The following will summarise key findings and limitations of the studies discussed in the article “Existing estimates of ME/CFS prevalence and their limitations“.
Evidence of the prevalence of ME/CFS in the UK – Samms and Ponting (2025)
The most recent study relating to the prevalence of ME/CFS in the UK is a paper by Samms and Ponting, published in April 2025.
This study uses information recorded in hospital records on 62,782,175 people with a valid NHS number in England to consider the number of people who had been diagnosed with the ICD-10 code for post-viral fatigue which includes ME/CFS (ICD-10: G93.3) between April 1, 1989, and October 7, 2023.
Results showed that although the prevalence of “post-viral fatigue” was higher for English females than males, and varied by ethnicity, age, and level of deprivation, overall, 100,055 people (a prevalence of 0.16%) had been assigned a diagnostic code for “post-viral fatigue”.
As people with ME/CFS – particularly men and those in ethnic minority groups – often experience delays in diagnosis or may delay seeking health care due to the stigma associated with the disease, the research team used complex methods to “estimate prevalence when taking into consideration social and healthcare barriers”. This analysis suggested that based on a UK population of 68.3 million, 83,626 males and 320,296 females would be given a G93.3 diagnosis in their lifetime – 403,922 people in total which equates to 0.6%.
Notably, 403,922 would be a 62% increase over a current prevalence estimate of 250,000.
As with all research studies, the paper has limitations:
The data are from England rather than the whole of the UK:
Although the prevalence rates calculated using data for England were extrapolated by the authors to provide an estimate of the number of people in the UK with the diagnostic code G93.3, prevalence rates vary by location, and the rates observed in England may not reflect those in Scotland, Wales or Northern Ireland.
The medical records are for a diagnosis for G93.3:
This refers to “post-viral fatigue syndrome” rather than ME/CFS specifically. It is possible that not all those with a G93.3 diagnosis would meet more stringent ME/CFS diagnostic criteria.
Only those with a G93.3 diagnosis are included in the estimate:
Many people with ME/CFS may be undiagnosed. It is likely that there are people yet to be diagnosed with ME/CFS (G93.3) who are not represented in the prevalence estimates calculated in this study. Regrettably, many people with ME/CFS experience a delay in diagnosis or misdiagnosis, and some – especially men with ME/CFS – have reported that they delay seeking medical help for symptoms due to the stigma associated with the disease.
The study period includes diagnoses made under both versions of the NICE guideline for ME/CFS:
Unlike the updated 2021 guidelines, the 2007 version did not require either cognitive difficulties or unrefreshing sleep for a diagnosis of ME/CFS to be made. The 2007 guidelines were also less clear about the requirement for the cardinal symptom of ME/CFS – post-exertional malaise.
The data are from hospital and outpatient appointments only:
The HES data used in the study do not include diagnoses made at GP practices that were not also recorded in hospital records, meaning the prevalence rate calculated in this study may be an underestimation. In the paper, the authors state that in the UK Biobank only 28% of those with an “ME/CFS code” in their GP records also have the code G93.3 in their hospital records. Despite this, “post viral fatigue syndrome” may be relatively less commonly diagnosed in the primary care setting.
The accuracy of HES data relies on how clearly and correctly information is recorded:
Accuracy and completeness of information recorded in medical records can vary between clinicians, over time, and by location.
Prevalence estimates are likely to be an underestimate:
Due to the limitations of the hospital data used, the researchers suggest that the prevalence estimates calculated in the study are likely to underestimate the number of people with ME/CFS.
2018 UK Biobank Study
This study did not aim to calculate the prevalence of ME/CFS specifically, but the data collected allowed the researchers to consider the prevalence of several different diseases – including “self-reported chronic fatigue syndrome” (CFS).
Limitations of the data from the UK Biobank include:
- Although from individuals across England, Scotland and Wales, the UK Biobank does not include individuals from Northern Ireland.
- The dataset lacks ethnic diversity as it is predominantly “individuals of largely European descent”.
- Only includes adults aged between 40 and 69 years at baseline (between 2006 and 2010), thus missing the two main peaks of onset; often between the ages of 10 to 19 years and 30 to 39 years, with the average age at onset at 33 years.
Systematic review evidence – Lim and colleagues (2020)
This systematic review – a specific type of study that uses repeatable methods to search for, select, and combine all available evidence to answer a specific research question – considered information from 46 studies (56 datasets) published between 1990 and 2018.
Results from this review were calculated by combining results from several different studies, and provided estimates of ME/CFS prevalence by case definition as follows:
| Criteria | Prevalence rate (95% confidence interval– gives an indication of the range in which we can be 95% certain the true value lies.) |
| CDC-1994 (Fukuda) | 0.89% (0.60 to 1.33) |
| Oxford | 1.41% (0.68 to 2.93) |
| Australian | 0.79% (0.05 to 12.55) |
| Holmes | 0.17% (0.06 to 0.49) |
| Canadian Consensus Criteria; Epidemiological case definition; Post-viral exhaustion syndrome; and the National Institute for Health and Care Excellence guideline 2007. | 0.17% (0.04 to 0.83) |
Note: the 95% confidence interval gives an indication of the range in which we can be 95% certain the true value lies – so for the CDC-1994 criteria, the prevalence estimate is 0.89% and we can be 95% certain that the true prevalence rate for this population lies somewhere between 0.60% and 1.33%.
Findings from this review also highlighted that estimates of ME/CFS prevalence varied greatly based on sex, age, study participants (community vs primary care), ME/CFS case definition implemented, and diagnostic method used.
Within the systematic review, although several studies considered the prevalence of ME/CFS in the UK – with estimates ranging from 0.1% to 2.62% – all were published before the 2018 UK Biobank study referenced by the 2021 NICE guidelines. It is important to note that these were not UK-wide studies, but rather estimated prevalence from samples of the population in areas of the UK.
Systematic Review evidence – Estévez-López and colleagues (2020)
This article provides ME/CFS prevalence estimates for Europe based on three cross-sectional studies (published between 2002 and 2011) using both the CDC-1994 and CCC-2003 criteria.
The overall prevalence estimate for Europe given by Estévez-López and colleagues was 0.1% to 2.2%.
The researchers also summarised estimates by diagnostic criteria:
| Criteria | Study | Prevalence rate (95% confidence interval – gives an indication of the range in which we can be 95% certain the true value lies) |
| CDC-1994 (Fukuda) | Rimes et al 2006; adolescents aged 11-15 years in the UK | 0.1% (no confidence interval provided) |
| Nacul et al 2011; adults aged 18 to 64 years in three regions of England | 0.20% (95% CI 0.18 to 0.23) | |
| Lindal et al 2002; adults aged 19 to 75 years in Iceland | 2.2% (no confidence interval provided) | |
| CCC-2003 | (Nacul et al 2011; adults aged 18 to 64 years in three regions of England) | 0.10% (95% CI 0.09 to 0.12) |
The three studies included in the review by Estévez-López and colleagues are also included in the systematic review by Lim and colleagues. However, here evidence is summarised for Europe alone, rather than worldwide.
Studies published elsewhere in the world.
Since the 2018 systematic review was published, there has been little peer-reviewed research on the prevalence of ME/CFS.
Australia
Most notably, in 2022, a study considering the prevalence of ME/CFS in Australia was published. The research team considered the medical records of patients in Australia aged 13 years or older for the years 2015 to 2019. Results showed that prevalence varied by year, and by participant age group, and was higher for females compared with males.
Considering findings for the most recent year (2019) among 2,991,489 people, 2,481 had a medical record that included “A diagnosis of “myalgic encephalomyelitis”, “chronic fatigue syndrome”, “ME”, “CFS”, or “ME/CFS”. This equated to a prevalence of 94.9 per 100,000 (or 0.095%) ranging from 91.5 to 98.5 per 100,000. The authors suggest that the prevalence rate observed here may be due to factors identified during interviews the researchers carried out with people who have ME/CFS including:
“barriers to reaching diagnosis, and limited interactions with primary care due to a lack of available treatments/interventions, stigma and disbelief in ME/CFS as a condition”
Canada
While not yet peer-reviewed, there is one pre-print article by Chuluunbaatar-Lussier and colleagues (2024) in British Columbia, Canada which reports:
- Self-reported prevalence of “CFS” was 1.1% (217 out of 19,145)
- Among those self-reporting CFS, 34% met the “CCC-2003 and/or the IoM 2015 criteria” which equates to a prevalence of 0.4% for “confirmed cases of ME/CFS”
It is essential to note that this study used self-reported cases of “CFS” rather than “ME/CFS”, and that the assessment against the “CCC-2003 and/or the IoM 2015 criteria” was based on a self-reported “Symptom assessment questionnaire” and not validated by a clinician. Additionally, the definition of “healthy control” is unclear in the paper, and it is not possible to ascertain whether “healthy” participants were also screened for ME/CFS using diagnostic criteria – it is possible that some people who reported not having “CFS” may meet ME/CFS diagnostic criteria. Importantly, there was only a small sample size used to assess confirmed cases of ME/CFS (187 people; 85 self-report CFS, 29 confirmed cases).
USA
A report by the Centers for Disease Control – published in 2023, using data from a “nationally representative household survey of the civilian noninstitutionalized U.S. population”, found that among 57,133 adults (29,482 in 2021 and 27,651 in 2022), 1.3% (95% CI 1.2, 1.4) self-reported a diagnosis of ME/CFS. Within this report, it was identified that the prevalence of ME/CFS was higher for women than men, and varied by age group, “Race and Hispanic origin”, family income, and by rurality.
As with the other studies discussed, it is important to consider the limitations of the prevalence estimates obtained using the NHIS data; these were predominantly that only self-reported ME/CFS status was used – rather than the implementation of diagnostic criteria requiring PEM for a diagnosis to be made. In addition, this evidence is from a government report rather than a peer-reviewed scientific study.
(Also see research led by Mariah Wood summairsed in the following section)
Studies on how long COVID might have impacted the prevalence of ME/CFS
The most recent research comes from a team led by Mariah Wood in the USA. This study considered the medical records and self-reported symptoms of 2,745,374 adults from October 2022 to assess:
- The prevalence of “ME/CFS-like illness” – cases were those whose self-reported symptoms were congruent with the 2015 IoM criteria.
- The proportion of those with ME/CFS-like illness identified following COVID-19 illness – History of COVID-19 was based on a positive SARS-CoV-2 test or ICD-10 diagnosis code in the medical record, or self-report of prior COVID-19 on a survey.
Results showed that 45,892 (1.67%) of the adults considered reported symptoms consistent with the 2015 IoM criteria – “ME/CFS-like illness”. Of these, 14.2% developed the illness after COVID-19 infection.
As with all studies, there are limitations worth considering, for example:
- This study considered the prevalence of “ME/CFS-like illness” rather than “ME/CFS” – the authors state that the designation “ME/CFS-like illness” was used “to indicate that a full clinical evaluation is a requirement for an ME/CFS diagnosis but is not feasible with this survey only study design”.
- The proportion of those who developed ME/CFS-like illness after COVID-19 infection may not include those who developed symptoms following an asymptomatic COVID-19 infection, for example where someone did not display any symptoms and therefore did not take a COVID test.
- While this study indicates the number of people with “ME/CFS-like illness”, ME/CFS was not diagnosed by a health professional – rather, self-reported symptoms were assessed against ME/CFS diagnostic criteria.
- Only those who stated that English was their preferred language were included in the study, this limits how well the results might represent the wider population in the USA.
Additionally, a study (behind paywall) from the USA by Mirin and colleagues, published in 2022, used existing ME/CFS prevalence estimates – taken from an earlier study by the research team – for the USA and recalculated them to reflect post-COVID increases.
The methods used were complex but the authors concluded that, using existing estimates of ME/CFS, if 10% of those with long COVID met ME/CFS diagnostic criteria, the prevalence of “ME/CFS” in the USA may have increased from 1.49 million to between “5 and 9 million individuals” – based on the 2022 population for the USA of 341.5 million, this would equate to approximately 1.46% to 2.64%, which includes the 1.67% observed by Wood and colleagues.
Again, there are limitations to this research that must be considered, these include:
- The initial prevalence estimates used for adults were for “CFS” according to the Fukuda criteria rather than a diagnosis of ME/CFS and therefore may include those without PEM in the estimation. (It is worth noting that although the Fukuda criteria were used, the diagnosis was validated by a clinician rather than self-report alone.)
- The increased post-COVID estimates add on the number of COVID-19 survivors who potentially meet ME/CFS diagnostic criteria to the estimates for ME/CFS published in 2021 – this does not appear to account for those who were diagnosed with ME/CFS without a COVID-19 infection.
- More recent research by Jason – published in 2023 – suggests that the number of those with long COVID who also meet diagnostic criteria for ME/CFS may be higher than 10%, with 58% of 465 participants with long COVID fulfilling the CCC ME/CFS diagnostic criteria. (Interestingly, research published in July 2024 which uses data from the USA, has identified that a diagnosis of ME/CFS was no more likely following COVID-19 than it was following other acute infections. The authors of this study suggest a prevalence of ME/CFS after infection-like illness – which includes COVID-19 – to be 3% to 4%, but that even this would have a high impact on society.)
One additional pre-print study by Jamal and colleagues, which has been published ahead of peer review, considered the the number of people who met either the CCC or IoM for ME/CFS 6-8 months following COVID-19 infection. The researchers looked at two different cohorts; 217 adults in the John Hopkins University (JHU) COVID recovery cohort, and the 388 adults in the long term Impact of Infection with Novel Coronavirus (LIINC).
Results suggested that in the JHU cohort, 97 (45%) of participants met the study definition for long COVID – of these, 5 (5.2%) met IoM criteria for ME/CFS, and 2 (2.1%) met the CCC – it is important to note that the majority of these individuals were also living with HIV.
In the LIINC cohort, 281 (72%) adults met the study definition for long COVID, 51 (18.2%) of these met the IoM criteria for ME/CFS, and 29 (10.3%) met the CCC – few of these people were living with HIV.
As with all research studies discussed thus far, there are limitations of this research. such as:
- The study is a pre-print and has not yet been through the full peer review process
- Even once studies have undergone peer review, and are published in full, findings must be interpreted with a critical eye.
- The LIINC and JHU studies were designed early in the pandemic and not designed to assess ME/CFS – in their discussion Jamal and colleagues state that this means that they are only able to report people with an “ME/CFS -like condition without clinical confirmation”.
- Both the LIINC and JHU cohorts had high proportions of people living with HIV – While this is an important population to investigate, it may mean that the findings to not represent long COVID and ME/CFS rates in the wider population.
- The definitions for long COVID differed between the cohorts.
- The LIINC cohort classified a participant as having long COVID if they reported “a presence of at least one or more COVID-related symptom at 240 days post COVID.
- The JHU COVID recovery cohort classified participants as having long COVID if “at a survey given at 4,2, or 12 months post COVID, the participant answered the question ‘were you at your usual (pre-COVID) health status in the last week?’ as ‘No’ and also reported at least one symptom”
- For the report by Jamal and colleagues, participants in the JHU cohort were also classified as ever having long COVID if they reported a higher severity of any symptom at 4, 6, or 12 months post-infection compared to the month prior to COVID-19.
- Neither the JHU recovery study, nor the LIiNC study administered ME/CFS specific surveys – rather information was extracted from other sources. In the paper by Jamal and colleagues, it is reported that of the 36 ME/CFS specific “data elements” – symptoms making up both the IoM and the CCC criteria, only 25 and 26 were captured by the information available in the JHU study and the LIINC study, respectively.
- Most notably, the LIINC study did not ask participants specifically about post exertional malaise (PEM) (although this information has been added to case report forms), and in the JHU study, the question which was used to assess PEM was “during the past week, I have found that exercise brings on my fatigue” – and while this does capture an element of PEM, it may not fully capture the impact of all types of exertion, such as mental and social, and the impact they can also have on PEM symptoms and severity.
Also of relevance is a systematic review – a type of study that uses repeatable methods to search for, select, and combine all available evidence to answer a specific research question, has been published and concluded that 51% of people with long COVID met ME/CFS diagnostic criteria. However, in a letter to the editor of the Journal of Infection, Nuno Sepúlveda and Francisco Westemeier highlight several potential methodological issues with the systematic review that may mean that the main finding is – that approximately 50% of people with long COVID meet ME/CFS criteria, is likely to be an overestimation that can be explained by “biased samples” recruited from long COVID support groups.
The limitations of the review pointed out by Sepúlveda and Westemeier include:
- The 13 included articles do not meet conditions specified by the European Network on Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (EUROMENE) which state that:
- Included studies should use data collected for the purposes of answering the specific research question at hand – primary data.
- Studies conducted in biased samples – samples that have been collected in such a way that means some members of the intended population are more or less likely to have been included than others, should be excluded.
- Studies included in the review should not use information based on self-reported ME/CFS diagnosis.
- Studies using inappropriate case definitions – such as “CFS-like illness”, should be excluded from the review.
- Duplicate reports – studies which use the same dataset, should also be excluded. Rather, where data from multiple studies overlap, the study with the largest number of participants should be selected for inclusion.
- The review considered the Depaul Symptom Questionnaire (DSQ) as a diagnostic criteria for ME/CFS.
- Sepúlveda and Westemeier state that this was surprising given that “the DSQ is simply a questionnaire that allows testing different case definitions in the same individual”.
- The search terms used by the review considered “ME/CFS” but did not also include “CFS/ME” as an alternative acronym for the disease – Something Sepúlveda and Westemeier say might explain the exclusion of a study by Simani and colleagues, that appears to meet the inclusion criteria for the systematic review.
Additional limitations of the review may also be that the researchers only searched four databases for published articles, and did not appear to examine the reference lists of relevant systematic reviews, or of the studies that met the review inclusion criteria. There was also no mention of “citation searching” – a method which takes relevant papers and looks for newer articles that have cited them. Another potential limitation of this review is that it only included studies published in the English language.
What is prevalence? | Existing estimates | Difficulties with prevalence estimates | Summary
