Fatigue
Fatigue is the most prevalent and one of the most quantitatively severe core symptoms of ME/CFS — present in nearly all patients and persistently severe across disease duration. While post-exertional malaise (PEM) is the illness’s hallmark feature (a delayed, exertion-triggered exacerbation), fatigue is the everyday symptom with the highest measured burden; the two are distinct and addressed separately (this section and the next). This section covers fatigue’s definition, assessment and severity, objective measurement, and phenotypic structure.
1 Definitional Distinction from Fatigability
Fatigue in ME/CFS is most usefully defined as the subjective perception of exhaustion, reduced energy, and impaired motivation to sustain activity. It is formally distinct from fatigability — the objective decline in physical or cognitive performance during sustained effort. The two constructs correlate only partially and require different measurement approaches.
Fatigue — the subjective perception of exhaustion, lack of energy, and reduced motivation — is formally distinct from fatigability — the objective decline in performance during sustained effort (Tankisi et al. 2024). These dimensions are assessed by different tools and reflect different neurophysiological processes: subjective scales capture perceived exhaustion, whereas objective tests (twitch interpolation, transcranial magnetic stimulation, surface EMG) quantify actual performance decline.
(Certainty: 0.60, narrative review; ME/CFS-relevant framework, not ME/CFS-specific cohort.)
Limitations: Narrative review without systematic methodology; framework applies broadly to fatigue states and is not ME/CFS-validated.
Reputation: Not yet replicated as an ME/CFS-specific diagnostic distinction, though the fatigue/fatigability split is standard in exercise physiology.
Severity applicability: unknown — the framework is not severity-stratified; it applies in principle across mild–very-severe patients but this is not directly evidenced.
Consequence: Clinicians who measure only subjective fatigue may miss objective fatigability, and vice versa; assessing both gives a more complete picture of the disability driving ME/CFS.
The World Health Organization’s International Classification of Functioning, Disability and Health (ICF) codifies fatigue and fatigability as separate body-function categories: energy level (b1300) captures fatigue as an impairment of subjective energy, while fatigability (b4552) captures the objective performance dimension (Bileviciute-Ljungar et al. 2020). In the preliminary ME/CFS ICF Core Set validation cohort (n=100), fatigue was impaired in 100% of patients and fatigability in 96%, indicating both dimensions are near-universally affected.
(Certainty: 0.55, preliminary single-center core set; discounted unchanged — ME/CFS cohort.)
Limitations: Preliminary core set developed by a single center’s rehabilitation team via consensus, not a full WHO Delphi process.
Reputation: Not yet independently replicated; the b1300/b4552 split is, however, standard ICF nomenclature.
Severity applicability: unknown — the preliminary ICF core-set cohort was not severity-stratified.
Consequence: Because an international classification system treats fatigue and fatigability as separate, internationally comparable categories, patients and researchers can use the same framework to describe both the experienced exhaustion and the measurable performance loss.
2 Severity of Fatigue
Fatigue is the most severe core symptom in ME/CFS by quantitative measurement. A systematic review and meta-analysis provides the most comprehensive severity benchmarks available.
A systematic review and meta-analysis of 60 randomized controlled trials (n=7,088 ME/CFS patients) estimated pooled fatigue severity at 77.9 out of 100 (95% CI 74.7–81.0) (Park et al. 2024). Physical fatigue (74.3) and cognitive fatigue (74.2) exceeded mental fatigue (70.1), indicating the subjective exhaustion is both physical and cognitive rather than purely motivational.
(Certainty: 0.75, systematic review with large pooled n; discounted 0.75 — ME/CFS cohort. Severity applicability: predominantly mild–moderate trial populations; severe/very-severe not stratified.)
Limitations: Heterogeneous RCT populations and diagnostic criteria; severity therefore ranges widely (54.2–88.6 on a normalized 100-point scale) depending on case definition and instrument.
Reputation: Single meta-analysis, not yet independently replicated, though the finding that fatigue is severe and near-universal is consistent across primary studies.
Consequence: Clinicians and trial designers can benchmark an individual patient’s fatigue (or a trial’s eligibility threshold) against a pooled estimate of about 78/100, giving a concrete reference point rather than an abstract impression.
Reported fatigue severity depends heavily on both the assessment instrument and the diagnostic criteria used. Across the pooled RCT population, normalized fatigue severity ranged from 54.2 (Mental Fatigue Scale; ICC-defined patients) to 88.6 (Checklist Individual Strength; Canadian-defined patients) (Park et al. 2024). This measurement heterogeneity means a single number has little meaning without specifying the instrument and case definition behind it.
(Certainty: 0.60, from the Park meta-analysis subgroup analysis; discounted 0.60.)
Limitations: Instrument-level values come from subgroup analysis of heterogeneous trials, not a head-to-head instrument validation.
Reputation: Consistent with the broader measurement-heterogeneity literature on fatigue questionnaires in chronic illness.
Severity applicability: applies to the pooled RCT trial populations (predominantly mild–moderate); severe/very-severe applicability unknown — study population not stratified.
Consequence: Because fatigue scores are not interchangeable across questionnaires, research teams must fix a single instrument for a given study or comparison, or results mislead.
3 Objective Measurement and Fatigability
Subjective fatigue scales measure what patients report; objective correlates — actigraphy-based activity and neurophysiological fatigability — measure what patients do. These provide independent evidence that fatigue and fatigability have measurable biological and behavioral substrates.
Actigraphy provides an objective behavioral correlate of fatigue. In a UK Biobank analysis, patients with chronic fatigue syndrome showed decreased overall movement (Cohen’s d=0.220), lower activity amplitudes (d=−0.377), and lower wrist-temperature amplitude (d=−0.173) relative to 63,133 controls (Liu et al. 2025). Community-based samples also showed a genetic overlap between CFS-associated variants and population-level fatigue and actigraphy measures, supporting a continuous biological substrate for fatigue rather than a categorical illness boundary.
(Certainty: 0.70, large population cohort; discounted 0.70. Caveat: CFS case definition was self-reported and the CFS group was small, n=295.)
Limitations: Self-reported CFS without clinical confirmation; actigraphy captures activity but not subjective fatigue directly; small CFS subgroup relative to controls.
Reputation: Actigraphy-based activity reduction in ME/CFS is broadly consistent with prior actigraphy studies, though replication of the specific genetic-continuity finding is pending.
Severity applicability: Population cohort, predominantly ambulatory; severe/very-severe patients (often bedbound) would show even more extreme reduction — untested.
Consequence: Wrist-worn sensors may one day give clinicians an objective, continuous measure of how profoundly fatigue limits a patient’s activity, complementing self-reported scores.
Multimodal neuroimaging during a sustained grip-force task showed that ME/CFS patients develop fatigue earlier than healthy controls despite equivalent maximum voluntary force (Bedard et al. 2026). Healthy brains progressively increase motor-cortical and subcortical output to sustain force, whereas ME/CFS brains show minimal fluctuation — a failure of central motor drive rather than peripheral muscle exhaustion. This conflicts with the common assumption that the disability is primarily peripheral.
(Certainty: 0.60, small but rigorous multimodal design, n=15 ME/CFS vs n=19 controls; discounted 0.60. Severity applicability: NIH intramural cohort, severity distribution not fully generalizable.)
Limitations: Small sample, single center, highly selected NIH intramural population; the grip-force paradigm tests mean-ability fatigability, not real-world exertion.
Reputation: A single study; the central-origin conclusion requires replication before it is weighted heavily.
Falsifiable prediction: Central fatigability in ME/CFS should respond to central-acting interventions (e.g. dopaminergic or wakefulness-promoting agents) but not to peripheral ergogenic aids. The mechanistic motor-drive cascade and its differential-diagnostic drug probes are traced in Section Systemic inflammation is downstream of a more upstream cause.
Consequence: If replicated, this reframes ME/CFS fatigability as a brain-level failure to push the body, directing research and treatment toward the central nervous system rather than muscle physiology.
Subjective fatigue (what patients report) and objective fatigability (what performance tests and actigraphy measure) may not correlate closely in ME/CFS. Objective measures show reduced activity and impaired central motor drive (Liu et al. 2025), (Bedard et al. 2026), while subjective scales capture the experienced exhaustion. Whether the two orders of measurement dissociate — and what a dissociation means clinically — is unresolved, because no large study has correlated both against each other within the same ME/CFS cohort.
(Certainty: 0.50, open-question framing; no direct discordance study exists for fatigue specifically.)
Limitations: The question is posed in the absence of a dedicated subjective-vs-objective fatigue discordance study in ME/CFS; related discordance evidence comes from other autonomic measures.
Severity applicability: unknown — no discordance cohort is severity-stratified.
Consequence: An answer would tell clinicians whether a patient reporting severe fatigue but normal performance (or vice versa) is an anomaly or a meaningful subtype — which currently cannot be distinguished.
4 Fatigue Phenotypes and Subgroups
Fatigue in ME/CFS does not behave as a single uniform construct. Symptom-structure and clustering analyses suggest it is embedded within coherent multi-system symptom clusters and that severity-based subgroups exist, with some molecular evidence of differentiation from other fatigue states.
Factor analysis in 748 adults with ME/CFS identified three distinct symptom clusters — Brain (brain fog, sensory hypersensitivity, visual disturbances, sleep disturbance, headaches), Gut-Immune (gastrointestinal complaints, food intolerances, flu-like symptoms, infection susceptibility), and Autonomic (orthostatic intolerance, palpitations, thermoregulatory dysfunction) — each with strong model fit (Habermann-Horstmeier and Horstmeier 2025). Fatigue was not a separate factor but distributed across domains, suggesting it is a cross-cutting symptom whose driver may differ by cluster.
(Certainty: 0.60, large n with rigorous EFA/CFA/SEM methodology; discounted 0.60. Cross-sectional self-report.)
Limitations: Cross-sectional design; latent factors do not establish causal mechanism; self-reported symptoms only.
Reputation: Single-sample factor solution; replication in an independent cohort is needed.
Severity applicability: unknown — the factor-analysis cohort was not severity-stratified.
Falsifiable prediction: In a pre-registered trial that stratifies ME/CFS patients by their dominant symptom cluster (Brain, Gut-Immune, Autonomic) and assigns a cluster-targeted intervention (e.g., anti-inflammatory for Gut-Immune-dominant), a significant cluster × treatment interaction on a validated fatigue instrument (e.g., Chalder Fatigue Scale, score range 0–33) must emerge — fatigue should improve more in patients receiving a cluster-matched intervention than in those receiving a cluster-mismatched intervention. Absence of a significant interaction (p > 0.05 after correction for multiple comparisons, or interaction effect size η²p < 0.01) would refute the claim that fatigue is driven by cluster-specific mechanisms. Critically, the prediction is falsified if the trial shows a main effect of treatment on fatigue with no differential benefit by cluster.
Consequence: Treating fatigue as a single target may miss cluster-specific mechanisms; patients whose fatigue is driven mainly by, say, autonomic dysfunction may need a different approach than those driven by gut-immune factors.
K-means clustering of the largest Australian ME/CFS registry (n=2,873 patients vs n=797 controls) identified four symptom-severity subgroups, with health-related quality of life (HRQoL) significantly impaired across all (Eaton-Fitch and Marshall-Gradisnik 2026). Patients meeting the ICC (International Consensus Criteria) definition had the worst HRQoL outcomes. The resulting cluster structure depended on which case definition was applied, meaning how patients are classified changes how severity subgroups appear.
(Certainty: 0.65, large registry; discounted 0.65. Voluntary registry self-selection bias.)
Limitations: Self-selected registry volunteers; cross-sectional; k-means clusters are descriptive and may not capture true latent subtypes.
Reputation: Single registry; needs replication in independently recruited cohorts.
Severity applicability: ambulatory registry cohort, predominantly mild–moderate; severe/very-severe under-represented.
Falsifiable prediction: Replication of k-means clustering in an independent cohort (n ≥ 500, recruited under the same case definition as the original) must recover ≥3 of the 4 original severity subgroups (by silhouette coefficient ≥ 0.25 for the recovered solution), and those recovered clusters must show statistically significant differences (p < 0.05, corrected for multiple comparisons) on at least one external measure NOT used in the clustering — such as 7-day actigraphy step count or a different validated fatigue instrument (e.g., MFI-20 subscales). If the external measure shows no significant between-cluster difference (all pairwise contrasts p > 0.05 after Holm–Bonferroni correction), the subgroup structure is likely an artifact of the clustering variables and the claim is refuted. The prediction is also refuted if applying a different case definition to the same cohort dissolves the original 4-cluster solution.
Consequence: Recognizing that fatigue severity subgroups depend on the diagnostic criteria used helps clinicians and researchers avoid treating “fatigue severity” as a single number — the same patient may look mild or severe depending on how they are classified.
An exploratory extracellular-vesicle (EV) microRNA (miRNA) analysis distinguished ME/CFS from both idiopathic chronic fatigue and depression with a 62-miRNA signature achieving 87% sensitivity and 94% specificity (Eguchi et al. 2026). ME/CFS-derived EVs also showed a unique size and fluorescence profile, and key miRNAs (miR-21-5p, let-7f-5p, miR-26b-5p, miR-20a-5p) differed between groups.
(Certainty: 0.50, very small discovery cohort, n=6 per group; discounted 0.50. Preliminary — requires replication.)
Limitations: Tiny exploratory sample (n=6 ME/CFS); no independent validation; early-stage marker, far from a clinical test.
Reputation: Not yet replicated; discovery findings of this size are high-risk for non-reproducibility.
Severity applicability: Unknown — discovery cohort not stratified by severity.
Consequence: If confirmed, a blood-based molecular signal could eventually help distinguish ME/CFS fatigue from the fatigue of depression or idiopathic chronic fatigue — conditions that appear similar to an observer but may need different care. This is early and may prove to be a dead end.
The convergent evidence indicates that ME/CFS fatigue is not a single uniform symptom but a cross-cutting feature embedded within coherent multi-system symptom clusters — Brain, Gut-Immune, and Autonomic — rather than a discrete factor of its own (Fatigue Is Embedded Within Coherent Multi-System Symptom Clusters). Severity-based subgrouping reinforces this: four clusters emerge that shift depending on the diagnostic criteria applied, so how patients are classified changes the apparent severity structure (Severity-Based Fatigue Subgroups Exist and Depend on Case Definition). Whether the subjective experience (fatigue) and the objective measurable performance loss (fatigability) align is itself unresolved, adding a measurement dimension to the phenotypic heterogeneity (Do Subjective Fatigue and Objective Fatigability Dissociate?). (Overall certainty: 0.60-0.65 for the cluster/subgroup structure; the subjective-objective dissociation remains an open question.)
This is a descriptive structural picture, not a mechanistic one. It says fatigue is heterogeneous in how it presents and clusters, not that different fatigue types map to distinct confirmed biological causes — the molecular differentiation hint (EV-miRNA) is preliminary (Molecular Evidence That ME/CFS Fatigue Is Distinct from Other Fatigue States). The cluster structure is cross-sectional and the severity subgroups are descriptive, so neither establishes causal subtypes.
Implication: Treating all fatigue as one target may obscure subgroup-specific drivers. Assessment that combines a subjective fatigue instrument, an objective fatigability measure, and awareness of the patient’s dominant symptom cluster gives a fuller picture than a single fatigue score.
Consequence: Because multiple independent lines converge on “fatigue is heterogeneous” — cluster membership, severity subgrouping, and the subjective-objective gap — clinicians and researchers should not treat a single fatigue score as a complete description, and should design studies and treatments with the possibility of fatigue subgroups in mind, while avoiding overclaiming that distinct biological types are established.