National Institutes of Health
Chronic Fatigue Syndrome
State-of-the-Science Consultation
February 6-7, 2000
Bethesda, Maryland
Introduction
Summary of Recommendations by the Consultants
Areas Addressed by the Consultants
Research Opportunities
Research Strategies
Administrative Strategies
An 11-member consultation panel, chaired by Dr. Gail Cassell, met on 6-7 February 2000 to review the state-of-the-science in Chronic Fatigue Syndrome (CFS) and to suggest to NIH staff and scientists optimal approaches to CFS research.
This review is part of an NIH process intended to improve the quality, direction and extent of CFS research supported by NIH, and is conducted in a manner consistent with other NIH consultations regarding other research efforts. Prior to the meeting, the consultants received a number of published reviews and papers in various areas of CFS research. They also received a list of currently funded NIH research projects. As part of the deliberative process, the consultants presented their thoughts on the state-of-the-science in CFS and on CFS research opportunities and directions. Additionally, the consultants had a presentation from a CFS patient, Ms. Kathy Rabin. Comments and questions from members of the DHHS CFS Coordinating Committee were also sought. A list of the participants and an agenda are attached (Appendix A).
Introduction
"The chronic fatigue syndrome is a clinically defined condition characterized by severe disabling fatigue and a combination of symptoms that prominently features self-reported impairments in concentration and short-term memory, sleep disturbances, and musculoskeletal pain. Diagnosis of the chronic fatigue syndrome can be made only after alternative medical and psychiatric causes of chronic fatiguing illness have been excluded. No pathognomic signs or diagnostic tests for this condition have been validated in scientific studies; moreover, no definitive treatments exist. Recent longitudinal studies suggest that some persons affected by the chronic fatigue syndrome improve with time but that most remain functionally impaired for several years."1
In many ways, this statement from the 1994 CFS case definition still summarizes the conundrum facing the CFS research and patient communities. CFS-like illnesses have been described in the medical literature for at least 150 years; thus, this type of illness is not new. It is also not alone in medicine as a symptom-based illness of unknown etiology. Fibromyalgia, irritable bowel syndrome, and Gulf War Syndrome (GWS) are only a few of the illnesses whose etiologies and pathogeneses have puzzled the medical and scientific communities. A key addition to the quote from the 1994 CFS definition is that the duration of symptoms is not limited to 'several years', but for most patients remains a lifelong issue.
Since the initial (1988) and revised (1994) case definitions of CFS were widely adopted in the U.S. and a number of other countries, researchers have pursued several possible etiologies and markers for CFS. No pathogen nor consistent primary immunologic abnormality has emerged. A variety of treatments have been tried, with many studies involving small case series or anecdotal reports. Cognitive behavioral therapy, cognitive behavioral stress management, graded exercise, and anti-depressant medications have benefited some patients, but their overall efficacies remain to be established. Of the true clinical trials that have been undertaken, benefits have typically been limited to subsets of patients.
Initial studies of CFS prevalence indicated rates of 7-10 per 100,000. More recent community-based studies using the 1994 CFS case definition have suggested rates of 200 or more per 100,000.
Although research progress has been slow, many ideas about etiology and pathogenesis have been explored. The positive results, as well as the negative and variable results, can provide ideas for new areas of research opportunity and new research strategies. The consultants noted that many of the current strategies for preventing and/or treating diseases were developed empirically.
Summary of Recommendations By The Consultants
After discussion of the current state-of-the-science, the consultants identified the following broad areas for research opportunities that would allow the better understanding of the pathogenesis of CFS, with the long-term goal of developing treatment and prevention strategies.
- Definitive trials of antidepressants, other currently-prescribed medications, interferon, cognitive-behavioral therapy, graded exercise, etc., to determine the effectiveness of treatments currently being prescribed as well as to provide insights into pathogenesis.
- Broader clinical, epidemiological, and behavioral comparisons of CFS and similar symptom-based and psychiatric syndromes.
- Prospective studies of natural history and of incident disease.
- Major epidemiological investigations of incidence in various heterogeneous populations.
- Studies of individual dimensions of CFS.
- Development of methodological approaches for measurements of disease severity, quality of life, and outcomes.
- Better studies of sleep abnormalities.
- Larger controlled studies of the HPA axis.
- Meaningful studies of brain imaging.
- Studies of exercise to identify physiological correlates of symptoms and performance.
- Gene expression studies using microarrays.
- Search for microbial agents using differential display technologies.
- Use of animal models to explore CFS pathology including the role of neurotropic viruses and of stress pathways.
- Blinded studies of the 2'5'-oligoA pathway.
- Health services research.
- Enhancement of research infrastructure
Areas Addressed By the Consultants
A central question that confounds CFS research and treatment is whether CFS is a single entity with a single cause, a single entity of multiple or complex causality, or an endpoint that can follow from a variety of physical or mental stresses (e.g., viral infections, emotional stress).
Research Opportunities
Epidemiology
Population-based studies allow one to characterize a disease or syndrome in terms of its frequency, natural history, and spectrum of signs and symptoms. However, large studies are limited by logistical issues, cost, and accuracy of information obtained. Molecular epidemiological components can provide additional information. Thus far, gene expression assessments have not been done in an epidemiological study of CFS. (The Centers for Disease Control and Prevention (CDC) are currently undertaking a gene expression assessment in specimens from an Atlanta CFS Case Control study.)
Major epidemiological investigations of incidence in various heterogeneous populations. These types of studies would identify populations at risk for CFS. Such large-scale studies could be made more cost effective if they included a broad group of overlapping syndromes, such as fibromyalgia, in which fatigue and other CFS symptoms are a predominant feature. Although difficult to do, studies of predisposing factors would be important to include. The issue of chronic life stresses also needs to be explored as a risk factor. Cells and serum derived from such large-scale studies could be studied by a multidisciplinary range of scientists.
Prospective studies of natural history. Information about persons who fully recover, partially recover, or whose conditions worsen is needed to better understand CFS and to develop treatment and prevention strategies.
Broader clinical, epidemiological, and behavioral comparisons of CFS and similar syndromes. Cross-comparisons of syndromes with symptoms overlapping those of CFS could provide insights into these illnesses individually and as a group.
Fatigue and its role in co-morbid conditions is also important. For example, persons with cancer and fatigue do worse than those without fatigue.
Pathogenesis
Search for microbial agents using differential display technologies. Although the evidence to date for a microbial cause of CFS is not strong, and the epidemiology is inconsistent with an infectious cause, newer differential display assays may allow the search for rarer microbes. Such studies would need incident case patients and thus interlink with the recommendation for investigations of CFS incidence. While the 6-month illness period of the case definition appears to be a time period that does clearly distinguish CFS from other types of fatigue, it means that an organism that triggers disease and then disappears will probably not be detected. Organisms whose persistence or expression perpetuates illness should remain, and presumably their nucleic acids could be detected. It is also noted that while it seems unlikely that infectious agents cause CFS, they may well be among a variety of stressors that precipitate it.
Gene expression studies using microarrays. Immune function abnormalities are an area where studies of the nature and degree of changes have yielded highly variable results. Microarray technology would allow studies of an entire regulatory network of cytokines, neurotransmitters, etc. Previous studies have of necessity been limited to the study of a few molecules at a time.
Biomarker and diagnostic studies are often considered "shotgun" approaches or "fishing expeditions". However, they are worthwhile if they can be done rapidly on appropriately controlled samples and using a small number of specimens. Nervous system markers may be important in this respect. Biological plausibility is an important consideration for selecting markers for these types of studies, as broad assays of multiple markers are bound to turn up significant associations by chance alone. This has already been problematic with immunologic studies of CFS.
Meaningful studies of brain imaging. If brain imaging studies are to be pursued, they must be hypothesis-based. Patients need to be carefully selected and patient factors such as medication use must be carefully considered and controlled for. In general, medication usage should be avoided, as it will confound the results. As an example, comparisons could be made between persons with CFS who do and do not have depression. Comparisons with healthy volunteers are not worthwhile. It was noted that studies with CFS-affected and unaffected twins did not identify any distinctions between the twin with CFS and the one without CFS. Approaches used in studying mood disorders may be helpful as models for CFS imaging studies. Ligands are available to study particular brain functions.
Studies of sleep abnormalities. Previous studies in this area have not been well controlled, and confounding factors such as medication use may not have been considered. The consultants noted that there is reciprocity of daily function and sleep. Thus, how one functions during the day affects how one sleeps, and how ones sleeps affects how one functions during the day. Function during sleep may hold clues in terms of CNS dysfunction. The issue of excess arousal during sleep may provide clues as to who may develop full-blown CFS. Additionally, the autonomic nervous system can be studied and probed for evidence of enhanced daytime or nighttime arousal, by such measures as hormone profiles over a 24-hour period. Anecdotally, there appear to be CFS patients whose circadian rhythms are disrupted and who appear to be helped when the rhythm is normalized.
Larger controlled studies of the HPA axis. Many CFS symptoms could be explained by changes in HPA axis function or regulation. The HPA axis has been well-studied in mood disorders and so there is a body of experience and published literature to draw on. Further studies of the system can lead to testing of hypotheses about CFS pathogenesis.
A number of approaches can be used to probe the stress axis mechanisms at all levels using hormone manipulation challenges, through waking and sleep, and with environmental manipulations, e.g., constant routine (to determine circadian and homeodynamic components). The role of hypocortisol as a vulnerability factor for disorders can be investigated, as can the effects of HPA activation patterns (e.g., hypocortisol) on neurobiology and behavior.
Studies of exercise to identify physiological correlates of symptoms and performance. Thus far, the only physiological differences detected in CFS patients are subtle. Advantage may be taken of the "CFS crisis" to try to study physiological factors. This is based on the hypotheses that whatever is abnormal in CFS would be more abnormal in a crisis situation, and that whatever precipitated CFS initially may initiate a crisis. While the ethical issues related to formal studies inducing CFS crisis are acknowledged, the etiology of the CFS crisis remains an important and understudied part of CFS.
Additional studies, such as those of Peter White and his colleagues2, which use graded exercise for the management of CFS, may provide additional insights on CFS patient management.
Studies of individual dimensions of CFS. A dimensional approach would consider symptoms, disability, beliefs, physiological abnormalities, mood disturbance, etc., as dimensions of illness. The research would then focus on the subgroups defined by this approach. These types of studies might, for example, explore fatigue versus effort.
Use of animal models to explore disease pathology including the role of neurotropic viruses and stress pathways. There are infectious agents that can result in neurological damage which is not seen for many years, and for which little or no residue of the agent may be detectable. In exploring the potential role of neurotropic viruses in disease precipitation, one needs to consider the natural history of the virus, e.g., the cell tropism or brain regions targeted, and then whether CFS symptomatology is concordant with these tropisms.
There are several animal models of chronic stress that could be used to characterize the stress activation pattern, link it to symptoms/illness, and to determine the contributing factors to this pattern.
Treatment
Development of methodological approaches for measurements of disease severity, quality of life, and outcomes. These methods are needed for better design and assessment of patient studies, whether epidemiological, pathological or treatment.
Definitive trials of antidepressants and other currently-prescribed medications, interferon, cognitive behavioral therapy, graded exercise, etc., to determine the effectiveness of treatments currently being prescribed as well as to provide insights into pathogenesis. For many patients, treatment is currently empirically guided and based on individual beliefs. It is important to undertake well-designed trials in order to demonstrate efficacy; there is merit in such studies even if the trials are negative, as such results would encourage patients to avoid taking ineffective and at times costly medications. Trials need a multidisciplinary approach and need input from neurobiology and the mental health disciplines.
As for any chronic illness, multicomponent rehabilitation strategies are useful for CFS. Strategies need to be demonstrated to work, need to be refined, and made acceptable and deliverable. Rehabilitation may also need to be specific for individual patients and targeted to their particular symptoms. Strategies can target problems such as fatigue, depression, pain, and/or abnormalities such as cortisol levels. Strategies need to consider the heterogeneity of patients and the multiple perpetuating factors including illness validation, beliefs, behavior, and the role of interpersonal and iatrogenic factors.
Other important insights that can develop from these trials include: characterizing individuals who respond to particular types of treatment; understanding the placebo effect and characterizing those individuals who are responsive to placebo; limited hypothesis testing; and the development of new hypotheses.
Blinded studies of the 2'5'-oligoA pathway. This remains an area of controversy and inconclusive studies. The availability of newer forms of Interferon-( may provide an opportunity to explore this area.
Infrastructure Enhancement
Health services research. The process of seeking health care for a symptom-based disease may in and of itself contribute to morbidity. Understanding this process and developing strategies to identify problems at the community and primary care level may provide for better patient outcomes and more effective delivery of complex services to those in need of them. There is a need to identify currently available useful therapies and develop innovative approaches to bring these treatments to people, even if these treatments are not cures.
The CFS case definition requires six months of illnesses; many cases are not seen until long after, often years after, the onset. The value of early management is not known, nor is it known whether treatment before six months is of value.
Enhancement of the research infrastructure. There is a need to increase the critical mass of CFS investigators and especially to involve investigators from different disciplines. Among the barriers to CFS research are money, acrimony (from patient support groups toward individual investigators), as well as methodological barriers (need for pre-morbid samples, need for subgroup studies, case definition issues).
Research Strategies
Future research strategies could employ two divergent tactics, broadening and/or narrowing of populations and clinical factors to be studied, depending on the goal of the research. Both of these approaches may provide new insights into CFS. Such strategies include the broadening of studies to include other conditions with overlapping symptomatology and unexplained etiology such as fibromyalgia and irritable bowel syndrome. The narrowing approach would involve studying particular dimensions of illness, such as fatigue, in more detail, for example by comparing fatigue versus effort, and by studying that dimension over a spectrum of illnesses, including those of known etiology in addition to CFS.
Patient Populations: A number of approaches and issues related to patient populations were raised.
It is important to distinguish between the research case definition that is intended to help design research studies and the identification of individual patients at the community and clinical level. The case definition was developed by physicians who see patients and by the characteristics of those patients who are seen by physicians. They are therefore likely to be highly selected and unrepresentative of the true spectrum of illness.
Empirical case definitions using information from people in population surveys may be derived from factor analysis, e.g., data from persons studied in population surveys. Factor analysis may be a useful exploratory tool for further refining the case definition. However, lacking any biological marker, there is no assurance that any complex of symptoms corresponds to a biological state.
Countries such as the United Kingdom having socialized medicine may provide a good milieu for cohort studies. National birth cohort studies as well as DOD-GWS studies may allow for the study of a whole population. The DOD studies of GWS may also allow for the identification of persons at high risk. It was noted that long-term follow up of persons with GWS reveal that some go on to meet diagnostic criteria for CFS. It is a cardinal premise of epidemiology that case-control studies should enroll incident cases. With CFS this is difficult to do, as many patients have already had the disease for quite a while and are from tertiary medical care. In any event, the case definition requires at least six months of fatigue.
In persons with chronic unwellness, there seem to be two categories of predominant symptom patterns that can be broadly classified as cognitive and inflammatory (although there is no definitive evidence of inflammation).
Some studies (e.g., neuroimaging studies) may benefit from very homogeneous and highly screened patient populations. Other studies, such as population-based incidence studies, may benefit from inclusion of overlapping illnesses such as fibromyalgia and irritable bowel syndrome.
CFS could be combined with overlapping syndromes and the multidimensional aspects studied from physiological and psychological perspectives, e.g., studies of the HPA axis and the role of stress.
Subdivisions of CFS populations could yield important insights. Thus, comparing persons with rapid onset of CFS to those with insidious onset may be helpful. Duration of illness may be another important factor. The patient subgroups selected should be relevant to the hypothesis to be tested.
Comparison groups for CFS studies should also be carefully selected. Persons with other illnesses such as rheumatoid arthritis may be more appropriate than, or used in addition to, healthy persons.
Studies of fatigue, even in groups where there is an underlying cause such as cancer or hepatitis C, could provide new insights for CFS.
Mechanistic Issues: In designing studies of CFS, a number of mechanistic issues and approaches were discussed.
Fatigue and sleep disorders may be the result of the two predominant symptom patterns, cognitive and inflammatory. Thus, requiring fatigue first could be approaching the issue backwards.
Stress reactions and immune regulation are more sensitive in women than in men. Women's physiology is more cyclical in nature. These factors need to be taken into account in study design.
The idea of a chronic stress activation pattern may be worthy of pursuit. There is much evidence for a stress component to CFS. Additionally, a range of conditions associated with stress or trauma has been related to hypocortisol. Thus, there is considerable symptom overlap in PTSD, nurses and "burnout," teachers who report living under chronic stress, CFS, FMS, and idiopathic chronic pain.
Predisposing factors, precipitating factors, and perpetuating factors need to be distinguished and studied. Understanding perpetuating factors and factors predicting recovery are important issues. Iatrogenic factors, such as medical treatment, involved in perpetuation or recovery need also to be considered.
There is a need to accept multifactoriality in the cause of illnesses such as CFS. Thus, a heart attack is a clinical entity. It may be an end result of a complex interaction of multiple risk factors such as smoking, hypertension, obesity and hypercholesterolemia.
When physical measures/illness factors are defined categorically, rather than dimensionally, an arbitrary categorical line must be drawn in which, e.g., individual are classified only as healthy vs. ill. Accepting heterogeneity and a disease continuum rather than imposing binary divisions may be helpful in studying CFS.
Studying physical measures/illness factors dimensionally may give a different perspective from classifying people into distinct syndromes. Differentiating dimensions such as type and number of symptoms, chronicity, disability, coping behavior, etc., may provide insights.
A focus on fatigue as a symptom of various illnesses, and on the mechanisms of fatigue, may provide information on specific illnesses as well as on general processes.
The physical versus psychiatric split has also been harmful in studying CFS (and illness in general). Both physical illness and psychiatric illness are real and based on symptoms. Moreover, scientists understand that every psychiatric syndrome must have an undying biological basis, even though none has been identified. Psychiatry is a discipline that historically deals with symptom-based illnesses such as CFS, which are experienced physically by patients but for which diagnostic markers have not been found.
Although making psychiatric diagnoses in CFS is problematic because of the uncertain etiology of symptoms, some possible bias in symptom presentation, and the inherent limitations of the categories themselves, research and clinical experience suggests a higher prevalence in CFS of depression, generalized anxiety disorder, and panic disorders than in other ill groups. Data suggest that the majority of CFS patients may have diagnosable psychiatric illnesses. But it is difficult to show whether these psychiatric conditions are primary, or secondary to CFS. Thus, processes related to these disorders and to somatization disorder should be considered for study-design and hypotheses development, as opposed to focusing on specific diagnoses (e.g., whether CFS is or is not a form of depression).
Beliefs about illness should be explored as an aspect of CFS; psychological and sociological factors affecting individual patients should be considered.
Development and evaluation of pragmatic treatments that can cope with heterogeneous problems independent of their etiology, such as cognitive behavioral therapy, may be of value. This type of strategy has been effective in the management of chronic pain.
Objectively measurable CFS abnormalities identified thus far have been variable and subtle. Studying the CFS crisis could provide insights into disease pathogenesis. This would include understanding factors that produce a crisis, predict the onset of a crisis, and define a standard crisis response. The concept of overtraining in athletics may also provide insights. (It was noted that the tilt-table test or even simulating its effect by having patients stand against a wall for a long period of time, might be useful to this end. For many patients, over-exertion will precipitate a crisis.)
Studies of sustainable activity levels in CFS may provide insights into physiological abnormalities. This is because the primary symptom of CFS is 'chronic fatigue', which is best characterized as a large reduction in the level of sustained activity. Changes in levels of sustained activity might serve as a good measure of CFS severity, as well as to evaluate the efficacy of interventional studies.
Methodological/Study Design Issues: A number of issues related to study design were put forward.
Studies should be hypothesis-driven and well-designed.
Measurement and study design must consider not only the symptoms, but also illness co-morbidities and appropriate comparison groups. Diagnostic and dimensional rating instruments are important.
In addition to the clear definition of patient populations, research methods and procedures (whether laboratory or survey-based) should have standard operating procedures and be reproducible. Measurement instruments should be shared and should include not only the items investigators wish to measure, but also items that are important to patients. Scales used in neurological studies may be helpful.
Measurements of treatment outcomes are especially important.
Timing is important when developing study design. Many trophic hormones peak at night. A number of features of host defense function on a circadian pattern.
In terms of specimens to be collected from patients, peripheral blood appears reasonable since CFS is systemic.
Correlation of degree of disability with changes in physiological markers should help define markers in which changes, even if subtle, might have physiological meaning. Subtle changes do not appear likely to reflect etiology.
Study results should be not only statistically significant, but must be biologically relevant.
Multidisciplinary approaches and collaborations should be encouraged and facilitated. The involvement of more neuroscience investigators and inclusion of psychiatric expertise is particularly important for studies.
Administrative Strategies
Information Issues
NIH needs to make the patient community and public better aware of how studies of physiological processes in normal persons (such as sleep and exercise) as well as other illnesses with overlapping symptomatology (such as fibromyalgia) provide a knowledge base for understanding CFS, and how study of CFS within the context of psychiatric disorders of unknown etiology is of critical importance.
HMO's need education about CFS and how it affects its service population. (HCFA is trying to develop CFS information for practitioners in order to get earlier case identification and treatment.)
Funding/Infrastructure
Funding mechanisms for CFS should address the need for multidisciplinary/trans-institutional research and the sharing of materials and reagents. For example, the CDC will be continuing to collect patient and control specimens during their epidemiological studies. While CDC does not have an R01 mechanism, the materials that they have collected could be used by NIH grantees for CFS studies.
Mechanisms and approaches to address the barriers for CFS research while maintaining peer review standards need to be developed. Making scientists aware of the fascinating issues that can be studied in this patient population may increase the interest of investigators from other fields and provide prestige and status to CFS research.
Infrastructure issues to support research, such as centers for excellence, and training of researchers, e.g., via training grants, need to be developed and/or enhanced.
Funding for systematic reviews of the literature should be considered. (This is done by organizations such as the AHRQ Evidenced-Based Practice Centers throughout the U.S.) One topic that might be reviewed is infectious agents that affect the CNS.
1Fukuda KF, Straus SE, Hickie I, et al. The chronic fatigue syndrome: a comprehensive approach to its definition and study. Ann Intern Med 1994;121:325-331.
2Fulcher and White. Randomised controlled trial of graded exercise in patients with chronic fatigue syndrome. British Medical Journal 1997;314:1967-52.
