The “Metabolic Trap” Hypothesis
Several researchers have proposed that ME/CFS involves metabolic “traps” — stable dysfunctional states that persist even after the initial trigger resolves.
1 IDO Metabolic Trap
One prominent hypothesis involves tryptophan metabolism:
- Trigger: Infection induces IFN-\(\gamma\), activating IDO
- Tryptophan depletion: IDO diverts tryptophan from serotonin to kynurenine
- Kynurenine effects: Metabolites may perpetuate immune activation
- Feedback loop: Chronic activation maintains the altered state
2 The “Dauer” Hypothesis
Drawing on C. elegans biology, some researchers propose ME/CFS represents a hypometabolic survival state:
- Dauer state: Nematode survival mode with reduced metabolism
- Human analog: ME/CFS as a protective metabolic downregulation
- Persistence: The hypometabolic state becomes self-perpetuating
- Treatment implications: May require specific signals to exit the state
3 The Serum-Induced Muscle Metabolic Trap: PDH → Glycolysis → Mitochondrial Collapse
Two bodies of in-vitro evidence appear to contradict each other. Fluge et al. report that ME/CFS serum drives myoblasts toward increased mitochondrial respiration with excess lactate secretion (Fluge et al. 2016), and a 3D biofabricated muscle model shows elevated oxygen consumption and mitochondrial hyperfusion after short (48 h) patient-serum exposure (Mughal et al. 2025) — a hypermetabolic, glycolytically-compensating state. In contrast, the metabolic-trap and PDH-impairment narrative predicts a hypometabolic, energetically-restricted muscle (reduced oxidative capacity, impaired pyruvate entry into the TCA cycle (Fluge et al. 2016) (Fernandez-Guerra et al. 2021)).
These are best reconciled as a temporal sequence rather than a contradiction: with pyruvate dehydrogenase impaired, the cell compensates by up-regulating glycolysis and alternative oxidative fuel oxidation, producing a transient hypermetabolic adaptation; as substrate supply, cofactor pools, or antioxidant defenses are exhausted, the system decompensates into the mitochondrial fragmentation and fragility observed after 96–144 h (Mughal et al. 2025). This maps onto the clinical PEM time-course — compensation during exertion, collapse in the recovery period. The two readings are not mutually exclusive; they describe different phases of one cycle.
Certainty: 0.26–0.52 — the biphasic temporal reading is an interpretation of concordant in-vitro findings (discounted by in-vitro population weight); the PDH impairment per se is better supported ((Fluge et al. 2016) 0.52).
Severity applicability: Unknown — the referenced in-vitro and serum-transfer studies are not stratified by severity.
Falsifiable prediction: In the 3D muscle model, measurements of PDH activity, lactate efflux, and mitochondrial morphology at serial time points (24–144 h) should show rising glycolysis and OXPHOS early, then declining electron-transport capacity and mitochondrial fragmentation after the transition point — matching, not contradicting, metabolic-trap predictions at the decompensation phase.
Consequence: What looks like a disagreement between “too active” and “too sluggish” muscle metabolism may be one cycle. That matters for treatment timing — metabolic support may need to be phase-specific — and it warns against discarding either the PDH-impairment or the hypermetabolic-adaptation evidence as incompatible.
The biphasic response to patient serum can be decomposed into three sequential phases with a specific underlying logic (Mughal et al. 2025). In Phase I (0–48 h), serum-driven pyruvate-dehydrogenase inhibition forces compensatory glycolysis and substrate-shift, producing the hypermetabolic picture (up-regulated glycolysis, elevated oxygen consumption, mitochondrial hyperfusion; matching Fluge’s finding of myoblasts in ME/CFS serum increasing respiration (Fluge et al. 2016)). In Phase II (48–96 h), sustained high-flux glycolysis depletes cofactor and antioxidant pools (NAD⁺/NADH balance, thiamine pyrophosphate, glutathione). In Phase III (96–144 h), this exhaustion tips into toroidal mitochondrial fragmentation and contractile failure — the fragile state.
This temporal decomposition reconciles what otherwise look like contradictory findings: “hypermetabolic” early data and “hypometabolic” PDH/metabolic-trap data describe different moments in one cycle, and it maps onto the clinical PEM time-course (relative compensation during exertion, collapse in recovery). Because the pyruvate-dehydrogenase node is central, cofactor repletion (thiamine, riboflavin, alpha-lipoic acid) is a mechanistically-motivated and low-risk candidate, though this remains a hypothesis to be tested in the in-vitro model before any clinical inference.
Certainty: 0.30 — a coherent reading of convergent in-vitro findings, discounted by the in-vitro population weight; the phase decomposition itself is not directly measured as a continuous trajectory.
Falsifiable prediction: Serial measurements of PDH activity, NAD⁺/NADH ratio, and mitochondrial morphology in the 3D model should show PDH decline first, then cofactor depletion, then toroidal fragmentation — and this specific ordering (not a simultaneous or reversed collapse) is what the model predicts. (Origin: brainstorm.)
Severity applicability: Unknown — in-vitro, not severity-stratified.
Consequence: If metabolic support is phase-specific, the timing of any cofactor or antioxidant intervention may matter as much as which agent is used — support given after the collapse point may arrive too late to prevent mitochondrial damage, whereas the same support during the compensatory phase could delay or lessen it.
The premise that ME/CFS serum factors are sufficient to transfer a mitochondrial phenotype to healthy myoblasts in vitro has not survived replication. A pre-registered study exposing healthy myoblasts to serum from 67 ME/CFS patients and 53 controls, across more than 1,700 Seahorse mitochondrial stress tests, found no difference in maximal respiratory capacity between the ME/CFS and control serum groups (Ryback et al. 2026). This replicates-and-extends the original Fluge serum/myoblast observation at larger scale and higher methodological standard, but with a null result (Fluge et al. 2016).
This null result narrows, rather than overturns, the serum-transfer literature. It specifically refutes the claim that a circulating ME/CFS serum factor alters healthy-myoblast mitochondrial function in the transfer model. It does NOT contradict the in-vivo and tissue-based evidence above — muscle-biopsy supercomplex/PDH findings, ³¹P-MRS flux abnormalities, exertion plasma proteomics (Germain et al. 2025), or brain ATP/PCr reduction (Weber-Fahr et al. 2026) — which rest on the patients’ own tissue rather than serum-cell transfer. The biphasic hypermetabolic/hypometabolic reading (Certainty 0.26-0.52) built partly on serum-transfer in-vitro data should therefore be weighted toward the tissue-derived evidence, and any claim resting specifically on the serum-transfer layer treated as unsupported pending further replication. (Raw certainty 0.75, pre-registered large-n replication; discounted to 0.30 for the in-vitro model gap.)
Consequence: This is a constructive negative — it removes one proposed mechanism layer (a transferable serum toxin) while leaving the tissue-intrinsic mitochondrial evidence intact, redirecting research toward what is measured in patients’ own muscle and brain rather than what patient serum does to donor cells.