Selective use of D-dimer based on pretest probability identified first suspected DVTs with less testing

More selective use of D-dimer testing allowed physicians to safely and efficiently diagnose first episodes of deep venous thrombosis (DVT), a new study found.

The randomized, controlled trial included more than 1,500 patients who presented to Canadian hospitals with symptoms of DVT. Physicians used the 9-point Wells clinical prediction rule to assess whether patients' clinical pretest probability of DVT was low, moderate or high. Then patients were randomized to one group in which all patients were uniformly given D-dimer tests (and given ultrasonography based on those results) or one in which pretest probability determined testing.

In the latter group, patients who had a low pretest probability and a D-dimer level below 1.0 µg/mL had DVT excluded as their diagnosis. For patients with a moderate pretest probability, the D-dimer cutoff was 0.5 µg/mL. Patients who scored below either of these levels did not receive ultrasonography. Outpatients with high pretest probability and all inpatients were not given D-dimer tests and instead all received ultrasonography. Patients were followed for three months, and results were published in the Jan. 15 Annals of Internal Medicine.

At three months, the selective and uniform testing groups had equal incidence of symptomatic VTE: 0.5% (difference between groups, 0 percentage points; 95% CI, −0.8 to 0.8 percentage point). Selective testing reduced the proportion of patients getting D-dimer tests by 21.8 percentage points (95% CI, 19.1 to 24.8 percentage points) and ultrasonography by 7.6 percentage points (95% CI, 2.9 to 12.2 percentage points). Outpatients with low pretest probability had a particularly steep drop in ultrasonography: 21 percentage points (95% CI, 14.2 to 27.6 percentage points).

Study authors concluded that the selective testing was as safe as and more efficient than uniform testing and resulted in a similar number of patients being diagnosed with VTE during testing. They noted that none of the patients with D-dimer levels between 0.5 and 1.0 µg/mL were diagnosed with VTE during the study and that a very small percentage of the high-risk patients in the control group had DVT excluded by D-dimer testing (15% of outpatients with high pretest probability and 2% of inpatients).

They cautioned that the results may not be generalizable to patients with a history of DVT or to other D-dimer tests but called for research on using selective testing in patients who present with suspected recurrences. For first suspected episodes of DVT, the results support basing testing choices on pretest probability, they concluded.