The EULAR 2022 Hybrid Congress provided an engaging and interactive forum for the presentation and discussion of the latest advancements and trends in the field of rheumatology. Four leading rheumatologists gave us their expert opinions on the key data presented at this year’s congress.

The views of the healthcare professionals expressed in this piece are their own and do not necessarily represent those of Pfizer.

In the EULAR 2022 Congress, gender has been my favourite research topic with its increasing importance in rheumatology. I would like to highlight three abstracts based on their impact on clinical practice.
 

Of late, there has been growing interest in interstitial lung disease (ILD) as a relevant comorbidity in rheumatoid arthritis (RA). A group of French researchers investigated the main sources of exposure to crystalline silica in patients with RA.¹ Crystalline silica is ubiquitous in the environment
so its exposure can occur in both men and women, in professional or occupational context and beyond.¹

The Dust Exposure Life-Course Questionnaire was used to longitudinally quantify occupational and non-occupational lifetime exposure to crystalline silica.¹ In women with RA, the main sources of exposure were cleaning activities, dusty clothes laundry and talcum powder handling, with significantly higher exposure levels from these sources compared with the general population (p<0.005).¹ Across the cohort, high silica exposure was found to be independently associated with ILD (odds ratio [OR] 6.5) and mediastinal lymphadenopathy (OR 6.3).¹

In an oral presentation, Ulas and colleagues published new perspectives in axial spondyloarthritis (axSpA) by assessing different magnetic resonance imaging criteria in axSpA separately for men and women.² They showed that the diagnostic performance of established image markers was significantly lower in female patients, underlining the need for gender-specific recommendations in imaging criteria.²

The third study conducted in Sweden reported that ankylosing spondylitis-related structural changes in the spine independently influenced patient-reported physical function in men but not in women.³
 

In the management of RA, methotrexate (MTX) is the first-line treatment but 30–40% of patients with RA do not respond to MTX.⁴ Clinicians often use tumour necrosis factor inhibitors (TNFis) in these patients, yet a similar proportion do not respond adequately to TNFis.⁵ As a result, the initiation of appropriate treatment is delayed.⁵ Two particular abstracts caught my attention, as their work may help us to predict patient response before prescription of such therapies.^4,5

A research group from France developed machine learning models based exclusively on parameters collected in routine clinical practice to predict patient response to MTX and TNFis.^4,5 Data from patient cohorts were utilised to train a machine learning model, which was in turn validated by a different population.^4,5 For both studies, two strategies were tested, prioritising the identification of either responders or non-responders.^4,5 In the study analysing response to TNFi, the positive predictive value (PPV; for TNFi responders) and negative predictive value (NPV; for TNFi non-responders) were 76% and 67%, respectively.⁵ The models used in the study analysing response to MTX also performed well, with a PPV (for MTX responders) and NPV (for MTX non-responders) of 95% and 68%, respectively.⁴ Of note, Disease Activity Score in 28 Joints was one of the parameters that was automatically chosen as predictors in both studies.^4,5

While these machine learning models will not predict all patient responses, this is a significant step forward in identifying which parameters to collect routinely from our patients to improve care. I look forward to reading the complete publication to learn the minimal dataset required to predict response. The aim will ultimately be to provide a treatment strategy that is personalised to each patient.
 

Significant progress in our understanding of RA has been made in recent years. Two RA studies at the EULAR 2022 Congress particularly stood out for me.

The first study by Qiao and colleagues potentially paves the way for the assessment of faecal microbiota as part of a personalised medicine approach in RA management.⁶ They analysed the intestinal microbiota of 145 Chinese patients with RA and identified two predominant patterns of dysbiosis in RA (i.e. RA-enterotypes).⁶ These two enterotypes have distinct bacterial distributions, primarily dominated by either Prevotella or Bacteroides.⁶ Notably, they were associated with different levels of CD8+ T cells.⁶

The stratification of microbiota variations as enterotypes instead of a continuous distribution suggests a limited number of stable host-microbial symbiotic states. Longitudinal studies of the intestinal microbiome, preferably in early untreated RA, would be required to confirm the weak correlations with selected clinical biomarkers. Of clinical relevance would be whether RA-enterotypes respond differently to antirheumatic drugs or differ in their long-term prognosis.

The second study conducted in Germany explored the effects of fasting and plant-based diet on RA.⁷ This nutritional trial randomised 50 patients with RA to either a 1-week fast followed by 11 weeks of a plant-based diet, or 12 weeks of conventional nutritional counselling.⁷ After 12 weeks, there was no significant difference in the primary outcome between both groups, even though post-hoc analyses suggested a beneficial trend on disease activity at 1 week (EULAR good or moderate response: 52% in the fasting group and 28% in the counselled group).⁷ Clinically relevant improvement in Health Assessment Questionnaire (HAQ) scores was reported in the fasting group after the initial fasting period too.⁷

A larger follow-up study would be needed to confirm these exploratory findings, although these findings indicate fasting may have a beneficial effect on inflammatory markers and patient-reported outcomes.⁷
 

Patients with psoriatic disease (PsD), including psoriatic arthritis and psoriasis, have an increased risk for cardiovascular (CV) disease, probably related to the interaction between traditional CV risk factors, inflammation and pharmacological treatments.⁸ I therefore believe it is important to further understand how this interaction can increase CV risk.

Colaco and colleagues enrolled a longitudinal cohort of 1,376 patients with PsD, without previous history of CV disease, and followed them between 1978 and 2020.⁸ A total of 10 traditional CV risk factors and 15 PsD-related risk factors were assessed.⁸ During the 10 years of follow-up, 102 (7.4%) patients developed incident CV events.⁸ In Cox regression models adjusted for age and sex, traditional CV risk factors such as age, diabetes, systolic blood pressure, body mass index, triglycerides, antihypertensive drugs and lipid-lowering agents were associated with increased CV risk.⁸ PsD-related risk factors including Psoriasis Area and Severity Index score, erythrocyte sedimentation rate, number of swollen joints, HAQ score, and daily use of non-steroidal anti-inflammatory drugs were also associated with increased CV risk; the use of biologic drugs was not found to be significantly protective against CV events.⁸ However, during the oral presentation for OP0030, it was reported that in a Cox regression model adjusted for traditional CV risk factors the use of biologic drugs was significantly protective against CV events.⁹

In this study, six PsD-related risk factors were identified to be significantly associated with incident CV events.⁸ As the current traditional CV risk scores underperform in predicting CV disease in patients with PsD, I am hopeful that these newly identified PsD-related risk factors may be useful for the future development of a PsD-specific CV disease risk prediction score.