Abstract. Evengård B, Schacterle RS, Komaroff AL (Karolinska Institute at Huddinge University Hospital, Huddinge, Sweden; and Brigham and Women’s Hospital, Harvard Medical School, Boston, USA). Chronic fatigue syndrome: new insights and old ignorance. J Intern Med 1999; 246: 455–469.
Chronic fatigue syndrome (CFS) is a condition characterized by impairment of neurocognitive functions and quality of sleep and of somatic symptoms such as recurrent sore throat, muscle aches, arthralgias, headache, and postexertional malaise. A majority of patients describe an infectious onset but the link between infections and CFS remains uncertain. Findings show an activation of the immune system, abberations in several hypothalamic-pituitary axes and involvement of other parts of the central nervous system. The origin is bound to be complex and it may well be that the solution will come together with a more generally altered view about mind–body dualism, and the concept of illness and disease.
The first descriptions in the medical literature of the illness now known as chronic fatigue syndrome (CFS) probably appeared during the middle of the 19th century, although some argue that similar illnesses were described as early as the 17th century . In the 19th century, an American neurologist named Charles Beard popularized the diagnosis of ‘neurasthenia’, and his descriptions of the illness bear a strong resemblance to what we call CFS today . Neurasthenia was one of the most frequently diagnosed conditions in medical practice during the late 19th century [3, 4]. However, the increasing sophistication of psychiatric diagnoses (such as the emergence of obsessive–compulsive disorder, anxiety, neurosis, and hysteria)  was accompanied by a decreased frequency of the diagnosis of neurasthenia. The outbreak at the Royal Free Hospital in 1955 [6, 7] was retrospectively regarded as a case of mass hysteria by McEvedy & Beard , but their conclusion has caused many objections and has been rejected . Thus, the issue of the cause of CFS has been one of controversy.
Since that diagnostic label fell out of favour in the early 20th century, other illness labels have been applied to similar conditions such as epidemic neuromyasthenia, myalgic encephalomyelitis, Icelandic disease, Royal Free disease, post-viral fatigue syndrome, and chronic mononucleosis.
Recently the Royal Colleges of Physicians, Psychiatrists, and General Practitioners in the United Kingdom recommended dropping the diagnostic term myalgic encephalomyelitis or ME, widely used in the United Kingdom in recent decades, because there is, so far, no recognized pathology in muscles and in the central nervous system – as is implied by the term ME. Instead, the term chronic fatigue syndrome was recommended, as defined by the United States’ Centers for Disease Control and Prevention (CDC, Atlanta, GA, USA)  and is today the most commonly used name. The current case definition is summarized in Figure 1. This definition stresses the appearance of somatic symptoms. No objective measurements are used – only symptoms reported by the patient. With the 1994 definition, consensus was reached on which tests would be necessary before making the diagnosis: complete blood count, erythrocyte sedimentation rate, alanine aminotransferase, total protein, albumin, globulin, alkaline phosphatase, calcium, phosphate, glucose, blood urea nitrogen, electrolytes, creatinine, thyroid stimulating hormone and urinalysis. Less common, and more expensive, testing such as lymphocyte phenotyping and neuroimaging studies are not necessary, although they have been used in research studies to help understand the biology of CFS.
Over the past decade, a good deal has been learned about the biological underpinnings of this illness, although its cause remains unknown.
Clinical picture and epidemiology
Although fatigue is a human experience, and although many people report experiencing chronic fatigue intermittently, CFS is quite different and quite unusual. Fatigue is a very common symptom in community-wide surveys performed in different countries in the Western world . In the Swedish study from Lundby  the lifetime prevalence of chronic fatigue (defined similarly to neurasthenia) was 33% for women and 21% for men. Whether there has been a real increase in fatigue prevalence has not been established, but in a study from 1960 only 9% of out-patients reported fatigue as a principal complaint, whilst some 30 years later the prevalence at two American medical clinics was 33% [13, 14]. An editorial  underscored the substantial impact of ‘minor’ illness symptoms, such as fatigue, on lost productivity and revenue. A large, population-based study in the United Kingdom found that 30% of women and 19% of men reported experiencing a persistent fatigue for at least the past month .
CFS, on the other hand, is less common and varies with the case definition used and the population under study. Studies of community-based populations, irrespective of whether people have sought medical care for any condition (including fatigue), in the U.S. have come to similar conclusions. In a community-based prospective study in the greater Seattle, Washington area  questionnaires were sent by mail to 4000 people, and follow-up testing was performed on potential chronic fatigue cases. The estimated prevalence of CFS was found to be 0.07–0.2%, whilst the prevalence of chronic fatigue alone was much higher at 1.8–6.3%.
The prevalence of CFS in populations seeking medical care is much higher than in community-based samples. In one U.S. study, the prevalence was found to be about 1% amongst patients seeking medical care for any reason (not just seeking care for fatigue). In a primary care practice study in the U.K., Wessely et al.  found the point prevalence of chronic fatigue was 11.3%, falling to 4.1% if patients with comorbid psychological disorders were excluded. The point prevalence of CFS was 2.6%, falling to 0.5% if patients with comorbid psychological disorders were excluded.
Whether CFS shares overlapping symptoms with other syndromes which include fatigue as a symptom (e.g. fibromyalgia and multiple chemical sensitivities) has not been well-characterized. One study , however, reported significant demographic and symptom overlap.
In any illness defined by a group of symptoms, two questions arise: do the patients with the illness in fact report the symptoms that investigators say they should, and do those symptoms distinguish patients with CFS from patients with other fatiguing illnesses. A recent study from the CDC reports that when a very large population of individuals was asked about a variety of chronic symptoms, there did indeed emerge a group of individuals who spontaneously described the symptoms of CFS . Another study compared the reporting of CFS symptoms by patients with CFS, by patients with other fatiguing illnesses, and by healthy control subjects. The symptoms of CFS, but not other ‘control’ symptoms, were reported much more frequently by patients with CFS than by patients with other diseases that produce chronic fatigue, such as multiple sclerosis and major depression, and by healthy subjects .
Most reports from hospitals or specialty practices that care for patients with CFS have found a higher fraction of women than men: about 70% women. This female predominance must be placed in context: about 60% of visits to doctors for all reasons are by women, and many diseases that produce fatigue have a predominance of females that is considerably greater – such as multiple sclerosis, systemic lupus erythematosus.
Moreover, when the prevalence of CFS is studied in the population at large, irrespective of whether they have sought medical care, it is almost as probable that men will meet the criteria for CFS as women [22, 23].
There are also differences in the socio-economic levels of patients with CFS who seek medical care, versus those identified in community surveys. As with female gender, patients who seek medical care for CFS may be of slightly higher socio-economic level, and are less often non-Caucasian. However, in community-based studies in the US, there appears to be a higher prevalence of CFS in people of lower socio-economic groups, and in African-Americans and Latino populations [22, 23].
CFS and psychiatric disorders
Some of the symptoms of CFS are commonly seen in patients with nonpsychotic psychiatric disorders, including fatigue, sleep disturbance and cognitive difficulties. For this reason, and because no objective diagnostic test for CFS has been found, some physicians believe that CFS is a psychiatric disorder. Whereas this belief was extremely common a decade ago, it is less common today, as more research has been conducted, and as physicians gain more experience in caring for these patients.
The question of whether the diagnosis of CFS should be applied in patients with comorbid psychiatric disorders has been hotly debated. In the current case definition, bipolar disorder, ‘melancholic depression’, and several comorbid psychotic disorders disallow the diagnosis of CFS, whereas major unipolar depression, dysthymia and various anxiety disorders do not. However, it is recommended that investigators segregate patients with and without comorbid psychiatric illness from other patients in analysing data, to be sure that the variables being measured do not reflect confounding by these psychiatric disorders.
The prevalence of ‘allowable’ psychiatric disorders amongst patients with CFS is also an area of controversy [9, 24–28]. Many investigators have tried to distinguish the presence of psychiatric disorders that were present before the onset of CFS, versus those that occurred after the onset of CFS and could be viewed as ‘reactive’ psychiatric illness in response to a chronic illness. Although there is some variability amongst studies, the following summary is a fair reflection of the literature: it is more probable that patients with CFS will have experienced at least one episode of major depression in the years before the onset of CFS than the typical member of the population; nevertheless, the majority of patients with CFS do not have such a history. Nearly half of patients with CFS develop depression in the years after the onset of CFS, but again a sizeable fraction do not. The frequency of panic disorder may also be higher in patients with CFS, both before and after the onset of CFS, but the majority of patients with CFS have never experienced that disorder.
One important distinction between major depression and CFS is that guilt, lack of self-esteem, and self-blame – core features of major depression – are much less frequent in patients with CFS . Instead of being apathetic and hopeless, as are most patients with major depression, patients with CFS are often frustrated, angry and eager to try anything to make them healthy.
One indirect indication that CFS is different from major depression is that – in contrast to patients with major depression – patients with CFS have very different scores on the widely used instrument that measures self-reported functional status, the Medical Outcomes Survey 36-Item Short-Form Health Survey (SF-36). On this instrument, CFS patients – in comparison to patients with major depression – showed significantly greater impairment in physical functioning, role limitations due to physical health problems, bodily pain, general health perceptions, vitality, and social functioning, and significantly higher scores for mental health and role limitations due to emotional health problems .
Several objective tests also seem to distinguish patients with CFS from those with major depression. Demitrack et al.  found that the up-regulation of the hypothalamic-pituitary–adrenal axis that is often present in patients with major depression was not seen in patients with CFS: indeed, patients with CFS had a down-regulation of the axis, with a central underproduction of corticotropin-releasing hormone (CRH) that leads ultimately to a slight hypocortisolism. Also, patients with CFS have an opposite response to patients with major depression in prolactin release after central stimulation [31–33]. Finally, in a comparison of SPECT imaging of the brain, the midcerebral uptake index (an objective measure of global radionuclide uptake) was significantly lower in CFS patients than in patients with major depression or healthy controls, and was similar to that seen in patients with AIDS encephalopathy .
Is CFS a post-infectious condition?
Many patients refer to an infectious type of onset with a flu-like illness. In the beginning of this century several authors pointed out that neurasthenia could follow infections such as influenza [35–38]. Since then different microorganisms have been postulated to play a role in the pathogenesis of CFS. However, the link between infections and CFS remains uncertain.
CFS following acute infections
There are a few reported cases in which CFS has followed in the wake of a well-documented acute infection – as contrasted to the nonspecific ‘infectious-like’ symptoms that many patients report at the onset of the illness. CFS can follow acute infectious mononucleosis, which is typically caused by Epstein–Barr virus, properly diagnosed and treated Lyme disease [39–41], and Q fever . Whilst these case reports provide strong evidence that CFS occasionally can be triggered by an acute infection with an agent (viral or bacterial) that has the capacity to produce a chronic infection, it is unclear what they may tell us about the aetiology of most cases of CFS. The role of these infectious agents in causing the ongoing, chronic illness remains unclear.
Acute infection with other infectious agents can be followed by CFS. These agents include other bacteria (Brucella) , rickettsial organisms and protozoal infection (toxoplasmosis and giardiasis) .
Apparent epidemics of CFS
Many apparent epidemics of an illness much like CFS have been reported in the medical literature of the past 70 years. The outbreaks have occurred all over the world [44, 45]. Some have occurred in hospitals – at the Los Angeles County Hospital in 1934  and at London’s Royal Free Hospital in 1955 [6, 7]. One of the latest reported outbreaks was in the Lake Tahoe/Incline Village area of Nevada in the late 1980s [47–49].
Reactivation of latent infections and CFS
The most recent wave of interest in CFS began in the mid-1980s with reports that antibody profiles indicated a chronic reactivation of latent infection with Epstein–Barr virus (EBV) [50, 51]. Whilst subsequent studies have largely found similar evidence of reactivation of EBV, the current general consensus is that rather than indicating a primary role for EBV in producing the symptoms of CFS, this more probably reflects a state of chronic immune dysregulation.
Human herpesvirus-6 (HHV-6) is activated more often in patients with CFS [49, 52–54]. HHV-6 is an attractive potential pathogenic agent in CFS, since it has a remarkably wide tissue tropism: several peripheral mononuclear white blood cells, respiratory and intestinal epithelial cells, fibroblasts, and central nervous system cells. Moreover, it is now clear that HHV-6 can produce clinical encephalitis and is associated with demyelinating diseases in immunosuppressed individuals and with multiple sclerosis in immunocompetent individuals [55–57]. However, like EBV, infection with HHV-6 is ubiquitous, and it is possible that the reactivation of HHV-6 in patients with CFS is a secondary event, reflecting immune dysregulation. There is as yet no compelling evidence that HHV-6 produces the pathology that leads to the symptoms of CFS.
Retroviruses have been reported to be involved , although this has not been confirmed . Another type of virus which has drawn a lot of attention is the enterovirus group [60–62] because clinical symptoms in CFS are similar to those found in enterovirus infections. So far, there is no clear evidence for an active involvement in studies based on either serology or muscle biopsies [60, 63–66]
Borna disease virus (BDV) is a newly classified neurotropic, negative-stranded RNA virus that infects warm-blooded animals causing profound neurological abnormalities [67, 68]. The infection results either in a meningo-encephalitis with neurological symptoms or in a mild persistent infection with cognitive and mood disorders in affected animals. Whether BDV naturally infects humans to cause neuropsychiatric disease remains controversial, however, there are now reports suggesting an involvement of BDV in CFS [69, 70]. This has not been the finding of others [70a] leaving unproven a BDV correlation with CFS.
Indirect evidence of a chronic viral infection in CFS
Patients with CFS may have an abnormality in an antiviral lymphocyte enzyme system called the 2–5 A pathway: the pathway appears to be chronically activated in patients with CFS [71, 72]. A case-control study, involving more than 700 patients from two different geographical areas, found a much higher frequency of atypical lymphocytes in patients with CFS than in healthy control subjects . These studies indirectly suggest the presence of a chronic viral infection, but are far from constituting proof of such an infection.
Thus, there are findings from blood testing consistent with involvement of an infectious agent. The exact nature of this involvement (e.g. a triggering event) remains an open question. So far, no microbes have been isolated and correlated to symptoms.
A growing number of immunological studies have been conducted, by different laboratories, on different groups of patients. Not all of the studies come to the same conclusions, although some common themes are emerging. One problem may be that, as the CDC criteria now used for diagnosis was established in 1994, after many of these studies were conducted, comparable patient populations may not have been studied.
Lymphocyte phenotyping studies
Landay and coworkers  reported a decreased percentage of CD8+/CD11b+ suppressor T cells and an increase in CD8+/CD11b-cytotoxic T cells. A marked activation in severely ill patients indicated by increased levels of CD38 and HLA-DR was also found. A decreased CD4+/CD8+ ratio has been reported as differing significantly from control groups [75, 76]. Also, a decrease in CD4+/CD45RA+ cells (naïve helper T cells) was reported in both studies and Straus’ group noticed a rise in memory cells (CD8+/CD45RO+). One study found an increased number of B cells . Reports of the total number of natural killer (NK) cells have varied, as have reports of the fraction of NK cells bearing activation antigens
Lymphocyte functional studies
Perhaps the most robust immunological finding in patients with CFS is the depressed function of natural killer cells [78–83]. Although not specific for CFS, depressed NK cell function has been consistently seen. The clinical implications of this finding remain uncertain. NK cells are thought to play a central role in defending against viral infections.
Another reasonably consistent finding in patients with CFS is that T cell responses to mitogens and specific antigens are depressed [75, 76, 84, 85]. Also, a depressed delayed-type hypersensitivity reaction has been described .
Serologic markers of immune activation
Other markers of immune activation also have been found: elevated levels of circulating immune complexes, and immunoglobulin G, and higher frequencies of various autoantibodies [73, 86]. It is not clear that these abnormalities have any relationship to the symptoms reported by patients with CFS .
An attractive but unproved hypothesis is that a state of chronic immune activation could lead to the production of cytokines that disrupt neurotransmitter function and result in the symptoms of CFS. Cytokines are a possible link between a possible infectious agent and CFS symptoms. Treatment with cytokines such as interleukin-1 (IL-1) can cause symptoms similar to those of CFS. Extensive studies searching for cytokines in serum, however, have not been conclusive. A recent publication shows increased levels of TGF-β in serum from patients compared to healthy controls and other patient comparison groups . An increase in this cytokine was reported earlier by Chao et al.  who also described an increase of neopterin, a marker of macrophage activation. Increased serum levels of IL-6 and IL-1α also have been reported [90, 91]. When spontaneous or stimulated cytokine release was measured from peripheral cells of patients, spontaneous release of TNF-α and IL-6 was increased, whilst both spontaneous and stimulated release of IL-10 was suppressed .
It is well known that some female patients have symptoms correlating to their menstrual cycle. Cannon and coworkers  showed an abnormality in IL-1β secretion in patients that may be related to altered sensitivity to estradiol and progesterone. They also showed increased release of IL-1Ra and sIL-1RII by cells from patients suggesting a low-level activation of the immune system.
Additional evidence for activation of the immune system is the up-regulation of the 2–5 A synthetase/RNase L antiviral pathway, an enzyme produced in leucocytes and induced by interferon, found in patients [71, 72]. It leads to increased RNase levels and decreased cell metabolism.
As muscular symptoms are a substantial part of the syndrome, researchers have looked for pathophysiological explanations, but no underlying pathology in muscle tissue has been identified . CFS patients show normal muscle physiology before and after exercise . In this study, the degree of perceived exertion was significantly greater, in relation to the increase in heart rate, during exercise in CFS patients. This has also been shown by Lloyd et al. . Perception of effort thus appears to be different in these patients.
Central nervous system abnormalities
Since many of the symptoms of CFS – fatigue, disrupted sleep, cognitive difficulties, pain – suggest involvement of the central nervous system (CNS), many investigators have looked for objective evidence of abnormalities. In our view, the evidence of CNS pathology is substantial.
Magnetic resonance imaging [MRI] demonstrates white matter abnormalities more often in patients with CFS than in healthy control subjects [49, 97]. Often, subcortical areas are involved, and sometimes deeper structures. No studies correlating tissue with MRI findings have been performed. Based on animal studies, such MRI abnormalities probably represent areas of inflammation and/or demyelination.
Single photon emission computed tomography [SPECT] abnormalities, indicating either defects in perfusion or CNS cellular function, have also been found more often in CFS than in healthy control subjects or patients with depression [34, 98]. Hypoperfusion of the brainstem has been reported , but technical objections to the evaluation of these findings have been raised . Studies using these techniques in a normal population, as well as in other patient groups having impaired cognitive function, have not been widely conducted. Therefore, the optimal use of these techniques in CFS patients is unknown.
Autonomic nervous system testing – such as vertical challenge on a tilt-table – has demonstrated abnormalities of the sympathetic and parasympathetic systems [101, 102], and a picture like that seen in neurally mediated hypotension and orthostatic tachycardia. Physical deconditioning does not appear to explain the abnormalities, nor does a coexistent depression . Studies are underway of treatments commonly used for neurally mediated hypotension – such as increasing salt intake, fludrocortisone, or beta-blockers.
Hypothalamic functional testing is abnormal in many patients with CFS. The stress axis has been studied most intensively, following the important report of Demitrack et al., and is mentioned in more detail below. Disruption of both serotonergic and noradrenergic pathways also has been demonstrated in patients with CFS [32, 103, 104]. Not only have these objective abnormalities been demonstrated reproducibly by several laboratories, but they are also in patterns opposite to those seen in major depression. Many of the symptoms of CFS could result from such hypothalamic dysfunction.
Central and peripheral balance centre testing also has found abnormalities in patients with CFS [105, 106], although no treatment studies based on these observations have yet been conducted.
Formal testing of cognition has been used in an attempt to make objective measurements of the cognitive debilitation found in this condition. In this type of study it is important to start with well-characterized patient and control groups. It is also important to include depressed patients who do not have CFS, diagnosed according to the criteria in the DSM-IV, as cognitive deficits in CFS patients are often attributed to a secondary depression [107, 108]. Cognitive deficits have also been described in patients having an ongoing viral infection such as mononucleosis .
Recent results have pointed to a specific impairment of information processing in CFS patients . Impairments in learning and memory were described in a subset of patients [111, 112]. When this patient group was compared with patients with multiple sclerosis and patients with depression, greater deficits were seen in the CFS patients in specific tests such as complex auditory information processing . DeLuca et al. concluded that impaired cognition in CFS cannot be explained solely by the presence of a psychiatric condition. When CFS patients were subgrouped into those reporting sudden versus gradual illness onset, a differentiation between groups emerged. The rate of concurrent psychiatric disease was significantly greater in the gradual onset group. Whilst both groups showed a significant reduction in information processing ability compared to controls, impairment in memory was more severe in patients with sudden onset, pointing to a need to stratify patients according to the type of onset . Although methodologies may differ, some consistent findings have been reported in CFS patients such as impairments in complex information processing speed and efficiency .
Sleep laboratory studies in patients with CFS have variable results. Moldofsky  found the same specific pattern of disordered sleep physiology (an alpha rhythm disturbance within the nonrapid eye movement sleep) in patients CFS as he had earlier found in patients with the similar syndrome, fibromyalgia. Krupp found that the sleep disturbances in CFS were less specific, but were different from those seen in patients with MULTIPLE SCLEROSIS or in healthy control subjects . Buchwald et al. found that sleep pathology in patients with CFS was unrelated to concomitant psychiatric diagnoses, and that treatment of frank sleep disorders never resolved the symptoms of CFS , indicating that the sleep pathology was secondary to, rather than a cause of the chronic fatigue.
Thus, there is considerable evidence of abnormalities in the CNS, particularly the limbic system, in patients with CFS. The aetiology of the abnormalities, and their relationship to the symptoms of CFS, remains obscure. One interesting hypothesis that links infection with central nervous system dysfunction is the possibility of a chronic viral encephalitis as an initiator of a process leading to CFS. It is known that viruses in animals [120, 121] and humans [122–125] can affect the HPA-axis, thus making infectious agents that cause CNS disturbances of particular interest. Dysfunction of the hypothalamus could explain many of the symptoms of CFS.
Influence of stress
The influence of stress, in conjunction with infection, also was introduced early on as a possible cause . The pace of ‘modern’ society was blamed  and still is . The patients themselves often believe that a virus together with stress were the causative agents .
The involvement of stress in the pathogenesis of this illness seems clear for most physicians, and many patients acknowledge the influence of this factor on the natural course of symptoms. Stress can be divided into long-term and short-term subgroups which have different effects on human physiology. There is also the influence of biological adaptation. At low levels, stress may actually improve some aspects of immune function . However, most types of stress have been shown to impair immune system function and, in general, the immune defence against viral illness . One example is herpes simplex which rests, latent, in the neurones, but is susceptible to reactivation by a variety of factors including immune suppression, fever, hormonal changes, and physical and emotional stress.
Perhaps the best studied activation system in the body is the hypothalamus–pituitary–adrenocortical (HPA) axis. This system is an example of a negative feedback circuit. Corticotropin-releasing hormone (CRH) released from the hypothalamus stimulates the pituitary gland to secrete adrenocorticotropin hormone (ACTH), which in turn stimulates the adrenal cortex to secrete cortisol into the blood. An increase in the concentration of cortisol in the blood results in a direct inhibition of CRH secretion by the hypothalamus.
Under chronic stress, different disturbances in this feedback can occur. In conditions of deep depression, the whole HPA axis seems to be hyperactive, where cortisol does not inhibit the secretion of CRH in the normal way [131, 132]. In post-traumatic stress disorder the basal secretion of cortisol is low, but when the individual is reminded of the horrible events he/she experienced, the axis is immediately awakened . In CFS it seems, instead, that the CRH activity has become attenuated. As CRH is the impetus, the activity of the whole axis is generally low . So far no one has been able to determine the mechanisms behind these three types of disturbances, but research in animals and humans has confirmed these abnormal patterns of function [30, 103, 134–141].
However, basal activity of the HPA axis in CFS patients has not been found to differ from controls by some authors .
A dysregulation of the HPA axis was found in patients with fibromyalgia, a condition sharing some characteristics with CFS. These patients also have increased levels of the neuropeptide substance P in cerebrospinal fluid which, however, was detected within normal range in CFS patients , thus making it a potential biological differential marker.
Recent data also present the possibility that there are abnormalities in the GH-IGF axis in CFS patients . Whether this is caused by a primary pathological process is not known. It is important to further clarify the observed variations in the relevant endocrinological parameters.
As an example of the effects of a single stress factor, the effects of Hurricane Andrew were studied in CFS patients in Florida . It was found that patients from the high impact area had significant increases in physician-related clinical relapses and an exacerbation in frequency of several categories of self-reported physical symptoms compared to patients from lower impact areas. Illness burden also was significantly increased. The patients’ post-hurricane distress response was the single strongest predictor of the probability and severity of relapse and functional impairment. Optimism and social support were significantly associated with lower illness burden after the hurricane.
We know that there is a common chemical language, only partly deciphered, for the brain, immune, and endocrine systems . The immune and neuroendocrine systems represent a totally integrated information circuit, a result of sharing ligands and their receptors. The immune system communicates with the neuroendocrine system and can, at a local level, itself act as an effective endocrine organ. It is probable that small perturbations in immunological stimuli may cause major fluctuations in endocrine status .
Studies of different treatment schemes are not easy to perform in this condition. This is probably a consequence of the heterogeneity of patient populations and complexity of origin. Therefore, data have to be analysed comparing different groups of patients (e.g. those with gradual versus sudden onset, symptom severity, etc.) and control groups are needed. The use of a common case definition in defining subjects of all treatment studies should help this problem, but not eliminate it: most investigators believe that the cases circumscribed by the current CDC case definition are probably still a heterogeneous group. The hope is that application of the case definition will reduce heterogeneity, although it probably will not eliminate it. As the definition of CFS becomes more accepted and used by international research groups, comparable studies should become easier to perform. The natural course of CFS is still largely unknown and the cyclic nature of the symptoms presents problems when treatments are being evaluated. Also, the absence of objective markers of this illness makes evaluation more difficult and dependent on subjective measurements. Thus, studies are not easy to design and evaluate.
Relatively few large randomized trials of therapy have been conducted. Many clinicians report, from uncontrolled experience, that low doses of tricyclic agents seem to help patients – particularly in alleviating the sleep problems. This is of interest because of the proven value of low-dose tricyclics in a very similar syndrome, fibromyalgia [148, 149]. Patients usually have difficulty tolerating doses used for treatment of depression, as sedative and anticholinergic effects cause disturbing side-effects. Therefore, other classes of antidepressant drugs have been tried. Selective serotonin reuptake antagonists (SSRA) should, theoretically, be beneficial, as serotonergic pathways also play a role in sleep disturbances and sympathetic drive. A double-blinded study using fluoxetine (Prozac), however, showed no effect . One small study of a monoamine oxidase inhibitor (phenelzine) showed some improvement of symptoms . In practice, many patients are reluctant to take these drugs.
The antiviral medication, acyclovir, was found ineffective ; however, this medication has minimal or no in vitro activity against the viruses that have most frequently been associated with CFS. In a large randomized trial, the immune modulating drug, Ampligen, was found to have an effect on both symptoms and objective parameters over a 26-week period, but there was no evaluation of the durability of the improvement following the end of therapy . Amantadine  has also been tried, not because of its antiviral properties, but because of its central effects in releasing noradrenaline and dopamine, retarding reuptake of these transmitters, and its demonstrated value in improving the fatigue seen in some patients with multiple sclerosis. None have proven efficient in relieving CFS symptoms.
Other drugs studied include the oral antihistamine terfenadine  and alpha-interferon . Neither showed any therapeutic effect in CFS patients. Carnitine, essential for mitochondrial energy production, has been found to be beneficial to the patients in one study . Immunologic therapy with high dosages of gammaglobulin given intravenously have not been found to be effective [157, 158]. Low-dose hydrocortisone was associated with some improvement in symptoms but the degree of adrenal suppression precludes its practical use for CFS .
Cognitive behavioural therapy has been used in different schemes – either in groups [160, 161] or, more successfully, on an individual basis . The latter form proved effective in moderating symptoms, at one-year follow-up. Cognitive behaviour therapy emphasizes self-help. The goal is to help the patient to change unhelpful cognitions and behaviour. Although the effect has rarely so far been shown to be curative, the effect is substantial. At present, intensive individual CBT administered by trained staff is one treatment of choice .
Clinical management of conditions such as CFS is not developed at all levels of the health care system. This causes problems for many clinicians coming into contact with this illness for the first time, as information regarding diagnosis and treatment is not readily available . Updated information on CFS and distribution of such to clinicians is important and needed.
Trust must be established between the patient and doctor, as a thorough clinical investigation, to exclude other probable illness causes, is of utmost importance. A complete detailed life history must include both physical and psychological problems, which often require the expertise of a specialist to exclude them.
If the diagnosis of CFS is established, information on current knowledge should be given to the patient and repeated. This includes possible triggering events (e.g. infections in a predisposed individual), and factors that can perpetuate the condition. Cognitive behavioural therapy should be used on an individual basis if possible. A consistent pattern of living – including work, rest, sleep, and physical activity – should be applied, and a slow increase of daily activities introduced. It should be explained that even a slow increase in physical exercise can cause an exacerbation of symptoms, but often these subside with time and there is improvement.
Since an internationally accepted case definition was established in 1994, few follow-up studies have been performed on patients meeting this criterion. However, a systematic review of these studies indicates that less than 10% of patients recover completely . Risk factors for poor prognosis appear to be older age, comorbid disorders, and holding to the belief that the illness is due to physical causes.
Chronic fatigue syndrome (CFS) is an illness that produces physical symptoms and, in some individuals, psychological morbidity. Although a large number of objective, biological abnormalities have been identified in the patients, no diagnostic test with adequate sensitivity and specificity is yet available. Moreover, the pathogenesis of the illness has not been completely delineated. When this is the state of affairs in understanding an illness, as it has been and is with many, observers have varying opinions about the illness, and controversy is generated.
This controversy is exaggerated by the mind–body dualism that still is present in Western thinking. Plato introduced, and Decartes reinforced, the concept of a dichotomy between body and soul, and of illnesses of the body versus illnesses of the soul. Moreover, illnesses of the body are often viewed as ‘real’, whereas diseases of the soul are not. Until an illness becomes a disease, with a clear pathogenesis and diagnostic markers, it is not taken seriously by some clinicians.
This separation often leads to failure in the care given to individuals experiencing ill health due to an illness, and it is mainly due to this philosophy laying the foundation for our medical education. With new techniques being developed, our capacity to investigate underlying biological causes of ill health is increasing, independent of the psychological or physical origins of the symptoms.
CFS can be viewed as a model of ill health in which biological disturbances lead to symptoms which our culture considers either psychosomatic or imaginative. Current research points to an intimate contact between the immune system and the central nervous system, with the possible involvement of a microorganism causing disturbances in systems affecting the function of the individual. The origin is bound to be complex, and it may well be that the solution will come together with a more generally altered view about mind–body dualism, and the concept of illness and disease.
Received 26 October 1998; accepted 30 November 1998.