Am Fam Physician. 2006;73(4):709-710
Prospectively validated clinical prediction rules such as the Wells rule for suspected deep venous thrombosis (DVT) allow for more accurate assessment of patients (see accompanying table). A Wells score of 1 or less, combined with a normal D-dimer test, is thought to lower the risk of DVT and make diagnostic lower leg ultrasonography unnecessary. Although the Wells rule has been validated in multiple groups of secondary care outpatients, concerns have been raised that this population is dissimilar to primary care populations. Oudega and colleagues conducted a cross-sectional study at 110 primary care practices in the Netherlands to evaluate the performance of the Wells prediction rule combined with D-dimer testing in DVT risk stratification.
The study population consisted of 1,295 consecutive adult patients seen between January 2002 and March 2003 who presented to their primary care physician with leg swelling, redness, or pain for 30 days or less. Patients with symptoms suggestive of pulmonary embolism were excluded. All primary care physicians in a geographically defined area of the Netherlands participated in the study. Before the study, physicians received instructions by mail on how to apply the Wells rule and attended a workshop on study logistics.
| Clinical feature | Score |
|---|---|
| Active cancer | 1 |
| Paralysis, paresis, or recent plaster immobilization of the lower extremity | 1 |
| Recently bedridden for more than three days or major surgery within four weeks | 1 |
| Localized tenderness along the distribution of the deep venous system | 1 |
| Entire leg swollen | 1 |
| Calf swelling by more than 3 cm when compared with the asymptomatic leg | 1 |
| Pitting edema (greater in the symptomatic leg) | 1 |
| Collateral superficial veins (nonvaricose) | 1 |
| Alternative diagnosis as likely or more possible than that of DVT | −2 |
Patients were assigned a Wells score in the office and then were referred to local hospitals for D-dimer testing and lower extremity duplex ultrasonography. Patients with initially normal ultrasound results had repeat ultrasound tests seven days later. These data were used to calculate the sensitivity and specificity of DVT risk predicted by Wells scores and D-dimer test results.
In contrast to previous studies performed at secondary referral centers, this study showed that 12 percent of patients usually defined as low risk (i.e., Wells score of 0 or less) had evidence of DVT on ultrasonography. Furthermore, a normal D-dimer result and a low-risk Wells score (the “rule out” combination) missed 2.9 percent of patients with DVTs. Lowering the low-risk score threshold to –1 or less would cut the number of missed DVTs in half, but more patients would need to undergo ultrasonography.
The authors conclude that an unacceptably high percentage of primary care patients with DVT were classified as low risk using the Wells prediction rule, even in combination with a normal D-dimer result. They recommend the development of a new prediction rule to exclude DVT that is specific to primary care patients.
editor’s note: Given this study’s conclusion, a question arises for family physicians: until a new prediction rule is validated in primary care patients, should we stop using the Wells rule to assess DVT risk? In an accompanying editorial,1 Douketis asserts that the Wells rule, in combination with D-dimer testing, remains a useful aid to clinical judgment in determining which patients to refer for ultra-sonography. A meta-analysis2 in the same issue found the Wells rule to be more valuable than any single clinical feature in identifying patients at high risk of DVT. Although the study by Oudega and colleagues provides a cautionary lesson in applying prediction rules to populations different from those in which it was derived, it would be unwise to throw out a useful tool.—k.w.l.