Time in remission as an alternative outcome measure for rheumatoid arthritis: a 10-year prospective study of 2618 new users of anti-TNF.

Abstract

Achieving targeted disease activity (DA) is the primary therapeutic strategy in RA. Point measurements of DA are done at out-patient visits, however true DA between visits remains unobserved. This study sought to describe and validate a new outcome measure, i.e. time in remission (TIR). Patients were enrolled in the Czech ATTRA-RA registry. TIR was calculated using linear interpolation of the DAS28-ESR determined at outpatient visits. Correlation coefficients were computed between TIR and DAS28-CRP, HAQ, Simple Disease Activity Index (SDAI), patient global assessment (PGA) and physician global assessment (PhGA). Using logistic regression, TIR was used as a predictor of remission (SDAI ≤3.3) and non-disability (HAQ <0.5). The predictive value of TIR was compared with point and sustained remission using the cross-validated area under receiver-operating curves. Since 2010, 2618 RA patients started anti-TNF therapy and were followed until 2020 or until treatment discontinuation. During the first 6 months of therapy, 56% of patients had no remission (TIR = 0), and 22% of patients reached sustained remission (TIR = 1), while 22% of patients had point remissions with 0 < TIR < 1. EULAR good responders and moderate/non-responders spent 64 ± 42% and 6 ± 18% of time in remission, respectively. The mean TIR grew during the follow-up and was correlated with DAS28-CRP, SDAI, HAQ, PGA, and PhGA (P < 0.0001). TIR at 3 and 6 months predicted remission (SDAI ≤3.3) and non-disability (HAQ <0.5) at 13 and 19 months better than point or sustained remission. TIR is an intuitive way of estimating unobserved DA between scheduled visits; its calculation only requires two consecutive DA values (https://www.medevio.cz/tir-calculator/). TIR is a valid predictor of RA outcomes.