In a randomized, multicenter controlled trial, researchers found that limiting D-dimer testing to outpatients with a low or moderate clinical pretest probability (C-PTP) of developing deep venous thrombosis (DVT), and adjusting the threshold to denote a positive result, did not lead to additional missed diagnoses. The study was published online Jan. 15 in Annals of Internal Medicine.
Currently D-dimer testing is used to rule out DVT. A result of 0.5µg/mL or more is considered positive and requires confirmation by venous ultrasound. Lori-Ann Linkins, MD, of McMaster University in Hamilton, Ontario, and colleagues set out to determine whether raising the threshold level that denotes a positive result for patients at low to moderate C-PTP of DVT would result in additional missed diagnoses. Pointing out that most hospitalized patients have high D-dimer levels due to other conditions, the researchers also tested the impact of omitting D-dimer testing in inpatients and others at high risk of developing DVT.
The study enrolled patients over 18 who presented at five Canadian tertiary care centers between October 2004 and January 2010 with suspected first acute DVT. Clinicians assessed appropriate patients using the nine-point Wells clinical prediction score and placed patients into low-risk, moderate-risk or high-risk categories. Patients from each risk category then were randomized 1:1 into a control group (uniform testing) or a selective testing group. All control group patients received D-dimer testing, as did patients in the selective testing group who were in the low or moderate C-PTP categories. Inpatients and all high-risk patients did not receive D-dimer testing but had ultrasonography of the proximal veins in the symptomatic leg.
A positive result for low-risk patients in the selective testing group was set at 1µg/mL or higher; for moderate-risk patients, a positive result was 0.5µg/mL or higher. If the patient received a negative D-dimer result, no further tests were performed and no anticoagulant therapy was prescribed. All patients who received positive results for D-dimer underwent ultrasonography. Low-risk patients with normal results received no further treatment; moderate-risk patients whose ultrasonography results were negative had repeat ultrasonography six to eight days later.
The study cohort comprised 1,723 patients, 863 randomized to uniform testing and 860 to selective testing. Outpatients made up 89 percent of the sample.
In the uniform testing group, 334 had low C-PTP, 319 moderate C-PTP, 119 high C-PTP and 91 were inpatients. D-dimer results were positive in 506 patients, including 88 inpatients. Initial ultrasonography led to a DVT diagnosis in 56 of them (11.1 percent); four additional DVT were diagnosed through follow-up (0.8 percent of those who had positive D-dimer results).
Of the 860 patients who underwent selective testing, C-PTP was low in 360, moderate in 310, high in 100, and 90 were inpatients. Of the low-risk group, 72 had positive D-dimer results, and DVT was confirmed by ultrasonography in eight of them. Of the moderate-risk patients, 176 had positive D-dimer results, and 23 were diagnosed with DVT by initial ultrasonography and one additional case developed during follow-up. One patient who had a negative D-dimer result developed DVT during follow-up.
Of the group that included 100 outpatients with high C-PTP and 90 inpatients, 20 were diagnosed with DVT during initial ultrasonography and two additional cases were diagnosed during follow-up.
The researchers found no difference in the number of DVT events between the uniform testing group and the selective testing group. Among patients with low C-PTP, the percentage of patients who underwent ultrasonography was 41 percent in the uniform testing group, compared to 20 percent in the selective testing group.
The researchers concluded that the selective testing strategy “was as safe and more efficient than the uniform testing strategy. …The frequency of VTE [venous thromboembolism] during the three-month follow-up did not differ between groups, with confidence bounds surrounding the estimate of no difference well below the clinically significant VTE event rate of 2 percent.”
They cautioned that their results may not be generalizable to patients with a history of DVT. They noted the small number of inpatients enrolled in the study, but claimed that the lower total enrollment did not reduce the power of the results on outpatients.