Finally an Optimal Strategy for Diagnosing Deep Vein Thrombosis

Abstract & Commentary

By Joseph E. Scherger, MD, MPH, Professor, University of California, San Diego. Dr. Scherger reports no financial relationship to this field of study.

Synopsis: A medical research unit in the United Kingdom analyzed 18 different strategies for managing patients with suspected deep vein thrombosis (DVT). Using systematic review, meta-analysis and cost effectiveness analysis, they settled on an algorithm based on clinical risk. The initial screen for all patients is a latex D-dimer test. If the D-dimer test is negative and the patient is low or intermediate risk (Wells rule), DVT is ruled out. If the clinical risk is high, and the D-dimer test is negative, an above-the-knee ultrasound is ordered, and, if positive, the patient is treated. If the D-dimer test is positive, an ultrasound is ordered regardless of risk. If the ultrasound is negative and the patient is high risk, it is repeated before a treatment or discharge decision is made.

Source: Goodacre S, et al. How should we diagnose suspected deep-vein thrombosis? QJM. 2006;99:377-388.

Deep vein thrombosis of the proximal lower extremity is one of the most important and challenging diagnoses to make. The treatment requires inpatient care and is potentially life saving. Multiple evolving diagnosis and treatment algorithms are available which make clinical decision making difficult. Variations in the management of patients with suspected DVT are an important quality issue. Credit the United Kingdom and its National Health Service to commission a meticulous detailed analysis of multiple strategies to come up with an approach which is both highly accurate and cost effective.

After a systematic review and meta-analysis of 18 different strategies, a hypothetical group of 1000 patients were managed using each algorithm. A mean survival of 11.6 quality adjusted life years (QALYs) was used after a diagnosis of DVT at age 60. Clinical risk was assigned using the Wells Clinical Prediction Rule using 9 factors to score the risk of DVT. A recent analysis of the Annals of Internal Medicine questioned the reliability of the Wells rule in ruling out DVT.1 However, in the British analysis, the Wells rule was mostly used to identify patients at high risk of DVT for further analysis even if the D-dimer test is negative.

The latex D-dimer test is a major advance in screening for DVT and other hypercoagulation states. The D-dimer reflects fibrin degradation products which indicate thrombus formation, and is positive in a number of conditions besides DVT, such as disseminated intravascular coagulation (DIC). The test is sensitive for DVT but not specific, and may rise in the elderly, with false positives in patients with rheumatoid arthritis, high triglycerides, and elevated bilirubin. The test is not diagnostic for DVT and should be followed up by more specific testing such as venous ultrasound.

Using the algorithms in this study, the percentage of patients with proximal DVT who would be treated appropriately ranged from 90.1% to 99.5%, and the patients without DVT treated inappropriately ranged from 0.6% to 6.0%. The final recommendations for the most accurate and cost effective algorithm are quite simple. The initial screen for all patients is a latex D-dimer test. If the D-dimer test is negative and the patient is low or intermediate risk (Wells rule), DVT is ruled out. If the clinical risk is high, and the D-dimer test is negative, an above the knee ultrasound is ordered and if positive the patient is treated. If the D-dimer test is positive, an ultrasound is ordered regardless of risk. If the ultrasound is negative and the patient is high risk, it is repeated before a treatment or discharge decision is made.

Commentary

Even though this study is based on hypothetical patients, the analysis is quite rigorous, and uses the best available evidence. No combination of tests is perfect and there is always a risk of a missed diagnosis. A main contribution of this analysis is that not all patients require ultrasound, which may be difficult to obtain acutely in the primary care setting. As the study in the Annals of Internal Medicine points out, some patients (2.9% in their analysis) will have a low Wells score and a negative D-dimer test and will be found on ultrasound to have DVT. Maybe clinical judgment here should also include the "thin slicing" of rapid clinical intuition described by Malcolm Gladwell in Blink.2 If our initial assessment makes one think of DVT, we should be quite certain it is not present before accepting that it is ruled out.

We now have coherence in the diagnosis of suspect DVT. The three-part assessment is easy to remember, D-dimer testing, the Wells rule and venous ultrasound. Combined with the clinical judgment of an experienced clinician, we have sound and consistent strategies for the diagnosis of DVT.

References

1. Oudega R, et al. The Wells rule does not adequately rule out deep venous thrombosis in primary care patients. Ann Intern Med. 2005;143:100-107.

2. Gladwell M. Blink. Little, Brown and Company. New York, NY: Time Warner Book Group, 2005.