Dihexa for Schizophrenia: KarXT/Cobenfy, Iclepertin CONNEX, BDNF, C4 Synaptic Pruning & the 2026 UK Review

Around 220,000 people in the UK live with diagnosed schizophrenia under the National Audit of Schizophrenia denominator, and the April 2026 OHID / NHS England prevalence model puts severe mental illness somewhere between 0.31% in the South East and 0.43% in London. Schizophrenia is the prototypical chronic psychotic illness: it almost always begins between the late teens and the early thirties, it carries a mortality gap of roughly 15-20 years compared with the general UK population, and despite seven decades of antipsychotic development the negative symptoms (avolition, social withdrawal, anhedonia, blunted affect) and the cognitive symptoms — collectively cognitive impairment associated with schizophrenia, CIAS — remain largely untreated by anything currently licensed. 2024-2026 has been the most consequential window in schizophrenia drug development in a generation: the FDA approved Bristol Myers Squibb’s KarXT (xanomeline-trospium), now Cobenfy, in September 2024 as the first non-D2 schizophrenia mechanism in roughly 30 years, with the company publicly targeting a 2026 UK MHRA launch at a ~$22,500 list price; Boehringer Ingelheim’s glycine transporter-1 inhibitor iclepertin (BI 425809), the most advanced dedicated CIAS programme in the industry, failed all three pivotal trials in the CONNEX programme in January 2025 across 1,840 patients in 41 countries, and the long-running extension CONNEX-X was discontinued immediately; AbbVie’s emraclidine (acquired with Cerevel for ~$8.7 billion in 2024) missed both EMPOWER Phase 2 trials in late 2024. At the same time, the C4 complement synaptic-pruning hypothesis — Sekar et al. Nature 2016 from the McCarroll lab at the Broad — has matured into a serious mechanistic frame for the cortical synapse loss seen in dorsolateral prefrontal cortex post-mortem tissue, and the 2025 Frontiers in Synaptic Neuroscience review made the case that this excessive C4-driven, microglial-pruning of dendritic spines is now a tractable target. Add the long-established BDNF Val66Met negative-symptom association, the CommonMind Consortium upregulation of HGF in schizophrenia post-mortem brain, and the Met receptor tyrosine kinase / HGF biology of GABAergic interneuron development, and the question that brings traffic to this page becomes legible: where, mechanistically, does a synaptogenic HGF/c-Met peptide like Dihexa sit in a disorder defined by excessive cortical synapse loss? A rigorous 2026 UK evidence review.

Schizophrenia in the UK in 2026: Scale, Demographics & the Treatment Gap

The Royal College of Psychiatrists’ National Clinical Audit of Psychosis (formerly the National Audit of Schizophrenia) puts the diagnosed UK schizophrenia population at around 220,000, with broader psychotic-disorder prevalence considerably higher. The April 2026 update of the OHID / NHS England Severe Mental Illness prevalence model reports schizophrenia point prevalence ranging from 0.31% in the South East to 0.43% in London, with a recognisable urban-deprivation gradient. The Mental Health Foundation summary of UK schizophrenia statistics sits broadly consistent with these numbers. Onset is bimodal but heavily concentrated between 16 and 30 years of age, with men typically presenting earlier than women.

Demographic differences are clinically important. Black African and Black Caribbean adults in the UK have markedly higher rates of first-episode psychosis presentation and compulsory admission — an inequality documented across more than four decades of UK epidemiology, including the Maudsley ÆSOP studies. Migration, urbanicity, social adversity, cannabis use, paternal age, obstetric complications and the C4 complement structural variant are all reasonably evidenced environmental and genetic contributors; the gene-by-environment story is the canonical one.

The treatment gap is structural rather than therapeutic. NHS England’s Mental Health Implementation Plan targets near-universal access to Early Intervention in Psychosis services with the published two-week access standard, but service variation across STPs and ICBs remains substantial, duration of untreated psychosis (DUP) varies more than fivefold between regions, and the negative-symptom and CIAS dimensions of illness are simply not addressed by any currently licensed treatment in the UK in 2026. That gap is what drives search interest in unlicensed synaptogenic peptides — Dihexa, Semax, Selank, NSI-189, cerebrolysin, sarcosine, N-acetylcysteine and adjacent compounds — from people who have run out of NICE-recommended options.

The Neurobiology of Schizophrenia: Dopamine, Glutamate, GABA & Cortical Synapse Loss

Schizophrenia is not one biology. The clinically useful frame is to think of it as three overlapping symptom dimensions — positive (hallucinations, delusions, disorganised thought and behaviour), negative (avolition, social withdrawal, anhedonia, blunted affect, alogia), and cognitive (working memory, attention, processing speed, social cognition) — each with a somewhat different underlying biology, all converging in the same disorder.

The dopamine hypothesis. The original mechanism: striatal D2 receptor hyperactivity drives positive symptoms. Every licensed antipsychotic except Cobenfy is built around D2 antagonism or partial agonism, from chlorpromazine in 1952 to aripiprazole, brexpiprazole and cariprazine today. The reformulated dopamine hypothesis (Howes & Kapur 2009) locates the primary dysfunction in presynaptic striatal dopamine synthesis, with mesocortical hypofunction contributing to negative and cognitive symptoms. This biology genuinely matters: at the population level, D2 antagonists produce reliable and large positive-symptom reductions, with little to no effect on negative or cognitive symptoms. The dimensional dissociation is the defining empirical fact of seven decades of antipsychotic pharmacology.

The NMDA hypofunction / glutamate hypothesis. Phencyclidine (PCP) and ketamine produce a schizophrenia-like syndrome in healthy adults, including negative and cognitive features. NMDA hypofunction on GABAergic parvalbumin-positive (PV+) interneurons disinhibits glutamatergic pyramidal output and produces the gamma-band cortical dysrhythmia seen on MEG/EEG in schizophrenia. This is the mechanistic rationale behind GlyT1 inhibition (iclepertin, bitopertin), D-serine adjuncts, sarcosine, N-acetylcysteine, NMDA modulators and the glycine-site agonist class — and is the same biology that makes ketamine paradoxically useful in treatment-resistant depression but psychotomimetic in healthy controls.

The GABAergic interneuron / parvalbumin biology. Post-mortem and imaging work consistently shows reduced parvalbumin interneuron density and reduced GAD67 expression in dorsolateral prefrontal cortex (Lewis, Glausier and others, over more than two decades). Cortical PV+ interneurons depend on the Met receptor tyrosine kinase for normal development — the same Met that Dihexa engages indirectly via stabilising hepatocyte growth factor. This is the cleanest single mechanistic link between Dihexa biology and schizophrenia biology, and we return to it below.

The synaptic pruning hypothesis (Feinberg, Sekar, Stevens, McCarroll). Healthy adolescent cortical development includes substantial activity-dependent pruning of dendritic spines, peaking in the prefrontal cortex in the late teens and early twenties — exactly the window where schizophrenia typically begins. Sekar et al. Nature 2016 showed that the schizophrenia GWAS signal at the MHC locus is driven by structural variants of complement component 4 (C4A in particular), that higher C4A copy number and expression confer schizophrenia risk, and that C4 directs microglial-mediated synapse elimination via the C1q/C3 cascade in mouse models. The 2025 Frontiers in Synaptic Neuroscience review (Is it possible to prevent excessive synaptic pruning in schizophrenia?) extended the picture: C4 overexpression also disrupts AMPA receptor (GluR1) trafficking, suggesting non-classical, non-microglial routes to synaptic dysfunction. Emerging therapeutic strategies discussed in that review include tetracycline antibiotics (minocycline, doxycycline), complement-pathway modulation, kynurenine modulation, BDNF / progranulin neuroprotection, and epigenetic approaches.

This last point is the one that matters for Dihexa. If the dominant cellular pathology in cortical schizophrenia is excessive synapse elimination, then a synaptogenic agent that promotes new spine formation in an already-pruned network is a non-trivial intervention — potentially restorative or potentially destabilising, depending on which neurons it preferentially affects and at what disease stage. There is no published data that resolves this question for Dihexa.

Where Does HGF/c-Met Sit in Schizophrenia Biology?

The case is more complete than most readers expect. The 2021 Frontiers in Cell and Developmental Biology review by Desbonnet and colleagues (HGF and MET: From Brain Development to Neurological Disorders) anchors the schizophrenia section in five threads:

1. Met-receptor SNPs associate with schizophrenia in Caucasian cohorts. The signal is modest and not consistently replicated across all ancestries, but it is real on meta-analysis.
2. The CommonMind Consortium reported HGF mRNA upregulation in schizophrenia post-mortem dorsolateral prefrontal cortex versus controls.
3. The Met receptor is required for GABAergic interneuron development. Conditional Met knockouts in mouse interneuron precursors produce GABAergic phenotypes reminiscent of schizophrenia post-mortem findings.
4. HGF promotes midbrain dopaminergic neuron survival and maturation — the same neurons whose A10/A9 projections to ventral striatum and prefrontal cortex are central to the dopamine hypothesis. The directionality of effect on circuit-level dopaminergic tone is non-trivial.
5. HGF / c-Met regulates microglial phenotype, and microglial dysfunction is now the leading cellular candidate for excessive synaptic pruning in adolescent-onset psychosis.

These five threads do not add up to “Dihexa treats schizophrenia”. They add up to: HGF/c-Met biology is genuinely involved in the schizophrenia phenotype, the involvement is most likely developmental and most easily framed at the level of GABAergic interneurons and microglial pruning, and the polarity of pharmacological intervention is unresolved. If the schizophrenia phenotype includes a partial Met loss-of-function in interneuron development, then potentiating Met signalling later in life is mechanistically defensible. If the schizophrenia phenotype includes Met over-activity in microglia driving excessive pruning, then potentiating Met signalling is the opposite of what you want. Without target-engagement data in schizophrenia patients, you cannot tell.

This is why the Dihexa vs BDNF and Mechanism of Action pages should be read together with this article: the synaptogenic story is more nuanced in psychotic disorders than it is in the neurodegenerative indications the rest of this site covers, including Alzheimer’s, Parkinson’s, vascular dementia, Lewy body dementia, FTD and Huntington’s disease.

BDNF, Val66Met & the Negative-Symptom Signal

BDNF, the canonical synaptic-plasticity neurotrophin, is reduced in serum and post-mortem brain in schizophrenia on most (though not all) meta-analyses. The cleanest mechanistic finding is the Val66Met polymorphism in the BDNF gene: the Met allele disrupts activity-dependent secretion of BDNF and is associated, reproducibly, with greater severity of negative symptoms in first-episode and early-onset schizophrenia spectrum disorders. Mas et al. 2021 showed the Met allele dose-dependent association in early-onset schizophrenia spectrum and other psychotic disorders. The 2021 mediation analysis connected age of onset, BDNF level, negative symptom severity and cognition in chronic schizophrenia.

This is the same BDNF axis that the Dihexa vs BDNF page covers in mechanistic detail and that the depression review, PTSD review and anxiety review all touch on. The point relevant for this page: the BDNF connection to schizophrenia is real, the connection is strongest for negative rather than positive symptoms, and the Dihexa preclinical literature loads on synaptogenesis and BDNF-adjacent biology rather than dopaminergic positive-symptom modulation. If there is a plausible Dihexa target in schizophrenia, on the current evidence, it is negative-symptom and CIAS biology rather than positive-symptom biology — which would put it in adjunctive territory, not antipsychotic territory.

NICE CG178, NG191 & the Current UK Standard of Care for Schizophrenia in 2026

NICE CG178 (Psychosis and schizophrenia in adults: prevention and management) remains the operative guidance, with the CG155 pathway for children and young people and NG191 for medicines optimisation and physical-health monitoring. The first-line offer is an oral antipsychotic, individual cognitive behavioural therapy for psychosis (CBTp) and family intervention, with attention to physical health, smoking cessation and metabolic risk — the latter taking on increasing importance as antipsychotic-driven weight gain, dyslipidaemia and type 2 diabetes drive the 15-20 year mortality gap.

Oral options include risperidone, paliperidone, olanzapine, aripiprazole, quetiapine, amisulpride, lurasidone and cariprazine, with long-acting injectables (paliperidone palmitate 1-monthly, 3-monthly and the newer 6-monthly Hafyera; aripiprazole monohydrate Maintena and Abilify Asimtufii; risperidone Perseris) for non-adherence and relapse prevention. Clozapine is the only licensed treatment for genuinely treatment-resistant schizophrenia (failure of two adequate antipsychotic trials), with the well-known agranulocytosis monitoring burden via the clozapine patient monitoring services.

Notably absent from CG178: any pharmacotherapy for negative symptoms or CIAS. Cariprazine carries a partial negative-symptom indication on the European label, lurasidone has cognitive signals on MATRICS, and cognitive remediation therapy (including the Wykes/King’s College London CIRCuiTS programme) sits in the NICE evidence base. But the practical reality on an NHS EIP service in 2026 is that negative symptoms and cognitive impairment remain the dominant drivers of disability after the first positive episode is controlled, and the licensed pharmacopoeia does very little for either.

Beyond CG178: The 2024-2026 Schizophrenia Pipeline

Three drugs and three failures define the current pipeline.

KarXT / Cobenfy (xanomeline-trospium) — Bristol Myers Squibb / Karuna

KarXT pairs xanomeline (an M1/M4 muscarinic acetylcholine receptor agonist originally developed by Eli Lilly in the 1990s for Alzheimer’s disease) with trospium (a peripherally-restricted muscarinic antagonist) to neutralise the cholinergic gastrointestinal and cardiovascular side effects that doomed earlier xanomeline development. Karuna Therapeutics ran the EMERGENT-1, -2 and -3 Phase 3 monotherapy trials in acutely psychotic adults with schizophrenia, with PANSS reductions of 9.6-11.6 points versus placebo and no clinically meaningful EPS, weight gain or metabolic effects — a clean separation from the D2-blocker class. The FDA approval came in September 2024, branded as Cobenfy. Bristol Myers Squibb acquired Karuna for $14 billion in March 2024, and has publicly committed to a 2026 UK MHRA submission and launch at a ~$22,500 annual list price, with the NICE health-technology appraisal expected to follow.

The 2025-2026 reads on Cobenfy have been mixed. The ARISE adjunctive Phase 3 trial in patients with inadequately controlled symptoms on existing atypical antipsychotics did not meet its primary PANSS endpoint at Week 6 — though signals on negative symptoms and tolerability were reported. The extension study of long-term safety and tolerability of adjunctive KarXT reached primary completion in March 2026. Phase 4 commitments include adolescent schizophrenia, bipolar I mania (covered in the Dihexa for Bipolar Disorder 2026 UK review), and the ADEPT-2 Alzheimer’s psychosis programme. The UK price point and the negative ARISE result will both feature heavily in the NICE appraisal.

Iclepertin (BI 425809) — Boehringer Ingelheim (CONNEX failure)

Iclepertin was the most advanced dedicated CIAS programme in the industry. A potent and selective glycine transporter-1 (GlyT1) inhibitor, it raised synaptic glycine and engaged the glycine modulatory site on NMDA receptors, with a Phase 2 proof-of-concept signal at 10 mg and 25 mg that supported the move into Phase 3. The CONNEX programme was a three-trial, double-blind, placebo-controlled programme in 1,840 adults with schizophrenia across 41 countries, 26 weeks of treatment, with MCCB composite score as the primary endpoint and the Schizophrenia Cognition Rating Scale as a secondary functional endpoint.

In January 2025 Boehringer announced that no statistically significant effect on cognition or functioning was observed at six months in any of the three CONNEX trials. The drug was generally well tolerated, with no safety surprises. The long-term extension trial, CONNEX-X, was discontinued effective immediately. The failure mirrored the earlier bitopertin GlyT1 failure at Roche and effectively ended GlyT1 inhibition as a near-term commercial CIAS strategy. The CIAS treatment gap remains unfilled.

Emraclidine — Cerevel / AbbVie (EMPOWER failure)

Emraclidine was an M4-selective muscarinic positive allosteric modulator — the mechanism that, in principle, should have been cleaner than the M1/M4-balanced xanomeline. AbbVie acquired Cerevel Therapeutics for ~$8.7 billion in 2024 in large part for this asset. The EMPOWER-1 and EMPOWER-2 Phase 2 trials in acute schizophrenia missed their primary PANSS endpoints in late 2024. The implications are still being digested: it may be that the M1 contribution in xanomeline is essential, or that the dose ranging in EMPOWER was wrong, or that the muscarinic schizophrenia story is more idiosyncratic than the KarXT data alone suggested.

Ulotaront (SEP-363856) — TAAR1 (DIAMOND failure)

Ulotaront was a non-D2 TAAR1 (trace amine-associated receptor 1) agonist with serotonergic activity that produced a clean Phase 2 signal (Koblan et al. NEJM 2020). Sumitomo / Sunovion ran the DIAMOND-1 and DIAMOND-2 Phase 3 trials in acute schizophrenia; both missed primary endpoints in 2023, the programme was discontinued, and Sumitomo took the goodwill write-down. TAAR1 remains an academically interesting target; commercially it is largely off the table.

Other 2024-2026 reads

Roluperidone (Minerva Neurosciences) for negative symptoms received an FDA Complete Response Letter in 2024 on labelling and study population grounds. Evenamide (Newron Pharmaceuticals) continues in treatment-resistant schizophrenia. RL-007 (Recognify / atai Life Sciences) is in earlier-stage CIAS development. Low-dose lithium augmentation, minocycline (anti-inflammatory), sarcosine (glycine site), N-acetylcysteine and pridopidine all have small-trial signals and no convincing Phase 3 data.

2026 News Context: What Has Just Changed in Schizophrenia

Three news stories matter for any 2026 UK schizophrenia readership.

1. Cobenfy UK launch positioning. Bristol Myers Squibb publicly confirmed a planned 2026 UK MHRA submission and launch of Cobenfy at a ~$22,500 annual list price — an order of magnitude above the generic atypical-antipsychotic baseline used by NHS commissioners. The NICE appraisal will need to weigh the metabolic-side-effect avoidance, the negative-symptom signal, and the ARISE adjunctive miss. A first MHRA marketing authorisation under a fast-track pathway is plausible during 2026; NICE-recommended NHS access is more likely 2027.

2. The synaptic-pruning hypothesis went mainstream. The 2025 Frontiers in Synaptic Neuroscience review (Is it possible to prevent excessive synaptic pruning in schizophrenia?) consolidated the C4-complement-microglia-spine-elimination frame into a tractable therapeutic agenda, including the surprising AMPA receptor (GluR1) trafficking finding that suggests non-classical, non-microglial routes. The clinical-trial implications are substantial: cohorts at high C4A copy-number / expression may be the natural enrolment target for any complement-modulating or anti-pruning intervention in the prodromal / early-illness window.

3. The CIAS treatment gap remained unfilled. The iclepertin CONNEX failure, the emraclidine EMPOWER failure, the ulotaront DIAMOND failure, and the ARISE adjunctive Cobenfy miss collectively mean there is still no licensed UK treatment for cognitive impairment associated with schizophrenia in 2026. Cognitive remediation therapy (CIRCuiTS, Wykes and colleagues, King’s College London) remains the best-evidenced non-pharmacological CIAS intervention and is increasingly available across NHS EIP services. Pharmacology has not caught up.

Dihexa Specifically in Schizophrenia: What the Evidence Actually Is

This is the section that does most of the cognitive work on this page. The structure mirrors the rest of the rolling 2026 series.

What Dihexa is. Dihexa (PNB-0408, N-hexanoic-Tyr-Ile-(6) aminohexanoic amide) is a small-molecule angiotensin IV analogue developed in the laboratories of Joseph Harding and John Wright at Washington State University. The Benoist 2014 JPET paper (The procognitive and synaptogenic effects of angiotensin IV-derived peptides are dependent on activation of the HGF/c-Met system) is the canonical mechanistic paper. Dihexa is reported to stabilise hepatocyte growth factor (HGF) and potentiate signalling through the Met receptor tyrosine kinase, with downstream MAPK/ERK and PI3K/Akt effects, synaptogenesis on cortical and hippocampal neurons, and reversal of scopolamine-induced spatial memory deficits in the Morris water maze.

What Dihexa is not, in 2026. Dihexa is not a licensed medicine anywhere in the world for any indication. There is no Investigational New Drug application, no published Phase 1, no published clinical-trial registration, and no published human PK or safety data for Dihexa itself in any therapeutic area, let alone schizophrenia. The clinical analogue in the same chemical family is fosgonimeton (ATH-1017) from Athira Pharma, which has been studied in Alzheimer’s disease (LIFT-AD), dementia with Lewy bodies (SHAPE) and Parkinson’s disease dementia — none of those programmes has produced positive Phase 2/3 readouts to date, and none has been studied in schizophrenia.

What the preclinical literature actually covers. The published Dihexa animal literature is dominated by Alzheimer’s-relevant cognitive paradigms (spatial water maze, passive avoidance, scopolamine reversal), Parkinson’s-relevant 6-OHDA models, and angiotensin IV / IRAP receptor pharmacology. There is no published Dihexa work in PCP, MK-801, ketamine, NMDA-hypofunction, amphetamine-sensitisation, social-defeat, neonatal-ventral-hippocampal-lesion, methylazoxymethanol (MAM), poly-I:C maternal immune activation, or any of the canonical schizophrenia preclinical models.

What the inference chain therefore actually is. To get from current Dihexa evidence to “Dihexa for schizophrenia” you must chain together: angiotensin IV peptides → HGF/c-Met activation → synaptogenesis (Benoist 2014) + Met required for GABAergic interneuron development (Frontiers 2021) + Met SNP / HGF mRNA upregulation in schizophrenia (CommonMind, Caucasian GWAS) + BDNF Val66Met negative-symptom association + cortical synapse loss documented by Sekar / Stevens / McCarroll. Each link is real. The composite is suggestive. No link is a controlled clinical trial of Dihexa in any psychotic disorder.

The polarity problem. Even granting every link, the directionality remains unresolved. Schizophrenia cortex shows reduced dendritic spine density — i.e. evidence of excess elimination of synapses. A synaptogenic intervention could in principle restore lost connections (a good thing) or amplify existing aberrant connections (a bad thing, theoretically destabilising in someone with active positive symptoms). The risk-benefit calculation in schizophrenia is genuinely different from the calculation in Alzheimer’s, vascular dementia or post-stroke recovery, where the dominant pathology is gross loss of structure rather than aberrant retained connectivity. This is why the Dihexa vs nootropics page, the side effects page and the BDNF page all emphasise that synaptogenic agents in psychotic illness are mechanistically non-trivial.

Safety: Antipsychotic Polypharmacy, QT, Clozapine & Theoretical Pro-Psychotic Risk

Practical safety considerations for any reader with schizophrenia, schizoaffective disorder or first-episode psychosis who is contemplating Dihexa fall into four categories.

1. Drug-drug interactions with antipsychotics. Most atypical antipsychotics are CYP1A2, CYP2D6 and CYP3A4 substrates; clozapine, olanzapine and quetiapine in particular have narrow therapeutic windows. There is no published Dihexa-antipsychotic drug-drug interaction data. Self-administered Dihexa alongside clozapine is contraindicated by absence of evidence.

2. QT prolongation. Several atypical antipsychotics (ziprasidone, iloperidone, sertindole, clozapine, paliperidone at high dose) prolong the QT interval. The Dihexa cardiovascular safety profile is essentially uncharacterised in humans. Combining unknown cardiovascular pharmacology with known QT-prolonging psychotropics is not a sensible plan.

3. Theoretical pro-psychotic risk. Several lines of indirect evidence raise the possibility that strongly synaptogenic agents could destabilise active psychosis. The PCP / ketamine models, the cannabis-precipitation literature, the BDNF / cortical-circuit reorganisation literature, and the post-mortem evidence of aberrantly retained rather than missing synapses in some cortical regions all converge on caution. There is no published evidence that Dihexa specifically precipitates or worsens psychosis. There is also no published evidence that it does not. The honest scientific position is uncertainty, and the honest clinical position is “not while symptomatic”.

4. Clozapine specifically. Patients on clozapine for treatment-resistant schizophrenia have, by definition, exhausted the first-line and second-line options. They are also the cohort with the most to lose from any destabilisation of a working regimen. Clozapine monitoring (mandatory FBC at week 0 / weekly to 18 weeks / 2-weekly to 1 year / 4-weekly thereafter via the clozapine patient monitoring service) reflects the agranulocytosis risk. Combining clozapine with any unlicensed peptide in this regulatory environment is not a sensible plan.

See also the Dihexa side effects & risks page and the dosage discussion — both written for the general nootropic / cognitive-enhancement reader rather than for patients with psychotic disorders, where the risk-benefit calculation is meaningfully different.

First-Episode Psychosis, EIP & the UK Prodromal Pathway

The NHS England Early Intervention in Psychosis access and waiting-time standard requires that ≥60% of new EIP referrals (ages 14-65, soon extending to older adults) start NICE-concordant treatment within two weeks of referral. The 2023 implementation guidance frames EIP around antipsychotic medication, CBTp, family intervention, vocational and educational support (including the Individual Placement and Support model rolled out by NHS England), and physical-health monitoring per NG191. EIP services are not free-standing: they are integrated with adult community mental-health services and crisis pathways. The headline national-audit metric is fidelity to the NICE-concordant care bundle, not antipsychotic prescribing alone.

The prodromal / clinical high risk for psychosis (CHR-P, formerly ultra-high-risk, UHR) population is served in a more uneven way across the UK. The flagship UK service is OASIS at the Maudsley (Outreach and Support in South London), with sister services such as Cambridge CAMEO and the EIP+CHR-P pathways in Manchester, Birmingham and Glasgow. International CHR-P guidelines (EPA, IEPA) do not recommend antipsychotic prophylaxis in unaffected CHR-P young people: the offer is monitoring, CBT, family intervention and treatment of comorbid depression, anxiety and substance use. Roughly 20-30% of CHR-P individuals transition to a first psychotic episode within two to three years; that transition rate has fallen across cohorts in recent decades for reasons that include broader screening and lead-time bias.

Critical safety point: the prodromal / CHR-P window is exactly the cohort where naive self-experimentation with unlicensed synaptogenic agents is theoretically most concerning. The cortex is mid-pruning, the genetic risk is concentrated, and the conversion-rate base rate is non-trivial. We are not aware of any published evidence that Dihexa precipitates psychosis. We are also not aware of any published evidence that it does not. In a population with a 20-30% three-year base-rate transition to a first psychotic episode, “we don’t know” is not a reassuring answer.

Practical Realities: What Actually Works in 2026

For a UK adult with diagnosed schizophrenia in 2026, the evidence-based offer reads:

1. NICE CG178 first-line: Oral atypical antipsychotic + CBT for psychosis + family intervention + physical-health monitoring per NG191. Most patients respond to a first or second adequate antipsychotic trial.
2. Long-acting injectable second-line for non-adherence: Paliperidone palmitate (1-monthly, 3-monthly, 6-monthly Hafyera), aripiprazole monohydrate (Maintena, Asimtufii), risperidone Perseris.
3. Clozapine for treatment resistance: The only licensed option after two adequate antipsychotic trials, with mandatory FBC monitoring.
4. Cognitive remediation therapy for CIAS: CIRCuiTS (Wykes, KCL) and equivalent computerised programmes; the best-evidenced non-pharmacological CIAS intervention.
5. Negative-symptom adjuncts (off-label): Cariprazine has a partial European negative-symptom indication; aripiprazole augmentation, amisulpride low-dose, NAC adjuncts and minocycline have small-trial signals.
6. Cobenfy (xanomeline-trospium): Watch for the MHRA decision in H2 2026 and the NICE appraisal that will follow. NHS access is more likely 2027 than 2026.
7. Clinical-trial enrolment for CIAS, negative symptoms and treatment resistance: Via NIHR Be Part of Research. The CIAS gap is genuine and trials are how it gets closed.
8. Vocational and educational support: NHS Individual Placement and Support, the SU vocational programmes, recovery colleges. The disability burden in schizophrenia is dominated by negative symptoms and CIAS rather than positive symptoms, and IPS evidence is robust.
9. Carer and family support: Rethink Mental Illness, Mind, Hearing Voices Network, family-intervention services within EIP and adult community mental-health teams.

Dihexa does not enter this list. Not because it has been tried and failed in schizophrenia — it has never been tried — but because the licensed evidence-based offer is more developed than the unlicensed peptide option in every dimension that matters.

What Would Need to Happen for Dihexa to Become a Real Schizophrenia Option?

The same checklist used on the Huntington’s, PTSD, FTD and vascular dementia pages applies here, with two schizophrenia-specific additions.

1. Published preclinical work in canonical schizophrenia models — PCP / MK-801 / ketamine sensitisation, neonatal-ventral-hippocampal-lesion, MAM, poly-I:C, social-defeat. None published to date.
2. Published preclinical work on C4 / microglial pruning interaction — does Dihexa rescue, exacerbate, or leave unchanged the C4-driven spine-loss phenotype in transgenic models? Open question.
3. Phase 1 healthy-volunteer PK/PD data — Dihexa has none. Fosgonimeton has Phase 1 PK from the Athira programme.
4. Phase 2 proof-of-concept in negative symptoms or CIAS, not in acute positive symptoms — the mechanistic case is for the residual / negative / cognitive dimensions, not for acute psychotic episodes.
5. Phase 3 with a recognised MCCB or NSA endpoint and an antipsychotic-adjunctive design — replacing antipsychotics on the basis of a synaptogenic peptide is not a defensible 2030 design.
6. MHRA and NICE pathway, with carer-involvement standards, before NHS uptake.

None of the above exists for Dihexa in schizophrenia in 2026. By contrast, all of the above (or its equivalent in dementia) did exist for Cobenfy by September 2024, which is why Cobenfy is approved and Dihexa is not.

The Bottom Line: A Plausible Mechanism, a Severe Disorder, and Zero Human Dihexa Data

Schizophrenia is a real and treatable condition. The cortical-synapse-loss biology and the GABAergic interneuron / HGF/c-Met biology genuinely overlap with Dihexa’s claimed synaptogenic mechanism. The BDNF Val66Met negative-symptom association is reproducible. The C4 complement synaptic-pruning hypothesis has matured into a tractable therapeutic frame. And yet none of this is a controlled clinical trial of Dihexa in any psychotic disorder, the directionality of any pharmacological intervention is unresolved because schizophrenia is a disorder of excessive synapse elimination, and the case for caution in symptomatic patients (and especially in the prodromal CHR-P window) is materially stronger than in any of the neurodegenerative indications elsewhere on this site.

The 2026 honest answer for someone living with schizophrenia in the UK is: NICE CG178 antipsychotic + CBT for psychosis + family intervention first; clozapine for treatment-resistance under monitoring; Cobenfy from H2 2026 if MHRA-licensed and NICE-recommended; CIRCuiTS / cognitive remediation for CIAS; vocational and educational support via Individual Placement and Support; clinical-trial enrolment for CIAS and negative symptoms via NIHR Be Part of Research; and Rethink, Mind, the Hearing Voices Network, and EIP family interventions for carers. Unlicensed synaptogenic peptides like Dihexa belong nowhere in that hierarchy in 2026.

Frequently Asked Questions: Dihexa & Schizophrenia

Can Dihexa help schizophrenia or psychosis?

There is no published controlled clinical trial of Dihexa in schizophrenia, schizoaffective disorder, first-episode psychosis, treatment-resistant schizophrenia, CIAS or prodromal psychosis. The mechanistic case via HGF/c-Met, GABAergic interneuron development, BDNF biology and the synaptic-pruning hypothesis is real but inferential. Self-experimenting with an unlicensed research peptide while acutely symptomatic, or in the CHR-P window, is not a sensible plan.

What is KarXT / Cobenfy and when does it reach the NHS?

Cobenfy is xanomeline-trospium, an M1/M4 muscarinic agonist + peripherally-restricted muscarinic antagonist combination. FDA-approved September 2024 as the first non-D2 schizophrenia mechanism in roughly 30 years. Bristol Myers Squibb has publicly committed to a 2026 UK MHRA submission and launch at a ~$22,500 annual list price. NICE appraisal will follow. NHS NICE-recommended access is more likely 2027 than 2026.

What happened with iclepertin and CIAS in 2025?

Boehringer Ingelheim’s glycine transporter-1 inhibitor BI 425809 missed primary and secondary endpoints in all three CONNEX Phase 3 trials in January 2025 across 1,840 patients in 41 countries. CONNEX-X long-term extension was discontinued immediately. The CIAS treatment gap remains unfilled and the GlyT1 mechanism is largely off the commercial table.

What does NICE CG178 say about schizophrenia treatment?

CG178 places oral atypical antipsychotic + CBT for psychosis + family intervention + physical-health monitoring (per NG191) as first-line care, with long-acting injectables for non-adherence and clozapine for genuinely treatment-resistant schizophrenia after two failed adequate antipsychotic trials. Dihexa is not mentioned anywhere in NG178, NG191 or CG155.

Is Dihexa contraindicated in schizophrenia?

Dihexa is not a licensed medicine for any indication, so “contraindicated” in the regulatory sense is the wrong frame. Pragmatically: Dihexa has no published human safety data, no drug-drug interaction data with antipsychotics, no QT data, no data in psychotic illness, and a mechanism (synaptogenesis via HGF/c-Met) whose interaction with the cortical-synapse-loss biology of schizophrenia is theoretically non-trivial. The honest answer is that it should not be used in schizophrenia outside a controlled clinical trial.

Where can I find UK schizophrenia support?

Your GP and your local NHS Early Intervention in Psychosis service for first-episode psychosis (ages 14-65). Rethink Mental Illness, Mind and the Hearing Voices Network for peer and carer support. OASIS Maudsley for South-London prodromal pathway. Samaritans 24-hour helpline 116 123 in crisis; 999 for immediate danger. The NIHR Be Part of Research portal for clinical-trial enrolment.

Sources & Citations

• NICE CG178. Psychosis and schizophrenia in adults: prevention and management.
• NICE CG155. Psychosis and schizophrenia in children and young people: recognition and management.
• NICE NG191. Antipsychotic medicines optimisation.
• FDA approves Cobenfy (xanomeline-trospium) for adults with schizophrenia (September 2024).
• Bristol Myers Squibb Plans 2026 UK Launch of Breakthrough Schizophrenia Drug Cobenfy at $22,500 Annual Price.
• Cobenfy Did Not Meet Primary Outcome as Add-On Treatment for Schizophrenia (ARISE adjunctive Phase 3).
• An Extension Study to Assess Long-Term Safety and Tolerability of Adjunctive KarXT in Inadequately Controlled Schizophrenia (primary completion March 2026).
• Boehringer provides update on iclepertin Phase III programme in schizophrenia (CONNEX failure, January 2025).
• Boehringer’s Phase 3 schizophrenia programme misses primary goal — Fierce Biotech.
• Fleischhacker WW et al. Efficacy and safety of iclepertin (BI 425809) in patients with schizophrenia: CONNEX, a Phase III randomised controlled trial programme. CNS Spectrums.
• Efficacy and safety of iclepertin for cognitive impairment associated with schizophrenia (CONNEX programme): results from three Phase 3 randomised controlled trials. Lancet Psychiatry (2025).
• AbbVie’s emraclidine fails pair of Phase 2 schizophrenia trials in blow to $8.7B Cerevel buyout (2024).
• Sumitomo pulls plug on ulotaront after Phase 3 failures in schizophrenia and depression (DIAMOND, 2023).
• Sekar A et al. Schizophrenia risk from complex variation of complement component 4. Nature (2016).
• Is it possible to prevent excessive synaptic pruning in schizophrenia? Possibilities and limitations. Frontiers in Synaptic Neuroscience (2025).
• Overexpression of complement component C4 in DLPFC, parietal cortex, superior temporal gyrus and associative striatum of patients with schizophrenia. Brain, Behavior, and Immunity.
• Multicausal disruption of complement system activity in schizophrenia: C4, complement control proteins and microglia-specific genes in brain and blood.
• Desbonnet et al. HGF and MET: From Brain Development to Neurological Disorders. Frontiers in Cell and Developmental Biology (2021).
• 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).
• BDNF Val66Met polymorphism is associated with negative symptoms in early-onset schizophrenia spectrum and other psychotic disorders.
• Association of BDNF Val66Met polymorphism with negative symptom severity in first-episode schizophrenia spectrum disorders. European Psychiatry.
• BDNF affects the mediating effect of negative symptoms on the relationship between age of onset and cognition in chronic schizophrenia.
• Royal College of Psychiatrists — National Clinical Audit of Psychosis.
• OHID / NHS England — Estimated prevalence of severe mental illness in England (April 2026 update).
• NHS England — Implementing the Early Intervention in Psychosis access and waiting-time standard.
• Mental Health Foundation — UK schizophrenia statistics.
• OASIS — Outreach and Support in South London (Maudsley CHR-P prodromal psychosis service).
• CIRCuiTS — Wykes et al. cognitive remediation therapy for schizophrenia (King’s College London).
• Rethink Mental Illness.
• Mind.
• Hearing Voices Network UK.
• Clozapine Patient Monitoring Service UK.
• NIHR Be Part of Research portal.