Phlebography as the Gold Standard in Thromboprophylactic Studies?

In 241 patients with total hip arthroplasty and entering a study on thrombosis prophylaxis, phlebography was adequately performed in 451 legs 7 to 11 days after surgery. The phlebograms were primary evaluated by 4 independent observers, and finally a consensus of the images in which disagreement primarily occurred was obtained. The diagnosis of thrombosis in the 4 primary observations varied between 65% and 83% (mean 70%) and the agreement on a negative diagnosis between 97% and 99% (mean 98%). Taking into account agreement by chance, kappa-values varied from 0.60 to 0.83 when the 6 different pairs of observations were compared. When comparing the primary evaluations with the final consensus, agreements on positive diagnosis varied between 70% and 90% (mean 80%) and on negative diagnosis between 97% and 99% (mean 98%). Kappa-values varied from 0.68 to 0.90. The factor of uncertainty in evaluation of phlebography may have to be considered when studies on postsurgical thromboprophylaxis are planned.