• arthritis
• rheumatoid
• magnetic resonance imaging
• synovitis

With great interest we read the article of Gessl et al that studied whether tenderness of non-swollen small hand joints was explained by subclinical inflammation as detected by ultrasound (US). Interestingly, the findings in rheumatoid arthritis (RA) were dependent on the disease stage: tenderness in longstanding RA (mean disease duration 7.2 years) was not associated with subclinical inflammation, while in early RA (disease duration <2 years) tenderness of hand joints was associated with subclinical inflammation (54.5% of tender joints vs 48.4% of non-tender joints showed inflammation; OR 2.22).1 Since validation of research findings is of utmost importance, we hypothesised that the findings observed in early clinical RA would correspond to similar findings in the symptomatic phase that precedes clinical inflammatory arthritis and RA.2 It is important to understand whether joint tenderness in patients with clinically suspect arthralgia (CSA) is related to subclinical inflammation, since joint swelling is per definition absent in this phase.

This prompted us to perform a validation study in CSA patients. We studied the association of joint tenderness in metacarpophalangeal (MCP) joints with subclinical inflammation as determined by MRI. We focused on MCP joints since these joints are included in the EULAR definition of arthralgia suspicious for progression to RA.3 On the contrary to Gessl et al, who used US examination for determining the presence of subclinical inflammation, we used MRI. According to the EULAR imaging taskforce, both MRI and US have an added value when diagnosing patients in the early disease phases of RA.4 US has advantages in terms of costs and availability. MRI in contrast is more sensitive; especially for tenosynovitis, the sensitivity of US compared with MRI is moderate-low.5 To prevent false negative findings due to missing subtle subclinical inflammation in the phase of CSA, MRI was implemented in our study.

Between April 2012 and February 2019, 602 patients were consecutively included in the Leiden CSA cohort and assessed for MCP tenderness by physical examination at baseline. Patient-reported joint pain was also assessed by using a mannequin. Unilateral MRI of the MCP (2–5) joints was made and scored for subclinical inflammation (synovitis, tenosynovitis, osteitis) by two experienced readers, blinded for clinical data, as described previously and in the online supplemental file.2 Presence of subclinical inflammation was determined after referring to MRI findings of MCP joints of symptom-free controls of the same age.6 Data of tenderness and inflammation in ipsilateral MCP joints were studied. Analyses were done in all patients presenting with CSA and in the subgroup that progressed to inflammatory arthritis within 2 years follow-up (in retrospect these patients were truly ‘pre-RA’ when presenting with arthralgia). A flowchart of the studied population and baseline characteristics are presented online supplemental figure 1 and table 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

(A) Association of subclinical inflammation with tender MCP joints (N=226 with tenderness vs N=376 without tenderness): OR 1.84, 95% CI 1.29 to 2.63. (B) Association of subclinical inflammation with tender MCP joints in patients who developed inflammatory arthritis (N=33 with tenderness vs N=54 without tenderness): OR 1.79, 95% CI 0.74 to 4.30.

In all CSA patients, 38% of patients with tender MCP joints had subclinical inflammation compared with 25% of patients with non-tender MCP joints (OR 1.84 (95% CI: 1.29 to 2.63); figure 1A, online supplemental table 2A). At patient level, MCP tenderness was associated with subclinical synovitis and tenosynovitis (OR 1.76 (95% CI: 1.10 to 2.81), OR 1.69 (95% CI: 1.12 to 2.55), online supplemental table 2A) but not with osteitis. In the subgroup of patients who progressed to clinical arthritis (n=87), 61% of patients with tender MCP joints had subclinical inflammation compared with 46% of patients with non-tender joints (OR subclinical inflammation 1.79 (95% CI: 0.74 to 4.30), figure 1B, online supplemental table 2A). Studying the 39% of patients who presented with MCP tenderness without subclinical inflammation at MRI and who later on developed IA revealed that clinical arthritis had mostly (85%) developed in joints that were not scanned, hence being left unanswered whether these patients had subclinical inflammation at presentation with CSA in joints that were not imaged. The remaining patients (15%) did develop clinical arthritis in MCP joints that were imaged but were normal at first presentation, suggesting that subclinical inflammation developed after CSA onset. Evaluating patient-reported MCP pain (mannequin) instead of MCP joint tenderness at physical examination showed similar associations with subclinical inflammation (online supplemental figure 2, table 2B).

Concluding, associations of joint tenderness with subclinical inflammation in the symptomatic phase preceding clinical arthritis had roughly similar effect sizes as observed in the phase of clinical arthritis. Hereby, we validated and expanded the findings of Gessl et al in early RA.