ABSTRACT
In order to investigate the influence of the interpretation of a postoperative venography, used as the scientific tool in a thromboprophylactic trial, twelve different evaluations of the same venography have been compared. Two hundred and nineteen patients, undergoing total hip replacement, fulfilled a randomized trial evaluating two different prophylactic regimens. Bilateral venography was performed 7-11 days after surgery, and the result of venography was defined as with or without deep venous thrombosis (DVT), irrespective of the location of a thrombus. The venograms were evaluated five times by different interpreters, and the criteria for the presence of DVT were recorded whether or not thrombi in minor veins were present. The incidences of DVT in the two prophylactic groups varied from 2.7 to 20.3% and from 13.5 to 34.2%, respectively, according to the diagnostic criteria. However, the risk reduction between the two groups was almost constant, no matter which type of evaluation was applied (7.9-15.1%), and was statistically significant at the 5% level in all but 4 of the 12 evaluations. It is concluded that despite uniformity in patient selection, the incidence of DVT is not uniform, even though it is stated that venography is used as the end point. The obvious interobserver variation did not change the conclusion when two regimens are compared in a specific randomized study.
