Effusion attenuates the effect of synovitis on radiographic progression in patients with hand osteoarthritis: a longitudinal magnetic resonance imaging study

Where OA was formerly considered a purely mechanical disorder, nowadays the consensus is that in OA also low-grade inflammation is implicated [1]. Imaging studies with ultrasound and magnetic resonance (MR) have shown that synovitis and bone marrow lesions (BMLs) play a role in the pathogenesis of hand osteoarthritis (OA), since these inflammatory features are associated with pain and radiographic progression [2‐6]. However, their precise role remains unclear since treatments with anti-inflammatory drugs such as biologicals and corticosteroids have largely been unsuccessful. Whether effusion, another inflammatory feature which is often seen with synovitis, is implicated, is unclear.

What causes joint effusion, i.e., excess fluid in the joint, is not understood. Osmotic pressure in the tissue versus the joint cavity, decreased drainage, and pressure of surrounding tissue seem of importance [7, 8]. Joint effusion is often seen with inflammatory reactions such as in erosive (inflammatory) hand OA [9, 10] but also in OA joints without warmth or tenderness or in clinically healthy joints [11]. These suggest different triggers for effusion, such as mechanics or inflammation.

A MR study that compared inflammatory features in finger joints between healthy volunteers and patients with rheumatoid or psoriatic arthritis suggested that joint effusion most likely reflects pathology [12]. In hand OA, the presence and role of effusion are rarely studied. Few studies scored effusion, likely since it is not included in the scoring methods as a separate feature [13, 14]. A study in finger OA that used both ultrasound and MR and an ultrasound study in hand OA suggested that assessment of effusion as a separate feature is valid and reliable and that effusion occurs frequently [2, 3, 15].

The aim of our exploratory study was to study the role of MRI-defined joint effusion in patients with hand OA. We defined effusion using 1.5-tesla MR images from our secondary care cohort and we studied the association of effusion, and effusion in combination with synovitis, with joint pain and radiographic progression. We hypothesized that the presence of effusion reflects inflammation and therefore reinforces the known association between synovitis and joint pain and progression.

We showed that MRI-defined effusion is present in interphalangeal joints of patients with hand OA, in joints with EST (enhanced synovial thickening), but remarkably also in joints without EST. Effusion on itself was not associated with pain or progression, although a tendency towards a protective effect was seen. Contrary to our hypothesis, the known association between synovitis (here: EST) and progression was attenuated by the presence of effusion.

Different mechanisms could explain our results. First, effusion could decrease mechanical stress, leading in the long term to less radiographic damage. Second, the fluid could contain substances that aid repair or could be anti-inflammatory. Third, the presence and absence of effusion could reflect different stages of the inflammatory disease process.

We assessed MR effusion apart from and in combination with EST. An ultrasound study that assessed effusion in patients with hand OA reported that it was associated with pain and radiographic progression, which is contrary to our results [2, 3]. Further differences are the imaging techniques and the hand joints studied. However, the ultrasound study did not investigate the role of effusion apart from synovitis.

Our study has its limitations. First, EST is not the same as synovitis. To assess synovitis, histology should be used, which is invasive and not feasible. We think that EST is the best available biomarker/proxy to assess synovial inflammation. Second, in small hand joints, synovitis and effusion might be difficult to distinguish. Therefore, we set predefined definitions and scored in a conservative manner. From our experience in this study, and supported by previous studies [2, 3, 12, 15], we feel that a distinction between synovitis and effusion is possible even in small finger joints. Third, we did not assess construct validity by comparing MR effusion with another measure such as ultrasound. However, the validity of MR effusion is supported by an earlier study assessing the construct validity of ultrasound effusion using MR imaging as a reference [15]. Nevertheless, validity and (interreader) reliability should be further confirmed before our definition of MR effusion could be used in hand OA MR scoring methods. Fourth, we did not ascertain effusion during follow-up while it is possible that effusion is a transient phenomenon [3, 18]. However, several studies have shown that baseline inflammatory features such as synovitis and BMLs are of clinical relevance for their association with pain and structural progression [3, 5, 6, 19], regardless of their possible transient nature [20]. We think the same applies to effusion and therefore, our results are relevant.

Up until now, effusion is not a feature in scoring methods for hand OA [13, 14]. Therefore, we made definitions in collaboration with experienced radiologists, and consensus scoring was performed with two readers. Therefore, we think our results are valid. We envision including effusion in the MRI scoring system; our scoring method could serve as a blueprint for this. From an earlier MRI study, we learned that increase/decrease in synovitis was associated with increase/decrease in pain [20]. We think that especially this last finding is interesting because the association of a decrease in synovitis with a decrease in joint pain makes synovitis a treatment target. In the present study, we showed that the presence of effusion has an attenuating effect on the effect of synovitis. Therefore, it would be interesting to investigate whether a change in effusion would be associated with a change in pain or a lower chance of radiographic progression over time and how this is related to the effect of synovitis. If future studies can show such effects, we expect that the inclusion of effusion will increase the responsiveness of the MRI scoring system. In conclusion, we found that effusion can be assessed as a separate feature on MR images and is present in interphalangeal joints apart from synovitis. Our study indicates that effusion seems to be of importance for its attenuating effect on the association between synovitis and radiographic progression. Our results warrant further study into the presence and role of effusion apart from synovitis in hand OA, to further elucidate the role of inflammation, and in particular effusion, in hand OA pathology and its treatment.