MCAS-Stack Drug Mechanisms: Receptor Arms, Clearance, and Cross-System Bridges
“MCAS stack” here is a working label for four drugs this subsection dissects arm by arm: H1/PAF antagonist (rupatadine, which also directly stabilises mast cells), H2 antagonist (famotidine), CysLT1 antagonist (montelukast), and acetylcholinesterase inhibitor (pyridostigmine, a cholinergic probe rather than a mast cell drug). They act on distinct arms of the mast cell and cholinergic systems. The label is narrower than established MCAS pharmacotherapy, which is H1 plus H2 blockade with mast cell stabilisers (ketotifen, cromolyn); the stabiliser arm is treated in its own section and is not repeated here. This subsection treats each drug not as a treatment but as a probe of a separate arm, and consolidates the mechanistic hypotheses that follow from that view. All content here is speculation, open-question, or limitation material (Phase-2 decision: PARTIAL — most evidence is preclinical, general-population, or adjacent-condition, with discounted certainty at or below 0.40). None of it is a treatment recommendation or a dosing guide; any actual use of these drugs must follow the safety, contraindication, and interaction information in the medication-reference chapters of this volume.
1 Receptor and mediator gaps
The histamine H4 receptor (H4R) is a fourth receptor arm expressed on immune cells — mast cells, basophils, eosinophils, T cells, and dendritic cells — where it drives chemotaxis and trafficking that the H1 receptor does not cover (Gibbs and Levi-Schaffer 2012) (Walter, Kottke, and Stark 2011). H4R mediates pruritus and Th2 inflammation independently of H1, so combined H1+H4 blockade outperforms H1 blockade alone in atopic dermatitis models (Cowden et al. 2010) (Ohsawa and Hirasawa 2014). A standard H1/H2/CysLT1 stack therefore leaves this arm pharmacologically untouched — as it also leaves the H3 arm untouched, the cognitive arm developed by the H3-receptor hypothesis in ch18 sec-09 (Mehta et al. 2020). H4R antagonists remain research-stage: izuforant has reached phase I pharmacokinetic/safety testing only, with no approved H4R antagonist available (Jin et al. 2024). All supporting evidence is general, allergic, or preclinical — no ME/CFS or MCAS cohort has been studied. (Certainty: 0.35)
Falsifiable prediction: In MCAS patients with incomplete symptom control on maximal H1/H2/CysLT1 blockade, an H4R antagonist produces additional symptom reduction that H1/H2/CysLT1 blockade alone cannot achieve (measured by pruritus or flare severity, not histamine release).
Consequence: If H4R blockade adds benefit over the H1/H2/CysLT1 stack, residual MCAS symptoms in treated patients would not indicate treatment failure of the stack but an unaddressed fourth arm. This is a mechanistic research claim, not a treatment recommendation — no H4R antagonist is approved for any indication.
Receptor blockade does not stop mast cell degranulation; it only blocks the downstream effect of released histamine and cysteinyl leukotrienes. Tryptase, a major mast cell granule protease, signals through PAR2 independently of histamine receptors; in rodent injury models this pathway has been shown to promote neuroinflammation and microglial activation (Ocak et al. 2020) (Qin et al. 2021), a route reinforced by the documented bidirectional mast-cell–microglia axis (Lakatos and Rosta 2025). Prostaglandin D2 (PGD2), co-released on degranulation, acts through DP1, CRTH2, and TP receptors — none of which is blocked by H1/H2 antihistamines or a CysLT1 antagonist such as montelukast (Kupczyk and Kuna 2017). Residual MCAS symptoms can therefore persist even under maximal H1/H2/CysLT1 blockade, not because blockade is incomplete but because the escape mediators are pharmacologically untouched. The tryptase–PAR2 evidence is preclinical (rodent cardiac arrest and subarachnoid haemorrhage models); there are no ME/CFS data. (Certainty: 0.35)
Falsifiable prediction: In MCAS patients maintained on maximal H1/H2/CysLT1 blockade, residual symptom severity correlates with tryptase and/or urinary PGD2 metabolites rather than with circulating histamine load.
Consequence: Persistent symptoms under full receptor blockade are consistent with two possibilities — non-histaminergic mediator escape, or simply that the drugs do not address the disease driver — and should not be assumed to prove either. This is a mechanistic framing, not a treatment recommendation — tryptase/PAR2 and PGD2/CRTH2 blockers are not clinically available for MCAS.
A histamine degradation defect is mechanistically distinct from receptor over-stimulation. Histamine intolerance arises when accumulation exceeds degradation capacity: DAO degrades extracellular and ingested histamine (mainly gut), while HNMT is cytosolic and degrades intracellular histamine (mainly CNS) (Maintz and Novak 2007). Receptor blockade lowers the symptomatic response to a given histamine load but does not remove that load, so H1/H2/CysLT1 blockade can mask an underlying clearance deficit without correcting it. Serum DAO measurement has not proven reliable as a diagnostic biomarker (Rentzos et al. 2024), and the AOC1/HNMT variant literature is reviewed in the context of dietary histamine management (Zybul et al. 2026) (Kadiyska et al. 2025). A fibromyalgia narrative review proposes DAO-deficiency histamine intolerance as a potentially treatable subtype (Protasio Netto et al. 2026), but this is narrative and the fibromyalgia population, not ME/CFS. The DAO enzyme and its measurement are treated in detail at dao; the specific HNMT clearance-failure argument sits at HNMT-Mediated Histamine Clearance Failure Subtype and is not duplicated here. (Certainty: 0.40)
Falsifiable prediction: Among histamine-symptomatic patients whose symptoms do not respond to receptor blockade, functional DAO/HNMT deficiency (genotype or measured activity) is more prevalent than among those who respond to blockade alone.
Consequence: Clearance defects and receptor over-stimulation imply different interventions (substrate reduction vs. blockade); treating a clearance defect with receptor blockers alone would mask, not resolve, the underlying load. This is a mechanistic distinction, not a treatment recommendation — dietary histamine restriction and DAO supplementation lack robust controlled evidence in ME/CFS.
DAO is a copper-dependent amine oxidase, established structurally by real-time crystallographic capture of copper amine oxidase catalysis (Murakawa et al. 2025). On that basis, functional copper deficiency could in principle lower histamine breakdown capacity. This bridge is not established in humans: the copper literature links copper status to tryptase and mast cell biology, not to DAO-mediated histamine clearance. Copper overload decreased tryptase content and starvation increased it, without changing histamine storage or IgE-mediated degranulation (Hu Frisk et al. 2017); dietary copper deficiency increased mast cell numbers without changing per-cell histamine content (Schuschke et al. 1994). No study directly connects copper deficiency to histamine intolerance through DAO, and no human “copper deficiency → histamine intolerance” link was found. The DAO enzyme itself is detailed at dao; this argument is distinct from the HNMT clearance model at HNMT-Mediated Histamine Clearance Failure Subtype. (Certainty: 0.25)
Falsifiable prediction: If the bridge exists, copper-deficient patients should show reduced DAO activity (with low ceruloplasmin) and elevated postprandial histamine, reversing on copper repletion; the bridge fails if copper status and DAO activity are dissociated. No published human dataset demonstrates the association.
Consequence: The copper→DAO→histamine path is a plausible mechanistic arrow with no direct human evidence; copper’s observed mast cell link runs through tryptase and mast cell number, not histamine clearance. This is a mechanistic hypothesis, not a basis for recommending copper supplementation for histamine symptoms.
2 Cross-arm and cholinergic interactions
Acetylcholine acts on mast cells in both directions: it activates resting mast cells through the M3 muscarinic receptor but suppresses degranulation in stimulated mast cells through the α7 nicotinic receptor, a receptor-pattern-dependent dual effect (Kutukova et al. 2025). Enhancing cholinergic tone can increase mast cell degranulation (neostigmine augmented degranulation and CGRP in a migraine model) (Kilinc, Torun, and Baranoglu Kilinc 2024), and mast cell activation required intact vagal signalling in an animal model (Woestemeier et al. 2026). The α7 arm is itself bidirectional: α7 agonism suppresses cysteinyl leukotriene release without affecting degranulation (Mishra et al. 2010), yet α7 activation on mast cells itself is pro-atherogenic in an atherosclerosis mouse model (Wang et al. 2017) — so α7-directed cholinergic effects can be pro-inflammatory, at least in animals; whether this operates at human drug doses is untested. The net effect of pyridostigmine (an acetylcholinesterase inhibitor) in a mast-cell-active patient is therefore unpredictable from mechanism alone. No direct pyridostigmine-plus-mast cell study exists; all evidence is in vitro or animal. (Certainty: 0.25)
Falsifiable prediction: In mast-cell-active ME/CFS patients, pyridostigmine produces no consistent directional effect on MCAS symptom severity, and the observed direction tracks the patient’s prevailing cholinergic receptor profile (M3 vs. α7 dominance).
Consequence: Pyridostigmine cannot be predicted to help or harm a mast-cell-active patient from mechanism alone; α7 activation on mast cells carries a pro-atherogenic risk signal in animals, of uncertain relevance at human drug doses. This is a mechanistic caution, not a treatment recommendation for or against pyridostigmine.
Histamine enhances Th1 responses through H1 receptors but suppresses IL-12 and stimulates IL-10 through H2 receptors, biasing T cells toward a Th2/anti-inflammatory profile (Jutel et al. 2001), (Elenkov et al. 1998), (Packard and Khan 2003). The receptor-differential model is consolidated in the authoritative review (Jutel, Akdis, and Akdis 2009). If H1 is blocked, free histamine can bind only the unblocked H2 receptor, so a net Th2 shift (a raised Th2/Th1 ratio) becomes mechanistically plausible. This means H1 monotherapy could be immunologically counterproductive in a mast-cell-driven state, whereas combined H1+H2 blockade would prevent the redistribution. Whether the redistribution actually occurs at clinical antihistamine concentrations in MCAS/ME/CFS — and whether any resulting Th2 bias is clinically meaningful rather than an in vitro artefact — is unresolved: the receptor model is well established, but the redistribution-to-Th2 consequence has never been measured in a treated patient cohort.
This is an open question; no certainty is assigned. The receptor-level mechanism is established in vitro and in knockout mice, but the redistribution consequence in treated humans is untested.
Falsifiable prediction: In MCAS/ME/CFS patients, H1 monotherapy raises the Th2/Th1 cytokine ratio (for example IL-4/IFN-γ or IL-10/IL-12) relative to baseline, whereas combined H1+H2 blockade leaves the ratio unchanged; if H1 monotherapy does not shift the ratio, the redistribution mechanism is not clinically operative.
Consequence: If real, the finding would overturn the default assumption that antihistamine monotherapy is immunologically inert and would argue for H1+H2 co-blockade whenever H1 blockade is used in mast-cell-driven illness — but until the shift is measured, H1 monotherapy remains the clinical default and no patient-management change is justified.
Histamine exerts autocrine negative feedback on its own release: histamine inhibits TNF-α release from mast cells through H2 and H3 receptors (Bissonnette 1996), and an autocrine H3-mediated brake on mast cell histamine release is postulated in proximity to C-fibre endings (Alstadhaug 2014). Famotidine is an H2 antagonist, so it tests only the H2 arm of such a brake: an H3-mediated brake would be untouched by it, and no H4 antagonist is clinically available. If an H2-mediated autocrine brake operates on human mast cells, an H2 antagonist would remove it and could paradoxically increase degranulation. The two classic 1982 papers frame the contradiction: H2 antihistamines augment antigen-induced histamine release from human basophils in vitro (Tung et al. 1982), whereas H2 receptors mediate inhibition of histamine release from isolated rat mast cells (Masini et al. 1982). The two papers differ in species and cell type (human basophils vs rat mast cells) and in the direction of the net effect they infer, and have never been reconciled in a single system. It is therefore unknown whether human mast cells carry a clinically meaningful H2/H4 autocrine brake, and whether famotidine could release it and worsen degranulation in MCAS patients.
This is an open question; no certainty is assigned. It rests on two conflicting 1982 in vitro studies in different cell types, with no human mast cell or in-vivo confirmation.
Falsifiable prediction: In purified human mast cells (or a MCAS provocation study), H2 blockade with famotidine increases spontaneous or IgE-mediated mediator release (tryptase/histamine) relative to vehicle, whereas H2 agonism suppresses it; if famotidine does not increase release, no clinically relevant autocrine H2 brake operates in humans.
Consequence: Famotidine is a first-line MCAS medication, so this bears directly on treatment: if the brake is real, H2 blockade could be a double-edged sword in a subset of patients, and its protective-versus-harmful direction would need re-examination — until then, famotidine’s clinical use rests on symptom evidence, not on a confirmed mechanism.
3 The drugs as diagnostic probes
Pyridostigmine inhibits acetylcholinesterase, raising synaptic acetylcholine. A symptom that improves on pyridostigmine is consistent with being at least partly cholinergic-reversible — but response alone does not prove a cholinergic mechanism, because the same drug also alters perfusion, orthostatic haemodynamics, gut motility, and neuromuscular transmission. Pyridostigmine at excess dose can also cause cholinergic crisis (reversed by atropine) and is contraindicated in mechanical gastrointestinal or urinary obstruction and cautionary in bronchial asthma; these standard precautions are detailed in the medication-reference chapters. If a ME/CFS subset carries autoantibodies that functionally block muscarinic signalling and if that blockade can be surmounted by raising acetylcholine — two premises the Sjögren’s evidence does not establish — then pyridostigmine would act as a functional bioassay for it: a response would mark M3R-mediated pathophysiology. Anti-muscarinic M3R autoantibodies have been reported in Sjögren’s disease, where they are proposed to help identify some seronegative cases (Elsaghir and Witte 2026), and M3R is a recognised tissue autoantigen with both a T-cell (Th17) and an antibody arm (Abe et al. 2020).
The cholinergic-reversibility claim is not yet testable in ME/CFS. The pyridostigmine evidence base is POTS-specific and hemodynamic only: single studies report beneficial hemodynamic effects (Schiweck et al. 2026), the randomized-trial base is limited (Kwok et al. 2025), and pyridostigmine appears among individualized options for orthostatic tachycardia (Chung and Raj 2026). No study has validated pyridostigmine as an M3-receptor bioassay in ME/CFS, and anti-M3R autoantibodies have not been measured in ME/CFS cohorts. This is a translation from Sjögren’s autoimmunity and POTS hemodynamics to an ME/CFS cholinergic mechanism. (Certainty: 0.35)
Falsifiable prediction: In ME/CFS patients with autonomic or secretomotor symptoms, serum anti-M3R positivity (ELISA or functional assay) predicts symptom improvement on pyridostigmine, responders showing higher anti-M3R titres than non-responders (OR \(\geq\) 2.0); pyridostigmine non-response in anti-M3R-positive patients argues the block is not rate-limiting.
Consequence: If validated, pyridostigmine would become a cheap, reversible probe that sorts ME/CFS patients into a cholinergic/M3R-autoantibody subtype. Until anti-M3R is measured in ME/CFS and a response bioassay is validated, a pyridostigmine response cannot be attributed to M3R blockade rather than to the drug’s nonspecific autonomic and perfusion effects.
The four pharmacologically distinct mediator arms — H1/PAF (blocked by rupatadine, a dual H1 + platelet-activating factor antagonist that also directly stabilises mast cells (Rehman et al. 2026)), H2 (famotidine), CysLT1 (leukotriene receptor antagonists), and cholinergic (acetylcholinesterase inhibition; anti-muscarinic agents act in the opposite direction, so the arm has no fixed sign) — could in principle be probed independently, and a patient’s pattern of responses across them might cluster into reproducible MCAS subtypes (a “four-arm mediator map”). A substantial fraction of POTS patients — 36.3% in a 19-study meta-analysis, though MCAS diagnostic criteria vary — report comorbid MCAS (Kulin et al. 2026), which would give a large candidate probe population in whom such clustering would be informative. But response-profile clustering across mediator arms is not an established method: no study has systematically measured differential response to all four arms within the same MCAS/POTS patients, so it is unknown whether the clusters would be stable, reproducible, or clinically actionable.
This is an open question; no certainty is assigned. The POTS–MCAS overlap is quantified and the arm pharmacology is real, but the clustering method itself is unvalidated.
Falsifiable prediction: In a within-subject N-of-1 or crossover design across the four arms, patients’ response profiles cluster into a small, stable number of subtypes (for example H1/PAF-dominant or CysLT1-dominant) that reproduce across independent cohorts and predict the direction of future response; if profiles are continuous with no reproducible clustering, a categorical “four-arm map” is not supported.
Consequence: A validated mediator-arm map would convert MCAS treatment from trial-and-error to a testable stratification, directing each patient to the arm their own response profile implicates — but until clustering is shown to reproduce, the map remains a research scaffold, not a clinical algorithm.
This is a research-method proposal, not a claim with direct supporting evidence. The core idea is inverse negative control: if each drug probes one mediator arm, then the pattern of which drugs do nothing (the “null fingerprint”) could constrain which arm is — and is not — driving a patient’s symptoms. Two methodologically relevant but indirect anchors exist. N-of-1 trials show that individual response heterogeneity can coexist with a null group mean, so single-drug group trials miss responder subsets and a per-patient fingerprint is the only way to capture them (Ortiz et al. 2026). Histamine-receptor blockade produces a clean null on exercise domain boundary and time-to-task-failure in healthy adults, demonstrating that histamine blockade can be genuinely inert for a given outcome (Abbotts et al. 2025). There is no MCAS-specific non-response literature, and no study has tested whether a four-arm null pattern maps onto mechanism — the method is unproven.
This is an open question; no certainty is assigned. The N-of-1 and null-result anchors are real but come from different fields; the fingerprint-to-mechanism inference is untested.
Falsifiable prediction: In MCAS/POTS patients, a four-arm N-of-1 design shows that the null-response pattern across H1/PAF, H2, CysLT1, and cholinergic arms is stable within a patient over repeated cycles and is mechanistically informative (for example a null on H2 but a response to CysLT1 predicts elevated CysLT-dependent, H2-independent activation); if the null pattern is unstable across cycles or uninformative about mechanism, inverse negative control fails as a diagnostic tool.
Consequence: If it works, non-response — usually discarded as “the drug did not help” — would become diagnostic information, giving clinicians a cheap, safe way to localise mechanism in a disease with no validated biomarker. Until then it must be treated strictly as a methods proposal with no direct evidence.
Histamine is a wake-promoting neuromodulator in the canonical arousal circuit (Jones 2020). First-generation H1 antihistamines cross the blood-brain barrier and measurably impair vigilance, working memory, and speed (Kay 2000) — the pharmacological signature of central H1 blockade removing a wakefulness drive. In mice, histamine (via histidine decarboxylase and H1) favours endurance and protects against exercise exhaustion, and H1 blockade or histamine deficiency impairs endurance (Niijima-Yaoita et al. 2012).
This predicts a subtype-specific, not uniform, effect. If such a hyperarousal subtype exists and its waking drive is partly maintained by histaminergic compensation, H1 blockade could remove that compensation and unmask fatigue; in patients without this compensatory drive, the same drug would do nothing or only sedate. The claim is deliberately split — it does not predict that antihistamines worsen all patients. The endurance evidence is mouse data, and the cognition evidence is from healthy adults given first-generation agents; neither establishes a hyperarousal ME/CFS subtype nor a human H1-blockade fatigue-unmasking effect (translation gap). (Certainty: 0.30)
Falsifiable prediction: Pre-treatment markers of central histaminergic tone (for example cerebrospinal-fluid histamine, or waking maintenance on a sleep-latency test) identify a hyperarousal subtype in which a standardized H1-blockade challenge produces an objectively larger drop in vigilance or exercise endurance than in non-hyperaroused ME/CFS controls, and than in healthy controls.
Consequence: If correct, antihistamine choice in MCAS overlap would have to be subtype-matched — non-sedating second-generation agents for the hyperarousal subtype — and “antihistamines worsen fatigue” could be reinterpreted not as toxicity but as a possible signal of compensatory histaminergic arousal — though ordinary first-generation H1 sedation is the simpler explanation.
4 Bridges to pain, arousal, and vascular systems
Histamine H1 signalling sensitises TRPV1 nociceptors: a TRPV1 antagonist abolished histamine-mediated TRPV1 sensitisation in dorsal-root-ganglion neurons and attenuated histaminergic itch in vivo (Nikolaeva-Koleva et al. 2021), and mast cell mediators acting on TRPV1+ sensory fibres drive allergic pain and itch, with repeated activation producing peripheral and central sensitisation (Kalangara et al. 2022). Independent mouse work shows H1 antagonists exert anti-nociception across inflammatory and neuropathic pain models, though that analgesia may be mediated by Kv7 channels rather than TRPV1 (Y. Chen et al. 2023).
This suggests an indirect route by which H1 blockade — as with rupatadine — might reduce TRPV1-mediated muscle pain or soreness in ME/CFS: not by blocking TRPV1 directly, but by removing the histaminergic sensitising drive that lowers the nociceptor threshold. This dovetails with, and does not duplicate, the arteriolar TRPV1/ischemia cascade (Inappropriate sinus tachycardia is dominant, not vascular failure.). The histamine-to-TRPV1 step is shown directly in vitro and in animals; no human trial has demonstrated that an antihistamine reduces pain via TRPV1 de-sensitisation (translation gap), and, as noted, the mouse analgesia may run through Kv7 channels rather than TRPV1. (Certainty: 0.35)
Falsifiable prediction: In a capsaicin or heat-pain quantitative sensory test, H1 blockade lowers TRPV1-mediated hyperalgesia in ME/CFS patients with concurrent mast cell symptoms more than in patients without them, and this reduction correlates with a fall in a histamine-sensitive marker of nociceptor sensitisation.
Consequence: If validated, a possible antihistamine benefit on ME/CFS pain or soreness would gain a candidate mechanism — H1-to-TRPV1 de-sensitisation — separable from its PAF arm and from the vascular TRPV1 cascade, though the Kv7 confound means it would not be proof of the TRPV1 route.
Histaminergic and orexinergic wake systems interact directly: orexin neurons act on monoaminergic targets — including histaminergic tuberomammillary neurons — to maintain arousal (Sakurai, Saito, and Yanagisawa 2021), and both systems are canonical wake-promoting neuromodulators (Jones 2020). Wake-promoting neuromodulators (histamine, orexin, norepinephrine) are further tied to sleep quality and glymphatic clearance (Pedersen et al. 2026).
Applied to MCAS: this bridge is not direct, and the compartment question is contested inside this corpus — the H3 hypothesis treats systemically released mast cell histamine as brain-active (ch18 sec-09), while other chapters state that histamine crosses the blood-brain barrier, and this speculation instead assumes peripheral mast cell histamine does not freely reach central histaminergic receptors, so that any contribution would require an indirect route (for example vagal or immune signalling to the tuberomammillary nucleus) that is not yet evidenced. If such a route exists, episodic mast cell histamine surges could transiently add to the central arousal load that contributes to the “wired but tired” mismatch — high arousal alongside the fatigue and non-restorative sleep that impaired clearance is proposed to produce (Pedersen et al. 2026). By symmetry, the sedation threshold of an H1 blocker — the dose at which central H1 blockade first emerges — could index a patient’s central histamine arousal load: a lower threshold would mark a larger compensatory histaminergic drive, a higher one a smaller reserve — the same direction the compensatory-arousal speculation predicts (Histamine as Compensatory Arousal: H1 Blockade Unmasks Fatigue in a Hyperarousal Subtype). The orexin-histamine interaction is established in animals and in reviews of human circuitry; no study has measured a histamine arousal load or an H1-blocker sedation threshold in ME/CFS (translation gap). (Certainty: 0.35)
Falsifiable prediction: A pharmacodynamic measure of the H1-blocker sedation threshold (dose at which objective vigilance first falls) correlates with a marker of central histaminergic or orexinergic tone (for example cerebrospinal-fluid orexin-A or histamine), and ME/CFS patients with a “wired but tired” profile show a lower threshold — an objectively larger vigilance drop at a given dose — than those without it, consistent with a larger compensatory drive.
Consequence: The sedation threshold would become a bedside readout of central histamine arousal load, giving “wired but tired” a specific neuromodulator correlate. This is a research measure, not a basis for dosing antihistamines.
Platelet-activating factor (PAF) activates platelets through the PAF receptor, a target with established molecular pharmacology (Hung et al. 2025). “Microclots” — amyloid fibrin(ogen) aggregates described by one research group — are reported to be composed of fibrin and platelets and to shrink with anticoagulation (Baker et al. 2023); this microclot construct is itself contested and not established in ME/CFS. Rupatadine is a dual H1-antagonist and PAF-receptor antagonist (Rehman et al. 2026), and its PAF arm has been used clinically against PAF-driven vascular permeability (Flores 2026). Rupatadine carries its own standard safety cautions (including QT-related and CYP3A4-interaction considerations) that must be checked in the medication-reference chapters before any use.
This invites the speculation that rupatadine’s PAF arm could, in principle, damp platelet activation and thereby modulate microclot formation or persistence. Honesty: each premise is individually plausible — a contested microclot construct (Baker et al. 2023), established PAF-receptor pharmacology (Hung et al. 2025), established rupatadine dual-arm pharmacology (Rehman et al. 2026) — but the conclusion “rupatadine reduces microclots” has zero direct evidence. No study has tested whether a PAF antagonist reduces microclot burden in any condition, let alone in ME/CFS; the dengue use is a single case report (Flores 2026). (Certainty: 0.30)
Falsifiable prediction: In ME/CFS patients with measurable microclot burden (amyloid-fibrin fluorescence microscopy), rupatadine reduces microclot density relative to a pure H1 antagonist of equal antihistamine potency, indicating a PAF-arm-specific effect.
Consequence: If the PAF arm did reduce microclots, rupatadine would gain a mechanism beyond antihistamine symptom control. But the claim is currently a chain of plausible premises with no connecting clinical evidence, and must not be used to justify off-label anti-thrombotic expectations.
Within the stack, montelukast is the member whose target arm is central: a computational screen predicts — without direct human measurement — that it binds a neuronal target (Cav3.1) and crosses the human blood-brain barrier (Fong et al. 2025). H1 antihistamines also cross the blood-brain barrier — the arousal speculations below rely on that — but their central target is the histamine receptor, not leukotriene signalling. and it engages CysLT1 in brain endothelial pathophysiology, restoring tight junctions and reducing inflammatory mediators (Zhou et al. 2019). This makes montelukast a probe of a central leukotriene/CysLT1 arm — a “brain MCAS” — that peripheral antihistamines cannot reach.
The probe is two-edged. Its central activity carries a harm signal: montelukast is associated with sleep disorders in children (OR 8.31) (Amezola-Herrera et al. 2026), and a quarter-century of use has produced strengthened regulatory warnings for neuropsychiatric adverse events and relegation to add-on status (Redfern and Smith 2026). The same CNS penetration that makes it a useful probe also makes it neuropsychiatrically risky. Montelukast carries other regulatory warnings — unmasking of eosinophilic granulomatosis with polyangiitis (Churg-Strauss) and rare hepatotoxicity — that any real-world use must also observe. This extends the montelukast entry in the pharmacodiagnostic compendium (ch25 sec-12), which already records the neuropsychiatric paradoxical-reactor pattern; the central-leukotriene-arm framing is what is added here. The “brain MCAS” target rests on in-silico prediction and in vitro/animal data (Fong et al. 2025) (Zhou et al. 2019); no study has shown a central leukotriene load in ME/CFS, and the sleep findings are pediatric (translation gap). (Certainty: 0.30)
Harm caution. Do not read a montelukast sleep or neuropsychiatric adverse effect as benign. These effects are real, can be serious, and in a population already burdened by unrefreshing sleep they may do net harm. Montelukast as a CNS probe should remain a research construct, not a treatment rationale.
Falsifiable prediction: In ME/CFS patients with a “brain MCAS” profile (for example elevated cerebrospinal-fluid leukotrienes or CysLT1-dependent neuroinflammation markers), montelukast improves brain-fog measures more than in patients without that profile while worsening sleep-disturbance measures above baseline — a cognitive-benefit/sleep-harm dissociation; if cognitive change occurs only alongside sleep worsening with no dissociation, the central-leukotriene reading is not supported.
Consequence: A differential central-symptom response would localise a leukotriene/CysLT1 arm inside the CNS; but the harm signal means any positive result must be weighed against the drug’s own capacity to worsen the very sleep it is proposed to probe.
5 Timing and dosing-window hypotheses
Allergic and asthmatic disease has a well-described circadian rhythm: symptoms and bronchial reactivity worsen in the overnight and early-morning hours, tracking a nocturnal peak in inflammatory mediators rather than random fluctuation (Smolensky, Lemmer, and Reinberg 2007). Urinary leukotriene E4, the stable end-product of cysteinyl leukotriene synthesis, peaks at night in both healthy subjects and nocturnal asthmatics (Kurokawa et al. 2001), and an inhaled corticosteroid that suppresses that early-morning leukotriene surge also improves nocturnal asthma control (Tanaka, Tanaka, and Abe 2003). Mast-cell reactivity itself shows a time-of-day signal, with skin-prick-test wheal size varying by clock time (Ochfeld et al. 2021). The chronopharmacology principle is that drug delivery timed to the rhythm of the target system can improve the effect-to-dose ratio (Lemmer 2005). Montelukast, a cysteinyl leukotriene receptor (CysLT1) antagonist, is already labelled for evening dosing, a rare instance where the nocturnal-mediator logic has reached routine practice. A caution against overgeneralising: brain histamine is a wake-promoting signal whose circadian peak runs opposite to the peripheral allergic mediators (Mochizuki 2022), so “histamine peaks at night” holds for the peripheral mast cell compartment, not for histamine everywhere in the body. If the same nocturnal-mediator rhythm governs the mast-cell-driven symptoms of ME/CFS, then shifting H1/H2 antihistamines and CysLT1 antagonists toward the evening could improve overnight and early-morning symptom control at an unchanged total daily dose. This is untested in ME/CFS: no trial has compared morning against evening dosing of these drugs in ME/CFS or MCAS, and the circadian evidence comes from asthma, rhinitis, and skin-testing cohorts, not from ME/CFS patients. (Certainty: 0.35)
Falsifiable prediction: In ME/CFS patients with mast-cell-type symptoms on a stable total daily dose of H1/H2 antihistamines or a CysLT1 antagonist, an evening-weighted schedule reduces overnight and early-morning symptom scores (and, where measured, nocturnal urinary leukotriene E4) relative to a morning-weighted schedule of the same total dose.
Consequence: If confirmed, time-of-day dosing would be a low-cost lever on symptom control. Until an AM-vs-PM trial is run in ME/CFS, this remains a scheduling hypothesis borrowed from asthma chronobiology, not a recommendation; moving sedating agents to the evening also carries an unmeasured next-day sedation trade-off.
Pyridostigmine, a reversible acetylcholinesterase inhibitor, raises acetylcholine at both nicotinic and muscarinic sites, which is why its benefit (ganglionic and neuromuscular) and its dose-limiting side effects (gastrointestinal cramps, diarrhoea, salivation, sweating, bradycardia, all muscarinic) arise from the same pharmacological action. Pyridostigmine also carries class warnings that follow from the same muscarinic action — it can provoke bronchospasm and worsen bronchial asthma, is contraindicated in mechanical gastrointestinal or urinary obstruction, and at excess dose produces cholinergic crisis reversed by atropine — so any use requires the standard contraindication screen (see the medication-reference chapters of this volume). This structure implies a per-patient window between the dose that produces benefit and the dose at which muscarinic side effects become intolerable. That window, the gap between the benefit dose and the side-effect dose, could be formalised as a “cholinergic reserve index”: a patient with a wide window has ample parasympathetic capacity and is likely to tolerate the drug, while a patient whose side effects appear at or below the benefit dose has little reserve and is likely to discontinue. The concept is assembled from adjacent populations. The most relevant ME/CFS datapoint is a trial reporting that pyridostigmine improved hand-grip strength in ME/CFS (Schlömer et al. 2025); it did not measure tolerability side-by-side or report a formal side-effect threshold. Dose-response and tolerability data are richer in autonomic failure and POTS: pyridostigmine is a listed option for POTS in recent review and systematic-review literature (Chung and Raj 2026) (Schiweck et al. 2026) (Raj et al. 2005), its orthostatic-hypotension efficacy has measurable clinical correlates (Okamoto et al. 2025), and its pressor effect is enhanced when combined with atomoxetine (Okamoto et al. 2019) or contrasted with yohimbine (Shibao et al. 2010). None of these studies computed a benefit-to-side-effect dose ratio, so the index is a proposed construct, not an existing measurement. The translation gap matters: the side-effect profile is similar across conditions, but the benefit target in ME/CFS (grip strength, possibly autonomic tone) differs from the pressor endpoint in orthostatic hypotension, so the index must be re-derived in ME/CFS rather than borrowed. (Certainty: 0.40)
Falsifiable prediction: In ME/CFS (and POTS) patients initiating pyridostigmine, the ratio of the lowest dose producing intolerable muscarinic side effects to the lowest benefit dose predicts continuation at 12 weeks: patients whose side-effect dose sits at or below the benefit dose (index \(\leq\) 1) discontinue, while those with a wider window (index \(\geq\) 2) continue.
Consequence: If validated, a simple titration observation could help gauge pyridostigmine tolerability before a long trial, sparing those with little cholinergic reserve a failed course. Until it is measured prospectively, the index is a descriptive hypothesis, not a validated predictor and not a dosing guide.
Chronic H1 receptor blockade can increase H1 receptor expression, a homeostatic response documented in receptor pharmacology (Miyoshi et al. 2006) and proposed as a target of antiallergic therapy (Fukui 2007). An analogous upregulation is described for H2 receptors, where the inverse-agonist property of H2 antagonists accounts for upregulation of spontaneously active H2 receptors (Smit et al. 1996). Clinically, a related phenomenon is reported as rebound pruritus and urticaria after stopping chronic antihistamines, in a scoping review and a cetirizine case report (Seng, Oka, and Tan 2026) (Seng, Cai, and Oka 2025). The picture is not uniform: tolerance to the sedative effects of H1 antihistamines develops within days (Richardson et al. 2002), while tolerance to the anti-allergic effect is debated, and the two are pharmacologically distinct. Together these lines raise the possibility that some apparent “the antihistamine stopped working” reports reflect receptor upregulation and withdrawal rebound rather than true loss of efficacy. A research protocol that deliberately interrupts dosing and measures the response on reintroduction could separate the two: if the response is restored after an interruption, the mechanism is reversible desensitisation and upregulation; if it stays blunted on reintroduction, it is true tolerance or an underlying worsening of disease. The caution is real: withdrawal rebound after chronic antihistamines is itself documented, so any drug holiday carries a possible worsening of pruritus and urticaria, and this specific risk has not been quantified in ME/CFS or MCAS patients. The receptor-upregulation evidence is molecular (cell and receptor-expression studies) and the rebound evidence is case-level and review-level; no controlled trial has tested interrupted antihistamine dosing in any population, and none in ME/CFS. Interrupting chronic antihistamines is not a recommended strategy: in mast-cell-driven disease the withdrawal rebound can be severe, and any structured interruption should occur only under clinical supervision. (Certainty: 0.35)
Falsifiable prediction: In patients reporting antihistamine tachyphylaxis, a supervised drug holiday followed by reintroduction of the same agent restores the anti-pruritic and anti-allergic response when the cause is receptor upregulation (reversible), and fails to restore it when the cause is true tolerance or disease escalation.
Consequence: If the distinction holds, some apparent antihistamine tolerance may be reversible desensitisation rather than true loss of efficacy. Interrupting antihistamines is not clinical guidance and carries a documented, though unquantified-in-ME/CFS, rebound risk that can be severe in mast-cell-driven disease. This is a research question, not a recommended strategy.
Estrogen can augment mast cell activity: estradiol activates mast cells through a non-genomic estrogen receptor-alpha and calcium influx in vitro (Zaitsu et al. 2007), and GPER-mediated estrogen signalling is linked to mast cell tryptase and histamine expression in a rat model of visceral hypersensitivity (Xu et al. 2020). In mice, mast cells accumulate in somatosensory thalamic nuclei during estrus (Kovács and Larson 2006). The human evidence is epidemiological and review-level: sex hormones are associated with immediate-type hypersensitivity (W. Chen et al. 2008), with urticaria (Kasperska-Zaja̧c, Brzoza, and Rogala 2008), and with respiratory and skin allergies more broadly (Ridolo et al. 2019). The full estrogen–progesterone–mast cell axis is treated elsewhere in this volume (HRT Cardiovascular and VTE Risk Is Route- and Regimen-Dependent); this speculation concerns only the dosing implication. If estrogen augments mast cell degranulation, then MCAS-type symptoms in premenopausal women should track the cycle, worsening around the estrogen peak. The direction is not settled inside this corpus: the estrogen-progesterone-mast cell axis treats estrogen as priming and progesterone as often suppressive (HRT Cardiovascular and VTE Risk Is Route- and Regimen-Dependent), whereas the menstrual-cycle dopaminergic-mast cell probe treats estrogen as mast-cell-stabilising and progesterone as triggering, with luteal rather than ovulatory worsening (menstrual cycle dopaminergic mast cell probe). These two readings cannot both be correct, and this speculation adopts the priming reading only for the sake of a testable prediction. If the priming reading holds, a schedule synchronised to cycle phase could outperform fixed daily dosing. The translation gap is large: the causal estrogen–mast cell data are in vitro and animal, the human data are associative, no study has shown that MCAS or ME/CFS symptoms track menstrual phase in a verified way, and no trial has compared cycle-synchronised against fixed mast cell dosing. (Certainty: 0.30)
Falsifiable prediction: In premenopausal women with MCAS-type ME/CFS symptoms, prospectively recorded symptom severity (flushing, pruritus, gastrointestinal, pain) is higher in the late-follicular-to-ovulatory phase than in the luteal phase, and a cycle-synchronised adjustment of mast cell therapy reduces that phasic worsening more than a fixed daily schedule.
Consequence: If confirmed, menstrual-cycle phase would become a variable to control in research on mast cell therapy in premenopausal patients. Until a cycle-timing trial exists, this is a testable hypothesis, not a dosing recommendation.
6 Possible iatrogenic harms
HARM-HYPOTHESIS (unsubstantiated): long-term famotidine could, in principle, raise total body histamine load by shifting the gut flora toward more histamine-producing bacteria. Each link in this chain is real but comes from a different source, and the chain itself is assembled by analogy, not by direct evidence. Gastric-acid suppression changes the gut microbiome: proton-pump inhibitors measurably alter the composition of the faecal microbiota (Imhann et al. 2016), and a broad drug-microbiota screen found that commonly used drugs, including acid suppressants, associate with microbial composition and metabolic function (Vich Vila et al. 2020). Many gut bacteria produce histamine (Barcik et al. 2017), the trait is distributed across phylogenetically diverse species (Engevik et al. 2024), and histamine-secreting taxa are widespread in the human gut (Mou et al. 2021). In mice, microbiota-produced histamine can drive visceral hypersensitivity through the histamine H4 receptor (De Palma et al. 2022). Strung together these yield: famotidine raises gastric pH, which shifts the flora, which may enrich histamine-producing taxa, which could raise luminal histamine, which could in turn raise systemic histamine exposure in an already histamine-sensitive patient. The load-bearing step is the weakest: luminal histamine reaching systemic exposure, which gut diamine oxidase (DAO) normally limits, is assumed rather than shown. The gaps are decisive. The microbiome evidence is built on proton-pump inhibitors, and famotidine is an H2-receptor antagonist, a weaker and mechanistically different acid suppressor, so the flora shift observed for PPIs cannot be assumed for famotidine. No study has measured the microbiome or histamine load in famotidine users, and no famotidine-specific harm of this kind has been demonstrated. This is a possible harm assembled by analogy, not a demonstrated one. (Certainty: 0.30)
Falsifiable prediction: Long-term famotidine users have a higher abundance of histamine-secreting bacterial taxa and/or higher urinary histamine metabolites than matched non-users, and that difference correlates with symptom burden; the hypothesis fails if no such floristic or metabolic shift is found, or if a shift occurs without clinical consequence.
Consequence: This is a possible harm, not a demonstrated one, and it does not by itself argue against using famotidine, a low-cost, widely used drug with a long safety record. Famotidine nevertheless has its own principal cautions — dose reduction in renal impairment, a QT-prolongation consideration, and the masking of gastric malignancy — which are detailed in the medication-reference chapters and are the safety issues that matter in practice, not the speculative ones raised here. It argues for measuring the microbiome and histamine load in chronic famotidine users before asserting any risk, and for caution when long-term acid suppression is combined with other histamine-loading factors.
HARM-HYPOTHESIS (unsubstantiated): blocking histamine receptors could remove an autocrine H2 negative-feedback loop and thereby increase mast cell mediator release. This is the clinical-harm reading of the mechanism posed as an open question at Do Mast Cells Carry an H2/H4 Autocrine Brake That Famotidine Could Release? — that entry asks whether the brake exists; this one asks what follows for patients if it does. The original observation is that histamine inhibits its own release through the H2 receptor, a negative-feedback control (Lichtenstein and Gillespie 1973), and that H1 and H2 antihistamines act on this histamine-mediated inhibition of release (Lichtenstein and Gillespie 1975). If H2 blockade removes that brake, mediator release should rise. The in vitro evidence is old and points in both directions. In human basophils, H2 antihistamines augmented antigen-induced histamine release, consistent with removal of negative feedback (Tung et al. 1982). In isolated rat mast cells, the finding had the opposite emphasis: evidence for H2-receptor-mediated inhibition of histamine release (Masini et al. 1982). These studies differ in cell type (basophils versus mast cells), species (human versus rat), and conclusion, so the direction of the net effect in vivo is genuinely unresolved. There is no clinical evidence for this hypothesis: it has not been shown in any human cohort, and certainly not in ME/CFS, and the very large real-world exposure to H2 antagonists without an obvious signal of increased mediator-driven disease argues against a large effect, while not excluding a subtle one. (Certainty: 0.25)
Falsifiable prediction: If H2 blockade removes autocrine negative feedback, ex vivo histamine release from a patient’s own cells (basophil or mast cell assay) increases after H2-antagonist exposure relative to before, and some H2-blocker users show a measurable rise in mediator markers; the hypothesis fails if a controlled ex vivo test shows no augmentation, or if the augmentation is confined to a single in vitro model.
Consequence: This is a possible harm, not a demonstrated one. The contradictory in vitro data cannot guide famotidine or other H2-antagonist prescribing, and they argue for a study, not for any change in practice.
7 Untested hypothesis and boundary
The proposition that chronic mast cell degranulation is a measurable energy drain — an ATP/Ca2+ cost of repeated exocytosis that shrinks the patient’s energy envelope — has no supporting literature. A search for “mast cell degranulation energy metabolism fatigue” returns zero directly relevant hits. The closest evidence is histamine-specific and does not measure mast cell metabolic cost: histamine favours endurance and protects against exhaustion in histidine-decarboxylase-knockout mice (Niijima-Yaoita et al. 2012), while histamine-receptor blockade leaves the exercise domain boundary and time-to-task-failure unchanged in healthy adults (Abbotts et al. 2025). Neither study measures the ATP or Ca2+ burden of degranulation itself, so the energy-drain claim is a plausible-but-unmeasured hypothesis, not an established mechanism, and must not be stated as if it were measured.
Certainty: N/A (no evidence exists to assign a probability; the claim is unquantifiable as stated).
Consequence: The energy-envelope drain should be presented strictly as an untested hypothesis to be investigated (for example, by measuring the ATP/Ca2+ cost of IgE-mediated degranulation in isolated human mast cells), never as an established contributor to ME/CFS fatigue, because doing so would overstate the evidence base.
8 Integrative model
The environments in this subsection assemble into a single organising idea: the four MCAS-stack drugs are not interchangeable symptomatic agents but probes of four distinct arms — H1/PAF (rupatadine, which also directly stabilises mast cells), H2/immune (famotidine), CysLT1/CNS (montelukast), and cholinergic (pyridostigmine) — and the pattern of response and non-response across the four is more informative than any single drug’s effect. Read together, the speculations and open questions describe where current blockade is incomplete and where it may act.
The gaps define the shape of the framework. Four arm-level entities are not targeted by the stack at all: the H3 cognitive receptor (developed by the H3 hypothesis in ch18 sec-09), the H4 immune-cell receptor (H4 Receptor as the Untargeted Fourth Histamine Arm), the tryptase/PGD2 mediators that escape H1/H2/CysLT1 blockade (Mast Cell Mediator Escape Beyond H1/H2/CysLT1 Blockade), and the histamine clearance machinery (DAO/HNMT) that is mechanistically distinct from receptor stimulation (Histamine Clearance Defect vs. Receptor Over-Stimulation) — and the proposed copper–DAO link (Copper–DAO–Histamine Bridge) is the weakest of these, since the direct human evidence was not found. The cross-arm interactions are the most mechanistically interesting: free histamine displaced from H1 may act on immunosuppressive H2 (Does H1 Blockade Redistribute Free Histamine onto Immunosuppressive H2 Receptors?), and mast cells may carry an H2/H4 autocrine brake whose blockade could paradoxically increase degranulation (Do Mast Cells Carry an H2/H4 Autocrine Brake That Famotidine Could Release?) — a possibility the old, contradictory in vitro literature cannot settle. Cholinergic signalling cuts both ways on mast cells (Cholinergic Modulation of Mast Cells Is Bidirectional).
This reframes the drugs as diagnostic instruments. A symptom that resolves on pyridostigmine is consistent with cholinergic reversibility (Pyridostigmine as a Functional M3-Receptor Bioassay); the four-arm response profile could define MCAS subtypes (Can Mediator-Arm Response Profiles Define a Reproducible “Four-Arm” MCAS Subtype Map?); and the pattern of nulls is a negative control that constrains mechanism (Can the Pattern of Nulls Across the Four Drugs Constrain Mechanism by Inverse Negative Control?). The framework also reaches the pain and arousal systems: H1 sensitises TRPV1 (H1-Mediated TRPV1 Nociceptor Sensitisation as an Antihistamine-Analgesia Bridge, adjacent to Inappropriate sinus tachycardia is dominant, not vascular failure.), histamine intersects the orexin arousal system (Histamine-Orexin Arousal Bridge and the H1-Blocker Sedation Threshold) and may supply compensatory arousal that H1 blockade unmasks (Histamine as Compensatory Arousal: H1 Blockade Unmasks Fatigue in a Hyperarousal Subtype), and rupatadine’s PAF arm could touch platelet/microclot biology (PAF-Mediated Platelet Activation and Microclot Modulation). Montelukast is the member most likely to engage a central target (a computational model predicts blood-brain barrier penetration) and is therefore the closest available probe of a central leukotriene arm — but its neuropsychiatric effects are a two-edged harm (Montelukast as a CNS-Penetrant Leukotriene Probe). Timing is a free variable: circadian mediator peaks (Circadian Timing of Mast-Cell-Targeted Drugs), the pyridostigmine therapeutic window (Cholinergic Reserve Index as a Tolerability Predictor for Pyridostigmine), H1 tachyphylaxis (Pulsed Dosing and a Drug-Holiday Restore-Test for Antihistamine Tachyphylaxis), and cycle-linked mast cell sensitivity (Cycle-Timed Mast Cell Dosing). Finally, two unsubstantiated harms warn against assuming receptor blockade is costless (Famotidine, Gastric pH, and a Microbiome-Histamine Load Hypothesis, Compensatory Mediator Release After Histamine-Receptor Blockade), and one hypothesis — that mast cell degranulation is a measurable energy drain — has no supporting literature at all (The Mast Cell “Energy-Envelope Drain” Is an Untested Hypothesis, Not a Finding).
The honest status is that the framework’s arms are well-described biology (H4, tryptase/PGD2, histamine H1/H2 immunology, TRPV1 sensitisation, cholinergic control of mast cells) while every ME/CFS application is speculative. Most supporting evidence is preclinical, general-population, or from adjacent conditions (POTS, Sjögren’s, asthma), discounted at or below 0.40; a few points (pyridostigmine in ME/CFS, POTS treatment reviews) reach 0.40 but are not ME/CFS mechanism studies. No ME/CFS cohort has measured H4, DAO activity, anti-M3R, PAF/microclots, or a four-arm response profile. The framework therefore does not support any treatment recommendation or ME/CFS dosing; its value is as a testable map — a set of response-pattern predictions that a single well-instrumented trial could begin to discriminate.
Consequence: If the four-arm dissection holds, the clinical question shifts from “which antihistamine” to “which arm is dominant in this patient,” and research probes (a pyridostigmine response, an H1-vs-H2 differential, a mediator panel) might stratify MCAS-positive ME/CFS. Because none of the arm-specific predictions has been tested in ME/CFS patients, this is a research agenda for stratified pharmacological probing, not a basis for altering current MCAS practice.