1. Future studies should not use the Oxford Criteria.

Although she had spent her life in banking, Nora Sark had always wanted to run a zoo of her own. One day, an unexpected windfall came her way and enabled her to buy a large island. At once she set about designing and constructing numerous enclosures and, consulting her copy of "Running Zoos for Dummies", she started to build up her stock by buying breeding pairs of various species.

She was fine most of the time, and could afford to pander to their particular needs, bear the costs, and consult specialists where necessary. But some of the animals were not in her Dummies Guide, and she didn't recognise them at all.

One group bothered her in particular. These were clearly birds, but they did not seem to want to fly. She tried shooing them up ramps (Gradient Enabling Therapy), putting them in trees (Clasping Branch Therapy), or even throwing them up in the air (Airborne Projectile Therapy), but all to no avail - despite having wings, they did not fly. So she put them all together in an area she called the Completely Flightless Species, and fed them donuts. Although all of her other animals thrived, these did not.

She tried all sorts of different donuts, but nothing seemed to be successful. It was so frustrating: she couldn't work out what was going wrong.

Yes, it is a silly story, but that is how the Oxford Criteria seems to work. If patients have profound fatigue for more than 6 months, and all the standard tests come back negative, under this set of criteria, they would all be diagnosed as having Chronic Fatigue Syndrome.

It is a diagnosis by default. Effectively it includes any illness of unknown cause that has profound fatigue as a persistent symptom. A diagnosed illness such as chronic heart disease, for which fatigue is a major problem, is excluded.

Around 40% of patients classified as having M.E./C.F..S according to these criteria would not be recognised as having M.E./C.F.S. elsewhere in the world, where they specify additional symptoms.

Yet in the U.K. we continue to fund research using this set of criteria

Underpinning the philosophy of the Oxford Criteria is a profile that paints the fatigue as being principally psychological in nature (due to the absence of any clear-cut medical diagnosis). Essentially this is a diagnosis of desperation. The diagnosis does not come from observation or analysis of a patient's behaviour, but simply from a failure to find evidence suggesting an alternative diagnosis (and of course a failure to find evidence is not the same as there being no such evidence).

Increasingly there is evidence of reliable biological markers for M.E., and once these are accepted, a diagnosis of M.E. will exclude a patient from this classification, just as it would with a diagnosis of heart disease.

Logically the current position makes no sense. A person's fatigue does not suddenly change in nature because of a diagnosis, or because a diagnosis is now possible that was not possible before. In effect a patient is judged as having a primarily psychological condition if doctors do not have the tools to diagnose an illness. The diagnosis is made from having no diagnosis, and so also includes patients for whom a full range of tests have not been carried out, those whose illness is yet to have a validated test, and those whose correct diagnosis has simply been missed. This is risky. Recently a friend, who had apparently had ME/CFS for many years, was diagnosed as having EHS/POTS (Ehlers-Danlos Syndrome, Postural orthostatic tachycardia syndrome): for him the prescribed CBT and GET had caused harm.

If it were the case that the Oxford criteria were used to define patients for whom long-term fatigue was a major factor, without any hidden assumption of a psychological profile, then there would be no logical reason for excluding other patients with such fatigue, but with a specific diagnosis.

It is perfectly worthwhile to study ways of relieving fatigue in a full range of illnesses. It is perfectly worthwhile to study fatigue that is specific to one particularly defined illness. But to separate out those illnesses for which we do not have clinical tests and to knowingly lump them together makes no scientific sense.

The recently published report on International Consensus Criteria develops this theme much more fully for patients who have what we would call "classic ME".

There are two clear practical reasons for not using the Oxford Criteria.

The first is that research into M.E./C.F.S. is international. Only the psychological studies on M.E./C.F.S. in the U.K. have used the Oxford Criteria as a matter of course. The most widely used international criteria are C.D.C./Fukuda. It is difficult to compare or combine the results of studies when different definitions of M.E./C.F.S. are used.

The second is that only 62% of the 640 patients in the PACE trial satisfied "international" criteria, which leads to the reasonable question of what would the international diagnosis be of the remaining 38% of patients? How can we say that they have M.E./C.F.S. when the rest of the world would not agree? Of course, that does not mean that the remaining 38% of patients should be ignored: their illness is just as real. But questions remain: how does the rest of the world assess (and treat) their illness? How can we be confident that the results of trials that use the Oxford Criteria apply to patients with M.E./C.F.S. when a little under half of the trial subjects do not have M.E./C.F.S. according to international agreement?

It must be remembered that the majority of the influential studies on M.E. funded by the U.K. taxpayer have used the Oxford Criteria, and form a core of our present "best practice".

For a brief description of the four different sets of criteria for M.E./C.F.S. go to the wikipedia entry. For further details of the C.D.C./Fukuda, London and Canadian Criteria, click here.

It is often argued that post-exertional fatigue is an important marker for M.E., and that the Canadian Criteria best represent that illness. This conflict between definitions needs to be tackled. A recent study by Nacul et. al. undertaken as part of the M.E./C.F.S. Observatory Research Programme, looking at the functional status of people with M.E., concluded that "the fact that the scores of cases meeting the Canadian criteria were consistently lower than those not meeting the criteria further suggests that diagnosis specificity is related to disease severity, and that diagnostic criteria such as the Canadian may be more appropriate for research studies investigating risk factors and disease biomarkers". If we were to use their figures from a related study, that could translate into approximately only a third of the patients in the PACE trial satisfying the Canadian Criteria - but of course, that is speculative.

Towards the end of a report by the Newcastle N.H.S. C.F.S. Service (the findings of which are covered in section 4), an interesting case is made for setting up Fatigue Clinics, rather like Pain Clinics, that are not devoted to any one particular illness. This sort of approach would free research to study M.E. itself, and put it on a par with other fatigue-associated illnesses, removing a lot of the confusion.

We also need to determine the nature of the illness(es?) classified as M.E./C.F.S. under the Oxford Criteria that would not be recognised as M.E./C.F.S. under the other criteria (38% of the PACE trial), as these patients should not be excluded from medical research.

A major part of the problem revolves around the way in which "fatigue" plays an important role in naming and specifying the illness. There are a range of other illnesses that have chronic fatigue as a severe consequence, such as heart disease and liver disease, but they do not carry the label C.F./Heart Disease or C.F./Liver Disease, and this irregularity for M.E./C.F.S. is both a source of confusion amongst the public, and a cause of frustration amongst the patients: the technical difference between C.F. and C.F.S. is too academic for most people to appreciate. For clarity – and accuracy – fatigue needs to be seen as a result of having M.E. rather than defining the illness.

There have been many conditions in the past that have been renamed both for reasons of accuracy and of perception: this illness needs to be renamed, and the report of the C.F.S./M.E. group touches on this in their recommendations in Chapter 6. In the meantime, it seems sensible to adopt the C.D.C./Fukuda suggestion that those people who experience long-term fatigue for unknown reasons and who do not satisfy the C.D.C. criteria should be classified as having Chronic Fatigue (dropping the "Syndrome" ). In that case it would be confusing for the general public if M.E./C.F.S. were to continue to be used for those who do satisfy the criteria: either M.E. alone should be used, or a suitably clear label devised.

A study by van Kessel et al. on the use of C.B.T. for Multiple Sclerosis Fatigue found much greater improvements in fatigue levels as measured by the Chalder scale than was the case in the PACE trial (from 21 to 8 vs 28 to 21). In a study on the use of CBT in rheumatoid arthritis, Hewlett, using the Multi-dimensional Assessment of Fatigue and the Visual Analogue Scale, found proportionately slightly smaller changes (than PACE) in fatigue levels. These studies are not robustly comparable with studies on M.E./C.F.S., but they put the M.E. results firmly in the group of major illnesses for which great fatigue is a consequence.

Stepping back from the situation, and looking back at the history of M.E./C.F.S., it is obvious that this lack of agreed definition lies at the core of the disagreements between the various factions. The psychologists and psychiatrists, clearly concerned about the increasing number of people suffering from chronic fatigue caused by social, work or psychological factors, devised therapies to help the situation. At the same time a different group of doctors were experiencing epidemic-type spreads of an illness for which profound fatigue was a major result. Through overlapping definitions and nomenclature, and a regrettable tendency to arrive at a diagnosis through a lack of alternative diagnosis, these two mutually exclusive aspects seemed to merge both in the eye of the public and across the broader medical front. There really is no excuse for such sloppy, unscientific thinking to continue.

We are very pleased that the recent report defining International Consensus Criteria for ME has come out so strongly with a way forward with to tackle this problem.

A rose by any other name ... or fifty flowers named rose?

How failure to perform differential diagnosis has marginalized an entire population of seriously ill and misdiagnosed patients and confused the research and medical communities for decades.

Official US Definiton: usual diagnostic definition internationally, research definition

The Fukuda definition requires profound debilitating fatigue, lasting or recurring for at least 6 months, and that is of new or definite onset, is not from ongoing exertion and does not resolve with rest, and that this fatigue is specifically defined as "mental and physical exhaustion, which differs from somnolence or lack of motivation". (Given the typical usages of the words, weakness probably would have been a better choice for what they describe than fatigue.) Additionally, there must be at least four items from the following list which must correlate with the disease (i.e. they must not have been present beforehand):

• debilitating problems with short-term memory or concentration;
• sore throat;
• tender cervical or axillary lymph nodes;
• muscle pain; multijoint pain without swelling or redness;
• headaches of a new type, pattern, or severity (including migraines);
• unrefreshing sleep;
• postexertional malaise (weakness and symptom flare, usually including dyspnea and flulike symptoms; pathologically long recovery of days, weeks, or more; triggered by physical and/or mental activity, which can be even trivial activity; the onset of this response may be delayed)

Major depressive disorders and other psychiatric-classified disorders which could result in serious fatigue are considered exclusionary, as are other medical-classified disorders which could explain the signs and symptoms, except fibromyalgia since it does not typically follow an infectious onset. If the possibly explanatory condition is fully treated, yet symptoms remain, or if symptoms are not explained by the other condition, the other condition is not supposed to be considered explanatory and a diagnosis of CFS is presumed (however, this is not generally how the exclusions are typically used, especially since CFS is often mistakenly taken to be a somatoform disorder). This inclusion was designed for research, but is also used to make clinical diagnosis.

Although the rubric does make a serious attempt at differential diagnosis, there is no requirement for any feature specific for ME/CFS, while the symptoms overlap with major depressive disorder (MDD) and other diseases, and it is thought that the Fukuda population contains some missed primary depression even in fairly good conditions. There is an ideological bias (without supporting evidence) among some researchers that it is unnecessary to differentiate between primary depression and CFS because CFS is a form of affective, maladjustment, hypochondriac, or somatoform disorder (see, for example, a discussion of how this came about and why it is incorrect in Jason et al. "Politics, Science, and the Emergence of a New Disease: the Case of Chronic Fatigue Syndrome"), and among these researchers, a Fukuda-defined population will have a high percentage of primary affective disorders, which the definition allows if one ignores those exclusions. The PACE study used the Fukuda for subgrouping only after pre-screening the entire enrolled population by use of Oxford inclusion. This raises the likelihood that those who met Fukuda in the PACE subgroups did so via symptomology inconsistent with a neurological disorder (this is how the World Health Organization classifies ME/CFS in G93.3 on the basis of the work of physicians like A Melvin Ramsay; notably, findings in early CFS definition cohorts include MRI abnormalities).

It is also possible that the Fukuda population contains several different types of post-infectious and chronic infectious disease, some of which resolve after a few years and some of which do not resolve. In addition, the population is agreed to be heterogeneous as to multiple characteristics, and there are likely to be additional factors which contribute to this heterogeneity which are not presently understood. It should be noted that the Fukuda article specifically directs research to be stratified according to several factors. This is very rarely done. Although everyone admits the heterogeneity, few seem to do much about it. Rather than stratify the group, the trend has been to broaden it further, still with no subgroups or stratification.

Common British Definition: research definition, diagnostic definition in Britain

A commonly-used definition in Britain is the Oxford Criteria. The only required symptom is severe disabling fatigue of new or definite onset, lasting at least six months; the fatigue of this definition may be related to lack of motivation and is permitted to be psychiatric in nature. Other permitted (but not diagnostic-contributory) symptoms include muscle pain, and mood and sleep disturbances. For research purposes, sometimes no clinically verifiable abnormalities (muscle twitches, bloodwork differences even if not clinically significant, etc.) are allowed, and sometimes abnormalities which are considered minor or of secondary significance are permitted. In clinical practice, persons diagnosed with or suspected to have CFS in Britain often have had a minimum of clinical investigations. There is a hypothesis, despite a lack of evidence, that too many investigations will encourage the patient in an abnormal illness belief and thereby impede recovery.

In effect, the use of the Oxford criteria is to give a diagnosis on the basis of a failure to find an alternative diagnosis. It focuses on fatigue which is a consequence of many chronic illnesses, and is found to a lesser extent in many healthy people, but treats fatigue as if it were a primary defining feature. By grouping together a mixture of different conditions, both medical-classified and psychiatric-classified disorders, it opens the way for a variety of treatments or therapies (including GET and CBT) to produce small average overall improvements simply by being helpful to a subset or by having a nonspecific effect on a variety of unrelated diseases (CBT, for instance, is used with moderate effect to impact the well-being of patients with diseases as diverse as cancer, multiple sclerosis, primary anxiety, and chronic obstructive pulmonary disease).

The conclusions to this proposed definition in 1990 were "The contributors hope that these guidelines will provide a basis for fruitful research studies, and for inter-disciplinary collaboration essential to this field of research. The guidelines are preliminary and will undoubtedly require further refinement and revision. The authors would welcome comments and suggestions" but the definition has remained as it is, in general use in the UK for the last 20 years, although its usage is less common now than it was in the 90s.

Unofficial US Research Inclusion Criteria, often described as Fukuda

A new rubric from the CDC, the CDC Empirical Definition, colloquially known as the Reeves criteria, is neither empirical nor a definition. Its objective is to operationalize the Fukuda criteria, but the actual effect is to destroy the requirements of Fukuda. Instead of requiring four diagnostic-contributory symptoms which are continuous or recurrent over the previous six months and associated with the disease, one still must endorse four items, but needs only to score a total of 25 points in symptoms over the single previous month (frequency x severity: any one item can score up to 16), and none of these need to be associated with the disease. This means that a form of post-viral fatigue which subsequently resolved as normal, a particularly long case of mono or other self-resolving form of post-viral fatigue, a case of unrelated back pain or of post-nasal drip, all could make a substantial contribution to the diagnosis. Instead of requiring a pathological reduction in activities, CDC Empirical considers a full quartile of the general population to qualify.

With this rubric, depression is considered to meet the Fukuda criteria, although Fukuda itself actually excludes primary depressive conditions that could be causing a serious level of fatigue. Jason has shown that 38% of patients with major depressive disorder can be misdiagnosed with CFS by use of this rubric. He has also argued that the reported incidence of CFS increased 10-fold coincidentally with the introduction of this rubric, and that the new prevalence of Empiric-CFS coincidentally approximates the prevalence of mood disorders (page 121).

Research using the CDC Empirical inclusion will say they used the CDC Fukuda definition. One has to check the full text to see where they have cited Reeves' 2005 paper.

Other Definitions

There is the Canadian definition which is more specific for the neurological and immune disease of myalgic encephalomyelitis (ME), classified by the WHO as a neurological condition (G93.3), and is closer to the disease that the first CFS definition was attempting to describe (this disease is also known as Epidemic Neuromyasthenia). There are also historic definitions of ME, such as the Ramsay definition, under which there is a large body of research and clinical data available which is largely ignored. These are also included in the second report by Jason mentioned above.

For a historical list of research literature and definitions, here is a wiki under preparation that may be useful.

Discussion

Using the Fukuda definition, and occasionally the Canadian criteria, clinically useful biological information about ME/CFS is available and a reasonable amount of this has been verified by replication studies.

The Fukuda definition is polythetic (select any 4 from a list of criteria) and does not necessarily require the characteristic traits of ME, so "dilutes" any patient study sample. However, the Fukuda is the more specific of the two definitions in common use today: the Fukuda and the Oxford. Current treatment policies are based upon the average improvements found in studies using these two definitions, but this means that they can make recommendations based upon a wider patient population that are inappropriate to a more specific patient population being treated.