Glutamate, Adenosine and Hyperarousal in RLS

This topic covers the non-dopaminergic "second pillar" of RLS pathophysiology: thalamic/cortical glutamatergic hyperactivity, adenosine A1-receptor hypofunction, endogenous opioid deficiency, and cortical (GABAergic) hyperexcitability, and how these together produce the hyperarousal, sleep-fragmentation, and sensory features of RLS that are only partly explained by the classic dopamine/iron model.

What is established (grade A/B)

• Cortical GABAergic disinhibition is a reproducible, systematic-review-level finding in RLS: a 2024 systematic review of 21 TMS studies (319 RLS patients vs 278 controls) found reduced short-interval intracortical inhibition (SICI) to be the single most consistent electrophysiological abnormality across the literature (2024-lanza-systematic-review-motor-cortex-tms.md, grade A).
• Boosting extracellular adenosine with the ENT1/ENT2 inhibitor dipyridamole improves RLS sensory and motor symptoms and objectively improves sleep in a placebo-controlled crossover RCT (IRLS change -13 vs -5, p<0.001; arousal index 23.5 vs 30.7/h, p=0.020; PLM-arousal index 2.5 vs 11.2, p<0.001; 2021-garcia-borreguero-dipyridamole-rct.md, grade B), building on an earlier open-label trial (2018-garcia-borreguero-open-label-dipyridamole.md, grade C).
• PLMS and cortical arousals are pharmacologically dissociable: a single dose of pramipexole suppressed PLMS without altering EEG arousal instability, while clonazepam did the reverse, in a placebo-controlled RCT (2012-manconi-plm-arousal-dissociation-rct.md, grade B) — evidence that the motor (dopaminergic/PLMS) and arousal (glutamatergic/GABAergic) systems in RLS are at least partly separable.
• A second, mechanistically distinct arousal-system drug class — the orexin antagonist suvorexant — also improves RLS motor/sensory symptoms and sleep architecture in an RCT (2024-garcia-borreguero-suvorexant-rct.md, grade B), and prior long-term dopamine-agonist exposure blunts the response to both dipyridamole and suvorexant even before clinical augmentation appears (2025-garcia-borreguero-subclinical-augmentation-posthoc.md, grade B) — an important anchor for sequencing decisions.
• Objective and subjective excessive daytime sleepiness (EDS) are, in fact, elevated in untreated RLS (self-reported EDS OR 13.7; objective MSLT-defined EDS OR 7.0 vs controls) and track specifically with periodic-limb-movement-associated arousals, not simply with insomnia (2025-chenini-daytime-sleepiness-arousal-rls.md, grade B).
• The AASM's 2025 clinical practice guideline gives dipyridamole a conditional recommendation (low certainty) as a treatment option, and separately recommends against pramipexole, ropinirole, rotigotine, and cabergoline as standard therapy (2025-winkelman-aasm-guideline-rls.md, grade D guideline).

Emerging C

• Thalamic glutamate/glutamine (Glx) on MRS is elevated in RLS and correlates with objective nocturnal wake time, but not with PLMS index (2013-allen-thalamic-glutamate-mrs.md) — the landmark but small (n=48) case-control study anchoring the "glutamatergic hyperarousal" hypothesis.
• Short-interval intracortical inhibition (SICI) is reduced across the full RLS severity spectrum, even in mild disease, and discriminates RLS from controls with AUC=0.79 (2019-magalhaes-sici-severity.md).
• Foundational TMS work (1999) first showed generalized (hand and foot) intracortical disinhibition in RLS (1999-tergau-motor-system-excitability.md), and later work showed impaired inhibitory (LTD-like) cortical plasticity on rTMS (2018-lanza-impaired-ltd-plasticity-rtms.md) and that dopaminergic treatment (cabergoline) reverses the SICI deficit (2006-nardone-cabergoline-cortical-hyperexcitability.md) — i.e., dopamine agonists partly work by correcting cortical disinhibition, a mechanistic clue for what a non-dopaminergic replacement would need to reproduce.
• Postmortem thalamic tissue shows reduced beta-endorphin (-37.5%, p=0.006) and Met-enkephalin (-26.4%, p=0.028) in RLS (2009-walters-postmortem-opioid-thalamus.md), and a 2024 review integrates this with opioid-receptor-knockout mouse models reproducing RLS-like circadian restlessness, anemia, and low iron (2024-walters-opioid-melanocortin-review.md, 2020-lyu-mu-opioid-receptor-knockout-mouse.md, 2019-lyu-triple-opioid-receptor-knockout-mouse.md).
• Animal work shows brain iron deficiency down-regulates striatal/cortical adenosine A1 receptors by roughly 20-29% alongside striatal D2 receptor loss (~20-35%) and A2A receptor upregulation, and that optogenetically-hypersensitive corticostriatal glutamate terminals are normalized by dipyridamole in an A1-receptor-dependent way (2016-quiroz-adenosine-a1-receptor-downregulation.md, 2019-ferre-adenosine-corticostriatal-terminals.md) — the mechanistic chain (iron deficiency → A1R hypofunction → glutamate hypersensitivity → facilitated dopamine release) is laid out in 2018-ferre-pivotal-adenosine-neurotransmission.md and updated in 2025-mogavero-adenosine-hypothesis-critical-review.md.
• A 2026 rodent study proposes that the same striatal-entopeduncular (striatal-pallidal) pathway, co-expressing mu-opioid and D1 receptors, links dopaminergic hypersensitivity to opioid responsiveness and to opioid-withdrawal-like restlessness (2026-valle-leon-striatal-entopeduncular-opioid-dopamine-model.md).
• A single case report of dextromethorphan (an NMDA antagonist) relieving severe, dopamine-agonist-refractory RLS is hypothesis-generating anecdotal support for glutamate-directed pharmacology (2026-khalil-dextromethorphan-case-report.md, grade E).
• A pilot "pharmaco-TMS" protocol combining dipyridamole/caffeine challenge with TMS biomarkers is registered but has not yet reported results (2024-lanza-tms-pharmaco-protocol-dipyridamole-caffeine.md).
• The foundational 1996 in-vitro work identifying gabapentin's molecular target (the alpha2delta-1 subunit of voltage-gated calcium channels) explains why alpha-2-delta ligands are, mechanistically, a glutamate-release-reducing drug class distinct from dopamine agonists (1996-gee-alpha2delta-gabapentin-binding.md) — relevant background for anyone already using gabapentin, pregabalin, or gabapentin enacarbil as their long-term therapy.

Contested or negative findings

• A follow-up MRS study (4T, MEGA-PRESS, three brain regions) found no significant RLS-vs-control difference in GABA, glutamate, or NAA in ACC, thalamus, or cerebellum — a null result that complicates the simple "RLS = high thalamic glutamate" story from Allen 2013 (2014-winkelman-gaba-glutamate-mrs.md). Region, field strength, and voxel placement differences may explain the discrepancy, but this has not been resolved.
• The classic clinical teaching that RLS-driven sleep loss does not translate into daytime sleepiness is directly contradicted by a large 2025 case-control study: both subjective and objective EDS are significantly increased in untreated RLS, and objective EDS tracks specifically with PLM-associated arousals (2025-chenini-daytime-sleepiness-arousal-rls.md). This reframes rather than confirms the "paradox" named in this topic's brief — the paradox may be smaller, or absent, once objective testing and arousal-specific analysis are used.
• Orexin blockade (suvorexant) improved subjective insomnia and diary-reported sleep in RLS patients whose motor symptoms were already medication-controlled, but produced no significant change on objective actigraphy (TST p=0.58, WASO p=0.99) — a clean negative result showing hyperarousal-targeted drugs do not uniformly fix objectively measured sleep even when patients feel better (2024-winkelman-suvorexant-actigraphy-null.md).
• A 2025 World Sleep Society international survey of 23 national sleep societies found broad global support for most AASM 2025 recommendations, but explicitly did not endorse the dipyridamole recommendation — the adenosine hypothesis's flagship drug remains the single most contested recommendation in the guideline (referenced within 2025-winkelman-aasm-guideline-rls.md).
• Dipyridamole and suvorexant trials to date have enrolled dopamine-agonist-naive (or minimally exposed) patients; efficacy in patients with established augmentation is essentially untested, and the 2025 post hoc analysis suggests prior dopaminergic exposure blunts response to both drugs even before augmentation is diagnosed (2025-garcia-borreguero-subclinical-augmentation-posthoc.md) — a caution against assuming these non-dopaminergic drugs will simply substitute for dopamine agonists in already-treated patients.

Mechanistic picture in 5 bullet points

1. Brain iron deficiency down-regulates adenosine A1 receptors (roughly 20-30% in animal models) in cortex and striatum, alongside the classic striatal D2-receptor down-regulation.
2. A1-receptor loss disinhibits corticostriatal and cortical glutamatergic terminals, producing a presynaptic hyperglutamatergic state; this hypersensitivity is reversible by dipyridamole (raising extracellular adenosine) in animal optogenetic-microdialysis experiments.
3. Cortical glutamate/GABA imbalance manifests as measurable cortical hyperexcitability — reduced short-interval intracortical inhibition on TMS, reproducible across 21 pooled studies, plus impaired inhibitory (LTD-like) cortical plasticity — and thalamic glutamate elevation correlates with objective nocturnal wake time.
4. This glutamatergic/cortical-arousal system is at least partly independent of the dopaminergic/PLMS system: dopamine agonists suppress PLMS without normalizing EEG arousal instability, while GABAergic drugs (clonazepam) do the reverse; separately, a parallel endogenous-opioid deficiency (reduced thalamic beta-endorphin) and orexin/hypocretin arousal signaling each contribute an additional, druggable arousal pathway.
5. The net clinical output is hyperarousal: increased cortical excitability, sleep-onset and maintenance disruption, frequent EEG arousals tied to PLMS, and — contrary to older teaching — measurably increased (not simply masked) objective and subjective daytime sleepiness in a meaningful subset of untreated patients.

Implications for a non-drug long-term plan for people currently on dopamine agonists or alpha-2-delta ligands

• Iron repletion remains foundational: because brain iron deficiency is the proposed upstream trigger for A1-receptor downregulation, correcting ferritin/transferrin saturation (per the AASM 2025 guideline's iron thresholds) is a prerequisite for any strategy aimed at this pathway, not a separate, optional step (2025-winkelman-aasm-guideline-rls.md).
• Adenosine- and orexin-targeted drugs (dipyridamole, suvorexant) are the most evidence-backed non-dopaminergic pharmacological bridges identified in this literature, but both have only been tested, at RCT quality, in dopamine-agonist-naive or lightly-exposed patients; the post hoc augmentation-precursor data (2025-garcia-borreguero-subclinical-augmentation-posthoc.md) suggest they may be less effective the longer someone has already been on a dopamine agonist — arguing for trying them earlier rather than later in a taper, always under physician supervision and never as a substitute for a supervised, gradual dopamine-agonist taper.
• Cortical hyperexcitability (reduced SICI) is a candidate biomarker and a candidate non-drug target: because reduced SICI is reproducible and reversible by dopaminergic treatment, and because low-frequency (inhibitory) rTMS pilot work exists in RLS, cortex-directed neuromodulation is a plausible non-pharmacological avenue worth tracking as it matures — but it remains investigational (pilot/protocol stage only), not a current standard of care.
• Residual insomnia despite motor-symptom control should not automatically be treated as "just RLS hyperarousal that needs a bigger dose": the suvorexant actigraphy-null trial shows that even a mechanistically appropriate hyperarousal-targeted drug can fail on objective sleep measures in this population, so expectations for any single-mechanism non-drug or non-dopaminergic add-on should be modest and outcomes should be tracked objectively (actigraphy/PSG), not by symptom report alone.
• Do not expect this literature to justify stopping medication abruptly: nothing here changes the master-plan rule that dopamine-agonist tapering must be physician-supervised and gradual, given augmentation and DAWS risk; the glutamate/adenosine/opioid pathways described here are best understood as candidate replacement or bridging mechanisms during a supervised taper, not as self-directed alternatives.

Open questions

• Why does thalamic MRS glutamate elevation (Allen 2013) fail to replicate as a global GABA/glutamate deficit in a different multi-region MRS study (Winkelman 2014)? Is this a field-strength/voxel issue, a sub-phenotype issue, or does the abnormality genuinely localize to specific circuits rather than being a global cortical/thalamic shift?
• Does correcting adenosine A1-receptor hypofunction (via dipyridamole or a more selective future agent) actually prevent or reduce dopamine-agonist augmentation risk when used as a first-line agent, or only treat symptoms in parallel? No trial to date has tested dipyridamole specifically as an augmentation-prevention strategy.
• What explains the discrepancy between improved subjective sleep/insomnia scores and unchanged objective actigraphy under suvorexant — is this a sensitivity limitation of actigraphy, a genuine subjective-only drug effect, or evidence that "hyperarousal" in treated RLS with residual insomnia is mechanistically different from the hyperarousal of untreated RLS?
• Is the endogenous-opioid deficiency in RLS (postmortem beta-endorphin loss) a cause or a consequence of chronic sensory/motor symptoms, and does it interact with the adenosine/glutamate axis directly (as the 2026 striatal-entopeduncular rodent model proposes) or run in parallel?
• Can TMS-measured cortical excitability (SICI) be used prospectively to predict which patients will respond to non-dopaminergic (adenosine/orexin/GABAergic) therapy, enabling a personalized, non-dopamine-agonist-first treatment algorithm? This has not been tested.
• Why did the World Sleep Society's international panel not endorse the AASM's dipyridamole recommendation — is this an access/availability issue, a genuine disagreement about evidence quality, or regional treatment culture, and what would change that assessment?

Notes in this topic

• 2013-allen-thalamic-glutamate-mrs.md — Landmark MRS study: thalamic glutamate/glutamine elevated in RLS, correlates with objective nocturnal wake time C.
• 2014-winkelman-gaba-glutamate-mrs.md — Null result: no group difference in GABA/glutamate/NAA across three brain regions; region-specific GABA correlations with PLM severity C.
• 2026-khalil-dextromethorphan-case-report.md — Single case of NMDA-antagonist dextromethorphan relieving dopamine-agonist-refractory RLS E.
• 2016-quiroz-adenosine-a1-receptor-downregulation.md — Animal model: brain iron deficiency down-regulates adenosine A1 receptors ~20-29% alongside D2 receptor loss E.
• 2018-ferre-pivotal-adenosine-neurotransmission.md — Integrative review proposing the iron→A1R→glutamate→dopamine mechanistic chain D.
• 2019-ferre-adenosine-corticostriatal-terminals.md — Rat optogenetic-microdialysis evidence that A1 receptors gate corticostriatal glutamate release, reversible by dipyridamole E.
• 2025-mogavero-adenosine-hypothesis-critical-review.md — 2025 update/critical review of the adenosine hypothesis and its treatment implications D.
• 2018-garcia-borreguero-open-label-dipyridamole.md — Open-label pilot: dipyridamole improves IRLS from 23.4 to 10.7 over 8 weeks C.
• 2021-garcia-borreguero-dipyridamole-rct.md — Placebo-controlled crossover RCT: dipyridamole significantly improves symptoms, arousal index, and slow-wave sleep B.
• 2025-garcia-borreguero-subclinical-augmentation-posthoc.md — Prior dopamine-agonist exposure blunts response to both dipyridamole and suvorexant, even pre-augmentation B.
• 2024-lanza-tms-pharmaco-protocol-dipyridamole-caffeine.md — Registered pilot protocol combining TMS with dipyridamole/caffeine challenge; no results yet E.
• 2026-valle-leon-striatal-entopeduncular-opioid-dopamine-model.md — Rodent model linking dopaminergic hypersensitivity to the mu-opioid-responsive striatal-pallidal pathway E.
• 2009-walters-postmortem-opioid-thalamus.md — Postmortem thalamus: beta-endorphin down 37.5%, Met-enkephalin down 26.4% in RLS C.
• 2024-walters-opioid-melanocortin-review.md — Review integrating opioid-deficiency and melanocortin-system hypotheses of RLS D.
• 2020-lyu-mu-opioid-receptor-knockout-mouse.md — Mu-opioid-receptor knockout mice show RLS-like anemia, iron changes, and rest-phase hyperactivity E.
• 2019-lyu-triple-opioid-receptor-knockout-mouse.md — Triple opioid-receptor knockout mice show hyperactivity, anemia, and dopamine-metabolism changes E.
• 2024-lanza-systematic-review-motor-cortex-tms.md — Systematic review (21 studies, 597 subjects): reduced SICI is the most consistent TMS finding in RLS A.
• 1999-tergau-motor-system-excitability.md — Foundational TMS study: generalized (hand+foot) intracortical disinhibition in RLS C.
• 2019-magalhaes-sici-severity.md — Reduced SICI across the full RLS severity spectrum; discriminates RLS from controls (AUC 0.79) C.
• 2006-nardone-cabergoline-cortical-hyperexcitability.md — Dopamine agonist cabergoline reverses reduced SICI in RLS C.
• 2018-lanza-impaired-ltd-plasticity-rtms.md — Pilot rTMS study: impaired LTD-like cortical plasticity in RLS C.
• 2012-manconi-plm-arousal-dissociation-rct.md — RCT showing PLMS and cortical arousals are pharmacologically dissociable (pramipexole vs clonazepam) B.
• 2002-sforza-cerebral-autonomic-plm-arousal.md — EEG/heart-rate analysis: graded brainstem-to-cortex arousal cascade around periodic leg movements C.
• 2025-chenini-daytime-sleepiness-arousal-rls.md — Large case-control study: both subjective and objective daytime sleepiness are increased in RLS, tracking with PLM-associated arousals B.
• 2024-garcia-borreguero-suvorexant-rct.md — RCT: orexin antagonist suvorexant improves RLS symptoms and sleep architecture B.
• 2024-winkelman-suvorexant-actigraphy-null.md — Negative result: suvorexant improves subjective but not actigraphic sleep in treated RLS with residual insomnia B.
• 2025-winkelman-aasm-guideline-rls.md — (pre-existing note found in the KB) AASM 2025 guideline: strong recommendation for IV iron/gabapentinoids; conditional recommendation for dipyridamole; against pramipexole/ropinirole/rotigotine; strong recommendation against cabergoline D.
• 1996-gee-alpha2delta-gabapentin-binding.md — (pre-existing note found in the KB) Foundational in-vitro paper: gabapentin binds the alpha2delta-1 calcium-channel subunit, the mechanistic root of why alpha2delta ligands reduce glutamate release independent of dopamine E.