Palindromic rheumatism (PR) is an episodic arthropathy characterised by recurrent attacks of short-lasting articular and periarticular inflammation without residual joint involvement. Although definite knowledge on the disease entity of PR is lacking, the clinical and immunological similarities and the frequent progression to chronic arthritis may suggest a pathogenic relationship with rheumatoid arthritis (RA).1 2 In particular, anticitrullinated protein antibody (ACPA)-positive PR has been regarded as a prodromic phase of RA.3 Compared with autoantibody-positive arthralgia, a well-established model of preclinical RA,4 ACPA-positive PR shows a clearly detectable articular phase and might thus represent a very early self-remitting stage of the disease.

Typical attacks of PR may escape definite joint effusion at clinical examination, in spite of considerable periarticular and para-articular swelling, pain and loss of function.1 More accurate identification of the anatomical targets of PR might expand our knowledge on the mechanisms of disease initiation in very early RA before its full-blown clinical stage. We have undertaken a study to characterise the imaging correlates of acute arthritis in ACPA-positive PR.

Inclusion criteria were: (1) fulfilment of the diagnostic criteria for PR1; (2) an ongoing attack of arthritis of the hands; (3) a positive ACPA test; and (4) exclusion of other causes of arthritis/associated rheumatic diseases. Fifteen patients who agreed to undergo ultrasonographic (US) assessment of bilateral wrists and metacarpophalangeal (MCP) joints5 within 24 h of the onset of arthritis were enrolled in the study. Prompt access to further examination by MRI was possible in four patients whose clinical characteristics were comparable to the overall group. Diagnoses other than PR were carefully excluded by an experienced rheumatologist (CM). All the patients were women. The involved locations were wrists (8/15) and MCPs (7/15).

US assessment of symptomatic joints showed signs of synovitis in nine of the 15 patients with a median gray scale (GS) score of 1 (range 0–2). Six patients had a positive power Doppler (PD) signal with a median score of 1 (range 0–2). The clinical characteristics of patients with and without US synovitis are summarised in table 1. Despite the fact that no significant differences were found, the small number of patients limits the ability to draw definitive conclusions. In the four patients who underwent MRI examination, bone marrow oedema was seen in all four patients and three had mild synovitis. Paired US data confirmed GS synovitis in the same three patients (score of 1 in two and score of 2 in one), with a positive PD signal (score 1) in two patients. Figure 1 shows representative examples of bone marrow oedema with and without US synovitis.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Paired MRI and power Doppler ultrasonography (PD-US) during transient arthritis in anticitrullinated protein antibody (ACPA)-positive palindromic rheumatism. (A, C) Coronal fat-suppressed T2-weighted MRI sequences of the wrist of the right hand (A) and the second metacarpophalangeal (MCP) joint of the right hand (C) in two representative patients during an active episode of arthritis. High-grade bone marrow oedema appears as a hyperintense signal (white arrowheads) in the scaphoid (A) and the second MCP head (C). (B, D) Corresponding longitudinal sections of PD-US scans of the same joint areas (wrist of the right hand (B) and second MCP of the right hand (D)). No gray scale (GS) and no PD signals (score 0 for both) are seen in the wrist despite bone marrow oedema of the scaphoid (B). Mild GS synovitis (score 1) without significant PD (score 0) is instead visible at the second MCP head (D), again despite overwhelming bone marrow oedema of the corresponding joint area. MRI and US performed 2 months after resolution of the acute episodes of arthritis showed complete reversal of the synovial and bone marrow alterations (not shown). US evaluations were performed by a single experienced operator using a GE Logiq9 scanner with a multifrequency linear array transducer (8–15 MHz). MRI was performed with a high field 1.5 Tesla device (Achieve, Philips, Amsterdam, NL). Sequences included a T1-weighted gradient echo (not shown) and a T2-weighted turbo spin echo with suppression of fat.

Our findings confirm and extend previous data in unselected patients with PR,6 indicating that synovial pathology can be mild even when the analysis is restricted to ACPA-positive patients (ie, those at higher risk of developing RA). However, we cannot exclude the possibility that US assessment might actually underestimate subtle or rapidly remitting synovial inflammation. Indeed, the neutrophilic infiltrate described in sporadic cases7 is probably below the resolution of current imaging techniques. In contrast, all patients assessed by MRI had florid bone marrow oedema. As the synovial and subchondral compartments in RA can be affected by similar immune inflammatory reactions8 and may virtually communicate through microscopic bone canals,9 identification of the primary target of the disease remains a major challenge. Indeed, synovitis and bone marrow oedema equally characterise RA from its earliest phases and predict erosive progression.10 In this context, longitudinal MRI and high-resolution CT studies in ACPA-positive PR may help to define the dynamics of reciprocal cross-talks between different joint compartments.

In conclusion, our data, although preliminary, suggest that subchondral bone marrow involvement may occur very early in RA and that ACPA-positive PR may be a fascinating human model to be explored in more comprehensive multicentre studies.