All the members (n=152) of GRESSER (GRupo de Estudio de ESpondiloartritis de la Sociedad Española de Reumatología), the Spanish working group with common interests in Spondyloarthritis, were invited to participate into the study.

The scientific committee generated 80 questions to be answered after an exhaustive literature review. A PubMed search of articles in English or Spanish since 01-01-2008 was performed. The following terms were included: “Spondylitis, Ankylosing”[Mesh] OR axial spondyloarthritis OR non-radiographic axial spondyloarthritis AND “Remission Induction”[Mesh] OR Remission OR “Recovery of Function”[Mesh] OR Minimal Disease Activity OR MDA OR Very Low Disease Activity OR VLDA OR Bath Ankylosing Spondylitis Disease Activity Index OR BASDAI OR Ankylosing Spondylitis Disease Activity Score OR ASDAS OR ASDAS-CRP OR Modified Stoke Ankylosing Spondylitis Spinal Score OR mSASSS OR Stoke Ankylosing Spondylitis Spine Score OR SASSS OR Spondyloarthritis Research Consortium of Canada OR SPARCC OR Ankylosing Spondylitis spine MRI score for activity OR ASspiMRI-a OR Berlin method OR treat to target. The selected articles were those that showed the validity of the indices, the associations among the different indices or of the indices with imaging tests or biomarkers, and the relationships with the disease prognosis or the response to treatment.

Also, clinical practice guidelines, systematic or narrative reviews of the last 5 years were sought in The Cochrane Library, US National Guidelines Clearinghouse, Tripdatabasem and the Biblioteca de Guías de Práctica Clínica del Sistema Nacional de Salud (GuiaSalud). The web pages of the Spanish Society of Rheumatology (SER), European League Against Rheumatism (EULAR), American College of Rheumatology (ACR) and The National Institute for Health and Care Excellence (NICE) were also reviewed in search of clinical practice guidelines.

A modified Delphi method was used in this study.5 According to this method, the opinions of a panel of experts were anonymously requested using an online questionnaire in two rounds of voting (in September and October 2018, respectively). The questionnaire consisted of assertions or items that addressed different aspects about axSpA and were scored according to the degree of the panellists’ agreement or disagreement with them. The results obtained after the first round were analysed and reviewed by the scientific committee. Then, these results were sent to the panellists and the items on which there was no consensus were subjected to a second round of voting. In this manner, the experts could reconsider their responses in light of the pooled results. The results obtained in this second round were analysed to determine which issues had finally achieved an adequate degree of consensus among the experts, and whether they were in agreement or in disagreement with each item presented.

The panellists assessed the items with a nine-point ordinal scale (1=full disagreement, 9=full agreement). Responses were grouped into three categories: 1–3=disagree; 4–6=neither agree nor disagree; and 7–9=agree. The median and interquartile range (IQR) of the responses were calculated. Consensus was reached if the median of the responses was over 7 or below 3 and less than one-third of the panellists voted outside these ranges. Also, the IQR should be less than 4. Results are shown in the tables as number of votes, median [IQR] of the answers, the degree of agreement among the panellists and the result regarding agreement or disagreement for each item. The degree of agreement indicates the percentage of panellists who voted within the median range.

The panellists agreed that currently there is not a consensus definition of remission in axSpA. However, they considered that it is possible to achieve remission in this entity. Remission in a chronic inflammatory disease, such as axSpA, does not correspond to complete healing or a state of symptoms acceptable to the patient. Instead, remission is a state in which the disease has little or no activity and during which there is no progression of the disease.6

Clinical remission/inactive disease is defined by the absence of clinical and laboratory evidence of significant inflammatory activity and the treatment target should be clinical remission/inactive disease of musculoskeletal (arthritis, dactylitis, enthesitis, axial disease) and extra-articular manifestations, as proposed by the recently updated treat to target recommendations.3 However, it remains unclear how to precisely assess this in clinical practice and different proposals have been made. In 2001 the Assessment of Spondyloarthritis International Society (ASAS) suggested a preliminary definition of partial remission as a value no greater than 2 on 0–10 visual analog scales (VAS) for the following four domains: patient global assessment, spinal pain, physical function (measured through the Bath Ankylosing Spondylitis Functional Index – BASFI) and inflammation (Bath Ankylosing Spondylitis Disease Activity Index – BASDAI – items 5 and 6).7 However, this definition includes a functional parameter (BASFI) and recovery of normal function becomes more difficult with increasing disease duration, as structural damage accumulates over time. In 2016, the ASAS/EULAR (European League Against Rheumatism) recommendations for axSpA described inactive disease as a “clinical remission-like definition” based on an ASDAS (Ankylosing Spondylitis Disease Activity Score) <1.3.8 In the 2017 updated consensus of the Spanish Society of Rheumatology (SER) for the treatment of axSpA, ASDAS-CRP was recommended as the main index to monitor disease activity, with the ASDAS-CRP <1.3 cut-off as the best available way to define remission.9 It has been established that an ASDAS-CRP <2.1, which is equivalent to low disease activity (LDA), may be considered acceptable. Therefore, ASDAS-CRP is recommended as the main index to monitor the disease activity, establishing thus the value of ASDAS-CRP <1.3 as the best available way to define remission. Also, when defining whether a patient with axSpA has reached remission or LDA, in addition to these indices, it is recommended to consider physician global assessment.9 The 2018 French10 and the 2016 Portuguese guidelines11 also recommended ASDAS <1.3 as the equivalent of remission. The 2017 guidelines of the UK's National Institute for Health and Clinical Excellence (NICE)12 and the 2015 American College of Rheumatology (ACR)13 did not provide a definition of remission in axSpA.

However, the state of ‘inactive disease’ does not always correspond to the remission of the disease. For example, a patient with absence of joint symptoms who presents with recurrent flares of uveitis should not be considered in remission6 and attention must be paid to extra-articular manifestations.14 As mentioned before, inactivity of the extra-articular manifestations is also highlighted as a treatment target.3

Regarding disease activity evaluation, different composite indices have been used, but currently there is no consensus on the optimal index. The BASDAI has been the most widely used in clinical trials and clinical practice. Although a definition of clinical remission or LDA has not been validated using this index,3 some clinical trials and observational studies have used different cut-off points (BASDAI <4, <3 or ≤3).14 BASDAI has the disadvantage of being a solely patient-reported outcome (PRO), so does not include objective parameters. The ASDAS index,15 in addition to including some questions from the BASDAI, offers a more objective assessment of disease activity by including C reactive protein (CRP), a marker of inflammation. Alternatively, it can be calculated using erythrocyte sedimentation rate (ESR), when CRP is not available. Cut-off points have been established and define disease activity states which have been validated in both clinical practice and clinical trials patients.16 As mentioned before, it is increasingly recommended to use the ASDAS for disease activity assessment in axSpA.

Although it is likely that rheumatologists have a good level of knowledge of the disease activity indices, this is often not reflected in their evaluations during the patient follow-up in clinical practice. Even the BASDAI that has been more broadly used until recently was not often assessed, or at least recorded in the clinical charts. As an example, in a 2012 Spanish study of clinical charts of 1168 SpA patients, even if the rheumatologists knew the indices to assess disease activity in such patients, they did not use them routinely.17 This is probably because they are subjective and time-consuming, but this finding points out that the evaluation and records of disease activity measurements in the clinical charts could improve.

Regarding treatment response, again there is no consensus on how it should be monitored in patients with axSpA. The BASDAI and ASDAS indices are commonly used, but they have limitations such as the reliance on subjective parameters, as well as not including extra-articular manifestations or functional assessment, which correlates with the development of structural damage. In 2001, ASAS developed a preliminary definition, the ASAS partial remission, as mentioned before.7 Furthermore, one of the most challenging aspects in the definition of remission/LDA in axSpA is how to incorporate the patient's perspective.14 This is because the assessment of remission by the patient is conditioned by multiple variables that may be unrelated to disease activity, such as the presence of associated degenerative pathology, fibromyalgia and other subjective parameters that modify the patient's perception of pain and well-being.

There was consensus that the assessment of remission in axSpA patients should include disease activity, physical function, impact on daily life activities and professional life, nocturnal pain, spinal mobility, peripheral arthritis, enthesitis, quality of life, response to treatment, extra-articular manifestations, inflammatory activity in imaging tests, CRP or ESR levels and patient and physician global assessments.

None of the currently available disease activity assessment tools for axSpA includes the presence of inflammation in imaging tests. Magnetic resonance imaging (MRI) is important in the initial assessment of axSpA, as it is part of the ASAS classification criteria for this entity.18 However, important limitations preclude the use of MRI in the definition of remission, such as the availability of the equipment, the radiologist's training, the sensitivity and specificity of the test (as bone marrow oedema can be associated with other causes, such as mechanical stress, and it is also seen in healthy people) and discrepancies between clinical remission and persistence of inflammatory lesions. The evaluation of enthesitis by ultrasound could also have some value in this aspect, although it requires special training and consumes additional time.

Although prevention of structural damage is considered a key therapeutic objective, it is not included in the composite indices used for the assessment of axSpA. In clinical practice, the BASFI or the spinal mobility have been used as indirect measures of structural damage. However, depending of the disease stage, BASFI or spinal mobility may be related to the presence of disease activity or structural damage. In this sense, in the early stages of the disease BASFI or mobility are more affected by disease activity19,20 while in more advanced stages of the disease radiographic damage impairs more the mobility of the patients.20 Radiographs are, to date, the first-choice method for assessing structural damage, due to their accessibility and low cost. Different methods for the assessment of structural damage based on X-rays have been developed, such as the Bath Ankylosing Spondylitis Radiology Index (BASRI)21 and the modified Stoke Ankylosing Spondylitis Spine Score (mSASSS).22

Comorbidities are prevalent in SpA and increase the burden of the disease by worsening its activity, increasing functional disability, as well as mortality. Osteoporosis, gastroduodenal ulcer, vertebral fractures, cardiovascular disease and infections are among the most frequent comorbidities.23 Although it is necessary to consider them when selecting a treatment or adjusting its dose, there is no consensus regarding their inclusion into the definition of remission in axSpA, because they do not always correlate with disease activity and may be a consequence of other risk factors. Therefore, their inclusion would be rather a confounder.

Fatigue is one of the most important symptoms in axSpA, along with pain and stiffness. However, this manifestation has not been studied extensively, probably because it is subjective, multifactorial and difficult to quantify.24 Consequently, there is no consensus regarding its inclusion within the concept of remission in axSpA, although it is recommended to monitor fatigue during follow-up visits.

Consensus was reached regarding the inclusion of a validated tool to measure disease activity during follow-up in outpatient clinics. This tool should include nocturnal pain, spinal mobility, peripheral arthritis, enthesitis, extra-articular manifestations, activity on imaging tests, CRP and ESR levels, favourable opinion of the patient, and physician global assessments. Also, the panellists agreed on the evaluation of physical function, impact on the activities of daily life and on professional life, quality of life and response to treatment. However, although we have tools that include some of the mentioned variables, none that include all of them is available, and most probably would not be feasible in clinical practice. This may be reflecting that even when clinicians do not use specific tools to measure all of the aspects of the disease they take into account all disease manifestations to evaluate the patient state.

Regarding other disease features, there was no agreement regarding fatigue, emotional well-being, structural damage on imaging tests, treatment toxicity and economic costs.

Nowadays, the best available biomarker for the assessment of axSpA is CRP, which is useful in both monitoring disease activity and predicting radiographic progression.4 ESR has traditionally been used for the assessment of disease activity in rheumatoid arthritis and is part of the original formula of the DAS28 index, but the role of ESR in axSpA is limited.25 However, it is also possible to calculate the ASDAS score using ESR (ASDAS-ESR) when CRP is not available.16

Outcomes such as physical function, quality of life, impact on the activities of daily life and on professional life indicate not only disease activity but also radiographic damage and other factors, such as comorbidity, that may also exert some influence in decreasing the functional status of the SpA patients.26 In this case, these outcomes would reflect not only the disease process, but also other factors related to the patients’ general health. This might be the reason why the panellists included them in the patient's follow-up, as they are also relevant for the patients’ management.

There is no consensus that in order to define remission in axSpA it is mandatory that the patient meets the criteria of ASAS partial remission. Although this index incorporates the assessment of physical function (not included in the BASDAI or the ASDAS), it also has disadvantages as it does not include objective parameters of inflammation, such as CRP levels.7 Furthermore, in advanced stages of the disease the recovery of normal function becomes more difficult because of the impact of structural damage.

Regarding the instruments to measure the quality of life in axSpA patients, the only one that reached consensus was the Ankylosing Spondylitis Quality of Life (ASQoL), which is the disease-specific measure of health-related quality of life most frequently used in axSpA studies. It includes items related to the impact of disease on sleep, mood, motivation, activities of daily living, independence, relationships, and social life.27 However, ASQoL presents an important limitation and is that it is a self-questionnaire licensed by a company, which makes it very difficult to use in non-funded research. Recently, the ASAS Health Index (ASAS-HI) has been developed; a validated tool for the assessment of health in patients with both axial and peripheral SpA.28 Although ASAS recommends using this tool both in clinical trials and in routine clinical practice, it has not yet been widely implemented in our setting. However, we foresee a more widespread use of the ASAS-HI in the future and thus the authors recommend considering its use when evaluating axSpA patients, and a Spanish validated version is already available in the ASAS website (https://www.asas-group.org/clinical-instruments/asas-health-index/).

As already mentioned, the consensual proposal for clinical remission in axSpA is based on the ASDAS cut-off value for inactive disease (ASDAS <1.3) associated to the absence of extra-articular (AAU, psoriasis, IBD) and peripheral (arthritis, enthesitis, dactylitis) manifestations, plus normal CRP serum levels. Furthermore, absence of radiographic progression was also suggested to be included in the proposal.

This study has the inherent limitations of the Delphi methodology: subjectivity linked to the personal evaluations and potential bias in the selection of the expert panel.

However, it should be noted that the study reflects the opinion of a significant and representative number of experts who are members of a Spanish working group focused on spondyloarthritis, so the results of the study may be considered valuable and relevant. In addition, the Delphi methodology prevents discussing the statements in detail. Nevertheless, we consider that the included indices and recommendations were the most useful in clinical practice.