Introduction Flares may be used as outcomes in axial spondyloarthritis (axSpA) trials or observational studies. The objective was to develop a definition for ‘flare’ (or worsening) in axSpA, based on validated composite indices, to be used in the context of clinical trial design. Methods (1) Systematic literature review of definitions of ‘flare’ in published randomised controlled trials in axSpA. (2) Vignette exercise: 140 scenarios were constructed for a typical patient with axSpA seen at two consecutive visits. Each scenario included a change in one of the following outcomes: pain, Bath Ankylosing Spondylitis Disease Activity Index (BASDAI), BASDAI plus C-reactive protein (CRP) or Ankylosing Spondylitis Disease Activity Score (ASDAS)-CRP. Each Assessment of Spondyloarthritis (ASAS) expert determined if every scenario from a random sample of 46 scenarios was considered a flare (yes/no). Receiver-operating characteristic (ROC) analyses were applied to derive optimal cut-off values. (3) ASAS consensus was reached. Results (1) The literature review yielded 38 studies using some definition of ‘flare’, with 27 different definitions indicating important heterogeneity. The most frequent definitions were based on BASDAI changes or pain changes. (2) 121 ASAS experts completed 4999 flare assessments. The areas under the ROC curves were high (range: 0.88–0.89). Preliminary cut-offs for pain (N=3), BASDAI (N=5) and ASDAS-CRP (N=4) were chosen, with a range of sensitivity 0.60–0.99 and range of specificity 0.40–0.94 against the expert's opinions. Conclusions This data-driven ASAS consensus process has led to 12 preliminary draft definitions of ‘flare’ in axSpA, based on widely used indices. These preliminary definitions will need validation in real patient data.