Am Fam Physician. 2010;82(2):203
Background: Contrast venography remains the standard for diagnosing suspected lower extremity deep venous thrombosis (DVT). Compression ultrasonography for diagnosing proximal DVT has been successfully implemented in many clinical settings. Its use in detecting distal lower extremity DVT is debatable. Current guidelines recommend follow-up compression ultrasonography five to seven days after an initial negative result because distal lower extremity DVT may propagate into proximal veins. This may not be effective because a small number (1 to 2 percent) of repeat images are positive for thrombus propagation, and patients may not return for their repeat procedure. For this reason, the ability of a single whole-leg compression ultrasound image to diagnose lower extremity DVT has been evaluated. Johnson and colleagues conducted a systematic review and meta-analysis to determine the risk of venous thromboembolism in patients for whom anticoagulant therapy was withheld following a single negative result on whole-leg compression ultrasonography.
The Study: The authors researched randomized controlled trials and prospective studies published from January 1970 through November 2009 that addressed whole-leg compression ultrasonography for suspected symptomatic DVT. Studies were included if they evaluated suspected lower extremity DVT with a single whole-leg compression ultrasound scan; had a follow-up period of at least 90 days, during which anticoagulant therapy was withheld after a negative result on whole-leg compression ultrasonography; and had objective confirmation of venous thromboembolism during the follow-up period. Studies that were retrospective, that included asymptomatic patients and those on anticoagulant therapy, or that used only partial-leg compression ultrasonography or interim data were excluded. Seven studies (six prospective and one randomized controlled trial) were used. Of the 11,851 patients screened, 10,090 were enrolled. The patients were from inpatient and ambulatory care settings, although the majority were from the latter.
Results: A total of 4,731 patients were included in the meta-analysis. Venous thromboembolism or suspected venous thromboembolism–related death occurred in 34 patients (0.7 percent). These included 11 distal DVT events, seven proximal DVT events, seven nonfatal pulmonary embolisms, and nine deaths in which thrombolism was the suspected cause but could not be proven because of the lack of autopsies. The combined venous thromboembolism risk at three months was 0.57 percent. For two of the seven studies included, the authors used the pretest probability for venous thromboembolism in the patients studied (1,618 patients). Wells scores categorized the patients as low risk (1,071 patients), moderate risk (467 patients), and high risk (80 patients). They found the pooled venous thromboembolism incidence rates from these subgroups were 0.29 percent, 0.82 percent, and 2.49 percent, respectively.
Conclusion: There is a low risk of venous thromboembolism for patients not started on anticoagulant therapy after a single negative result on whole-leg compression ultrasonography for a suspected lower extremity DVT. Their conclusion can be upheld only for the low-risk pre-test probability subgroup, when the pretest probability of venous thromboembolism is incorporated into their assessment. Further studies are needed to address the moderate- and high-risk subgroups because a relatively lower number of these patients were studied.
editor's note: It is useful to note that what the authors refer to as whole-leg compression ultrasonography also has been termed “complete venous ultrasound” and “comprehensive duplex ultrasonography” in the literature. The technique involves real-time imaging in which failure to compress vein walls under pressure constitutes diagnostic criterion for venous thrombosis. It is 95 percent sensitive and 98 percent specific for above-the-knee veins and has a low technical failure rate.1 Doppler waveform analysis is incorporated into ultrasound imaging as either duplex or color flow ultrasonography. An accompanying editorial on the study done by Johnson and colleagues warns that meta-analyses such as this may be oversimplifying medical care by averaging variable clinical scenarios in which the individual studies were performed (i.e., this meta-analysis included studies from both ambulatory and inpatient clinical settings).2--SUMI SEXTON, MD, Associate Medical Editor, American Family Physician