Submission to the Working Group
Draft Clinical Practice Guidelines on the Evaluation of Prolonged Fatigue
and the Diagnosis and Management of Chronic Fatigue Syndrome
A Review of Draft Version 1
Jim Oakley
January 1998
----------------------------------------
Contents
1 Introduction
2 Overview
3 Strengths
4 Major Areas for Improvement
Diagnostic Criteria
CBT and Graded Aerobic Exercise
Management of Sleep Disorder
Natural History
Treatment and Placebo Effect
Pathophysiology
Symptoms
Diet
Heterogeneity
Support Groups
Depression
Nomenclature
References
Consumer Input
5 Minor Areas for Improvement
6 Conclusion
----------------------------------------
1 Introduction
I would like to thank the Working Group for the significant effort that has gone
into producing the Draft Clinical Practice Guidelines. I have reviewed the guidelines
in detail, and this submission documents my response. I was unable to use the evaluation
form supplied with the draft guidelines since it did not allow adequate reporting
of the nature and extent of my comments.
As with any draft, I believe corrections and refinements are needed to the guidelines
so that they can serve their intended purpose in the best possible manner, and so
that they accurately reflect the evidence in the published literature for all topics
discussed. To the best of my ability I have focused my comments on the contents of
the guidelines and the published medical literature. I believe that this approach
is appropriate for a review of an evidence-based document. I have avoided philosophical
discussions, speculation and emotional arguments.
I have undertaken this review in a spirit of goodwill and cooperation, and have documented
my comments in the hope that they will help to strengthen the Working Group's already
substantial efforts. I have divided this submission into five sections: Overview,
Strengths, Major Areas for Improvement, Minor Areas for Improvement, and Conclusion.
2 Overview
The draft guidelines document is very attractive and comprehensive, and has the potential
to be the benchmark by which other similar documents are judged. Most relevant topics
are covered, and the content is supported by a large number of references from the
mainstream medical literature. The guidelines will be a valuable resource for the
medical profession in Australia for the next few years.
It must be recognised that, once published, the guidelines will become the 'bible'
of CFS in this country. The final document will, for the foreseeable future, be the
authoritative source on CFS for general practitioners and other health service providers,
government departments, insurance companies, schools and employers. The guidelines
will potentially be used as a source of evidence in court cases involving people
with CFS. The Working Group thus has an enormous professional and ethical responsibility
to minimise the chances of the guidelines being used as a basis for the unfair, discriminatory
or harmful treatment of people with CFS.
* Criteria for evidence-based guidelines
I believe it is imperative that the guidelines fulfil the following criteria:
1) They must present a fair and balanced view of the published literature relating
to the topics which are covered. They must accurately reflect the results and findings
of the referenced studies.
2) The principles used for developing the guidelines must be applied firmly and consistently
at all times (eg. quality of evidence ratings, exclusion of data from studies using
a single case series with no control group, etc). 3) They must be worded in such
a way that minimises opportunities for non-experts to draw inappropriate conclusions
about the topics and evidence that are presented.
4) They must be free of typographical and referencing errors, which ultimately reduce
the utility of the document.
In many instances, the guidelines satisfy these criteria very well. I have, however,
identified a number of important areas where these criteria are not satisfied, which
causes me great concern. My concerns, and the evidence to support them, are detailed
in this submission. I request that the Working Group carefully reviews all of the
issues and points that are raised, so that they can be fully addressed before the
final document is published.
3 Strengths
The draft guidelines have many strengths, which together make this document one of
national and (potentially) international significance. In my view, the strengths
are:
Presentation and Format
The guidelines document is particularly well presented. Boxes and diagrams have been
used to maximum effect, there is plenty of white space, and the text is easily readable.
When printed in colour, the presentation will be further enhanced. This is important
because, regardless of the content of the guidelines, if they are not presented in
an attractive and easy-to-read manner, they will be only partially read, or even
ignored, by many of the intended readership.
Clinical Overview
This is concise and pertinent, and provides a good overview of the subject.
Content
The information in the body of the document is presented in a logical order, is very
comprehensive, and is generally well supported by references. Most relevant issues
of interest to general practitioners are covered.
References
An extensive list of studies is referred to in the text, consistent with the requirements
for an evidence-based document. Providing references is important since it gives
credibility to the statements in the text. General practitioners can then feel confident
that the guidelines contain the best available information. The list of references
is an important resource in its own right.
Illness Severity
The guidelines treat CFS as a serious illness, and do not attempt to trivialise the
suffering of people with the disorder.
Diagrams
The diagrams are generally well thought out, and in all cases convey their message
with clarity. The diagram in Box 5.5 is outstanding.
Text Boxes
The text boxes also provide a valuable dimension to the guidelines, by summarising
the important points in the body of the text. Boxes 4.1 and 4.2 are extremely useful
in that they summarise a large amount of information, allowing readers to quickly
obtain a clear idea of the merits of different treatments which have been scientifically
tested in people with CFS. Boxes 5.1-5.4 are also excellent for the concise overview
they provide of CFS research findings.
4 Major Areas for Improvement
This section deals with the non-trivial aspects of the guidelines which I believe
could be improved. Where I have made explicit or implied recommendations or suggestions
for change, the text is underlined for ease of identification. References are the
same as those presented in the guidelines; in cases where a reference is not mentioned
in the guidelines, the full citation is included where it first appears in this submission.
Diagnostic Criteria
* Use of research diagnostic criteria
The Working Group has selected the most recent Centers for Disease Control and Prevention
(CDC) diagnostic criteria (Fukuda et al. 1994) to define CFS in a clinical setting.
This is expressly stated in the guidelines in Box 1 on page 1, and separately on
pages 5 and 6. There are two important consequences of this decision, which are described
below.
The 1994 CDC diagnostic criteria were never intended for use in a clinical setting,
but were designed to help identify more uniform groups of patients for research studies.
It is stated in the published paper (Fukuda et al. 1994, page 956) that:
"the overall purpose of the proposed conceptual framework and guidelines is
to foster a more systematic and comprehensive approach toward the collection of data
about the chronic fatigue syndrome and similar illnesses."
On 27th September 1993, the CDC in Atlanta hosted a public meeting to review the
case definition for CFS, in advance of publishing the revised case definition in
1994. One of the researchers who gave a presentation was Dr Ian Hickie, who noted
the need for separate clinical and research criteria. According to The CFIDS Chronicle
of November 1993, page 7, Dr Hickie said:
"There are different needs in criteria sets. One is a clinical one. It is inherently
patient-focused. It's about disability. It's about issues of duration and it's tied
up in the issues related to the health care system...Quite separately, one needs
to have research criteria that allow specific questions to actually be asked that
match hypotheses and criteria. So we would strongly recommend that you can't simply
come up with one size fits all. You've got to have some separation that allows the
research to continue dealing with fundamental issues and similarly allows the patients
to get on with their treatment and not feel that they have to fit the research criteria
in order to receive care and disability."
The 1994 CDC diagnostic criteria were devised by researchers including those in the
International Chronic Fatigue Syndrome Study Group. Following the publication of
the paper of Fukuda et al., one member of the study group, Dr Jonathan Rest, wrote
that:
"reliance on the new definition to apportion health insurance benefits or make
other socially relevant decisions may be utilising the definition in ways inappropriate
to its intent." (The CFIDS Chronicle, Fall 1995, page 51).
As a result of the Working Group's decision to adopt CFS research criteria for clinical
application by general practitioners, the following issues arise:
1) Many people who currently have a diagnosis of CFS will now no longer have CFS,
but will have 'chronic fatigue' (or more correctly, idiopathic chronic fatigue).
Similarly, in future, many people who would have received a diagnosis of CFS will
be diagnosed with 'chronic fatigue'.
2) Many of the studies referred to in the draft guidelines as studies of CFS are
actually studies of 'chronic fatigue', since the patients were diagnosed according
to either Oxford or Australian diagnostic criteria, which are not as exclusive as
the 1994 CDC case definition for CFS.
The logical approach for the Working Group is to adopt a broader definition of CFS
for clinical purposes. There is little merit in diagnosing some people with CFS and
some with chronic fatigue simply on the basis of the presence or absence of some
symptoms when, with the current state of knowledge, the distinction is arbitrary.
The distinction is present for research purposes, but has no meaning in clinical
practice. The adoption of a broader set of diagnostic criteria would also help to
justify the practice in the guidelines of referring to research studies employing
Oxford or Australian diagnostic criteria as CFS studies, when some or many of the
people in these studies did not meet the current CDC research criteria for CFS.
If the Working Group decides to keep the 1994 CDC case definition as the criteria
for diagnosing CFS in general practice, then it is essential that readers of the
guidelines are made fully aware that people with idiopathic chronic fatigue are not
'less sick' than people with CFS simply because their illness has a different name.
People with idiopathic chronic fatigue should, on an as-needs basis, have access
to the same medications, disability allowances, doctor support and other services
as people with CFS who have similar levels of disability. The only place in the draft
guidelines where this principle is alluded to is on page 5, where it is stated:
"Some people with severe fatigue syndromes will fail to meet the research diagnostic
criteria for CFS but may still benefit from the assessment and intervention strategies
described in these guidelines."
This one sentence is grossly inadequate to protect the interests of those many people
who have idiopathic chronic fatigue but who do not meet all the research criteria
for CFS. It is important that this issue is presented more prominently in the guidelines,
for example in a highlighted text box.
* Source of confusion
The classification of people who have clinically evaluated but unexplained chronic
fatigue is handled poorly in the draft guidelines. A potential source of confusion
about appropriate diagnostic categories arises in Box 1.1 on page 6 and in Box 2.2
on page 10. The information in the two boxes is not internally consistent, and is
not consistent with the paper containing the most recent CDC case definitions for
CFS and idiopathic chronic fatigue (Fukuda et al. 1994). The definitions, as given
by Fukuda et al., are:
Prolonged fatigue: self-reported, persistent fatigue lasting one month or longer.
Chronic fatigue: self-reported persistent or relapsing fatigue lasting six or more
consecutive months.
Idiopathic chronic fatigue: clinically evaluated, unexplained chronic fatigue that
fails to meet criteria for the chronic fatigue syndrome.
Chronic fatigue syndrome: as per Box 1, page 1 of the draft guidelines.
Instead of four definitions, Box 1.1 lists only three definitions. The definitions
are for prolonged fatigue, chronic fatigue, and CFS. The definitions of prolonged
fatigue and chronic fatigue are consistent with that of Fukuda et al. There is no
definition for idiopathic chronic fatigue in Box 1.1, but there is a definition for
CFS.
In the paper of Fukuda et al., there is an important distinction between chronic
fatigue and idiopathic chronic fatigue. Chronic fatigue refers to unevaluated chronic
fatigue. Upon evaluation, patients with chronic fatigue will either be given an alternative
diagnosis, a diagnosis of CFS, or a diagnosis of idiopathic chronic fatigue. Idiopathic
chronic fatigue is similar to CFS, except that in this case patients do not meet
all the symptom criteria for CFS.
The guidelines need to clearly define idiopathic chronic fatigue, because many people
fit into this category. If the Working Group is using the term 'chronic fatigue'
to mean idiopathic chronic fatigue, then the definition in Box 1.1 needs to be changed
to include the words "unexplained by other medical or psychological conditions",
and an explanation needs to be given in the text.
Problems with terminology also arise in Box 2.2 on page 10. On the flow chart, fatigue
present for more than one month is labelled "persistent fatigue" instead
of "prolonged fatigue". This is not consistent with the information in
Box 1.1. I suggest that the title of Box 2.2 be changed to "Flow Chart for the
Evaluation of Persistent Fatigue", where the word "persistent" has
replaced the word "prolonged", because "prolonged fatigue" has
its own restricted definition. It appears that on this flow chart the term 'chronic
fatigue' refers to idiopathic chronic fatigue, since it has been evaluated and remains
unexplained at six months. This is not consistent with the definition of chronic
fatigue in Box 1.1 or in the paper of Fukuda et al. 1994.
Cognitive Behavioural Therapy (CBT) and Graded Aerobic Exercise Therapy
* Nature of studies
The experience and consequences of severe and/or long term CFS can be horrific. I
support the use of any ethical management strategy, therapy or treatment for people
with CFS which has been proved to be safe and efficacious. The central issue with
CBT and graded aerobic exercise therapy is, of course, whether or not they have been
'proved' to be safe and effective.
My conclusion, on the basis of the published studies quoted in the guidelines in
Box 4.1 on page 18, and again on page 20, is that CBT and graded aerobic exercise
therapy should be regarded as separate treatments (they are in fact listed separately
in Box 4.1). The reasons for this view are as follows:
1) The graded aerobic exercise in the study of Fulcher and White 1997 was much more
regimented and controlled than the graded activity described in the CBT papers. For
example, Fulcher and White employed an exercise physiologist to conduct clinic-based
exercise and to prescribe home-based exercise. Patients were given ambulatory heart
rate monitors to ensure they reached, but did not exceed, target heart rates. The
focus of Fulcher and White's program was specifically on graded aerobic exercise,
not graded activity. In the paper of Deale et al. 1997, graded activity includes
activities as diverse as walking, reading, visiting friends or gardening.
2) In the last paragraph on page 1651 of their paper, Fulcher and White clearly regard
their exercise treatment as separate to and different from CBT.
3) The study of Fulcher and White 1997 is the only one in which it is possible to
assess the effect of graded exercise or graded activity on CFS. With the CBT studies,
it is not possible to determine the relative contribution of graded activity to the
improvement (or non-improvement) experienced by people with CFS who undertook CBT.
4) In support of this last point, the following statement appears in one of the studies
(Deale et al. 1997, page 414) which reported benefit for people with CFS from CBT
incorporating graded activity:
"However, further research is necessary in order to determine the efficacy of
specific components in, and the optimal delivery of, cognitive behaviour therapy
for chronic fatigue syndrome. The issues of who benefits from such treatment and
how the response rate can be maximised merit further attention."
* Graded aerobic exercise therapy
It is essential that the quality of evidence ratings are applied firmly and consistently
throughout the guidelines. In light of the comments above, I believe that graded
aerobic exercise should correctly be given a quality of evidence rating of Level
III-3. Graded aerobic exercise should be listed separately in Box 4.2 on page 19,
and to be consistent with other Level III-3 treatments, the comments in the 'Recommendation
for treatment use' column should be: "Not recommended, Further studies indicated".
In Box 4.2 under comments relating to CBT, it is stated that:
"Replicated studies show benefit of graded exercise component".
This statement is not correct and should be removed. Only one study has specifically
examined graded aerobic exercise, and it is not possible with the CBT studies to
determine how much influence, if any, graded activity had on the outcome of these
trials.
On page 20 of the guidelines, referring to the study of Fulcher and White 1997, the
following statement appears:
"Only one of the 66 people with CFS in this exercise study reported worsening
of symptoms over the 12 months of the program." This statement is incorrect,
and is a misinterpretation of the information in the box on page 1651 of Fulcher
and White's paper. Four of the original 33 people in the exercise group dropped out
during the twelve weeks of treatment. One of these four dropped out because he said
the treatment made him worse. This is reported on page 1649 of Fulcher and White's
paper under 'Treatment Results'. As can be seen from Table 1 on page 1649, one person
out of the remaining 29 in the exercise group rated himself as worse after completing
12 weeks of treatment. This is also stated in the text on page 1649 under 'Treatment
Results'. Similar results were found for the flexibility treatment group.
Thus by the time the 12 weeks of exercise treatment was completed, two people out
of 33 (6%) had been adversely affected by the treatment. Follow-ups were conducted
at three and 12 months after the completion of the exercise program. Fulcher and
White state that, at the three month follow-up, 68% of people who completed the exercise
program rated themselves as better. At the 12 month follow-up, 74% of patients who
completed the exercise program rated themselves as better. No data are provided in
the paper about how many people who completed the exercise program rated themselves
as worse at these particular times. This was a crossover trial, and only 47 (not
66) people completed exercise therapy.
It can be seen then that the above quote on page 20 of the guidelines is incorrect
and should be removed.
In the third dot point in the text box titled 'Management' on page 17 of the guidelines,
it is stated that:
"Graded aerobic exercise is safe and effective for people with CFS (Level II)".
The evidence is Level III-3 as explained above, and the quoted statement cannot be
justified on the basis of one study. It should be modified or removed. A similar
statement on page 3 in the Clinical Overview should also be modified or removed.
More trials are required to fully determine if this form of treatment is safe and
effective for the general population of people with CFS. Furthermore, all the participants
in the study of Fulcher and White 1997 were ambulatory at trial entry. Five people
were unable to participate in the trial because they were too incapacitated to attend
for outpatient treatment. It is not known how graded aerobic exercise therapy will
affect people with more severe forms of CFS.
* Cognitive behavioural therapy
It is essential that the quality of evidence ratings are applied firmly and consistently
throughout the guidelines. Two studies have found CBT to be beneficial for people
with CFS (Sharpe et al. 1996, Deale et al. 1997), while two others have found no
such benefit (Lloyd et al. 1990, Friedberg and Krupp 1994). The only quality of evidence
rating that can be assigned to CBT is Level III-4. This rating is the same as that
given by the Working Group to intravenous immunoglobulin G, where two studies showed
benefit and two didn't (see Box 4.2, page 19).
In Box 4.2, for CBT, the level of evidence rating should be changed from Level II
to Level III-4. The comments in the 'Recommendation for treatment use' column should
be "Not recommended, Further studies indicated".
In the second dot point in the text box titled 'Management' on page 17, it says:
"Cognitive-behaviour therapy is effective for people with CFS (Level 1)".
This statement cannot be justified, and should be modified or removed.
On page 26 of the guidelines, in the section headed 'How were these clinical practice
guidelines developed?', it is stated that:
"Data from a single case series without control subjects provide little more
than a stimulus for subsequent hypothesis testing. Such reports were not included
in the systematic analysis of evidence upon which these guidelines are based."
The reference to the study of Butler et al. 1991 on page 20 should therefore be removed,
since this study is a single case series without control subjects. Other CFS studies
which had no control group have not been mentioned in the guidelines, so the paper
by Butler et al. should be no exception.
* Context of studies
One of the weaknesses of the discussion in the guidelines about both CBT and graded
aerobic exercise therapy is that the referenced studies are not put in context for
the reader. It is not stated how sick participants were at trial entry, or how long
they had been ill, and it is not stated how much improvement was made by treatment
responders. Based on the discussion in the guidelines, general practitioners will
have no real idea about what to expect if they implement a CBT or graded aerobic
exercise program with their CFS patients.
In the studies which reported benefit for people with CFS who underwent CBT or graded
aerobic exercise, the outcomes at long term follow-up were as follows: 74% improved
(Fulcher and White 1997); 70% substantially improved (Deale et al. 1997); 60% significant
improvement (Sharpe et al. 1996). This is not mentioned in the guidelines. Unless
general practitioners study the referenced papers, they will have no idea that even
in the 'positive' studies, anywhere from 26% to 40% of people with CFS did not improve
significantly with these treatments. Some deteriorated (13% in the study of Sharpe
et al. 1996).
The guidelines present CBT and graded aerobic exercise therapy almost as a panacea
for CFS. General practitioners who adopt these approaches, and who are seduced by
the upbeat mood of this section of the guidelines (and who do not refer to the original
references), may tend to push their CFS patients to do more if they are not improving
during and after treatment. This could be highly detrimental to some people with
CFS.
* Can CBT and graded aerobic exercise therapy programs be implemented effectively
and safely in general practice?
Effectiveness
The CBT and graded aerobic exercise studies referred to in the guidelines were carefully
planned and conducted. The CBT programs were administered by psychiatrists or psychologists
experienced in this therapy; the graded aerobic exercise therapy was supervised by
an exercise physiologist. In the studies reporting benefit for people with CFS, there
was significant contact time between patients and those providing the treatment:
patients attended sessions at the clinic every week or fortnight, and attendance
was required for between 12 and 16 sessions, depending on the study. For the CBT
programs, each session was about an hour in length, and on a one-to-one basis. Literature
was provided to the patients in some of the studies.
The obvious question that arises is whether general practitioners have the time and
the expertise to treat their patients with CBT or graded aerobic exercise therapy.
Poorly designed and conducted programs are likely to benefit no-one. It would seem
sensible for these therapies to be administered by professionals who are experienced
in the techniques (eg. psychiatrists or psychologists for CBT), and who can structure
their working day so that one-hour sessions can be accommodated.
If it is the intention of the Working Group for general practitioners to implement
behavioural therapies with their CFS patients, careful thought needs to be given
as to how they can do this. A clear explanation of procedures will then have to be
provided in the guidelines. If it is the intention of the Working Group for experienced
psychologists or psychiatrists to administer behavioural therapies to people with
CFS upon the recommendation of a general practitioner, then this must be stated explicitly
in the guidelines in order to prevent half-baked, ineffective, or even harmful programs
from being devised and used on people with CFS.
Safety
In each of the three trials which found benefit for people with CFS from CBT or graded
aerobic exercise therapy, there were still some people in the active treatment groups
whose health had deteriorated at the time of follow-up compared with at trial entry.
In the study of Sharpe et al. 1996, deterioration occurred in 13% of the CBT group;
for Deale et al. 1997, 11% were either worse or unchanged. For the graded aerobic
exercise therapy of Fulcher and White 1997, 2 patients out of 33 (6%) were worse
by the end of 12 weeks of treatment (one of whom was unable to complete the treatment).
Adverse outcomes were likely to have been minimised in the published trials because
of the expertise of the people administering the treatments, and because the programs
were carefully designed in the first place. In the hands of less experienced people,
behavioural therapies for CFS are likely to be accompanied by higher rates of adverse
outcome. This again emphasises the need for general practitioners to utilise competent
and experienced therapists for the implementation of behavioural therapies in people
with CFS.
It is also important to note that adverse outcomes can be serious. Even experienced
psychiatrists working in multidisciplinary teams have caused major problems for people
with CFS using behavioural therapies. Consider the words of Dr Ian Hickie (see Abbey
1993, discussion, page 256), who in 1992 told the Ciba Foundation Symposium on CFS
that:
"in trials of cognitive behaviour therapy, we have made the mistake alluded
to; we have taken CFS patients who were coping well, using their somatic form, and
occasionally unleashed a disaster by attempting to move them to a psychological form.
Some have become more self-destructive, so we have created a major problem."
* Model of persistent disability
The diagram in Box 4.4 on page 21 of the guidelines attempts to demonstrate how persistent
disability arises in CFS, and how CBT can be used to break the "vicious circles".
According to Box 4.4, people with CFS who have no CBT are likely to enter a spiral
of disability perpetuated by depression, anxiety, social isolation and increased
physical illness. This theory about how persistent disability arises is at odds with
the natural history data in Box 3.2 on page 16, which show that many people with
CFS improve over time in the absence of CBT. It is my experience that people are
usually sickest in the early stages of CFS and that they gradually improve, without
any CBT. Even people who are bedridden for several years eventually become more functional
without any intervention in the form of CBT.
The model of CFS disability depicted in Box 4.4, and espoused on page 20 of the guidelines,
applies only to a minority of people with CFS. This should be made clear in the guidelines.
Whether CBT is beneficial for people with CFS who are caught in the 'vicious circle',
and whether it is beneficial for people with CFS whose health is stable or improving,
can only be determined by future controlled trials.
It is also disingenuous to discuss persistent disability in CFS without any reference
to the underlying pathophysiological processes. It is possible that in some or most
people with CFS, the physical disease process (rather than psychological/behavioural/social
factors) plays a very major role in causing persistent disability. With the current
state of knowledge about CFS, there is no way that biological aspects can be discounted.
This should be acknowledged prominently in the guidelines.
In Box 4.4 under the minor heading 'Education and Support', the second dot point
says:
"encourage return to work, social activities, physical activity".
This should only be done after reference to the patient's level of health. For example,
it is pointless encouraging someone to return to work if they are not well enough
to attend. This will only cause additional problems. The guidelines should avoid
catch-all statements, and should continually emphasis to readers that each patient
should be treated according to his or her level of health and circumstances.
* How should CBT and graded aerobic exercise therapy be presented in the guidelines?
Based on the correct level of evidence ratings for these therapies (see above), there
should be minimal discussion on CBT and graded aerobic exercise therapy in the guidelines,
since at this stage there is not enough evidence to recommend them for use in a clinical
setting. This approach is consistent with that applied in the guidelines to other
treatments which have the same level of evidence ratings.
However, symptomatic drug therapy is currently used widely in the management of CFS,
and its continued use is endorsed on pages 3 and 22 of the draft guidelines. This
has been done in the absence of any proven benefit for these treatments in people
with CFS; in fact, in most cases no trials have been done. General practitioners
would essentially have nothing to offer their CFS patients if the use of unproven
symptomatic treatments was discontinued.
The question then arises as to whether CBT and graded aerobic exercise therapy should
be completely discarded until indisputable evidence of their benefit or otherwise
for people with CFS is published. I have noted the high positive response rate to
graded aerobic exercise in the trial of Fulcher and White 1997, and the fact that
this form of therapy has been used to reduce incapacity and symptoms in conditions
such as post-polio syndrome, chronic back pain and fibromyalgia.
I have also noted the high positive response rates to CBT in the controlled studies
of Deale et al. 1997, and Sharpe et al. 1996. In particular, the results reported
in Table 4 of the paper by Deale et al. are interesting. At the six month follow-up
interview, people with CFS who underwent CBT were asked about their opinions of the
treatment. 78% were satisfied or very satisfied with the treatment outcome, while
96% thought the treatment was useful or very useful. This appears to be a good endorsement
of the CBT approach used in this trial.
Future trials may (or may not) demonstrate conclusively that CBT and/or graded aerobic
exercise are beneficial for many people with CFS. So that people with CFS who would
like to try these therapies in the hope of making some improvement in their condition
are not denied the opportunity, I suggest that the Working Group could adopt the
following approach:
1) State that the use of CBT and graded aerobic exercise therapy in CFS remains experimental,
and that conclusive results await further trials.
2) State clearly that patients should never be coerced or pressured into undertaking
these unproven therapies, and that there should be no expectation for them to do
so.
3) Endorse the use of these therapies in cases where both patient and doctor would
like to trial them.
4) Put more emphasis on the following statement which appears on page 20 of the guidelines:
"The physical activity programs are individually designed to take account of
the patient's current level of disability and avoid immediate worsening of symptoms."
Management of Sleep Disorder
I recognise that sleep disorder occurs commonly in people with CFS, as reported in
the literature. I agree in general with the principles of sleep management, and accept
the desirability of normalising circadian rhythms in people with CFS. However, I
am concerned about two of the recommendations made on page 21 of the draft guidelines
(which also appear on page 3 in the Clinical Overview).
* Restricting the sleep period to about eight night-time hours
People with CFS often sleep significantly longer than eight hours as a result of
the disease process, ie. they are not sleeping longer by choice, but as a result
of the pathophysiological processes occurring in their bodies. People who are particularly
ill may 'need' the most sleep. It is silly to recommend that sleep should be restricted
to about eight hours per night. This is like saying that muscle pain or headaches
should be restricted to eight hours per day. People who are very sick with CFS will
simply fall asleep during the day if they are forced to wake up after only eight
hours of sleep at night; for these people, sleeping is not a matter of choice.
It is stated on page 21 of the guidelines that:
"In CFS, sleeping longer does not improve physical or mental functioning".
As far as I am aware, there are no published studies to support this statement. Furthermore,
the figure of eight hours sleep per night is arbitrary. Many healthy people naturally
sleep longer than eight hours per night. This part of the sleep guidelines should
be replaced with a more general statement which, while encouraging the resumption
of normal night-time sleep, does not prescribe a specific number of hours of sleep.
* Reducing or abolishing daytime naps
Again, this is a sweeping statement which does not take into account individual needs.
Some people with CFS who do not sleep well at night need daytime naps simply to give
them enough energy to be able to undertake their daily physical and mental activities.
This also applies to some people with CFS who do sleep well at night. This guideline
needs to be broadened to inform general practitioners that the individual circumstances
of each patient should be taken into account before any drastic reduction in daytime
sleeping is implemented.
It is also stated on page 21 of the guidelines that:
"If people with CFS have a concurrent primary sleep pathology (eg. sleep apnoea),
this requires specific intervention."
However, on pages 9, 11 (Box 2.3), and 12 of the guidelines, sleep apnoea is listed
as a specific alternative diagnosis to CFS, ie. one that excludes a diagnosis of
CFS. This is in conflict with the above statement on page 21. It needs to be clarified
in the guidelines whether or not someone can have CFS and sleep apnoea concurrently,
or whether the two diagnoses are mutually exclusive.
I also recommend that the guidelines should contain reference to the following recent
paper on CFS and sleep disorder:
Morriss R K, Wearden A J, and Battersby L, 'The Relation of Sleep Difficulties to
Fatigue, Mood and Disability in Chronic Fatigue Syndrome', Journal of Psychosomatic
Research 1997; 42(6): 597-605.
Natural History
* Outcome of fatigue states
On page 15 of the guidelines, under the heading 'What is the outcome of fatigue states?',
it is stated:
"The long term outcome of CFS has been evaluated mostly in people treated within
tertiary referral settings (Box 3.2). Such patient samples are biased towards chronic
illness and limited patterns of recovery...Patient reports drawn from self-help group
populations show similar biases (Sharpe et al. 1992)."
The paper of Sharpe et al. 1992 cannot be used to support this last sentence. Sharpe
et al. found a correlation between functional impairment and membership of a patient
organisation. Length of illness was not correlated with membership of a patient organisation.
On the basis of this evidence it is unfair to say that patient reports drawn from
self-help groups are biased towards chronic illness.
In addition, there is only one reference given in the guidelines (ie. Sharpe et al.
1992) which refers to the characteristics of people who are members of CFS/ME self-help
groups. It is misleading to talk about "patient reports" as if there were
multiple papers dealing with this subject. If there are in fact more papers, they
should be referenced so that readers can assess the quality of the evidence for themselves.
On page 15 of the guidelines, referring to the study of Wilson et al. 1994b, it is
also stated that:
"During follow-up, patients were very unlikely to develop other medical disorders
(2%) but a significant number did develop other psychological disorders (19%), notably
major depression and anxiety. Similar outcomes were confirmed in several other retrospective
studies from tertiary referral centres."
It is stated clearly by Wilson et al. 1994b, that 19% of the 103 people in this study
had an alternative primary psychiatric diagnosis at follow-up. This diagnosis then
excludes a diagnosis of CFS. It is quite possible that some or all of these 20 people
were misdiagnosed in the first place, and never should have been included in a study
of chronic fatigue syndrome. The wording in the guidelines quoted above implies that
these patients developed psychological disorders that were concurrent with their
CFS, which is not an accurate reflection of the findings of Wilson et al. It should
be clearly stated in the guidelines that these 20 people had an alternative primary
psychiatric diagnosis, not a concurrent psychological or psychiatric condition.
Table II in the study of Wilson et al. lists the alternative primary psychiatric
diagnoses of the 20 patients as: personality disorder with somatoform disorder (7);
primary somatisation disorder (3); recurrent unipolar depressive disorder (6); anxiety
disorder (3); schizophrenia (1). It is hardly justifiable for the guidelines to use
the words "notably depression and anxiety". Why notably? Furthermore, if
similar outcomes were confirmed in several other retrospective studies, then these
studies should be referred to on page 15.
In the FAQs box on page 4 of the guidelines, under the heading 'How long will it
take to recover?', it is stated that:
"people with CFS who have been ill for less than two years have a high likelihood
of recovery within the following two to three years."
There is no evidence to support this statement in the papers listed in Box 3.2 on
page 16. As for the studies listed in Box 3.1 on page 16, all participants had been
fatigued for less than six months at the time of study entry and, in the studies
with adults, follow-up was a maximum of 15 months. There is thus no information on
how people with CFS who have been sick for one-to-two years fare over the subsequent
two or three years. The sentence above should be revised or removed.
* Factors delaying recovery
Of the eight studies on the natural history of CFS listed in Box 3.2 on page 16 of
the guidelines, only three (Sharpe et al. 1992; Wilson et al. 1994b; Vercoulen et
al. 1996; and excluding Bonner et al. 1994 due to very small sample size) examined
predictors of poor outcome. The discussion in the guidelines on the natural history
of CFS would be considerably strengthened with the inclusion of references to the
two large studies listed below, both of which also examined predictors of poor outcome:
Bombardier C H and Buchwald D, 'Outcome and Prognosis of Patients with Chronic Fatigue
vs Chronic Fatigue Syndrome', Archives of Internal Medicine 1995; 155: 2105-2110.
Ray C, Jefferies S, and Weir W R C, 'Coping and Other Predictors of Outcome in Chronic
Fatigue Syndrome: A 1-Year Follow-Up', Journal of Psychosomatic Research 1997; 43(4):
405-415.
Table 1 overleaf documents the main findings of the five studies that have examined
predictors and associations of outcome in CFS (including the two studies listed immediately
above). These five studies all included people who met one of the published sets
of CFS diagnostic criteria. They are by far the most comprehensive studies of their
kind, and each included more than 100 patients. Other follow up studies are reported
in Joyce et al. 1997. Of these, the study of Clarke et al. 1995 was also very comprehensive,
but has not been included here since no published case definition was used for inclusion
of patients, and the sample size - particularly for those who did meet CDC or NIH
CFS criteria - was relatively small.
It can be seen from Table 1 that the factors that are consistently associated with
poor outcome, for which there is little or no conflicting evidence, are:
* emotional disorder at follow-up
* high levels of fatigue or functional impairment at entry
* a low sense of control over symptoms at entry
Emotional disorder at follow-up was found to be associated with poor outcome in the
studies of Sharpe et al. and Bombardier et al.; Wilson et al. found that an alternative
primary psychiatric diagnosis at follow-up was associated with poor outcome. The
data of Vercoulen et al. indicate a positive correlation between poorer psychological
wellbeing at follow-up and poorer functional status, but it is not stated if the
apparent association reached significance. Ray et al. did not measure emotional disorder
at follow-up.
Levels of levels of fatigue and functional impairment at entry were found to predict
poor outcome in the studies of Vercoulen et al. and Ray et al. These parameters were
not measured at entry in the other three studies. Similarly, Vercoulen et al. and
Ray et al. found that having a sense of lack of control over symptoms, or a poor
internal locus of control, was associated with poor outcome. The other three studies
did not measure this parameter.
The factors that are consistently found to be not associated with poor outcome, and
for which there is little or no conflicting evidence in Table 1, are:
* age
* gender
* emotional disorder at entry.
Age, whether at study entry or illness onset, was found to have no association with
outcome in CFS in all of the studies in Table 1. Bombardier et al. found that older
age at study entry predicted poorer outcome in a group of people with chronic fatigue,
but this finding did not apply to the CFS group in the same study. In the review
of Joyce et al. 1997, four out of eight studies (not four out of six as stated by
Joyce et al.) found that increasing age was associated with poorer outcome. The four
that did find an association included the study of Bombardier et al. mentioned above,
and the study of Hinds and McCluskey 1993, which simply found that patients under
20 years of age had a better outcome. The weight of evidence clearly supports the
contention that older age does not predict poorer outcome.
Gender has never been found to predict poorer outcome, and emotional disorder at
study entry was found in four studies in Table 1 to have no association with outcome.
============================================================
Table 1: Summary of Findings of Studies on the Natural History of CFS
--------------------------------------------------
Paper: Sharpe et al. 1992
Number of Participants: 144
Duration of Follow-Up (Months): Median: 12, Ave: 15#
Predictors/Associations of Poor Outcome:
At follow-up only:
* Belief in a viral cause
* Avoiding alcohol
* Belonging to a self-help organisation
* Current emotional disorder
Factors Not Associated with Poor Outcome:
* Age
* Gender
* Marital status
----------------------------------------
Paper: Wilson et al. 1994
Number of Participants: 103
Duration of Follow-Up (Months): 38
Predictors/Associations of Poor Outcome:
* Disease conviction at entry
* Primary psychiatric diagnosis at follow-up
Factors Not Associated with Poor Outcome:
* Age at illness onset
* Duration of illness
* Premorbid psychiatric diagnoses
----------------------------------------
Paper: Bombardier et al. 1995
Number of Participants: 226 (CFS)
Number of Participants: 216 (CF)
Duration of Follow-Up (Months): 18
Predictors/Associations of Poor Outcome:
CFS
* Dysthymia at follow-up
CF
* Older age at entry
* Longer duration of symptoms
* Lifetime history of dysthymia
Factors Not Associated with Poor Outcome:
For CFS and CF
* Gender
* Psychiatric diagnoses at entry
----------------------------------------
Paper: Vercoulen et al. 1996
Number of Participants: 246
Duration of Follow-Up (Months): 18
Predictors/Associations of Poor Outcome:
At entry:
* High fatigue severity
* Lack of control over symptoms
* Functional impairment
Predictors of Improvement
At entry:
* Sense of control over symptoms
* Lower fatigue severity
* Shorter duration of illness
* Relative absence of physical attributions
Factors Not Associated with Poor Outcome:
At entry:
* Age
* Physical attributions
* Level of psychological wellbeing
* Sleep problems
----------------------------------------
Paper: Ray et al. 1997
Number of Participants: 137
Duration of Follow-Up (Months): 12
Predictors/Associations of Poor Outcome:
At entry:
For impairment at follow-up:
* Functional impairment
* High level of fatigue
* Illness duration
* Internal locus of control
For fatigue at follow-up:
* Fatigue
* Cognitive difficulties
* Information seeking
Factors Not Associated with Poor Outcome:
At entry:
* Age
* Emotional distress
* Depression
* Illness attribution
--------------------------------------------------
#Average calculated from Table 1 on page 149 of paper by Sharpe et al. 1992.
============================================================
On page 15 of the guidelines, it is stated that:
"The factors associated with a poorer outcome include older age, concurrent
psychiatric disorder, and the person's belief that the illness is purely physical
in origin."
As discussed above, it is not correct to claim that older age is associated with
poorer outcome. It is, however, correct to say that concurrent psychiatric disorder
is associated with a poor outcome. The third item, ie. the person's belief that the
illness is purely physical in origin, cannot be justified on the basis of the evidence
in Table 1.
In Table 1, it can be seen that the studies of Sharpe et al. and Wilson et al. found
that belief in a physical cause of the illness was associated with poorer outcome.
In the study of Sharpe et al., however, the belief was assessed only at follow-up,
and (as pointed out by the authors themselves) was thus not predictive. The studies
of Vercoulen et al. and Ray et al. found that illness attributions were not predictive
of poor outcome (eg. higher fatigue severity). Thus the level of evidence for this
parameter is Level III-4.
Vercoulen et al. did find that physical attributions were predictive of improvement
(p<0.05), but this was a small influence compared with the correlation that sense
of control over symptoms had with improvement (p<0.0001). In the same study, lack
of sense of control over symptoms was also predictive of fatigue severity (p<0.01),
whereas physical attributions were not (as mentioned above). These data are in Tables
2 and 3 on page 492 of Vercoulen's paper. In the study of Ray et al., a low internal
locus of control was predictive of impairment (p<0.03). This parameter is very
similar to the sense of control parameter of Vercoulen et al. The level of evidence
to support the predictive qualities of this parameter is therefore Level II.
Table 2 below summarises the levels of evidence for the predictors/associations of
outcome recommended here for inclusion in the guidelines, and those which are currently
in the draft guidelines.
Data with a higher level of evidence should take precedence over data with a lower
level of evidence. The wording on page 15 of the guidelines should be changed to
reflect the recommendations contained in Table 2, to be consistent with the evidence
in the literature. The fourth dot point in the text box titled 'Natural history'
on page 15 also needs to be revised in light of the above evidence.
============================================================
Table 2: Predictors and Associations of Poor Outcome in People with CFS
----------------------------------------
Recommended: Emotional disorder at follow-up
Evidence Level: Level II
Draft Guidelines: Older age
Evidence Level: Level III-4 (at best)
----------------------------------------
Recommended: High levels of fatigue or functional impairment at entry
Evidence Level: Level II
Draft Guidelines: Psychological disorder at follow-up
Evidence Level: Level II
----------------------------------------
Recommended: Low sense of control over symptoms at entry
Evidence Level: Level II
Draft Guidelines: Belief in a purely physical cause of CFS at entry
Evidence Level: Level III-4
============================================================
On page 20 of the guidelines, the following statement appears:
"Given that simplistic attributions of a purely physical basis for the illness
are associated with poor outcome, people with CFS should be encouraged to adopt the
widest possible view of the psychosocial, medical and rehabilitative strategies to
promote recovery."
This is an excellent sentiment, but the wording referring to attributions of a purely
physical basis for CFS being associated with poor outcome should be replaced with
wording relating to lack of sense of control over symptoms being associated with
poor outcome.
Information contained in the guidelines will invariably be conveyed to people with
CFS. Telling medical practitioners that older age is associated with poor outcome
is likely to be unhelpful for older people with CFS, since they can have no control
over this factor, and believing this will increase their sense of powerlessness.
This is quite apart from the fact that the weight of evidence indicates that older
age is not associated with poorer outcome, as discussed above.
The third dot point in the text box on page 15 also needs to be revised. The natural
history studies referred to in the draft guidelines and in this submission reported
that anywhere from 17% to 64% of people with CFS improved over time, with a small
additional percentage of people recovering completely. Three studies in Table 1 (Wilson
et al. 1994; Bombardier et al. 1995; and Ray et al. 1997) reported that more than
60% of people with CFS were improved at follow-up compared with their condition at
study entry. Improvement inherently encompassed psychological as well as physical
dimensions. Vercoulen et al. 1996, reported that psychological wellbeing was significantly
improved at follow-up in those who had recovered, in those who had improved, and
even in those who had not improved. Thus the frequency and severity of psychological
disorders in people with CFS reduces over time (Level I). It is not justifiable to
say that people with CFS are at increased risk of developing psychological disorders!
The statement in the text box probably derives from an interpretation of the study
of Wilson et al. 1994, but Wilson et al. found that 19% of people in their study
had an alternative primary psychiatric diagnosis at follow-up, and it is thus quite
possible that these individuals never had CFS in the first place. Furthermore, no
other study in Table 1 reported alternative primary psychiatric diagnoses at follow-up.
The common finding that emotional disorder at follow-up is associated with poor outcome
is completely different to claiming that people with CFS are at increased risk of
developing psychological disorders. In the absence of any evidence to support it,
the third dot point in the text box on page 15 should be rewritten or removed.
The discussion relating to factors delaying recovery makes no mention of possible
biological factors that may contribute to poor outcome. Just because the pathophysiological
basis of CFS is unknown does not mean that it can be disregarded as an important
influence on illness outcome. This should at least be acknowledged in this section
of the guidelines. According to the data in Box 3.2 on page 16, people who have been
sick with CFS for several have a very limited chance of regaining their pre-illness
level of health over the subsequent few years. Psychological factors may explain
some, but not all, components of persisting ill-health in people with CFS, and may
play a larger role in some individuals than in others.
* Prolonged fatigue
With respect to prolonged fatigue, the following sentence appears on page 15 under
the heading 'What factors may delay recovery?': "In a study of the relationship
of non-specific viral illness and the development of prolonged fatigue in general
practice, the person's view of the illness and the doctor's behaviour, rather than
the viral infection, were predictive of the development of prolonged fatigue (Cope
et al. 1994)."
This is a highly selective use of the available evidence. Another larger study (Wessely
et al. 1995), which followed people from before the onset of a viral illness until
six months after their initial presentation to a general practitioner, reported different
results. Wessely et al. 1995 found that the person's view of the illness (ie. belief
that fatigue was due to a physical cause) was not a significant predictor of fatigue
at six months' follow-up. This is in direct contrast to the findings of Cope et al.
1994.
Wessely et al. also found that the presence of fatigue before presentation with a
viral illness was a strong predictor of fatigue at six months' follow-up. This is
also in direct contrast to the findings of Cope et al., who found that fatigue reported
before the initial consultation was not associated with fatigue six months after
the initial consultation.
The two studies cited here reported conflicting results on two important issues.
The guidelines should either present a balanced view of the evidence, or they should
avoid discussion of the results altogether.
* Text box 3.2, page 16
This text box needs to be expanded and revised. The studies of Bombardier et al.
1995, and Ray et al. 1997, should be included. The study of Calder et al. 1987, should
be in moved into Box 3.1, since patients in this study had been ill for less than
six months at the time of entry. In fact, 65% of patients were enrolled in Calder's
study less than one month after becoming ill. This was a study of prolonged fatigue,
not CFS.
Sample size should refer to those patients who were initially evaluated and who were
evaluated at follow-up. For example, in the study of Wilson et al. 1994, of 139 people
who were sent follow-up questionnaires, only 103 returned completed questionnaires.
Thus the sample size was 103, as correctly reported in Box 3.2 of the guidelines.
However, the same principle has not been applied to all other studies listed in the
box. Box 3.2 in the guidelines needs to be updated with the information in Table
3 below:
============================================================
Table 3: Corrections to Sample Size Data in Box 3.2 on Page 16
----------------------------------------
Study: Calder et al. 1987*
Sample Size: 59 (not 65)
----------------------------------------
Study: Sharpe et al. 1992
Sample Size: 144 (not 177)
----------------------------------------
Study: Hinds and McCluskey 1993
Sample Size: 291** (not 393)
----------------------------------------
Study: Peterson et al. 1991
Sample Size: 62 (not 68)
----------------------------------------
Study: Vercoulen et al. 1996
Sample Size: 246 (not 298)
----------------------------------------
* This study belongs in Box 3.1
** According to Joyce et al. 1997
============================================================
The length of follow-up for the study of Sharpe et al. 1992, was not 24 months as
stated in Box 3.2, but was a median of 12 months. The average length of follow-up,
as calculated from Table 1 on page 149 of the paper by Sharpe et al., was 15 months.
The outcome figures in Box 3.2 are correct as given, because in the papers these
data have been calculated according to the correct sample sizes listed in Table 3
above. The exception is the study of Calder et al. 1987. In this study, 36 people
were still unwell at the 12 months' follow up. Thus (59-36) people, ie. 23 people
had improved. This translates to a percentage improvement of (23/59x100)%, ie. 39%,
not 45% as listed in Box 3.2.
In the study of Sharpe et al. 1992, it is stated that 13% of people had recovered
at follow-up, but no figure is given in the paper for the overall percentage of people
who improved. This can be calculated from Table 1 on page 149 of Sharpe's paper.
Of the whole sample of 144 people, 93 (or 65%) remained functionally impaired at
follow up, whereas 100% were functionally impaired at study entry. Thus 35% had moved
from being functionally impaired to being not functionally impaired over the period
of follow-up. Sharpe et al. state that 13% of patients regarded themselves as fully
recovered. This means that (35-13)%, ie. 22% of people improved but did not recover.
In Box 3.2 it is indicated that 63% of people in this study improved. This is not
correct.
The correlates of persistent debility presented in Box 3.2 also need some revision.
For the study of Sharpe et al., "limiting exercise" is listed as a correlate
of persistent debility, but this variable did not reach significance when the logistic
regression adjusted for confounding associations. "Limiting exercise" should
thus be removed from Box 3.2. It also needs to be explained to readers of the guidelines
that the other correlates found by Sharpe et al. which are listed in Box 3.2 were
assessed only at follow-up.
Similarly with the study of Wilson et al. 1994, the correlation of persistent disability
with psychiatric disorder (which should correctly be expressed as primary psychiatric
disorder) was for psychiatric disorder at follow-up, not at study entry. The correlation
with belief in a purely physical cause of CFS was for belief at study entry, not
at follow-up.
In the study of Hinds and McCluskey 1993, the correlation of persistent disability
was with age over 20 years (according to Joyce et al. 1997). In Box 3.2, the words
"Younger age" should be replaced with "Age over 20 years".
As discussed previously, the study of Vercoulen et al. 1996 found that the correlates
of persistent disability were (in order of importance): fatigue severity at study
entry, lack of sense of control over symptoms at study entry, and functional impairment
at study entry. None of these appears in Box 3.2, but "belief in a purely physical
cause of CFS" does. This is not correct, and should be removed and replaced
with the true correlates.
* Box 2, page 2
This box presents a graph of the number of patients with continuing symptoms of CFS
versus time, under the heading of 'Model of natural history of chronic fatigue syndrome'.
There are no references given to support this model, and the information in Boxes
3.1 and 3.2 on page 16 is not sufficient to allow such a graph to be constructed.
In an evidence-based document, there is little scope for inclusion of unsubstantiated
models. If the purpose of the graph in Box 2 is to convey an idea or a general trend
which is known intuitively, but does not have hard evidence to support it, it would
be much better to delete the scale on the vertical axis so that specific recovery
rates are not implied. Alternatively, the graph could be removed altogether.
Furthermore, does the phrase "Patients with continuing symptoms of CFS"
include people who have substantially improved but still have residual symptoms of
CFS, no matter how minor; or does it refer only to people who still meet diagnostic
criteria for CFS? In the absence of any references, it is difficult to interpret
the graph.
* Doctor-patient relationship
On page 15 of the guidelines, under the heading 'Does the doctor-patient relationship
affect the outcome of CFS?', it says:
"Simplistic notions offered by doctors, with unjustified medical labels such
as 'chronic viral infection' or pseudomedical diagnoses such as 'hypoglycaemia',
are inappropriate and unhelpful."
Readers of the guidelines should also be informed that unjustified labels such as
'malingering' and 'depression' are also inappropriate and unhelpful for people with
CFS.
Treatment and Placebo Effect
* Extent of placebo effect
On page 17 of the guidelines it is stated that:
"At least 30%-50% of people with CFS typically demonstrate improvement in the
non-specific (or 'placebo') treatment arm of controlled trials (Wilson et al. 1994a,
Hickie et al. 1995b)."
There is no basis whatsoever for this statement, and it should be removed from the
guidelines. The evidence is presented in Table 4 overleaf, where the controlled treatment
trials listed in Box 4.1 on page 18 of the guidelines are again presented, but this
time with their respective placebo response rates. Reference to the hydrocortisone
trial of Straus et al. should be removed from Box 4.1 because it has not been published,
and therefore has no place in the evidence-based guidelines. It is not included in
Table 4.
It is also questionable whether the moclobemide trial of Wilson et al. 1994 should
be referenced, since the only published information is an abstract, and it is not
possible for readers to obtain full details of the trial. It is included in Table
4 overleaf nonetheless.
In the trials for which it was possible to calculate a placebo-response rate, those
with the highest rates were Rowe 1997 (44%) and Straus et al. 1988a (42%). The total
length of Rowe's study was nine months, including follow-up at six months after the
end of treatment. The incidence of natural improvement in CFS over time - particularly
in children - can be high, and this may have been a factor in the 44% of children
who received placebo and who were significantly improved at the end of the study.
Straus et al. also noted the potential impact of the natural history of the illness
on the high placebo response found during their small 30 week trial.
============================================================
Table 4: Placebo Response Rate in Published Controlled Trials of Treatments for CFS
----------------------------------------
Immunological/Virological
Treatment: Intravenous immunoglobulin G
Study: Peterson et al. 1990
Sample Size: 28
Response Rate to Placebo or 'Non-Active' Treatment (%): 0*
Study: Lloyd et al. 1990
Sample Size: 49
Response Rate to Placebo or 'Non-Active' Treatment (%): 12
Study: Rowe 1997
Sample Size: 71
Response Rate to Placebo or 'Non-Active' Treatment (%): 44
Study: Vollmer-Conna et al. 1997
Sample Size: 99
Response Rate to Placebo or 'Non-Active' Treatment (%): Not stated but statistically
significant
----------------------------------------
Treatment: Ampligen
Study: Strayer et al. 1994
Sample Size: 92
Response Rate to Placebo or 'Non-Active' Treatment (%): 0**
----------------------------------------
Treatment: Acyclovir
Study: Straus et al. 1988a
Sample Size: 24
Response Rate to Placebo or 'Non-Active' Treatment (%): 42
----------------------------------------
Treatment: Transfer factor
Study: Lloyd et al. 1993
Sample Size: 88
Response Rate to Placebo or 'Non-Active' Treatment (%): 33***
----------------------------------------
Treatment: Interferon-alpha
Study: See and Tilles 1996
Sample Size: 30
Response Rate to Placebo or 'Non-Active' Treatment (%): Paper not sighted
Behavioural
Treatment: Cognitive behaviour therapy
Study: Lloyd et al. 1993
Sample Size: 88
Response Rate to Placebo or 'Non-Active' Treatment (%): 33***
Study: Friedberg and Krupp 1994
Sample Size: 42
Response Rate to Placebo or 'Non-Active' Treatment (%): 0
Study: Sharpe et al. 1996
Sample Size: 60
Response Rate to Placebo or 'Non-Active' Treatment (%): 23
Study: Deale et al. 1997
Sample Size: 53
Response Rate to Placebo or 'Non-Active' Treatment (%): 19
----------------------------------------
Treatment: Graded exercise
Study: Fulcher and White 1997
Sample Size: 59
Response Rate to Placebo or 'Non-Active' Treatment (%): 27
CNS active agents
Treatment: Galanthamine hydrobromide
Study: Snorrason et al. 1996
Sample Size: 49
Response Rate to Placebo or 'Non-Active' Treatment (%): 4****
----------------------------------------
Treatment: L-carnitine
Study: Plioplys and Plioplys 1997
Sample Size: 30
Response Rate to Placebo or 'Non-Active' Treatment (%): No placebo arm
----------------------------------------
Treatment: Amantidine
Study: Plioplys and Plioplys 1997
Sample Size: 30
Response Rate to Placebo or 'Non-Active' Treatment (%): No placebo arm
----------------------------------------
Treatment: Moclobemide
Study: Wilson et al. 1994
Sample Size: 90
Response Rate to Placebo or 'Non-Active' Treatment (%): Abstract not sighted
----------------------------------------
Treatment: Fluoxetine
Study: Vercoulen et al. 1996
Sample Size: 96
Response Rate to Placebo or 'Non-Active' Treatment (%): 10
----------------------------------------
Treatment: Phenelzine
Study: Natelson et al. 1996
Sample Size: 18
Response Rate to Placebo or 'Non-Active' Treatment (%): 0
Metabolic/other
Treatment: Essential fatty acids
Study: Behan et al. 1990
Sample Size: 63
Response Rate to Placebo or 'Non-Active' Treatment (%): 17
----------------------------------------
Treatment: Magnesium sulphate
Study: Cox et al. 1991
Sample Size: 32
Response Rate to Placebo or 'Non-Active' Treatment (%): 18
----------------------------------------
Treatment: Liver extract + folic acid + vitamin B12
Study: Kaslow et al. 1989
Sample Size: 14
Response Rate to Placebo or 'Non-Active' Treatment (%): Not stated but ***** statistically
significant
----------------------------------------
Treatment: Terfenadine
Study: Steinberg et al. 1996
Sample Size: 28
Response Rate to Placebo or 'Non-Active' Treatment (%): 0
----------------------------------------
* Some symptoms improved in some patients, but no clinically significant benefit
was seen.
** Based on Karnofsky performance score. Minor improvement in other measures.
*** Stated that up to one third of people reported more than one standard deviation
of improvement in physical and psychological status.
**** Placebo-controlled phase only lasted two weeks. Only one patient out of 24 (4%)
receiving placebo improved more than 10% during this time.
***** Two week trial only (crossover design).
============================================================
Notwithstanding this it can be calculated from Table 4, for the studies for which
figures are available, that the average placebo response rate is just 17%. Five studies
had a zero percent placebo response rate. For the studies of Vollmer-Conna et al.
1997, and Kaslow et al. 1989, it is indicated that a statistically significant placebo
response was found, but this response was not quantified in terms of the percentage
of individuals who improved. If it is assumed that in each of these studies the rate
of response to placebo was 50%, then the overall average placebo response rate becomes
20%.
For blinded controlled trials (ie. excluding trials of CBT and graded aerobic exercise
therapy), the average placebo response rate is 15%; or 20% if the trials of Vollmer-Conna
et al. and Kaslow et al. are included with an assumed 50% placebo response rate.
The studies of See and Tilles 1996, and Wilson et al. 1994, were not sighted and
so are not included in this analysis. Even so, it can be seen that the statement
in the guidelines that at least 30%-50% of people with CFS typically demonstrate
improvement in the placebo arm of trials is unsupportable.
* Inadequate supporting references
Two references are given in the guidelines to support the above quote: Wilson et
al. 1994a, and Hickie et al. 1995b. (The paper of Hickie et al. 1995b is titled 'Chronic
fatigue syndrome: current perspectives on evaluation and management'). To justify
comments about high placebo response rates in CFS, Hickie et al. 1995b in turn refer
to two papers: Lloyd et al. 1993 (transfer factor trial), and Wilson et al. 1994a.
Wilson's paper is a review, and presents no new data. To justify their own comments
about high placebo response rates in CFS, Wilson et al. 1994a refer to two trials:
Straus et al.1988a (acyclovir), and Kaslow et al. 1989 (liver extract plus vitamins).
The study of Kaslow et al. is the smallest and shortest double-blind, placebo-controlled
trial ever done with people with CFS - only 14 people participated in the trial,
which lasted for just two weeks.
The paper of Hickie et al. 1995b was published at a time when eleven of the 23 trials
listed in Box 4.1 had been published. Yet the authors directly or indirectly referred
to only three of the eleven published trials. The comments in this paper about high
placebo response rates in people with CFS were not valid in 1995, and they are even
less valid today.
* Inconsistent ratings
Of the treatment trials listed in Box 4.1 in the guidelines, 15 out of 23 (65%) (excluding
the hydrocortisone trial of Straus) are given a number 6 in the column headed 'Quality
of study design and analysis'. The number 6 refers to the comment "Response
rate to placebo therapy consistent with standard clinical care (ie. 15%-40%)".
It is not stated where the figures for response rate to standard clinical care come
from. Are they derived from trials with people with CFS, or are they the normal response
rates to standard clinical care which occur across a range of medical illnesses?
Either way, if 65% of CFS treatment trials have placebo response rates between 15%
and 40%, it is inconsistent with the claim that the typical placebo response rate
is at least 30%-50%. However, of the trials that are given a number 6 rating, two
had placebo response rates lower than 15%. These are the gammaglobulin studies of
Peterson et al. 1990 (0%) and Lloyd et al. 1993 (12%). The number 6 should be removed
from the 'Quality of study design and analysis' column for these two trials.
* Sample size
Sample size in Table 4 of this submission refers to the number of people who completed,
or who partially completed, the trial and who were included by the authors in the
analysis of results. Sample size in Box 4.1 in the draft guidelines refers to the
number of people initially enrolled in the trial, regardless of whether or not they
completed it, or whether or not they were included in the analysis of results. On
either basis, the sample size for the study of Vollmer-Conna et al. 1997 in Box 4.1
should be 99 rather than 96.
* Further studies indicated
It is correctly noted in the first dot point in the text box on page 17 of the guidelines
that no single pharmacological treatment has been shown to be effective for people
with CFS. The absence of any effective treatment is the single greatest problem for
people with CFS and their doctors. It is thus essential that controlled and blinded
treatment studies are encouraged. Initial trials of four substances which have shown
benefit for people with CFS need to be followed up with further trials. The four
substances are: essential fatty acids; phenelzine; L-carnitine; and galanthamine.
It is correctly stated in Box 4.2 on page 19 of the guidelines that these drugs cannot
currently be recommended for treatment use. There is limited knowledge about their
efficacy and side effects when used by people with CFS. However, further studies
are definitely indicated, and this should be stated in Box 4.2, as has been done
for moclobemide and Ampligen.
* Speculation
It is stated on page 17 of the guidelines, under the heading 'What are the expectations
of a treatment trial for CFS?', that:
"any claim that a particular treatment can cure most people with CFS is likely
to be spurious, or the treatment will be acting via a non-specific mechanism."
This is speculation, and has no place in evidence-based guidelines. It is pre-empting
the outcome of research into the pathophysiology of CFS, and is based on assumptions
about the biological heterogeneity of people with CFS. There is little or no evidence
to support these assumptions (see the sub-section on Heterogeneity below).
For example, in the essential fatty acid trial of Behan et al. 1990, there was a
treatment response rate of 85%, and a placebo response rate of 15%. It is hard to
see how a "non-specific mechanism" could account for the difference in
response rates between the treatment and placebo groups, given that all patients
took capsules that looked identical, and there was no regurgitation of fishy-tasting
material. In addition, although most people with CFS who received the active treatment
did report improvement, there is no good reason to deride the results as "spurious".
The Working Group obviously thought that this study was of good enough design to
be reported in the guidelines. So, although essential fatty acids did not 'cure'
anyone with CFS, they did apparently help to improve the condition of a high percentage
of people. This example is illustrative of the absurdity of sweeping statements like
the one quoted above.
Pathophysiology
* Leading hypothesis?
On page 22 of the guidelines under the heading 'What are the major research findings
in people with CFS?', it is noted that one of the five leading hypotheses about the
pathophysiological basis of CFS is that the illness is the outcome of:
"a psychologically determined response to infection or other stimuli occurring
in 'vulnerable' individuals (Imboden et al. 1961; Abbey 1993)."
There is no evidence offered in support of this statement in Boxes 5.1 to 5.4. While
the study of Imboden et al. 1961 looked at convalescence in 26 people after influenza,
this pathogen has not recently been associated with CFS (see Box 5.1, page 23). The
follow-up period in this study was only three to six weeks after the initial illness(!),
so it has little relevance to CFS. The studies of White et al. 1995 and Bruce-Jones
et al. 1994, refute the above hypothesis in relation to the Epstein-Barr virus, which
is a recognised viral trigger for CFS.
The paper of Abbey 1993 is a review, and offers no new evidence to support the hypothesis.
Thus it is very difficult to see how the above quote is regarded as one of the "leading
hypotheses" about the pathophysiological basis of CFS. After more than a decade
of research into the illness, a leading hypothesis should have some good evidence
to support it, and this one doesn't. The sentence quoted above should be removed
from the guidelines.
* Retroviruses
It is stated on page 22, under the heading 'What are the gaps in our knowledge and
priorities for future research' that:
"there is good evidence excluding several candidate mechanisms for the disorder,
including retroviral infection..."
In the text box titled 'Phenomena associated with CFS' on page 22, it says:
"Retroviruses do not cause CFS (Level 1)".
And in Box 5.1 on page 23 under the heading 'Retroviruses' is the conclusion of:
"Strong evidence against a role for retroviruses in CFS."
It is frequently acknowledged in the literature on retroviruses in CFS that the findings
of the studies do not preclude a role for unknown retroviruses. For example, Heneine
et al. 1994 state that their very comprehensive study
"clearly does not exclude the involvement of other retroviruses in this population
of CFS patients or of the assayed retroviruses in other such populations."
Therefore, the statements in the guidelines concerning retroviruses should refer
to known retroviruses, and should not make any assumptions about as-yet unidentified
retroviruses.
* Patient selection for research studies
It is also stated on page 22, under the heading 'What are the gaps in our knowledge
and priorities for future research' that:
"Future studies must seek to identify potentially homogeneous patient groups,
including subjects collected prospectively from defined infectious or other putative
causes of CFS."
This is an important and valid approach to CFS research. However, in the paper of
Fukuda et al. 1994 (page 956), common epidemiologic or laboratory features in people
with CFS are regarded as optional subgrouping variables. There is no requirement
for studies to be conducted with groups of individuals who had a similar onset to
CFS, although researchers are free to pursue this approach if they wish. The wording
in the guidelines should reflect this established consensus view.
It is valid to study groups of patients who meet diagnostic criteria for CFS, but
who don't have a common identifiable cause for their illness. Valuable information
and insights can be obtained from studying such cohorts. The pathophysiology of CFS
in these groups may or may not be heterogeneous.
Observations about heterogeneity in CFS populations are based on patient symptomatology,
and not on firm laboratory evidence (see Heterogeneity sub-section below). It may
be that all people with CFS share some or all of the disease processes which contribute
to the disorder. Just because the apparent triggering factor is different for individuals
within the CFS population, it does not therefore follow that the ongoing disease
process is different.
* Inconsistent levels of evidence
There are a couple of inconsistencies in the quality of evidence ratings for dot
points in the text box titled 'Phenomena associated with CFS' on page 22, and their
counterpart ratings in Box 5.3 on page 24.
For example, the fifth dot point in the text box on page 22 says:
"Neuroendocrine changes indicating hypothalamic-pituitary axis disturbance are
common in people with CFS (Level III-4)..."
Yet in Box 5.3 under the heading 'Neuroendocrine function' it says:
"Impaired hypothalamic-pituitary-adrenal (HPA) axis activation has been demonstrated
(Level III-2)."
There is clearly a discrepancy in the ratings given for what is essentially the
same statement.
The sixth dot point in the text box on page 22 says:
"Sleep disturbance is very common in people with CFS (Level 1)..."
In Box 5.3 under the heading 'Sleep' it says:
"Disturbances of sleep maintenance (eg. frequent awakenings) are prevalent (Level
III-2)."
Other evidence under the 'Sleep' heading is rated as Level III-3. So again there
is a discrepancy between the ratings on page 22 and page 24.
These inconsistencies need to be reviewed and corrected.
* Pathogens not related to CFS
An extensive list of infectious agents has been investigated in people with CFS,
and most have been found to have no association with the ongoing illness. It would
be helpful for general practitioners to be given an idea of the range of infectious
agents that have effectively been excluded. These are in addition to the pathogens
listed in Box 5.1 on page 23 of the guidelines. In particular, the following reference
lists more than forty microorganisms that have been screened for in people with CFS
in a controlled study, and for which no association with CFS was found:
Mawle A C, Nisenbaum R, Dobbins J G, et al., 'Seroepidemiology of Chronic Fatigue
Syndrome: A Case-Control Study', Clinical Infectious Diseases 1995; 21: 1386-1389.
* Other research studies
The guidelines would be strengthened with the inclusion of additional references
relating to several areas of research. These references are discussed below. Their
inclusion would be of interest to general practitioners and others who are using
the guidelines as a resource. I recommend that Boxes 5.1-5.4 on pages 23 and 24 should
be expanded as necessary to accommodate the additional studies, and that further
boxes be added if it makes sense to do so.
Cardiac Function
Several papers have been published which examine cardiac function in people with
CFS. These are not related to the neurally mediated hypotension studies which are
mentioned in the guidelines. People with CFS often complain of chest pain or heart
palpitations (as well as fatigue of course), so these particular studies have relevance
to any discussion of the pathophysiology of CFS. The references are:
Montague T J, Marrie T J, Klassen G A et al., 'Cardiac Function at Rest and with
Exercise in the Chronic Fatigue Syndrome', Chest 1989; 95: 779-784.
Lerner A M, Lawrie C, and Dworkin H S, 'Repetitively Negative Changing T Waves at
24-h Electrocardiographic Monitors in Patients with the Chronic Fatigue Syndrome',
Chest 1993; 104: 1417-1421.
Lerner A M, Goldstein J, Chang C et al., 'Cardiac Involvement in Patients with Chronic
Fatigue Syndrome as Documented with Holter and Biopsy Data in Birmingham, Michigan,
1991-1993', Infectious Diseases in Clinical Practice 1997; 6: 327-333.
Genetics
In Box 5.5 on page 25 of the draft guidelines, genetic make-up is cited as one of
the risk factors for CFS. In the FAQs box on page 4, genetic influences are cited
as being likely to contribute to the cause of CFS. However, there are no references
given to support these statements. One published paper which does provide some evidence
of genetic influences in CFS is:
Keller R H, Lane J L, Klimas N et al., 'Association Between HLA Class II Antigens
and the Chronic Fatigue Immune Dysfunction Syndrome', Clinical Infectious Diseases
1994; 18 (Supplement 1): S154-S156.
This study should be referred to in the guidelines.
Biochemistry
Four studies looking at biochemical aspects of CFS are of general interest, and the
guidelines would benefit from their inclusion:
Lieberman J and Bell D S, 'Serum Angiotensin-Converting Enzyme as a Marker for the
Chronic Fatigue-Immune Dysfunction Syndrome: A Comparison to Serum Angiotensin-Converting
Enzyme in Sarcoidosis', The American Journal of Medicine 1993; 95: 407-412.
Kuratsune H, Yamaguti K, Takahashi M et al., 'Acylcarnitine Deficiency in Chronic
Fatigue Syndrome', Clinical Infectious Diseases 1994; 18 (Supplement 1): S62-S67.
Suhadolnik R I, Peterson D L, O'Brien K et al., 'Biochemical Evidence for a Novel
Low Molecular Weight 2-5A-Dependent Rnase L in Chronic Fatigue Syndrome', Journal
of Interferon and Cytokine Research 1997; 17: 377-385.
Burnet R B, Yeap B B, Chatterton B E et al., 'Chronic Fatigue Syndrome: Is Total
Body Potassium Important?', Medical Journal of Australia 1996; 164: 384.
Neuroendocrine function
I believe that four additional neuroendocrine studies should be evaluated and referred
to in the guidelines:
Bakheit A M O, Behan P O, Dinan T G et al., 'Possible Upregulation of Hypothalamic
5-Hydroxytryptamine Receptors in Patients with Postviral Fatigue Syndrome', British
Medical Journal 1992; 304: 1010-1012.
Dinan T G, Majeed T, Lavelle E et al., 'Blunted Serotonin-Mediated Activation of
the Hypothalamic-Pituitary-Adrenal Axis in Chronic Fatigue Syndrome', Psychoneuroendocrinology
1997; 22(4): 261-267.
Chaudhuri A, Majeed T, Dinan T et al., 'Chronic Fatigue Syndrome: A Disorder of Central
Cholinergic Transmission', Journal of Chronic Fatigue Syndrome 1997; 3(1): 3-16.
Bennett A L, Mayes D M, Fagioli L R et al., 'Somatomedin C (Insulin-Like Growth Factor
1) Levels in Patients with Chronic Fatigue Syndrome', Journal of Psychiatric Research
1997; 31(1): 91-96.
Cytokines
A surprising omission from the list of cytokine studies presented in Box 5.2 on page
23 of the guidelines is:
Bennett A L, Chao C C, Hu S et al., 'Elevation of Bioactive Transforming Growth Factor-?
in Serum from Patients with Chronic Fatigue Syndrome', Journal of Clinical Immunology
1997; 17(2): 160-166.
This was a large study which compared TGF-? serum levels in 93 people with CFS, 80
healthy controls, 46 people with major depression, 50 people with systemic lupus
erythematosus, 41 people with relapsing/remitting multiple sclerosis and 16 people
with chronic progressive multiple sclerosis. It was found that TGF-? levels were
significantly elevated in people with CFS compared with the healthy controls and
the disease comparison groups.
All studies that have examined the issue (including those of Chao et al. 1991, and
Mawle et al. 1997) have found statistically significant elevations in levels of TGF-??in
people with CFS. The level of evidence for increased concentrations of this cytokine
in people with CFS is Level III-2, as opposed to Level III-4 for other cytokines.
This is worth a mention in the guidelines.
Another cytokine study should also be referenced, since it is one of only three that
have examined cytokine release in response to exercise in people with CFS:
Lloyd A, Gandevia S, Brockman A et al., 'Cytokine Production and Fatigue in Patients
with Chronic Fatigue Syndrome and Healthy Control Subjects in Response to Exercise',
Clinical Infectious Diseases 1994; 18 (Supplement 1): S142-S146.
Symptoms
Post-exertional malaise is one of the characteristic symptoms of CFS, as reported
in Box 1 on page 1, and on page 5, of the guidelines. Malaise or post-exertional
malaise have been reported by high percentages of people with CFS in studies involving
large numbers of patients. For example, in a study of 281 people with chronic fatigue
(of whom 91% met the original 1988 CDC criteria for CFS), 79% reported post-exertional
malaise (Komaroff et al. 1996b). In another study of 217 people with postviral fatigue
syndrome, 98% reported malaise (Miller et al. 1991).
Characteristic CFS symptoms are also listed on page 8 of the guidelines, under the
heading 'How should a doctor evaluate fatigue?'. Post-exertional malaise has been
omitted from this list and I request that it be included. Similarly, Box 5.5 on page
25 would benefit from the inclusion of post-exertional malaise as a symptom under
either the immune pathway or perhaps the neuroendocrine pathway.
On page 9 of the guidelines, it is observed that:
"When adults present for medical assessment with fatigue states the most common
alternative diagnosis to consider is major depression."
Komaroff et al. 1996b found that post-exertional malaise was much more commonly reported
in people with CFS than in those with major depression (p<0.001), a finding that
also applied to other flu-like symptoms in the 1994 CDC CFS diagnostic criteria set,
including sore throat and headaches. Another smaller study also found that people
with CFS were more likely to report recent flu-like symptoms than people with major
depression (p<0.001). The reference for this latter study is: Lynch S P J, Seth
R V and Main J, 'Monospot and VP1 Tests in Chronic Fatigue Syndrome and Major Depression',
Journal of the Royal Society of Medicine 1992; 85: 537-540.
Thus it is important that at least post-exertional malaise, but preferably also sore
throat and headaches, are listed among CFS symptoms on page 8 of the guidelines.
I ask the Working Group to seriously consider whether the ability of flu-like symptoms
to help discriminate between CFS and depression should be drawn to the attention
of readers of the guidelines. Consistent with the evidence, I believe that in Box
2.5 on page 13, the symptom 'post-exertional malaise' should be included under the
heading 'Chronic Fatigue Syndrome'.
Diet
On page 4 of the guidelines, in the text box on FAQs, under the heading 'Does a modified
diet help?', it is stated that:
"There is no evidence to suggest that dietary treatment or 'megavitamin' therapy
is effective in treating CFS."
At the Cambridge conference on CFS/ME, held on April 9-12, 1990, Dr Robert Loblay
presented data from a study of food intolerance in people with CFS. Carefully designed,
double-blinded challenges were undertaken after a strict elimination diet, and the
study involved more than three hundred people with CFS suspected of having food intolerance.
This work was reported in the following reference:
Loblay R H and Swain A R, 'The Role of Food Intolerance in Chronic Fatigue Syndrome',
in The Clinical and Scientific Basis of Myalgic Encephalomyelitis/Chronic Fatigue
Syndrome (Byron M Hyde, ed.), The Nightingale Research Foundation, Ottawa, 1992.
Dr Loblay's research found that the most common reactions were to salicylates, preservatives,
glutamate and amines. Of 102 people with CFS who were followed up, 38% rated themselves
as much better or completely well after their elimination diet at the time of initial
testing. This paper estimates that food intolerance is a significant factor for 20-30%
of people with CFS. Commenting on this study in The Medical Journal of Australia,
Dr Andrew Lloyd noted:
"Elimination dietary therapy was conducted in a double-blind and placebo-controlled
trial, in that the challenges (including placebo challenges) were presented to the
patients in a random fashion generated by code; obviously the elimination diet itself
could not be administered double-blinded. The results of this study showing that
a minority of subjects do benefit therefore do partially control for a placebo effect
from the treatment programme."
The reference for Dr Lloyds comment is:
Lloyd A R, 'Muscle Versus Brain: Chronic Fatigue Syndrome (reply)', The Medical Journal
of Australia 1991; 154: 220.
In light of Dr Loblay's study, it is not justifiable to say in the guidelines that
there is no evidence that dietary treatment is effective in treating CFS. There is
evidence, and its rating is Level III-3. This is certainly stronger evidence than
exists for some of the symptomatic treatments advocated on page 3 of the guidelines.
I request that the wording in the text box on FAQs on page 4 be revised, and that
some discussion of elimination diets be included in the text of the guidelines. Dr
Loblay's study should be considered for inclusion in Box 4.1 on page 18.
Heterogeneity
It is stated in several places in the guidelines that CFS is a heterogeneous disorder.
For example, on page 12:
"In fact, as the diagnosis of CFS currently identifies a heterogeneous group
of people (Hickie et al. 1995), it is unlikely that any one diagnostic test will
emerge."
and on page 17:
"Patient cohorts in CFS treatment trials are likely to be heterogeneous because
of the relatively subjective and non-specific criteria used to make the diagnosis
(Hickie et al. 1995a)."
and on page 22:
"The recognised heterogeneity within patient groups labelled as having CFS makes
it highly likely that there are multiple contributing factors in the disorder."
The paper by Hickie et al. 1995 (titled 'Can the chronic fatigue syndrome be defined
by distinct clinical features?'), which is used to support statements about patient
heterogeneity, is contentious because it claims to be able to separate people with
CFS into two different groups based on statistical differences, using a range of
independent variables as measures. Closer examination of Table 2 on page 931 of the
paper of Hickie et al. 1995, indicates that most of the variables are not independent,
but are functions of illness severity. Hickie et al. correctly note that:
"Given the novel statistical approach used, however, the proposed solution and
our interpretations require replication within independent samples."
To date, there has been no published replication of the study of Hickie et al. 1995.
The evidence for patient symptom heterogeneity as presented in the guidelines is
only Level III-3. The real issue is whether symptom heterogeneity reflects aetiological
or pathophysiological heterogeneity. The statements in the guidelines imply that
people with CFS are heterogeneous with respect to these two factors. However, at
this stage there is no good published evidence to directly support this proposition.
Clarification of this issue is unlikely to come until biological markers for CFS
or sub-groups of CFS are identified.
The wording in the guidelines should emphasise that there is limited or preliminary
evidence for symptomatic heterogeneity in populations of people with CFS, and that
biological heterogeneity is an assumption based not on firm laboratory evidence,
but simply on reports of patient symptomatology.
Support Groups
On page 21 of the draft guidelines, this sentence appears:
"Inevitably, members of CFS support groups tend to include those with the most
prolonged illnesses (Sharpe et al. 1992)."
The paper of Sharpe et al. 1992 cannot be used to support this statement. Sharpe
et al. found a correlation between functional impairment and membership of a patient
organisation. Length of illness was not correlated with membership of a patient organisation.
On page 151 of their paper, Sharpe et al. also noted that:
"the findings do not indicate the direction of cause underlying the associations".
Also on page 21 of the guidelines, it says that:
"Depending on the nature of the groups, some may serve to increase alienation
from medical and government agencies and encourage forms of treatment that lack scientific
evaluation."
I wish to point out that I am unaware of any instance in which any CFS Society has
served to increase alienation from medical and government agencies. On the contrary,
CFS Societies maintain lists of knowledgeable and sympathetic doctors so that Society
members, and members of the general public who think they may have CFS, can be evaluated
and treated in a fair and compassionate manner. Similarly, government agencies such
as the Department of Social Security are important to people with CFS, and CFS Societies
provide information about relevant government benefits and services to people who
enquire. The above potentially damaging statement in the guidelines should be removed
unless there is comprehensive evidence to support its inclusion.
As for encouraging forms of treatment that lack scientific evaluation, I also wish
to emphasise that in its 171/2 year history, the CFS/ME Society of Victoria Inc.
(of which I am a Life Member) has never advertised, promoted, endorsed or sold any
form of treatment or therapy for CFS/ME.
Depression
On page 12 of the guidelines is the statement:
"Up to two-thirds of adults with CFS have either a prior, or concurrent, diagnosis
of major depression, as do people with fibromyalgia and irritable bowel syndrome.
By comparison, the lifetime rate of comparable depressive disorders in the general
community is 15%-25%."
It would be more helpful for general practitioners and other readers if the figures
for prior and concurrent depression in people with CFS were quoted separately. It
would be useful for GPs to have an estimate of the likelihood of current depression
in their CFS patients, but lumping prior depression in with concurrent depression
simply confuses the issue and makes the incidence of depression in CFS appear higher
than it actually is. Also, to be consistent with the presentation of data for community
lifetime depression, an upper and lower range should be quoted for the incidence
of prior and concurrent depression in people with CFS.
Readers should also be informed whether the CFS depression data come from community-based
studies, or from highly selected tertiary referral samples. Readers were told this
in relation to studies of the natural history of CFS (see page 15 of the guidelines),
and it is only reasonable that this information is also included in the discussion
on depression.
Nomenclature
On page 6 of the guidelines, nomenclature for CFS is discussed under the heading
'What other names are commonly used for CFS?'. There remains a need to address the
term 'post-viral syndrome', since some patients are still being told by general practitioners
that they have this condition. These patients are then confused about the relationship
between post-viral syndrome and CFS. Since 'post-viral syndrome' has been abandoned
as an alternative name for CFS or idiopathic chronic fatigue, this should be pointed
out to general practitioners in the guidelines. This will then help to avoid any
further confusion for people with either prolonged or idiopathic chronic fatigue,
or CFS.
References
The numerous references are the foundation of the guidelines, and are a major resource
for future readers. To maximise the validity and utility of the guidelines, it is
desirable for the reference to be updated to March 1998, when the draft is revised.
Some additional references have been suggested in this submission, but other highly
relevant papers may now be available as well.
Consumer Input
A major effort was made by the Consumer Health Forum representative in compiling
the 80 submissions originally received by the Working Group. A 'Compilation of Submissions'
document, more than 200 pages long, was provided to members of the Working Group.
Yet the draft guidelines contain very little direct information from the consumer
submissions, apart from a few text boxes. Consumers were invited to comment on a
range of issues, some of which received no coverage in the guidelines (eg. the needs
of special groups such as children and adolescents).
Effective management of people with CFS requires empathy on the part of the treating
medical practitioner. Empathy can be engendered by a better understanding of the
experiences and the needs of people living with CFS. In the context of the evidence-based
guidelines, this understanding can best be provided by including an Appendix (perhaps
three or four pages long) which incorporates and summarises material from Craig Ellis'
'CFS Health Consumer Perspective'. This should be in addition to the consumer quotes
already present in the draft guidelines.
5 Minor Areas for Improvement
The items presented in this section are listed in ascending page number order for
ease of reference. Underlining is not used in this section to highlight explicit
or implied recommendations for change since there is little discussion associated
with each separate issue, and recommendations for change are therefore easy to identify.
Page 2, Understanding the illness
It is stated that:
"Both the doctor and the patient should avoid simplistic attributions of CFS
to 'a virus', 'immune dysfunction', 'malingering', or 'mere depression'".
The words "mere depression" should be replaced with the word "depression".
As pointed out elsewhere in the guidelines, CFS cannot be attributed to major depression,
and laboratory findings in people with CFS differ from those in major depression
(see Box 5.3, page 24). The presence of the word "mere" implies that perhaps
CFS can be attributed to some other more complex form of depression, although no
evidence for this is presented in the guidelines. In the absence of any such evidence,
the word "mere" is inappropriate.
Page 3, Physical activity
It is stated that:
"In the early stages of the illness, many people with CFS make the mistake of
putting off chores or social engagements until they feel better, then pushing themselves
too hard on 'good days' to make up for lost time. The subsequent worsening of symptoms
and delayed recovery can establish a cyclic pattern of illness and disability."
It should also be noted in the guidelines that some people with CFS try to 'push
through' the illness in the early stages, and attempt to maintain a high level of
activity in defiance of their symptoms. This exacerbates their CFS and can ultimately
lead to a greater level of disability than would have otherwise occurred.
Page 4, FAQs
Under the heading 'What causes CFS?', the words "as well as genetic influences"
should be replaced with the words "as well as genetic and environmental influences",
to be consistent with the information in Box 5.5 on page 25.
Under the heading 'Is CFS a psychological disorder?', it needs to be stated clearly
that CFS is not a primary psychological disorder.
Under the heading 'Is there a cure for CFS?', the words "there are a range of
symptomatic and supportive treatments which are beneficial" should read "there
are a range of symptomatic and supportive treatments which may be beneficial."
This question should logically be placed above the question 'Will bed rest cure the
illness?'.
Under the heading 'Will exercise cure the illness?', it should also be noted that
excessive physical activity may also worsen symptoms in the medium term, particularly
if the exercise provokes a major relapse.
In and under the heading 'Are people with CFS eligible for Sickness Benefits?', the
words "sickness benefits" should be replaced with the correct words "sickness
allowance". The other benefit which people with CFS may be able to receive is
the disability support pension.
Page 5, Epidemiology box
It is stated in the second dot point in the text box titled 'Epidemiology' that:
"The prevalence of CFS in the community is 0.2%-0.5%...".
To be consistent with the data presented on pages 6 and 7, the prevalence figure
should be "0.2%-0.7%".
Page 6, What other names are commonly used for CFS?
In the first paragraph under this heading, it is stated that the term 'chronic fatigue
and immune dysfunction syndrome', or CFIDS, is inappropriate because it implies that
the cause of CFS is immune deficiency. I wish to point out that immune dysfunction
covers not only immune deficiency but also overactive immunity.
Page 6, Box 1.1
In Box 1.1, in the list of studies of the prevalence of fatigue in primary care,
the study of Hickie 1996 should be entered above the study of Wessely 1997, to be
consistent with the rest of the list which is in chronological order.
Page 9, How should a doctor evaluate fatigue?
The words "(see Boxes 3.1 and 3.3)" should read "(see Boxes 2.1 and
2.3)".
Page 11, Box 2.3
I suggest that in Box 2.3 titled 'Alternative causes of chronic fatigue', two more
alternative causes are added. Under 'Sleep disorders', narcolepsy should be included
so as to be consistent with Box 2.1. I also suggest that organophosphates be included
under 'Occupational and environmental factors', to be consistent with Box 5.4.
Page 11, How should the context of the illness be assessed?
The following sentence appears under this heading:
"For example, it may be evident that an adolescent's 45-minute walk to school
produces fatigue and other symptoms that last all day."
It is highly unlikely that anyone who can walk for 45 minutes will meet the CDC
case definition for CFS. For people who do meet the CDC case definition, a 45 minute
walk - if it could be completed - would be very likely to cause a significant exacerbation
of symptoms for much more than the rest of the day. In fact, one of the symptoms
listed in the CDC diagnostic criteria is post exertional malaise lasting more than
24 hours. It would be more suitable for a more realistic example to be used for someone
with a mild form of CFS.
It is also noted that:
"At the severe end of the spectrum of CFS, people may be housebound and experience
profound fatigue simply from the necessities of self-care such as showering or dressing."
It is important for general practitioners to be aware that in CFS the fatigue is
often not the most distressing or disabling symptom. Other symptoms such as muscle
pain, headaches or nausea can be worse at different times. I suggest that the above
sentence should read "people may be housebound and experience profound fatigue
and other symptoms from the necessities...".
Page 14, What are the benefits of a diagnosis of CFS?
The sentence:
"Making the diagnosis should mark the end of investigations to exclude alternative
diagnoses."
should be expanded to remind general practitioners that a comprehensive fatigue assessment
(which includes excluding alternative diagnoses) is required every twelve months,
as per Box 2.2 on page 10 of the guidelines.
Page 15, What is the outcome of chronic fatigue states?
It is stated under this heading that:
"One death by suicide and two unrelated deaths occurred in 2075 people followed
up in 19 published studies of the outcome of prolonged fatigue and CFS (Joyce et
al. 1997)."
In fact, the words of Joyce et al. are:
"In some studies, it was not clear whether the vital status of the non-respondents
was checked, hence the death rate could be higher. One of the known deaths was from
an unrelated physical illness, another from an unspecified cause, and the other was
by suicide."
The statement in the guidelines should be altered to correctly reflect the findings
of Joyce et al.
Page 18, Box 4.1
In the column labelled 'Intervention', the words "Interferon alfa" should
read "Interferon alpha" to be consistent with the title of the paper and
with conventional spelling.
Page 19, Reference for moclobemide study
In the discussion concerning the moclobemide trial, reference is made to a paper
by Hickie et al. 1998. No such paper has been published. The reference should be
to the abstract of Wilson et al. 1994.
Page 19, Box 4.2
The asterisk at the end of the heading 'Recommendation for treatment use' does not
have an associated footnote at the bottom of Box 4.2 (although it does appear in
the Web version of the guidelines).
Page 22, What are the gaps in our knowledge and priorities for future research?
The word 'patholological', which appears near the bottom left hand corner, should
be 'pathological'.
Page 24, Box 5.4
Under the heading 'Metabolic disturbance', the words "(McGregor et al. 1996a,
1996b)" are in a font which is too large compared with the other dot points
in the Box. Immediately under this are the words "of CFS" in large bold
type. These words should be part of the heading for Box 5.4, and they correctly come
after the word "pathophysiology" in the heading.
Page 26, How were these clinical practice guidelines developed?
The words "see Clinical Practice Guidelines Box 1" should read "see
Quality of Evidence Ratings Box on this page."
Page 32, Australian patient support groups
These groups should be listed in alphabetical order, or in order of size, but not
in the haphazard manner which is evident on page 32.
6 Conclusion
The draft guidelines are comprehensive and attractive, and will be used widely once
they are published in final form. There are, however, a number of areas in which
the draft guidelines need improvement. These have been described in this submission.
I believe that the changes required are extensive enough to warrant a second version
of the draft being made available for public comment.
The Working Group should not be concerned if any changes cause the guidelines to
be expanded by a few pages. There is a very good clinical overview at the front,
and it doesn't matter if the document as a whole is a few pages longer than the first
draft. It is important that readers are provided with up-to-date and accurate information
in a comprehensible and succinct form.
Thank you for taking the time to read this submission.
Yours sincerely
Jim Oakley
Acknowledgment: I would like to thank Bernhard Liedtke for his assistance in obtaining
some of the literature referred to herein.