Recommendations for research

The guideline committee has made the following recommendations for research. The committee's full set of research recommendations is detailed in the full guideline.

1 Referral criteria for people with suspected axial spondyloarthritis

What are the optimal referral criteria for people with suspected axial spondyloarthritis?

Why this is important

The Dutch CaFaSpA study (van Hoeven et al. 2014, 2015) should be repeated in a UK population. This would involve examining GP databases to identify a cohort of people who have a diagnosis of non-specific back pain who first consulted their GP for back symptoms under the age of 45. These people would be invited for a full rheumatological assessment (including identifying signs and symptoms relevant to axial spondyloarthritis, X‑ray, MRI and HLA‑B27 test). All participants would be given a reference-standard diagnosis of axial spondyloarthritis or not (ideally using expert clinician opinion, or if this is not possible, using the ASAS [Assessment of Spondyloarthritis International Society] classification criteria). The cohort would be split into a development and validation set, to derive and validate optimal rules for case-finding from the available data, with each candidate strategy judged according to expected cost per quality-adjusted life year (QALY) gained (the NICE economic model developed for this guideline could easily be used to estimate these).

As a result of the large number of permutations of possible referral strategies, it is impractical to run separate validation studies for all referral criteria that are developed. Therefore, a single large, representative cohort study would, provided it measured the predictor variables for all reasonable referral strategies, provide the ability to develop and validate any number of possible referral strategies. The study would need to be large enough that sufficient data are available to derive new referral rules and to validate those rules in a separate, independent subset of the data. A UK‑specific dataset would provide more relevant data to do this than is currently available from the Dutch CaFaSpA study. For example, that study found an HLA‑B27 prevalence of 20% in people with axial spondyloarthritis and 2% in people without; much lower than the estimates found elsewhere (75% and 20% respectively). This lowers the validity of extrapolating any results found to the UK, and reinforces the need for UK‑specific data to address this question.

2 Long-term complications of spondyloarthritis

What is the incidence of long-term complications, in particular osteoporosis, cardiovascular disease (CVD) and metabolic syndrome, in people with spondyloarthritis, and how does this compare with the general population? Are any specific spondyloarthritis features or risk factors associated with the incidence and outcomes of these complications?

Why this is important

Spondyloarthritides are a group of systemic inflammatory conditions, and as such it is thought that people with these conditions may have an elevated risk of CVD, particularly if their disease is not adequately controlled. This may have direct vascular effects as well as precluding maintenance of a good level of cardiovascular fitness.

There is also clinical uncertainty around the long-term use of non-steroidal anti-inflammatory drugs (NSAIDs): whether the long-term CVD risks associated with this class of drugs are observed in this population, or whether the suppression of inflammation with these drugs mitigates some of the CVD risks associated with these conditions. In addition, risks of osteoporosis and fracture are known to be higher in people with axial spondyloarthritis than the general population, and the prevalence of axial manifestations in people diagnosed with peripheral disease implies the risks may also be high in peripheral spondyloarthritis.

The longer-term complication rates in the spondyloarthritides need to be established, as well as whether standard biological disease-modifying anti-rheumatic drug (DMARD) therapies and biological DMARDs influence these outcomes. Research that evaluates incidence of osteoporosis, CVD and metabolic syndrome in people with either axial or peripheral spondyloarthritis compared with the general population would therefore be of value. This research should take into account disease stage, personal activity levels and medicine use, and look to address how frequently it is appropriate to monitor people with spondyloarthritis for long-term complications.

3 Educational intervention to improve healthcare professionals' awareness of spondyloarthritis

What is the effectiveness and cost effectiveness of educational interventions for healthcare professionals in order to increase the number of prompt diagnoses of spondyloarthritis?

Why this is important

One of the major reasons for the delays in diagnosing spondyloarthritis is a lack of awareness of the condition by healthcare professionals. This can take many forms, such as a lack of awareness of different spondyloarthritis subtypes, lack of knowledge about associated clinical features (for example, the differences between inflammatory and mechanical back pain) or characteristics of the patient populations (for example, that spondyloarthritis affects similar numbers of men and women, or that a substantial proportion of people with spondyloarthritis are HLA‑B27 negative). Educational interventions to improve the level of awareness may therefore lead to reductions in diagnosis delays, but there is a lack of evidence as to the efficacy of these interventions. Randomised controlled trials of structured educational interventions are therefore needed to assess both whether they reduce the length of time it takes for people to be correctly diagnosed, and whether they represent a cost-effective use of NHS resources.

4 Pharmacological management of peripheral spondyloarthritis

What is the comparative effectiveness and cost effectiveness of standard DMARDs for managing peripheral spondyloarthritis, and is this effectiveness affected by differences in dose escalation protocols?

Why this is important

The committee noted that, although there are a number of randomised controlled trials comparing standard DMARDs with placebo for managing peripheral spondyloarthritis, there is a lack of evidence comparing individual standard DMARDs to other standard DMARDs. This lack of evidence makes it difficult to optimise initial therapy, either by specifying specific drugs within the class or optimising dose, administration and monitoring protocols. There is therefore the need for randomised controlled trials looking at alternative drug, dosing and administration route alternatives for the administration of standard DMARDs for managing peripheral spondyloarthritis. These trials should ensure NSAIDs and steroids are available to participants as needed, and should include (as outcome measures) both health-related quality of life (measured using the EQ‑5D) and health service resource use, to enable the results to be used to assess the cost effectiveness of the interventions.

5 Biological therapies for peripheral spondyloarthritis

What is the effectiveness and cost effectiveness of biological DMARDs in people with persistent peripheral spondyloarthritis (excluding psoriatic arthritis) or undifferentiated spondyloarthritis?

Why this is important

Although there have been trials conducted of biological therapies for psoriatic arthritis, which have led to positive recommendations in NICE technology appraisals, no such good-quality evidence exists in enteropathic arthritis, reactive arthritis or undifferentiated spondyloarthritis. The substantial side effects possible with biological therapies, and their significant cost, means it is difficult to justify offering them to these groups without good evidence of efficacy. There is therefore the need for randomised controlled trials, with a sufficient sample size to identify possible benefits, in these 3 populations. If trials were to recruit participants from multiple spondyloarthritis subpopulations, results should be clearly stratified by diagnosis to enable any differences in benefits or harms between the groups to be identified. These trials should include (as outcome measures) both health-related quality of life (measured using the EQ‑5D) and health service resource use, to enable the results to be used to assess the cost effectiveness of the interventions.

  • National Institute for Health and Care Excellence (NICE)