5-HTP (5-Hydroxytryptophan)
Appears in: Ion Channel Hypotheses Step A4 (serotonin precursor probe), Autonomic Hypotheses, cross-reference matrix Phase 4 Group C.
Serotonin synthesis is rate-limiting; IDO tryptophan drain confirmed. Cognitive or mood improvement confirms serotonin depletion.
- Certainty
- Low
- Level of action
- Partial root cause
Serotonin synthesis not rate-limiting, or IDO drain is absent.
- Certainty
- Low
- Level of action
- Partial root cause
GI distress at 50 mg → gut serotonin hypersensitivity (Pattern 3, Side Effects as Diagnostic Probes). Serotonin syndrome at 50–100 mg → MAO-A dysfunction or AADC overactivity.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (if serotonin syndrome → metabolic cost of autonomic instability → PEM), E (discontinuation of chronic 5-HTP). Protracted risk: Low. Permanent risk: Very Low. Serotonin syndrome at standard doses is self-limited with drug clearance. The main avoidable risk: combining 5-HTP with SSRIs, SNRIs, MAOIs, or other serotonergics without washout period. Rechallenge: Yes, with careful serotonergic drug audit first.
1 Acyclovir
Appears in: Cross-reference matrix, antiviral cascade sections. Restorative — suppresses active herpesvirus replication; suppression only, not eradication — relapse on discontinuation
Herpesvirus reactivation rate-limiting — similar diagnostic to valacyclovir but with lower oral bioavailability. If acyclovir works where valacyclovir did → the patient preferentially converts the prodrug; acyclovir is the active drug. If both work → viral reactivation confirmed.
- Certainty
- Low
- Level of action
- Partial root cause
Viral reactivation not rate-limiting; or bioavailability too low. If acyclovir fails but valacyclovir works → acyclovir bioavailability (10–20%) is the bottleneck; valacyclovir prodrug conversion is required for adequate plasma levels.
- Certainty
- Low
- Level of action
- Partial root cause
CNS toxicity (confusion, hallucinations, myoclonus) at standard dose → occult renal impairment with acyclovir accumulation (Pattern K, pharmacokinetic catastrophe). The same principle as valacyclovir — acyclovir is the common active metabolite.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: K (acyclovir accumulation from impaired renal clearance → CNS toxicity) + F (unmasks occult renal impairment). Same as valacyclovir — CNS toxicity resolves with drug clearance; the renal impairment is permanent but not caused by acyclovir. Protracted risk: Low. Permanent risk: Very Low. Rechallenge: Yes, with dose adjustment for renal function. If CNS toxicity at standard dose: measure cystatin C, CrCl; adjust all renally cleared medications.
2 AIP (Autoimmune Protocol Diet)
Appears in: Cross-reference matrix Group C. Corrective — eliminates food antigens driving immune activation; if food antigens drive pathology, addresses antigen source — diagnostic probe via reintroduction
Food-antigen-driven immune activation rate-limiting. Structured elimination removes immunogenic triggers → multi-system improvement confirms food antigens are driving immune/inflammatory pathology.
- Certainty
- Low
- Level of action
- Partial root cause
Food antigens not rate-limiting. If AIP fails but IVIG or rituximab works → autoimmunity is primary, not food-antigen-driven.
- Certainty
- Low
- Level of action
- Partial root cause
SE No diagnostic information from side effects per se — the elimination/reintroduction protocol itself IS the diagnostic probe. Specific reintroduction reactions (joint pain, brain fog, GI symptoms 24–72h post-reintroduction) map individual food-antigen sensitivities.
W Mechanism: B (caloric deficit from over-restrictive elimination → metabolic PEM if intake is inadequate). The diet itself causes no pharmacologic harm. Protracted risk: Very Low. Permanent risk: Very Low. Nutritional deficiency only if maintained without guidance. Rechallenge: Always acceptable — an elimination diet is a zero-pharmacologic-risk diagnostic probe. The reintroduction phase provides the diagnostic information.
3 ALA (Alpha-Lipoic Acid)
Appears in: DMF AND vitamin C/NAC produce no improvement E2, cross-reference matrix Phase 4 Group C. HD Substrate-repletion — PDH cofactor + antioxidant; supports PDH function.
PDH cofactor + antioxidant → PDH activation rate-limiting (lactate reduction, improved exercise tolerance) and/or antioxidant capacity is the bottleneck.
- Certainty
- Low
- Level of action
- Partial root cause
Neither PDH cofactor nor antioxidant capacity rate-limiting. Lesion is elsewhere in mitochondrial chain.
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — well-tolerated.
W Mechanism: None plausibly causing protracted or permanent worsening. Paradoxical worsening (Pattern 5) identifies ETC bottleneck — diagnostic, not harmful. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable.
4 Allopregnanolone
Appears in: Cross-reference matrix Phase 4 Group C. HD Threshold-modulatory — GABA-A neurosteroid site PAM; if deficiency confirmed, replacement is substrate-repletion.
GABA-A neurosteroid site functional and rate-limiting. Allopregnanolone is a potent endogenous positive allosteric modulator of GABA-A at neurosteroid binding sites — improvement confirms GABAergic neurosteroid deficiency is the lesion.
- Certainty
- Low
- Level of action
- Partial root cause
GABA-A neurosteroid site not rate-limiting; or receptor desensitized. If allopregnanolone fails but benzodiazepines work → the GABA-A benzodiazepine site is intact; neurosteroid site is the specific lesion.
- Certainty
- Low
- Level of action
- Partial root cause
Sedation at low dose → GABA-A neurosteroid supersensitivity (Pattern 1). Sedation threshold is the ordinal readout of GABAergic neurosteroid reserve — lower threshold = greater deficiency with compensatory receptor sensitization. No sedation at therapeutic dose → GABA-A neurosteroid site desensitized (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: J (exogenous neurosteroid → potential feedback suppression of endogenous neurosteroid synthesis — theoretical; magnitude unknown). Protracted risk: Very Low. Permanent risk: Very Low. Effects reverse with drug clearance. Rechallenge: Yes. If diagnostic response: the lesion is at neurosteroid-GABA-A site specifically — distinguishes from benzodiazepine-site lesions. The combined probe (allopregnanolone + benzodiazepine) maps which GABA-A modulatory site is dysfunctional.
5 Amantadine
Appears in: Cervical collar works — structural CCI component, Neuroinflammatory Hypotheses Step K2d, Systemic inflammation is downstream of a more upstream cause Step C3, cross-reference matrix. HD Threshold-modulatory — enhances presynaptic DA release; distinguishes from DAT blockers diagnostically.
Presynaptic DA release is the bottleneck — DA stores exist but release is impaired. Amantadine enhances DA release without blocking reuptake, distinguishing from methylphenidate.
- Certainty
- Low
- Level of action
- Partial root cause
DA release machinery is intact; lesion is postsynaptic or non-dopaminergic. If methylphenidate worked but amantadine didn’t → reuptake is the bottleneck, not release.
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — well-tolerated at standard doses. Dizziness or confusion at 100 mg → subclinical renal impairment (renally cleared).
W Mechanism: B (metabolic cost of sustained dopaminergic tone — modest compared to methylphenidate), E (withdrawal after chronic use). Protracted risk: Low. Permanent risk: Very Low. Unlike amphetamines, amantadine does not reverse DAT or deplete VMAT2. Unlike methylphenidate, it does not block DAT — DA levels rise from enhanced release, not blocked clearance, producing lower peak synaptic DA and lower metabolic cost. Rechallenge: Yes, at lower dose.
6 Ambrisentan
Appears in: Cross-reference matrix Phase 4 Group C. HD Threshold-modulatory — endothelin receptor antagonist reducing pulmonary vasoconstriction; limited to documented PAH context.
Endothelin-1 antagonism → pulmonary vasodilation → improved PAH-related symptoms (limited relevance outside documented PAH with ME/CFS).
- Certainty
- Low
- Level of action
- Partial root cause
Endothelin pathway not rate-limiting.
- Certainty
- Low
- Level of action
- Partial root cause
SE Hepatotoxicity (ET receptor antagonist class effect), fluid retention, anemia.
W Mechanism: A (hepatotoxicity, teratogenicity), F (unmasking subclinical hepatic impairment). Protracted risk: Moderate. Permanent risk: Low (hepatotoxicity reversible if caught early). Rechallenge: Only if PAH is confirmed and other ET antagonists are the only option; LFT monitoring mandatory.
7 Ambroxol
Appears in: Cross-reference matrix Phase 4 Group C. HD Threshold-modulatory — TRPV1 antagonist reducing muscle pain and PEM; mucolytic if airway symptoms.
TRPV1 antagonism → muscle pain reduction, PEM reduction. Mucolytic effect (if airway symptoms present).
- Certainty
- Low
- Level of action
- Partial root cause
TRPV1 not rate-limiting for pain/PEM.
- Certainty
- Low
- Level of action
- Partial root cause
GI upset → enteric TRPV1 hypersensitivity. Confirms systemic TRPV1 engagement.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: None known. TRPV1 antagonism is reversible. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable.
8 Amitriptyline (Low-Dose)
Appears in: Inappropriate sinus tachycardia is dominant, not vascular failure. Step C1, Neuroinflammatory Hypotheses Step K2d, Dramatic cervical collar improvement confirms CCI as rate-limiting, Shoe lift improves COMPASS-31 — sympathetic chain compression driving autonomic dysfunction, cross-reference matrix. HD Threshold-modulatory — H1 + M1 + NE/5-HT reuptake inhibition at low dose; multimodal sensory modulation.
H1 + M1 + NE/5-HT reuptake inhibition → multimodal benefit for neuropathic pain, sleep, and mood if these systems are deficient.
- Certainty
- Low
- Level of action
- Partial root cause
None of these systems rate-limiting.
- Certainty
- Low
- Level of action
- Partial root cause
SE Sedation at 5–10 mg (below antidepressant dose) → H1 receptor supersensitivity. Anticholinergic symptoms at low dose → subclinical AChE deficiency or undiagnosed Sjögren’s.
W Mechanism: E (withdrawal-PEM convergence after chronic use), J (anticholinergic effects on HPA and autonomic regulation). Protracted risk: Low–Moderate. Permanent risk: Low. Withdrawal risk from chronic use (anticholinergic rebound, sleep deterioration). Rechallenge: Yes, with very slow taper planned in advance if chronic use is anticipated.
9 Amlodipine
Appears in: Cross-reference matrix. HD Threshold-modulatory — L-type Ca²⁺ channel blocker reducing TRPV1-mediated vasoconstriction downstream; does not correct TRPV1 dysfunction.
L-type Ca²⁺ channel-mediated vasoconstriction rate-limiting for microvascular perfusion deficits. TRPV1 activation on vascular smooth muscle opens L-type Ca²⁺ channels → vasoconstriction. Amlodipine blocks this downstream effect and restores perfusion.
- Certainty
- Low
- Level of action
- Partial root cause
L-type Ca²⁺ channels not rate-limiting for vascular dysfunction. If amlodipine fails but cilostazol works → microvascular pathology is platelet/endothelial, not L-type channel-mediated.
- Certainty
- Low
- Level of action
- Partial root cause
Peripheral edema without perfusion benefit → venous capacitance vessel dilation exceeding arteriolar benefit (Pattern 4) — confirms the microvascular lesion is arteriolar; the venous side is not the bottleneck. Flushing at therapeutic dose → dihydropyridine-mediated vasodilation confirms intact endothelium (informative null: absence of flush → severe endothelial dysfunction). BP drop without symptom improvement → the patient’s orthostatic tolerance depends on L-type Ca²⁺-mediated vasoconstriction (Pattern 1).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: C (paradoxical — excessive vasodilation in POTS patients with borderline cerebral perfusion → worsening orthostasis). Reversible with drug clearance (half-life 30–50h). Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Yes, at 1.25–2.5 mg with orthostatic monitoring. If BP drop without symptom improvement: L-type Ca²⁺-mediated vasoconstriction is load-bearing for orthostatic tolerance — diagnostic done; use alternative vasodilators (PDE5i, pentoxifylline).
10 Amphetamines (Dextroamphetamine, Lisdexamfetamine)
Appears in: Neuroinflammatory Hypotheses Step K2b, Systemic inflammation is downstream of a more upstream cause Step C2, cross-reference matrix. HD Symptomatic — reverses DAT + VMAT2 releasing DA/NE; most dangerous stimulant class: VMAT2 depletion, permanent terminal damage, PMC, metabolic PEM. Diagnostic value: confirms presynaptic DA terminals functional.
VMAT2 + DAT reversal → DA release confirms presynaptic DA terminals are present and functional. Cognitive improvement confirms DA is rate-limiting.
- Certainty
- Low
- Level of action
- Partial root cause
DAT absent or saturated, or DA system not bottleneck.
- Certainty
- Low
- Level of action
- Partial root cause
Cluster-dependent stimulant benefit: in the \(n = 3{,}925\) patient-reported survey, ADHD prescription stimulants helped brain fog (77.1%) and general fatigue (71.7%), but did NOT improve PEM (effect −1.5%) or POTS (−3.0%); they were beneficial in the cognitive/sleep-pain cluster (62.1% positive) where POTS is low (3.7%), yet showed no significant benefit in the multisystemic cluster (Eckey et al. 2025). A minority (17–21%) reported these medications worsened their condition. This empirically grounds the recommendation to trial stimulants ONLY in a low-POTS, cognitive-predominant profile, never in POTS-dominant or multisystemic patients.
- Certainty
- Medium
- Level of action
- Partial root cause
Severe post-dose crash indistinguishable from PEM → VMAT2 terminal compromise. See Side Effects as Diagnostic Probes.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (VMAT2 depletion → direct terminal damage) + B (metabolic PEM from sustained increased energy expenditure + tachycardia) + C (paradoxical excitation if D2 supersensitivity) + E (withdrawal dopamine crash → PEM) + H (PMC — cognitive benefit masks PEM warning, enables sustained overexertion). Protracted risk: Very High. Permanent risk: High. VMAT2 terminal dropout, mitochondrial DNA damage, and epigenetic consolidation converge — the single most dangerous medication class in this chapter. Rechallenge: NEVER. Document as severe intolerance. Counsel patient that all catecholamine-releasing agents (pseudoephedrine, ephedrine, high-dose caffeine) carry the same risk.
11 Anakinra
Appears in: Cross-reference matrix Group A, β-blocker benefit declines at a stable dose and a weekend drug holiday restores it U2. Corrective — blocks IL-1 receptor, interrupting IL-1β-driven autoinflammatory amplification; does not address IL-1β source
IL-1α and IL-1β are rate-limiting for the sickness-behavior / inflammatory phenotype. IL-1 receptor antagonism blocks both IL-1α and IL-1β — broader than canakinumab (IL-1β only). Improvement confirms IL-1-driven pathology dominates.
- Certainty
- Low
- Level of action
- Partial root cause
IL-1 not rate-limiting. If anakinra fails but canakinumab works → IL-1β alone is the driver; IL-1α blockade adds no benefit (or is counterproductive). If anakinra fails but colchicine works → the inflammasome is the lesion, but IL-1 is not the dominant downstream effector (IL-18 or pyroptosis-driven).
- Certainty
- Low
- Level of action
- Partial root cause
SE Injection-site reaction → intact innate immune recognition — confirms immune competence within the context of autoimmune pathology (Pattern 2). Severe infection on anakinra → the IL-1 pathway was load-bearing for host defense — narrows the therapeutic window: IL-1 blockade is immunosuppressive in this patient (Pattern 4).
W Mechanism: D (IL-1 blockade → immunosuppression → infection susceptibility during treatment window). IL-1 blockade is reversible with drug clearance (half-life 4–6h — short; daily dosing). Protracted risk: Low. Permanent risk: Very Low. Infection risk resolves with discontinuation. Rechallenge: Only if IL-1-driven pathology is confirmed by response and no severe infection occurred. Short half-life is an advantage — rapid reversibility if infection develops.
12 Apixaban
Appears in: Cross-reference matrix Group A. Corrective — Factor Xa inhibitor reducing microthrombi formation — bleeding risk; research-stage
Factor Xa-mediated microclot pathology rate-limiting for tissue hypoperfusion and PEM. Direct factor Xa inhibition prevents fibrin microclot formation — improvement confirms coagulation-driven microvascular occlusion is the dominant lesion.
- Certainty
- Low
- Level of action
- Partial root cause
Microclot pathology not rate-limiting, or microclots are fibrinolytic-resistant (not Factor Xa-mediated). If apixaban fails but nattokinase/lumbrokinase works → the microclot pathology is fibrinolytic-sensitive, not anticoagulation-preventable.
- Certainty
- Low
- Level of action
- Partial root cause
SE Bleeding at prophylactic dose → hemostatic reserve critically low — the coagulation system is already strained; microclots may be compensatory, not pathological (Pattern 4). No bleeding at therapeutic dose → coagulation system intact — safe to continue (Pattern 5).
W Mechanism: A (anticoagulation → bleeding risk — clinically significant if supratherapeutic or in patients with pre-existing hemostatic defects). Protracted risk: Low. Permanent risk: Very Low (unless intracranial or GI hemorrhage occurs). Rechallenge: Yes, with careful monitoring for bleeding signals. The diagnostic probe is obtained at the first bleeding signal — discontinue immediately. DOAC preferred over warfarin for predictability and lower monitoring burden. Never combine with antiplatelet agents unless specifically indicated and monitored.
13 Arginine
Appears in: Cross-reference matrix Group C. HD Substrate-repletion — NO precursor; supports endothelial function if arginine is rate-limiting.
NO-mediated vasodilation is rate-limiting for microvascular perfusion. Arginine is the NOS substrate — if supplementation improves perfusion, NOS is functional but substrate-limited.
- Certainty
- Low
- Level of action
- Partial root cause
NOS not substrate-limited; or endothelial NOS is dysfunctional (uncoupled — producing superoxide instead of NO). If citrulline works but arginine doesn’t → first-pass metabolism limits arginine; the urea cycle is intact.
- Certainty
- Low
- Level of action
- Partial root cause
Herpesvirus outbreak at standard dose → arginine feeding viral replication — confirms viral reactivation is the dominant pathology (Pattern 2). The outbreak IS diagnostic: it confirms arginine was absorbed and bioavailable, and the patient has active herpesvirus replication. GI distress → rapid bacterial arginine metabolism producing polyamines. Worsening without viral outbreak → eNOS uncoupling — arginine feeds superoxide production rather than NO (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: C (paradoxical — arginine feeding herpesvirus replication → viral flare → PEM cascade), D (if eNOS uncoupled: arginine → superoxide → oxidative damage). Protracted risk: Low. Permanent risk: Very Low (viral outbreak resolves with cessation of arginine loading; acyclovir/valacyclovir can suppress). Rechallenge: Never if herpesvirus outbreak occurred — arginine is feeding active viral replication. Use citrulline as alternative NO precursor — bypasses first-pass arginase but does not bypass viral arginine uptake. If no outbreak but worsening: eNOS likely uncoupled — NO pathway probe is completed; do not rechallenge.
14 Aripiprazole (Low-Dose)
Appears in: Cervical collar works — structural CCI component, Neuroinflammatory Hypotheses Step K2d, Cross-Hypothesis Convergence Patterns, cross-reference matrix. Clinical reference: Chapter Medication Response Reference: From Drug Response to Mechanism Identification, LDA section — response interpretation, dose-response differential (0.2–2 mg), combination inference with LDN. HD Threshold-modulatory — microglial D2 partial agonism raising activation threshold; symptomatic dopaminergic cognitive support. Does not remove what primed the microglia.
D2/D3 partial agonism → DA dysfunction rate-limiting. Lesion at or above D2/D3 receptor level.
- Certainty
- Low
- Level of action
- Partial root cause
Lesion below D2/D3 (VMAT2, ATP-dependent release, or non-dopaminergic).
- Certainty
- Low
- Level of action
- Partial root cause
Akathisia at microdose (0.25–0.5 mg) → severe DA deficiency with supersensitive D2/D3 receptors (Pattern 3). Duration of akathisia after clearance: ≤1 week → simple receptor occupancy; 2–6 weeks → receptor-state consolidation; >6 weeks → microglial triggering. If neither akathisia nor therapeutic response → D2/D3 neither deficient nor responsive (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (receptor-state consolidation — D2/D3 driven into persistent activated state), C (paradoxical — supersensitive receptors produce full-agonist-level activation from partial agonism), D (microglial triggering via D2/D3 on microglia). Protracted risk: Low–Moderate (akathisia persisting weeks beyond clearance). Permanent risk: Low (microglial triggering becoming self-sustaining — rare but theoretically plausible). Rechallenge: At 0.05–0.1 mg (liquid titration) if prior akathisia resolved with drug clearance. NEVER if akathisia persisted weeks beyond clearance or if microglial triggering suspected.
15 Ashwagandha (Withania somnifera)
Appears in: Cross-reference matrix Group C. HD Threshold-modulatory — adaptogen with GABA-A modulation + cortisol reduction.
Cortisol-mediated stress-axis pathology is rate-limiting. Ashwagandha reduces stress-induced cortisol elevation — improvement confirms cortisol excess (not deficiency) is the dominant HPA lesion.
- Certainty
- Low
- Level of action
- Partial root cause
HPA axis not cortisol-excess driven; patient may have low cortisol. If worsening → ashwagandha pushed already-low cortisol lower — confirms HPA axis is deficient, not excessive (Pattern 1).
- Certainty
- Low
- Level of action
- Partial root cause
Worsening of fatigue, orthostatic intolerance → low-cortisol phenotype confirmed — the patient has adrenal insufficiency or blunted HPA axis; cortisol-lowering agents are contraindicated (Pattern 1). This worsening IS diagnostic: it localizes the HPA lesion to deficiency, not excess. Sedation at low dose → GABAergic component potent in this patient — ashwagandha’s GABA-mimetic effects are dominant (Pattern 4).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: J (cortisol suppression in HPA-deficient patients → worsening of hypocortisolism). Protracted risk: Very Low. Permanent risk: Very Low. Cortisol suppression reverses with discontinuation. Rechallenge: Never if fatigue/orthostasis worsened — the hypocortisol phenotype is confirmed. If sedation only: lower dose; the GABAergic component is dominant. The worsening response is diagnostic information obtained at minimal risk.
16 Aspirin / NSAIDs
Appears in: Side Effects as Diagnostic Probes, cross-reference matrix. HD Threshold-modulatory — COX inhibition reducing prostaglandin-mediated inflammation/pain; does not address trigger.
COX-1/COX-2 inhibition → prostaglandin-driven inflammation is rate-limiting (TRPV1 pain, PGE2-mediated vasodilation in POTS).
- Certainty
- Low
- Level of action
- Partial root cause
Prostaglandin pathway not rate-limiting. Inflammation/pain is non-prostaglandin-mediated.
- Certainty
- Low
- Level of action
- Partial root cause
MCAS worsening at antiplatelet dose → COX-1 inhibition removes prostaglandin brake on mast cell degranulation (PGD2 was tonically inhibiting mast cells). Confirms prostaglandin-buffered MCAS subtype.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: C (paradoxical — mast cell destabilization via prostaglandin brake removal with sustained MCAS flare), D (mast cell sensitization from uncontrolled degranulation episode), G (cumulative GI mucosal damage → chronic iron-deficiency anemia in already-fatigued patient). Protracted risk: Low. Permanent risk: Very Low. MCAS flare resolves on discontinuation. Rechallenge: If aspirin worsened MCAS: trial celecoxib (COX-2 selective) — if tolerated, the protective prostaglandin is COX-1-derived (PGD2 from mast cells). If celecoxib also worsens, avoid all NSAIDs.
17 Atomoxetine
Appears in: Neuroinflammatory Hypotheses Step K2d, Aripiprazole + shoe lift response localizes dopamine lesion to Gerlier Pathway 2, cross-reference matrix. HD Threshold-modulatory — selective NET inhibitor increasing NE tone; does not correct why NE is deficient.
NE reuptake inhibition → NE deficiency rate-limiting. Cognitive benefit (PFC-dependent tasks) confirms prefrontal NE deficit.
- Certainty
- Low
- Level of action
- Partial root cause
NE not rate-limiting. Did NE levels rise? If yes but no cognitive benefit → postsynaptic α2A lesion. If no → presynaptic synthesis failure.
- Certainty
- Low
- Level of action
- Partial root cause
Tachycardia/BP elevation at subtherapeutic dose without cognitive benefit → PFC α2A desensitization (Pattern 4). Cognitive benefit without BP/HR change → DA system is the bottleneck (NET block raises NE but not DA; benefit is from NE reuptake alone).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (metabolic PEM from HR increase + increased sympathetic tone — less than methylphenidate but still present), E (withdrawal after chronic use). Protracted risk: Low–Moderate. Permanent risk: Low. Lower than methylphenidate because atomoxetine is pure NRI (no DAT block → no DA depletion). Rechallenge: Yes, at lower dose, with HR monitoring.
18 Azathioprine
Appears in: Cross-reference matrix Group B. Corrective — purine synthesis inhibitor; broad immunosuppression — high side-effect burden; requires TPMT testing
Purine-dependent lymphocyte proliferation rate-limiting — T-cell- and B-cell-driven autoimmunity confirmed. Azathioprine blocks purine synthesis → reduced lymphocyte expansion → improvement confirms lymphocyte-driven pathology is dominant.
- Certainty
- Low
- Level of action
- Partial root cause
Lymphocyte expansion not rate-limiting — non-proliferative autoimmune mechanisms (autoantibody production by long-lived plasma cells, complement activation). If azathioprine fails but rituximab works → B-cell-driven, but the lesion is mature antibody-secreting cells, not proliferating precursors.
- Certainty
- Low
- Level of action
- Partial root cause
Myelotoxicity at subtherapeutic dose (ANC < 1,500 at 50 mg/day) → TPMT deficiency / intermediate metabolizer — 6-MP accumulation from impaired methylation → the patient cannot safely metabolize azathioprine (Pattern K). The myelotoxicity at subtherapeutic dose is the diagnostic signal: TPMT genotype/phenotype is the bottleneck.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (direct hematopoietic toxicity — cumulative myelosuppression), D (prolonged immunosuppression → infection susceptibility, lymphoma risk — cumulative). Protracted risk: Moderate. Permanent risk: Low–Moderate (secondary malignancy risk with prolonged use). Rechallenge: Only if TPMT normal and response confirmed. Weekly CBC monitoring mandatory. If myelotoxicity at subtherapeutic dose: TPMT deficiency confirmed — azathioprine permanently contraindicated. Use mycophenolate (different metabolic pathway) as alternative.
19 B12 (Vitamin B12 / Cobalamin)
Appears in: Cross-reference matrix. HD Substrate-repletion — repletes B12 if deficient; supports methylation and mitochondrial function.
B12 deficiency rate-limiting for methylation, DNA synthesis, or neurological function. Improvement confirms B12 was the bottleneck — one of the few fully reversible lesions in the ME/CFS differential.
- Certainty
- Low
- Level of action
- Partial root cause
B12 not rate-limiting. Normal serum B12 + no improvement → methylation cycle competent; consider folate, SAMe, or non-methylation pathology.
- Certainty
- Low
- Level of action
- Partial root cause
No response despite documented low serum B12 → functional B12 deficiency — serum B12 is normal or elevated but cellular B12 is unavailable (transcobalamin deficiency, autoantibodies to intrinsic factor, or impaired B12 transport across the BBB). Check MMA and homocysteine — if both normal, B12 status is adequate despite low serum level.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: None plausibly causing protracted or permanent worsening. B12 is a water-soluble vitamin. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. B12 should be the first nutritional deficiency excluded — fatigue, neuropathy, and cognitive dysfunction are indistinguishable from ME/CFS symptoms. Methylcobalamin preferred over cyanocobalamin for neurological symptoms (does not require hepatic conversion).
20 Baclofen
Appears in: Cross-reference matrix. HD Threshold-modulatory — GABA-B agonist reducing muscle spasticity; does not address what causes the spasticity.
GABA-B-mediated muscle spasticity rate-limiting. GABA-B agonism reduces muscle hypertonia — confirms spinal GABA-B circuits are intact and spasticity is the dominant muscle symptom driver.
- Certainty
- Low
- Level of action
- Partial root cause
Muscle symptoms not GABA-B-mediated; or spasticity is non-neural. If baclofen fails but cyclobenzaprine works → 5-HT2-mediated muscle tension is dominant.
- Certainty
- Low
- Level of action
- Partial root cause
Sedation at 5 mg → central GABA-B supersensitivity — GABA-B circuits are hypersensitive from chronic GABA deficiency (Pattern 1). Intrathecal baclofen without sedation but with spasticity relief → the problem is spinal, not supraspinal — confirms the GABA-B lesion is at the spinal level. Paradoxical muscle stiffness → GABA-B desensitization with compensatory glutamatergic upregulation (Pattern 2).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: E (withdrawal after chronic use → GABA-B receptor downregulation → severe spasticity rebound, autonomic instability, seizures at high chronic dose). Protracted risk: Low. Permanent risk: Very Low (withdrawal resolves with slow taper). Rechallenge: Yes, at 2.5–5 mg. Never discontinue abruptly if chronic use >2 weeks — taper over weeks. If sedation at low dose: GABA-B supersensitivity confirmed — the dose reduction is the diagnostic signal.
21 BC007 / Immunoadsorption
Appears in: Autoimmune Hypotheses I1, Midodrine works but fludrocortisone does NOT → vasomotor failure, cross-reference matrix Phase 4 Group A. Restorative — removes pathogenic GPCR autoantibodies from circulation; plasma cells continue producing AAb — effect transient
GPCR AAb rate-limiting — AAb removal restores receptor function. If multi-channel improvement → PIP2 convergence. If BC007 works but IA doesn’t → BC007-targeted AAb subtype (specific β2/M2/AT1 etc.) confirmed.
- Certainty
- Low
- Level of action
- Partial root cause
AAb absent, non-pathogenic, not rate-limiting; or AAb not removed by IA column type (column mismatch). If IA fails but BC007 works → different AAb removed.
- Certainty
- Low
- Level of action
- Partial root cause
BC007 vs IA dissociation → narrows AAb subtype. No flare + no response → profound immune exhaustion or zero pathogenic AAb (Pattern 5). Hypotension during procedure → POTS with borderline cerebral perfusion.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (metabolic PEM — procedure time + hemodynamic stress + complement activation from blood-membrane contact in IA). Unlike IVIG, IA does not introduce exogenous IgG — no immune-complex formation risk. The risk is hemodynamic/metabolic: extracorporeal circuit stresses the cardiovascular system. Protracted risk: Low–Moderate. Permanent risk: Low. Procedure-induced catastrophic PEM is possible in the most severe patients but uncommon. Rechallenge: Acceptable with hemodynamic support (IV fluids, supine positioning, midodrine pre-treatment for POTS). BC007 (IV infusion without extracorporeal circuit) eliminates the hemodynamic stress of IA.
22 BCAA (Branched-Chain Amino Acids)
Appears in: Cross-reference matrix. HD Substrate-repletion — supplies branched-chain amino acids; compensates for catabolic muscle breakdown.
Catabolic amino acid depletion rate-limiting for energy and muscle function. BCAA supplementation restores branched-chain amino acid pools — confirms catabolic metabolism is actively degrading muscle for gluconeogenic substrate.
- Certainty
- Low
- Level of action
- Partial root cause
BCAA depletion not rate-limiting. Amino acid pools are adequate; energy failure is mitochondrial, not substrate-limited.
- Certainty
- Low
- Level of action
- Partial root cause
Worsening of brain fog → BCAA competition with tryptophan at the large neutral amino acid transporter → reduced brain tryptophan → reduced serotonin synthesis. Confirms the blood-brain barrier amino acid transporter is the bottleneck — BCAAs outcompete tryptophan for CNS entry (Pattern 4). No effect → amino acid depletion is not the bottleneck.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: None plausibly causing protracted or permanent worsening. BCAAs are dietary amino acids. Brain fog from tryptophan competition resolves with dose reduction or discontinuation. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Yes, with lower dose or BCAA/tryptophan co-administration. If brain fog: the CNS amino acid transporter competition is confirmed — reduce BCAA dose or take with tryptophan.
23 Belzutifan
Appears in: Connective Tissue Hypotheses N1, cross-reference matrix Group B. Corrective — HIF-2α inhibitor; if HIF-2α drives vascular pathology, targets the hypoxic signaling cascade — research-stage
HIF-2α-driven pathology rate-limiting — the endothelial glycolytic shift and angiogenic imbalance are HIF-2α-dependent. Belzutifan blocks HIF-2α transcription → improvement confirms viral-induced mitochondrial ROS → HIF-2α stabilization is the dominant pathway.
- Certainty
- Low
- Level of action
- Partial root cause
HIF-2α not rate-limiting; or the lesion is HIF-1α-driven (not HIF-2α-selective). If belzutifan fails but DMF works → the dominant HIF isoform is HIF-1α; DMF activates Nrf2 and suppresses both HIF-1α and HIF-2α downstream.
- Certainty
- Low
- Level of action
- Partial root cause
Anemia at standard dose → EPO is HIF-2α-dependent — confirms systemic HIF-2α engagement (Pattern 4). The anemia IS diagnostic: HIF-2α was load-bearing for erythropoiesis. Hypoxia symptoms without anemia → HIF-2α inhibition unmasks pre-existing pulmonary or cardiovascular limitation.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (anemia from EPO suppression → reduced oxygen-carrying capacity → tissue hypoxia → metabolic PEM in already energy-depleted patient). Protracted risk: Low–Moderate. Permanent risk: Low (anemia reverses with drug clearance or dose adjustment). Rechallenge: Only with hemoglobin monitoring and at lowest effective dose. The anemia confirms the drug is working at the molecular target — if benefit exceeds anemia cost, continue with supportive care.
24 Benzodiazepines / Z-Drugs
Appears in: Cross-reference matrix Group A (Z-drugs as zolpidem), sleep cascade sections. HD Symptomatic — GABA-A potentiation suppressing anxiety/spasm/insomnia without addressing cause — PMC (masking PEM), withdrawal risk.
GABA-A α1 system intact and responsive — sleep initiation is GABA-A-dependent. Confirms sleep onset is GABA-A modulated.
- Certainty
- Low
- Level of action
- Partial root cause
GABA-A system not rate-limiting for sleep. Sleep pathology is orexin, histaminergic, circadian, or pain-driven.
- Certainty
- Low
- Level of action
- Partial root cause
Paradoxical excitation → GABAergic inversion (NKCC1/KCC2 chloride gradient reversal, Pattern 2). Absent amnesia → hippocampal α1 desensitized from chronic GABAergic tone. Complex behaviours at therapeutic dose → striatal GABA-A dysregulation (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: C (paradoxical excitation from chloride gradient inversion), E (withdrawal-PEM convergence — benzodiazepine withdrawal is among the most severe withdrawal syndromes, with anxiety, insomnia, tachycardia, and autonomic instability persisting for months). Chronic use also causes H (PMC — anxiolysis and muscle relaxation suppress PEM warning signals). Protracted risk: Low–Moderate. Permanent risk: Low. The withdrawal-triggered PEM cascade can advance the ratchet; additionally, chronic use may cause cumulative cognitive impairment (GABAergic suppression of glutamatergic plasticity) that does not fully reverse. Rechallenge: Yes with very slow taper planned in advance (months, not weeks). A patient on clonazepam 0.5 mg nightly for 2 years should taper over 6–12 months (0.0625–0.125 mg every 4–8 weeks).
25 Berberine
Appears in: Cross-reference matrix Group C. Corrective — natural metformin analog; AMPK activation restoring metabolic efficiency — bioavailability-limited
AMPK/mTOR pathway rate-limiting for metabolic dysfunction. Berberine is a natural metformin analog — AMPK activation restores metabolic efficiency.
- Certainty
- Low
- Level of action
- Partial root cause
AMPK pathway not rate-limiting; or GI side effects prevent titration to therapeutic dose.
- Certainty
- Low
- Level of action
- Partial root cause
SE GI ceiling before metabolic benefit → gut AMPK activation is the limiting factor, not systemic (Pattern 4). Gut mitochondrial vulnerability prevents reaching therapeutic systemic dose.
W Mechanism: B (GI distress → reduced oral intake → metabolic PEM — minimal), G (cumulative GI mucosal effects with prolonged use — theoretical). Protracted risk: Very Low. Permanent risk: Very Low. Reversible on discontinuation. Rechallenge: Yes, at lower dose with gradual titration.
26 Beta-Blockers (Propranolol, Metoprolol, Bisoprolol, Nebivolol)
Appears in: Autonomic Hypotheses Q1, cross-reference matrix. HD Threshold-modulatory — reduces sympathetic tone at cardiac/vascular β-receptors; compensates for excessive adrenergic signaling without correcting the source.
Sympathetic overactivation rate-limiting for HR. β1 blockade controls tachycardia without the metabolic cost of full sympathetic suppression (if cardioselective). Nebivolol → additional NO-mediated vasodilation may benefit endothelial dysfunction.
- Certainty
- Low
- Level of action
- Partial root cause
HR not the bottleneck; or β2 metabolic cost unacceptable. Fatigue worsening → β2-AR lipolysis was a fallback energy pathway; PDH/ETC impairment (Pattern 1).
- Certainty
- Low
- Level of action
- Partial root cause
Real-world heterogeneity: in a \(n = 3{,}925\) patient-reported survey, 17–21% of patients reported that beta-blockers (alongside ADHD stimulants and Mestinon) worsened their condition, even though a majority benefited (beta-blockers/ivabradine NAS 47.1%). This quantifies the β2-metabolic-cost failure mode — a minority experience fatigue- or PEM-worsening that the mean response obscures.
- Certainty
- Medium
- Level of action
- Partial root cause
Fatigue worsening at low dose → β2-AR blockade removes lipolysis backup in PDH-impaired patients (Pattern 1). Nebivolol fatigue less than propranolol → β1-selectivity + NO benefit confirms β2 metabolic cost. Cold extremities → β2-mediated vasodilation in skin was load-bearing.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (β2 blockade → loss of lipolysis backup → mitochondrial fuel shortage), B (reduced cardiac output may trigger PEM in patients dependent on compensatory tachycardia for perfusion), F (unmasking dependency on sympathetic drive for cardiac output in stiff, hypovolemic ventricle). Protracted risk: Low. Permanent risk: Very Low. Effects reverse with drug clearance. Rechallenge: If fatigue from non-selective β-blocker: trial cardioselective (bisoprolol) or nebivolol (β1-selective + NO). If all β-blockers cause fatigue: β2-lipolysis is load-bearing → avoid β2 blockade.
27 Bromocriptine / Rotigotine
Appears in: Cross-reference matrix Group B, Systemic inflammation is downstream of a more upstream cause Step C2. HD Threshold-modulatory — dopamine agonist compensating for deficient dopamine tone.
D2/D3 postsynaptic receptors intact and responsive — same pharmacodiagnostic as pramipexole. Rotigotine patch provides continuous delivery to distinguish PK vs PD ceiling.
- Certainty
- Low
- Level of action
- Partial root cause
Lesion not at D2/D3 — synthesis, storage, or downstream signaling is the bottleneck.
- Certainty
- Low
- Level of action
- Partial root cause
Nausea/OH at RLS dose → D2 supersensitivity in area postrema and vasculature (Pattern 1). Rotigotine patch tolerated but oral bromocriptine not → delivery kinetics matter — continuous vs pulsatile D2 activation.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (DA agonist → sustained D2/D3 activation → potential receptor-state consolidation in supersensitive patients), C (paradoxical — D2 agonism in a DA-deficient system may trigger agitation or dyskinesia). Protracted risk: Low–Moderate. Permanent risk: Low (same receptor-state consolidation risk as aripiprazole). Rechallenge: At lower dose with slow titration. Rotigotine patch preferred for smoother PK.
28 Butyrate / Tributyrin
Appears in: Autonomic Hypotheses R1, cross-reference matrix Group B. HD Substrate-repletion — SCFA supporting gut barrier integrity; repletes if microbiome production inadequate.
Enteric-chromaffin-vagal pathway functional — gut dysbiosis → low butyrate was the vagal lesion. Improvement confirms gut-brain axis is intact and butyrate is the missing signal.
- Certainty
- Low
- Level of action
- Partial root cause
Enterochromaffin cells damaged (Long COVID pattern); or the lesion is efferent (DMV), not afferent. If butyrate fails but pyridostigmine works → efferent lesion.
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — GI tolerability is the main signal. No GI benefit → CNS/systemic mast cells or efferent DMV lesion dominate (Pattern 5).
W Mechanism: None plausibly causing protracted or permanent worsening. Butyrate is a gut metabolite already produced endogenously. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable.
29 Caffeine
Appears in: Cross-reference matrix. HD Symptomatic — adenosine antagonist; transient alertness — does not restore ATP production or address fatigue cause.
Catecholamine release intact — presynaptic pool not exhausted. Adenosine A2A antagonism confirms adenosine-mediated fatigue is rate-limiting for alertness.
- Certainty
- Low
- Level of action
- Partial root cause
CNS sympathetic activation failure (CNS NE deficiency). No tachycardia → central sympathetic failure — the presynaptic NE pool cannot be released. No alertness → A2A receptors desensitized from chronic adenosine elevation (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
No tachycardia at 200+ mg → central sympathetic failure (Pattern 5). No alertness despite tachycardia → postsynaptic noradrenergic lesion (α2A desensitization). Jitteriness/anxiety at low dose → supersensitive adrenergic system (Pattern 1).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (at very high doses, caffeine mobilizes intracellular calcium → sustained excitation → metabolic drain in energy-depleted neurons), B (metabolic cost of sustained sympathetic activation — modest). Protracted risk: Very Low. Permanent risk: Very Low. Caffeine’s half-life is 3–7h; effects resolve with drug clearance. Rechallenge: Always acceptable. The diagnostic information (central vs peripheral sympathetic failure) is obtained at no permanent risk.
30 Canakinumab
Appears in: Cross-reference matrix Group A, β-blocker benefit declines at a stable dose and a weekend drug holiday restores it U2. Corrective — blocks IL-1β specifically, interrupting autoinflammatory amplification; does not address IL-1β source
IL-1β specifically rate-limiting for the inflammatory sickness-behavior phenotype. Canakinumab neutralizes IL-1β without affecting IL-1α — improvement confirms IL-1β (not IL-1α) is the dominant cytokine. More specific probe than anakinra (which blocks both).
- Certainty
- Low
- Level of action
- Partial root cause
IL-1β not rate-limiting. If canakinumab fails but anakinra works → IL-1α is the dominant IL-1 driver (or dual blockade is required). If canakinumab fails but colchicine works → the NLRP3 inflammasome is rate-limiting but IL-18 or pyroptosis (not IL-1β) is the dominant downstream effector.
- Certainty
- Low
- Level of action
- Partial root cause
SE Infection on canakinumab → IL-1β host defense was load-bearing — same principle as anakinra but narrower (IL-1β-specific immunosuppression). The infection rate relative to anakinra informs whether IL-1α is load-bearing for host defense.
W Mechanism: D (IL-1β blockade → immunosuppression). The clinical trial data (CANTOS) show modest infection increase but significant cardiovascular benefit — the risk-benefit in ME/CFS is unknown. Protracted risk: Low. Permanent risk: Very Low. Rechallenge: Only with response confirmation and infection surveillance. Long half-life (~26 days) means effects persist for weeks after last dose — advantage for sustained effect, disadvantage if adverse event.
31 CBD (Cannabidiol)
Appears in: Cross-reference matrix Group C. HD Threshold-modulatory — multi-target cannabinoid system modulator reducing inflammation/anxiety.
Endocannabinoid deficiency or TRPV1-driven pathology rate-limiting. CBD is multi-target: 5-HT1A agonist (anxiolytic), TRPV1 desensitizer (pain), PPARγ activator (anti-inflammatory), adenosine reuptake inhibitor (neuroprotective). Multi-domain improvement confirms multi-target convergence.
- Certainty
- Low
- Level of action
- Partial root cause
None of CBD’s targets rate-limiting. If CBD fails but LDN works → the pathology is TLR4-dependent, not endocannabinoid/TRPV1. If CBD fails but THC-containing products work → CB1 receptor is the dominant cannabinoid target.
- Certainty
- Low
- Level of action
- Partial root cause
Fatigue worsening at low dose → 5-HT1A agonism dominant — serotonergic activation overshoots, producing sedative effects indistinguishable from PEM (Pattern 4). Anxiety at low dose → paradoxical 5-HT1A activation — the serotonergic system is in a fragile state (Pattern 2). CYP450 drug interactions → CBD inhibits CYP3A4 and CYP2C19 — reduces clearance of many ME/CFS medications.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: C (paradoxical — anxiety from 5-HT1A activation in a fragile serotonergic system). Protracted risk: Very Low. Permanent risk: Very Low. CBD does not cause structural damage. Rechallenge: Yes, at lower dose. If fatigue worsening: the serotonergic component was dominant — diagnostic; use non-serotonergic alternatives. If anxiety: the paradoxical reactor phenotype for 5-HT1A is identified — avoid serotonergic agents.
32 CBT (Cognitive-Behavioral Therapy)
Appears in: Cross-reference matrix. HD Symptomatic — modulates illness perception and coping; does not correct pathophysiology (NICE 2021) — harmful if GET-based.
Coping deficit rate-limiting for quality of life — CBT improves illness management. Does NOT address core pathophysiology. Positive response identifies the patient has reserve to engage in therapy without triggering PEM.
- Certainty
- Low
- Level of action
- Partial root cause
Coping is not the bottleneck; or cognitive demand of therapy exceeds energy envelope. If CBT fails because sessions trigger PEM → the patient’s cognitive exertion ceiling is below the therapy demand — confirms severe cognitive PEM threshold (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
PEM triggered by therapy sessions → cognitive exertion is rate-limiting — confirms the patient’s cognitive energy envelope is extremely narrow. This is a diagnostic output, not a failure of the patient.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (cognitive exertion → PEM if sessions exceed energy envelope), H (PMC — if therapist attributes symptoms to “deconditioning” or “fear-avoidance” → patient ignores PEM warnings → PEM spiral). Protracted risk: Low. Permanent risk: Very Low (PEM from sessions is reversible if caught early). Rechallenge: Only if sessions did not trigger PEM and the therapist understands ME/CFS pathophysiology. Never if therapist promotes GET-based models — this is iatrogenic harm. CBT has no role in treating core ME/CFS pathology (NICE 2021).
33 Celecoxib / Etoricoxib
Appears in: Inappropriate sinus tachycardia is dominant, not vascular failure. B2, cross-reference matrix. HD Threshold-modulatory — selective COX-2 inhibition; anti-inflammatory with cardiovascular risk.
COX-2-driven inflammation rate-limiting for pain. Selective COX-2 inhibition avoids COX-1-mediated mast cell destabilization (safer than aspirin in MCAS).
- Certainty
- Low
- Level of action
- Partial root cause
COX-2 not rate-limiting — inflammation is COX-1, complement, mast-cell, or non-inflammatory.
- Certainty
- Low
- Level of action
- Partial root cause
Fatigue worsening at 100 mg BID → COX-2-derived PGE2 was maintaining perfusion, mitochondrial function, or HPA axis integrity (Pattern 4). Distinguish: if aspirin also causes fatigue → prostaglandin pathway is load-bearing; if aspirin tolerated but celecoxib causes fatigue → COX-2-specific PGE2 was protective.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: G (cumulative COX-2 inhibition → impaired PGE2-mediated tissue repair and HPA axis support), J (COX-2-derived PGE2 tonically supports HPA axis → chronic COX-2 inhibition → gradual HPA desuppression). Protracted risk: Low. Permanent risk: Very Low. Cardiovascular risk from COX-2 inhibitors (thrombotic) is the population-level concern but not specific to ME/CFS worsening mechanism. Rechallenge: Yes, with monitoring. If aspirin worsens MCAS but celecoxib is tolerated → confirms COX-1-derived PGD2 was mast-cell protective. If both worsen → avoid all NSAIDs.
34 Cervical Collar
Appears in: Timed vitamin C + NAC response identifies ROS-mediated PHD inhibition O1, cross-reference matrix. Corrective — mechanical stabilization of craniocervical junction; if CCI is the lesion, directly addresses structural instability — diagnostic probe
CCI present and rate-limiting. If collar resolves dizziness but not POTS → non-cervical POTS mechanism. Complete resolution of brain fog and orthostatic symptoms → CCI is the dominant lesion.
- Certainty
- Low
- Level of action
- Partial root cause
CCI absent or not rate-limiting. If collar resolves nothing → non-mechanical etiology.
- Certainty
- Low
- Level of action
- Partial root cause
SE None — mechanical device without pharmacologic side effects. Skin irritation at most.
W Mechanism: None. A cervical collar is a passive mechanical support. No pharmacodynamic, metabolic, immune, or endocrine mechanism applies. Protracted risk: Very Low (muscle deconditioning from prolonged collar use — reversible with PT). Permanent risk: Very Low. Rechallenge: Always acceptable. If collar is diagnostic-positive, proceed to upright MRI with flexion/extension for surgical evaluation.
35 Cilostazol
Appears in: Cross-reference matrix Group B. HD Threshold-modulatory — PDE3 inhibitor improving microvascular perfusion; addresses perfusion deficit without correcting the cause.
PDE3-mediated microvascular pathology rate-limiting. Cilostazol increases cAMP in platelets and vascular smooth muscle → vasodilation + antiplatelet effects → improved microvascular perfusion. Confirms the combined platelet-endothelial lesion.
- Certainty
- Low
- Level of action
- Partial root cause
Microvascular pathology not PDE3-mediated. If cilostazol fails but pentoxifylline works → the perfusion deficit is RBC deformability-mediated, not cAMP/platelet-mediated. If cilostazol fails but L-type CCB works → the vascular lesion is Ca²⁺-channel-mediated, not cAMP-mediated.
- Certainty
- Low
- Level of action
- Partial root cause
Tachycardia ceiling → PDE3 inhibition in cardiac myocytes increases HR — confirms cardiac PDE3 engagement. If the HR increase triggers PEM → the perfusion benefit of vasodilation is outweighed by the metabolic cost of tachycardia (Pattern 4). Headache → PDE3-mediated cerebral vasodilation — confirms drug engagement.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (tachycardia → increased metabolic cost → PEM if benefit doesn’t exceed HR burden). PDE3 inhibition is reversible with drug clearance. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Yes, at lower dose. If tachycardia ceiling: the patient’s metabolic compensation for increased HR is insufficient — diagnostic done; use pentoxifylline (no HR effect) or L-arginine. The tachycardia identifies the PDE3-cardiac isoform sensitivity.
36 Citrulline
Appears in: Cross-reference matrix Group C. HD Substrate-repletion — arginine precursor with better bioavailability; supports NO synthesis.
Arginine-depletion-driven NO deficiency rate-limiting. Citrulline bypasses hepatic arginase → sustained arginine elevation → sustained NO production. If citrulline works but arginine doesn’t → first-pass arginase was degrading arginine in the liver; citrulline bypass confirms the NO pathway is intact downstream.
- Certainty
- Low
- Level of action
- Partial root cause
NO pathway not substrate-limited. If citrulline fails despite confirmed arginine elevation → eNOS is dysfunctional (uncoupled) or the NO receptor (sGC) is desensitized.
- Certainty
- Low
- Level of action
- Partial root cause
SE GI tolerance better than arginine → confirms arginine GI intolerance was osmotic, not NO-mediated — the patient can tolerate NO precursor therapy if the right precursor is used. This is a therapeutic advantage, not a diagnostic signal.
W Mechanism: None plausibly causing protracted or permanent worsening. Citrulline is a non-proteinogenic amino acid; excess is excreted renally. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. Citrulline is the preferred NO-donor probe over arginine — avoids first-pass metabolism, GI side effects, and herpesvirus arginine feeding. The citrulline-to-arginine conversion tests renal argininosuccinate synthetase function.
37 Clonazepam
Appears in: Cross-reference matrix, sleep cascade sections (GABA-A modulation, myoclonus/sensory hypersensitivity). HD Symptomatic — GABA-A potentiation (long half-life); suppresses symptoms without addressing cause — high PMC/withdrawal risk.
GABA-A-mediated sensory hypersensitivity or myoclonus rate-limiting. Clonazepam’s additional serotonergic effects may benefit where other benzodiazepines do not. High potency + long half-life provides sustained GABA-A potentiation.
- Certainty
- Low
- Level of action
- Partial root cause
GABA-A not rate-limiting; or the specific GABA-A subtype (non-α1) is the lesion. If clonazepam works but zolpidem (α1-selective) doesn’t → the lesion is at α2/α3/α5 GABA-A subtypes — non-sleep GABA-A functions (anxiolysis, muscle relaxation, sensory gating) are the dominant pathology.
- Certainty
- Low
- Level of action
- Partial root cause
Paradoxical excitation → GABAergic inversion (NKCC1/KCC2 chloride gradient reversal, Pattern 2) — same principle as all benzodiazepines. Tolerance development within 4–8 weeks → receptor internalization from sustained GABA-A activation — the most rapid tolerance among benzodiazepines (Pattern 4). Severe withdrawal within 24h of missed dose → high potency + long half-life creates profound dependence — the GABA-A system has compensated dramatically during chronic use.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: E (withdrawal-PEM convergence — clonazepam withdrawal is among the most severe: protracted anxiety, insomnia, dysautonomia, sensory hypersensitivity lasting months) + C (paradoxical excitation from chloride gradient inversion). Protracted risk: Moderate. Permanent risk: Low. The high potency creates a deeper dependence — clonazepam is the most dependence-producing benzodiazepine. Rechallenge: Only with 12–18 month taper planned from day 1. A patient on clonazepam 0.5 mg nightly should taper by 0.0625 mg every 8–12 weeks. For diagnostic use only: the information gained (GABA-A subtype localization) must be weighed against the withdrawal burden. Lorazepam or diazepam preferred for purely diagnostic probes.
38 Clonidine
Appears in: Cognitive dysfunction is cholinergic — basal forebrain pathology, cross-reference matrix. HD Threshold-modulatory — central α2 agonism reduces sympathetic outflow; suppresses compensatory drive without addressing why it is elevated.
Central sympathetic overactivation is rate-limiting. Brainstem baroreflex reset or GPCR AAb at area postrema → central α2 agonism reduces sympathetic outflow.
- Certainty
- Low
- Level of action
- Partial root cause
Central sympathetic drive is NOT load-bearing; α2 autoreceptors desensitized (hyperadrenergic POTS). If clonidine drops BP without symptom improvement → central α2 agonism is lowering an already-deficient sympathetic drive.
- Certainty
- Low
- Level of action
- Partial root cause
BP crash at 0.05 mg → zero sympathetic reserve — supine NE critically low, orthostatic BP completely sympathetically dependent (Pattern 1). No BP change at 0.1 mg → α2 receptors desensitized — hyperadrenergic POTS subtype confirmed (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (if BP crash → cerebral hypoperfusion → PEM from ischemic stress), C (paradoxical — in patients with already-low sympathetic reserve, clonidine suppresses the remaining drive → cardiovascular collapse). Protracted risk: Low. Permanent risk: Very Low (effects reverse with drug clearance, half-life 12–16h). Rechallenge: At 0.025 mg (quarter tablet) if initial BP crash was mild and recovered fully. Never if BP crash was severe or accompanied by syncope.
39 Clopidogrel
Appears in: Cross-reference matrix Group A. Corrective — P2Y12 inhibitor reducing platelet aggregation — research-stage for microthrombi
P2Y12-mediated platelet activation rate-limiting for microclot pathology. P2Y12 blockade reduces ADP-induced platelet aggregation — improvement confirms platelet-driven microthrombus formation is the dominant microvascular lesion.
- Certainty
- Low
- Level of action
- Partial root cause
Platelet activation not P2Y12-mediated — thrombin or collagen-driven. If clopidogrel fails but apixaban works → the coagulation cascade (fibrin formation) is dominant, not platelet aggregation. If clopidogrel fails but aspirin works → COX-1/TXA2 pathway is the dominant platelet activation route.
- Certainty
- Low
- Level of action
- Partial root cause
SE Bleeding at standard dose → hemostatic reserve minimal — same diagnostic principle as apixaban. CYP2C19 poor metabolizer → clopidogrel is a prodrug — zero antiplatelet effect confirms CYP2C19 genotype is the bottleneck (Pattern K). Bruising without clinical bleeding → platelet inhibition confirmed — the drug is working.
W Mechanism: A (antiplatelet effect → bleeding risk). Protracted risk: Low. Permanent risk: Very Low (unless intracranial or GI hemorrhage). Rechallenge: Yes, with monitoring. If zero antiplatelet effect: CYP2C19 poor metabolizer confirmed — use ticagrelor (non-prodrug) or aspirin as alternative. The CYP2C19 null response is diagnostic information about the patient’s pharmacogenetics.
40 Coenzyme Q10 (CoQ10)
Appears in: Mitochondrial Hypotheses D2, cross-reference matrix. HD Substrate-repletion — repletes CoQ10 consumed by oxidative stress; supports Complex III — does not correct the ROS source.
ETC inefficiency is functional (cofactor-compensable). CoQ10 flooding restores electron flux through Complex III — confirms the SC disruption is reversible, not structural.
- Certainty
- Low
- Level of action
- Partial root cause
SC disruption severe — cofactor flooding inadequate. Lesion is structural (Complex III protein loss) or downstream.
- Certainty
- Low
- Level of action
- Partial root cause
Worsening of fatigue/energy → Complex III block — added electrons from CoQ10 overload produce ROS rather than ATP (Pattern 5). Combined with carnitine worsening → lipid peroxide production from FAO overload meeting ETC block.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: None plausibly causing protracted or permanent worsening. Paradoxical worsening (Pattern 5) identifies Complex III bottleneck — diagnostic, not harmful. CoQ10 is an endogenous cofactor. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable — worsening is diagnostic information, not iatrogenic damage.
41 Colchicine
Appears in: Cross-reference matrix Group A, β-blocker benefit declines at a stable dose and a weekend drug holiday restores it U2. Corrective — microtubule inhibitor with NLRP3 inflammasome suppression; anti-inflammatory with narrow therapeutic window
NLRP3 inflammasome-driven pathology rate-limiting. Colchicine blocks microtubule polymerization → prevents NLRP3 inflammasome assembly + inhibits neutrophil chemotaxis. Improvement confirms the inflammasome is the dominant inflammatory driver.
- Certainty
- Low
- Level of action
- Partial root cause
NLRP3 not rate-limiting for inflammation. If colchicine fails but anakinra works → IL-1 is rate-limiting but NLRP3 is not the dominant source (other inflammasomes or non-inflammasome IL-1). If colchicine fails but prednisone works → inflammation is steroid-responsive, not inflammasome-driven.
- Certainty
- Low
- Level of action
- Partial root cause
SE GI toxicity (diarrhea) at subtherapeutic dose → microtubule-dependent enterocyte turnover is colchicine-sensitive — the drug’s therapeutic window is too narrow; the GI ceiling prevents anti-inflammatory dosing (Pattern 4). Diarrhea IS diagnostic: the patient’s enterocyte microtubule dynamics are highly colchicine-sensitive — the GI lesion may itself be microtubule-related. Neuromyopathy with prolonged use → cumulative microtubule toxicity — pre-existing SFN makes this patient vulnerable to microtubule toxins.
W Mechanism: A (cumulative microtubule toxicity → neuromyopathy — can be permanent if unrecognized, especially in patients with pre-existing SFN) + G (cumulative dose-dependent — narrow therapeutic index). Protracted risk: Moderate. Permanent risk: Low–Moderate. Colchicine-induced neuromyopathy can be permanent in patients with pre-existing IENFD reduction. Rechallenge: Only if GI ceiling was not reached and anti-inflammatory benefit confirmed. CK monitoring mandatory. If GI ceiling prevents therapeutic dose: NLRP3 inflammasome present but inaccessible via colchicine — use anakinra or canakinumab to target IL-1 downstream.
42 Cordyceps
Appears in: Cross-reference matrix Group C. HD Threshold-modulatory — fungal supplement with adenosine-like compounds — variable evidence.
Mitochondrial or immune dysfunction rate-limiting. Cordyceps (cordycepin + polysaccharides) enhances mitochondrial function and modulates immunity — improvement confirms the mitochondrial-immune axis is accessible to adaptogenic support.
- Certainty
- Low
- Level of action
- Partial root cause
Mitochondrial/immune axis not cordyceps-responsive. Low-potency nutraceutical — null non-informative.
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — well-tolerated at standard doses. No specific diagnostic side effects. Null non-informative given variable potency and standardization.
W Mechanism: None plausibly causing protracted or permanent worsening. Medicinal mushroom supplement. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. Cordyceps militaris (cultivated, standardized to cordycepin) preferred over C. sinensis (wild, inconsistent). The null is always possible from poor-quality product — source quality is a confound.
43 Corticosteroids (Prednisone, Methylprednisolone, Hydrocortisone)
Appears in: Cross-reference matrix, Hypothalamic-Pituitary-Adrenal (HPA) Axis, β-blocker benefit declines at a stable dose and a weekend drug holiday restores it U1. HD Threshold-modulatory — suppresses inflammation broadly via glucocorticoid receptor; does not address inflammatory driver — HPA suppression risk.
Inflammation is rate-limiting (does NOT identify source). Broad anti-inflammatory effect confirms inflammatory component is present and steroid-responsive.
- Certainty
- Low
- Level of action
- Partial root cause
Steroid-resistant inflammation — inflammasome, complement, or mast-cell-driven. If inflammation improves with colchicine but not prednisone → inflammasome-driven. No improvement + no crash on taper → steroid-resistant but HPA axis intact.
- Certainty
- Low
- Level of action
- Partial root cause
Crash on taper → HPA axis fragile — confirmed by morning cortisol/ACTH stimulation test (Pattern 5). Improvement during treatment + crash on taper → steroid-responsive inflammation but HPA axis was load-bearing and now suppressed. No improvement + crash on taper → HPA axis was already near-failure; steroids pushed it over the edge.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: J (HPA axis suppression → adrenal zona fasciculata atrophy → prolonged hypocortisolism) + E (corticosteroid withdrawal syndrome indistinguishable from severe ME/CFS crash — fatigue, myalgia, orthostatic intolerance, mood collapse). The two mechanisms converge: HPA axis suppression (J) produces withdrawal (E) that triggers PEM (B). Protracted risk: Moderate–High. Permanent risk: Low–Moderate. Prolonged HPA suppression → adrenal atrophy → permanent iatrogenic adrenal insufficiency if atrophy is severe. Rechallenge: Short courses only (≤5 days); taper over weeks, not days (decrease by 5–10 mg every 3–5 days for prednisone). If permanent worsening confirmed by low morning cortisol + blunted ACTH stimulation: treat with physiologic hydrocortisone replacement (15–25 mg/day divided doses). List corticosteroids as relative contraindication — life-threatening indications only.
44 CPAP
Appears in: Cross-reference matrix. Corrective — restores airway patency in OSA; if OSA confounds sleep quality, directly addresses mechanical obstruction
Obstructive sleep apnea rate-limiting for sleep quality — CPAP restores airway patency. If CPAP resolves unrefreshing sleep → OSA was the dominant sleep pathology, not ME/CFS-specific sleep dysfunction.
- Certainty
- Low
- Level of action
- Partial root cause
OSA absent or not rate-limiting. If CPAP normalizes AHI but unrefreshing sleep persists → sleep pathology is intrinsic to ME/CFS (alpha-delta intrusion, reduced SWS, glymphatic impairment), not OSA-related.
- Certainty
- Low
- Level of action
- Partial root cause
SE No symptomatic improvement despite normalized AHI → ME/CFS sleep pathology confirmed. This is the key diagnostic output — CPAP serves as a probe to exclude OSA as the confounder. Mask intolerance → standard CPAP compliance issue; not ME/CFS-specific.
W Mechanism: None. CPAP is a pneumatic device without pharmacologic effects. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. CPAP should be used whenever OSA is diagnosed in an ME/CFS patient — OSA is a treatable confounder. The null (normal AHI without symptom resolution) confirms ME/CFS-specific sleep pathology.
45 Creatine
Appears in: Mitochondrial Hypotheses D3, cross-reference matrix. HD Substrate-repletion — supplies creatine for ATP buffering via phosphocreatine system.
ATP pool insufficient — consistent with reduced steady-state ATP production. Creatine phosphate buffering capacity is rate-limiting for high-demand activities.
- Certainty
- Low
- Level of action
- Partial root cause
ATP buffering not rate-limiting. Muscle creatine stores may already be saturated or the lesion is in ATP production (not buffering).
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — well-tolerated. GI bloating at loading doses → reduce to maintenance dose (3–5 g/day).
W Mechanism: None plausibly causing protracted or permanent worsening. Creatine is an endogenous metabolite. Protracted risk: Very Low. Permanent risk: Very Low. See Mitochondrial Supplements group entry for combined analysis. Rechallenge: Always acceptable.
46 Cromolyn Sodium
Appears in: Inappropriate sinus tachycardia is dominant, not vascular failure. C2, cross-reference matrix. HD Threshold-modulatory — stabilizes mast cells preventing degranulation; does not address what activates mast cells.
Gut mast cell degranulation rate-limiting for GI symptoms. Stabilization confirms MCAS with enteric-dominant phenotype.
- Certainty
- Low
- Level of action
- Partial root cause
Gut mast cells not dominant MCAS source. CNS or systemic mast cells dominate.
- Certainty
- Low
- Level of action
- Partial root cause
SE No GI benefit → systemic MCAS (CNS, vascular, or cutaneous mast cells are the dominant source (Pattern 5)). The null is informative: directs treatment toward systemic mast-cell stabilizers (ketotifen, oral montelukast).
W Mechanism: None plausibly causing protracted or permanent worsening. Cromolyn is minimally absorbed — acts topically on gut mast cells. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable.
47 Curcumin
Appears in: Cross-reference matrix Group C. HD Threshold-modulatory — NF-κB inhibitor + antioxidant; reduces inflammatory signaling — limited bioavailability.
NF-κB-driven inflammation rate-limiting. Curcumin’s pleiotropic anti-inflammatory effects (NF-κB inhibition, COX/LOX inhibition, Nrf2 activation) confirm the NF-κB-inflammatory axis is dominant and curcumin-responsive.
- Certainty
- Low
- Level of action
- Partial root cause
NF-κB or COX/LOX not rate-limiting. If curcumin fails but celecoxib works → COX-2 inflammation is dominant but curcumin’s COX inhibition is too weak — confirms the lesion but requires pharmaceutical potency.
- Certainty
- Low
- Level of action
- Partial root cause
SE No response despite confirmed bioavailable formulation → null non-informative or the lesion is curcumin-resistant. With bioavailable curcumin (Meriva, Longvida, Theracurmin): null response means NF-κB is not the bottleneck. GI intolerance → standard curcumin GI sensitivity — switching to a more bioavailable formulation may resolve.
W Mechanism: None plausibly causing protracted or permanent worsening. Polyphenol supplement. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. Only bioavailable formulations should be used for diagnostic probing — non-bioavailable curcumin (standard turmeric powder) produces false negatives. If standard curcumin fails: trial Meriva or Longvida to exclude bioavailability confound before concluding the inflammatory pathway is curcumin-resistant.
48 Cyclobenzaprine
Appears in: Cross-reference matrix. HD Threshold-modulatory — 5-HT2 receptor antagonist + central muscle relaxant; modulates 5-HT2-mediated muscle tension.
5-HT2-mediated muscle hypertonia or central sensitization rate-limiting for pain and sleep. 5-HT2 antagonism + tricyclic-adjacent structure provides dual muscle relaxant and serotonergic effects.
- Certainty
- Low
- Level of action
- Partial root cause
Muscle symptoms not 5-HT2-mediated. If cyclobenzaprine fails but baclofen works → the lesion is GABA-B-mediated spinal spasticity, not 5-HT2-driven. If cyclobenzaprine fails but tizanidine works → α2-mediated spinal spasticity, not 5-HT2.
- Certainty
- Low
- Level of action
- Partial root cause
Next-day sedation at 5 mg → H1 + 5-HT2 supersensitivity — serotonin and histamine tone are load-bearing for wakefulness (Pattern 1). Anticholinergic symptoms (dry mouth, constipation) at low dose → M1 supersensitivity — confirms the tricyclic-adjacent anticholinergic component is dominant (Pattern 3). No sedation + no muscle relaxation at 10 mg → 5-HT2 desensitized — chronic central sensitization has desensitized serotonin receptors (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: E (withdrawal after chronic use — serotonergic rebound, sleep deterioration), H (PMC — sedation masks PEM warning, enabling overexertion). Protracted risk: Low. Permanent risk: Very Low. Rechallenge: Yes, at lower dose (2.5 mg). If next-day sedation: 5 mg is above the therapeutic ceiling for this patient — the receptor supersensitivity is the dose-limiting factor. The sedation threshold localizes the serotonin/histamine arousal reserve.
49 Cyclophosphamide
Appears in: Cross-reference matrix Group A. Corrective — broad immunosuppression depleting B and T cells; high toxicity limits use to severe refractory cases
B-cell- and T-cell-driven autoimmunity rate-limiting — severe, proliferative autoimmune disease confirmed. Cyclophosphamide’s alkylating action depletes proliferating lymphocytes — improvement confirms autoimmune pathology is severe enough to justify the toxicity.
- Certainty
- Low
- Level of action
- Partial root cause
Autoimmunity not cyclophosphamide-responsive; or the AAb source is non-proliferating (long-lived plasma cells). If cyclophosphamide fails but daratumumab works → plasma-cell-driven — cyclophosphamide cannot deplete quiescent plasma cells.
- Certainty
- Low
- Level of action
- Partial root cause
Hemorrhagic cystitis → acrolein metabolite toxicity — the bladder is the canary for alkylating toxicity (Pattern A). Bone marrow suppression at subtherapeutic dose → HSC pool exhausted — same principle as valganciclovir but via alkylation rather than viral targeting. Infertility → germ cell alkylation — confirms cumulative toxicity at the gonadal stem-cell level.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (alkylating toxicity — direct DNA damage in hematopoietic stem cells, germ cells, and bladder epithelium → cumulative and potentially permanent), G (cumulative — each cycle depletes HSC and germ-cell pools; secondary malignancy risk (MDS/AML) increases with cumulative dose). Protracted risk: High. Permanent risk: High (infertility, bone marrow failure, secondary malignancy). The most toxic medication in this chapter — reserved for life-threatening autoimmune disease refractory to all other agents. Rechallenge: Only if prior response was transformative and no permanent toxicity occurred. Permanent toxicity (infertility, MDS) is irreversible — the decision to use at all is the critical one. Use MESNA for bladder protection; cryopreserve gametes before treatment if fertility matters.
50 Cyclosporine
Appears in: Cross-reference matrix Group B. Corrective — calcineurin inhibitor suppressing T-cell activation; broad immunosuppression — nephrotoxicity risk
NFAT-mediated T-cell activation rate-limiting — confirms T-cell-driven autoimmunity. Cyclosporine blocks calcineurin → prevents NFAT dephosphorylation → blocks IL-2 transcription → T-cell suppression.
- Certainty
- Low
- Level of action
- Partial root cause
T-cell-driven autoimmunity absent or calcineurin-independent. If cyclosporine fails but tacrolimus works → the specific calcineurin inhibitor matters (FKBP12 vs cyclophilin complex). If both fail → T-cell autoimmunity is not rate-limiting.
- Certainty
- Low
- Level of action
- Partial root cause
Hypertension at low dose → calcineurin-mediated renal vasoconstriction — confirms the renal vasculature is calcineurin-sensitive (Pattern 4). Tremor → calcineurin-mediated cerebellar effect — CNS calcineurin engagement confirmed. Nephrotoxicity → cumulative calcineurin-mediated renal vasoconstriction + tubulointerstitial fibrosis.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (nephrotoxicity — cumulative calcineurin inhibition → renal vasoconstriction → ischemic injury → irreversible tubulointerstitial fibrosis), G (cumulative — each year of therapy incrementally reduces GFR). Protracted risk: Moderate. Permanent risk: Low–Moderate. Nephrotoxicity is partially reversible early; irreversible fibrosis develops with prolonged use. Rechallenge: Only with documented T-cell autoimmunity, monthly renal function monitoring, and lowest effective trough level. If nephrotoxicity or hypertension: the renal calcineurin sensitivity is confirmed — dose reduction or tacrolimus switch. The T-cell-autoimmune confirmation must justify renal risk.
51 D-Ribose
Appears in: Cross-reference matrix. HD Substrate-repletion — supplies ribose for de novo purine synthesis; bypasses impaired ATP regeneration.
Purine salvage pathway is rate-limiting for ATP regeneration. D-ribose bypasses the de novo purine synthesis bottleneck.
- Certainty
- Low
- Level of action
- Partial root cause
ATP regeneration not purine-limited. Lesion is in ETC, PDH, or substrate availability — not purine synthesis.
- Certainty
- Low
- Level of action
- Partial root cause
Hypoglycemia → impaired gluconeogenesis — ribose may stimulate insulin release or compete with glucose metabolism (Pattern 4). GI symptoms at low doses → rapid absorption intolerance.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: None plausibly causing protracted or permanent worsening. D-ribose is a sugar molecule already present in every cell. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable.
52 Daratumumab (Anti-CD38)
Appears in: Autoimmune Hypotheses I1, cross-reference matrix Group B. Corrective — depletes CD38+ long-lived plasma cells; targets the source rituximab spares — research-stage in ME/CFS
AAb from CD38+ long-lived plasma cells — works where rituximab (CD20) fails. Confirms plasma-cell-driven autoimmunity is rate-limiting. If rituximab transient response → daratumumab sustained response → plasma cells, not B cells, drive chronicity.
- Certainty
- Low
- Level of action
- Partial root cause
AAb source not accessible to B-cell or plasma-cell depletion — AAb from short-lived plasmablasts (rituximab-sensitive but not daratumumab-specific), CD38− cells, or non-immune AAb source.
- Certainty
- Low
- Level of action
- Partial root cause
Infusion reactions → intact immune recognition and complement activation (Pattern 2). Neutropenia → marrow CD38+ myeloid precursors affected — hematopoietic reserve narrow (Pattern 4).
- Certainty
- Low
- Level of action
- Partial root cause
Viral dUTPase drives the TFH/LLPC axis daratumumab depletes — EBV/HHV-6A dUTPases promote TFH differentiation, activin A and IL-21 secretion, and extrafollicular antibody responses in ME/CFS, plausibly generating the CD38+ long-lived plasma cells (LLPCs) that produce pathogenic autoantibodies (Cox et al. 2022). Provides the viral-trigger → cell-type bridge explaining why plasma-cell depletion (daratumumab) is mechanistically targeted where B-cell depletion (rituximab) failed: the relevant antibody source is the LLPC, not the cycling CD20+ pool.
- Certainty
- Medium
- Level of action
- Partial root cause
Pre-existing NK-cell deficiency predicts daratumumab non-response — NK cells are an on-target CD38+ population depleted by daratumumab, and Long COVID studies show marked CD56+CD16+ NK depletion and reduced inflammatory capacity in post-viral fatigue (Ray et al. 2026). Consistent with the Fluge pilot observation that low baseline NK count predicted poor response (Fluge et al. 2025). In NK-depleted patients the on-target immune loss is compounded and infection-risk is elevated during the immunosuppressed window.
- Certainty
- Medium
- Level of action
- Partial root cause
W Mechanism: A (CD38+ hematopoietic stem cell depletion → marrow suppression), B (infusion reaction → inflammatory flare → PEM), D (immune dysregulation from plasma-cell depletion → loss of protective antibodies → infection susceptibility). Protracted risk: Moderate. Permanent risk: Low–Moderate. Marrow suppression and infection risk are reversible with time and supportive care, but severe infection during the immunosuppressed window can cause permanent decline through PEM. Rechallenge: Only with infection prophylaxis, immunoglobulin monitoring, and careful risk-benefit assessment. Reserved for severe, rituximab-refractory, AAb-confirmed cases.
53 DCA (Dichloroacetate)
Appears in: DMF AND vitamin C/NAC produce no improvement E2, cross-reference matrix Group B. Corrective — PDK inhibitor restoring PDH activity and pyruvate entry into TCA cycle — neurotoxicity risk at cumulative dose
PDH phosphorylation rate-limiting for metabolic block. PDK inhibition → PDH activation → lactate reduction confirms PDH is the bottleneck.
- Certainty
- Low
- Level of action
- Partial root cause
PDH not rate-limiting. Metabolic block is downstream of PDH (ETC, TCA, substrate).
- Certainty
- Low
- Level of action
- Partial root cause
Neuropathy at low cumulative dose → pre-existing SFN vulnerability (Pattern 4). No lactate reduction at therapeutic dose → PDH not the bottleneck (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (direct cumulative axonal toxicity) + G (cumulative dose-dependent — above ~25–30 g lifetime dose, irreversible neuropathy risk increases substantially; ME/CFS patients with pre-existing IENFD reduction have lower threshold). Protracted risk: Moderate. Permanent risk: Moderate. DCA-induced neuropathy can be permanent — nerve regeneration machinery is already stressed in ME/CFS. Rechallenge: Only if no neuropathy develops, PDH is the last remaining hypothesis, and cumulative lifetime dose is monitored. Thiamine (which supports PDH without neurotoxicity) should be trialed first. If pre-existing IENFD reduction is documented, DCA is relatively contraindicated.
54 Devil’s Claw (Harpagophytum procumbens)
Appears in: Cross-reference matrix Group C. HD Threshold-modulatory — anti-inflammatory via COX-2 and iNOS inhibition — variable evidence.
COX/TNF-α-mediated inflammation rate-limiting. Herbal anti-inflammatory effect confirms inflammatory component is present and COX/TNF-α-dependent.
- Certainty
- Low
- Level of action
- Partial root cause
COX/TNF-α not rate-limiting for pain. Inflammation is non-COX, non-TNF-α mediated.
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — mild GI upset at high doses. Null is non-informative given low potency.
W Mechanism: None plausibly causing protracted or permanent worsening. Herbal supplement with COX/TNF-α inhibition far weaker than pharmaceutical NSAIDs. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable.
55 DHEA (Dehydroepiandrosterone)
Appears in: Cross-reference matrix Group C. HD Substrate-repletion — neurosteroid/androgen precursor; repletes DHEA if adrenal synthesis deficient.
Neurosteroid or androgen deficiency rate-limiting. DHEA supplementation restores GABAergic neurosteroid tone and/or androgen-dependent mitochondrial function.
- Certainty
- Low
- Level of action
- Partial root cause
Androgen/neurosteroid axis not rate-limiting.
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — androgenic side effects (acne, hirsutism) possible at supraphysiologic doses but not diagnostically informative in ME/CFS. Null is non-informative.
W Mechanism: J (exogenous DHEA → suppression of endogenous adrenal DHEA production via feedback, though weaker than corticosteroid HPA suppression). If supraphysiologic dosing → possible HPA axis perturbation. Protracted risk: Very Low. Permanent risk: Very Low. Endogenous production recovers after discontinuation. Rechallenge: Yes, with physiologic monitoring (DHEA-S levels).
56 Diazepam
Appears in: Cross-reference matrix, muscle spasm / sleep cascade sections. HD Symptomatic — GABA-A potentiation (very long half-life); suppresses symptoms — high dependence risk.
GABA-A-mediated muscle spasm or anxiety rate-limiting. Diazepam’s long half-life (20–100h including active metabolites) provides sustained GABA-A potentiation — useful for continuous symptom coverage.
- Certainty
- Low
- Level of action
- Partial root cause
GABA-A not rate-limiting. If diazepam works but zolpidem doesn’t → the lesion is at non-α1 GABA-A subtypes (muscle relaxation, anxiolysis), not sleep-specific α1 GABA-A.
- Certainty
- Low
- Level of action
- Partial root cause
Accumulation with repeated dosing → active metabolite (nordiazepam) half-life up to 100h — next-day sedation is cumulative, not single-dose (Pattern 4). The accumulation pattern IS diagnostic: it maps the patient’s hepatic metabolism rate. Paradoxical excitation → chloride gradient inversion (Pattern 2). Withdrawal sensitivity → same as all benzodiazepines but extended withdrawal duration due to long half-life.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: E (withdrawal-PEM convergence — extended withdrawal duration from active metabolite accumulation), H (PMC — chronic muscle relaxation suppresses PEM warning signals from muscle tension). Protracted risk: Low–Moderate. Permanent risk: Low. Rechallenge: Yes, with taper planned in advance over 6–12 months. Long half-life is both advantage (self-tapering, smoother withdrawal) and disadvantage (accumulation, extended withdrawal). For purely diagnostic use: lorazepam preferred (shorter half-life, no active metabolites). Never use diazepam for muscle spasm diagnosis if the patient has not been counseled about the 100h active metabolite half-life. ## DMF (Dimethyl Fumarate)
Appears in: Connective Tissue Hypotheses, Mitochondrial Hypotheses, cross-reference matrix Group B. Corrective — Nrf2 activator + immunomodulator; activates endogenous anti-inflammatory programs via Keap1 modification
HIF-1α-driven pathology rate-limiting. Nrf2 activation reduces oxidative stress; improvement confirms HIF-1α → ROS → tissue damage as the dominant pathway.
- Certainty
- Low
- Level of action
- Partial root cause
HIF-1α not rate-limiting — or Nrf2 activation insufficient to overcome HIF-1α-driven pathology.
- Certainty
- Low
- Level of action
- Partial root cause
Severe flushing → mast cell hyper-reactivity — DMF triggers mast cell PGD2 release. Flush prevented by aspirin → PGD2-mediated MCAS confirmed (Pattern 3). The flush is the diagnostic signal.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: D (flush → mast-cell degranulation → potential mast-cell sensitization from repeated degranulation episodes — theoretical). The Nrf2 activation mechanism does not cause structural damage. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Yes, with aspirin pre-treatment for flush. The aspirin prevention of DMF flush confirms prostaglandin-mediated MCAS — diagnostic information obtained at minimal risk.
57 Domperidone
Appears in: Cross-reference matrix, Gastrointestinal Symptoms, Does ME/CFS Gastroparesis Share the Enteric-Nerve and Pacemaker-Cell Loss Seen in Diabetic and Idiopathic Gastroparesis? (prokinetic response discriminates receptor-mediated from structural ICC/enteric-neuron loss). HD Threshold-modulatory — peripheral D2 antagonist; prokinetic — QT prolongation risk.
Peripheral D2-mediated gastroparesis rate-limiting. Domperidone blocks D2 receptors in the GI tract (does not cross BBB) → improved gastric emptying confirms the gastroparesis is D2-mediated and the enteric D2 receptor population is intact.
- Certainty
- Low
- Level of action
- Partial root cause
Gastroparesis not D2-mediated — vagal efferent lesion, interstitial cells of Cajal dysfunction, or structural. If domperidone fails but erythromycin works → motilin receptor-mediated gastric emptying is intact; D2 was not the lesion.
- Certainty
- Low
- Level of action
- Partial root cause
QT prolongation at 30–40 mg/day → hERG channel sensitivity — confirms the cardiac potassium channel is domperidone-sensitive (Pattern 4). No prolactin elevation → D2 receptors in the pituitary are inaccessible (domperidone is peripheral) or pituitary D2 is desensitized — confirms peripherally-restricted action.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (QT prolongation → risk of torsades de pointes — potentially fatal if unrecognized). ECG monitoring at baseline and during treatment is mandatory. Protracted risk: Low. Permanent risk: Very Low (QT prolongation reverses with drug clearance). Rechallenge: Yes, with ECG monitoring and K⁺/Mg²⁺ supplementation. Avoid concurrent QT-prolonging medications (fluoroquinolones, macrolides, many antipsychotics). If QT prolongation at standard dose: GI D2 receptors are more sensitive than anticipated; reduce dose or switch to metoclopramide (with extrapyramidal monitoring).
58 Donepezil
Appears in: Neuroinflammatory Hypotheses Step K2d, cross-reference matrix. HD Threshold-modulatory — acetylcholinesterase inhibitor enhancing cholinergic transmission.
Central cholinergic deficiency rate-limiting for cognition. Donepezil’s CNS AChE inhibition enhances cholinergic tone → cognitive improvement confirms the cholinergic anti-inflammatory pathway and prefrontal cholinergic networks are functional and ACh-deficient.
- Certainty
- Low
- Level of action
- Partial root cause
Central cholinergic deficiency absent; or nicotinic/muscarinic receptors desensitized. If donepezil fails but pyridostigmine works → peripheral cholinergic lesion is dominant; CNS cholinergic system is intact but not the bottleneck.
- Certainty
- Low
- Level of action
- Partial root cause
GI cramping + bradycardia → peripheral AChE inhibition at CNS doses — confirms the drug crosses into the periphery at cognitive-effective dose (Pattern 4). The GI cramping IS diagnostic: it confirms the peripheral and central cholinergic systems are both deficient — the drug is filling both. Vivid dreams at 5 mg → REM cholinergic activation — confirms CNS AChE engagement. No GI effect + no cognitive effect → AChE non-responsive — cholinergic lesion is postsynaptic (receptor desensitization) (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: C (bradycardia from peripheral AChE inhibition → if excessive → cerebral hypoperfusion in POTS patients). Reversible with drug clearance (half-life ~70h). Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Yes, at 2.5 mg with HR monitoring. If GI cramping + bradycardia: the peripheral cholinergic hypersensitivity is confirmed — start at 1.25 mg with very slow titration (2–4 weeks per dose increment). If only GI effect without cognitive benefit: the cholinergic lesion is peripheral, not central — switch to pyridostigmine.
59 DORAs (Daridorexant, Suvorexant, Lemborexant)
Appears in: Cross-reference matrix Group A. HD Corrective — dual orexin receptor antagonists reducing orexin-mediated wakefulness drive; may have neuroprotective effects but in ME/CFS addresses only the sleep symptom.
Nocturnal orexin surges rate-limiting for sleep maintenance. OX1R/OX2R antagonism improves sleep continuity — confirms orexin-mediated hyperarousal as sleep pathology.
- Certainty
- Low
- Level of action
- Partial root cause
Orexin tone not rate-limiting — or already at floor. If DORA produces no effect → orexin is already critically low, consistent with the orexin-suppression model in ME/CFS (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
Sleep paralysis at low dose → orexin critically low — DORA is contraindicated because orexin tone is already at narcolepsy-range levels (Pattern 4). Complex sleep behaviours → orexin-mediated sleep-wake boundary fragile. No effect → orexin already at floor (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: C (paradoxical — in patients with critically low orexin tone, DORA produces near-narcolepsy states with sleep paralysis). These resolve with drug clearance. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Never if sleep paralysis occurred — the patient’s orexin tone is too low; diagnostic done. The sleep paralysis localizes the sleep pathology to orexin deficiency with high specificity and no permanent consequence.
60 Dornase alfa
Appears in: Cross-reference matrix Group A. Corrective — DNase enzyme degrading NET DNA; if NETosis confirmed, clears NET-derived DAMPs — research-stage
NET-derived extracellular DNA rate-limiting for microclot pathology. Dornase alfa cleaves NET-derived DNA → degradation of the DNA scaffold in microclots → improved microvascular perfusion confirms NETs are the structural component of microclot pathology.
- Certainty
- Low
- Level of action
- Partial root cause
NETs not the dominant component of microclot pathology — fibrin, amyloid, or platelet-derived. If dornase alfa fails but lumbrokinase works → fibrin is the structural component. If dornase alfa fails but nattokinase works → the microclot is fibrinolytic-accessible but NET concentration is not rate-limiting.
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — inhaled administration limits systemic effects. No systemic bleeding risk — DNase is specific for extracellular DNA, not circulating coagulation factors. Minimal diagnostic side effects.
W Mechanism: None plausibly causing protracted or permanent worsening. Recombinant human DNase — inhaled administration; minimal systemic absorption. Protracted risk: Very Low. Permanent risk: Very Low. One of the safest microclot probes — zero bleeding risk because it targets DNA, not the coagulation cascade. Rechallenge: Always acceptable. If positive: NET-derived microclot pathology confirmed; combine with anticoagulant or fibrinolytic for comprehensive microclot treatment.
61 Doxepin (Low-Dose)
Appears in: Cross-reference matrix. HD Symptomatic — potent H1 antagonist at 3–6 mg for sleep initiation; does not restore restorative sleep.
Histaminergic arousal is dominant cause of nocturnal awakenings. Pure H1 antagonism at 3–6 mg improves sleep maintenance — confirms histaminergic sleep disruption.
- Certainty
- Low
- Level of action
- Partial root cause
Awakenings are non-histaminergic — orexin, adrenergic, or pain-driven.
- Certainty
- Low
- Level of action
- Partial root cause
Next-day sedation at 3 mg → H1 receptor supersensitivity (Pattern 1). No sedation at 6 mg → H1 receptors desensitized from chronic MCAS histamine exposure (Pattern 5). Anticholinergic symptoms at low dose → M1 supersensitivity.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: E (withdrawal after chronic use — antihistamine rebound insomnia). At 3–6 mg (pure H1) the risk is far lower than higher tricyclic doses. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Yes, with slow taper if chronic use anticipated.
62 Doxycycline (Low-Dose)
Appears in: GPCR IgG AAb are not the dominant mechanism J3, cross-reference matrix Group B. Corrective — MMP inhibition at sub-antimicrobial doses reduces ECM degradation; does not address what drives MMP-9
MMP-9-driven ECM degradation rate-limiting. MMP inhibition by doxycycline (independent of antimicrobial effect) stabilizes connective tissue.
- Certainty
- Low
- Level of action
- Partial root cause
MMP-9 not rate-limiting for CT symptoms. ECM degradation is MMP-independent or structural (genetic EDS).
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal at sub-antimicrobial dose (40 mg/day). GI upset → standard doxycycline GI intolerance.
W Mechanism: G (cumulative photosensitivity, esophageal erosion with prolonged use — standard doxycycline toxicity). At sub-antimicrobial dose, these risks are lower. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Yes, with food and sun protection.
63 DPP-4 Inhibitors (Sitagliptin)
Appears in: Cross-reference matrix Group C. HD Threshold-modulatory — preserves endogenous incretins via DPP-4 inhibition; research-stage.
Endogenous incretin preservation rate-limiting — GLP-1/GIP degradation reduction restores incretin signaling. Suggests metabolic-immune link through incretin axis.
- Certainty
- Low
- Level of action
- Partial root cause
Incretin axis not rate-limiting.
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — well-tolerated at standard doses. Null is non-informative at supplement-level dosing.
W Mechanism: None known. DPP-4 inhibitors have a favorable safety profile. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable.
64 Droxidopa
Appears in: Cross-reference matrix. HD Threshold-modulatory — norepinephrine prodrug restoring central NE; compensates for deficient synthesis — does not correct the pathway.
NE synthesis capacity intact. Presynaptic NE stores loadable via AADC conversion of droxidopa to NE. Improvement confirms NE deficiency is synthesis-limited, not storage or release.
- Certainty
- Low
- Level of action
- Partial root cause
AADC deficient or VMAT2 cannot store NE — presynaptic lesion. If droxidopa fails but midodrine works → postsynaptic α1 intact, presynaptic NE machinery is the lesion.
- Certainty
- Low
- Level of action
- Partial root cause
Supine HTN without orthostatic benefit → postsynaptic α1 desensitized — confirms hyperadrenergic POTS (Pattern 4). NE cannot further activate already-maximally stimulated α1 receptors.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (supine HTN from NE overload in non-upright positions → potential baroreflex reset), B (metabolic cost of increased sympathetic tone — modest). Protracted risk: Low. Permanent risk: Very Low. Effects reverse with drug clearance (half-life ~2.5h for droxidopa, ~1.5h for NE). Rechallenge: Yes, with supine BP monitoring. If supine HTN without orthostatic benefit: the diagnostic information is obtained — no rechallenge needed.
65 Duloxetine
Appears in: Cross-reference matrix. HD Threshold-modulatory — SNRI increasing synaptic 5-HT and NE; compensates for deficient monoamine tone.
NE or 5-HT deficiency rate-limiting for pain, cognition, or mood. SNRI dual action confirms monoamine deficit is multimodal.
- Certainty
- Low
- Level of action
- Partial root cause
NE/5-HT not rate-limiting — or lesion is postsynaptic. If atomoxetine worked but duloxetine didn’t → NE is rate-limiting; the 5-HT component adds no benefit and may cause side effects. If duloxetine worked but atomoxetine didn’t → the 5-HT component is dominant; NE alone insufficient.
- Certainty
- Low
- Level of action
- Partial root cause
Anticholinergic symptoms at low dose → subclinical AChE deficiency (Pattern 3). Nausea at 30 mg → 5-HT3 gut hypersensitivity — confirms enteric serotonin receptor supersensitivity. Insomnia ceiling → NET-mediated NE increase is activating; the sleep architecture depends on low NE tone (Pattern 4).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: E (withdrawal-PEM convergence — SNRI discontinuation syndrome with brain zaps, dizziness, dysautonomia). Venlafaxine and duloxetine produce more severe withdrawal than SSRIs due to shorter half-life + dual monoamine withdrawal. Protracted risk: Low–Moderate. Permanent risk: Low. Withdrawal-triggered PEM can advance the ratchet if taper is too rapid. Rechallenge: Yes, with very slow taper planned in advance. If severe withdrawal on discontinuation: reinstate low dose to stabilize, then taper over months (5–10% dose reduction every 2–4 weeks). For future trials, prefer fluoxetine (self-tapering due to long half-life) over SNRIs.
66 EGCG (Epigallocatechin Gallate)
Appears in: Cross-reference matrix Group C. HD Substrate-repletion + Corrective — antioxidant + Nrf2 activator + COMT inhibitor; hepatotoxicity at supratherapeutic doses.
NF-κB-driven or Nrf2-deficient pathology rate-limiting. EGCG inhibits NF-κB and activates Nrf2 → dual anti-inflammatory + antioxidant effect. Improvement confirms the NF-κB/Nrf2 axis is the dominant lesion.
- Certainty
- Low
- Level of action
- Partial root cause
NF-κB/Nrf2 axis not rate-limiting; or potency insufficient. If EGCG fails but sulforaphane works → Nrf2 is the bottleneck but EGCG’s Nrf2 activation is too weak. If EGCG fails but curcumin works → NF-κB is the bottleneck but requires stronger inhibition.
- Certainty
- Low
- Level of action
- Partial root cause
SE Hepatotoxicity at high doses (800+ mg/day green tea extract) → idiosyncratic EGCG hepatotoxicity — confirms the patient has EGCG-sensitive hepatic metabolism (Pattern A). This is a known but rare toxicity — not diagnostically useful for ME/CFS; standard hepatotoxicity monitoring is required. GI upset at standard dose → standard catechin GI intolerance.
W Mechanism: A (hepatotoxicity at supratherapeutic doses — idiosyncratic; resolves with discontinuation if caught early). Protracted risk: Very Low. Permanent risk: Very Low. Hepatotoxicity reverses with discontinuation. Rechallenge: Never if hepatotoxicity occurred — the idiosyncratic sensitivity is permanent. Use LFT monitoring at baseline and every 3–6 months if used chronically. Prefer EGCG isolate over whole green tea extract (reduced hepatotoxic component load).
67 EMDR (Eye Movement Desensitization and Reprocessing)
Appears in: Cross-reference matrix. HD Threshold-modulatory — trauma processing therapy; addresses psychological amplifier only.
PTSD/trauma-related hyperarousal rate-limiting for the sympathetic component of ME/CFS. EMDR reduces trauma-driven sympathetic activation → improved HRV and sleep confirms trauma-induced autonomic overload was rate-limiting.
- Certainty
- Low
- Level of action
- Partial root cause
Trauma not rate-limiting for autonomic dysfunction. If EMDR reduces PTSD symptoms without ME/CFS improvement → PTSD was comorbid but not causal; the autonomic lesion is intrinsic to ME/CFS.
- Certainty
- Low
- Level of action
- Partial root cause
PEM from session emotional intensity → cognitive-emotional PEM confirmed (Pattern 5). The session intensity revealing PEM IS diagnostic: the patient’s emotional processing ceiling is below the EMDR demand — confirms emotional exertion triggers PEM. Sessions should be paced within the patient’s energy envelope.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (emotional exertion → PEM if sessions exceed energy envelope). Protracted risk: Very Low. Permanent risk: Very Low. PEM from emotional exertion is reversible with rest. Rechallenge: Yes, with shorter sessions and pacing within energy envelope. EMDR is a psychotherapeutic tool, not a pharmacological agent — the only risk is PEM from session demand exceeding capacity.
68 Epoprostenol
Appears in: Cross-reference matrix Group C. HD Threshold-modulatory — prostacyclin analog for pulmonary vasodilation; limited to documented PAH context.
Prostacyclin deficiency rate-limiting for pulmonary vasodilation. IP receptor agonism restores pulmonary vascular tone — limited relevance outside documented PAH with ME/CFS.
- Certainty
- Low
- Level of action
- Partial root cause
Prostacyclin pathway not rate-limiting.
- Certainty
- Low
- Level of action
- Partial root cause
SE Jaw pain, flushing, headache → prostacyclin-mediated vasodilation — confirms IP receptor engagement. Minimal diagnostic value for ME/CFS without PAH.
W Mechanism: A (vasodilation → hypotension → potential cerebral hypoperfusion in POTS patients). Protracted risk: Low. Permanent risk: Very Low. PAH-specific; minimal ME/CFS relevance. Rechallenge: Only if PAH is confirmed and IP agonist is the indicated therapy.
69 Erythromycin (Low-Dose, Prokinetic)
Appears in: Cross-reference matrix, Gastrointestinal Symptoms, Does ME/CFS Gastroparesis Share the Enteric-Nerve and Pacemaker-Cell Loss Seen in Diabetic and Idiopathic Gastroparesis? (motilin-pathway response distinguishes functional motilin-driven emptying from structural ICC/enteric-neuron loss). HD Threshold-modulatory — motilin agonist; prokinetic — tachyphylaxis within 2–4 weeks.
Motilin receptor-mediated gastroparesis rate-limiting. Erythromycin is a motilin receptor agonist at low doses (50–100 mg) — the prokinetic effect is independent of antibiotic action. Improved gastric emptying confirms the motilin pathway is functional.
- Certainty
- Low
- Level of action
- Partial root cause
Motilin receptor absent, desensitized, or accessory. If erythromycin fails but domperidone works → the gastroparesis is D2-mediated, not motilin-mediated. If erythromycin fails but pyridostigmine works → the lesion is parasympathetic, not motilin receptor.
- Certainty
- Low
- Level of action
- Partial root cause
Tachyphylaxis within 2–4 weeks → motilin receptor desensitization — confirms the receptor is present but downregulates rapidly with sustained agonism (Pattern 4). The tachyphylaxis IS diagnostic: the motilin receptor is functional but cannot sustain prolonged activation — intermittent dosing (3–4 days/week) is the correct strategy. QT prolongation → hERG channel sensitivity — same risk as domperidone; ECG monitoring mandatory.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (QT prolongation from hERG channel block → torsades risk). Protracted risk: Very Low. Permanent risk: Very Low. QT prolongation reverses with drug clearance. Drug interactions: erythromycin is a potent CYP3A4 inhibitor — check all concurrent medications. Rechallenge: Yes, with ECG monitoring and intermittent dosing schedule (3–4 days/week to prevent tachyphylaxis). The tachyphylaxis is the key diagnostic output — it confirms the motilin pathway is the lesion and informs the correct dosing strategy. Avoid concurrent CYP3A4 substrates and QT-prolonging drugs.
70 Eszopiclone
Appears in: Cross-reference matrix, sleep cascade sections. HD Symptomatic — selective GABA-A α1 agonism for sleep onset; does not restore restorative sleep.
GABA-A α1-mediated sleep initiation AND maintenance rate-limiting. Eszopiclone’s longer half-life (~6h) vs zolpidem (~2.5h) targets sleep maintenance — confirmed by improved WASO if sleep maintenance is the dominant insomnia component.
- Certainty
- Low
- Level of action
- Partial root cause
GABA-A not rate-limiting for sleep maintenance. If eszopiclone induces sleep onset but doesn’t maintain → the sleep-maintenance pathology is non-GABAergic (orexin hyperarousal, adrenergic night-time surges, pain). If DORA works for maintenance but eszopiclone doesn’t → orexin is the dominant maintenance pathology.
- Certainty
- Low
- Level of action
- Partial root cause
SE Metallic taste (dysgeusia) → eszopiclone-specific — confirms drug engagement but not ME/CFS-diagnostic. Complex sleep behaviors at therapeutic dose → striatal GABA-A disinhibition — same as all Z-drugs (Pattern 5). Next-day sedation → CYP3A4 slow metabolizer — eszopiclone accumulation confirms pharmacogenetic variation.
W Mechanism: E (withdrawal after chronic use — same GABA-A downregulation as all Z-drugs/benzodiazepines), H (PMC — sleep is artificially induced; sleep quality may not be restorative, masking the unrefreshing-sleep PEM warning). Protracted risk: Low. Permanent risk: Very Low. Same risks as the Benzodiazepines/Z-Drugs group entry. Rechallenge: Yes, with the same caveats as all Z-drugs — plan the exit, never use chronically without a taper plan, and monitor for PMC masking of unrefreshing sleep.
71 Famciclovir
Appears in: Cross-reference matrix, antiviral cascade sections. Restorative — suppresses active herpesvirus replication; suppression only, not eradication — relapse on discontinuation
Herpesvirus reactivation rate-limiting — similar diagnostic to valacyclovir but with penciclovir as the active drug. If famciclovir works where valacyclovir failed → the patient’s viral strain is penciclovir-sensitive, acyclovir-resistant; or famciclovir’s different pharmacokinetics achieve better tissue penetration.
- Certainty
- Low
- Level of action
- Partial root cause
Herpesvirus not rate-limiting. If famciclovir fails but valganciclovir works → broader-spectrum antiviral needed; the virus is CMV or HHV-6, which famciclovir does not cover.
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — well-tolerated. Headache, nausea → standard nucleoside analogue GI intolerance. CNS effects less common than acyclovir/valacyclovir (no renal accumulation of penciclovir to the same degree).
W Mechanism: K (penciclovir accumulation if renal impairment — milder than acyclovir). Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Yes, with renal function monitoring. Famciclovir’s safety profile is better than valacyclovir and valganciclovir — preferred as a first-line antiviral probe if valacyclovir is not tolerated.
72 Famotidine / Cimetidine (H2 Antagonists)
Appears in: Cross-reference matrix. HD Threshold-modulatory — H2 receptor antagonism; famotidine: gastric acid; cimetidine: T-cell enhancement via H2 on suppressor T cells.
H2-mediated gastric histamine or T-cell modulation rate-limiting. Cimetidine’s additional immunomodulatory effect (T-cell suppression) may provide benefit where famotidine doesn’t.
- Certainty
- Low
- Level of action
- Partial root cause
H2 not rate-limiting for gastric or immune symptoms.
- Certainty
- Low
- Level of action
- Partial root cause
SE Depression or suicidal ideation → paradoxical reactor in CNS histaminergic pathways (Pattern 2). Famotidine worse depression than cimetidine → off-target effect. Cimetidine worse depression than famotidine → CYP450 drug interaction (cimetidine inhibits multiple CYPs).
W Mechanism: C (paradoxical — CNS H2 receptor blockade → histaminergic pathway disruption → depressive episode → secondary HPA stress cascade → sustained mood disturbance). Protracted risk: Low–Moderate. Permanent risk: Very Low. Mood disturbance resolves with drug clearance, but the depressive episode can trigger a PEM cascade. Rechallenge: Never if psychiatric reaction occurred — the paradoxical reactor phenotype for histaminergic systems is a permanent trait (same principle as LDN for opioid systems).
73 Fludrocortisone
Appears in: Autonomic Hypotheses Q3, cross-reference matrix. HD Threshold-modulatory — mineralocorticoid expanding plasma volume; compensates for hypovolemia without addressing volume regulation defect.
Hypovolemia present (RAAS paradox). Mineralocorticoid-mediated volume expansion increases cardiac preload and improves orthostatic tolerance — confirms hypovolemic POTS subtype.
- Certainty
- Low
- Level of action
- Partial root cause
Not hypovolemic. CT laxity or hyperadrenergic POTS. If midodrine works but fludrocortisone doesn’t → neuropathic POTS (α1 agonists restore vascular tone; volume expansion alone insufficient).
- Certainty
- Low
- Level of action
- Partial root cause
Worsening orthostasis → venous pooling — volume expansion pools in legs rather than increasing cardiac preload in CT-lax patients (Pattern 2). Hypokalemia → RAAS defect localized to renin — confirms the RAAS paradox: renin is high but aldosterone is inappropriately normal or low. The hypokalemia from fludrocortisone reveals that aldosterone receptor activation is intact but endogenous aldosterone is deficient.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: C (paradoxical — in hEDS patients with venous compliance from CT laxity, volume expansion pools peripherally; hypokalemia worsens vascular smooth muscle function). Protracted risk: Low. Permanent risk: Very Low. Effects on volume and electrolytes are fully reversible with drug clearance (~3.5h half-life). Rechallenge: Yes, at 0.025 mg (half standard starting dose) with concurrent compression garments and potassium monitoring. If orthostatic symptoms worsen: the worsening confirms CT-driven venous pooling, not hypovolemia — compression garments should work; fludrocortisone should not.
74 Fluvoxamine (Low-Dose)
Appears in: Cross-reference matrix Group A. Corrective — S1R agonism reduces ER stress and ISR activity; anti-inflammatory via S1R, not serotonin reuptake at low dose
Sigma1R agonism rate-limiting — ER stress and autophagy deficit. Low-dose fluvoxamine (25–50 mg) preferentially activates sigma1R over 5-HT reuptake. Improvement confirms ER stress pathway is dominant, distinct from SSRI mechanism.
- Certainty
- Low
- Level of action
- Partial root cause
Neither sigma1R nor 5-HT reuptake is rate-limiting.
- Certainty
- Low
- Level of action
- Partial root cause
GI distress at 25–50 mg → 5-HT3 hypersensitivity blocks the sigma1R window — the drug cannot reach sigma1R concentration before 5-HT side effects emerge (Pattern 3). Benefit lost at 100+ mg → sigma1R is dominant mechanism; 5-HT is counterproductive — confirms the sigma1R/5-HT selectivity window.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: E (withdrawal after chronic use — same SSRI withdrawal risk, but lower at low dose). Sigma1R agonism is reversible. Protracted risk: Low. Permanent risk: Very Low at low dose (25–50 mg). Rechallenge: Yes, at 25 mg. If GI distress at 25 mg prevents titration: the sigma1R window is inaccessible — use alternative sigma1R agonists (donepezil, if cholinergic system intact).
75 FMT (Fecal Microbiota Transplantation)
Appears in: Cross-reference matrix. Corrective — restores donor microbiome; if gut dysbiosis drives systemic inflammation, targets the microbial source — research-stage
Gut microbiome dysbiosis rate-limiting for systemic symptoms. FMT restores donor microbiota → multi-system improvement confirms gut-brain axis pathology originates in the microbiome.
- Certainty
- Low
- Level of action
- Partial root cause
Gut dysbiosis not rate-limiting; or the lesion is host-dependent (immune response to microbiota, not microbiota composition). If FMT fails but low-FODMAP diet works → microbial metabolites are rate-limiting, not microbial composition. If FMT fails but butyrate works → the missing metabolite is rate-limiting, not the ecosystem.
- Certainty
- Low
- Level of action
- Partial root cause
Transient symptom flare 24–72h post-FMT → microbiome engraftment with immune recognition — the flare IS diagnostic: it confirms donor microbes are engrafting and host immunity is competent (Pattern 2). No flare + no improvement → engraftment failure — donor microbiota did not colonize, or host immune response rejected the transplant (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (procedure-day energy cost + potential PEM from immune response to new microbiota), D (risk of pathogen transmission from inadequately screened donor — mitigated by rigorous screening). Protracted risk: Low. Permanent risk: Very Low (screened donor; pathogen transmission risk is the only permanent risk pathway). Rechallenge: Yes, if donor screening is rigorous and prior response was positive. The flare is diagnostic information, not a failure — it confirms engraftment and immune competence.
76 FODMAP (Low-FODMAP Diet)
Appears in: Cross-reference matrix. Corrective — reduces fermentable substrates driving SIBO/fermentation; addresses bacterial substrate source
Fermentable carbohydrate-driven GI symptoms rate-limiting for IBS/SIBO component. Low-FODMAP diet reduces bacterial fermentation → reduced bloating, distension, pain confirms bacterial overgrowth/fermentation is the dominant GI driver.
- Certainty
- Low
- Level of action
- Partial root cause
Fermentation not the dominant GI pathology. If low-FODMAP fails but cromolyn works → mast-cell-driven GI symptoms, not fermentation. If low-FODMAP fails but pyridostigmine works → GI dysmotility is dominant, not fermentation.
- Certainty
- Low
- Level of action
- Partial root cause
No response during elimination phase → SIBO absent or fermentation is not rate-limiting (Pattern 5). Response during elimination + specific food reintroduction triggers → maps the patient’s individual FODMAP sensitivities. The reintroduction phase IS the diagnostic probe — each reintroduced FODMAP subgroup identifies which carbohydrate class triggers symptoms.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: None. Dietary intervention — no pharmacologic risk. B (caloric deficit from over-restrictive elimination → metabolic PEM) is the only risk. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. Low-FODMAP is a zero-risk diagnostic probe for the fermentation component of GI symptoms. Should be done with dietitian guidance to avoid nutritional deficiency.
77 Gabapentin / Pregabalin
Appears in: Cross-reference matrix. HD Threshold-modulatory — α2δ Ca²⁺ channel subunit binding reducing neurotransmitter release; reduces central sensitization without correcting the driver.
α2δ-CaV channels functional and rate-limiting for neuropathic pain. Gabapentinoid binding reduces calcium channel trafficking → reduced neurotransmitter release → central sensitization suppression.
- Certainty
- Low
- Level of action
- Partial root cause
Central sensitization not α2δ-mediated. Consider PEA, LDN, or ketamine for non-gabapentinoid-responsive pain pathways.
- Certainty
- Low
- Level of action
- Partial root cause
Sedation at minimal dose → severe central sensitization — calcium-channel-mediated neurotransmission is load-bearing for consciousness (Pattern 4). Paradoxical agitation → GAD dysfunction — GABA synthesis failure means α2δ block removes what little GABA tone exists (Pattern 2).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: E (withdrawal-PEM convergence — gabapentinoid withdrawal: anxiety, insomnia, tachycardia, sweating; can persist weeks to months). H (PMC — pain suppression + sedation masks PEM warning signals, enabling overexertion). Protracted risk: Low–Moderate. Permanent risk: Low. Withdrawal-triggered PEM can advance the ratchet. Rechallenge: Yes, with very slow taper planned in advance (months, not weeks). Use lowest effective dose. For ME/CFS-related anxiety, consider safer alternatives first: midodrine, ivabradine, or pyridostigmine for autonomic component + LDN for neuroinflammation.
78 GET (Graded Exercise Therapy)
Appears in: Cross-reference matrix. HD Symptomatic — HARMFUL in ME/CFS; ignores PEM and energy-ratchet model; contraindicated (NICE 2021).
GET is contraindicated in ME/CFS (NICE 2021). Any perceived improvement from GET represents PMC — the patient is borrowing energy against future PEM. The “improvement” IS the diagnostic signal of PMC.
- Certainty
- Low
- Level of action
- Partial root cause
GET-induced PEM at low exercise increments → confirms the patient has ME/CFS with a narrow energy envelope. This is the expected and diagnostic outcome — GET functions as an iatrogenic CPET: it demonstrates the PEM threshold through harm.
- Certainty
- Low
- Level of action
- Partial root cause
Active PEM induction — GET IS the side effect. The rate of deterioration with incremental exercise maps the patient’s energy-ratchet sensitivity. GET functions diagnostically only in the perverse sense — it confirms ME/CFS by making the patient worse.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (structured progressive exertion → repeated PEM → each PEM episode ratchets the energy threshold down), H (PMC — the program explicitly trains the patient to ignore PEM signals → sustained overexertion → permanent decline). Protracted risk: Very High. Permanent risk: High. GET is the clinical equivalent of amphetamines in its potential to cause permanent worsening through repeated PEM induction. Rechallenge: NEVER. GET is contraindicated per NICE (2021) — the risk of permanent deterioration through repeated PEM is unacceptable. Patients who have been harmed by GET should have it documented as a severe adverse treatment reaction.
79 Ginkgo Biloba
Appears in: Cross-reference matrix Group B. HD Threshold-modulatory — multi-target neurovascular modulator improving microcirculation and cognition.
Microvascular platelet aggregation is rate-limiting. Ginkgo improves microvascular perfusion through platelet inhibition and vasodilation.
- Certainty
- Low
- Level of action
- Partial root cause
Platelet aggregation not rate-limiting. Microvascular pathology is non-platelet-mediated.
- Certainty
- Low
- Level of action
- Partial root cause
SE No bleeding at 240+ mg/day → platelet count and function intact (Pattern 5). Bleeding at standard dose → pre-existing platelet dysfunction or concurrent anticoagulant use — reveals occult hemostatic vulnerability.
W Mechanism: A (excessive bleeding risk from platelet inhibition — theoretically reversible but clinically dangerous if CNS or GI bleeding occurs). Protracted risk: Very Low. Permanent risk: Very Low (unless intracranial hemorrhage occurs from severe overdose — extremely rare). Rechallenge: Yes, with monitoring. Stop before surgery.
80 GLP-1 Agonists (Semaglutide, Tirzepatide)
Appears in: Cross-reference matrix Group B. Corrective — incretin axis modulation restoring metabolic-immune interface signaling — research-stage in ME/CFS
GLP-1R-mediated inflammation reduction is rate-limiting. GLP-1 agonism suppresses systemic inflammation (reduced TNF-α, IL-6) and may improve metabolic efficiency.
- Certainty
- Low
- Level of action
- Partial root cause
GLP-1R pathway not rate-limiting for inflammation. Or GI side effects prevent titration.
- Certainty
- Low
- Level of action
- Partial root cause
Lean mass loss > anti-inflammatory benefit → sarcopenic ME/CFS — contraindicated because loss of muscle mass reduces mitochondrial capacity and metabolic reserve (Pattern 4). Severe GI → pre-existing gastroparesis unmasked — confirms autonomic GI involvement.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (GI side effects → reduced oral intake → metabolic PEM from caloric deficit), G (cumulative muscle loss with prolonged use → sarcopenia → permanently reduced metabolic reserve). Protracted risk: Low–Moderate. Permanent risk: Low. Muscle loss during sustained caloric deficit can compound existing mitochondrial dysfunction. Rechallenge: Only if lean mass is preserved and anti-inflammatory benefit is documented. If sarcopenic ME/CFS phenotype: contraindicated — the metabolic cost of muscle loss exceeds any anti-inflammatory benefit.
81 Glutathione
Appears in: Mitochondrial Hypotheses D2, Connective Tissue Hypotheses N1, cross-reference matrix. HD Substrate-repletion — exogenous glutathione; limited oral bioavailability — IV/liposomal better.
Glutathione depletion rate-limiting for oxidative stress. Direct glutathione supplementation (IV or liposomal) bypasses the synthesis bottleneck — improvement confirms GSH deficiency is the dominant antioxidant lesion.
- Certainty
- Low
- Level of action
- Partial root cause
Glutathione not rate-limiting; or the lesion is in glutathione recycling (GPX, GR), not synthesis. If glutathione fails but NAC works → the synthesis precursor (cysteine) is rate-limiting, not direct GSH. If glutathione fails but sulforaphane works → Nrf2-mediated antioxidant enzyme induction is dominant, not GSH pool size.
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — well-tolerated. Oral glutathione has negligible bioavailability — if oral GSH produces benefit, the effect is likely placebo or prebiotic (gut microbiome metabolizes glutathione). The bioavailability confound is critical: response to oral GSH cannot be interpreted as systemic GSH repletion. IV or liposomal GSH required for systemic probe.
W Mechanism: None plausibly causing protracted or permanent worsening. Endogenous tripeptide. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. Only IV or liposomal forms should be used for diagnostic probing — oral GSH produces false negatives due to negligible systemic bioavailability. Combine with NAC and glycine for a full glutathione synthesis probe rather than direct GSH supplementation.
82 Glycine
Appears in: Cross-reference matrix Group C. HD Substrate-repletion — amino acid supporting glutathione synthesis and collagen.
Glycine-mediated inhibitory neurotransmission or glutathione precursor deficiency rate-limiting. Glycine is dual-action: inhibitory neurotransmitter (glycine receptor) and glutathione precursor (with glutamate + cysteine). Sleep improvement confirms glycinergic tone is rate-limiting for sleep maintenance.
- Certainty
- Low
- Level of action
- Partial root cause
Glycine not rate-limiting for sleep or antioxidant synthesis. If glycine fails but theanine works → the inhibitory amino acid lesion is GABAergic, not glycinergic. If glycine fails but NAC + glutamine works → glutathione synthesis is rate-limiting but glycine is not the bottleneck.
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — well-tolerated amino acid. Sedation at 3 g → glycinergic tone was critically low; supplementation restores inhibitory neurotransmission (Pattern 1). No sedation at 3 g → glycinergic system is intact or desensitized (Pattern 5).
W Mechanism: None plausibly causing protracted or permanent worsening. Glycine is the simplest amino acid — endogenously produced and widely distributed. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. The sedation threshold at 3 g maps the patient’s glycinergic inhibitory reserve — lower threshold = greater glycinergic deficiency.
83 Glycyrrhizin
Appears in: Cross-reference matrix Group B. Corrective — HMGB1 inhibitor targeting a causal DAMP mediator in PEM — research-stage
HMGB1-driven TLR4/RAGE pathology rate-limiting. Glycyrrhizin directly binds and inhibits HMGB1, the key DAMP linking cellular stress to neuroinflammation. Improvement confirms HMGB1-mediated sterile inflammation is the dominant driver.
- Certainty
- Low
- Level of action
- Partial root cause
HMGB1 not rate-limiting for neuroinflammation. If glycyrrhizin fails but LDN works → TLR4 is rate-limiting but the ligand is not HMGB1 (LPS or other DAMPs). If glycyrrhizin fails but anakinra works → IL-1 is the dominant downstream effector regardless of HMGB1 status.
- Certainty
- Low
- Level of action
- Partial root cause
Hypertension + hypokalemia (pseudohyperaldosteronism) at standard dose → 11β-HSD2 inhibition — glycyrrhizin’s metabolite (glycyrrhetinic acid) blocks cortisol inactivation in the kidney, producing a mineralocorticoid-excess state (Pattern 4). The hypertension IS diagnostic: it confirms glycyrrhizin is absorbed and bioavailable; the 11β-HSD2 enzyme is sensitive. If no HTN at sustained dose → the patient is a rapid metabolizer — glycyrrhizin may be ineffective.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (pseudohyperaldosteronism — hypertension, hypokalemia from 11β-HSD2 inhibition — reversible with discontinuation but dangerous if unrecognized). Protracted risk: Low. Permanent risk: Very Low. Hypertension and hypokalemia reverse with discontinuation, but sustained undiagnosed hypertension can cause vascular damage. Rechallenge: Yes, with BP and K⁺ monitoring. If hypertension at standard dose: 11β-HSD2 sensitivity confirmed — dose reduction or intermittent dosing. Never use glycyrrhizin long-term without BP and K⁺ monitoring. The hypertension threshold maps 11β-HSD2 sensitivity — diagnostic information obtained at manageable risk.
84 Guanfacine
Appears in: Neuroinflammatory Hypotheses K2d, cross-reference matrix. HD Threshold-modulatory — α2A-adrenergic agonist with prefrontal specificity; modulates catecholamine inverted-U in prefrontal circuits.
PFC NE deficiency rate-limiting for cognition. α2A agonism enhances PFC network connectivity by reducing cAMP and strengthening working-memory-related firing. Cognitive benefit confirms prefrontal noradrenergic lesion.
- Certainty
- Low
- Level of action
- Partial root cause
PFC lesion is not NE-mediated — glutamatergic or circuit-level. If atomoxetine improves cognition but guanfacine doesn’t → presynaptic NE release is the bottleneck, not postsynaptic α2A.
- Certainty
- Low
- Level of action
- Partial root cause
BP crash at 1 mg → sympathetically dependent BP — requires midodrine to access cognitive benefit. This combination (guanfacine + midodrine) isolates the PFC cognitive node: midodrine prevents peripheral hypotension; guanfacine improves central cognition (Pattern 4). No BP change + no cognitive benefit → PFC α2A desensitized or the cognitive lesion is non-noradrenergic (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (BP crash → cerebral hypoperfusion → PEM from ischemic stress in POTS patients with borderline cerebral perfusion), E (withdrawal rebound hypertension after chronic use). Protracted risk: Low. Permanent risk: Very Low. Effects reverse with drug clearance. Rechallenge: Yes, at 0.5 mg or lower, with BP monitoring. If BP crash but cognitive benefit is significant: add midodrine to maintain BP and isolate the central cognitive effect.
Combination note: Guanfacine is reported in fixed combination with N-acetylcysteine (NAC) for post-infectious/post-TBI cognitive deficits (Speculation, (clinical-finding:guanfacine-nac-postinfectious?)). The combination spans the catecholamine cluster (α2A PFC) and Nrf2 cluster (NAC redox support) — two non-overlapping bottlenecks. No ME/CFS data; hypothesis-generating only, not a treatment recommendation. Dose-related hypotension and bradyarrhythmia cautions apply as for guanfacine monotherapy.
85 H1 Antihistamines (Diphenhydramine, Hydroxyzine, Ketotifen)
Appears in: Inappropriate sinus tachycardia is dominant, not vascular failure. C2, cross-reference matrix. HD Threshold-modulatory — H1 receptor blockade preventing histamine-mediated symptoms without reducing mast cell degranulation.
Mast cell histamine is rate-limiting for symptoms. H1 blockade confirms histamine-driven pathology. Ketotifen additionally stabilizes mast cells (dual mechanism).
- Certainty
- Low
- Level of action
- Partial root cause
Histamine not the dominant mast cell mediator — PGD2, tryptase, or leukotrienes are dominant. If antihistamines fail but montelukast works → leukotrienes are dominant. If antihistamines fail but aspirin works → PGD2 is dominant.
- Certainty
- Low
- Level of action
- Partial root cause
SE No sedation at standard dose → CNS H1 desensitized from chronic MCAS histamine exposure (Pattern 5). Sedation at low dose → histaminergic wakefulness depends on MCAS histamine (Pattern 1). Ketotifen sedation threshold identifies CNS mast cell involvement specifically — the dose at which sedation occurs is an ordinal readout of histaminergic arousal reserve.
W Mechanism: None plausibly causing protracted or permanent worsening. Sedation resolves with drug clearance (hours to days). No known mechanism for H1 blockade → structural damage, immune sensitization, or HPA axis suppression. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. If sedation is severe: lower dose or non-sedating antihistamines (cetirizine, fexofenadine, loratadine).
86 HBOT (Hyperbaric Oxygen Therapy)
Appears in: Cross-reference matrix Group C. Corrective — increases tissue oxygen delivery; may overcome microvascular oxygen extraction deficit — research-stage
Tissue hypoxia rate-limiting for PEM and mitochondrial dysfunction. HBOT increases plasma-dissolved oxygen → bypasses RBC-dependent oxygen delivery → improvement confirms tissue hypoxia is the dominant perfusion lesion.
- Certainty
- Low
- Level of action
- Partial root cause
Tissue hypoxia not rate-limiting — mitochondrial dysfunction is intrinsic, not oxygen-delivery-dependent. If HBOT fails but mitochondrial supplements work → the lesion is in mitochondrial utilization of oxygen, not oxygen delivery.
- Certainty
- Low
- Level of action
- Partial root cause
PEM from procedure-day energy cost → the physical demand of traveling to/from and undergoing HBOT exceeds the patient’s energy envelope. This IS diagnostic: the patient’s exertion ceiling is below the treatment-access threshold. Barotrauma → standard HBOT complication; not ME/CFS-specific.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (metabolic PEM from procedure-day exertion — the treatment process itself is the risk). The oxygen delivery itself is not harmful; the exertion to access it is. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Yes, with accommodations (home-based mild HBOT if available, reduced session frequency). The procedure-access exertion is the dose-limiting factor — if the patient cannot access HBOT without triggering PEM, the treatment is inaccessible.
87 Heparin
Appears in: Cross-reference matrix Group A. Corrective — anticoagulant reducing microthrombi; if microclots are pathological, targets the coagulation cascade — bleeding risk; research-stage
Antithrombin III-mediated coagulation rate-limiting for microclot pathology. Heparin potentiates antithrombin III → inhibits thrombin and factor Xa → fibrin formation blocked. Improvement confirms thrombin-driven microclot formation is the dominant lesion.
- Certainty
- Low
- Level of action
- Partial root cause
Thrombin-driven coagulation not rate-limiting. If heparin fails but apixaban works → Factor Xa inhibition without thrombin inhibition is sufficient; or heparin’s dependence on antithrombin III is the bottleneck (AT-III deficiency). If both fail → coagulation is not the dominant microclot driver — NETs or platelet aggregation dominant.
- Certainty
- Low
- Level of action
- Partial root cause
SE Bleeding at prophylactic dose → antithrombin III sensitivity — confirms AT-III is present and heparin-responsive but the therapeutic window is narrow (Pattern 1). Heparin-induced thrombocytopenia (HIT) at 5–14 days → PF4-heparin antibody formation — the patient forms antibodies to the heparin-PF4 complex; HIT is permanently contraindicated for all heparins (Pattern D).
W Mechanism: A (bleeding risk), D (HIT → PF4-heparin immune complexes → platelet activation → thrombosis + thrombocytopenia — paradoxical thrombosis during anticoagulation, potentially fatal). Protracted risk: Low (bleeding). Permanent risk: Very Low (unless HIT develops — then permanent thrombosis risk during acute phase, and permanent contraindication to all heparins). Rechallenge: Never if HIT occurred — all heparins (UFH, LMWH, fondaparinux with caution) are permanently contraindicated. If bleeding only: dose adjustment. The HIT risk makes heparin a riskier microclot probe than DOACs — use apixaban or rivaroxaban first. Endogenous heparin from mast cells in MCAS may contribute to the coagulation phenotype.
88 Immunoadsorption / Plasmapheresis
Restorative — removes pathogenic autoantibodies and immune complexes from circulation; plasma cells continue producing — effect transient
See medication reference entry for BC007 above (BC007 / Immunoadsorption). The worsening risk is identical — mechanism B, protracted risk Low–Moderate, permanent risk Low. IA does not introduce exogenous IgG (unlike IVIG), so immune-complex formation risk is absent. Rechallenge acceptable with hemodynamic support.
89 Iron
Appears in: Cross-reference matrix. Serotonin Supplementation for ME/CFS: No Demonstrated Benefit, with a Differentiated Mechanistic Assessment, iron redox polarity diagnostic bifurcation, ferritin tsat ratio diagnostic. HD Substrate-repletion — repletes iron if deficient; supports hemoglobin and ETC — contraindicated if iron overload.
Iron deficiency rate-limiting for oxygen transport, ETC function, and neurotransmitter synthesis. Iron repletion restores haemoglobin, mitochondrial complexes, and catecholamine synthesis — confirms iron deficiency was the dominant correctable lesion. Phenotype before prescribing: if ferritin \(<\) 30 µg/L or TSAT \(<\) 16% (+ normal CRP) → genuine iron deficiency → iron supplementation indicated regardless of ME/CFS or LC diagnosis.
- Certainty
- Low
- Level of action
- Partial root cause
Iron not rate-limiting. Ferritin normal + no improvement → fatigue is not iron-mediated — one of the most common ME/CFS mimics excluded. If iron supplementation causes worsening → iron overload, ferroptosis risk, or functional iron deficiency (ferritin \(>\) 150 µg/L, TSAT \(<\) 20% — the body has iron but it is trapped in macrophages, iron redox polarity diagnostic bifurcation). In functional iron deficiency, iron supplementation may be harmful: iron is sequestered via hepcidin-independent mechanisms (hepcidin independent ferroportin blockade) and additional oral iron fuels oxidative stress and ferroptosis risk.
- Certainty
- Low
- Level of action
- Partial root cause
SE GI intolerance → standard iron GI side effects — switch to iron bisglycinate or heme iron polypeptide. No ferritin rise at 4 weeks → ferroportin blockade or hepcidin-independent iron trapping — the patient may have functional iron deficiency (Pattern 4, hepcidin independent ferroportin blockade). NOTE: hepcidin is LOW in ME/CFS ((Kavyani et al. 2024)), not high — “hepcidin blockade” is not the mechanism. The non-response IS diagnostic: the patient has functional iron deficiency (trapped iron), not simple iron deficiency — IV iron may bypass the enteral block, or iron chelation (deferiprone, deferiprone functional iron deficiency) may be more appropriate if the problem is iron trapping rather than iron shortage.
W Mechanism: A (iron overload if supplemented in the absence of deficiency — hemochromatosis gene carriers; ferroptosis risk if supplemented in functional iron deficiency with elevated ferritin + low TSAT). G (GI mucosal irritation with prolonged oral iron). Protracted risk: Very Low for genuine iron deficiency; MODERATE if supplemented in functional iron deficiency (chronic oxidative stress, lipid peroxidation, ferroptosis amplification via NTBI). Permanent risk: Very Low (unless undiagnosed hemochromatosis, or unless chronic iron supplementation in functional iron deficiency drives cumulative ferroptotic tissue damage). Rechallenge: Yes, with ferritin and TSAT monitoring. Iron deficiency must be excluded before diagnosing ME/CFS — the fatigue, cognitive dysfunction, and exercise intolerance are indistinguishable. MEASURE FULL PANEL BEFORE PRESCRIBING: serum Fe, ferritin, TSAT, sTfR, Hb, CRP. If ferritin \(>\) 150 µg/L and TSAT \(<\) 20% → functional iron deficiency → do NOT supplement oral iron; consider IV iron referral or deferiprone investigation (deferiprone functional iron deficiency, prominent safety caveats apply). If ferritin \(<\) 30 µg/L or TSAT \(<\) 16% + normal CRP → genuine iron deficiency → supplement iron and recheck at 4 weeks.
90 Ivabradine
Appears in: Autonomic Hypotheses Q1, cross-reference matrix. HD Threshold-modulatory — reduces heart rate via If channel inhibition; does not address why compensatory tachycardia is needed.
HR control without β2 blockade → rules out metabolic cost of beta-blockers. Pure If-channel blockade confirms SA node overactivation is rate-limiting, and the patient can tolerate HR reduction without β2-mediated fatigue. Patient-reported data (\(n = 3{,}925\)) align: ivabradine reported a significantly higher positive impact (NAS 66.8%) than beta-blockers — suggestive, though a single unblinded survey, that ivabradine may carry less fatigue burden as an HR-control probe in POTS-dominant ME/CFS (Eckey et al. 2025).
- Certainty
- Medium
- Level of action
- Partial root cause
SA node intrinsic dysfunction — chronotropic incompetence, not autonomic. If ivabradine produces bradycardia without symptom improvement → the tachycardia was compensatory for low stroke volume.
- Certainty
- Low
- Level of action
- Partial root cause
Bradycardia ceiling → SA node lower limit reached before therapeutic HR is achieved — chronotropic incompetence confirmed (Pattern 4). At that point, the drug cannot lower HR further without causing symptoms. Visual phosphenes → If-current in retinal cells affected — benign, resolves with dose reduction.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (if bradycardia is excessive → cerebral hypoperfusion → PEM from ischemic stress). The If-current blockade is fully reversible. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Yes, at lower dose. If bradycardia without symptom improvement: the diagnostic information (tachycardia was compensatory) is obtained — ivabradine is not the right drug.
91 IVIG (Intravenous Immunoglobulin)
Appears in: Cross-reference matrix, Autoimmune Ganglionic Acetylcholine-Receptor Blockade Could Be a Reversible Cause of Some ME/CFS Gastroparesis (immunotherapy response distinguishes reversible autoimmune ganglionic AChR blockade from structural gastroparesis), Gastrointestinal Symptoms. Corrective — neutralizes autoantibodies + immunomodulation; does not eliminate the plasma cell source — transient effect
GPCR AAb present and rate-limiting. IgG-mediated neutralization + immunomodulation → functional improvement confirms autoimmune mechanism. Transient response → B-cell repopulation restores AAb production.
- Certainty
- Low
- Level of action
- Partial root cause
No AAb, AAb not rate-limiting, or IgG/IgM mismatch.
- Certainty
- Low
- Level of action
- Partial root cause
Transient flare 48–72h → confirms AAb neutralization — immune complex formation + complement activation (Pattern 2). Flare >7 days → complement-MCAS amplification loop — C3a/C5a → mast cell degranulation → histamine, tryptase, PGD2 → further complement activation → sustained inflammatory amplification. No flare + no response → immune exhaustion (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (metabolic PEM from infusion-day energy cost + immune flare — the flare IS diagnostically useful but IS also metabolically costly) + D (immune sensitization — anti-IgA antibodies in IgA-deficient patients; complement-MCAS amplification loop that may not self-terminate). Protracted risk: Moderate. Permanent risk: Low–Moderate. If complement-MCAS loop becomes self-sustaining, it can permanently lower the patient’s baseline. Rechallenge: Only with mast-cell stabilization pre-treatment (ketotifen 1–2 mg BID for 2–4 weeks + aspirin if COX-1/COX-2 PGD2 confirmed). Consider SCIG (subcutaneous immunoglobulin) as alternative — slower absorption, lower peak IgG, reduced complement activation. If permanent worsening: IVIG is no longer accessible — consider plasmapheresis or IA as alternatives that remove AAb without introducing exogenous IgG.
92 Ketamine (Sub-Anesthetic)
Appears in: Cross-reference matrix. HD Threshold-modulatory — NMDA receptor antagonist with rapid neuroplastic effects — research-stage; not established in ME/CFS.
NMDA-mediated excitotoxicity (QUIN-driven) rate-limiting for pain and cognition. Low-dose ketamine blocks NMDA receptor overactivation from quinolinic acid in the kynurenine pathway.
- Certainty
- Low
- Level of action
- Partial root cause
NMDA system not rate-limiting.
- Certainty
- Low
- Level of action
- Partial root cause
Psychotomimetic reaction at 0.1 mg/kg → NMDA hypofunction — kynurenic acid dominance blocks basal NMDA tone; ketamine further reduces NMDA signaling below the threshold for normal perception (Pattern 4). Pain relief without psychotomimetic effects → QUIN-driven excitotoxicity confirmed — NMDA blockade is therapeutic, not disruptive.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (infusion-day metabolic cost + potential PEM from procedure stress), C (paradoxical — psychotomimetic reaction → traumatic psychological experience → PTSD-like chronic hypervigilance → permanently lowered PEM threshold through chronic sympathetic overactivation). Protracted risk: Low. Permanent risk: Very Low (psychological trauma from adverse reaction is the main permanent-risk pathway — rare but serious). Rechallenge: If psychotomimetic reaction occurred: NEVER — the NMDA hypofunction phenotype is identified, and the psychological trauma risk is unacceptable. If pain relief without psychotomimetic effects: yes, with appropriate setting and monitoring.
93 Ketotifen (Mast Cell Stabilizer)
Appears in: Inappropriate sinus tachycardia is dominant, not vascular failure. C2, Autoimmune Hypotheses I2, cross-reference matrix. HD Threshold-modulatory — mast cell stabilizer + H1 antagonist; prevents degranulation without addressing trigger.
Mast cell degranulation rate-limiting. Dual mechanism — H1 antagonism + mast cell membrane stabilization — confirms MCAS is treatable with mast-cell-directed therapy.
- Certainty
- Low
- Level of action
- Partial root cause
Mast cells not rate-limiting; or IgE-driven MCAS not ketotifen-responsive. If ketotifen fails but cromolyn works → gut MCAS with differential drug sensitivity. If ketotifen fails but montelukast works → leukotrienes are dominant mediator, not histamine.
- Certainty
- Low
- Level of action
- Partial root cause
Sedation at 0.5 mg → histaminergic wakefulness depends on MCAS histamine release — confirms MCAS is load-bearing for arousal (Pattern 1). No sedation at 2 mg → CNS H1 desensitized from chronic MCAS histamine (Pattern 5). Weight gain at low dose → H1-mediated appetite regulation is histamine-dependent — confirms systemic MCAS.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: None plausibly causing protracted or permanent worsening. Same as H1 Antihistamines — sedation resolves with drug clearance. Mast cell stabilization is reversible. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. The sedation threshold is diagnostic — lower dose for titration.
94 L-Carnitine
Appears in: Mitochondrial Hypotheses D2, cross-reference matrix. HD Substrate-repletion — supplies carnitine for fatty acid transport into mitochondria.
FAO is functional and rate-limiting for ATP. L-carnitine shuttles fatty acids into mitochondria — improvement confirms FAO is the bottleneck and the shuttle mechanism is intact.
- Certainty
- Low
- Level of action
- Partial root cause
FAO not bottleneck — PDH/ETC defect dominates. If carnitine fails but CoQ10 works → the bottleneck is ETC, not FAO.
- Certainty
- Low
- Level of action
- Partial root cause
Worsening of fatigue → lipid peroxide production — FAO is functional but ETC downstream is blocked; fatty acid oxidation produces ROS without ATP (Pattern 5). Try NAC first to boost glutathione before re-challenging carnitine — if NAC prevents the worsening, it confirms lipid peroxide was the mechanism.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: None plausibly causing protracted or permanent worsening. Paradoxical worsening (Pattern 5) identifies the FAO-to-ETC bottleneck — diagnostic, not harmful. See Mitochondrial Supplements group entry for combined analysis. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable — NAC pre-treatment may prevent lipid peroxide worsening and unmask the therapeutic benefit.
95 L-DOPA / Carbidopa
Appears in: Cross-reference matrix. HD Threshold-modulatory — dopamine precursor compensating for deficient DA synthesis; peripheral decarboxylase inhibitor limits systemic effects.
DA synthesis is the bottleneck — AADC functional, substrate-limited. If L-DOPA works → the lesion is upstream of AADC (tyrosine hydroxylase, GTP cyclohydrolase, or substrate). If pramipexole worked but L-DOPA didn’t → the DA system works postsynaptically but synthesis is not the bottleneck.
- Certainty
- Low
- Level of action
- Partial root cause
Lesion not at DA synthesis — postsynaptic D2/D3 absent or desensitized, or non-dopaminergic. If L-DOPA fails but methylphenidate works → presynaptic release is intact but synthesis is not the bottleneck (DAT-mediated recycling is dominant).
- Certainty
- Low
- Level of action
- Partial root cause
Nausea/OH at 1/4 standard tablet → D2 supersensitivity in area postrema and vasculature (Pattern 1) — same supersensitivity pattern as pramipexole and aripiprazole. No benefit at full dose → lesion is not at DA synthesis (Pattern 5). Dyskinesia at low dose → DA receptor supersensitivity with postsynaptic hypersensitivity.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (metabolic cost of increased dopaminergic tone — modest compared to stimulants because L-DOPA does not reverse DAT or block reuptake), E (withdrawal after chronic use — DA synthesis downregulation). Protracted risk: Low. Permanent risk: Very Low. Unlike amphetamines, L-DOPA does not deplete VMAT2 (it loads vesicles, doesn’t empty them). Rechallenge: Yes, at microdoses (1/8–1/4 tablet) with carbidopa to prevent peripheral conversion.
96 Lamotrigine
Appears in: Cross-reference matrix. HD Threshold-modulatory — voltage-gated Na⁺ channel inhibitor reducing neuronal hyperexcitability; does not address what creates it.
Voltage-gated sodium channel-mediated glutamate release rate-limiting for neuropathic pain and sensory hypersensitivity. Lamotrigine stabilizes presynaptic Na⁺ channels → reduced glutamate release → improvement confirms glutamatergic hyperexcitability is the dominant pain pathway.
- Certainty
- Low
- Level of action
- Partial root cause
Glutamatergic hyperexcitability not rate-limiting. If lamotrigine fails but gabapentin works → the pain is α2δ-Ca²⁺ channel-mediated, not Na⁺ channel-mediated. If lamotrigine fails but LDN works → the pain is microglial/TLR4-mediated, not synaptic.
- Certainty
- Low
- Level of action
- Partial root cause
Rash at subtherapeutic dose (25 mg) → risk of Stevens-Johnson syndrome — the slow titration is mandatory; rash at low dose is the canary (Pattern A). Worsening of brain fog → reduced glutamate release suppresses already-deficient glutamatergic cognition — confirms the cognitive lesion is glutamate-deficient, not glutamate-toxic (Pattern 4). This is the same principle as memantine — both unmask a glutamate-deficient cognitive network.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (Stevens-Johnson syndrome — rare but potentially fatal; risk highest in first 8 weeks, especially with rapid titration). Protracted risk: Low. Permanent risk: Very Low (SJS is rare; slow titration reduces risk to near zero). Rechallenge: Never if rash with fever or mucosal involvement occurred — SJS/TEN risk confirmed; lamotrigine permanently contraindicated. If brain fog worsening only: reduced glutamate release confirmed as the mechanism — the cognitive lesion is glutamate-deficient; lamotrigine is the wrong probe (diagnostic done). Slow titration (25 mg every 2 weeks) mandatory — rapid titration increases SJS risk.
97 Larazotide
Appears in: Cross-reference matrix Group B. Corrective — zonulin antagonist reducing intestinal permeability; if leaky gut drives inflammation, targets a structural barrier defect — research-stage
Zonulin-mediated intestinal permeability rate-limiting for metabolic endotoxemia and systemic inflammation. Larazotide blocks zonulin → tight junction closure → reduced LPS translocation → improvement confirms gut barrier dysfunction is the source of systemic inflammation.
- Certainty
- Low
- Level of action
- Partial root cause
Intestinal permeability not rate-limiting. If larazotide fails but low-FODMAP works → fermentation is dominant, not permeability. If larazotide fails but cromolyn works → mast cells are the gut driver, not zonulin-mediated permeability.
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — zonulin antagonist with local gut action; minimal systemic absorption. No systemic side effects. The null response is the diagnostic output — rules out zonulin-mediated permeability as the dominant mechanism.
W Mechanism: None plausibly causing protracted or permanent worsening. Locally-acting tight junction regulator without systemic absorption. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable — zero-risk diagnostic probe for zonulin-mediated gut permeability. Investigational status limits access. The null excludes one gut permeability pathway without risk.
98 LDN (Low-Dose Naltrexone)
Appears in: Ion Channel Hypotheses A2, Neuroinflammatory Hypotheses K2d, β-blocker benefit declines at a stable dose and a weekend drug holiday restores it U3, Cross-Hypothesis Convergence Patterns, cross-reference matrix. Clinical reference: Chapter Medication Response Reference: From Drug Response to Mechanism Identification, LDN section — response interpretation, side effects per dose zone, combination inference, dose-dependent differential. HD Mixed — Restorative (TRPM3 Ca²⁺ influx restoration at 3.0–4.5 mg) + Corrective (TLR4/Nrf2 hormetic anti-inflammatory priming at 0.5–1.5 mg + compensatory endorphin upregulation at 1.5–3.0 mg). Dose-dependent: higher doses are more restorative; lower doses are threshold-modulatory/corrective.
Dose-response structure: LDN targets four mechanisms with non-overlapping dose optima — TLR4 hormetic window (Nrf2-mediated microglial M1→M2 priming peaking at 0.5–1.5 mg), opioid compensatory upregulation (endorphin ceiling at 1.5–3.0 mg), TRPM3 calcium flux restoration (possibly requiring 3.0–4.5 mg), and orexin disinhibition (tracking TLR4 microglial effects in hypothalamus). Dose-finding is therefore diagnostic: a patient whose optimal dose is 0.5–1.5 mg has TLR4/Nrf2 as the dominant therapeutic mechanism; a patient needing 3.0–4.5 mg has TRPM3 as dominant. (Calabrese and Kozumbo 2021) (Kučić et al. 2021) (Dara et al. 2023) (Toljan and Vrooman 2018)
Dose-response diagnostic logic:
- Benefit peaks at 0.5–1.5 mg and declines at 3.0–4.5 mg → TLR4 hormetic window is the dominant mechanism. Microglial Nrf2 reserve is narrow — this patient will likely show narrow hormetic windows for other Nrf2-activating drugs (sulforaphane, lithium). Neuroinflammation is confirmed but the anti-inflammatory benefit requires the priming-toxicity balance, not maximal TLR4 blockade.
- Benefit monotonic up to 4.5 mg with no decline → TRPM3 or endorphin mechanism is dominant (opioid compensatory ceiling not yet reached). This patient may tolerate well above 4.5 mg but risks catastrophic inversion at the μ-blockade threshold (~50 mg).
- Benefit plateau from 1.5 mg upward → endorphin ceiling reached early; further dose escalation adds nothing. Try dropping to 1.5 mg — if benefit is preserved, the therapeutic mechanism is opioid compensatory, not TRPM3.
- No dose matters (equivalent benefit at all doses or no benefit at any dose) → either LDN effective via a saturating mechanism (receptor reserve high) or LDN ineffective for this patient.
- Worsening at 0.5 mg (paradoxical reactor) → opioid systems critically load-bearing for mood/pain homeostasis; avoid all opioid-modulating drugs.
Channelopathy rate-limiting — TRPM3 dysfunction confirmed. TLR4 antagonism on microglia → reduced neuroinflammation. Combined with clinical improvement across multiple domains (pain, cognition, PEM) suggests TRPM3/TLR4 pathway is dominant.
- Certainty
- Low
- Level of action
- Partial root cause
Patient-reported symptom profile: in a \(n = 3{,}925\) survey, LDN improved fatigue (41.5%), PEM (33.2%), and brain fog (42.3%) but did NOT improve POTS (Eckey et al. 2025). Differential implication: a POTS-dominant patient whose fatigue/brain fog persists on LDN may still be improving on LDN-targeted axes (immune/TLR4) while the POTS component — mechanistically distinct — requires autonomic/hemodynamic targeting (see Autonomic Hypotheses). LDN non-response for POTS symptoms must NOT be read as overall LDN failure.
- Certainty
- Medium
- Level of action
- Partial root cause
SE Dose-dependent mechanism identification: the dose at which benefit peaks reveals which target is rate-limiting. 0.5–1.5 mg peak → TLR4/Nrf2 hormetic window dominant. 3.0–4.5 mg peak → TRPM3 restoration dominant. Benefit lost on dose increase within LDN range → hormetic window crossed; the inversion point measures Nrf2 transcriptional reserve.
- Certainty
- Low
- Level of action
- Partial root cause
TRPM3 absent, non-responsive, or not rate-limiting. If LDN fails but aripiprazole works → DA-specific lesion, not TRPM3. If LDN fails but IVIG works → AAb-driven, not TLR4-driven. Critical: non-response when only high doses (3.0–4.5 mg) were trialled is uninterpretable — the TLR4/Nrf2 hormetic window (0.5–1.5 mg) may be the patient’s therapeutic range and was never tested. Non-response cannot be concluded unless both the low-dose window and the 3.0–4.5 mg range have been trialled.
- Certainty
- Low
- Level of action
- Partial root cause
Dose-specific side effects (diagnostic):
Systemic sickness (malaise, flu-like, nausea) at a specific dose, absent below → within-range hormetic crossing. TLR4 blockade threshold crossed between tolerated dose and sickness dose → Nrf2-mediated M2 microglial programme collapses → M1 reversion → cytokine surge. The inversion dose identifies the patient’s TLR4 hormetic ceiling. Narrow window (sickness at 4.5 mg in a patient fine at 4.0 mg) → Nrf2 reserve is narrow, consistent with the tipping-point being near-maximal TLR4 blockade. Broad window (sickness only at 4.5 mg, fine at all lower doses) → Nrf2 reserve adequate but TLR4 signal is strong. Sickness at 1.5 mg → very narrow reserve; patient may need 0.5 mg.
- Certainty
- Low
- Level of action
- Partial root cause
GI-predominant nausea without systemic sickness → gut enteric TLR4 blockade dominates. Enteric glia express TLR4; first-pass gut concentration may exceed CNS concentration. Split dosing or transdermal LDN bypasses gut first-pass — if nausea resolves on transdermal → gut-localized. If nausea persists → CNS-mediated TLR4 threshold crossed.
- Certainty
- Low
- Level of action
- Partial root cause
Sedation at 0.5–1.5 mg (especially at initiation) → orexin critically low — opioid blockade further suppresses already-low orexin (Pattern 1). Sedation resolving after 2–4 weeks → endorphin compensation restabilized orexin — functional orexin reserve confirmed. Persistent sedation → structural orexin deficit.
- Certainty
- Low
- Level of action
- Partial root cause
Dysphoria/depression at any dose → opioid-dependent mood regulation — paradoxical reactor phenotype identified (Pattern 2). Permanent trait: avoid all opioid-modulating drugs (naltrexone, naloxone, buprenorphine, tramadol, tapentadol). Occurs at any dose, including 0.5 mg — this is pharmacodynamic, not dose-rate-dependent.
- Certainty
- Low
- Level of action
- Partial root cause
Vivid dreams/nightmares → REM disinhibition from opioid blockade — confirms opioid tone was suppressing REM. Switch to morning dosing if sleep disruption unacceptable (concedes endorphin upregulation benefit, preserves TLR4/TRPM3).
- Certainty
- Low
- Level of action
- Partial root cause
SE5 No side effects at any dose up to 4.5 mg + no therapeutic benefit → none of LDN’s four receptor targets (TLR4, TRPM3, opioid, orexin) produced any perceptible signal. Strong null — these systems are not rate-limiting. But requires that ALL dose zones were tested.
W Mechanism: C (paradoxical — opioid blockade unmasks severe endogenous opioid deficit → depressive episode → secondary HPA stress cascade → sustained mood disturbance). Protracted risk: Low. Permanent risk: Very Low (the most plausible permanent pathway is psychological trauma from severe adverse reaction, not pharmacodynamic damage). Rechallenge: Never if psychiatric reaction occurred — the paradoxical reactor phenotype for opioid systems is a permanent trait. The patient should avoid all opioid-modulating medications (naltrexone, naloxone, buprenorphine, tramadol, tapentadol). If LDN was tolerated but not therapeutic: the null diagnostic information is obtained; no rechallenge needed unless dosing window was incorrect.
99 LDN + Aripiprazole (Combo)
Appears in: Cross-reference matrix Group A. HD Mixed — LDN: Restorative/TRPM3 + Corrective/TLR4-endorphin. Aripiprazole: Threshold-modulatory/microglial. Complementary targets on same cell population.
Neuroinflammation + DA deficiency are independent co-rate-limiting mechanisms. If the combination produces greater improvement than either alone → dual-pathway convergence confirmed — TLR4-microglial + D2/D3 are separate bottlenecks that reinforce each other.
- Certainty
- Low
- Level of action
- Partial root cause
If LDN produced a dose-specific ceiling signal (sickness at 4.5 mg) with no therapeutic window at any tested dose, but LDA produces sharp threshold benefit (PEM protection at 1 mg, absent at 0.5 mg) → microglial involvement confirmed by two independent probes on the same cell population, but the therapeutically accessible receptor is D2, not TLR4. LDN’s ceiling proves receptor engagement: the microglial TLR4 population is present and functional. But the hormetic collapse (sickness at 4.5 mg) occurred before any therapeutic window appeared — TLR4 blockade cannot produce benefit in this patient because the ceiling precedes the window. LDA’s sharp threshold independently confirms the same microglial population is D2-responsive. The asymmetry — TLR4 has a ceiling with no window, D2 has a window with no ceiling below 2 mg — IS the diagnostic signal. This pattern excludes TLR4 as a therapeutic target while confirming D2 as the dominant microglial activation pathway. It is stronger evidence than simple LDN non-response, which could reflect inadequate dose-range testing or receptor inaccessibility rather than true mechanism exclusion.
- Certainty
- Low
- Level of action
- Partial root cause
If additive only (combo benefit = LDN + aripiprazole individually) → shared pathway — LDN reduces microglial TNF-α → partially restores DA synthesis; aripiprazole works downstream on D2/D3. No synergy means the pathways converge on the same node, not independent.
- Certainty
- Low
- Level of action
- Partial root cause
If LDN alone causes sedation and aripiprazole alone causes akathisia but combo neutralizes both → cross-modulation: LDN suppresses microglial D2/D3 overactivation; aripiprazole stabilizes DA tone that LDN’s opioid blockade destabilized. If combo worsens both → antagonistic interaction — opioid blockade + D2/D3 partial agonism produces dysphoric dopaminergic-opioid imbalance (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: Same as individual agents (C for LDN, A+C+D for aripiprazole). The combination risk is additive, not synergistic — each drug’s worsening risk is independent. Protracted risk: Low–Moderate (aripiprazole component). Permanent risk: Low. Rechallenge: If each alone was tolerated but combo failed → trial separately; the combination antagonism is pharmacokinetic or receptor-level, not permanent.
100 LDN + Midodrine (Combo)
Appears in: Cross-reference matrix Group A. HD Mixed — LDN: Restorative/TRPM3 + Corrective/TLR4. Midodrine: Threshold-modulatory/vasoconstriction. Parallel targeting of neuroinflammation + orthostatic intolerance.
Both TLR4/TRPM3 + α1 adrenergic are rate-limiting simultaneously. Parallel system dysfunction confirmed — autoimmune/inflammatory + autonomic.
- Certainty
- Low
- Level of action
- Partial root cause
If each alone works but combo fails → pharmacodynamic antagonism (midodrine pressor effect counteracted by pyridostigmine bradycardia if also on pyridostigmine). Not a class effect — identify the specific antagonist drug.
- Certainty
- Low
- Level of action
- Partial root cause
If midodrine’s HTN ceiling is lowered by LDN → LDN is reducing sympathetic outflow (unexpected — possibly through orexin-mediated sympathetic suppression). If LDN’s therapeutic window narrows on midodrine → BP changes alter LDN’s CNS penetration (theoretical).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: Same as individual agents. The combination does not introduce new mechanisms. Protracted risk: Low. Permanent risk: Very Low. Rechallenge: Yes — trial each alone first to confirm individual benefit before combining. If combo fails due to antagonism, identify and remove the antagonist drug.
101 Levetiracetam
Appears in: Cross-reference matrix Group A. HD Threshold-modulatory — SV2A binding modulating neurotransmitter release; empirical in CNS hyperexcitability.
PEM is a kindling/neuroplastic process — SV2A binding reduces PEM frequency. Levetiracetam’s anti-kindling mechanism confirms PEM has a neuroplastic sensitization component.
- Certainty
- Low
- Level of action
- Partial root cause
Kindling not SV2A-mediated; or PEM is metabolic, not neuroplastic. If levetiracetam fails but mitochondrial supplements work → PEM is metabolic (ETC, PDH), not neuroplastic kindling.
- Certainty
- Low
- Level of action
- Partial root cause
Psychiatric symptoms at low dose → limbic DA/5-HT reserve minimal — SV2A modulation unmasks limbic dopamine/serotonin insufficiency (Pattern 4). Sedation at low dose → cortical hypometabolism — SV2A binding reduces already-low cortical synaptic activity. No effect → PEM is not neuroplastic (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: C (paradoxical — psychiatric adverse effects from SV2A modulation in a system with minimal monoamine reserve). These resolve with drug clearance. Protracted risk: Low. Permanent risk: Very Low. Rechallenge: If psychiatric reaction at low dose: do NOT rechallenge — the limbic system’s monoamine reserve is critically low. If no effect: diagnostic done — PEM is not kindling-mediated; no rechallenge needed.
102 Lion’s Mane (Hericium erinaceus)
Appears in: Cross-reference matrix Group C. HD Threshold-modulatory — NGF/BDNF induction supporting neuronal health — research-stage; limited ME/CFS evidence.
NGF-mediated neurogenesis/nerve repair rate-limiting for cognitive dysfunction and peripheral neuropathy. Lion’s Mane stimulates NGF synthesis → improved cognition and reduced neuropathic symptoms confirms the lesion is neurotrophin-deficient and NGF-responsive.
- Certainty
- Low
- Level of action
- Partial root cause
Neurotrophin deficiency absent; or NGF pathway is intact but the lesion is structural (axon loss, not trophic factor deficiency). If Lion’s Mane fails but donepezil works → the cognitive lesion is cholinergic, not neurotrophin-mediated.
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — well-tolerated. Rare allergic sensitivity in those with mushroom allergy. No specific ME/CFS diagnostic side effects. The null is non-informative given variable extract standardization.
W Mechanism: None plausibly causing protracted or permanent worsening. Medicinal mushroom supplement. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. Extracts standardized to erinacines (mycelium) and hericenones (fruiting body) may have differential NGF-stimulating potency — standardization variability is a confound for null interpretation.
103 Lithium (Low-Dose)
Appears in: Ion Channel Hypotheses A1, cross-reference matrix. Corrective — GSK-3β inhibition → Nrf2 nuclear localization + IMPase/inositol depletion + mTORC1 modulation; triple mechanism
PIP2 depletion rate-limiting. Lithium inhibits IMPase → increases PIP2 availability → improves GPCR signaling. Confirms GPCR AAb are upstream and PIP2 is the bottleneck — the Gq/11 → PLC → PIP2 → TRPM3 pathway is intact but PIP2-depleted.
- Certainty
- Low
- Level of action
- Partial root cause
PIP2 not bottleneck. Lesion is downstream of PIP2 (channel defect is PIP2-independent) or the GPCR AAb are not rate-limiting.
- Certainty
- Low
- Level of action
- Partial root cause
Polydipsia/polyuria at 2 mg → occult nephrogenic diabetes insipidus — ADH receptor or aquaporin dysfunction confirmed (Pattern 3). Tremor at 2 mg → cerebellar NE sensitivity — lithium unmasks noradrenergic tremor pathway at a dose far below the standard psychiatric tremor threshold.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (nephrogenic DI from chronic lithium → aquaporin-2 downregulation), G (cumulative nephrotoxicity at standard psychiatric doses — far above the ME/CFS probe dose of 1–5 mg). At low-dose (1–5 mg), these risks are theoretical. Protracted risk: Very Low. Permanent risk: Very Low at low dose. Rechallenge: Yes, with monitoring of urine output and renal function. The low-dose probe (1–5 mg) carries orders of magnitude less risk than standard psychiatric doses (600–1200 mg).
104 LLLT (Low-Level Laser Therapy / Photobiomodulation)
Appears in: Cross-reference matrix Group C. Corrective — cytochrome c oxidase photostimulation enhancing ATP production — research-stage
Cytochrome c oxidase photostimulation rate-limiting for mitochondrial function. LLLT delivers red/NIR light → photon absorption by Complex IV → increased electron transport → improved ATP production confirms the ETC bottleneck is photostimulable.
- Certainty
- Low
- Level of action
- Partial root cause
ETC not photostimulable — the lesion is structural (Complex IV protein loss), not functional. If LLLT fails but CoQ10 works → the bottleneck is upstream (Complex I–III shuttle), not Complex IV.
- Certainty
- Low
- Level of action
- Partial root cause
SE No objective response → null non-informative given limited evidence base and variable device parameters (wavelength, fluence, treatment site). The null is often a device/dosing confound, not a mechanism exclusion.
W Mechanism: None plausibly causing protracted or permanent worsening. Non-ionizing radiation; no thermal tissue damage at therapeutic fluences. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. The evidence base in ME/CFS is limited — the diagnostic value is low compared to mitochondrial cofactor supplementation. If positive: Complex IV photostimulable lesion confirmed. If null: confounded by device parameters and dosing.
105 Lorazepam
Appears in: Cross-reference matrix, sleep cascade sections. HD Symptomatic — GABA-A potentiation (intermediate half-life); suppresses symptoms — dependence risk.
GABA-A-mediated acute anxiety or sleep initiation rate-limiting. Lorazepam’s intermediate half-life (10–20h) and absence of active metabolites makes it pharmacokinetically cleaner than diazepam or clonazepam — preferred for diagnostic GABA-A probes.
- Certainty
- Low
- Level of action
- Partial root cause
GABA-A not rate-limiting. If lorazepam works for anxiety but zolpidem doesn’t for sleep → non-α1 GABA-A subtypes are the dominant lesion. Lorazepam engages all GABA-A subtypes (pan-GABA-A), distinguishing from α1-selective Z-drugs.
- Certainty
- Low
- Level of action
- Partial root cause
Paradoxical excitation → chloride gradient inversion (Pattern 2) — same as all benzodiazepines. Cognitive impairment at low dose (0.5 mg) → GABA-A-dependent cognition is load-bearing — the patient’s cognitive network is GABA-A-dependent; further potentiation suppresses already-low glutamatergic tone (Pattern 4). Amnesia at therapeutic dose → hippocampal GABA-A supersensitivity.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: E (withdrawal after chronic use — shorter withdrawal than clonazepam due to shorter half-life but still significant), H (PMC — anxiolysis masks anxiety-driven PEM warning). Protracted risk: Low. Permanent risk: Very Low. Rechallenge: Yes, as a diagnostic probe with planned discontinuation. Lorazepam is the preferred benzodiazepine for diagnostic GABA-A probing — no active metabolites (safe in hepatic impairment), shorter half-life than diazepam/clonazepam, easier to discontinue. Chronic use still carries benzodiazepine dependence risk — the diagnostic probe should be converted to a taper plan from day 1.
106 Losartan
Appears in: Cross-reference matrix Group B. HD Threshold-modulatory — ARB reducing angiotensin II-mediated vasoconstriction and inflammation; modulates one arm of the RAAS.
Angiotensin II AT1R-mediated pathology rate-limiting — RAAS dysregulation with TGF-β overproduction. Losartan blocks AT1R → reduced TGF-β signaling + vasodilation → improvement confirms AT1R-driven TGF-β overproduction is the dominant fibrotic/inflammatory pathway.
- Certainty
- Low
- Level of action
- Partial root cause
AT1R-TGF-β axis not rate-limiting. If losartan fails but beta-blockers work → HR control is dominant, not AT1R-TGF-β. If losartan fails but fludrocortisone works → hypovolemia is dominant; AT1R blockade is irrelevant.
- Certainty
- Low
- Level of action
- Partial root cause
BP crash at 12.5 mg → RAAS-dependent BP — the patient’s orthostatic tolerance is entirely AT1R-mediated; angiotensin II vasoconstriction is load-bearing (Pattern 1). The BP crash IS diagnostic: it confirms AT1R dependence; RAAS blockade is contraindicated. No BP change → RAAS not load-bearing for BP — AT1R blockade is safe (Pattern 5). AKI → renal perfusion AT1R-dependent — unmasks occult renal artery stenosis (Pattern F).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (hypotension → cerebral hypoperfusion in POTS patients with borderline cerebral perfusion), F (unmasking occult renal artery stenosis → AKI). Protracted risk: Very Low. Permanent risk: Very Low (AKI reversible if caught early; hypotension resolves with drug clearance). Rechallenge: Yes, at 6.25 mg with orthostatic BP and renal function monitoring. If BP crash: AT1R-dependent BP confirmed — diagnostic; the RAAS is load-bearing; avoid all RAAS blockers. If BP stable: trial for TGF-β anti-fibrotic effect with ongoing monitoring.
107 Low-Dose IL-2
Appears in: Cross-reference matrix Group A. Corrective — expands regulatory T cells; if autoantibodies fall, Treg deficiency was permissive — research-stage
Treg deficiency drives autoimmunity — Treg-suppressible pathology confirmed. Low-dose IL-2 preferentially expands Tregs (CD4+CD25+FoxP3+), suppressing autoimmune effector T cells. Improvement confirms Treg-population-driven autoimmunity.
- Certainty
- Low
- Level of action
- Partial root cause
Tregs not rate-limiting — B-cell-driven or Treg-resistant autoimmunity, or non-immune pathology. If low-dose IL-2 fails but rituximab works → B-cell-driven, not Treg-driven.
- Certainty
- Low
- Level of action
- Partial root cause
Flare at 24–48h → effector T-cell activation precedes Treg expansion — confirms the immune system is competent and T-cell-mediated autoimmunity is present (Pattern 2). No flare + no improvement → Tregs not the bottleneck — immune exhaustion or non-T-cell pathology (Pattern 5).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: D (effector T-cell flare → temporary autoimmune exacerbation → if severe, can trigger PEM cascade while awaiting Treg expansion). The flare is expected and diagnostic, but it is metabolically costly. Protracted risk: Low. Permanent risk: Very Low. The flare resolves with Treg expansion (peak at 5–7 days post-dose). Rechallenge: Yes, if flare was mild and diagnostic. If flare was severe (prolonged PEM, functional decline): the Treg expansion window is too slow relative to the effector T-cell activation — low-dose IL-2 is inaccessible for this patient.
108 Lumbrokinase
Appears in: Cross-reference matrix Group A. Corrective — fibrinolytic enzyme degrading fibrin microthrombi — research-stage; limited quality control
Fibrinolytic-accessible microclot pathology rate-limiting. Lumbrokinase directly degrades fibrin and fibrinogen — improvement confirms fibrin is the dominant structural component of microclots and is lumbrokinase-sensitive.
- Certainty
- Low
- Level of action
- Partial root cause
Microclot fibrin not lumbrokinase-sensitive; or fibrinolytic resistance. If lumbrokinase fails but nattokinase works → the fibrinolytic enzyme specificity matters — different cleavage sites, different fibrinolytic profile. If lumbrokinase fails but apixaban works → the microclot formation rate (coagulation) exceeds the fibrinolytic capacity — anticoagulation is needed, not fibrinolysis.
- Certainty
- Low
- Level of action
- Partial root cause
SE Bleeding at fibrinolytic dose → hemostatic reserve narrow — the patient’s coagulation system is already near the bleeding threshold; microclots may be compensatory (Pattern 4). No bleeding at therapeutic dose → platelet count and coagulation intact (Pattern 5).
W Mechanism: A (fibrinolysis → bleeding risk — clinically significant at supratherapeutic doses or in patients with pre-existing coagulopathy). Protracted risk: Very Low. Permanent risk: Very Low (unless intracranial or GI hemorrhage — rare). Rechallenge: Yes, with monitoring for bleeding. Stop 1–2 weeks before surgery or dental procedures. If bleeding at standard dose: the hemostatic threshold is confirmed — dose reduction; the microclot pathology may require a gentler fibrinolytic approach.
109 Luteolin
Appears in: Neuroinflammatory Hypotheses K1, cross-reference matrix. HD Threshold-modulatory — mast cell stabilizer with BBB penetration; implicates CNS mast cells if benefit beyond quercetin.
Mast cell degranulation and microglial activation rate-limiting — dual mast-cell and microglial stabilization. Luteolin inhibits mast cell degranulation and microglial activation → improvement in brain fog + MCAS symptoms confirms the mast-cell-microglial axis is dominant.
- Certainty
- Low
- Level of action
- Partial root cause
Mast cells and microglia not luteolin-responsive. If luteolin fails but ketotifen works → mast cells are rate-limiting but luteolin’s mast-cell stabilization is insufficient. If luteolin fails but LDN works → microglial TLR4 is the dominant pathway.
- Certainty
- Low
- Level of action
- Partial root cause
SE Minimal — flavonoid supplement. Liposomal luteolin preferred for CNS penetration — non-liposomal form may produce false negative for brain fog due to limited BBB penetration. The liposomal vs non-liposomal difference is diagnostic: if liposomal works but non-liposomal doesn’t → the lesion is CNS (requires BBB penetration); if both fail → mast-cell-microglial axis not rate-limiting.
W Mechanism: None plausibly causing protracted or permanent worsening. Flavonoid supplement. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. Combined with quercetin (for CD38 inhibition) and PEA (for PPAR-α microglial modulation), luteolin is part of a multi-target microglial/mast-cell probe. Use liposomal formulation for CNS penetration.
110 Lysine
Appears in: Cross-reference matrix. HD Substrate-repletion — amino acid supporting collagen synthesis — research-stage.
Arginine-competitive herpesvirus replication rate-limiting. Lysine competes with arginine for cellular uptake → reduced intracellular arginine → reduced viral replication. Improvement during lysine loading confirms herpesvirus replication is active and arginine-dependent.
- Certainty
- Low
- Level of action
- Partial root cause
Herpesvirus not active; or viral replication is arginine-independent (or the arginine pool is so large that lysine competition is insignificant). If lysine fails but acyclovir works → viral replication confirmed but direct antiviral required; competitive arginine antagonism insufficient.
- Certainty
- Low
- Level of action
- Partial root cause
SE GI upset at high dose (3+ g/day) → standard amino acid osmotic GI effect. No diagnostic-specific side effects for ME/CFS.
W Mechanism: None plausibly causing protracted or permanent worsening. Dietary amino acid. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. Lysine is a low-potency antiviral — the null response does not exclude herpesvirus reactivation (acyclovir/valacyclovir may be needed to achieve sufficient antiviral effect). The arginine-lysine competition ratio is the relevant variable — high-arginine diet may undermine lysine’s efficacy.
111 Magnesium
Appears in: Cross-reference matrix. HD Substrate-repletion — repletes magnesium if deficient; supports enzyme function and neuronal signaling.
Magnesium deficiency rate-limiting for ATP-dependent enzymes, muscle function, and sleep. Magnesium repletion restores enzyme cofactor function — improvement confirms magnesium deficiency was a correctable lesion.
- Certainty
- Low
- Level of action
- Partial root cause
Magnesium not rate-limiting. Serum Mg normal + no improvement → fatigue is not Mg-mediated. RBC Mg may be low despite normal serum Mg — functional magnesium deficiency.
- Certainty
- Low
- Level of action
- Partial root cause
Diarrhea at therapeutic dose → magnesium osmotic GI effect — the GI ceiling (usually 300–400 mg/day of certain salts) prevents therapeutic dosing (Pattern 4). Switch to Mg glycinate (less GI effect, better absorption) — if tolerated, confirms the GI ceiling was salt-specific. Sedation → Mg is an NMDA antagonist and GABA-A potentiator — confirms NMDA overactivity is dominant (Pattern 1).
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: None plausibly causing protracted or permanent worsening. Essential mineral. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. Mg glycinate, threonate, or malate preferred over oxide or citrate for GI tolerability. Mg threonate specifically crosses BBB — preferred for cognitive symptoms. The GI ceiling identifies which salt form and dose is tolerable — diagnostic for the patient’s Mg absorption capacity.
112 Maraviroc
Appears in: Cross-reference matrix Group B. Corrective — CCR5 antagonist; if CCR5-mediated inflammatory cell trafficking drives neuroinflammation, targets a specific chemokine pathway — research-stage
CCR5-mediated monocyte chemotaxis rate-limiting for neuroinflammation. Maraviroc blocks CCR5 → prevents monocyte migration into inflamed CNS tissue → improvement confirms monocyte-driven neuroinflammation is the dominant CNS pathology.
- Certainty
- Low
- Level of action
- Partial root cause
CCR5 not rate-limiting. If maraviroc fails but LDN works → microglial TLR4 activation (resident microglia, not infiltrating monocytes) is dominant. If maraviroc fails but minocycline works → microglial suppression is needed regardless of CCR5 status.
- Certainty
- Low
- Level of action
- Partial root cause
SE Hepatotoxicity → standard CCR5 antagonist class effect. Minimal other side effects — CCR5 blockade is tolerated (CCR5-Δ32 homozygotes are healthy). The hepatotoxicity is not diagnostically useful for ME/CFS — standard monitoring required.
W Mechanism: A (hepatotoxicity). Protracted risk: Low. Permanent risk: Very Low (hepatotoxicity reversible if caught early). Rechallenge: Only if CCR5-driven monocyte pathology is confirmed by prior response and no hepatotoxicity. LFT monitoring mandatory. Maraviroc + pravastatin combination (as tested in Long COVID) targets dual monocyte-endothelial axis — monocyte migration (maraviroc) + endothelial inflammation (pravastatin).
113 MBSR (Mindfulness-Based Stress Reduction)
Appears in: Cross-reference matrix. HD Threshold-modulatory — modulates stress response and pain perception; adjunctive, not disease-modifying.
Stress-driven sympathetic overactivation rate-limiting. MBSR reduces sympathetic tone through mindfulness practice → improved HRV and anxiety confirms the sympathetic component has a stress-modifiable component.
- Certainty
- Low
- Level of action
- Partial root cause
Stress not rate-limiting for sympathetic overactivation — the autonomic lesion is intrinsic and mindfulness-resistant. If MBSR improves anxiety without HRV improvement → the autonomic lesion is organic, not stress-driven.
- Certainty
- Low
- Level of action
- Partial root cause
PEM from cognitive demand of the program → the 8-week schedule and daily practice may exceed the cognitive energy envelope (Pattern 5). Meditation-induced depersonalization in trauma patients → mindfulness is counterproductive in some PTSD phenotypes.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (cognitive exertion → PEM if program demand exceeds energy envelope). Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Yes, with shortened sessions and pacing within energy envelope. MBSR is a supportive intervention, not a disease-modifying treatment. The null does not exclude ME/CFS — it confirms the autonomic lesion is not stress-modifiable.
114 Melatonin
Appears in: Cross-reference matrix. HD Substrate-repletion at low-dose (0.3–0.5 mg) — circadian resynchronization + antioxidant; symptomatic at high-dose (MT1/MT2 desensitization).
Circadian phase disorder rate-limiting for sleep onset. MT1/MT2 agonism phase-advances the sleep-wake cycle — confirms circadian rhythm is the dominant sleep pathology.
- Certainty
- Low
- Level of action
- Partial root cause
Patient-reported differential: among the top-20 treatment groups in a \(n = 3{,}925\) survey, melatonin was the ONLY treatment that significantly improved unrefreshing sleep (43.0% symptom-specific improvement) (Eckey et al. 2025) — whereas pacing, fluids, stimulants, LDN, and others did not on this symptom. This isolates unrefreshing sleep as a melatonin-responsive (circadian/resynchronization) domain distinct from other core symptoms. Differential implication: if a patient’s unrefreshing sleep does not improve on trial of melatonin, the sleep pathology is non-circadian (orexin hyperarousal, α1-GABA-A insensitivity, or sleep-structure defect) rather than a dose/substrate shortfall.
- Certainty
- Medium
- Level of action
- Partial root cause
Sleep-onset insomnia is hyperarousal-driven (not circadian). If melatonin fails but DORA works → orexin hyperarousal, not circadian phase disorder.
- Certainty
- Low
- Level of action
- Partial root cause
Paradoxical alertness → severely delayed sleep phase or CYP1A2 rapid metabolism — melatonin is being cleared before it can exert phase-advancing effect (Pattern 2). Nightmares → REM disinhibition — melatonin enhances REM propensity, unmasking REM-sleep vulnerability. No effect at 5–10 mg → MT1/MT2 desensitized or circadian amplitude too blunted.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: None plausibly causing protracted or permanent worsening. Melatonin is an endogenous hormone. Protracted risk: Very Low. Permanent risk: Very Low. Rechallenge: Always acceptable. If paradoxical alertness: try immediate-release at lower dose (0.5–1 mg) or switch to ramelteon (MT1/MT2 agonist without CYP1A2 metabolism issue). If nightmares: reduce dose or discontinue — the REM disinhibition phenotype is identified.
115 Memantine
Appears in: NMN/NR does NOT work H2b, cross-reference matrix. HD Threshold-modulatory — NMDA receptor antagonist reducing glutamatergic excitotoxicity; does not address why glutamate is elevated.
Quinolinic acid-driven NMDA excitotoxicity rate-limiting. Memantine’s uncompetitive NMDA antagonism blocks tonic overactivation while preserving phasic signaling — improvement confirms QUIN-driven excitotoxicity dominant.
- Certainty
- Low
- Level of action
- Partial root cause
Excitotoxicity not rate-limiting — glutamate tone may be low (kynurenic acid dominance). If memantine causes sedation at low dose → basal glutamate tone low (see SE).
- Certainty
- Low
- Level of action
- Partial root cause
Sedation at low dose (5 mg) → basal glutamate tone critically low — even weak NMDA blockade further reduces already-deficient glutamatergic signaling (Pattern 4). Brain fog worsening → NMDA-dependent working memory networks are glutamate-deficient, not glutamate-toxic.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: A (NMDA blockade in a glutamate-deficient system → cognitive impairment — reversible on discontinuation). Protracted risk: Very Low. Permanent risk: Very Low. Memantine has a long half-life (60–80h) but effects fully reverse with drug clearance. Rechallenge: Yes, at 2.5 mg. If sedation at 5 mg: the diagnostic information (basal glutamate tone is low, not high) is obtained — NMDA antagonism is contraindicated; consider NMDA co-agonists or glycine-site modulators instead.
116 Metformin
Appears in: Cross-reference matrix Group B. Corrective — AMPK activation restoring metabolic efficiency + reducing mTORC1 activity; multi-mechanism metabolic intervention
AMPK pathway rate-limiting for metabolic dysfunction. Metformin activates AMPK → improved mitochondrial efficiency, reduced hepatic gluconeogenesis, and anti-inflammatory effects.
- Certainty
- Low
- Level of action
- Partial root cause
AMPK → MFF phosphorylation (Ser155/Ser172) blocks DRP1 recruitment → anti-fission effect (Toyama 2016, Science). Shankar et al. 2025 demonstrated metformin attenuates T-cell hyperproliferation in ME/CFS lymphocytes in vitro (Shankar et al. 2025). AMPK → PGC-1α → mitochondrial biogenesis may restore spare respiratory capacity in memory T cells (Pearce 2013, Science). If metformin improves fatigue + CD8+ TEM ΔΨm (TMRE flow cytometry) → CD8+ mitochondrial fusion/fission imbalance may be rate-limiting for immune-mediated symptoms. If fatigue improves but CD8+ ΔΨm unchanged → AMPK benefit is through non-T-cell mechanisms (hepatic, CNS, endothelial). If CD8+ ΔΨm improves but fatigue unchanged → T-cell mitochondria are permissive but not sufficient for symptom improvement. Ex vivo test: metformin (1 mM, 24h) on sorted CD8+ TEM cells with p-AMPK(T172) + mitochondrial network morphology readout — direct probe of AMPK-MFF-DRP1 axis. Ex vivo mdivi-1 comparator (DRP1 inhibitor, 10 µM) → if mdivi-1 restores mito fusion but metformin does not → AMPK-MFF is not the dominant pathway; DRP1 is constitutively active downstream. Clinical caveat: the in-vitro T-cell signal is not yet supported by clinical data in established fatigue — the REVIVE-TOGETHER Long COVID RCT (Long COVID proxy population) halted the metformin arm for futility in established fatigue ((Reis et al. 2026)), while COVID-OUT supports only early-treatment prevention (41% reduction vs. placebo, (Bramante et al. 2023)). Certainty: LOW — T-cell-specific clinical endpoint data pending; in vitro T-cell signal from Shankar 2025; AMPK-MFF-DRP1 pathway documented in non-ME/CFS contexts.
- Certainty
- Low
- Level of action
- Partial root cause
AMPK not rate-limiting; or GI side effects prevent titration to therapeutic dose. Or CD8+ ΔΨm normal despite fatigue → T-cell mitochondria not involved.
- Certainty
- Low
- Level of action
- Partial root cause
SE GI ceiling before metabolic benefit → gut mitochondrial vulnerability — enterocytes’ AMPK activation produces GI symptoms before systemic AMPK activation occurs (Pattern 4). Lactic acidosis risk → pre-existing mitochondrial dysfunction — metformin accumulation from impaired clearance; check renal function and lactate.
W Mechanism: B (GI distress → reduced oral intake → metabolic PEM from caloric deficit), G (cumulative vitamin B12 deficiency from impaired ileal absorption → subacute combined degeneration of spinal cord if unrecognized — compound existing neurocognitive deficits). K (lactic acidosis if renal impairment → impaired metformin clearance). Protracted risk: Low. Permanent risk: Very Low (B12 deficiency is treatable if recognized; lactic acidosis is uncommon at standard doses in patients with normal renal function). Rechallenge: Yes, at low dose (250–500 mg) with gradual titration, B12 monitoring, and renal function checks. If GI ceiling prevents therapeutic dose: berberine (natural AMPK activator) may be better tolerated.
117 Methylphenidate
Appears in: Cervical collar works — structural CCI component, Neuroinflammatory Hypotheses K2d, Systemic inflammation is downstream of a more upstream cause Step C1/C2, cross-reference matrix. HD Threshold-modulatory — DAT/NET blocker increasing synaptic DA/NE; +7% REE; diagnostic probe for DAT functional status.
NE/DA deficiency rate-limiting for cognition and fatigue. DAT/NET blockade confirms presynaptic terminals are present and functional — reuptake is the dominant clearance mechanism, and blocking it restores synaptic tone.
- Certainty
- Low
- Level of action
- Partial root cause
DA/NE not rate-limiting. Or cognitive lesion is postsynaptic, not presynaptic. If atomoxetine works but methylphenidate doesn’t → the lesion is NE-specific; the DA component (DAT block) provides no additional benefit.
- Certainty
- Low
- Level of action
- Partial root cause
Tachycardia ceiling → hyperadrenergic POTS — NET-mediated NE increase unmasks already-elevated sympathetic drive (Pattern 4). Crash ceiling → DA depletion + ATP deficit — DAT-mediated DA increase is metabolically costly and cannot be sustained (Pattern 4). Insomnia ceiling → DAT occupancy too high — supraphysiologic DA tone at night (Pattern 4). Three ceilings define the therapeutic window.
- Certainty
- Low
- Level of action
- Partial root cause
W Mechanism: B (metabolic PEM — estimated +7% energy expenditure increase, sustained tachycardia) + E (withdrawal-PEM convergence — dopamine crash on cessation superimposed on accumulated metabolic debt) + H (PMC — cognitive benefit masks PEM warning signals). Unlike amphetamines, methylphenidate does NOT cause direct VMAT2 depletion (Mechanism A absent). Protracted risk: Moderate. Permanent risk: Low–Moderate. Sustained metabolic PEM can cross the ratchet threshold. Rechallenge: Only after full recovery, at lower dose (2.5–5 mg once daily), with concurrent mitochondrial support (CoQ10 + NADH). If methylphenidate crash but atomoxetine tolerated → the DA component was the metabolic stressor; pure NRI is safer. If permanent worsening (>12 months without recovery): all DNRI medications are relatively contraindicated.
118 Metoclopramide
Appears in: Cross-reference matrix, Gastrointestinal Symptoms, Does ME/CFS Gastroparesis Share the Enteric-Nerve and Pacemaker-Cell Loss Seen in Diabetic and Idiopathic Gastroparesis? (D2/5-HT4 response discriminates receptor-mediated from structural ICC/enteric-neuron gastroparesis). HD Threshold-modulatory — D2 antagonist + 5-HT4 agonist; prokinetic — extrapyramidal risk.
D2-mediated gastroparesis and 5-HT4-mediated prokinesis rate-limiting. Metoclopramide is dual-action: D2 antagonism (like domperidone) + 5-HT4 agonism (prokinetic). Improvement confirms the gastric emptying lesion is D2 and/or 5-HT4-mediated.
- Certainty
- Low
- Level of action
- Partial root cause
D2 and 5-HT4 not rate-limiting. If metoclopramide fails but erythromycin works → motilin receptor is the dominant prokinetic pathway. If metoclopramide fails but domperidone works → D2-mediated but the patient needs a peripherally-restricted agent (CNS penetration causes side effects).
- Certainty
- Low
- Level of action
- Partial root cause
Extrapyramidal symptoms (EPS: akathisia, dystonia, parkinsonism) at standard dose → nigrostriatal D2 supersensitivity — the patient’s dopamine system has compensatory D2 upregulation from chronic DA deficiency (Pattern 1). The EPS IS diagnostic: it confirms the patient has a supersensitive nigrostriatal D2 system — the same DA supersensitivity identified by aripiprazole/pramipexole. Tardive dyskinesia with prolonged use → cumulative D2 receptor-state consolidation — potentially irreversible (Pattern A). Prolactin elevation → pituitary D2 blockade — confirms drug engagement at all D2-receptor populations.
- Certainty
- Low
- Level of action
- Partial root cause