ROME — Candidate international consensus criteria for diagnosis of polymyalgia rheumatica performed well in their first large prospective validation study. But there's room for further improvement with tweaking of the criteria, according to Dr. Bhaskar Dasgupta, the project leader.
The core diagnostic criteria under consideration in a joint ACR/EULAR (American College of Rheumatology/European League Against Rheumatism) project exhibited an overall 85% specificity and 68% sensitivity in discriminating the multitude of conditions that mimic polymyalgia rheumatica (PMR) from true PMR.
Not bad. But the 94% specificity for discriminating rotator cuff tears and other shoulder conditions from PMR was substantially better than the 78% specificity for discriminating rheumatoid arthritis from PMR. And a small percentage of patients—less than 5%—were consistently misclassified using the candidate diagnostic algorithm, Dr. Dasgupta noted at the congress.
“What this means is we have a group of patients—both PMR cases and controls—who are currently unclassifiable. So we need to do more work,” said Dr. Dasgupta, director of rheumatology at Southend University Hospital in Westcliff-on-Sea, England.
He stressed that the consensus diagnostic criteria are still evolving and remain very much a work in progress.
Preliminary results from the study suggest ultrasound could well end up in the final diagnostic algorithm. For the 89 PMR patients and 117 controls whose ultrasound results have been analyzed to date, a finding of bicipital tenosynovitis was present in 68% of the PMR group and 49% of controls.
Subdeltoid bursitis was identified in 60% of PMR patients compared with 35% of controls. Both differences were significant.
Ultrasound had particular value in separating PMR from rheumatoid arthritis, an area where the core diagnostic criteria need shoring up. Subdeltoid bursitis was present in 60% of the PMR group but only 35% with rheumatoid arthritis, according to Dr. Dasgupta.
For all these reasons, the rheumatologist continued, “PMR is a very, very important condition to get right.”
The ACR/EULAR prospective study includes 120 patients with new-onset presumptive PMR as agreed by experts plus 240 controls with new-onset proximal shoulder pain due to various PMR mimickers. These mimicking disorders include myeloma and other malignancies, active infections, fibromyalgia and other chronic pain conditions, osteoarthritis, adhesive capsulitis and other shoulder conditions, rheumatoid arthritis, thyroid disease and other endocrinopathies, and Parkinson's disease. Participants were followed via clinical examinations and lab tests at 1, 4, 12, and 24 weeks.
Seventy-three percent of PMR patients presented with all three of the following features: bilateral shoulder aching, morning stiffness lasting more than 45 minutes, and an elevated C-reactive protein level or erythrocyte sedimentation rate. That was the case for only 29% of controls. All three of these features made it into the core diagnostic inclusion criteria.
Along the way, numerous candidate diagnostic criteria have been weeded out as insufficiently discriminatory to make the short list of core criteria. These include gender, weight loss, neck pain, and limitation of movement.
Features that best discriminated rheumatoid arthritis from PMR included an abnormal rheumatoid factor or anti-citrullinated protein antibody test and peripheral synovitis lasting more than 6 weeks. In contrast, synovitis lasting less than 6 weeks pointed more toward PMR.
The committee is developing a stepwise approach to the diagnosis of PMR. Fully establishing the diagnosis in this way takes 4-6 weeks.
▸ Step 1 of the diagnostic algorithm requires that a patient meet all of the core inclusion criteria: age 50 years or more, bilateral shoulder and/or bilateral pelvic girdle ache, morning stiffness lasting more than 45 minutes, duration of symptoms greater than 2 weeks, and an acute phase response in the form of an elevated erythrocyte sedimentation rate and/or C-reactive protein level.
▸ Step 2 involves evaluating the patient who has successfully passed step 1 for the core exclusion criteria, including active infection, neoplasia, and giant cell arteritis.
▸ Step 3 is a brief therapeutic trial of low-dose steroids; that is, 15-20 mg/day of prednisone. A positive result consistent with a diagnosis of PMR is a clinical response within 1 week marked by at least 70% global improvement, plus laboratory resolution of the acute phase response in 3-4 weeks.
The proposed algorithm calls for a series of lab tests to be done prior to giving low-dose steroids. These include dipstick urinalysis, a full blood count, rheumatoid factor, liver function tests, a bone profile, and measurements of thyroid-stimulating hormone, anti-cyclic citrullinated peptide antibody, and creatine kinase. Selectively ordering a chest x-ray in those patients with systemic symptoms is deemed appropriate, but use of chest computed tomography to screen for malignancies is not, Dr. Dasgupta said.