Dihexa & GLP-1 Drugs (Ozempic, Wegovy, Mounjaro) and the Brain: Cognition, Dementia Risk, the EVOKE Failure & the 2026 UK Review

Around one million UK adults are now prescribed a GLP-1 receptor agonist — Ozempic, Wegovy, Mounjaro, Rybelsus, Trulicity or Victoza — under NHS or private cover, with NHS England aiming for roughly 220,000 NHS Mounjaro patients in the first three years of the phased weight-management rollout. The class has reshaped cardiometabolic medicine in five years: SELECT showed semaglutide reduced major adverse cardiovascular events by 20% in adults with obesity and established cardiovascular disease without diabetes; FLOW showed semaglutide reduced major kidney events in type 2 diabetes with CKD; STEP-HFpEF showed semaglutide improved heart-failure outcomes in obesity-related HFpEF; SURMOUNT-OSA drove an FDA approval for tirzepatide in obstructive sleep apnoea with obesity. The brain has been the unfinished chapter. Real-world data are striking — a 2025 JAMA Network Open cohort of 60,000+ US adults with type 2 diabetes and obesity reported a hazard ratio of 0.63 for incident dementia on semaglutide or tirzepatide versus other antidiabetics, the US Department of Veterans Affairs analyses across roughly 200,000 patients link GLP-1 receptor agonists to lower neurocognitive-disorder incidence, and a target-trial emulation across 1,710,995 patients showed semaglutide associated with a substantially lower three-year Alzheimer's-disease-related dementia incidence than insulin or any other antidiabetic. The randomised data tell a different story: on 24 November 2025 Novo Nordisk announced that the EVOKE and EVOKE+ Phase 3 trials of oral semaglutide (Rybelsus) versus placebo on top of standard of care, in 3,808 adults with early symptomatic Alzheimer's disease, did not demonstrate statistically significant superiority on the CDR-SB primary endpoint or any secondary cognitive or functional outcome over two years — despite a ~10% reduction in neuroinflammation- and AD-related biomarkers. The 1-year extension was discontinued. The full readout came at AD/PD 2026. The question that brings traffic to this page: do GLP-1 drugs protect the brain, do they cause “Ozempic brain fog”, where does NHS access actually stand in 2026, and where does a synaptogenic HGF/c-Met peptide like Dihexa sit on the converging BDNF / synaptic-plasticity axis downstream of GLP-1 receptor signalling? A rigorous 2026 UK evidence review.

GLP-1 Drugs in the UK in 2026: Scale, Rollout & NICE Pathway

GLP-1 receptor agonists in the UK in 2026 are no longer a niche endocrinology prescription. Around one million UK adults are now on a GLP-1 RA across NHS and private channels combined, with private prescriptions still outnumbering NHS dispensing while the NHS England rollout phases in. The class includes:

• Semaglutide (Novo Nordisk): Ozempic (subcutaneous, type 2 diabetes), Wegovy (subcutaneous, weight management, higher dose), Rybelsus (oral tablet, type 2 diabetes — the formulation tested in EVOKE).
• Tirzepatide (Eli Lilly): Mounjaro in the UK for both type 2 diabetes and weight management (the global brand name Zepbound is used in the US for weight management).
• Liraglutide (Novo Nordisk): Victoza (diabetes), Saxenda (weight management). Daily injection. Older agent.
• Dulaglutide (Lilly): Trulicity (diabetes).
• Exenatide (AstraZeneca, originating Amylin): Byetta (twice daily), Bydureon (weekly). Older agent, the one studied most extensively in Parkinson's disease.
• Pipeline: Retatrutide (triple-agonist GLP-1/GIP/glucagon, Eli Lilly Phase 3), orforglipron (oral non-peptide GLP-1 RA, Lilly Phase 3), CagriSema (semaglutide + cagrilintide, Novo Nordisk).

The NICE pathway is the gating step for NHS access. NICE TA1026 recommended tirzepatide for managing overweight and obesity, with implementation phased over twelve years to manage NHS capacity. NICE TA875 recommended semaglutide (Wegovy) in specialist weight management. Ozempic, separately, is recommended by NICE NG28 for type 2 diabetes within the NICE diabetes pathway alongside other GLP-1 RAs.

The rollout milestones are the practical story. From 23 March 2025, tirzepatide could only be prescribed for weight management through NHS specialist weight management services. The NHS England interim commissioning guidance set the rules. From 23 June 2025, primary-care access expanded under a phased twelve-year rollout. From 1 April 2026, prescribing for obesity was incorporated into the 2026/27 GP contract through new QOF indicators — the operational lever that converts a NICE recommendation into reliable primary-care delivery. NHS England has set a public target of around 220,000 NHS Mounjaro-for-obesity patients in the first three years.

Eligibility tightens the cohort. Initial NHS access prioritises BMI ≥40 with multiple weight-related health conditions (type 2 diabetes, hypertension, dyslipidaemia, OSA, knee osteoarthritis), or BMI ≥35 with an urgent clinical need such as preparation for surgery, fertility treatment, or organ transplant. For South Asian, Chinese, other Asian, Middle Eastern, Black African and African-Caribbean populations the BMI thresholds are reduced by 2.5 kg/m² to account for the higher cardiometabolic risk at a given body mass.

Private channels remain the dominant access route in 2026. Wegovy and Mounjaro are widely prescribed by online pharmacies (Boots, Lloyds Direct, Numan, Voy, Juniper, MyJuniper, heySlim and others) at price points typically £150-£250 per month depending on dose. The MHRA and the General Pharmaceutical Council have issued guidance on appropriate clinical assessment before private GLP-1 RA prescribing. The patient-safety risk is overprescription without weight-related comorbidity screening and inadequate nutrition counselling on top of the rapid caloric restriction that defines successful GLP-1 RA use.

How GLP-1 Drugs Reach the Brain: cAMP, PKA, AMPK & BDNF

The peripheral story is well understood — pancreatic β-cell GLP-1 receptor activation, glucose-dependent insulin secretion, glucagon suppression, delayed gastric emptying, central satiety signalling. The brain story is more interesting and is where the dementia data come from.

GLP-1 receptors in the brain. The 2025 Frontiers in Neuroscience review on GLP-1 and cognitive function in diabetes mellitus catalogued GLP-1 receptor expression on neurons in hippocampus, cortex, hypothalamus, ventral tegmental area, nucleus accumbens, and on microglia and astrocytes. GLP-1 RAs reach the central nervous system either by crossing the blood-brain barrier directly (semaglutide does so modestly; liraglutide and exenatide more so) or by engaging GLP-1 receptors on circumventricular organs and vagal afferents that project centrally. The hypothalamic appetite-suppression effect is well-characterised; the cognitive and reward-system effects are more recent in the literature.

The cAMP → PKA → BDNF axis. The cleanest mechanistic frame, supported by the 2025 Neuroscience review on semaglutide neuroprotection, is that neuronal GLP-1 receptor activation raises intracellular cyclic AMP, activates protein kinase A, and engages CREB-mediated transcription of BDNF and other neurotrophic factors. The downstream effects look familiar from the depression and neurodegeneration literature: increased dendritic spine density, increased long-term potentiation in hippocampal slices, increased adult hippocampal neurogenesis in the dentate gyrus, and improved cognition in animal models of diabetes and Alzheimer's disease.

The AMPK axis. Semaglutide also activates AMP-activated protein kinase, the cellular energy sensor that drives the metabolic-rebalancing effects on hepatic gluconeogenesis, muscle insulin sensitivity, and neuronal autophagy. Brain Sciences 2025 showed that semaglutide and high-intensity interval exercise both improved cognition in diabetic mice, with both increasing hippocampal BDNF and decreasing amyloid-β and phospho-tau — semaglutide acting primarily via PKA, exercise primarily via AMPK, with convergent downstream effects.

The neuroinflammation axis. Microglial GLP-1 receptor activation pushes microglia from a pro-inflammatory M1-like state toward an anti-inflammatory M2-like state. This reduces IL-1β, TNF-α and IL-6 secretion, and is one of the more reproducible mechanisms across animal models of stroke, traumatic brain injury, Parkinson's disease, Alzheimer's disease and metabolic syndrome. The EVOKE biomarker data — ~10% reductions in glial fibrillary acidic protein and other neuroinflammation markers — are consistent with this mechanism even though the clinical CDR-SB primary endpoint missed.

The amyloid and tau axis. In cell and mouse models, semaglutide reduces amyloid-β aggregation and clearance, and reduces tau hyperphosphorylation, in part through improvement of brain insulin signalling. The conceptual frame here is the “type 3 diabetes” hypothesis: Alzheimer's disease as a brain-restricted insulin-resistance disorder. The peripheral metabolic correction with GLP-1 RAs should, on this account, restore central insulin signalling and reverse the amyloid / tau / synapse-loss cascade. The EVOKE failure is the strongest randomised challenge to this hypothesis to date.

The reward axis. GLP-1 receptors in the ventral tegmental area and nucleus accumbens reduce dopaminergic responses to palatable food, addictive drugs (nicotine, alcohol), and gambling rewards. This is the mechanism behind the food-noise-reduction phenomenology that patients describe on Ozempic and Mounjaro, and the same mechanism is being formally tested in alcohol use disorder, nicotine dependence (the NIHR Oxford Health BRC 2025 study), and gambling disorder.

These mechanisms map onto the broader BDNF / synaptic-plasticity story that runs through this site — depression, anxiety, PTSD, Alzheimer's, vascular dementia, diabetic brain fog, post-stroke recovery. The GLP-1 RA story is not orthogonal to the rest of the cognitive-enhancement and neurodegeneration evidence base — it largely converges on the same downstream BDNF axis.

The EVOKE / EVOKE+ Phase 3 Failure (November 2025) and AD/PD 2026

EVOKE and EVOKE+ were the most ambitious clinical test of the metabolic / inflammatory / type-3-diabetes Alzheimer's hypothesis ever attempted. Together they randomised 3,808 adults with early symptomatic Alzheimer's disease, double-blinded, placebo-controlled, on top of standard of care. The intervention was oral semaglutide (Rybelsus) 14 mg, escalated from 3 mg over eight weeks, with a primary endpoint of change in CDR-SB (Clinical Dementia Rating — Sum of Boxes) at 104 weeks. Secondary endpoints covered ADAS-Cog13, ADCS-ADL-MCI, the Mini-Mental State Examination, neuroimaging (volumetric MRI, amyloid PET in a substudy) and CSF biomarkers (Aβ42/40, p-tau181, GFAP, neurofilament light).

On 24 November 2025, Novo Nordisk announced topline results: neither EVOKE nor EVOKE+ demonstrated statistically significant superiority of oral semaglutide over placebo on CDR-SB. The secondary cognitive and functional outcomes were also not meaningfully different. Treatment was well-tolerated, with the GI side-effect profile consistent with prior semaglutide data. Biomarker analyses showed up to ~10% reductions in neuroinflammation- and AD-related biomarkers — statistically significant in the modified intention-to-treat population but not clinically translated. The 1-year extension period was discontinued; the semaglutide-in-Alzheimer's clinical programme has been ended.

The AD/PD 2026 readout provided the granular picture. Pre-specified subgroup analyses across age, sex, APOE genotype, baseline CDR, MMSE strata and amyloid status showed no meaningful between-group differences. The amyloid PET substudy showed no significant reduction in centiloids. The volumetric MRI substudy showed no meaningful preservation of hippocampal or whole-brain volume. The biomarker reductions (GFAP, p-tau181) were real but small and were not associated with any clinical endpoint at the individual-patient level.

The implications matter for the broader brain-health story:

1. Established Alzheimer's disease is not GLP-1-modifiable at the doses and the formulation tested. The window for metabolic-neuroinflammatory intervention may be earlier — pre-symptomatic, MCI, or restricted to higher-risk genotypes such as APOE4 homozygotes (see Section 5).
2. Real-world dementia-prevention signals remain credible but are mechanistically different. The cohorts of GLP-1-treated patients showing lower dementia incidence are predominantly people without AD pathology starting GLP-1 RAs for diabetes or obesity, over five-to-ten-year horizons. EVOKE tested a different question.
3. Biomarker reductions without clinical benefit echo the Alzheimer's-disease-modifying anti-amyloid story (aducanumab, lecanemab, donanemab): biomarker movement is necessary but not sufficient for clinical translation. The 10% GFAP / p-tau181 reductions are real biology; they were not large enough to slow the disease at 104 weeks.
4. Formulation matters. Oral semaglutide (Rybelsus) achieves lower bioavailability and lower CNS exposure than the higher-dose subcutaneous Wegovy formulation. Whether higher CNS exposure would have changed the outcome is unknowable from EVOKE alone; Eli Lilly, with retatrutide and the higher-potency next-generation incretin programmes, will likely run the comparable trial in due course.

The Alzheimer's Research UK statement was sober and correct: the trials closed an important question, do not invalidate the broader neuroprotection signal in metabolic disease, and shift the focus to earlier-stage prevention and APOE4 enrichment.

The Real-World Cohort Data: Where the Dementia Signal Actually Lives

Real-world observational data and randomised trial data have to be read together. EVOKE asked: does oral semaglutide slow established Alzheimer's? Answer: no. The real-world cohorts ask: do people who start GLP-1 RAs for diabetes or obesity develop dementia at a lower rate than matched controls over the following years? Answer: yes, consistently.

JAMA Network Open 2025 (n >60,000). The JAMA Network Open cohort analysed US adults with type 2 diabetes and obesity who initiated semaglutide or tirzepatide between 2017 and 2023, with a propensity-matched comparator on other antidiabetic medications. Over a follow-up of up to seven years, the hazard ratio for incident dementia was 0.63 (95% confidence interval excluding 1), with parallel reductions in stroke and all-cause mortality. The benefits were largest in adults aged 60 years or older, women, and those with BMI in the 30-40 range. The signal held after adjustment for HbA1c, blood pressure, lipid profile and weight change.

US Veterans Affairs cohort (~200,000 patients). The VA database has been mined repeatedly for GLP-1 RA neurocognitive outcomes, with consistent results — roughly 20-35% lower incidence of neurocognitive disorders compared with DPP-4 inhibitors or SGLT2 inhibitors, with the largest effect size for semaglutide.

Target-trial emulation across 1.7 million patients. The 2024 target-trial emulation study used electronic health records from 1,710,995 US patients with type 2 diabetes to emulate a randomised comparison of semaglutide versus seven other antidiabetic classes. Semaglutide was associated with a 40-70% lower three-year incidence of Alzheimer's-disease-related dementia, with the largest effect versus insulin.

The reconciliation with the EVOKE failure is straightforward in principle: the cohorts are intervention-in-people-without-AD-pathology over five-to-ten-year horizons, with the dominant mechanism being metabolic-vascular-inflammatory prevention; EVOKE tested established-AD treatment over two years. The two are different therapeutic windows. The cohorts also have all the well-known limitations of observational data — confounding by indication, prescribing-channel selection, missing data on lifestyle covariates. They are not a substitute for randomised data. They are, however, the strongest pre-randomised signal that the metabolic-neuroinflammatory frame is right and that the timing is critical.

For the UK reader, the practical question is: if you are 55, BMI 34, type 2 diabetic, no current cognitive symptoms, does it matter for your future dementia risk whether you are on semaglutide versus a sulphonylurea? The honest 2026 answer, on the totality of the cohort evidence, is: probably yes, with a modest effect size, in the right direction. That does not constitute medical advice, and treatment-choice conversations belong with your GP and diabetes team. It does mean that the brain-health angle is now part of the antidiabetic decision in a way it was not five years ago.

The APOE4 Homozygote Case for Prevention: Daly 2025

The Daly et al. 2025 Alzheimer's & Dementia paper articulated the cleanest post-EVOKE case for a focused randomised prevention trial: APOE4 homozygotes — the ~2-3% of the population with two copies of the APOE ε4 allele, lifetime AD risk approaching 60-80%, and the strongest documented metabolic-neuroinflammatory phenotype across cohorts. The argument: in APOE4 homozygotes, the combination of disrupted brain lipid handling, attenuated cholesterol clearance, exaggerated microglial inflammatory tone and brain insulin resistance is exactly the substrate that a GLP-1 RA could plausibly modify, before symptoms emerge. The proposed trial design is biomarker-stratified, cognitively-normal or MCI APOE4-homozygote enrichment, with a five-to-seven-year follow-up to capture dementia incidence rather than two-year CDR-SB change.

No such randomised trial has yet read out. The closest in design are the Cognitive Vitality / Alzheimer's Drug Discovery Foundation semaglutide review and the registered exploratory analyses in the Lilly retatrutide and orforglipron programmes. The 2026-2030 window is where the APOE-stratified prevention question is likely to be answered.

For UK readers carrying APOE ε4 (whether one copy or two), the conventional NICE NG97 dementia-prevention guidance applies: blood pressure control, lipid optimisation, smoking cessation, physical activity, hearing-aid use, sleep, mediterranean-style diet, alcohol moderation, cognitive engagement, social engagement. The 2024 Lancet Commission on dementia prevention added LDL cholesterol and uncorrected vision to the 14-factor list. GLP-1 RAs are not on the NICE prevention pathway in 2026 outside their licensed type 2 diabetes and weight-management indications.

“Ozempic Brain Fog”: What Is Actually Happening?

The search-volume curve for “Ozempic brain fog” tracked the Wegovy and Mounjaro UK launches almost month-for-month. The phenomenology is consistent across patient reports: slow thinking, fatigue, difficulty concentrating, occasionally word-finding pauses, typically in the first 4-8 weeks of dose escalation. Brain fog is not listed as an adverse reaction in the Ozempic SmPC, the Wegovy SmPC, or the Mounjaro SmPC. That does not mean it is not real — it means the cognitive complaint is mediated by mechanisms that are documented (hypoglycaemia, dehydration, nutritional inadequacy) rather than by direct GLP-1 RA neurotoxicity.

The probable mechanisms, in rough order of contribution:

1. Hypoglycaemia. GLP-1 RAs cause glucose-dependent insulin secretion — meaningful hypoglycaemia is unusual unless co-prescribed with insulin or a sulphonylurea, which is common in type 2 diabetes management. The early-stage GP review for a new GLP-1 RA prescription should include sulphonylurea down-titration; in private weight-loss prescribing without diabetes, the risk is lower but not zero in someone with prediabetes or significant carbohydrate restriction.
2. Dehydration. The most common GLP-1 RA side effects are nausea, vomiting and constipation. Combined with appetite suppression and reduced food intake, total fluid intake commonly drops. Mild dehydration is a robust cognitive-function depressor.
3. Macronutrient and micronutrient gaps. A 30-40% reduction in caloric intake without corresponding attention to protein, B-vitamin, magnesium, electrolyte and iron intake produces real cognitive symptoms within weeks. The single most important nutrition message for anyone starting Wegovy or Mounjaro is: prioritise protein (1.2-1.6 g/kg/day), micronutrient-dense food, hydration, and a B-complex or food-first equivalent. The NHS wraparound care framework is built around this point.
4. Metabolic transition. Rapid weight loss and improved insulin sensitivity move the body toward partial fat-oxidation and intermittent ketogenesis. Some patients describe a 2-4 week brain-adjustment window comparable to the early phase of a low-carbohydrate dietary intervention.
5. Sleep changes. Weight loss and improved obstructive-sleep-apnoea control should improve sleep; in some patients the early phase of GLP-1 RA use disturbs sleep through nocturnal GI symptoms or appetite-related early-morning waking. Disturbed sleep is a cognitive-function depressor.
6. Mood symptoms. The EMA has had a non-conclusive review of suicidal ideation with GLP-1 RAs since 2023. The signal has not been confirmed; the regulator has not changed labelling at the EU or UK level beyond the existing psychiatric monitoring guidance. Persistent low mood while on a GLP-1 RA is a clinical issue that warrants GP review.

The pragmatic answer for “Ozempic brain fog”, drawing on the Bolt Pharmacy patient guidance and the Drugs.com clinical summary: most cases resolve within 4-8 weeks as the dose stabilises and food intake normalises. The reversible-mechanism shortlist (hydration, protein, B-vitamins, electrolytes, sleep) is the right place to intervene first. Persistence or worsening warrants a GP review and, if necessary, dose adjustment or temporary pause.

None of this is “take a peptide on top of Ozempic” territory. For more on the cognitive-fog literature in general, see the Long COVID brain fog, menopause brain fog, chemo brain and diabetic brain fog reviews on this site, where the same underlying mechanisms (neuroinflammation, BDNF, mitochondrial bioenergetics, sleep) are explored in adjacent contexts.

Where Dihexa Converges with GLP-1 Biology: Two Pathways, One BDNF Destination

The most useful frame for any reader wondering how Dihexa fits the GLP-1 RA brain story is to think of two parallel pathways with different upstream receptors that converge on the same downstream synaptic-plasticity machinery.

The GLP-1 pathway (above): neuronal and microglial GLP-1 receptors → Gα_s-coupled adenylate cyclase → cAMP → PKA → CREB → BDNF transcription. In parallel, AMPK activation, anti-inflammatory microglial polarisation, reduced amyloid-β aggregation, reduced tau hyperphosphorylation, improved brain insulin signalling, and ventral-tegmental dopamine modulation in reward circuits.

The Dihexa / HGF / c-Met pathway: Dihexa (PNB-0408) is a small-molecule angiotensin IV analogue, the canonical mechanistic paper being Benoist CC et al. JPET 2014 — “The procognitive and synaptogenic effects of angiotensin IV-derived peptides depend on activation of the HGF/c-Met system”. Dihexa is reported to stabilise hepatocyte growth factor (HGF), the natural ligand for the Met receptor tyrosine kinase, and potentiate Met signalling. Met activation engages MAPK / ERK1/2 and PI3K / Akt downstream cascades, with reported effects on synaptogenesis (new dendritic-spine formation), angiogenesis (relevant to vascular cognitive impairment and stroke recovery), microglial polarisation toward a regenerative phenotype, and BDNF-adjacent plasticity. The mechanism-of-action page walks through this in detail and the Dihexa vs BDNF page covers the famous “10 million times more potent than BDNF” potency claim in its proper context (acute spinogenesis in cortical neuron cultures, not chronic disease modification).

The convergence point: BDNF expression, dendritic-spine formation, long-term potentiation, and the structural and functional changes that we collectively call synaptic plasticity. Both pathways are pro-plasticity; both reduce neuroinflammation; both have animal-model evidence of cognitive enhancement.

The disagreements: GLP-1 RAs have decades of human data, billions of patient-months of exposure, regulatory approval across three major indications, and randomised cardiovascular / kidney / sleep-apnoea / heart-failure / obesity outcomes data. Dihexa has no published Investigational New Drug application, no published Phase 1, no published clinical trial of any kind, and no published human pharmacokinetic data. The cousin compound, fosgonimeton (ATH-1017) from Athira Pharma, has Phase 2/3 readouts in Alzheimer's disease (LIFT-AD; negative on primary), dementia with Lewy bodies (SHAPE; trended positive) and Parkinson's disease dementia, but no published GLP-1 RA combination data. The clinical translation of the Dihexa story is a fraction of the GLP-1 RA story.

The practical implication: the convergence on BDNF is mechanistically interesting and is one reason synaptogenic peptides are searched for by people who are already on GLP-1 RAs and want to push the cognitive-enhancement dial further. It is not a stacking recommendation. Combining an unlicensed peptide with an active prescription medicine that has its own psychiatric monitoring profile is the lowest-evidence and highest-uncertainty option in 2026. The strongest argument for any reader considering this is that the licensed and the unlicensed pathways look mechanistically complementary; the strongest argument against is that one has clinical evidence and the other does not.

GLP-1 RA Safety in 2026: GI, Pancreatitis, Thyroid, NAION, Mood & Sarcopenia

The randomised safety database for GLP-1 RAs in 2026 is vast and is dominated by the gastrointestinal profile. Nausea, vomiting, diarrhoea, constipation, abdominal pain and dyspepsia are the dominant adverse events, typically dose-dependent and tachyphylactic over weeks. The MHRA and EMA have specific guidance on each of the following lower-frequency but more serious signals:

• Acute pancreatitis: small absolute-risk increase, in patients with prior pancreatitis history or risk factors (alcohol, gallstones, hypertriglyceridaemia). GLP-1 RAs should be discontinued at the first sign of pancreatitis pending evaluation.
• Gallbladder disease: cholelithiasis and cholecystitis are modestly increased; rapid weight loss is a known gallstone risk factor and probably accounts for most of the signal.
• Medullary thyroid cancer / MEN2: contraindicated in personal or family history of medullary thyroid carcinoma or MEN2. The signal originates from rodent C-cell tumour data; the human signal is unconfirmed.
• Diabetic retinopathy: rapid HbA1c reduction (true of GLP-1 RAs and of insulin intensification) is associated with short-term worsening of retinopathy in patients with pre-existing diabetic eye disease. Pre-treatment retinal screening matters in type 2 diabetes.
• NAION (non-arteritic anterior ischaemic optic neuropathy): under EMA review with semaglutide following a 2024 JAMA Ophthalmology case series. Sudden painless visual loss while on semaglutide should prompt urgent ophthalmological assessment.
• Mood and suicidal ideation: the EMA review (2023-2026) has not confirmed a causal signal, but psychiatric monitoring — particularly in patients with a personal history of depression, anxiety, eating disorders, bipolar disorder or self-harm — is reasonable. The 2025 NIHR Oxford Health BRC study found semaglutide safe for brain health and possibly beneficial for nicotine dependence; the suicidal-ideation signal in the EMA review is non-conclusive.
• Sarcopenia and lean-mass loss: rapid weight loss with GLP-1 RAs is roughly 20-40% lean mass and 60-80% fat mass, depending on baseline composition and protein intake. Resistance training, adequate protein (1.2-1.6 g/kg/day), and avoidance of very rapid dose escalation in older patients matter for preserving muscle and bone.
• Gastroparesis: pre-operative anaesthesia guidance now considers GLP-1 RAs as a delayed-gastric-emptying drug and recommends a pre-procedure pause for elective endoscopy and general anaesthesia.

For UK readers on private prescriptions, the most important safety question is the appropriateness of the prescription itself. The MHRA clampdown on illegally sold weight-loss medicines in 2024-2025 highlighted the risk of unregulated channels. NHS or registered private channels with proper clinical assessment, weight-related comorbidity screening, and structured wraparound nutrition and physical-activity advice are the safe route.

Beyond Cognition: GLP-1 RAs in Parkinson's, Nicotine, Alcohol & Reward Disorders

The 2024-2026 GLP-1 RA neurology pipeline goes well beyond Alzheimer's disease.

Parkinson's disease. The longest-running GLP-1 RA story in neurology is exenatide in Parkinson's. The Athauda et al. 2017 Lancet exenatide-PD2 Phase 2 trial reported a positive motor signal versus placebo at 48 weeks, with a sustained between-group MDS-UPDRS-III difference of 3-4 points. The mechanism is hypothesised to be GLP-1 receptor-mediated dopaminergic neuron protection plus reduced microglial neuroinflammation in the substantia nigra. The Cleveland Clinic-sponsored NLY01 Phase 2 in Parkinson's disease subsequently read out negative. The Foltynie / UCL exenatide Parkinson's programme continues. Tirzepatide in Parkinson's is not yet in formal trials. The Parkinson's review on this site covers the HGF/c-Met and BDNF cases for synaptogenic intervention in PD; the GLP-1 RA case is more developed clinically and is currently the most plausible disease-modifying chemical strategy in Parkinson's outside the alpha-synuclein-targeted antibodies (prasinezumab PADOVA, minzasolmin ORCHESTRA).

Nicotine dependence. The 2025 NIHR Oxford Health Biomedical Research Centre study found semaglutide safe for brain health with possible benefits for nicotine dependence (smoking cessation). Mechanistically this fits the ventral-tegmental / nucleus-accumbens reward-circuit GLP-1 receptor biology. Formal nicotine-cessation indications are not yet under regulatory review.

Alcohol use disorder. Multiple small randomised trials of semaglutide and liraglutide in alcohol use disorder have shown reduced craving and reduced drinking in 2024-2025. The mechanism is the same reward-circuit one. The Yale and Pennsylvania programmes have larger ongoing trials. The phenomenology — reduced reward-pull from alcohol, reduced ruminative drinking — is consistent with the food-noise-reduction reports from weight-loss patients.

Obstructive sleep apnoea. SURMOUNT-OSA (Malhotra et al., NEJM 2024) showed tirzepatide reduced apnoea-hypopnoea index in adults with obesity and moderate-to-severe OSA, leading to FDA approval of tirzepatide for OSA in obesity in December 2024. The brain-fog implication is non-trivial — OSA is a major cause of daytime cognitive dysfunction, and the GLP-1 RA effect on weight reduces OSA and probably reduces OSA-driven cognitive symptoms.

Cardiovascular outcomes that protect the brain. SELECT (Lincoff et al., NEJM 2023) showed a 20% reduction in major adverse cardiovascular events with semaglutide in adults with obesity and established cardiovascular disease without diabetes. FLOW (Perkovic et al., NEJM 2024) showed major kidney-event reduction in T2DM with CKD. STEP-HFpEF (Kosiborod et al., NEJM 2023) showed heart-failure-with-preserved-ejection-fraction symptom improvement. Stroke risk reduction is a likely contributor to the dementia-incidence reduction in the real-world cohorts — vascular cognitive impairment and Alzheimer's-disease-related dementia overlap considerably in old age. See the vascular dementia review and post-stroke recovery review for the related cerebrovascular biology.

The Next-Generation Pipeline: Retatrutide, Orforglipron, CagriSema

The 2026-2030 window is dominated by next-generation incretin pharmacology.

Retatrutide (Eli Lilly). A triple agonist of GLP-1, GIP and glucagon receptors, in Phase 3 in obesity, type 2 diabetes, MASH and obstructive sleep apnoea. Phase 2 weight-loss results suggested ~24% weight reduction at 48 weeks — the largest pharmacological weight-loss effect to date. Brain-health subprogrammes are exploratory.

Orforglipron (Eli Lilly). An oral, non-peptide, small-molecule GLP-1 receptor agonist in Phase 3 in obesity and type 2 diabetes. Reads more like a conventional small molecule than a peptide, with implications for cost, manufacturing, and access. Brain-health subprogrammes are exploratory.

CagriSema (Novo Nordisk). Semaglutide combined with cagrilintide (an amylin analogue), in Phase 3 in obesity and type 2 diabetes. Aiming for incremental weight loss above semaglutide monotherapy.

Eli Lilly's Alzheimer's positioning. Following the EVOKE failure with semaglutide, the question of whether tirzepatide, retatrutide or orforglipron would have done better in early Alzheimer's is open. Eli Lilly has been publicly cautious about committing to a parallel Phase 3 in established AD without stronger biomarker rationale, and the most likely next move is APOE-stratified prevention rather than symptomatic AD.

The implications for cognitive enhancement and dementia prevention are: incretin pharmacology will continue to deliver weight, cardiovascular, renal and possibly cognitive benefits at population scale, the magnitude of weight loss and metabolic correction is increasing with each generation, and the cleanest 2030 prevention-trial design is APOE4-enriched cognitively-normal-to-MCI populations on the highest-potency available incretin combination, with five-to-seven-year follow-up.

Dihexa Specifically for Someone Already on Ozempic, Wegovy or Mounjaro

This is the section that does most of the cognitive work on this page for a reader who is themselves prescribed a GLP-1 RA and is searching for synaptogenic / nootropic options on top.

The pre-question: should you stack at all? The strongest argument for any cognitive-enhancement intervention is unmet need. If you are on Mounjaro for obesity and your cognition is fine, the marginal cognitive return on adding an unlicensed peptide is small and the uncertainty cost is real. If you are on Ozempic for type 2 diabetes and notice clear brain fog that has not resolved after 8-12 weeks of dose stabilisation with adequate hydration, protein, B-vitamins and sleep, the right next step is a GP review and a structured workup for hypoglycaemia, OSA, depression, hypothyroidism, anaemia, B12/folate deficiency — not a peptide.

What is theoretically interesting. The convergence on BDNF and synaptic plasticity from two different upstream receptors is genuinely interesting biology. The Dihexa preclinical literature is dominated by Alzheimer's-relevant paradigms (Morris water maze, scopolamine reversal, spatial memory) and the GLP-1 RA preclinical literature is dominated by the same paradigms in diabetic and high-fat-fed mice. The two interventions plausibly hit different upstream receptors that converge on the same hippocampal synaptic-plasticity machinery.

What is missing. There is no published Dihexa × GLP-1 RA combination preclinical study. There is no published Dihexa human PK study. There is no published Dihexa drug-drug interaction data with any prescription medicine. The fosgonimeton (ATH-1017) clinical analogue has not been studied in combination with any GLP-1 RA. The total Dihexa × GLP-1 RA evidence base in 2026 is the null set.

What the practical risk-benefit looks like. Adding an unlicensed peptide to a GLP-1 RA introduces uncharacterised cardiovascular, hepatic, renal and psychiatric pharmacology on top of an active intervention that already has a documented psychiatric-monitoring profile (the EMA suicidal-ideation review), an active cardiovascular and hepatic safety profile, and an active GI side-effect profile. The risk floor is non-zero. The benefit ceiling, in the absence of any human data, is also unknown.

The honest framing. If you are searching for ways to push the cognitive dial on top of a GLP-1 RA, the highest-evidence, lowest-risk options are well-established: sleep optimisation, structured aerobic and resistance training (the Brain Sciences 2025 paper showed that high-intensity interval exercise matched semaglutide on hippocampal BDNF in diabetic mice), Mediterranean-style nutrition, cognitive engagement, smoking cessation, alcohol moderation, vision and hearing correction, social engagement. The 2024 Lancet Commission 14-factor list is the cleanest summary. Unlicensed synaptogenic peptides are below this list, not above it.

For broader context on the Dihexa risk-benefit conversation, see the side-effects page, the dosage page, the UK legal status page, the research-chemicals UK overview and the rolling series on specific indications. For metabolic-cognitive overlap specifically, the diabetic brain fog review is the closest neighbour.

2025-2026 News Context: Five Stories That Matter

1. EVOKE / EVOKE+ topline (24 November 2025). Novo Nordisk announced that oral semaglutide did not slow Alzheimer's disease progression in 3,808 adults with early symptomatic AD. The semaglutide-in-AD programme has ended. The biomarker reductions were real but small. The implication for the broader brain-health story is that the metabolic-neuroinflammatory frame is right, but the therapeutic window is upstream of established AD.

2. AD/PD 2026 full readout. The Vienna AD/PD International Conference in 2026 published the full EVOKE and EVOKE+ subgroup analyses, biomarker substudies and volumetric MRI data, confirming the topline. Clinical Trials Arena’s analyst comment framed the read as closing the symptomatic-AD chapter and opening the prevention chapter.

3. NHS England Mounjaro phased rollout milestones. 23 March 2025 (specialist services), 23 June 2025 (primary care expansion), 1 April 2026 (QOF integration). NHS England targeting 220,000 NHS Mounjaro-for-obesity patients in the first three years under NICE TA1026. The brain-health implications are at population scale.

4. NIHR Oxford Health BRC semaglutide brain-health study (2025). The University of Oxford / NIHR Oxford Health BRC study concluded semaglutide was safe for brain health with possible benefits for cognition and nicotine dependence. The most useful UK-relevant data point for the general patient population.

5. EMA suicidal-ideation review and the NAION optic-nerve signal. Both ongoing in 2026. The mood-and-vision safety conversation has not changed prescribing in the UK at scale; it has changed prescribing-conversation expectations, and it is part of why the “Ozempic brain” search-volume curve is what it is.

What Would Need to Happen for Dihexa to Become a Real GLP-1-Adjacent Option?

The same checklist used on the Huntington's, schizophrenia, vascular dementia and FTD reviews applies here, with two GLP-1-specific additions:

1. Preclinical work in a GLP-1 RA-treated diabetic or obese animal model, showing additive or synergistic effects on hippocampal BDNF, spinogenesis and behavioural cognitive endpoints. None published to date.
2. Preclinical pharmacokinetic and drug-drug interaction work: does Dihexa alter GLP-1 RA exposure or vice versa, do they share CYP metabolism, do they share renal clearance pathways. None published to date.
3. Phase 1 healthy-volunteer PK/PD data for Dihexa. Fosgonimeton has Phase 1 data; Dihexa itself does not.
4. Phase 2 proof-of-concept in a defined cognitive endpoint — either in GLP-1 RA-treated patients with metabolic disease and brain-fog complaints, or in MCI / Alzheimer's-disease populations, with a recognised cognitive battery primary endpoint and biomarker substudy.
5. Phase 3 trial in a regulatory-meaningful indication, MHRA / EMA / FDA approval, and a NICE health-technology appraisal pathway before NHS uptake.
6. NICE technology appraisal that places Dihexa or its successors in a coherent position in the cognitive-decline pathway, either alongside or downstream of GLP-1 RAs, with explicit cost-effectiveness modelling against the lecanemab / donanemab pathway and the conventional NICE NG97 dementia-prevention list.

None of the above exists for Dihexa in any GLP-1 RA-related indication in 2026. By contrast, the GLP-1 RA class has produced six positive Phase 3 cardiovascular, renal, sleep-apnoea and weight-management outcome trials in the same period, and one Phase 3 Alzheimer's-disease failure. The clinical evidence asymmetry is large.

The Bottom Line: A Real Brain Story, a Real Failed Trial, and Where Dihexa Honestly Sits

GLP-1 receptor agonists are now the most clinically successful systemic intervention in cardiometabolic medicine of the last decade. They have moved from diabetes to weight management to cardiovascular outcomes to kidney outcomes to sleep apnoea, and they have a credible — if not yet randomly-proven — signal for dementia prevention in metabolic disease populations over five-to-ten-year horizons. They are not Alzheimer's disease-modifying once symptoms are established: EVOKE and EVOKE+ closed that chapter in November 2025. They probably are causally upstream of part of the lower dementia incidence seen in the real-world cohorts. The mechanism — cAMP / PKA / AMPK / BDNF / anti-neuroinflammation / improved brain insulin signalling — is real and overlaps with the broader BDNF and synaptic-plasticity literature that runs through this site.

For the UK adult who is currently on Ozempic, Wegovy or Mounjaro and worried about brain fog or curious about cognitive enhancement, the honest 2026 read is: stabilise the basics (sleep, hydration, B-vitamin and protein intake), use the NIHR Be Part of Research portal for legitimate trials, and treat unlicensed peptide self-experimentation on top of a GLP-1 RA as the lowest-evidence and highest-uncertainty option in your toolkit. The Dihexa story is mechanistically convergent on the same BDNF / synaptic-plasticity downstream — but the upstream evidence base, the clinical translation and the safety data are an order of magnitude thinner than the GLP-1 RA story, and combining the two introduces uncharacterised pharmacology on top of an already-active prescription medicine. The mechanistic case is interesting; the evidence case is not yet there.

The right next steps depend on the reason you arrived at this page: if you are on a GLP-1 RA and worried, talk to your GP; if you are interested in the mechanistic story, read the BDNF, mechanism and diabetic brain fog pages alongside this one; if you are interested in trials, the NIHR portal is the place to start; if you are interested in the licensed clinical analogue, read the fosgonimeton page. We will update this article as the AD/PD 2026 post-hoc analyses, the retatrutide and orforglipron readouts, and the APOE4-stratified prevention work mature.

Frequently Asked Questions: Dihexa & GLP-1 Drugs

Do Ozempic, Wegovy and Mounjaro protect against dementia?

Real-world cohorts of more than two million people consistently link GLP-1 receptor agonists to 20-40% lower dementia incidence in type 2 diabetes and obesity, with the largest effect for semaglutide and tirzepatide. The 2025 JAMA Network Open cohort reported HR 0.63 for incident dementia. The November 2025 EVOKE and EVOKE+ Phase 3 trials of oral semaglutide in 3,808 adults with established early Alzheimer's disease, however, did not slow clinical progression. The cleanest interpretation is that GLP-1 RAs are likely preventive in metabolic disease populations over years, not disease-modifying in established symptomatic Alzheimer's.

Why did EVOKE and EVOKE+ fail if the cohort data are positive?

Different therapeutic windows. The cohorts are years-to-decades of GLP-1 RA exposure in people without AD pathology, with the dominant benefit being metabolic-vascular-inflammatory prevention. EVOKE was 104 weeks in people with established symptomatic Alzheimer's disease, where the amyloid-tau-synapse-loss cascade is far advanced. Biomarker reductions of ~10% were real but were not large enough to translate clinically. Formulation (oral Rybelsus rather than higher-dose subcutaneous Wegovy) and dose may also matter; whether higher-CNS-exposure incretins (Mounjaro, retatrutide) would have done better is an open question.

Is Ozempic brain fog real?

Brain fog is not listed in the official SmPC for Ozempic, Wegovy or Mounjaro, but transient cognitive symptoms in the first 4-8 weeks of dose escalation are widely reported. The mechanisms are mundane and reversible: hypoglycaemia (with insulin or sulphonylurea co-prescription), dehydration from GI side effects, protein / B-vitamin / electrolyte / micronutrient gaps from rapid caloric restriction, and metabolic adjustment to ketone-shifted fuel supply. Most cases resolve on dose stabilisation and adequate nutrition. Persistent or worsening cognitive symptoms warrant a GP review.

How does semaglutide affect BDNF?

Semaglutide activates neuronal GLP-1 receptors → Gα_s-coupled adenylate cyclase → cAMP → PKA → CREB → BDNF transcription. It also activates AMPK and shifts microglia toward an anti-inflammatory phenotype, reducing IL-1β, TNF-α and IL-6. Animal models in 2025 (Brain Sciences) showed semaglutide increased hippocampal BDNF and reduced amyloid-β and phospho-tau, with effects comparable to high-intensity interval exercise. The clinical translation is positive in real-world cohorts and negative in EVOKE.

When is Mounjaro available on the NHS for weight loss?

From 23 March 2025 (specialist weight management services only), expanded into primary care from 23 June 2025, with full GP-contract integration via the 2026/27 QOF indicators from 1 April 2026. NHS England targets 220,000 NHS Mounjaro patients in the first three years of the phased rollout. Initial eligibility prioritises BMI ≥40 with multiple weight-related conditions or BMI ≥35 with urgent clinical need; thresholds are reduced by 2.5 kg/m² for South Asian, Chinese, other Asian, Middle Eastern, Black African and African-Caribbean populations.

Can I take Dihexa with Ozempic, Wegovy or Mounjaro?

There is no published Dihexa × GLP-1 RA drug-drug interaction or pharmacokinetic data. Dihexa is not a licensed medicine for any indication, has no published human PK or safety data, and is not approved by the MHRA, FDA or EMA. Self-combining an unlicensed peptide with a prescribed GLP-1 RA — which has its own active psychiatric monitoring and safety profile under EMA review — introduces uncharacterised risk on top of an active intervention. The mechanistic convergence on BDNF is interesting but does not constitute a stacking recommendation.

What about GLP-1 RAs in Parkinson's disease?

The longest-running neurology story for GLP-1 RAs is exenatide in Parkinson's. Athauda et al. 2017 Lancet reported a positive Phase 2 motor signal at 48 weeks; the Cleveland Clinic-sponsored NLY01 Phase 2 read out negative. The Foltynie / UCL exenatide programme continues. Tirzepatide and the next-generation incretins are not yet in formal Parkinson's trials. See the Parkinson's review for the related HGF/c-Met and BDNF biology.

Should I take a GLP-1 RA to prevent dementia if I have APOE4?

No randomised trial has yet tested GLP-1 RAs for dementia prevention in APOE4 carriers (heterozygotes or homozygotes). The Daly et al. 2025 paper made the case for such a trial in APOE4 homozygotes specifically. Until those data exist, GLP-1 RAs in the UK are licensed for type 2 diabetes and obesity in defined eligibility categories, and APOE genotype is not part of the prescribing criteria. The NICE NG97 dementia-prevention guidance (vascular risk factors, lifestyle, sensory correction, social engagement) remains the operative framework.

Where can I find UK trials of GLP-1 drugs for brain disorders?

The NIHR Be Part of Research portal lists open UK studies in dementia prevention, Parkinson's disease, obesity, type 2 diabetes and related comorbidities. The University of Oxford NIHR Oxford Health Biomedical Research Centre and the UCL clinical-trials network are the most active sites for neurology-relevant GLP-1 RA research in the UK in 2026.

What is the difference between Ozempic, Wegovy and Rybelsus?

All three are semaglutide (Novo Nordisk). Ozempic is the subcutaneous weekly injection licensed for type 2 diabetes and (in the SELECT-indicated population) cardiovascular risk reduction in established CVD with obesity. Wegovy is the higher-dose subcutaneous weekly injection licensed for weight management. Rybelsus is the oral tablet (lower CNS exposure) licensed for type 2 diabetes — the formulation tested in EVOKE / EVOKE+.

What is the difference between Mounjaro and Zepbound?

Both are tirzepatide (Eli Lilly). Mounjaro is the UK brand name across both type 2 diabetes and weight-management indications. Zepbound is the US brand name for the weight-management indication. The molecule and the device are the same; the brand name and the licensing pathway differ by region.

Sources & Citations

• Novo Nordisk A/S: Evoke phase 3 trials did not demonstrate a statistically significant reduction in Alzheimer's disease progression (24 November 2025).
• Results of EVOKE and EVOKE+ trials show no effect of oral semaglutide on AD progression — Alzheimer Europe (November 2025).
• AD/PD 2026: Novo Nordisk’s semaglutide fails to demonstrate any benefit for early Alzheimer’s — Clinical Trials Arena.
• Weight loss drug does not slow Alzheimer’s, says drug maker — Alzheimer’s Research UK.
• Semaglutide, Tirzepatide Cut Dementia, Stroke Risk in Type 2 Diabetes — JAMA Network Open cohort (n > 60,000), 2025.
• Associations of semaglutide with Alzheimer’s disease-related dementias in patients with type 2 diabetes: A real-world target trial emulation study (n = 1,710,995).
• Daly K et al. Targeting dementia prevention with semaglutide: The case for APOE4 homozygotes. Alzheimer’s & Dementia 2025.
• Semaglutide and High-Intensity Interval Exercise Attenuate Cognitive Impairment in Type 2 Diabetic Mice via BDNF Modulation. Brain Sciences 2025.
• Neuroprotective and cognitive benefits of Semaglutide: Insights into the underlying molecular mechanisms. Neuroscience 2025.
• Effect and mechanism of GLP-1 on cognitive function in diabetes mellitus. Frontiers in Neuroscience 2025.
• Alzheimer’s Disease as Type 3 Diabetes: The Impact of Insulin Resistance and the Therapeutic Potential of Semaglutide in Cognitive Decline.
• NIHR Oxford Health BRC: New study finds semaglutide safe for brain health with possible benefits for cognition and nicotine dependence (2025).
• Cognitive Vitality — Semaglutide drug report (Alzheimer’s Drug Discovery Foundation).
• ALZFORUM Therapeutics — Semaglutide.
• GLP-1 Receptor Agonists Show Translational Potential in Alzheimer Disease Despite Mixed Study Results — NeurologyLive.
• BrightFocus — How GLP-1s Could Transform Alzheimer’s Treatment.
• Lincoff AM et al. Semaglutide and Cardiovascular Outcomes in Obesity Without Diabetes (SELECT). NEJM 2023.
• Perkovic V et al. Effects of Semaglutide on Chronic Kidney Disease in Patients with Type 2 Diabetes (FLOW). NEJM 2024.
• Kosiborod MN et al. Semaglutide in Patients with Heart Failure with Preserved Ejection Fraction and Obesity (STEP-HFpEF). NEJM 2023.
• Malhotra A et al. Tirzepatide for the Treatment of Obstructive Sleep Apnea (SURMOUNT-OSA). NEJM 2024.
• FDA approves first medication for obstructive sleep apnea (tirzepatide, December 2024).
• Athauda D et al. Exenatide once weekly versus placebo in Parkinson’s disease (exenatide-PD2). The Lancet 2017.
• NLY01 Phase 2 in Parkinson’s Disease — clinicaltrials.gov.
• NICE TA1026. Tirzepatide for managing overweight and obesity.
• NICE TA875. Semaglutide for managing overweight and obesity.
• NICE NG28. Type 2 diabetes in adults: management.
• NICE NG97. Dementia: assessment, management and support for people living with dementia and their carers.
• NHS England — Weight management injections.
• NHS England — Interim commissioning guidance for the implementation of NICE TA1026 tirzepatide (March 2025).
• Ozempic (semaglutide) Summary of Product Characteristics — eMC.
• Wegovy (semaglutide) Summary of Product Characteristics — eMC.
• Mounjaro (tirzepatide) Summary of Product Characteristics — eMC.
• EMA — Ongoing review of GLP-1 receptor agonists (suicidal ideation signal).
• Bolt Pharmacy — Ozempic and Brain Fog: Causes, Management and When to Seek Help.
• Drugs.com — Does Ozempic cause brain fog?
• Drugs.com — Do GLP-1 drugs like Ozempic reduce dementia risk?
• Livingston G et al. Dementia prevention, intervention, and care: 2024 report of the Lancet Standing Commission. The Lancet 2024.
• Benoist CC et al. The procognitive and synaptogenic effects of angiotensin IV-derived peptides depend on activation of the HGF/c-Met system. JPET 2014.
• UK Weight Loss Medication Statistics 2026: 60+ Data Points on Adoption, Demographics, and the NHS vs Private Market.
• NIHR Be Part of Research portal.
• MHRA — Safety clampdown on illegally sold weight-loss medicines.