2 wk9 wk20 dID, Infectious Diseases; Adol, adolescent.
*Not all patients met CDC case definitions for CFS.
Copyright © 1999 by Mosby, Inc.
The Chronic Fatigue Syndrome Coordinating Committee was chartered in 1996 to advise The Secretary of Health and Human Services regarding chronic fatigue syndrome research and to assure coordination of effort between governmental agencies, the biomedical community, and patient organizations. Acknowledging the paucity of information regarding CFS in adolescents, the committee recently called for a state-of-the-science workshop. The meeting, cosponsored by the National Institutes of Health and the Centers for Disease Control and Prevention, was held in Washington, DC, on April 28, 1998. This article summarizes the proceedings of that meeting. The intent is to call attention to the problem of CFS in adolescents and the state of knowledge of its epidemiology, natural history, and pathogenesis and to provide a framework for research.
Persistent illness characterized by fatigue and constitutional complaints is not new, being described under various names since the 18th century.[1] The term chronic fatigue syndrome, arrived at by a panel of experts,[2] emphasizes 3 important aspects of the illness: its chronicity (despite a tendency to wax and wane), the importance of profound fatigue as the overriding complaint for most patients, and our inability to ascribe the symptoms to a defined series of pathogenetic events ("syndrome"). Objections to the term have been raised, primarily by advocacy groups, citing its pejorative nature and its inaccurate portrayal of the symptoms experienced by patients. However, other terms such as chronic fatigue and immune dysfunction syndrome,[3] [4] postinfectious neuromyasthenia,[5] [6] and myalgic encephalomyelitis[5] [7] imply pathogenetic mechanisms that are not firmly established. A Joint Committee of the Royal Colleges of Physicians, Psychiatrists, and General Practitioners recently upheld the term CFS for this reason,[8] and most experts favor its retention until pathogenetic mechanisms are better understood or laboratory markers identified.
The concept of CFS was explored in the 1980s following the description of persistent fatiguing illness associated with serologic evidence of Epstein-Barr virus infection[9] [12] and the report of a cluster of cases in Incline Village, Nevada.[13] Shortly thereafter, it was found that EBV serologies did not discriminate outbreak cases from controls,[14] a fact that has been borne out in subsequent studies.[15] [19] In fact, virus isolation, genome detection, and lymphocyte transformation are generally not different between cases and controls,[20] [21] prompting some investigators to consider EBV as, at most, an innocent bystander. Nevertheless, some patients with acquired EBV infection experience fatigue for many months.[22] New information regarding the natural history of EBV mononucleosis was presented by Dr Dedra Buchwald (University of Washington, Seattle). Young adults (n = 100), identified by positive heterophile antibody test results, were studied within 2 weeks of illness onset. Only 38% and 12% had not recovered at 2 and 6 months, respectively, and neither EBV serologic markers nor psychologic variables predicted the outcome.
One legacy of EBV studies in CFS is the perseverant belief in a simple
medical model of pathogenesis, namely that of chronic or persistent viral infection.
Initial enthusiasm for a variety of causative agents, such as retroviruses,[3]
enteroviruses,[23]
and human herpesvirus
6[24]
has been tempered by the inability
to replicate preliminary observations.[25]
[27]
A host of other infectious agents have been studied without proof of causation, despite
the consistent observation that many cases are preceded by a flu-like illness. The
list of suspect agents includes: EBV;[9]
[21]
[28]
[32]
retroviruses (including human immunodeficiency virus)[3]
[19]
[25]
[27]
[33]
[36]
;
enteroviruses[19]
[23]
[26]
[29]
[37]
[45]
; human herpesvirus
6
*References [7]
[19]
[20]
[24]
[27]
[29]
[30]
and [46]
[51]
.
; human herpesvirus 7[47]
;
cytomegalovirus[14]
[19]
[20]
; herpes simplex virus[19]
[29]
; Borna disease virus[52]
;
Inoue-Melnick virus[48]
; hepatitis
A, B, and C viruses[19]
[53]
[54]
; arboviruses[19]
;
"stealth" virus[55]
[56]
;
respiratory viruses[19]
; measles[19]
;
parvovirus[19]
; rubella[19]
;
Borrelia burgdorferi[19]
[57]
[58]
; Mycoplasma
fermentans[59]
; Chlamydia species[19]
[60]
;
Candida albicans[19]
[61]
;
Toxoplasma gondii[20]
; Bartonella species[19]
; rickettsiae
and ehrlichiae[19]
; and Babesia
microti.[58]
The relationship
between EBV and illness does serve to illustrate the complexity of interaction between
pathogen and host. EBV titers tend to be higher in stressed individuals than in control
subjects,[62]
suggesting that psychologic
states can influence the ability to limit viral reactivation. Moreover, pre-morbid
psychosocial variables may affect clinical disease expression in EBV seroconverters.
[63]
Thus in EBV infection, as in
CFS,
the concept of mind-body dualism may detract from a complete understanding of the
illness.[64]
Other important observations from studies of adults set a framework for discussion of CFS in children. First, persistent fatigue is exceedingly common among primary care patients, ranging from 11% to 27%.[65] [68] This does not mean, however, that the prevalence of CFS is that high, because most fatigued adults do not meet CFS case definitions.[69] Second, many studies document significant psychologic disturbance among patients with CFS.[70] [76] However, CFS is not strictly a psychological problem, because differentiation of reactive from pre-morbid features is problematic.
Because there are no validated biomarkers for CFS, the syndrome remains
operationally defined. Four definitions have been offered
(Table I); none of these definitions specify age limits or specifically address issues
pertinent to pediatric cases.[2]
[77]
[79]
The question of whether the 1994 CDC case definition could be applied to pediatric cases was raised at this workshop and elsewhere.[8] In particular, the 6 months of fatigue required was considered by some to be overly rigorous, inasmuch as each month of disability and social isolation has more significance in the dynamic world of the adolescent than it does in the comparatively stationary world of the adult. Likewise, the psychosocial milieu of the adolescent cannot be overlooked in considering the pathogenesis of CFS. Adolescents struggle with issues related to puberty, socialization, autonomy, body image, self-concept, sexuality, peer pressure, goal uncertainty, and academic demands--issues that by and large have been resolved by adulthood. Moreover, although an adolescent's activity may be self-determined, it is heavily influenced by parents, in both positive and negative ways.
There have been several attempts to estimate the prevalence of CFS in adults. Early studies based on physician referrals in US cities estimated the prevalence at 2 to 7 per 100,000.[86] [87] In a Seattle health maintenance organization population-based study, the point prevalence of CFS was estimated at 75 to 267 cases per 100,000.[88] Estimates of 2200 to 2300 per 100,000 for CFS-like illness were derived from population-based studies in Michigan[89] and San Francisco, [90] but rates for CFS could not be established because physical examinations and laboratory tests were not done. The economic impact of CFS may be substantial.[91]
The
prevalence of CFS in Australian children, aged 10 to 19 years, was estimated at 48
per 100,000,[77]
and a recent controversial
study in the United Kingdom suggested that CFS was the most common cause of prolonged
medical leave from school.[92]
Dr
William C. Reeves (CDC, Atlanta, Ga) summarized efforts to estimate the prevalence
of CFS among children in the United States. Three studies were conducted: (1) physician-based
surveillance in 4 US cities that included patients 12 to 17 years of age, (2) a cross-sectional
telephone survey of 8004 randomly selected households in San Francisco, and (3) a
telephone survey of school nurses in Reno and Wichita.[93]
Prevalence estimates from these studies are given in Table III.
Dr Karen M. Jordan (DePaul University, Chicago, Ill) conducted a random-digit dialing survey of nearly 13,000 homes in Chicago, wherein one adult was asked about fatigue in all children, aged 5 to 18 years, living in the household. The study found CFS-like illness in 80 children, for a prevalence rate of 2060 per 100,000. There were approximately equal numbers of boys and girls, and the average age was 11.6 years. Latinos were strikingly over-represented among cases, a finding that differs from other epidemiologic studies. Individuals with CFS-like illness are currently being recalled for psychological evaluations, physical examinations, and laboratory studies in an effort to identify cases of CFS.
Dr Mark S. Smith (University of Washington, Seattle) extended an earlier report[81] on the clinical features and natural history of CFS in adolescents, describing 87 patients who were fatigued for over 6 months. The mean age was 15 years, and 71% were female. Four-factor Hollingshead scores were normally distributed, arguing against over-representation of upper socioeconomic strata. Sixty-six percent of patients reported abrupt onset, and the mean duration of symptoms at time of study entry was 19 months. The average number of school days missed by all referred patients during a 6-month period was 42, and 16 patients were home schooled. Although Spielberger State-Trait Anxiety Inventory/Trait form, Child Depression Inventory, and Child Somatization Inventory scores were in the normal range, psychiatric diagnoses were made in 35 of 76 referrals--most commonly major depression (n = 16), anxiety disorder (n = 8), and adjustment disorder (n = 8). Follow-up at 12 to 60 months showed self-rated improvement in 40 of 54 cases. Only 42 patients met 1994 CDC criteria, and common associated symptoms among them were headache, exercise intolerance, sore throat, concentration problems, myalgia, sleep disturbance (in many cases "phase shift"), and arthralgia. Dr Timothy V. Coy (University of Washington, Seattle) reported results from another study suggesting that symptoms of CFS reported by adolescents are similar to those reported by adults. The study group, however, was derived from referrals to an adult medicine facility and thus may have been biased in the direction of similarity.
Dr Henry M. Feder (University of Connecticut, Farmington) reported his recent experience with CFS referrals. Of 31 new patients seen between 1995 and 1997, 61% were honor students before illness onset. After onset, 48% became homebound and only 13% were able to attend school regularly. Demographic features and symptoms were similar to those reported from Seattle. In contrast to an earlier series from Farmington that reported excellent long-term outcomes,[82] only 48% reported marked improvement or resolution over 3 years. This may be due to changes in the types of patients referred today as compared with times when CFS was not well recognized. Fifteen-year follow-up data for 35 children involved in the Lyndonville cluster[80] were reported by Dr Bell. Interestingly, no new medical diagnoses that would have explained the initial illness were reported, consistent with previous observations that alternative diagnoses are rare.[22] [30] Overall, 80% reported a satisfactory outcome, although the time to resolution was variable and did not correlate with age at onset, gender, or initial illness severity. Many subjects who considered themselves completely recovered nevertheless considered their health to be suboptimal, indicating the possibility of health status confusion. Eight subjects reported having home schooling for more than 2 years.
These experiences suggest that the clinical presentation of CFS in adolescents is similar to that in adults. Affected youth are often socially isolated because of school absenteeism but in general tend to recover more readily. Questions remain about the biases that determine who is referred for subspecialty evaluation and how discrepancies between actual and self-reported activities can be resolved.
Beyond studies of infectious agents, a number of biologic associations have been proposed for CFS. Although interesting with respect to biologic plausibility, most of these associations remain unvalidated and fall short of establishing markers for the syndrome. Reported observations include elevated serum angiotensin-converting enzyme levels,[4] brainstem perfusion abnormalities,[94] impaired central muscle activation,[95] and the identification of unique urinary metabolic profiles.[96] [97] Immunologic, neurocognitive, and cardiovascular findings are discussed in more detail below, as are recently reported psychologic associations.
For some time, immunologic dysfunction has been postulated as a cause of CFS. Numerous studies have been done in different populations with varying degrees of methodological rigor, with inconsistent results. For example, some studies show increases in activated suppressor T cells,[51] [98] whereas others show levels similar to those of control subjects.[99] [101] Mitogen responses have been normal in some studies[20] [100] [104] but decreased in others.[98] [99] [105] [106] Studies have shown normal,[51] [99] [107] increased,[98] [102] [108] or decreased[100] [109] numbers of natural killer cells. Studies of cytokines have likewise been inconsistent. [102] [104] [110] [112]
As summarized by Dr James F. Jones (National Jewish Medical and Research Center, Denver, Colo), inconsistencies in the data and overinterpretation of laboratory findings have led to confusion about the role of immunologic phenomena in the pathogenesis of CFS. Moreover, given the effects of psychologic state on neuroendocrine-immune interactions, [113] it is difficult to know whether immunologic perturbations, if they exist, are primary or secondary, much less how they contribute to or correlate with illness expression. Adolescence per se probably does not constitute a unique immunologic setting, save for the possible effects of stress and hormonal changes.
Intriguing early observations suggested that CFS was associated with decreased numbers of circulating naive (CD4/CD45RA+ ) T cells and increases in memory (CD45RO+ ) T cells[99] (conflicting data[107] have been presented on this issue). Dr Jones reported that similar phenotype changes are seen in atopic individuals during allergy season and that allergic symptoms are frequently reported in patients with CFS.[9] [12] [80] [98] [114] A study at National Jewish involving approximately 20 subjects with CFS showed increased cell-associated interferon-alpha, interferon-alpha messenger RNA, tumor necrosis factor-alpha, and TNF-alpha mRNA in both subjects with CFS and subjects with allergy when compared with healthy control subjects and subjects with clinical depression. Interestingly, diminished levels of interleukin-10, an anti-inflammatory cytokine, were seen in CFS, allergy, and depression groups, suggesting that these conditions may involve unopposed synthesis of pro-inflammatory cytokines. Subjects with CFS were grouped into those with immune activation (elevated TNF-alpha and decreased interleukin-10) and those without. To determine the extent to which psychologic measures could predict immunologic status, the presence or absence of DSM-III-R Axis I or Axis II diagnoses and the SCL-90R Positive Symptom Distress Index were used as predictors in a multiple regression analysis. These 3 variables accounted for 66% of the variance in immunologic status, suggesting again an interplay between biologic host factors and psychologic well-being.
Dr Richard L. Bruno (The Post-Polio Institute, Englewood, NJ) drew parallels between CFS and post-polio syndrome, an illness characterized by new muscle weakness, musculoskeletal pain, fatigue, and functional decline occurring many years after acute poliomyelitis.[115] [116] Parallels include (1) the occurrence of lethargy and mental status changes during acute poliomyelitis, correlated with damage to the reticular activating system[117] ; (2) historical descriptions of CFS-like illness associated with polio outbreaks[5] ; (3) clinical similarities, such as impaired concentration and memory exacerbated by physical exertion and emotional stress[118] ; and (4) associations with abnormal brain lesions on magnetic resonance imaging,[119] [121] hypothalamo-pituitary-adrenocortical axis impairment,[122] [123] and autonomic dysfunction[124] [126] in both syndromes. To further elucidate the neuropsychology of CFS, 13 adolescents who met 1994 CDC criteria for CFS and 10 healthy control subjects were studied at The Post-Polio Institute. Subjects with CFS were found to be clinically impaired on the Paced Auditory Serial Addition and Gordon Distractibility tests, scored significantly lower on the Trail Making and IVA Visual Consistency tests, and were significantly more fatigued during testing. Scores on 2 tests, Trail Making and the Visual Memory Scale, correlated with fatigue severity. Interestingly, electroencephalographic power frequency analysis did not correlate with fatigue, as it had in subjects with post-polio syndrome, nor did it predict CFS diagnosis. The attention deficits seen in patients with CFS were relatively mild and consistent with subjective complaints but did not mirror precisely those seen in patients with post-polio syndrome.
Orthostatic stress results in venous pooling, reduced ventricular preload, and catecholamine release causing increased heart rate and peripheral vasoconstriction; the end result is maintenance of blood pressure. In patients with orthostatic intolerance syndromes, exaggerated sympathetic responses lead to excessive chronotropic and inotropic activity (postural tachycardia syndrome), stimulation of mechanoreceptors in the left ventricle, and ultimately withdrawal of sympathetic tone, resulting in vasodilation, bradycardia, and hypotension (neurally mediated hypotension).[127] The clinical result ranges from syncope to light-headedness, and in some patients, includes persistent fatigue.[128] Based on the latter association, Dr Peter C. Rowe and colleagues (Johns Hopkins University, Baltimore, Md) studied 7 adolescents with chronic fatigue by using the tilt-table test and found hypotension and exacerbation of fatigue in all of them.[125] Moreover, chronic symptoms resolved in 4 subjects with increased salt intake and use of beta-blocking agents or disopyramide. In a subsequent controlled study, abnormal tilt-table responses were seen in 22 of 23 subjects with CFS, most of whom were adults.[126] Sixteen of 21 treated patients demonstrated improvement in symptoms, usually within 1 month. Dr Rowe reported a recent series of 50 adolescents (mean age, 15 years; 37 female) who met the 1994 CDC criteria. Twenty of 27 had POTS and 5 had NMH while standing against a wall for 10 minutes. Twenty-two of 29 subjects experienced NMH during the tilt-table test, although only 6 had POTS as well. When the results of both tests were combined, 92% of subjects with CFS had some form of orthostatic intolerance.
Orthostatic intolerance among patients with CFS has also been demonstrated in studies by Dr Julian M. Stewart (New York Medical College, Valhalla). In addition, these studies suggest a lack of heart rate variability among subjects with CFS during both rest and provocative maneuvers. Although studies of orthostatic intolerance are intriguing, tilt test data should be interpreted cautiously. Some healthy individuals experience hypotension with tilting, particularly because increasing doses of isoproterenol are used to provoke vaso-vagal reflexes. In addition, subjects in these studies, by necessity, are not blinded to the test conditions, and this could affect the results.
There is little argument that psychologic distress is highly prevalent
among adults with CFS. In fact, many of the vegetative symptoms of depression such
as sleep disturbance, changes in appetite, loss of energy, difficulty concentrating,
and musculoskeletal complaints are commonly reported by patients with CFS. (To some
extent, current CFS definitions, which insist on minimum symptom criteria, may actually
select for psychologic comorbidity.[74]
)
Several studies point to psychologic disturbance predating CFS,[71]
[74]
[76]
implying a predisposing psychology analogous to the predisposing biology suggested
by some immunologic studies (Table IV).[22]
[81]
[129]
[131]
CFS is a complex psychobiologic phenomenon, which at present is better conceptualized as an illness (a disvalued change in quality of life) rather than a disease (a series of pathologic events).[133] Simple medical or psychologic models of pathogenesis seem implausible, and the weight of evidence points to a complex interaction between biologic factors and psychosocial variables. The results of studies in adults cannot be indiscriminately applied to pediatric patients. Especially for adolescents, the social context in which the illness occurs cannot be ignored, in its potential contribution to facilitating the illness and the potential impact of the illness on the individual.
The need for sensitive case ascertainment in epidemiologic studies must be balanced with the need for homogeneity in study groups. Current case definitions of CFS may need to be modified to achieve these ends for studies of adolescents. The mandatory nature of school attendance provides a uniquely quantifiable measure of dysfunction, which could be incorporated into a pediatric case definition. However, some children with chronic fatigue remain in school at the expense of other activities and would be excluded by definitions requiring nominal periods of school absence. The need for population-based samples, which avoid referral and treatment-seeking biases,[134] was underscored, and the need for appropriate denominators and validated, sensitive screening instruments was stressed. Linguistic and cultural differences in expression of fatigue need to be explored. Issues peculiar to pediatric studies were raised,[135] including confidentiality, consent, the age-appropriateness of interview questions, and the validity of second-hand reporting by caretakers.
The natural history of CFS in adolescents is favorable, but studies need to examine outcomes by demographic, clinical, biologic, and psychologic subtypes. The spectrum of illness needs to be defined, and the region of overlap with prolonged post-viral or post-mononucleosis fatigue explored, if indeed these are distinct entities. [136] More objective measures of functioning, such as daily phone diaries[137] or motion-sensing devices,[138] should be applied. There is clearly more room for further qualifying and quantifying the neurocognitive features of the syndrome. Dose-response relationships should be sought between proposed biologic associations and the degree of impairment or psychologic dysfunction. Cohort and case-control studies may be inadequate to address the issues of premorbid biologic and psychologic functioning. If large enough pre-existing databases could be found, nested case-control studies could potentially identify premorbid factors.
In physiologic studies such as those that examine orthostatic intolerance, more appropriate comparison groups for subjects with CFS might include physically deconditioned patients without fatigue. Factors that predict CFS among those with POTS or NMH should be studied. A double-blind, placebo-controlled trial of fludrocortisone (a potent mineralocorticoid), currently underway, should shed light on the relationship between orthostatic intolerance and CFS. Any physiologic benefits or symptom relief should be correlated with changes in psychologic well-being and social functioning.
Despite exhaustive research, no single agent has fulfilled Evans' criteria for disease causation.[139] With the sophistication of modern molecular diagnostic techniques, capable of identifying the agents causing acquired immunodeficiency syndrome in 2 years and hantavirus pulmonary syndrome in 5 weeks, it seems unlikely that a "chronic fatigue virus" will be discovered. Nevertheless, there may be a role for infectious agents as triggers for onset, along with other biologic insults such as allergen exposure. CDC is currently looking at differential gene expression in cases and controls in an effort to identify biomarkers for the syndrome.
Adolescents with CFS are psychologically distressed, but the direction of causality has not been established. Future studies will need to broaden the scope of investigation from individual psychologic characteristics to understanding the patient in the context of his or her family, the social milieu, and the health care community. The bidirectional relationship between psychologic functioning and biologic processes such as immune activation, neurocognitive impairment, and neuroendocrine dysregulation needs to be further explored. The selection of comparison groups that control for chronic inactivity, debilitation, and medical validation of illness is important. Finally, findings in small cohorts at single centers need to be replicated across locations and time periods.
CFS can be devastating for patients and their families, and physicians can become frustrated over the complexities of evaluation and treatment. Although there are no definitive interventions at this time, principles for multidisciplinary management have been suggested, centering on validation of the illness, screening for alternative diagnoses, psychologic evaluation, reassurance, reconditioning, symptomatic relief, graded increases in activity, and avoidance of unproven remedies.[82] [140] [142] More specific therapies await a better understanding of pathogenesis.
The author acknowledges significant contributions from Drs Leigh A. Sawyer and George T. Curlin and Ms. Lillian Abbey (NIAID), as well as Dr Brian W. J. Mahy and Ms Lisa C. Blake-DiSpigna (NCID).
