Background/Purpose: A BASDAI ≥4 has often been required to start TNF inhibitors (TNFi) therapy in patients with axial SpA (axSpA). However, this cut-off of high disease activity (HDA) is largely arbitrary. Unlike BASDAI, the Ankylosing Spondylitis Disease Activity Score (ASDAS) incorporates objective measures (e.g. CRP) and has a validated definition of HDA (≥2.1). It has thus been suggested that ASDAS could also be used to guide treatment decisions, but evidence to support this is still scarce.

Our objective was to compare the impact of applying the ASDAS and BASDAI definitions of HDA in selecting patients for TNFi-treatment in daily clinical practice.

Methods: Patients from Reuma.pt (Rheumatic Diseases Portuguese Register), with diagnosis of axSpA according to their rheumatologists (both treated and not treated with their first TNFi), with complete baseline BASDAI and ASDAS data and complete 6-month of follow-up (i.e.  baseline, 3 and 6 months visits available) were included. Four subgroups [cross-tabulation between ASDAS (≥2.1) and BASDAI (≥4) definitions of HDA], were compared according to baseline demographic and clinical characteristics in the ‘eligible population’ (i.e. irrespective of TNFi-treatment). In addition, for patients starting TNFi and with complete follow-up BASDAI/ASDAS data (‘efficacy population’), the subgroups were also compared according to different response criteria (Table 1), at 3 and 6 months.

Results: In total, 466 patients were included (59% males and 66% HLA-B27 positive). The large majority (n=382; 82%) fulfilled the definition of HDA according to both BASDAI and ASDAS at baseline (i.e. ASDAS≥2.1 and BASDAI≥4). The frequency of ASDAS≥2.1, if BASDAI<4, was much higher than the opposite condition (i.e. ASDAS<2.1, if BASDAI≥4) (70% vs 0.5%). Compared to patients fulfilling both definitions, those who were ASDAS≥2.1 only were more likely to be male (82.5% vs 54%), HLA-B27 positive (79% vs 54%), to show higher levels of CRP (2.6 (SD 2.5) vs 2.2 (2.8) mg/dL) and lower BASFI (3.1 (2.6) vs 5.6 (2.3)). In the ‘efficacy population’ (n=296), better responses were observed among patients with ASDAS≥2.1 only, especially for the most ’stringent’ outcomes [e.g. ASDAS inactive disease (ASDAS ID): 59% and 50%, at 3 and 6 months, respectively], compared to patients fulfilling both definitions (ASDAS ID:  26% and 25% at 3 and 6 months, respectively) (Table 1).

Conclusion: Our results show that the ASDAS-HDA definition (ASDAS≥2.1) is more inclusive than the BASDAI-HDA definition (≥4) in selecting axSpA patients for TNFi treatment. Importantly, the additionally ‘captured’ patients respond better and have higher likelihood of predictors thereof. These results support the use of ASDAS≥2.1 as a selection criterion for treatment decisions.