Objectives: In this randomized management study, we examined the safety of withholding anticoagulation on the basis of negative impedance plethysmography (IPG) compared to negative contrast venography (CV) in symptomatic patients with a first episode of clinically suspected deep vein thrombosis (DVT), and we determined the impact of the limitations of IPG or CV on their clinical utility.
Methods: Patients at a university teaching hospital presenting with a first episode of clinically suspected DVT were randomized to one of two management strategies at study entry: (1) IPG: if positive, confirmatory CV was performed. If CV was positive, anticoagulants were administered, if CV was negative, anticoagulants were held. If negative, IPG was repeated serially and if it remained negative, anticoagulants were held (n = 165). (2) CV: if positive, anticoagulants were administered, if negative, anticoagulants were held (n = 159). The negative predictive value (NPV) of IPG and CV, positive predictive value (PPV) of IPG, and the failure rate of each strategy were assessed.
Results: Among IPG patients, 28 of 37 with positive IPG initially or during serial testing and evaluable CV had confirmed DVT (PPV 76%; 95% confidence interval, CI [62%, 90%]). DVT was diagnosed during serial testing in 2.1% of patients with initially negative IPG who completed testing. The NPV overall of negative IPG was 98.3%. During follow-up, two patients in the IPG group (1.2%) and two patients in the CV group (1.3%) developed venous thromboembolism (VTE). Death during follow-up occurred in 11% of IPG patients compared to 6% of CV patients (P =.13) The investigation strategy failed in 25% of IPG patients and in 14% of CV patients.
Conclusions: Our findings demonstrate that the two diagnostic strategies we studied are equivalent methods for ruling out DVT in patients with a first episode of suspected DVT. The PPV of IPG was too low to permit its use alone as a test to rule in DVT. Both strategies had surprisingly high failure rates.