflicting recommendations.
- The most recent American Thoracic Society guidelines, from 1999, recommend imaging with ultrasound or impedance plethysmography for all patients with suspected DVT.7
- In contrast, the Institute for Clinical Systems Improvement (ICSI) 2006 guidelines recommend first determining the clinical pretest probability of DVT using the Wells score, and then using a D-dimer test to determine which patients with a low probability test need to proceed to ultrasound. This algorithm recommends ultrasound for all patients with either a moderate or high Wells score.8
Variety of D-dimer tests, range of sensitivities
One meta-analysis of 12 studies compared a highly sensitive ELISA D-dimer assay to the less sensitive SimpliRED D-dimer assay. In studies using the highly sensitive ELISA assay, in patients with negative D-dimer and low or moderate Wells score, the 3-month incidence of DVT was 0.5%. However, using the SimpliRED assay, while the 3-month incidence of DVT was 0.5% with negative D-dimer and low Wells score, it was 3.5% with negative D-dimer and intermediate Wells score.7
Study Summary: Seeking convenience and economy
This systematic review, meta-analysis, and decision analysis sought the most practical, cost-effective strategy to detect DVT. The researchers compared the findings of 18 studies of diagnostic strategies (or algorithms) that combined Wells score (TABLE), D-dimer, ultrasound, or venography, and that followed up patients with negative results for at least 3 months. They developed a decision analysis model to compare the algorithms in a hypothetical cohort of 1000 outpatients with suspected DVT.
Applying the estimated sensitivity and specificity of each algorithm to the hypothetical population, they determined the proportions of patients with and without DVT who would receive treatment and which patients would suffer events relating to DVT or treatment, and then estimated lifetime health outcomes (quality-adjusted life years [QALYs]) and costs of testing and treatment.
Using thresholds for willingness to pay of £10,000, £20,000, and £30,000 per QALY, the study identified 2 optimal diagnostic strategies, both of which incorporated D-dimer testing and Wells score. While one strategy starts with Wells score and the other starts with D-dimer, both recommend that patients with a combination of a negative D-dimer test and an intermediate or low Wells score can be safely discharged without further testing.
One weakness of this study is that the authors made the assumption that ultrasound results are independent of Wells score or D-dimer. While it is unlikely that ultrasound results are related to D-dimer results, there is some evidence that ultrasound is more accurate in patients with higher Wells scores. However, if this true, the authors would have underestimated the cost-effectiveness of the favored strategies. Also, they did not include algorithms that involved plethysmography.