Department of Health Skip to content

Please note that this website has a UK government access keys system.

You are here:

Annex 1: Epidemiology of CFS/ME

1. Introduction

Information on the incidence and prevalence of CFS/ME is fragmentary and contradictory. Prevalence estimates vary by a factor of eight in the UK, and by a factor of 1000 worldwide. There are several reasons for this variation, some real (natural variation between populations) and some artefactual. Real variations mainly reflect the extent to which populations experience epidemic outbreaks of the illness, which occurs in both epidemic and sporadic forms. Artefactual variations may be due to differences in research methodologies and case definitions or to selection bias.

Much epidemiological research has used complex clinical research definitions, which are designed primarily to enable identification of homogeneous groups of individuals for clinical trials (table). Such studies generally have exclusion criteria of varying stringency, and consequently tend to underestimate the public-health burden of the disease when used in epidemiological research. Most epidemiological studies have been on a small scale, giving inconclusive results with wide confidence intervals. In some cases, inferences about population proportions have been made on the basis of studies that themselves lack a population base. Thus a study of hospital referrals, which used a fairly extensive case definition with no major exclusions, suggested an annual population incidence of three to five cases per 100,000 population,1 which is much lower than is generally thought.

Table - Existing diagnostic criteria (adults)

US Centers for Disease Control and Prevention, 1988 (CFS)(Holmes 1988)
6 months duration of fatigue
Functional activity - 50% decrease in activity
Six or eight symptoms required; physical signs sometimes required
Neuropsychiatric symptoms - may be present
New onset required
Exclusions: Extensive list of known physical causes,
psychosis, bipolar disorder, substance abuse
US Centers for Disease control and Prevention, 1994 (CFS)(Fukuda 1994)
6 months duration of fatigue
Substantial functional impairment
Four symptoms required
Cognitive or neuropsychiatric symptoms may be present
New onset required
Exclusions: Clinically important medical conditions, melancholic depression, substance abuse, bipolar disorder, psychosis, eating disorders
Australia, 1990 (CFS)(Lloyd)
6 months duration of fatigue
Substantial functional impairment - disruption of daily activities
Postexertional fatigue
No symptoms specified
Cognitive or neuropsychiatric symptoms required
New onset not required
Exclusions: Known physical causes, psychosis, bipolar disorder, substance abuse, eating disorders
UK, 1991 "Oxford Criteria" (CFS)(Sharpe)
6 months duration of fatigue
Disabling functional impairment - affects physical and mental functioning
No symptoms specified
Cognitive or neuropsychiatric symptoms - may be present
Definite onset required
Exclusions: Known physical causes, psychosis, bipolar disorder, eating disorder, organic brain disease, substance abuse
Other psychiatric disorders (depressive illness, anxiety disorders) are not reasons for exclusion
London, 1990
(ME)(Derived from Dowsett & Ramsay)
Complaint of general or local muscular fatigue following minimal exertion with prolonged recovery time
Neurological disturbance, especially of cognitive, autonomic, and sensory functions
Variable involvement of cardiac and other systems, a prolonged relapsing course
Syndrome commonly initiated by respiratory and/or gastrointestinal infection but an insidious or more dramatic onset after neurological, cardiac, or endocrine disability
World Health Organization, 1994
The WHO's International Classification of Diseases (ICD) provides a system of categories for international systematic recording. These are not diagnostic criteria, and are not used by clinicians as such. The current version, ICD-10, includes categories for 'Neurasthenia'/ 'Fatigue Syndrome' and for 'Post-Viral Fatigue Syndrome'/'Benign Myalgic Encephalomyelitis'.

The single outbreak in Asia was in Lebanon. In general, less developed countries are very under-represented, perhaps because there is a genuine difference in the occurrence of epidemics, but perhaps also owing to the absence of health-care and investigative facilities, and a high prevalence of other fatiguing conditions, which may mask the presence of CFS/ME.

Most cases occur sporadically, with a lifetime incidence rate in the region of 3-7%.4 Various small-scale studies 5-7 have suggested a prevalence of one to two per 1000 population (or between about 50,000 and 100,000 cases in the UK population at any time), with a peak rate in the 20-40 age group, and a preponderance of females. There seems to be a strong age association. Ramsey and Dowsett 8 confirmed the peak incidence in young adults and females, asserted a lifetime involvement of between 3% and 7% of the exposed population in the UK and the USA, and pointed to an association with social class, with professionals being more frequently affected than others. North American experience indicates a large increase in the number of cases in the late 1980s and early 1990s.9 This rise apparently paralleled a similar increase in the UK, where it was followed by a small decline in the late 1990s.10

This lack of consistent and reliable data explains the substantial difference in estimates of prevalence between those of the National Task Force Report 11 and the Royal Colleges Working Party.12 The latter estimate, which most authorities consider to be much higher than the actual number, was based on only two studies. One of these was a case-control study, which is not an appropriate way to find out the prevalence of a condition. The other 13 was a well conducted prevalence study of a population of 2376 people, aged 18 to 45 years. It showed a point prevalence for CFS/ME of 2·6%; the estimate was 0·5% when comorbid psychological disorders were excluded, with a modest preponderance of females and no association with social class. The investigators suggested that an excess of people in higher socioeconomic groups in other studies may be the result of selection bias. These rates, if applicable to the whole country, suggest a total of between 250,000 and 1,250,000 people with CFS/ME in the UK, depending on the case definition used. This total is much higher than the numbers suggested by other researchers, and it has not been corroborated by other studies. However, the study was carried out in Swindon, a New Town. In their study of long-term school absence due to illness, Dowsett and Colby 14 noted that rates of absence were higher in the New Towns included than elsewhere, and postulated a possible effect of the mixing of populations with differing degrees of immunity to viral infection (cf, the Kinlen hypothesis on childhood leukaemia in proximity to nuclear installations).

2. Worldwide variation

The worldwide research evidence on the epidemiology of CFS/ME must be interpreted with caution. In particular, application to the UK population of prevalence rates determined elsewhere is problematic, because the extent of natural variation between populations is simply not known. In the USA, the Centers for Disease Control and Prevention (CDC) reported a prevalence rate in the adult (18 and older) population of four to ten cases per 100,000, but a much higher prevalence of CFS/ME-like illness was detected in a study in San Francisco.15 The difference was essentially due to use of a more restrictive case definition in the CDC study. A study in a health-maintenance organization in Seattle found a population prevalence of between 0·08% and 0·30%,16 but this is viewed as a conservative estimate because of the restrictive case definition used.12

A nationwide survey of hospital referrals in Japan 17 suggested point prevalences of 0·4 and 0·6 per 100,000 population in 1992 and 1993, respectively, with an annual incidence rate of 0·46 per 100,000 population, and a slight preponderance of females but no relation to age. However, hospital referrals are the tip of an iceberg, so inferences as to the true population prevalence are fraught with difficulty.

A recent community-based study in Chicago 18 made an important contribution to the epidemiological literature. The researchers pointed out that few previous epidemiological studies have had community-based samples. They reported the results of a two-stage sample survey of 18,668 adults from an ethnically and socioeconomically diverse population. Overall, a prevalence rate of 422 per 100,000 population was found, with the 1994 Fukuda definition, though only 75% of patients reported postexertional malaise.19 This study has attracted particular interest in the UK, because such a rate, if applied to the UK population, would suggest a total number of cases of around 250,000, which more or less agrees with the general consensus view of the burden of the illness in the population. However, although the initial sample was large, only 32 people with CFS/ME were identified, and the confidence interval was wide. The researchers also commented on prevalence rates in subsets of the population defined by sex, age, ethnicity, and population, and pointed out differences between them. For example, the highest rates were found in females, people of Hispanic origin, people aged 40-49, and skilled workers. However, the numbers of cases in each subset of the population were small, 95% confidence intervals of the proportions (not presented in the paper) were very wide, and none of the differences was statistically significant.

This lack of significance of variations between subsets of the population is borne out by a similar study, in San Francisco,20 in which the overall prevalence rate was similar, but a different pattern of variation between population subsets was observed, with the highest prevalence being found in women, people with low incomes, African Americans, and Native Americans.

3. Case definitions and other variations

Various attempts have been made to define the syndrome considered here as CFS/ME on the basis of its clinical features. The original description of ME was published by Ramsay and colleagues. Since the first definition of CFS in 1988,21 one revision and other alternatives have been proposed.22-24 An internationally agreed definition has now been proposed by Fukuda and colleagues,25 for the International Chronic Fatigue Syndrome Study Group, which includes CDC, Australian, and British representatives, and this definition is gaining general approval for clinical research purposes.

Levine 26 confirmed that epidemiological studies of CFS/ME have been hampered by the lack of a specific diagnostic test, and that changes in case definitions have been a major factor in the widely varying estimates of prevalence rates. Levine also made the point 27 that many fatiguing syndromes overlap, and to define boundaries is difficult, and to some extent artificial. These syndromes include CFS/ME, chronic fatigue/immune dysfunction syndrome, low natural killer cell syndrome, post-infectious CFS/acute-onset ME, epidemic neuromyasthenia, Lyme disease, brucellosis, and chronic Epstein-Barr virus infection. This overlapping makes incidence and prevalence studies difficult, and means that the findings are very prone to variation due to differences in case definitions used. Risk factors identified to date were summarised by Levine as age (young adults are most at risk), sex (female), ethnicity (whites more than blacks), higher socioeconomic status (though this may reflect referral patterns rather than true incidence), and predisposing factors such as genetic factors and endocrine dysfunction.

Another problem of definition arises from the fact that other disorders, such as fibromyalgia and multiple chemical sensitivity, have similar clinical features. There is much overlap, and comorbidity is common.28 CFS/ME is clearly not a single diagnostic entity, but a symptom complex in which dysfunction is multifactorial in origin.29 Therefore, concern exists, especially for many patients with CFS/ME, that higher priority be given to research directed at identifying specific subgroups, particularly where these can be characterised by biological markers.

Results from a physician-based surveillance system in four US cities 30 enabled prevalence to be estimated for different ethnic groups (whites and blacks), and demonstrated the impact of different case definitions on prevalence. In the group of patients defined according to the CDC definition, the overall prevalence was 5·9 per 100,000 population, whereas in a group with a less restrictive definition the prevalence rate was 9·1. A third group, with a case definition approximating to the Oxford definition, had a prevalence rate of 25·4 per 100,000 population. The disease was much more common in all groups among whites than among blacks, with an overall prevalence in all groups with no major exclusions of 29·1 per 100,000 population in whites and 3·1 in blacks.

Manu and colleagues 31 examined the impact of the 1988 Holmes definition 21 on case ascertainment. They found that, of 135 patients with chronic fatigue of at least 6 months' duration, only six met the definition. Most did not meet the second major criterion, because of the presence of other diseases, particularly psychiatric disorders.

Another factor that is likely to affect reported incidence and prevalence of CFS/ME is the willingness of doctors to diagnose the disorder. Evidence suggests that many doctors remain ignorant of the condition. A 1997 survey of Irish GPs found that only 58% accepted CFS/ME as a genuine clinical entity.32 Similarly, in a survey of physicians in Recife, Brazil, only about a third of respondents mentioned CFS/ME as a possible diagnosis when presented with a CFS/ME-like case scenario.33 However, medical practice may be changing in terms of the extent to which doctors are prepared to make a diagnosis of CFS/ME, or indeed the illness may be changing its nature. In 1995, Shanks and Ho-Yen 34 concluded that the recorded incidence of psychiatric disorder was lower among patients who had become ill more recently. Possible explanations for this finding include: that patients are being given an earlier diagnosis, which together with more appropriate advice and management reduces consequent stress, anxiety, and depression; or that doctors are becoming more accurate at diagnosing CFS/ME and distinguishing it from other psychological illnesses which may have previously been misdiagnosed under the umbrella term CFS/ME.

4. Children

CFS/ME occurs in children in both epidemic and sporadic forms,14 but it seems to be less common than in adults, particularly young adults. It is uncommon under the age of 5 years. It occurs equally in boys and girls. Dowsett and Colby 14 undertook a survey in six Local Education Authority areas to identify all causes of medically certified long-term sickness absence in pupils and staff over a 5-year period. 1098 schools were surveyed, covering 333,024 pupils and 27,327 staff. The investigators found period prevalence rates of 69/100,000 among pupils and 520/100,000 for staff, confirming the impression from other research that CFS/ME is less common among children than among young adults. The female/male ratio among the pupils with CFS/ME was 2·15 to 1. Dowsett and Colby concluded that CFS/ME was the single most important medical cause of long-term school absence. This conclusion seems likely, but since the diagnoses could not be validated, the data should be interpreted with caution; it is not certain that consistent diagnostic criteria were being applied.

The impression of CFS/ME affecting boys and girls equally is confirmed by a community-based survey of more than 12,000 households in Chicago,36 which was undertaken to assess the prevalence of fatigue and CFS/ME-like illness in a sample of 5-17-year-olds. 2·05% were diagnosed with CFS/ME-like illness (modified CDC criteria 1994). Adolescents had a slightly higher rate of CFS/ME-like illness (2·91%) than children aged 5-12 (1·96%). There were roughly equal numbers of male and female patients. Prevalence rates were highest among children and young people of Hispanic origin, but numbers were small and confidence intervals wide.

Certain differences are apparent between the syndrome as described in adults and that encountered in children and adolescents. A review of 58 child and adolescent patients in the USA showed that these patients tended to present earlier in the course of the illness and to recover more quickly.37 These findings were confirmed by Rangel and co-workers.38 In their study, 25 children attending two London tertiary referral centres over the previous 5 years, who met the Oxford case definition for CFS/ME, were assessed by means of case-note examination and interview. 75% of the patients had developed CFS/ME between September and December, with an average age of onset of 11 years. 76% were from social classes I or II, and 50% were receiving private education. Two-thirds of the patients had recovered by the time of interview (mean 45·5 months after onset), and the mean time between onset and recovery was 38 months. Factors associated with prolonged illness included poorer socioeconomic background and development of the disease at a time other than the autumn period. Acute-onset disease, with a specific trigger event (infectious in most cases), was associated with better prognosis. Most children made a full recovery within 4 years, but that time may be crucial for child development, and the disease may therefore have serious long-term repercussions. These findings are interesting, but the results should be treated with caution. The sample size was very small, and appears unrepresentative of the population, which suggests the possibility of selection biases. The researchers noted that in many children, the onset coincided with the stressful event of transfer to secondary school. This finding does not, of course, prove causation, and they reported also that outcomes were unrelated to the presence or absence of emotional problems.

As regards epidemic cases, Bell 35 reported the impact of the 1985 Lyndonville outbreak on children. The modal age at onset (15 of 44 cases, or 34%) was 9-11. In contrast to other studies in children, there was a clear preponderance (61%) of female cases. 34% had an illness of acute onset. Bell pointed out that CFS/ME is commonly misdiagnosed as school phobia, and advised that this diagnosis should not be made without evidence of pre-existing and concurrent emotional disorder.

5. Conclusion

We know very little about the incidence and prevalence of CFS/ME. Thus, one of the key pieces of information needed for health-needs assessment, as a prelude to provision of an adequate network of services, is missing. This gap must be filled if the NHS and other statutory agencies are to meet the needs of CFS/ME patients in a comprehensive and equitable way.

Some useful studies have been done, but even in these, the numbers of cases identified were small and the 95% confidence limits of estimates wide. Therefore limited reliance should be placed on their findings. Even though the study by Jason and colleagues 18 produced a result compatible with many preconceptions about the scale of the problem in the UK, its use of a restrictive case definition would make underestimation of the true scale of the disease likely. The researchers themselves concede that their methodology may well have led to underestimates of the number of people with severe CFS/ME in the study population. Furthermore, given the likelihood of natural variation between populations, results obtained in one population cannot be accurately extrapolated to another.

Many estimates of incidence and prevalence are made by extrapolation of results from different populations because there have been so few epidemiological studies. There is, however, no evidence that similar incidence and prevalence rates occur in different populations. Indeed, the likelihood is that biological variation in incidence and prevalence occurs between populations, but in the present state of knowledge even this cannot be said with certainty.

Local community-based studies, while providing useful pointers, are not sufficient in themselves to answer outstanding questions about incidence and prevalence of the disease. In the UK, organisation of primary care services offers a unique opportunity to carry out prevalence studies on the national scale necessary to generate requisite data. Such studies could usefully be augmented by community-based studies to demonstrate and quantify variations in prevalence between communities, and to validate predictions based on national studies.


1 Behan PO, Behan WMH, Bell E. The post-viral syndrome: an analysis of the findings in 50 cases. J Infect 1985; 10: 211-22.

2 Acheson ED, The clinical syndrome variously called Benign Myalgic Encephalomyelitis, Iceland Disease and Epidemic Neuromyasthenia. Am J Med 1959; 26: 569-95.

3 Hyde BM, Parish G. A bibliography of ME/CFS epidemics. In: Hyde BM, Goldstein J, Levine P, eds. The clinical and scientific basis of myalgic encephalomyelitis/chronic fatigue syndrome. Ottawa: Nightingale Research Foundation, 1992: 176-86.

4 Calder BD, Warnock PJ. Chronic fatigue syndrome. BMJ 1991; 304: 181.

5 Dowsett EG, Ramsay AM, McCartney RA, Bell EJ. Myalgic encephalitis - a persistent enteroviral infection? Postgrad Med J 1990; 66: 526-30.

6 Hinds GME, McCluskey DR. A retrospective study of the chronic fatigue syndrome. Proc Roy Coll Phys Edinb 1993; 23: 10-14.

7 Ho-Yen DO, McNamara I. General practitioners' experience of the chronic fatigue syndrome. Br J Gen Pract 1991; 41: 324-26.

8 Ramsay AM, Dowsett MB. In: Hyde BM, Goldstein J, Levine P, eds. The clinical and scientific basis of myalgic encephalomyelitis/chronic fatigue syndrome. Ottawa: Nightingale Research Foundation, 1992: 81-84.

9 Hyde B. Unravelling the enigma (preface). In: Hyde BM, Goldstein J, Levine P, eds. The clinical and scientific basis of myalgic encephalomyelitis/chronic fatigue syndrome. Ottawa: Nightingale Research Foundation, 1992: x-xiii.

10 Dowsett EG. Personal communication.

11 National Task Force on Chronic Fatigue Syndrome/Post Viral Fatigue Syndrome/Myalgic Encephalomyelitis (1994). Report. Bristol: Westcare, 1994.

12 Royal Colleges of Physicians, Psychiatrists, and General Practitioners. Chronic fatigue syndrome. Report of a Joint Working Group. London, 1996.

13 Wessely S, Chalder T, Hirsch S, et al. The prevalence and morbidity of chronic fatigue and chronic fatigue syndrome: a prospective primary care study. Am J Public Health 1997; 87: 1449-55.

14 Dowsett EG, Colby J. Long-term sickness absence due to ME/CFS in UK schools. J Chron Fatigue Syndr 1997; 3: 29-42.

15 National Institute of Allergy and Infectious Diseases. Chronic fatigue syndrome -information for physicians. Bethesda: National Institutes of Health, 1996.

16 Buchwald D, Umali P, Umali J, et al. Chronic fatigue and the chronic syndrome: prevalence in a Pacific Northwest Health Care System. Ann Intern Med 1995; 123: 81-88.

17 Minowa M, Jiamo M. Descriptive epidemiology of chronic fatigue syndrome based on a nationwide survey in Japan. J Epidemiol 1996; 6: 75-80.

18 Jason LA, Richman JA, Rademaker AW, et al. A community-based study of chronic fatigue syndrome. Arch Intern Med 1999; 159: 2129-37.

19 Goudsmit E. Capita Selecta Quarterly, December 1999; URL :-

20 Steele L, Dobbins JG, Fukuda K, et al. The epidemiology of chronic fatigue in San Francisco. Am J Med 1998; 105: 83S-90S.

21 Holmes GP, Kaplan JE, Gantz MN, et al. Chronic fatigue syndrome: a working case definition. Ann Intern Med 1988; 108: 387-89.

22 Schluderberg A, Straus SE, Peterson P, et al. Chronic fatigue syndrome research: definition and medical outcome assessment. Ann Intern Med 1992; 117: 325-31.

23 Lloyd AR, Hickie I, Boughton CR, et al. Prevalence of chronic fatigue syndrome in an Australian population. Med J Aust 1990; 153: 522-28.

24 Sharpe MC, Archard LC, Banatvala JE, et al. A report - Chronic fatigue syndrome: guidelines for research. J Roy Soc Med 1991; 84: 118-21.

25 Fukuda K, Straus S, Hickie I, Sharpe MC, Dobbins JG, Komaroff AL, and the International Chronic Fatigue Syndrome Study Group. Chronic fatigue syndrome: a comprehensive approach to its definition and study. Ann Intern Med 1994; 121: 953-59.

26 Levine PH. Epidemiologic advances in chronic fatigue syndrome. J Psychiatr Res 1997; 31: 7-18.

27 Levine PH. Epidemiological aspects of CFS/ME. In: Hyde BM, Goldstein J, Levine P, eds. The clinical and scientific basis of myalgic encephalomyelitis/chronic fatigue syndrome. Ottawa: Nightingale Research Foundation, 1992: 196-205.

28 Jason LA., Taylor RR, Kennedy CL. Chronic fatigue syndrome, fibromyalgia, and multiple chemical sensitivities in a community-based sample of persons with chronic fatigue syndrome-like symptoms. Psychosom Med 2000; 62: 655-63.

29 Wessely S. Chronic fatigue syndrome. Summary of a report of a joint committee of the Royal Colleges of Physicians, Psychiatrists and General Practitioners. J Roy Coll Phys Lond 1996; 30: 497-504.

30 Gunn WJ, Connell DB, Randall B. Epidemiology of chronic fatigue syndrome: the Centers for Disease Control study. In: Ciba Foundation Symposium no. 173 - Chronic fatigue syndrome. Chichester: Wiley, 1993: 83-93.

31 Manu P, Lane TJ, Matthews DA. The frequency of the chronic fatigue syndrome in patients with symptoms of persistent fatigue. Ann Intern Med 1988; 109: 554-56.

32 Fitzgibbon EJ, Murphy D, O'Shea K, Kelleher C. Chronic debilitating fatigue in Irish general practice: a survey of general practitioners' experience. Br J Gen Pract 1997; 47: 618-22.

33 Nacul LC, Pereira FLR, Jaques GA, et al. What do doctors and health workers of North-East Brazil know and so about chronic fatigue syndrome? (Personal communication - unpublished).

34 Shanks MF, Ho-Yen DO. A clinical study of chronic fatigue syndrome. Br J Psychiatry 1995; 166: 798-801.

35 Bell DS. Children with myalgic encephalomyelitis/chronic fatigue-immune dysfunctionsyndrome: overview and review of the literature. In Hyde, B (ed.). The Clinical and Scientific Basis of Myalgic Encephalomyelitis/Chronic Fatigue Syndrome, Ottawa, The Nightingale Research Foundation, 1992. Chapter 19, p 209-216.

36 Jordan KM, Ayers PM, Jahn SC, et al. Prevalence of fatigue and chronic fatigue syndrome-like illness in children and adolescents. J Chron Fatigue Syndr 2000; 6: 3-21.

37 Krilov LR, Fisher M, Friedman SB, Reitman D, Mandel FS. Course and outcome of chronic fatigue in children and adolescents. Pediatrics 1998; 102: 360-66.

38 Rangel L, Garralda ME, Levin M, Roberts H. The course of severe chronic fatigue syndrome in childhood. J Roy Soc Med 2000; 93: 129-34.

Access keys