Vascular-Centric Brainstorm: Novel Hypotheses, Treatments, and Mathematical Approaches
The growing vascular pathology literature in ME/CFS opens a set of creative mechanistic questions not yet addressed in the clinical or research literature. This section explores them speculatively.
1 Novel Mechanistic Hypotheses
In healthy subjects, the endothelial glycocalyx acts as a mechanosensor: blood flow-induced shear stress deforms glycocalyx heparan sulphate chains, activating eNOS and producing NO-mediated vasodilation calibrated to local metabolic demand. If glycocalyx degradation (via sheddases, SASP, inflammatory mediators) impairs this mechanosensing, the vascular system loses the ability to accurately calibrate local blood flow to exercise demand.
The predicted consequence: even sub-threshold exertion that should produce proportional local vasodilation instead fails to do so, delivering inadequate oxygen to contracting muscle. This would produce the characteristic ME/CFS pattern of disproportionate post-exertional symptoms at low absolute workloads — not because mitochondria are failing, but because the blood flow regulatory system cannot respond accurately.
Certainty: 0.30. Glycocalyx mechanosensing is established physiology; its impairment in ME/CFS is inferred from endothelial dysfunction data, not directly measured.
Falsifiable predictions: + ME/CFS patients should show blunted flow-mediated dilation responses that are proportionally worse during incremental exercise (not just at rest) + Syndecan-1 shedding (glycocalyx damage marker) should correlate with VE/VCO2 slope on CPET + Glycocalyx restoration (sulodexide, or recombinant heparanase inhibitor) should normalise the flow-mediated exercise vasodilation response before improving resting FMD
Research priority: Intra-exercise flow-mediated dilation measurement (using Doppler during low-level ergometry) has not been performed in ME/CFS.
The documented orthostatic hypocapnia in ME/CFS (PETCO2 26–30~mmHg during tilt ) may contribute to microclot persistence via a novel biochemical pathway: hypocapnia (alkalosis) shifts pH toward higher values. Based on general coagulation physiology (alkalosis-associated coagulopathy is well documented in trauma and critical care settings, though not specifically studied in the orthostatic context), higher pH is expected to activate several coagulation factors and inhibit tissue plasminogen activator (tPA), shifting the balance toward coagulation over fibrinolysis. No citation specific to the orthostatic/ME/CFS context is available; this link is an inference from general physiology.
In this model, every orthostatic episode in ME/CFS produces: (1) cerebral hypoperfusion via CO2-mediated vasoconstriction, (2) mild systemic alkalosis favoring coagulation over fibrinolysis, (3) incremental microclot accumulation, (4) worsened capillary obstruction, (5) worse tissue hypoxia and PEM. Over months to years, this cycle could explain the progressive capillary basement membrane thickening documented in ME/CFS .
Certainty: 0.20. Each individual link is physiologically grounded; the integrated cycle in ME/CFS has not been directly tested.
Falsifiable prediction: Tilt-induced alkalosis (measured by arterial or capillary pH) should correlate with post-tilt microclot burden (measured by fluorescence microscopy) in a within-patient comparison.
Senescent cells shed exosomes (senescence-associated extracellular vesicles, SA-EVs) that can induce senescence in adjacent non-senescent cells — the “bystander effect.” If senescent endothelial cells in ME/CFS shed SA-EVs into the circulation, these vesicles could: (1) induce senescence in distant endothelial beds not directly infected by the initial virus (2) suppress NK and T cell function via SASP cytokines carried in the exosome cargo (3) transfer procoagulant surface proteins (phosphatidylserine, tissue factor) to circulating platelets
This would explain why ME/CFS cardiovascular pathology is systemic (affecting cerebral, gut, and skeletal muscle vasculature simultaneously) despite localised initial viral injury, and why the disease is self-propagating.
Certainty: 0.15. SA-EV biology is an active research field; their role in ME/CFS is entirely speculative. No ME/CFS study has characterised extracellular vesicle senescence markers.
Research proposal: Characterise EV cargo in ME/CFS plasma for SA-EV markers (p21, p16, SA-\(\beta\)-galactosidase-containing EVs) and test whether ME/CFS EVs induce senescence in healthy endothelial cells in vitro.
Under oxidative stress, RBC membrane phosphatidylserine (PS) flips from the inner to the outer leaflet — a signal that normally marks old RBCs for phagocytic clearance. PS externalisation also directly activates platelets and promotes fibrin polymerisation on the RBC surface. ME/CFS RBCs show elevated ROS production and morphological abnormalities consistent with membrane stress (the latter documented in Long COVID RBCs , and inferred by analogy for ME/CFS given shared post-infectious context).
If ME/CFS RBCs externalise PS at elevated rates, they could act as continuous nucleation sites for fibrin microclot formation — explaining the microclot burden documented by Nunes et al. without requiring primary coagulation factor defects. This would also explain why microclot burden correlates with RBC morphological abnormality: the RBC membrane defect is the primary driver of microclot nucleation.
Certainty: 0.30. PS externalisation on ME/CFS RBCs has not been directly measured; it is inferred from documented ROS elevation and morphological abnormalities.
Falsifiable prediction: Annexin-V binding to ME/CFS RBCs (the standard assay for PS externalisation) should be elevated, and should correlate with microclot area in fluorescence microscopy of the same patient sample.
Peripheral inflammation in ME/CFS may reduce blood-to-brain transport of TSPO PET tracers (Barzon et al. 2026), creating false-negative neuroinflammation findings. Negative TSPO PET results in ME/CFS (Raijmakers et al. 2021) may reflect BBB dysfunction rather than absence of central inflammation.
Barzon et al. (2026) analyzed 358 TSPO PET scans across three tracers ([11C]-PK11195, [18F]-DPA714, [11C]-PBR28) and demonstrated that peripheral inflammation (elevated CRP) correlates with reduced tracer brain influx rate constant (K1). This effect was consistent across all tracers tested, indicating a fundamental BBB transport limitation rather than tracer-specific issue.
If ME/CFS involves systemic inflammation (as suggested by plasma proteomics and cytokine studies), TSPO PET may systematically under-detect brain microglial activation in ME/CFS patients. This confounding factor complicates interpretation of the single published ME/CFS TSPO-PET study, which found no significant difference in TSPO binding between ME/CFS patients and healthy controls (Raijmakers et al. 2021).
Under this hypothesis, three interpretations of negative TSPO PET findings in ME/CFS are possible: 1. No neuroinflammation: ME/CFS pathology is entirely peripheral (muscle, bone marrow, immune system) without central microglial involvement 2. Masked neuroinflammation: Central microglial activation exists but TSPO tracer cannot reach it due to BBB transport limitation caused by systemic inflammation 3. Mixed pathology: Peripheral immune activation drives symptoms via non-neuroinflammatory mechanisms (e.g., vagal signaling, peripheral afferent sensitization) while central inflammation is present but invisible to TSPO-PET
Falsifiable prediction: TSPO tracer brain influx rate (K1) will be lower in ME/CFS patients compared to healthy controls, particularly during PEM flare-ups when peripheral inflammation is expected to be highest. Patients with highest CRP or plasma cytokine levels will show lowest brain TSPO K1 values. If K1 values are normal in ME/CFS, BBB transport confounding is not the explanation for negative TSPO-PET findings.
Limitations: No study has directly measured TSPO tracer K1 (brain influx rate) in ME/CFS. Barzon et al. (2026) finding was in depression and schizophrenia patients, not ME/CFS. Replication status: BBB transport limitation well-established across three tracers; ME/CFS application untested. Single negative ME/CFS TSPO-PET study (Raijmakers et al. 2021) did not measure K1 and cannot distinguish between “no inflammation” and “masked inflammation” interpretations.
The paradoxical pattern suggested by preliminary Michelle James data (elevated peripheral TSPO, decreased brain TSPO) may reflect a resource reallocation mechanism: systemic inflammation drives massive TSPO upregulation in peripheral immune cells (monocytes in bone marrow, muscle macrophages), while mitochondrial dysfunction in brain microglia prevents compensatory TSPO upregulation. The brain’s high energy demand (20–25% of resting metabolism) combined with ME/CFS mitochondrial impairment may exhaust capacity for microglial activation, leaving CNS “immune-quiet” despite peripheral inflammation driving symptoms via vagal and humoral signaling.
Under this hypothesis:
- Peripheral inflammation: Drives TSPO upregulation in immune-accessible tissues (bone marrow, muscle, spleen) where cells still have mitochondrial capacity
- Central starvation: Mitochondrial dysfunction in brain microglia limits ability to upregulate TSPO even if microglia are “activated” in a functional sense
- Pathophysiological implication: ME/CFS symptoms are driven by peripheral immune activation (vagal signaling, afferent sensitization) rather than central neuroinflammation, explaining why negative TSPO PET findings coexist with systemic inflammation
This model predicts that treatments reducing peripheral immune activation (immunomodulators, senolytics) should improve symptoms without changing brain TSPO signal, and that mitochondrial support therapies may fail to improve brain TSPO binding because mitochondrial dysfunction limits capacity for activation.
Evidence Link:
- Direct: Michelle James preliminary data (conference-only) shows peripheral TSPO elevation with brain signal decrease
- Direct: Barzon 2026 shows peripheral inflammation reduces TSPO tracer brain influx (BBB transport confounding)
- Direct: Heng 2025 shows immune cell energy deficits (reduced ATP/ADP ratio)
- Supporting: Van Campen 2020-2024 series shows 91% of ME/CFS patients have CBF decline during orthostasis
- Supporting: Walitt 2024 shows central catecholamine deficiency in ME/CFS
Certainty: 0.25 (Mechanistically plausible; Michelle James data not peer-reviewed; alternative explanations exist)
Falsifiable predictions: 1. ME/CFS patients should show elevated TSPO signal in bone marrow, spleen, and postural muscles on whole-body TSPO PET, with brain signal equal to or below healthy controls 2. TSPO expression in ME/CFS peripheral monocytes should be elevated (flow cytometry), while brain microglia from post-mortem tissue (if available) should show normal or reduced TSPO 3. Peripheral inflammation markers (CRP, IL-6, TNF-alpha) should correlate positively with peripheral TSPO signal but negatively with brain TSPO signal (supporting trade-off hypothesis) 4. If peripheral TSPO elevation drives symptoms, treatments that reduce peripheral immune activation (immunomodulators, senolytics) should improve symptoms without changing brain TSPO signal
Alternative interpretation: The observed brain signal decrease could be BBB transport limitation (Barzon 2026 mechanism) rather than true absence of microglial activation. This can be distinguished by measuring tracer influx rate K1 directly.
2 Novel Treatment Proposals
If virus-induced endothelial senescence is causal , senolytics represent a rationally-targeted intervention. The dasatinib + quercetin (D+Q) combination has the strongest senolytic evidence in other conditions. A pragmatic ME/CFS pilot design:
Design: Intermittent dosing (D 100~mg + Q 1000~mg orally for 2 consecutive days per month for 3 months) — the intermittent protocol used in IPF trials, chosen because senolytics need only be present when killing senescent cells, and chronic dosing risks off-target toxicity (dasatinib is a BCR-ABL inhibitor with haematological and pleural effusion risks).
Primary outcome: SASP markers at 3 months (GDF-15, IL-6, IL-8, syndecan-1 as glycocalyx shedding index).
Secondary outcomes: EndoPAT reactive hyperemia index, 6-minute walk distance, SF-36 vitality, CPET peak VO2.
Safety monitoring: Complete blood count at each cycle (dasatinib myelosuppression risk), pleural effusion symptoms.
Key eligibility: Post-infectious ME/CFS (to select for virus-induced endothelial senescence) with elevated SASP marker baseline. Navitoclax (a BCL-2/XL inhibitor with a distinct senolytic mechanism from D+Q) would be a future escalation if D+Q shows inadequate SASP reduction, given its different target profile — though thrombocytopenia risk from BCL-XL inhibition requires dose modification protocols.
This design generates mechanistic data (does senolytic treatment reduce SASP in ME/CFS?) and safety data regardless of clinical outcome.
Prescription note: Dasatinib is a BCR-ABL tyrosine kinase inhibitor licensed for leukaemia. Off-label prescribing requires oncology familiarity with its safety profile; it is not appropriate for GP-level off-label use. The dosing detail above describes a research protocol, not a clinical recommendation for individual patients outside a supervised trial.
The Scheibenbogen-Wirth temporal model implies that early post-COVID vascular intervention might prevent ME/CFS progression. Sulodexide (glycocalyx precursor) has an established safety record and showed endothelial function improvement in Long COVID within 21 days .
Research proposal: A prevention trial enrolling COVID-19 patients within 4 weeks of infection who show early markers of endothelial glycocalyx shedding (plasma syndecan-1 \(>\) 30~ng/mL), randomised to sulodexide 500~LSU daily for 90 days vs. placebo. Primary outcome: ME/CFS incidence at 6 months (by consensus criteria). If the vascular-primary model is correct, early glycocalyx restoration should reduce ME/CFS incidence by interrupting the capillary damage-to-mitochondrial injury cascade before it becomes structural.
This trial design is feasible with existing infrastructure, uses an approved medication, and would simultaneously test the vascular-primary causal hypothesis and provide clinical benefit data.
The orthostatic symptom burden in ME/CFS appears to be amplified by at least two independent mechanisms: preload failure (reduced blood volume, venous pooling) and carotid body sensitisation (exaggerated hypoxic/hypercapnic reflexes driving dysfunctional breathing during orthostasis ). Current management addresses only the first (fludrocortisone, midodrine, IV saline).
A combination intervention targeting both:
- Volume expansion: Oral rehydration or fludrocortisone to address preload failure
- P2X3 antagonism (gefapixant): To reduce carotid body sensitisation and attenuate orthostatic hypocapnia
The hypothesis is that the two components are additive because they target mechanistically independent amplifiers. Orthostatic symptoms would be expected to improve more with the combination than either alone, measurable by tilt-table PETCO2 trajectories and OI symptom scores.
3 Endothelial Senescence as Unifying Mechanism (Long COVID ↔︎ ME/CFS)
Nunes et al.~(Nunes and Kruger 2026) propose that acute viral infection induces endothelial dysfunction and senescence at blood-brain barrier, cerebral arteries, gastrointestinal tract, and skeletal muscle. The endothelial senescence-associated secretory phenotype (SASP) is characterized by proinflammatory, prooxidative, procoagulant, and vasoconstriction-prone properties. Critically, authors distinguish endothelial dysfunction from coagulopathy: SASP is procoagulant but does not equate to overt clotting disorder (McAlpine et al. 2026) (Nunes and Kruger 2026). Study: (hypothesis framework, no primary data, certainty: Low). Testable predictions: (1) senescent endothelial cells accumulate in ME/CFS/Long COVID patients; (2) SASP factors elevated in plasma; (3) immune dysfunction prevents clearance of senescent cells; (4) senolytic therapy improves symptoms.
4 Vascular Dysfunction Biomarkers in ME/CFS (First Direct Evidence)
Heng et al.~(Heng et al. 2025) conducted multimodal analysis of ME/CFS patients, identifying elevated plasma proteins associated with thrombus formation and vascular reactivity. Importantly, study distinguishes vascular endothelial dysfunction from overt coagulopathy, aligning with both McAlpine (McAlpine et al. 2026) in Long COVID and Nunes (Nunes and Kruger 2026) theoretical framework. The study also documented immune cell energy deficits (reduced ATP/ADP ratio) and T/NK subset skewing, providing cross-system evidence of ME/CFS pathophysiology. Study: (multimodal design, proteomics, certainty: Medium-High).
5 Cross-Disease Comparison: Long COVID Vascular Biomarkers
McAlpine et al.~(McAlpine et al. 2026) found that vascular biomarkers in neuropsychiatric Long COVID differ from recovered controls and acute COVID-19. Elevated endothelial adhesion markers (sL-selectin, ADAMTS13, sP-selectin, sICAM-1) and vascular reactivity proteins (fetuin, α-2 macroglobulin) characterize Long COVID vascular pathology. Note: ADAMTS13 is a VWF-cleaving protease primarily synthesized in hepatic stellate cells; its elevation reflects thrombotic risk rather than direct endothelial adhesion. Importantly, coagulation markers (D-dimer, fibrinogen) did not differ from recovered controls, supporting endothelial dysfunction WITHOUT overt coagulopathy. These vascular biomarkers correlated with cognitive measures (fluency, verbal learning, memory) and psychiatric symptoms (depression, anxiety), linking endothelial dysfunction to “brain fog.” Study: (n=50 Long COVID, n=29 recovered controls, n=28 acute COVID-19, certainty: Medium-High).
6 Novel Hypotheses from Vascular Biomarker Findings
McAlpine et al. (McAlpine et al. 2026) demonstrated that elevated endothelial adhesion markers (sP-selectin, sICAM-1, sL-selectin, ADAMTS13) correlate with impaired cognitive performance in Long COVID, including reduced verbal fluency, learning, and memory. If ME/CFS shares the same endothelial dysfunction pathophysiology , the same biomarker-cognition relationship should hold.
The proposed mechanism: endothelial activation increases vascular permeability and reduces cerebral perfusion efficiency through microvascular adhesion and inflammatory signaling. This could explain “brain fog” and cognitive processing speed deficits in ME/CFS through a vascular-mediated pathway rather than primary neural damage.
Certainty: 0.40. The biomarker-cognition correlation is established in Long COVID; ME/CFS extrapolation is inferred from shared endothelial dysfunction patterns but not yet tested.
Falsifiable predictions: + ME/CFS patients should show the same inverse correlation between sP-selectin/sICAM-1 levels and cognitive test performance (fluency, memory) as observed in Long COVID + Vascular biomarker levels should correlate with functional MRI measures of cerebral blood flow velocity + Anti-adhesion therapy (targeting P-selectin or ICAM-1) should improve cognitive processing speed before improving subjective fatigue
Research priority: Cross-sectional ME/CFS cohort measurement of endothelial adhesion markers vs comprehensive cognitive battery — directly tests whether Long COVID biomarker-cognition pattern extends to ME/CFS.
McAlpine et al. (McAlpine et al. 2026) found that lower α1-acid glycoprotein (AGP) levels were strongly associated with poorer verbal memory, verbal learning, fluency, depression, and anxiety in Long COVID. This inverse relationship (higher AGP = better cognition) is unexpected, as AGP is typically considered an acute-phase reactant.
Potential interpretations: (1) AGP may have neuroprotective properties in chronic post-viral states independent of its role as an inflammatory marker (2) Low AGP could reflect chronic immune exhaustion or dysregulated acute-phase response (3) AGP might modulate endothelial permeability or neuroinflammation through unknown pathways
If this pattern holds in ME/CFS, AGP could serve as both a prognostic cognitive biomarker and a therapeutic target (supplementation or upregulation strategies).
Certainty: 0.25. The inverse AGP-cognition relationship is observed in Long COVID; ME/CFS relevance is entirely speculative. No mechanistic studies explain AGP’s neuroprotective role.
Research proposal: Measure AGP levels in ME/CFS patients vs controls and correlate with cognitive performance. Conduct in vitro studies of AGP’s effects on neuronal survival, endothelial permeability, and neuroinflammation models.
McAlpine et al. (McAlpine et al. 2026) found that vascular biomarker dysregulation normalized in a “late Long COVID” cohort (>3 years post-infection), while cognitive and psychiatric symptoms persisted in some individuals. This suggests a dissociation between acute vascular inflammation and chronic symptom maintenance.
If ME/CFS follows a similar temporal pattern: (1) Acute phase (0-12 months): Endothelial activation markers elevated, correlate with symptom severity (2) Subacute phase (1-3 years): Vascular markers partially normalize, but microstructural capillary damage persists (3) Chronic phase (>3 years): Vascular biomarkers return to baseline, but secondary consequences (mitochondrial dysfunction, dysautonomia, neuroinflammation) become primary drivers
This model predicts that anti-inflammatory or endothelial-protective interventions would only be effective in early disease stages, while later-stage ME/CFS requires addressing downstream damage.
Certainty: 0.35. Temporal evolution pattern observed in Long COVID; ME/CFS temporal trajectory not characterized. The dissociation hypothesis is plausible but untested.
Falsifiable predictions: + Longitudinal ME/CFS cohorts should show declining endothelial biomarker levels over time despite persistent symptoms in chronic cases + Early-stage ME/CFS patients (less than 12 months) should respond better to anti-endothelial therapies than chronic-stage patients (greater than 3 years) + Vascular biomarker levels should predict treatment response to senolytics or glycocalyx restoration therapies
Research priority: Longitudinal ME/CFS cohort with serial vascular biomarker measurement at 6, 12, 24, and 36 months post-onset — maps disease trajectory and identifies therapeutic windows.
7 Novel Treatment Proposals from Vascular Findings
If endothelial adhesion markers (sP-selectin, sICAM-1, sL-selectin) mediate cognitive impairment through vascular inflammation as suggested by McAlpine et al. (McAlpine et al. 2026), then targeting these pathways could improve cognitive function.
Potential agents:
- Crizanlizumab (anti-P-selectin monoclonal antibody): Approved for sickle cell vaso-occlusive crises; directly targets P-selectin-mediated platelet adhesion
- Alicaforsen (ICAM-1 antisense): Investigational for inflammatory conditions; reduces ICAM-1 expression on endothelium
- Small molecule selectin inhibitors: Oral agents targeting selectin-mediated leukocyte rolling
Proposed ME/CFS pilot design:
- Population: ME/CFS patients with elevated endothelial adhesion markers (sP-selectin >75th percentile)
- Intervention: Crizanlizumab 5~mg/kg IV at weeks 0, 2, 4, then monthly for 6 months total
- Primary outcome: Cognitive performance change (GNA battery: fluency, memory, processing speed)
- Secondary outcomes: Vascular biomarker levels, brain fog VAS scores, fatigue severity scales
Rationale: If endothelial adhesion drives cognitive dysfunction via reduced cerebral perfusion or neuroinflammation, then blocking this pathway should improve cognitive metrics independent of fatigue improvement. This would establish a mechanistic link and identify patients who might benefit from anti-adhesion therapy.
Safety considerations: Crizanlizumab has established safety profile in sickle cell; monitor for infusion reactions, thrombocytopenia, and infection risk. Alicaforsen has more limited safety data but oral bioavailability offers advantage.
Prescription note: Both agents are off-label for ME/CFS. Crizanlizumab requires hematology/oncology prescribing familiarity. Alicaforsen remains investigational.
McAlpine et al. (McAlpine et al. 2026) suggest that vascular biomarkers correlate with specific symptom domains (cognition, psychiatric symptoms) in Long COVID. This implies that ME/CFS patients could be stratified by biomarker profiles into mechanistically distinct subtypes:
Proposed ME/CFS vascular subtypes:
- Endothelial Adhesion-Dominant: Elevated sP-selectin/sICAM-1 → cognitive impairment, brain fog; target with anti-adhesion therapy
- Angiogenic Failure-Dominant: Reduced angiogenic capacity markers (VEGF, angiopoietin) → exercise intolerance, PEM; target with pro-angiogenic agents
- Glycocalyx Damage-Dominant: Elevated syndecan-1, heparanase → dysautonomia, orthostatic intolerance; target with glycocalyx restoration
- Senescence-Dominant: Elevated p21/p16, SASP cytokines → multisystem symptoms; target with senolytics
Research design: Multi-biomarker panel (endothelial adhesion, angiogenic, glycocalyx, senescence markers) measured in ME/CFS cohort (n=200+). Cluster analysis identifies biomarker-defined subtypes. Each subtype receives targeted therapy in randomized arms.
Clinical implications: Biomarker-guided subtype identification could explain heterogeneous treatment responses in ME/CFS, enable precision medicine approaches, and improve clinical trial design by reducing unstratified patient enrollment.
Implementation challenges: Requires validated assay panels for all biomarker classes; cost considerations for multi-marker testing; potential overlap between subtypes (patients may have mixed profiles).
8 Research Priority Matrix
The following priorities are ranked by mechanistic novelty \(\\times\) clinical impact / feasibility:
- Senescent cell burden in ME/CFS tissue: p21/p16 immunostaining in skin punch or muscle biopsies — fundamental unknowns for the senescence hypothesis; low risk, moderate cost
- HVR measurement in ME/CFS: Replicate El-Medany carotid body protocol in ME/CFS cohorts — 30~min test per patient, no intervention risk
- Intra-exercise FMD: Glycocalyx mechanosensing assessment during low-level ergometry — novel methodology, requires Doppler expertise
- PS externalisation on ME/CFS RBCs: Annexin-V flow cytometry — 2~h assay from standard blood draw, links RBC oxidative damage to microclot nucleation
- Longitudinal post-COVID biopsy series: Serial muscle biopsy at 3, 6, 12~months post-infection stratified by ME/CFS development — the definitive test of vascular-primary temporal model; invasive but scientifically decisive
- D+Q pilot senolytic trial: 30~patients, 3~months, primary endpoint SASP markers — feasibility trial that could be initiated rapidly
9 Mathematical Model Sketch: Vascular-Mitochondrial Bistable Loop
The following ODE sketch formalises the Scheibenbogen-Wirth vascular-to-mitochondrial causal chain as a dynamical system. It is intended as a framework for future quantitative modelling, not a calibrated simulation.
Let:
- \(V(t)\) = microvascular flow adequacy (dimensionless, 0 = complete obstruction, 1 = normal)
- \(M(t)\) = mitochondrial structural integrity (dimensionless, 0 = full damage, 1 = intact)
- \(\text{Na}(t)\) = intracellular sodium concentration (mM)
- \(\text{Ca}(t)\) = intramitochondrial calcium concentration (\(\mu\)M)
The coupled system:
\[ (d V)/(d t) = r_V dot (1 - V) - \delta_V dot C_\text{clot}(t) - \delta_\text{ROS} dot F(M) \]
\[ (d M)/(d t) = -k_\text{Ca} dot \text{Ca}(t) + r_M dot V(t) dot (1 - M) \]
\[ (d \text{Na})/(d t) = k_\text{NHE1} dot (1 - V) / V_0 - k_\text{NKA} dot M dot \text{Na} \]
\[ (d \text{Ca})/(d t) = k_\text{NCX} dot \text{Na}^2 - k_\text{NCLX} dot M dot \text{Ca} - k_\text{out} \]
Where:
- \(r_V\) = vascular repair rate (driven by angiogenic capacity; reduced in ME/CFS )
- \(\delta_V dot C_\text{clot}\) = microclot-driven capillary obstruction
- \(\delta_\text{ROS} dot F(M)\) = ROS from damaged mitochondria further impairing endothelium
- \(k_\text{NHE1} dot (1 - V)/V_0\) = sodium influx proportional to ischaemic hypoperfusion
- \(k_\text{NKA} dot M dot \text{Na}\) = Na/K-ATPase clearance dependent on mitochondrial integrity
- \(k_\text{NCX} dot \text{Na}^2\) = reverse-mode NCX calcium import (quadratic sodium dependence from thermodynamics)
- \(k_\text{NCLX} dot M dot \text{Ca}\) = mitochondrial calcium efflux blocked by damaged NCLX
Conjectured behaviour (not proven): This system is constructed to be consistent with bistability. For low initial microclot burden and intact \(r_V\), a fixed-point analysis would predict return to \((V, M, \text{Na}, \text{Ca}) = (1, 1, \text{Na}_0, \text{Ca}_0)\) — healthy state. Once microclot burden exceeds a threshold and/or angiogenic repair capacity (\(r_V\)) is sufficiently reduced, the structure of the equations allows a second stable fixed point at low \(V\), low \(M\), elevated \(\text{Na}\) and \(\text{Ca}\). Whether the system actually exhibits bistability — and for what parameter ranges — requires numerical bifurcation analysis and empirical parameter estimation, neither of which has been performed. This is a structural conjecture, not a demonstrated result.
Therapeutic implication from the model: If bistability is confirmed by analysis, interventions that simultaneously reduce microclot burden (restoring \(V\)) AND prevent ROS-mediated endothelial damage (the \(\delta_\text{ROS}\) term) AND restore angiogenic repair (\(r_V\)) would be needed to shift the system back to the healthy attractor. This would predict combination therapy superiority over single-agent approaches. However, this prediction is only as strong as the model’s assumptions and the yet-to-be-estimated parameters. The prediction should be treated as a hypothesis-generating result, not a clinically actionable conclusion.
Research direction: Parameter estimation for this system requires: (1) longitudinal measurement of microclot burden and capillary flow (to calibrate \(\delta_V\)), (2) serial measurement of intracellular sodium by 23Na-MRI before and during therapeutic intervention (to calibrate \(k_\text{NHE1}\) and \(k_\text{NKA}\)), (3) mitochondrial membrane potential measurements in patient biopsies (to estimate \(M(t)\) directly).
This ODE system is entirely speculative and constructed to be consistent with bistability, not to demonstrate it. No parameter values have been empirically estimated from ME/CFS patient data. Whether the system actually exhibits bistability — and for what parameter ranges — requires numerical bifurcation analysis and empirical parameter estimation, neither of which has been performed. The therapeutic implications and phase-transition predictions are hypothesis-generating, not clinically actionable, until the model is validated.
10 Research Priority Matrix
The following priorities are ranked by mechanistic novelty \(\\times\) clinical impact / feasibility:
- Senescent cell burden in ME/CFS tissue: p21/p16 immunostaining in skin punch or muscle biopsies — fundamental unknowns for the senescence hypothesis; low risk, moderate cost
- HVR measurement in ME/CFS: Replicate El-Medany carotid body protocol in ME/CFS cohorts — 30~min test per patient, no intervention risk
- Intra-exercise FMD: Glycocalyx mechanosensing assessment during low-level ergometry — novel methodology, requires Doppler expertise
- PS externalisation on ME/CFS RBCs: Annexin-V flow cytometry — 2~h assay from standard blood draw, links RBC oxidative damage to microclot nucleation
- Longitudinal post-COVID biopsy series: Serial muscle biopsy at 3, 6, 12~months post-infection stratified by ME/CFS development — the definitive test of vascular-primary temporal model; invasive but scientifically decisive
- D+Q pilot senolytic trial: 30~patients, 3~months, primary endpoint SASP markers — feasibility trial that could be initiated rapidly
11 May Thurner Syndrome: Venous Compression as a Vascular Substrate
Certainty: 0.25. May Thurner syndrome (MTS)—compression of the left common iliac vein by the overlying right common iliac artery—warrants investigation as a mechanical contributor to ME/CFS cardiovascular pathology. No direct MTS-ME/CFS evidence exists; the connection is inferred from overlapping pathophysiology.
Pathways from iliac compression to ME/CFS.
- Asymmetric preload failure. MTS sequesters ~5–10% of effective blood volume in the left lower extremity, compounding the 10–15% total blood volume deficit documented in ME/CFS (Newton et al. 2016). Unlike de novo hypovolemia, MTS-driven preload failure is mechanically asymmetric and gravity-dependent—worsening specifically during upright posture. This could explain why some ME/CFS patients show left-leg-predominant edema and disproportionate orthostatic symptoms.
- IJV flow amplification. Hartung et al. demonstrated that MTS patients lose 36% of internal jugular vein flow velocity on moving from supine to upright, versus 24% in controls (Hartung et al. 2019). If an ME/CFS patient with pre-existing cerebral hypoperfusion also has MTS, the upright IJV flow reduction is predicted to be synergistic (interaction effect, not merely additive), potentially driving greater cognitive symptoms and PEM than either condition alone. Note that “synergistic” is a structural prediction from cerebrovascular physiology, not an empirically observed effect—the specific magnitude of interaction is unknown.
- Glycocalyx shedding feedback loop. Sustained venous hypertension from MTS damages the endothelial glycocalyx, increasing capillary permeability and plasma extravasation—which further depletes circulating volume, increases orthostatic intolerance, and triggers compensatory sympathetic activation. The cycle is: MTS → venous hypertension → glycocalyx shedding → capillary leak → worsened hypovolemia → worse orthostasis → more sympathetic activation → venous pooling → back to MTS (Section Extracranial Venous Compression Impairs Glymphatic Clearance).
- Autonomic remodeling. Anderson et al. found reduced HRV, impaired Valsalva response, elevated plasma norepinephrine, and 45% orthostatic hypotension prevalence in MTS patients (Anderson et al. 2021). The COMPASS-31 autonomic score in MTS (34.2) overlaps with ME/CFS ranges, suggesting MTS alone can produce autonomic dysfunction resembling ME/CFS—and when both are present, autonomic pathology may be amplified.
Treatment evidence as indirect validation. Iliac vein stenting improves fatigue in 32–68% of MTS patients, with SF-36 vitality gains of 19 points (Wolpert et al. 2020) (O’Sullivan et al. 2018) (Ferreira et al. 2023). While these patients did not have ME/CFS, the magnitude of improvement exceeds what would be expected from regression to the mean or placebo, suggesting venous drainage can be a reversible fatigue contributor.
Falsifiable predictions. At least two of these must be confirmed for the MTS-ME/CFS link to merit clinical investigation: MTS prevalence on CT venography >35% in ME/CFS (vs. 22–24% general population); upright IJV flow velocity lower in MTS+ vs. MTS− ME/CFS, controlling for orthostatic severity; DTI-ALPS glymphatic index inversely correlated with iliac vein compression severity; stenting improves SF-36 vitality ≥10 points in MTS+ ME/CFS vs. sham or medical management.
Limitations. Anatomical MTS may be incidental. Reversible venous fatigue is distinct from ME/CFS fatigue. No study has screened an ME/CFS cohort for MTS. Stenting carries procedural risks (thrombosis, stent migration, bleeding). Causal counter-argument: if MTS is causal, why don’t 22–24% of the general population with anatomical MTS have ME/CFS? MTS must be a co-factor, not a cause, amplifying existing ME/CFS pathology rather than generating it de novo. Screening carries radiation burden (CT venography), overdiagnosis risk, and is not feasible for bedbound patients. No validated cutoff for what degree of compression is clinically significant in ME/CFS exists; the >50% stenosis threshold used in interventional MTS literature may not apply to orthostatic/neurocognitive outcomes.
Research priority. Systematic CT venography or MRV screening of an ME/CFS cohort (n ≥ 200), stratified by orthostatic severity.
12 Hypoxia-Mediated Mechanisms
Certainty: 0.50. ME/CFS may involve a failure of HIF-1\(\alpha\) stabilization and transcriptional programme activation despite chronic tissue hypoxia. Under normal physiology, tissue hypoxia triggers HIF-1\(\alpha\) stabilization, transcriptional activation of >200 target genes including EPO (erythropoiesis), VEGF (angiogenesis), GLUT1 (glycolysis), and BNIP3 (mitophagy). The predicted consequence: ME/CFS patients show blunted EPO and VEGF responses to hypoxia, with miRNA-mediated silencing of the HIF transcriptional programme (Kaczmarek 2023). Winkler et al. reported normal basal EPO in ME/CFS, suggesting the deficit may be in inducibility rather than baseline production (Winkler et al. 2004).
Subtype qualification: The pattern may differ by disease subtype. The HIF-2\(\alpha\) post-viral evidence suggests that post-viral ME/CFS patients may show elevated basal VEGF (from sustained HIF-2\(\alpha\)) rather than blunted VEGF, while gradual-onset patients may show the predicted blunted pattern (Section Immune Complex–Endothelial Injury as a Central Vascular Mechanism in Cardiovascular Dysfunction). The HIF inertia hypothesis should be understood as primarily describing the HIF-1\(\alpha\) arm; the HIF-2\(\alpha\) arm may follow a separate, subtype-dependent trajectory. Future hypoxia-challenge studies should therefore measure both HIF isoforms and stratify by onset type.
Falsifiable predictions: + ME/CFS patients exposed to mild normobaric hypoxia (FiO2 0.14, 2~h) will show peak plasma EPO elevation less than 60% of control median and peak VEGF elevation less than 50% of control median at 4–8~h + ME/CFS PBMCs under 1% O2 for 6~h will show leq 2-fold induction of HIF-1\(\alpha\) target genes (EPO, VEGF, PDK1, BNIP3) vs geq 5-fold in healthy PBMCs + Baseline miRNA-155/-210 levels will correlate with blunted HIF response (corr coeff leq -0.5 for peak EPO fold-change) and classify responders vs non-responders with AUC geq 0.80
Limitations. HIF-1\(\alpha\) protein measurement is technically challenging (rapid proteasomal degradation on reoxygenation); miRNA-HIF interaction data are from in vitro models, not patient tissue. Blunted inducibility is inferred from single-timepoint basal EPO measurements — serial hypoxia-challenge data do not exist.
Certainty: 0.40. Moderate altitude (1500–2500~m) may trigger beneficial hypoxic conditioning via mild HIF-1\(\alpha\) stabilization, improved mitochondrial efficiency, and increased erythropoiesis — an adaptive response similar to intermittent hypoxic training in athletes (Gangwar et al. 2019). However, higher altitude (>3000~m) may overwhelm the impaired cerebral autoregulation documented in ME/CFS, producing disproportionate cerebral hypoperfusion, worsening cognitive symptoms, and triggering PEM (Medow and Stewart 2024) (Badhwar et al. 2025). This biphasic dose-response curve predicts a narrow therapeutic window: moderate hypoxia may benefit, severe hypoxia harms.
Falsifiable predictions: + ME/CFS patients exposed to simulated altitude at 2000~m (FiO2 0.165) for 2~h will show composite cognitive score improvement \(>=0.5\)~SD (CANTAB/COGBAT) and CBFv within 10% of pre-exposure baseline on NIRS + The same patients exposed to 3500~m (FiO2 0.135) will show cognitive score decline \(>=0.5\)~SD, CBFv decline above 20%, and PEM incidence at least 50% at 24~h + ETCO2 decline at 3500~m will exceed the alveolar gas equation prediction by >=3~mmHg, confirming intact hypoxic ventilatory response but inadequate vasodilatory compensation
Limitations. No altitude-chamber study exists in ME/CFS. The dose-response curve is extrapolated from healthy altitude physiology and POTS/hypocapnia studies. Individual variability in hypoxic ventilatory response may mask group effects.
The hypoxia environments across this chapter and the energy-metabolism and core-mechanistic chapters assemble a candidate mechanism — one whose every component is itself inferred from other conditions or is an untested prediction, not drawn from direct ME/CFS measurement — distinct from the post-viral HIF-2\(\alpha\) endothelial story (Immune Complex–Endothelial Injury as a Central Vascular Mechanism) and from the VEGF-convergence synthesis (VEGF Pathway Dysfunction Converges Across HIF-1α/HIF-2α Dichotomous Responses: Angiogenic Failure, Sustained HIF-2α, and the sFlt-1/PlGF Biomarker Model): a failure of the HIF-1\(\alpha\) adaptive programme itself. The organising claim, labelled HIF pathway inertia (HIF Pathway Inertia: Failure of Hypoxic Adaptation in ME/CFS) — though “inertia” strictly fits only the failed-to-switch-off mode and is a loose fit for the failed-to-switch-on (blunted-induction) mode, so read it as an umbrella for HIF-1\(\alpha\) adaptive failure in either direction — is the hypothesis that, if ME/CFS tissues experience chronic hypoxia, they may fail to mount, or fail to switch off, the normal HIF-1\(\alpha\) transcriptional response, which would produce the paradox of hypoxia without adequate glycolytic/angiogenic adaptation (neither the chronic hypoxia nor the HIF-response failure has been directly measured in ME/CFS). Three speculative extensions cohere around this — each multiply determined and none an ME/CFS observation. First, the altitude paradox (The Altitude Paradox: Biphasic Response to Hypoxia in ME/CFS): a predicted, untested biphasic dose-response in which mild hypoxia might retrain a sluggish HIF response while higher altitude harms via failed cerebral autoregulation. Second, PEM as maladaptive HIF persistence (PEM as Maladaptive Persistence of HIF-Mediated Metabolic Suppression): the mirror-image speculation in which exertion-induced HIF activation fails to deactivate on reoxygenation — but PEM is itself driven by immune, metabolic, and autonomic factors, so this would at most be one contributing thread. Third, a downstream HIF-1\(\alpha\)–mitochondria–ECM triad (HIF-1alpha-Mitochondria-ECM Pathogenic Triad, HIF-1\(\alpha\)-Mediated Connective Tissue Remodeling) inferred largely from tendinopathy analogy, in which sustained HIF-1\(\alpha\) suppresses oxidative phosphorylation and drives MMP-mediated connective-tissue remodelling. A gut-derived modulator, butyrate-mediated HIF-1\(\alpha\) stabilisation via HDAC inhibition (Butyrate-Deficient Hypoxia Signaling as Persistent Inflammatory Driver), links microbiome state to this axis. If the mechanism is real, its adaptive-failure framing points to therapeutic corollaries about retraining or restoring dynamic range — intermittent hypoxic training, breathing retraining, and (with explicit caution) HIF-PH inhibitors — developed as investigational proposals in the treatment chapters. Where the mechanism would plausibly apply, if at all, is an autonomic/post-exertional subset rather than all patients.
The honest status is that HIF-1\(\alpha\) pathway inertia is a unifying hypothesis in search of its confirming measurement: no study has measured the post-exertional HIF-1\(\alpha\) target-gene time course in ME/CFS, no altitude-chamber study exists, and the ECM-triad rests on tendinopathy and connective-tissue analogy rather than primary ME/CFS tissue. The model’s value is that its two failure modes make opposite, discriminable predictions (blunted induction versus failed de-activation) that a single post-CPET transcriptional time-course study could separate — and that it explicitly competes with, rather than assumes, the alternative that tissue hypoxia in ME/CFS is adequately compensated and the pathology lies elsewhere. Where the HIF-1\(\alpha\) and HIF-2\(\alpha\) arms meet — whether, assuming either isoform is causally involved in ME/CFS at all (which has not been demonstrated), early disease is HIF-1\(\alpha\)-driven and late disease HIF-2\(\alpha\)-driven — is left as the open question already posed in VEGF Pathway Dysfunction Converges Across HIF-1α/HIF-2α Dichotomous Responses: Angiogenic Failure, Sustained HIF-2α, and the sFlt-1/PlGF Biomarker Model, not re-adjudicated here.
Consequence: This proposes that part of ME/CFS may be the body’s low-oxygen “adaptation switch” being stuck — either failing to turn on when it should or failing to turn off after exertion — which would reframe post-exertional crashes as a metabolic state that won’t reset and would make “retraining” approaches (carefully dosed intermittent hypoxia, breathing techniques) worth formally testing; but every step is currently inferred from other tissues and conditions, so the immediate payoff is a specific before/after-exercise gene-expression experiment that would confirm or refute it, not a treatment.
13 HIF-2α Mechanism and Endothelial Dysfunction Subtyping
Ribeiro et al. (2026) (Ribeiro et al. 2026) demonstrated that SARS-CoV-2 spike S1 protein drives a dichotomous hypoxia-inducible factor response in endothelial cells: transient HIF-1\(\alpha\) activation followed by sustained HIF-2\(\alpha\) nuclear localization persisting for days, with VEGF overproduction, intercellular gap formation, and increased permeability rescued by belzutifan. This establishes a direct molecular pathway from viral protein exposure to chronic endothelial dysfunction — relevant beyond COVID-19, as multiple viruses exploit HIF-2\(\alpha\) for replication (HCV (Couteaudier et al. 2025), RSV (Morris et al. 2025)) and the antiviral signalling protein IRF3 directly regulates HIF-\(\alpha\) cytoplasmic retention (Deng et al. 2026). The sustained HIF-2\(\alpha\) mechanism explains several features of post-viral ME/CFS: (1) endothelial dysfunction in the ~70% of patients without \(\beta_2\)AR autoantibodies (Stein et al. 2025), the mechanism of which has been unexplained; (2) persistence of vascular dysfunction after viral clearance; (3) the pattern of microvascular impairment documented by OCT-A and retinal vessel analysis.
Certainty: 0.55. The mechanistic chain from spike → sustained HIF-2\(\alpha\) → barrier impairment is established in vitro (Ribeiro2026, cert 0.50). Antiviral-HIF linkage (Deng2026IRF3HIFalpha, cert 0.65) and viral HIF exploitation (Couteaudier2025HIF2HCV, cert 0.55) provide convergent support. Direct ME/CFS endothelial HIF-2\(\alpha\) measurements are lacking.
Falsifiable predictions: + Nuclear HIF-2\(\alpha\) protein is elevated in dermal endothelial cells from post-viral ME/CFS patients vs healthy controls (skin biopsy IHC) + Post-viral ME/CFS patients (but not gradual-onset) show elevated plasma VEGF, vWF, and sFlt-1 compared to healthy controls + Belzutifan reduces VEGF and improves endothelial function (FMD) in post-viral ME/CFS: predicted effect size ΔFMD ≥ 20% over 8 weeks
Research priority: Skin biopsy or PBMC nuclear fraction for HIF-2\(\alpha\) IHC combined with plasma VEGF, vWF, sFlt-1 panel — cross-sectional post-viral ME/CFS (n=40) vs gradual-onset (n=20) vs healthy (n=20).
If post-viral ME/CFS involves sustained HIF-2\(\alpha\) activation (Section HIF-2\(\alpha\) Sustained Activation as the Molecular Basis for Post-Viral Endothelial Dysfunction), a single bout of maximal exercise (CPET) should produce distinct temporal profiles of nuclear HIF-1\(\alpha\) and HIF-2\(\alpha\) in PBMCs across subtypes. Prediction: healthy controls show transient HIF-1\(\alpha\) nuclear peak at 0~h post-CPET returning to baseline by 2~h; post-viral ME/CFS shows delayed HIF-2\(\alpha\) nuclear rise at 24–48~h persisting at 48~h; gradual-onset ME/CFS shows an intermediate pattern. This could serve as a dynamic functional test analogous to the glucose tolerance test. Certainty: 0.55 (temporal HIF dynamics after exercise are unmeasured in ME/CFS; the prediction follows from Ribeiro2026 in vitro kinetics applied to the post-exertional context).
Falsifiable: In serial PBMC nuclear fractions collected before and after CPET (baseline, 0 h, 2 h, 6 h, 24 h, 48 h), post-viral ME/CFS shows HIF-2\(alpha\) AUC at 2 fold controls over 0-48h, with sustained elevation at 48 h (p less than 0.01).
In preeclampsia, HIF-2\(\alpha\) activation drives excessive sFlt-1 (soluble VEGFR1, an anti-angiogenic VEGF trap) production, causing systemic endothelial dysfunction. The sFlt-1:PlGF ratio is validated clinically for preeclampsia diagnosis and risk stratification. A parallel mechanism may operate in post-viral ME/CFS: viral-triggered HIF-2\(\alpha\) activation may produce elevated sFlt-1 that functionally neutralises VEGF, creating an anti-angiogenic state despite normal or elevated total VEGF levels. The net effect would be impaired angiogenic compensation — consistent with the failed angiogenic response of ME/CFS serum (Flaskamp et al. 2022). Certainty: 0.45 (sFlt-1:HIF-2\(\alpha\) link in preeclampsia is established; no ME/CFS sFlt-1 data exist).
Falsifiable: sFlt-1:PlGF ratio is elevated in post-viral ME/CFS vs controls (mean ratio \(\geq\) 38 vs \(\leq\) 38 in controls), correlating with PEM severity and impaired angiogenic serum activity. Ratio normalises in patients who respond to belzutifan.
Certainty: 0.50. PEM may not be merely tissue damage from ischaemia-reperfusion injury after exertion (Kell and Pretorius 2022) but a maladaptive persistence of the HIF-mediated hypoxic metabolic programme. HIF-1\(\alpha\) normally suppresses oxidative phosphorylation (PDK1 inhibits pyruvate dehydrogenase), shifts metabolism to glycolysis, and induces a hibernation-mimetic state that conserves ATP during hypoxia (Hoel et al. 2021). In ME/CFS, this programme may fail to deactivate upon reoxygenation — cells remain locked in a hypoxic metabolic state despite adequate oxygen, producing sustained lactate production, impaired oxidative ATP synthesis, and prolonged recovery. miRNA-mediated HIF stabilisation may underlie this persistence (Kaczmarek 2023).
Falsifiable predictions: + After submaximal exercise, ME/CFS PBMCs will show prolonged HIF-1\(\alpha\) target gene expression (PDK1, BNIP3, LDHA) at 24–48~h compared to controls where expression returns to baseline by 6~h + PDH activity will be \(<=70\)% of pre-exercise baseline at 24~h and 48~h post-exercise in ME/CFS vs \(>90\)% at 24~h in controls + Serum lactate remains \(>=2\)~mmol/L above pre-exercise baseline at 24~h post-exercise in ME/CFS and correlates with sustained HIF target gene expression (\(r >= 0.5\))
Limitations. Post-exercise HIF target gene time-course has not been measured in ME/CFS. The hibernation-mimetic model is derived from ischaemia-reperfusion and cardiac preconditioning literature, not primary ME/CFS studies. PBMC findings may not reflect muscle or brain tissue.
The near-complete separation between healthy controls and patients by percentage basement membrane coverage in the Charlton/Slaghekke cohort (Charlton et al. 2025) (max 62.7% in HC vs min 63.2% in patients) is extraordinary — it exceeds the discriminative power of any existing ME/CFS blood-based biomarker. If validated in a larger independent cohort with formal receiver operating characteristic (ROC) analysis, EM-measured %BM coverage could serve as a histological gold-standard diagnostic biomarker for ME/CFS — analogous to duodenal biopsy for coeliac disease or renal biopsy for glomerulonephritis.
Certainty: 0.45. The core finding is replicated in 3 countries. However: (a) the diagnostic cutoff has not been formally derived by ROC analysis, (b) the Charlton/Slaghekke paper is a preprint, (c) muscle biopsy is invasive and cannot serve as a routine screening test — its diagnostic role would be as a confirmatory gold standard in equivocal cases, not as a first-line test, (d) a non-invasive surrogate (serum HA/TSG-6 panel, NIRS, or contrast-enhanced ultrasound) would be needed for routine clinical deployment.
(Origin: brainstorm.)
Falsifiable predictions: + Formal ROC analysis in a combined dataset (Amsterdam + Berlin + Aarhus biopsies, pooled n > 70 patients) yields AUC > 0.90 for %BM coverage discriminating ME/CFS from healthy controls + The optimal cutoff (Youden index) correctly classifies > 85% of patients and > 90% of controls + Serum HA and TSG-6 levels correlate with EM-measured %BM coverage (r > 0.6), enabling development of a non-invasive screening surrogate + Falsified if: %BM coverage does not discriminate ME/CFS from other fatiguing conditions (multiple sclerosis fatigue, post-cancer fatigue, depression with fatigue) — i.e., the finding is specific to fatigue states generally, not ME/CFS specifically. Falsified if: the near-complete separation in Charlton/Slaghekke is a winner’s-curse artifact of a small single-study sample and AUC < 0.75 in a pooled independent dataset
Consequence: If validated, this would give clinicians their first histological test for ME/CFS — a muscle biopsy that can objectively confirm the disease, analogous to how a duodenal biopsy confirms coeliac disease. This would end the “it’s all in your head” era by providing tissue-level proof visible under a microscope. However, the test is invasive and would be reserved for diagnostically ambiguous cases, not routine screening.
Whether capillary BM thickening is restricted to skeletal muscle or represents a systemic microvascular disease is unknown — no multi-organ biopsy data exist. If BM thickening is systemic, skeletal muscle biopsy findings provide a window into capillary pathology in every organ: cardiac muscle (explaining reduced cardiac output and chronotropic incompetence via impaired myocardial O₂ delivery), brain microvasculature (contributing to neuroinflammation and cognitive dysfunction via impaired cerebral O₂ extraction), GI smooth muscle (contributing to dysmotility and SIBO via impaired peristaltic energetics), and skin (explaining slow wound healing). The same capillary-level diffusion barrier that limits O₂ extraction in quadriceps during exercise would limit O₂ delivery to every tissue in the body — a single structural lesion producing multi-system symptoms.
Certainty: 0.30. No multi-organ capillary BM data exist for ME/CFS. The inference is from the consistency of limb muscle findings (vastus lateralis, biceps brachii, anterior tibialis all show BM thickening) plus the clinical observation that ME/CFS symptoms are multi-system. Systemic involvement is plausible but unproven. (Evidence source: limb skeletal muscle EM — Inference target: cardiac, brain, GI, skin capillaries. Link is indirect — no non-muscle capillary BM measurement in ME/CFS.)
(Origin: brainstorm.)
Falsifiable predictions: + Skin punch biopsy dermal capillary BM thickness correlates with vastus lateralis BM thickness in the same patient (r > 0.7) — confirming systemic involvement via the most accessible capillary bed + Cardiac MRI with T1 mapping (extracellular volume fraction as a fibrosis/BM proxy) correlates with skeletal muscle BM thickness + Patients with higher skeletal muscle BM thickness have slower wound healing rates, measured by standardized punch biopsy closure time + Falsified if: skin capillary BM thickness is normal in patients with confirmed skeletal muscle BM thickening — then the pathology is muscle-specific and multi-system symptoms must be explained by other mechanisms. Multi-system inference is rejected.
Consequence: If capillary wall thickening affects every organ, not just leg muscles, the disease makes sense as a unified entity — the same structural problem limiting oxygen in your muscles during exercise also limits oxygen in your brain (causing brain fog), in your gut (causing digestive problems), and in your skin (slowing wound healing). A single tissue-level defect could explain the bewildering range of symptoms patients report.
The capillary BM thickening documented in ME/CFS and Long COVID skeletal muscle (Charlton et al. 2025) (Aschman et al. 2023) is a striking structural finding, but should it be considered primary or secondary? Several alternative models are compatible with the same evidence:
Epiphenomenon of mitochondrial dysfunction. Chronic mitochondrial impairment → reduced ATP for ECM maintenance → failure to clear accumulated BM collagen → BM thickening as a consequence, not a cause. In this model, BM thickening is a marker of disease duration and severity but the rate-limiting step is mitochondrial, not microvascular.
Immune-mediated endothelial injury as primary. IgG immune complexes (Liu et al. 2026) injure endothelial cells → endothelial hypertrophy, degeneration → BM thickening as a scar-like reparative response. In this model, the immune assault on endothelium is the root cause; BM thickening is the histological scar of chronic immune-mediated injury.
Neurovascular dysregulation as primary. Impaired cholinergic vasodilation (pyridostigmine-responsive (Joseph et al. 2022)) → chronic capillary-level flow heterogeneity → hypoxic BM remodeling (HIF-1α-driven collagen IV deposition). In this model, the autonomic dysfunction documented throughout the paper is upstream of the structural pathology.
Deconditioning — incompletely excluded. The 60-day bed rest comparator (Charlton et al. 2025) showed a different phenotype (atrophy + reduced OXPHOS vs BM thickening + glycolytic shift), but bed rest in healthy young volunteers may not capture the multi-year deconditioning pattern of severe ME/CFS. A 60-day deconditioning model cannot refute a 5-year deconditioning model.
All of the above — convergent pathology. Each mechanism may contribute differently in different patients. BM thickening may be mitochondrial-driven in some, immune-driven in others, autonomic-driven in still others — but the shared endpoint (thickened capillary BM → impaired O₂ diffusion → exercise intolerance) is the same. This “final common pathway” model predicts that treatments targeting different upstream mechanisms should all partially improve O₂ extraction, but none completely normalizes it.
These alternatives are not mutually exclusive. The field’s task is to determine the temporal sequence and relative contribution of each mechanism in individual patients. Consequence: The finding that capillaries are structurally abnormal is robust, but why they’re abnormal matters for treatment. If it’s a scar from immune attack, you treat the immune system. If it’s a result of years of inactivity, you treat deconditioning. If it’s autonomic dysregulation, you treat the nerves. The right treatment depends on which explanation is correct — and we don’t yet know.
The skeletal muscle capillary basement membrane thickening documented by Charlton, Slaghekke et al. (Charlton et al. 2025) and independently replicated in Berlin (Aschman et al. 2023) and Aarhus (Agergaard et al. 2023) represents the most consistently replicated tissue-level structural abnormality in ME/CFS to date — three independent cohorts, two muscle groups, consistent core finding. The near-complete separation between patients and controls by percentage BM coverage (max HC 62.7% vs min patient 63.2%) exceeds the discriminative power of any existing blood-based biomarker and is definitively distinguished from deconditioning by a 60-day bed rest comparator showing a structurally different phenotype.
The functional consequences are independently demonstrated: impaired peripheral O₂ extraction on gold-standard invasive CPET (0.69 vs 0.77, (Joseph et al. 2022)) and reduced tissue O₂ uptake on NIRS (Charlton et al. 2025), with a pharmacological probe (pyridostigmine) confirming neurovascular dysregulation as a contributing mechanism. The endothelial hypertrophy, microvacuolization, and degenerative changes seen on EM (Charlton et al. 2025) suggest that capillary pathology extends beyond passive BM accumulation to active endothelial injury — potentially creating heterogeneous perfusion patterns (Endothelial Hypertrophy and Heterogeneous Perfusion — Functional Capillary Dropout) that NIRS averages cannot detect.
Three critical open questions remain. First, the temporal sequence: does BM thickening precede mitochondrial dysfunction (the microvascular-primacy model (Scheibenbogen and Wirth 2024)), follow it (the epiphenomenon model), or arise in parallel from a shared upstream driver (immune complexes (Liu et al. 2026), autonomic dysregulation (Joseph et al. 2022))? Second, the anatomical scope: is this a systemic capillary disease affecting every organ — potentially explaining multi-system ME/CFS symptoms from a single structural lesion (Multi-System Microvascular BM Pathology — Skeletal Muscle as the Most Accessible Window) — or a muscle-specific finding limited to exertional pathophysiology (Skeletal Muscle Specificity — A Local or Systemic Microvascular Disease?)? Third, the diagnostic potential: can EM-measured %BM coverage serve as a histological gold standard for ME/CFS diagnosis — analogous to duodenal biopsy for coeliac disease — and can a non-invasive surrogate (serum HA/TSG-6, NIRS, or CEUS) be developed for routine clinical use (Capillary Basement Membrane Coverage as a Histological Diagnostic Biomarker)? The alternative explanations enumerated in Alternative Explanations for Capillary-Level O₂ Extraction Impairment are plausibly compatible with the same evidence — the field’s task is to differentiate them.
Consequence: Tissue-level proof of structural capillary damage in ME/CFS muscle — independently confirmed in three countries — provides compelling evidence against psychosomatic/deconditioning explanations. Whatever causes ME/CFS, the capillaries are demonstrably abnormal under a microscope, and this abnormality directly impairs the muscles’ ability to extract oxygen during exercise. The question is no longer whether capillaries are structurally remodelled, but why — and whether the same remodelling affects every organ or just skeletal muscle.