Objectives: To assess the role of short-term response to first anti-TNF in long-term prediction of disability.Methods: In nationwide registry ATTRA, we identified ankylosing spondylitis patients starting anti-TNF between 01/2003 and 12/2016. Full disability and work impairment (WI; WPAI questionnaire) were predicted via the Cox- and lagged-parameter mixed-effect regression.Results: 2,274 biologicals-naïve patients newly indicated to anti-TNF were prospectively followed (6,333 patient-years; median follow-up 1.9 years). Reaching BASDAI < 4 (77.4%) and ASDAS-CRP < 2.1 (61.1%) after 3 months of anti-TNF both decreased the risk of future disability by ≈2.5-fold. ASDAS-CRP < 2.1 predicted non-disability better than BASDAI < 4 & CRP < 5 mg/L (p = 0.032). BASDAI < 4 & CRP < 5 mg/L was comparable to BASDAI < 4 (p = 0.941) and to BASDAI change by >50% or by >2 points (p = 0.902). ASDAS-CRP change >1.1 and >2.0 both failed to predict non-disability. Once on anti-TNF therapy, the strongest predictor of WI was Pain (SF36). Yearly increase in indirect costs remains below €3,000 in those reaching ASDAS-CRP < 2.1.Conclusions: Low disease activity measured by ASDAS-CRP ≤ 2.1 should be used to measure the outcome of new anti-TNF therapy. Continuous WI could be decreased through pain management.
Keywords: ASDAS; Ankylosing spondylitis; BASDAI; anti-TNF; disability; multivariate modeling; predictive power; work impairment.