Context: Outpatients with suspected deep vein thrombosis (DVT) have nonspecific signs and symptoms. Missed DVT diagnosis may result in fatal pulmonary embolism. Since many patients may have DVT, a selective and efficient diagnostic process is needed.
Objective: To systematically review trials that determined the prevalence of DVT using clinical prediction rules either with or without D-dimer, for the diagnosis of DVT.
Data sources: English- and French-language studies were identified from MEDLINE from 1990 to July 2004 and supplemented by a review of all relevant bibliographies.
Study selection: We included studies that prospectively enrolled consecutive, unselected outpatients with suspected DVT and applied clinical prediction rules before D-dimer testing or diagnostic imaging. All studies included sufficient information to allow the calculation of the prevalence of DVT for at least 1 of the 3 clinical probability estimates (low, moderate, or high). We required that patients be followed up for a minimum 3-month period. Unless the clinical model incorporated prior DVT, studies were excluded if patients with a history of prior DVT were enrolled.
Data extraction: Two reviewers independently reviewed and abstracted data for estimating the prevalence of DVT, sensitivity, specificity, and likelihood ratios (LRs) of D-dimer in each of the 3 clinical probability estimates. Data for the D-dimer in all studies were pooled and analyzed as high-sensitivity/low-specificity test or a moderate-sensitivity/moderate-specificity test.
Data synthesis: Fourteen studies involving more than 8000 patients used 1 clinical prediction rule for diagnosing DVT, of which 11 incorporated D-dimer testing in the diagnostic algorithm. The prevalence of DVT in the low, moderate, and high clinical probability groups was 5.0% (95% CI, 4.0%-8.0%), 17% (95% CI, 13%-23%), and 53% (95% CI, 44%-61%), respectively. The overall prevalence of DVT was 19% (95% CI, 16%-23%). Pooling all studies, the sensitivity, specificity, and negative LRs of D-dimer testing in the low probability group were 88% (95% CI, 81%-92%), 72% (95% CI, 65%-78%), and 0.18% (95% CI, 0.12-0.18); in the moderate probability group: 90% (95% CI, 80%-95%), 58% (95% CI, 49%-67%), and 0.19% (95% CI, 0.11-0.32); and in the high probability group: 92% (95% CI, 85%-96%), 45% (95% CI, 37%-52%), and 0.16% (95% CI, 0.09-0.30). The LRs for a normal result on a high or moderately sensitive D-dimer assay among patients with: (1) low clinical suspicion were 0.10 (95% CI, 0.03-0.37) and 0.20 (95% CI, 0.12-0.31); (2) moderate clinical suspicion were 0.05 (95% CI, 0.01-0.21) and 0.23 (95% CI, 0.13-0.39); and (3) high clinical suspicion were 0.07 (95% CI, 0.03-0.18) and 0.15 (95% CI, 0.10-0.38).
Conclusions: Diagnostic accuracy for DVT improves when clinical probability is estimated before diagnostic tests. Patients with low clinical probability on the predictive rule have prevalence of DVT of less than 5%. In low-probability patients with negative D-dimer results, diagnosis of DVT can be excluded without ultrasound; in patients with high clinical suspicion for DVT, results should not affect clinical decisions.