A comparison of the sensitivity of diagnostic criteria for polymyalgia rheumatica

Objective: To compare the performance of the several different diagnostic criteria sets currently in use for polymyalgia rheumatica (PMR). Methods: 213 patients attending eight rheumatological centres in eight different European countries were studied. All had recently been referred and were considered by the senior investigator at each centre, selected because of their experience in treatment of PMR, to have this condition. By use of a standard international proforma, the requisite diagnostic points in each criteria set were sought. Sensitivity for each criterion from each set was then calculated, as well as the sensitivity of each criteria set as a whole. Results: Of four criteria sets compared, the Bird (1979) criteria performed best with a sensitivity of 99.5%, and the Hunder (1982) criteria second best, with sensitivity of 93.3%. These both performed significantly better than the two other criteria sets, though each of these was admittedly developed for rather specialised reasons. Conclusions: Although this study compares homogeneity, we suggest the Bird 1979 or Hunder 1982 criteria should be used whenever possible. Studies that have used alternative criteria may have less sensitivity in diagnosis.