From Claude Opus 5.5 (High Effort):
What drugs currently show promise as an glymphatic enhancer?
Here’s the landscape as of now, tiered by evidence strength.
Tier 1: Human clearance data exists
ACX-02 (dexmedetomidine plus midodrine) is the only pharmacological intervention with published human evidence of enhanced brain-to-blood clearance of amyloid and tau.
The backstory matters because it is a direct test of the paper you just sent me. Applied Cognition, with Jeff Iliff and Brendan Lucey as co-authors, ran dexmedetomidine alone first, exactly as the 2022 opinion proposed. It failed. DEX increased EEG slow waves substantially, but it did not reduce brain parenchymal resistance and produced no treatment effect on plasma amyloid or tau ratios. The authors explicitly note they “were surprised” by this, given the rodent literature and the suggestion that DEX was a leading glymphatic enhancer candidate. medrxivmedrxiv
Their explanation: DEX dropped mean arterial pressure by about 19.8 mmHg, which triggered compensatory cerebral autoregulatory vasodilation, increasing cerebral blood volume and effectively closing the perivascular spaces upstream of the interstitium. The drug opened the sleep channel and closed the plumbing. medrxiv
So they added midodrine, a peripherally acting alpha-1 agonist that does not cross the blood-brain barrier, to hold blood pressure up while DEX suppressed central noradrenergic tone.
Results, in 19 healthy adults aged 55 to 64, crossover design, single 4-hour-15-minute sleep opportunity:
| Measure |
ACX-02 vs placebo |
| Parenchymal resistance |
-51.4% |
| Slow wave count |
765 vs 427 |
| Mean arterial pressure |
no significant change |
| Aβ42/Aβ40 mass ratio clearance |
+8.45% (p = 0.024) |
| %p-tau217 clearance |
+9.66% (p = 0.049) |
Bayesian mediation found parenchymal resistance was the dominant mediator, accounting for 68 to 81 percent of the treatment effect under ACX-02 versus 12 to 31 percent under placebo. medrxiv
What to hold against it: n=19 completers, single dose, healthy people with no amyloid pathology and no sleep disorder, fixed treatment-then-placebo order rather than randomized sequence, no correction for multiple comparisons, intravenous dexmedetomidine with an arterial line (not something you take at home), and the clearance endpoint is a plasma mass-balance inference rather than direct imaging. Four authors including the first author hold stock options in Applied Cognition. It’s a preprint posted in March 2026 and not yet peer-reviewed. medrxivmedrxiv
Their extrapolation is the part I’d flag hardest: they argue a sustained 10 percent clearance increase would translate to “a multi-year delay in reaching pathological Aβ thresholds” given accumulation rates around 3.0 centiloids per year. That is a chain of assumptions about durability, tolerance, and linearity that a single 4-hour session cannot support [Confidence: Low]. medrxiv
Still, this is the first time anyone has shown pharmacological glymphatic enhancement in humans with a clearance readout, and the fact that it corrected a failed first attempt by a specific mechanistic hypothesis makes it more credible, not less.
Tier 2: Human biomarker data, mechanism probably not glymphatic
Dual orexin receptor antagonists (suvorexant, lemborexant, daridorexant). Suvorexant 20 mg acutely decreased CSF amyloid-beta by about 10 to 20 percent versus placebo starting 5 hours after dosing, and reduced the p-tau181/tau181 ratio by about 10 to 15 percent, though phosphorylation at S202 and T217 was unchanged. wustl
The interesting wrinkle: Lucey has noted that suvorexant did not increase sleep compared to placebo, which led him to hypothesize that orexin signaling itself, not improved sleep, mediated the tau reduction. Supporting this, lemborexant in tauopathy mice reduced tau phosphorylation via protein kinase A, and treated animals had hippocampus and overlying cortex 30 to 40 percent larger than controls. So this may be production reduction rather than clearance enhancement, which is a different lever. alzforumalzforum
Two Phase 2 trials are running: SToP-AD testing 24 months of suvorexant 20 mg in healthy elderly with amyloid and tau biomarker endpoints, and a lemborexant trial with 6-month CSF endpoints. These are the longest-duration human tests of any sleep-based anti-amyloid strategy. nih
Tier 3: Non-drug, but further along clinically
40 Hz gamma sensory stimulation has the best mechanistic package and the largest trial. Murdock’s 2024 Nature paper showed multisensory 40 Hz stimulation promotes amyloid clearance through AQP4-dependent glymphatic transport, and that both genetic AQP4 reduction and the AQP4 blocker TGN-020 abolished the effect. Reported magnitudes in 5xFAD mice: meningeal lymphatic vessel diameter up 19 percent, cortical CSF tracer accumulation up 34 percent, amyloid efflux to cervical lymph nodes up 47 percent. The chain runs through VIP interneuron peptide release driving arterial pulsatility; chemogenetically silencing those neurons abolished both the pulsatility increase and the amyloid clearance. Multisensory gamma stimulation promotes glymphatic clearance of amyloid +2
Cognito Therapeutics is running HOPE, a Phase 3 trial of Spectris AD in 670 patients with mild-to-moderate Alzheimer’s. That readout is the most consequential near-term test of the glymphatic hypothesis in humans, from any modality. gamma40
Tolerability is not nothing: in OVERTURE, headache occurred in 20.7 percent and tinnitus in 15.2 percent of the active group versus 10.7 percent and 0 percent in sham. biorxiv
Also in early human testing: intermittent hypercapnia combined with non-invasive vagus nerve stimulation in Parkinson’s and healthy older adults, and centromedian thalamic low-intensity focused ultrasound in amyloid-positive MCI at UCSF. Both are pilot-stage. clinicaltrialsclinicaltrials
Tier 4: Preclinical only
-
AQP4 modulators. The obvious target, and the most frustrating. TGN-020 is used experimentally as an inhibitor; what you’d want is a polarization enhancer, and nothing clinical-grade exists. The field also has not settled whether AQP4 inhibitors work through AQP4 at all.
-
Statins, GLP-1 receptor agonists, SGLT2 inhibitors. The ACX-02 authors themselves invoke the epidemiological signal that these cardiovascular drugs are associated with reduced dementia risk, and propose cerebrovascular compliance as the shared mechanism. That is a hypothesis built on observational data, not a glymphatic claim [Confidence: Low]. medrxiv
-
Omega-3, melatonin, various TCM compounds. Rodent tracer studies, no human clearance data.
What appears to work against clearance
This is the more actionable half. Zolpidem suppressed norepinephrine oscillations and reduced glymphatic CSF tracer inflow in mice, with fluid transport dropping more than 30 percent despite faster sleep onset. Norepinephrine waves during deep sleep were 50 percent lower in zolpidem-treated mice than in naturally sleeping ones. Benzodiazepines reduce slow-wave sleep and carry their own dementia association. Norepinephrine-mediated slow vasomotion drives glymphatic clearance during sleep: Cell +2
The pattern across zolpidem, dexmedetomidine alone, and ACX-02 is consistent and worth stating plainly: inducing sleep does not enhance clearance. Delta power is not the mediator. What separates ACX-02 from both failures is vascular, not hypnotic.
The honest bottom line
Four years ago the hypothesis was “find a drug that makes deep sleep.” Today the operative model is “find a drug that simultaneously drops interstitial resistance, preserves vascular pulsatility, and does not trigger autoregulatory vasodilation.” Three constraints, not one, and two of them are vascular.
Nothing is available to use. ACX-02 requires IV infusion and arterial monitoring in its current form. The DORAs are prescription insomnia drugs with biomarker data but unproven mechanism and no outcome data. 40 Hz devices are investigational. The only thing with a plausible claim to being both available and mechanistically aligned is avoiding GABAergic hypnotics if you can.
If you’re tracking this for publication, the two dates that matter are the HOPE Phase 3 readout and whether the ACX-02 preprint survives peer review with those effect sizes intact.
One caveat on everything above: all human clearance endpoints here are plasma or CSF biomarker proxies measured over hours. Not one study has shown that pharmacologically enhanced clearance changes amyloid PET burden, cognition, or dementia incidence. The field is currently at the target-engagement stage, which is roughly where anti-amyloid antibodies were two decades before lecanemab.
References
- Dagum P, Satterfield BC, Giovangrandi L, Feng TR, Corbellini A, Yarasheski K, Lucey BP, Van Dongen H, Iliff JJ, Cheung AT. Pharmacological enhancement of glymphatic function in humans increases the clearance of Alzheimer’s disease-related proteins. medRxiv 2026. Pharmacological enhancement of glymphatic function in humans increases the clearance of Alzheimer’s disease-related proteins
- Hauglund NL, et al. Norepinephrine-mediated slow vasomotion drives glymphatic clearance during sleep. Cell 2025;188(3):606-622. https://www.cell.com/cell/fulltext/S0092-8674(24)01343-6
- Murdock MH, et al. Multisensory gamma stimulation promotes glymphatic clearance of amyloid. Nature 2024. Multisensory gamma stimulation promotes glymphatic clearance of amyloid | Nature
- Lucey BP, Liu H, Bateman RJ. Suvorexant acutely decreases tau phosphorylation and Aβ in the human CNS. Annals of Neurology 2023. https://profiles.wustl.edu/en/publications/suvorexant-acutely-decreases-tau-phosphorylation-and-aβ-in-the-hu/
- Orexin receptor antagonists for the prevention and treatment of Alzheimer’s disease. PMC. Orexin Receptor Antagonists for the Prevention and Treatment of Alzheimer’s Disease and Associated Sleep Disorders - PMC
- Night-Night, Neurodegeneration: Sleep Aid Lulls Tau Pathology. Alzforum. https://www.alzforum.org/news/research-news/night-night-neurodegeneration-sleep-aid-lulls-tau-pathology
- Research progress on 40 Hz sensory stimulation for Alzheimer’s disease. PMC. Research progress on 40 Hz sensory stimulation for the treatment of Alzheimer’s disease - PMC
- How sensory gamma rhythm stimulation clears amyloid in Alzheimer’s mice. MIT News. How sensory gamma rhythm stimulation clears amyloid in Alzheimer’s mice | MIT News | Massachusetts Institute of Technology
- Synthetic Vasomotion for Glymphatic Enhancement (NCT07814534). ClinicalTrials.gov
- CMN LIFU for Enhanced Glymphatic Clearance in Amyloid-Positive MCI (NCT07866768). ClinicalTrials.gov
- Sleep aid disrupts norepinephrine-mediated waste clearance in mice. GEN. Sleep Aid Disrupts Norepinephrine-Mediated Waste Clearance in Mice
- Using magnetoencephalography to track 40 Hz invisible spectral flicker (OVERTURE adverse event data). bioRxiv. https://www.biorxiv.org/content/10.1101/2025.01.07.631693.full.pdf