Background The gold standard for diagnosis of calcium pyrophosphate (CPP) deposition (CPPD) is the identification of CPP crystals in synovial fluid. However, aspiration of synovial fluid can be challenging in small joints such as the wrist, a usual location of arthritis in CPPD. Despite its low sensitivity, the most widely used imaging modality is conventional radiography but ultrasound (US) seems a useful tool for diagnosis of CPPD.
Objectives We aimed to compare the performance of US and conventional radiography of the wrist for diagnosis of CPPD.
Methods Patients with joint effusion (knee, hip, shoulder, ankle or wrist) were consecutively included. CPPD was diagnosed by CPP crystals identified in synovial fluid. Patients without CPP crystals in synovial fluid were controls. As recommended, we used the term chondrocalcinosis (CC) to assess imaging features suggesting CPPD. Two blinded operators assessed CC in all patients by US and conventional radiography of the wrist. The presence of CC in triangular fibrocartilage (TFC) and wrist hyaline cartilage in US and TFC and radiocarpal (RC) joint in radiography was noted. A patient was considered to have CC if at least one wrist had imaging features of CC.
Results We included 58 patients with joint effusion (32 with CPPD). The remaining 26 patients, controls, had rheumatoid arthritis (n=13), spondyloarthritis (n=6), gout (n=6), and osteoarthritis (n=1). The mean age was 67.1±16.3 years. Location of joint effusion was as follows: 34 knees, 15 wrists, 3 shoulders, 2 ankles, 3 hips and 1 elbow. Among CPPD patients, US revealed CC in 30 (93.7%) and radiography in 17 (53.1%) (p<0.001). The sensitivity (Se) and specificity (Sp) of US for the diagnosis of CPPD were 94% and 85%, respectively; the positive likelihood ratio was 6.1. When analyzing US features of CC separately, US Se was higher at the TFC than RC joint, 81% and 50% respectively. The Se and Sp of radiography were 53.1% and 100%, respectively. Intraobserver reliabilities for US and radiographic CC were almost perfect: κ coefficient 0.832 [95% confidence interval 0.651–1.0] and 0.880 [0.314–0.880], respectively. In all 58 patients, 113 joints were analyzed (3 patients had radiography of only one wrist). The κ coefficient between US and radiography for CC was moderate: 0.33 [0.171–0.408].
Conclusions Our study suggests that wrist US should be considered a relevant tool for the diagnosis of CPPD, with higher sensitivity than radiography.
Disclosure of Interest None declared