Polymyalgia Rheumatica: New Classification Criteria
At the end of the study period, the investigators developed a scoring algorithm that combined "mandatory" features of disease (age ≥ 50 years, bilateral shoulder aching, and abnormal ESR and CRP) with other clinical features that were most strongly predictive of PMR vs other conditions. Points were assigned to these other features as follows:
The total score was used to determine the accuracy of the PMR diagnosis. Of interest, rapid improvement of symptoms in response to corticosteroid therapy was not a significant factor in distinguishing PMR from other conditions.
In this algorithm, a score of ≥ 4 was 68% sensitive and 78% specific for discriminating the control group from PMR. The specificity was higher (88%) when distinguishing PMR from presumably nonsystemic inflammatory conditions, such as mechanical rotator cuff disease, but lower when distinguishing PMR from rheumatoid arthritis (65%). Furthermore, the investigators found that adding ultrasonography findings of tenosynovitis, bursitis, or synovitis in or around the shoulders or hips changed the diagnostic accuracy for PMR to a sensitivity of 66% and a specificity of 81%.
The investigators concluded that with this approach, patients ≥ 50 years with bilateral shoulder pain, abnormal ESR or CRP, and no alternative pathology could be classified with fair diagnostic accuracy as having PMR. However, they were careful to state that these provisional criteria are only for research purposes and, at this juncture, are not intended to be used for clinical diagnosis of PMR.
The multiyear effort of these investigators to establish provisional classification criteria for PMR is to be applauded. However, PMR is still a disease with protean manifestations, it lacks an infallible gold standard for diagnosis, and apparently even rapid improvement of symptoms in response to corticosteroid therapy may not be sufficient to distinguish it from other conditions. As such, establishing a widely accepted set of classification criteria that can ultimately be used for diagnostic purposes will be difficult. However, let us hope -- as proposed by Dasgupta and colleagues -- that these provisional criteria will set the stage for further studies that can shed light on the pathophysiology of PMR and perhaps lead to both improved diagnostic approaches and improved therapies.
Abstract
A Scoring Algorithm
At the end of the study period, the investigators developed a scoring algorithm that combined "mandatory" features of disease (age ≥ 50 years, bilateral shoulder aching, and abnormal ESR and CRP) with other clinical features that were most strongly predictive of PMR vs other conditions. Points were assigned to these other features as follows:
Morning stiffness > 45 minutes (2 points);
Hip pain or limitation of movement (1 point);
Absence of rheumatoid factor or antibodies to citrullinated protein antigens (2 points); and
Absence of other joint pain (1 point).
The total score was used to determine the accuracy of the PMR diagnosis. Of interest, rapid improvement of symptoms in response to corticosteroid therapy was not a significant factor in distinguishing PMR from other conditions.
In this algorithm, a score of ≥ 4 was 68% sensitive and 78% specific for discriminating the control group from PMR. The specificity was higher (88%) when distinguishing PMR from presumably nonsystemic inflammatory conditions, such as mechanical rotator cuff disease, but lower when distinguishing PMR from rheumatoid arthritis (65%). Furthermore, the investigators found that adding ultrasonography findings of tenosynovitis, bursitis, or synovitis in or around the shoulders or hips changed the diagnostic accuracy for PMR to a sensitivity of 66% and a specificity of 81%.
The investigators concluded that with this approach, patients ≥ 50 years with bilateral shoulder pain, abnormal ESR or CRP, and no alternative pathology could be classified with fair diagnostic accuracy as having PMR. However, they were careful to state that these provisional criteria are only for research purposes and, at this juncture, are not intended to be used for clinical diagnosis of PMR.
Viewpoint
The multiyear effort of these investigators to establish provisional classification criteria for PMR is to be applauded. However, PMR is still a disease with protean manifestations, it lacks an infallible gold standard for diagnosis, and apparently even rapid improvement of symptoms in response to corticosteroid therapy may not be sufficient to distinguish it from other conditions. As such, establishing a widely accepted set of classification criteria that can ultimately be used for diagnostic purposes will be difficult. However, let us hope -- as proposed by Dasgupta and colleagues -- that these provisional criteria will set the stage for further studies that can shed light on the pathophysiology of PMR and perhaps lead to both improved diagnostic approaches and improved therapies.
Abstract