Contrast venography, the gold standard for the diagnosis of deep-vein thrombosis: improvement in observer agreement.

To determine whether the Rabinov-Paulin or the long-leg venography technique should be preferred in the diagnostic management of patients with clinically suspected deep-vein thrombosis, two independent experienced radiologists blindly assessed two different series of venograms of consecutive outpatients with clinically suspected deep-vein thrombosis. Venograms were obtained from two outpatient clinics of primary referral centres. In one centre the venograms were performed according to the technique of Rabinov and Paulin with the use of 100 ml of radiographic material and spot films of the calf, popliteal and more proximal veins. In the other centre, long-leg films were obtained after the administration of 150 ml of contrast material. The percentage venograms adjudicated as inadequate by at least one radiologist and inter-observer disagreement for both series were used as the main study outcome measures. Prior to the study, both radiologists agreed on the standardized criteria for a normal, abnormal and inadequate test result using a separate set of films. An inadequacy rate of 20% was found for the Rabinov-Paulin venography series (n = 123), whereas only 2% of the 126 long-leg films were inadequate for interpretation (p less than 0.001). The inter-observer diagreement for inadequacy, presence or absence of deep-vein thrombosis was 21% for the Rabinov and Paulin venograms and 4% for the long-leg films (kappa, 0.65 and 0.92; 95% confidence intervals: 0.53 to 0.77 and 0.84 to 0.99, respectively; p less than 0.002).(ABSTRACT TRUNCATED AT 250 WORDS)