Background Sustained remission is an important goal to of therapy in rheumatoid arthritis (RA). In daily clinical practice, RA patients in remission are assessed by various clinical outcomes.
Objectives The purpose of this study is to identify clinical prognostic factors for sustained remission and build a predictive model from a large cohort database.
Methods Data from RA patients registered in a nationwide Japanese cohort database (NinJa : National Database of Rheumatic Diseases by iR-net in Japan) in 2009 and 2010 were used to evaluate baseline characteristics, treatment profiles, and the following clinical outcomes: tender joint count (TJC), swollen joint count (SJC), C-reactive protein (CRP), ESR, Visual Analog Scale (VAS), pain VAS, patient VAS, physician VAS, and Modified Health Assessment Questionnaire (MHAQ). All patients with SDAI remission in 2009 were categorized based on whether remission was maintained in 2010. A multivariate Cox regression model was constructed for predictive factors and a prognostic prediction model was constructed.
Results We enrolled 4215 patients in 2009 of whom 930 patients had SDAI remission. After 1 year, 623 patients (67.0%) had sustained remission. Compared to non-sustained remission group, disease duration, presence of erosion, SJC, CRP, ESR, VAS, pain VAS, physician VAS, and MHAQ were significantly lower in the sustained remission group (p<0.05). A prognostic prediction model with a total score of 4 points for remission failure after 1 year was constructed as follows: 2 points for MHAQ >1, and 1 point each for presence of erosion and SJC >1. For a total scores of 0, 1, 2, 3, and 4 points, remission failure rates after 1 year are 17.2%, 25.8%, 37.7%, 50.8%, 60.0%, respectively. Area under the receiver operating characteristic (ROC) curve for this model was 66.0% (95% CI: 61.5–70.5%).
Conclusions We found that presence of erosion, SJC, and MHAQ were significantly less in those with sustained remission after 1 year. A prediction model was successfully built and validated using clinically relevant parameters.
Disclosure of Interest None Declared