Pitfalls in double contrast knee arthrography

Two hundred and forty-two patients investigated by double-contrast knee arthrography between 1973 and 1977 have been reviewed. Of these 134 have come to arthrotomy and an initial arthrographic accuracy of 85% is reported. An attempt to explain the disparities between arthrography and arthrotomy has revealed that in half the cases (ten patients) there were simple reporting errors on the arthrograms. These are shown all to be false negative reports. The other group of ten patients comprised presumed errors in arthrotomy interpretation in four and arthrographic interpretation or technical failure in six. The causes and implications of these errors are analysed and discussed.