The patterns of abnormality (both site and intensity of uptake) on bone scintigraphy in 76 total hip replacements were studied and graded to determine whether specific patterns of abnormality were associated with greater accuracy in deciding the need for revision surgery. Patients were divided into three groups: 1) 45 hips were considered to have required surgery, 2) in 15 hips surgery was retrospectively considered unnecessary, and 3) 16 prostheses were painless. There was a trend for fewer abnormal sites and a lower intensity of abnormality from group 1 to group 3. By considering patterns of uptake, the specificity of bone scintigraphy could be improved from 65% to 95%, but at the expense of lower sensitivity, falling from 90% to 75%. The most accurate criteria for determining the need for revision surgery (sensitivity 88%, specificity 85%) were increased uptake at the tip or a total scintigraphy score of greater than 3.