Background Rheumatoid Arthritis (RA) patients in the maintenance phase should be treated by escalating antirheumatics up to stable low disease activity and then often are treated with tumor necrosis factor inhibitors (TNFi) and conventional synthetic disease modifyers (csDMARD). Once this phase has been reached the risk of flare is an issue once drugs are to be tapered or discontinued.
Objectives To examine the ability of the multi-biomarker disease activity (MBDA) score as predictor for disease flare in rheumatoid arthritis (RA) patients with stable low disease activity using tumor necrosis factor inhibitors (TNFi) and conventional synthetic disease modifyers (csDMARD).
Methods Data were used from the continuation control group of the Dutch multicenter, open-label, POET trial, in which patients with stable low disease activity (disease activity score [DAS28] <3.2 for at least 6 months] were randomized to either stop or continue TNFi treatment. Three indicators of disease relapse were assessed: 1) flare based on DAS28 (DAS28 ≥3.2 with ΔDAS28 >0.6), 2) flare based on escalation of any DMARD therapy, and 3) flare based on physician-reported disease activity. Associations between baseline MBDA score and meeting the different criteria for disease flare within 6 or 12 months of follow-up were examined using univariate analysis and multivariate logistic regression adjusting for baseline DAS28 score.
Results For this post-hoc analysis, baseline serum samples to measure MBDA scores were available for 225/286 (78.7%) of the patients randomized to the TNFi continuation group (88.9% also used methotrexate, another 8.0% used another csDMARD and 3.1% used no csDMARD): 86.2% with a first TNFi, 11.6% with second TNFi and 2.2% with a third TNFi. Within 12 months, 19.1% of patients had experienced a DAS28 flare, 12.0% had medication escalation and 8.0% had ≥1 physician-reported flare. Median time to DAS28-based flare was 26 weeks (IQR:13–28). Univariately, patients with high baseline MBDA (>44) scores (n=31) were at increased risk of experiencing a DAS28 flare within 6 (OR =4.39, P=0.001) or 12 (OR =2.78, P=0.015) months. MBDA scores were not associated with increased risk of medication escalation or physician-reported flare. After adjustment for baseline DAS28 scores, high MBDA score remained predictive for risk of flare within 6 months (OR =3.15, P=0.017), but not for flare within 12 months (OR =2.05, P=0.107).
Conclusions MBDA score may be of additional value in predicting DAS28 flares but not in predicting medication escalations or physician-reported flares in RA patients on TNFi in stable low disease activity.
Acknowledgements We wish to acknowledge all POET investigators and all who gave their contributions to the POET project.
Disclosure of Interest P. M. Klooster: None declared, M. Ghiti-Moghadam: None declared, F. Lamers-Karnebeek: None declared, H. Vonkeman: None declared, E. Sasso Shareholder of: Myriad Genetics, Employee of: Crescendo Bioscience, P. Riel: None declared, M. Laar: None declared, T. L. Jansen: None declared