Background New provisional classification criteria for polymyalgia rheumatica (PMR) were published in 2012, and those criteria are the first to contain musculoskeletal ultrasound (US) in an additional algorithm. However, several studies validating the diagnostic ability of the criteria suggested that differentiation between PMR and elderly-onset rheumatoid arthritis with PMR-like onset (polymyalgic-EORA) is difficult even though US criteria were included. Through the clinical experience, we noted that there seems to be differences not in the kind or the frequency but in the severity of synovial inflammation between the two diseases.
Objectives To evaluate the discriminating ability of US scoring of shoulder synovitis and its surrogate maker to distinguish between PMR and polymyalgic-EORA.
Methods We analyzed consecutive records of 15 PMR patients and 15 polymyalgic-EORA patients. All of them were underwent US examination before treatment-start. The severity of tenosynovitis of the long-head of the biceps, bursitis of shoulder including subdeltoid, subacrominal and subcoracoid bursitis, and joint synovitis of grenohumeral joint were subjectively scored for GS and PD on a four-point scale: 0 = absent, 1 = mild, 2 = moderate or 3 = severe. Sum of the all scores of both shoulders was defined as “patient-shoulder synovitis score (PSSS)”, and the correlation between PSSS and serum markers was assessed. Indices contributing to discriminating between PMR and polymyalgic-EORA were explored.
Results PSSS in PMR tended to be lower than that in polymyalgic-EORA (7.73±3.77 vs 12.3±8.57, p=0.072 [mean ± SD]). PSSS were positively correlated with serum MMP3 (|R|=0.707, p<0.0001). Both PSSS (|R|=0.602, p=0.0175) and MMP3 (|R|=0.463, p=0.095) were positively correlated with serum CRP in polymyalgic-EORA but not in PMR. The ratio of PSSS to CRP (mg/dL) (PSSS/CRP) was significantly lower in PMR than in polymyalgic-EORA (1.35±1.63 vs 5.55±4.46, p=0.003 [mean ± SD]). The ratio of MMP3 (ng/mL) x104 to CRP (MMP3/CRP) was significantly lower in PMR than in polymyalgic-EORA (32.2±35.9 vs 101±87.1, p=0.006 [mean ± SD]). Sensitivity and specificity of 2012 PMR criteria scoring algorithm without US for PMR diagnosis in our patients were 93.3% and 40.0%, respectively. ROC analysis for PSSS/CRP and MMP3/CRP demonstrated areas under the curve (AUCs) of 0.902 and 0.852, respectively. By adding 1 point to the points from 2012 PMR criteria scoring algorithm without US in cases where the ratio is lower than the threshold, the optimal cut point became 5. Adding the criterion of either PSSS/CRP or MMP3/CRP improved the specificity of the 2012 criteria in our patients (80.0% and 71.4%, respectively) without much influence on the sensitivity (86.7% and 86.7%, respectively).
Conclusions Sonographic semi-quantitative assessment of shoulder synovitis provided useful information for discriminating PMR from polymyalgic-EORA. Serum MMP3 can be a surrogate maker for US-shoulder synovitis scoring.
Disclosure of Interest None declared