Time in Remission as an Alternative Outcome Measure for Ankylosing Spondylitis: A 4-Year Prospective Study of 1900 Users of Anti-TNF

OBJECTIVES:

We have recently validated time in remission (TIR; https://www.medevio.cz/tir-calculator/) as novel outcome measure for rheumatoid arthritis patients. This study aimed to introduce TIR in ankylosing spondylitis (AS).

METHODS:

The ATTRA-AS registry cohort of AS patients treated with anti-TNF between 2012 and 2016 has been described previously. Point remission and sustained remission were defined as ASDAS<1.3 in one or both of two consecutive visits, respectively. TIR (0-100%) was interpolated between each two ASDAS values measured at two consecutive visits.

Following patients over time, Spearman correlation coefficients were calculated between CRP, BASFI, BASDAI, HAQ, SF-36 bodily pain, EQ-5D utility and WPAI work impairment (WI). Additionally, we used TIR, point remission and sustained remission to predict EQ-5D utility and WI in a training sample via mixed effect clustered linear regression. Adjusted R² and mean squared error (MSE) of the prediction were calculated in the test set (split 70/30).

RESULTS:

TIR was significantly correlated (p<0.001) with CRP (coefficient -0.528), BASFI (-0.566), BASDAI (-0.653), HAQ (-0.511), SF-36 bodily pain (0.563), EQ-5D utility (0.485) and WI (-0.565).

During the follow-up, TIR predicted EQ-5D utility in the test set (R²=0.17; MSE=0.056) better than sustained remission (R²=0.13; MSE=0.057) and somehow worse than point remission (R²=0.19; MSE=0.053). WI was also predicted by TIR more reliably (R²=0.27; MSE=0.033) compared to sustained remission (R²=0.20; MSE=0.035) and slightly less reliably than with point remission (R²=0.28; MSE=0.031).

CONCLUSIONS:

TIR estimated in AS patients during the 4 years of anti-TNF therapy was correlated with established measures of disease activity, disability, quality of life and WI. TIR is an independent predictor of the quality of life and the work impairment. In patients with AS, TIR can be used to describe their health state over time without losing information about achieving the treatment target compared to widely accepted outcome measures (point and sustained remission assessed in two consecutive visits).