Defining ME/CFS

There have been over 25 published case definitions for ME/CFS.5 There is no single diagnostic test to establish a diagnosis. It is instead a diagnosis based on the exclusion of other disorders, and the recognition of a cluster of characteristic symptoms, many of which are individually also seen in other conditions. Some criteria are based on an explicit definition of these symptoms and their duration and so may exclude people with ME/CFS who are not ‘typical’ in their presentation. Other criteria are less specific and may define not only people with ME/CFS, but also people with other fatigue-related conditions. The lack of a ‘gold-standard’ diagnostic test means it is impossible to assess the validity of the diagnostic criteria in terms of accuracy and precision.

The guideline committee examined the criteria currently in use to assess which might be most appropriate for suspecting and establishing a diagnosis of ME/CFS. The development of the criteria was appraised using Appraisal of Guidelines for Research & Evaluation II Instrument (AGREE II)2 to establish whether they demonstrated an unbiased, clearly reported, evidence-based and consensus-driven process using the expertise of patients, clinicians and researchers. All criteria were found to have serious or very serious limitations. Most criteria included post-exertional malaise (PEM), severe and prolonged fatigue unexplained by activity, cognitive difficulties and unrefreshing sleep, although there were differences in the way these symptoms were defined and whether they were compulsory for diagnosis.

The committee considered the balance between overdiagnosis and underdiagnosis. While the Institute of Medicine (IoM) criteria6 are potentially more encompassing than the International Consensus Criteria,7 reducing the probability of missing a diagnosis, the IoM criteria are narrower than others such as the Fukuda criteria,8 reducing the risk of overdiagnosis. The decision of the committee to use a slightly modified version of the IoM criteria derived from assessment of the quality, benefits and harms, and usability of the criteria. It does not represent a new set of diagnostic criteria.

NICE evaluated scientific evidence for ME/CFS as it is currently defined, based on the IoM criteria. This is the definition that the US Center for Disease Control uses9 when providing information to healthcare providers. Other reviews on ME/CFS, such as the recent one by IQWiG (Institut für Qualität und Wirtschaftlichkeit im Gesundheitswesen) in Germany, have used a similar approach.10

The committee’s criteria for diagnosing ME/CFS are more restrictive than in the previous NICE guideline,11 since patients are required to have all the following: debilitating fatigue, PEM, unrefreshing sleep and cognitive difficulties. Symptoms such as fatigue and sleep problems are generic to many conditions and the current NICE guideline therefore contains a definition and explanation of these terms. The fact that these individual symptoms are not specific to ME/CFS does not invalidate their mandatory inclusion in the diagnostic criteria. Fever, for example, is a mandatory feature in the American Heart Association diagnostic criteria for Kawasaki disease,12 but those criteria are not invalidated just because fever also occurs in tonsillitis.

PEM is a required component of the Revised Canadian Consensus Criteria,13 the International Consensus Criteria7 and the IoM diagnostic criteria.6 Wider definitions, such as the Oxford criteria14 or the London criteria,15 both used in the PACE trial,16 do not mandate the inclusion of PEM and therefore risk including people with fatigue from other causes. In one study of over 6000 participants, 85% of those identified by the Oxford criteria as having CFS were healthy subjects with mild fatigue who had been inappropriately labelled.17 The committee agreed that papers that had used diagnostic criteria that did not include PEM would be downgraded due to ‘indirectness’ in the GRADE (Grading of Recommendations, Assessment, Development, and Evaluations) assessment. This does not mean that papers were ignored or excluded, but that the uncertainty as to whether some trials accurately represent the population with ME/CFS was taken into account when assessing the evidence.

In summary, the diagnostic criteria chosen by the committee were based on those of the IoM, the most recent and widely accepted criteria. It adds another set of consensus criteria to the literature but is not new. The guideline committee recommended further research to define clear diagnostic criteria18 to differentiate ME/CFS from other causes of chronic fatigue.

Endpoints to assess efficacy

All NICE guidelines follow the process for evidence synthesis set out in ‘Developing NICE guidelines: the manual’.1 This process defines what outcomes and endpoints will be used when determining clinical and cost effectiveness. These are defined before evidence is examined. Any inference that the committee looked at the evidence review and picked the outcomes which best matched their pre-conceptions is simply incorrect.

There is no standard approach to choosing time points for NICE reviews. In this guideline, the longest follow-up time point for which data were available from each study was extracted. The committee considered long-term data of treatments for ME/CFS to be more reflective of real-world efficacy, and so more helpful for decision-making and implementation in clinical practice. ME/CFS is, after all, a long-term condition.

In 2019, stakeholder feedback regarding the NICE guideline on depression19 expressed concern that NICE had not considered long-term outcomes. They requested that ‘NICE should conduct a proper analysis of 1 and 2-year follow-up data where available and prioritise treatment recommendations made on the basis of this data’20 and subsequently stated that ‘long-term follow-up, where available, must be included and prioritised’.21 This is precisely what NICE has done in the ME/CFS guideline.

The guideline committee found that the early differences between interventions and controls reported in some studies were not maintained over time. There are several possible explanations for this. Initial improvements might be due to response bias or a placebo effect, or due to participants receiving alternative treatments after the initial trial interventions. However, in two major trials,22 23 exploratory analysis suggested that additional treatment received after randomisation did not explain the lack of long-term benefit from the intervention. One trial reported that ‘there is no evidence that the improvements observed in the Standard Medical Care (SMC) group were due to them having received more exposure to therapy than the Graded Exercise Therapy (GET) group after trial completion’.23 Similarly, in the long term follow-up of the PACE trial,22 the control group caught up on the intervention group and an exploratory analysis suggested that this improvement was not associated with receipt of additional treatment after the trial ended.

Data on treatment harm

Treatment harm and side effects are routinely considered in evaluations of interventions, and qualitative evidence, appropriately analysed, is important for assessing patient impact in NICE guidelines.

Twenty-five qualitative studies were included in the review of non-pharmacological interventions.24 These were analysed using robust and accepted methodologies such as the ‘Confidence in the Evidence from Reviews of Qualitative Research’ (CERQual) Approach.25 In studies relating to GET or other exercise interventions, participants reported that following the exercise programme was ‘hard work’ and in some, this led to worsening of symptoms after each session. For others, trying to persist with the exercise programme led to a worsening of their symptoms in the longer term. Debilitating exacerbation of symptoms was a reason for discontinuation.24 In one study, up to 74% of patients have reported experiencing harms after GET.26

The committee considered clinical trial data when assessing treatment harm.24 A recent meta-analysis found that harms of graded exercise versus inactive controls were not well reported in clinical trials.27 The Cochrane review on exercise therapies in ME/CFS28 concluded that: ‘We are uncertain about the risk of serious adverse reactions because the certainty of the evidence is very low’. Previous reports have also expressed concern that the reporting of adverse events in psychological treatments is weak and the criteria used may not be appropriate.29

In conclusion, the committee was presented with contested evidence of only limited and short-term benefits of GET (compared with ‘SMC’), benefits that were found to be even less outside of the setting of a clinical trial.30 This was contrasted against considerable, but methodologically low-quality evidence of patient harm,31 which by some accounts was quite devastating.32 After reviewing all this evidence, the committee agreed that GET should not be offered to patients.

Fatigue as an outcome

A lack of blinding in clinical trials can introduce performance and detection bias due to the participants’ and assessors’ knowledge of the intervention. This bias is likely to be exacerbated if the outcomes of interest are measured subjectively rather than objectively. For unblinded trials, all subjective outcomes, including fatigue measured subjectively, were downgraded consistently for this risk of bias in the GRADE analysis. This is standard practice.33 Blinded trials that used fatigue as an outcome were not downgraded for this risk of bias. ‘Downgrading’ in critical appraisal indicates the quality of the evidence, it does not mean exclusion of the evidence. Fatigue was one of ‘the outcomes that matter most’ in the evidence review.24

Many of the intervention studies for ME/CFS rely on unblinded subjective outcomes. Where reported, objective outcomes data such as activity levels, employment levels, disability payments or fitness data showed no or minimal improvement in these trials.

Evidence synthesis methodologies

NICE guidelines follow the process for evidence synthesis set out in ‘Developing NICE guidelines: the manual’1 and accepted methodologies, such as GRADE.34 GRADE methodology is not used to exclude or reject evidence, but to assess the quality and certainty of the evidence. All studies that met the inclusion criteria in the protocols were included in the evidence reviews, and these were synthesised and presented to the committee for discussion. The evidence review was quality assured externally, and the rating of the evidence as ‘low to very low quality’ is consistent with other reviews, including the Cochrane review on exercise therapy for ME/CFS.28 The guideline committee placed greater emphasis on longer-term outcomes where these were reported.

In accordance with GRADE methodology,35 trials were downgraded due to indirectness, for example, where trials included patients with different diagnostic criteria; due to imprecision, where a CI crossed the minimally important difference; or due to risk of bias, such as where therapist encouragement might reasonably lead to response bias rather than therapeutic effect. Given the heterogeneity of different types of exercise included in the review, and each type is a different multicomponent intervention, the decision to not aggregate such heterogeneous data was entirely appropriate, set out in the pre-agreed protocol and consistent with the NICE Methods Guide.1

Graded exercise therapy

Aside from the different definitions of the term ‘graded exercise therapy’, the content and application of such programmes also differ. GET was defined in the guideline as consisting of establishing a baseline of achievable exercise or physical activity and then making fixed incremental increases in the time spent being physically active, based on the deconditioning and exercise avoidance theories of ME/CFS.16 This definition reflects the descriptions of GET included in the evidence reviewed. For instance, the GET Therapists manual from the PACE Trial instructs practitioners to ‘negotiate meaningful goals’ and then, as its next step, to increase the duration of exercise by a fixed increment, such as ‘add 20% duration, up to 30 min’.36 If people have exacerbation of their symptoms, they are encouraged to maintain the same exercise level that led to their deterioration and told that ‘hurt does not equal harm’.36

Some stakeholders stated that GET, as defined in the guideline, is not what is delivered in practice.37 Others report that GET, as defined in the guideline, is still used and that has caused significant harm.37

Cross-referencing NICE guidance

NICE recommends different therapies for different conditions, even though many share common symptoms such as pain or breathlessness, with each condition requiring specific appropriately focused guidance, particularly when comorbidities may have treatment advice that conflicts. Clinicians must always balance the benefits against possible treatment harms for the individual patient.

Pain is commonly experienced in many conditions, including in people with ME/CFS who comprise less than 1% of the total population with chronic pain.38 Some describe neuropathic-type pain, some have headache or other pain symptoms and may benefit from support from specialist pain services, as is recommended in the ME/CFS guideline.3

People with ME/CFS differ from people with chronic primary pain who are the main focus of the Chronic Pain Guideline,39 making other guidance more appropriate for these individuals. Although exercise has been shown to improve chronic pain, exercise in individuals with ME/CFS can exacerbate both fatigue-related and pain-related PEM.40 41 Management strategies differ depending on the underlying condition.

The guideline recommends a personalised care and support plan which should include the management of pain.

Evidence about energy management

There is a lack of evidence supporting energy management in people with ME/CFS, and no evidence that it is a ‘cure’. But there is also no signal of harm or deterioration with energy management programmes, unlike with GET.

The guideline recommends that clinicians ‘discuss with people with ME/CFS the principles of energy management, the potential benefits and risks and what they should expect’. One principle of energy management is an attempt to avoid PEM, and so there is no drive to increase activity by setting graded goals and no suggestion that worsening symptoms are a natural response to increased activity, as there is in GET.16

People with ME/CFS may feel ready to progress their physical activity beyond their current level, and the guideline provides support for this, recommending that people may be offered a physical activity or exercise programme, overseen by a physiotherapist in an ME/CFS specialist team. Furthermore, the guideline recognises the importance of maintaining physical functioning and mobility, particularly in those with severe or very severe ME/CFS, and gives guidance on this.3