# Schizophrenia

Genetic, developmental and social contributions to schizophrenia, the role and limits of dopamine explanations, and treatments for psychosis, relapse prevention and everyday recovery.

Information, not medical advice. Discuss personal concerns with a qualified health professional.

[Watch the film](https://www.youtube.com/watch?v=Zli0mwlxSdc) · [All films and essays](https://evidencepress.org/why-minds-break/films/)

Schizophrenia can change how a person understands what is happening around them. They may hear a voice that other people cannot hear, become convinced that ordinary events carry a threatening personal message, or find their thoughts difficult to organise. Some people also struggle with concentration, memory, motivation or expressing emotion. These difficulties do not all have the same explanation, and they do not all respond equally well to the same treatment.

The broad scientific picture is of several interacting influences: inherited susceptibility, development, experiences and circumstances, and changes in the way brain cells communicate. There is strong evidence for parts of this picture. There is much less certainty about the sequence that connects them in any one person. Researchers have not found a single cause, scan or blood test that explains or identifies every case. [Schizophrenia: a broad scientific review](https://doi.org/10.1038/s41572-025-00667-6)

An important distinction concerns *psychosis*: experiences such as hallucinations and strongly held beliefs that are not shared by others. Psychosis can occur in conditions other than schizophrenia. A study of people having their first psychotic episode may therefore tell us something important without establishing a finding specific to schizophrenia. The treatment research discussed here mainly concerns adults; pregnancy, older age, serious physical illness and complex substance use need additional clinical assessment.

## Inherited susceptibility is substantial, but it is not destiny

Schizophrenia runs in families, and genetic research shows that inherited differences contribute substantially to susceptibility. A Danish study of 31,524 twin pairs estimated that inherited differences accounted for about 79% of variation in susceptibility in that population. Yet when one identical twin had schizophrenia, the other also had it in about 33% of pairs; the corresponding figure for non-identical twins was about 7%. Strong genetic influence clearly coexisted with considerable variation in what happened to people. [The Danish twin study](https://doi.org/10.1016/j.biopsych.2017.08.017)

The 79% estimate does not mean that 79% of one person's illness is genetic. Nor does it mean that 79% of cases are inherited, or that changing circumstances can help only the remaining 21%. It describes differences between people under the conditions and assumptions of a particular study. The twin figures and the estimate answer different questions.

Most inherited susceptibility comes from many common genetic differences, each having a small effect. Larger studies are identifying more of them. A 2026 study including people from a wider range of ancestries added more than 100 implicated regions of the genome. Much of the underlying genetic pattern was shared across populations. However, earlier research concentrated heavily on people of European ancestry, limiting how reliably its findings can be applied elsewhere. [Research across ancestrally diverse populations](https://doi.org/10.1038/s41586-025-10000-6)

Some very rare genetic changes have larger associations. A study comparing 24,248 people with schizophrenia with 97,322 controls identified damaging variants in ten genes. Their reported associations ranged from roughly threefold to fiftyfold differences in the odds of schizophrenia. These are uncommon changes, and large relative differences do not make illness certain. Several implicated genes affect the connections through which nerve cells communicate, including receptors that respond to the chemical messenger glutamate. They offer biological clues, not a single “schizophrenia gene”. [The rare genetic variant study](https://doi.org/10.1038/s41586-022-04556-w)

Development is also important. Some vulnerabilities are present long before a first psychotic episode. But development does not mean only events in the womb or early childhood. Genetic findings concern processes running from fetal life through to the continuing work of adult brain connections, across different cell types and regions. There is no established account in which one early injury to one brain circuit explains every case. [What genomic research changes about developmental explanations](https://doi.org/10.1038/s41380-026-03654-9)

An early vulnerability does not guarantee later illness. Nor does a developmental explanation mean that everyone will progressively deteriorate. People have different courses of illness and different opportunities for recovery. Some genetic findings also overlap with autism, epilepsy and other developmental conditions; shared biological influences can lead to different outcomes. [Clinical review of schizophrenia and its course](https://doi.org/10.1016/S0140-6736%2821%2901730-X), [Shared and distinct genetic findings](https://doi.org/10.1038/s41586-022-04556-w)

## Experiences and circumstances matter, with different degrees of certainty

Research links schizophrenia and other psychotic disorders with a range of pregnancy and birth circumstances, childhood adversity, cannabis exposure and social disadvantage. These findings deserve attention. They also need careful interpretation: circumstances occur together, inherited influences can affect exposure, and early illness can itself change behaviour or social conditions.

A 2026 review brought together 69 reviews covering 970 studies and 221 possible risk or protective factors for psychotic disorders. It found credible associations in several areas, although some became weaker when researchers used more cautious analyses or restricted attention to studies that followed people forward in time. A strong association is not automatically a measurement of how much one factor causes illness. [The review of risk and protective factors](https://doi.org/10.1002/wps.70093)

Low birthweight provides an example. Birthweight below 2.5 kilograms was associated with about one-and-a-half times the odds of later psychotic disorder. That could reflect the effects of restricted growth, an underlying pregnancy-related process, or other connected influences. The observation supports attention to development; it does not identify the causal pathway by itself.

Childhood adversity, meaning serious harmful experiences during childhood, is another important example. A review of 183 samples, involving 349,265 people, found about 2.8 times the odds of psychosis among those who had experienced adversity. The association appeared across different study designs, but results varied considerably. The main analysis favoured estimates before adjustment for other influences; adjusted estimates were generally smaller. Some evidence also suggested that published studies might overstate the association. Adversity is an important warning sign for later difficulties and probably contributes to some pathways into psychosis, but this pooled number is not its isolated causal effect. [Four decades of research on childhood adversity and psychosis](https://doi.org/10.1176/appi.ajp.20240456)

Cannabis research raises particular concern about frequent use of high-potency products. In the EU-GEI study, daily use was associated with about 3.2 times the odds of a first psychotic disorder compared with never using cannabis. Daily use of high-potency cannabis was associated with about 4.8 times the odds. The study compared 901 people with first-episode psychosis with 1,237 controls. The pattern across frequency and potency strengthens the case for a contributory causal effect, but people were not randomly assigned to cannabis use. Differences between groups, selection into the study and recall of past use remain possible influences. [The EU-GEI cannabis study](https://doi.org/10.1016/S2215-0366%2819%2930048-3)

Family comparisons show why the qualification matters. In a Swedish study, the association between registered cannabis abuse and later schizophrenia fell from an odds ratio of 4.24 in the population comparison to 1.98 when full siblings were compared. At least seven years separated registered abuse from diagnosis. Shared family influences explained part of the association; the remainder was compatible with cannabis contributing to risk, although differences between siblings could still matter. [The Swedish family-comparison study](https://doi.org/10.1017/S0033291714001524)

These are relative comparisons, not a forecast for a particular person. As a purely mathematical illustration, multiplying the odds of a 1% baseline risk by five produces a risk of about 4.8%, not certainty. That is not an estimate for an individual cannabis user: a useful estimate would need a suitable starting risk, timescale and evidence that the study applies to them. Nor can these studies tell us precisely how many cases would be prevented by reducing cannabis exposure or adversity.

Migration, being a minority within particular social settings, and living among fewer people from one's own group are also associated with psychosis. Exclusion, discrimination and access to support are plausible contributors. Differences in diagnosis and access to services may matter too. These findings concern social context; they do not establish ethnicity as an inherited biological cause. [Social and developmental risk evidence](https://doi.org/10.1002/wps.70093)

## What dopamine can explain

Dopamine is a chemical messenger used by nerve cells. The strongest repeated chemical finding in schizophrenia concerns increased dopamine production and release in parts of the *striatum*, a group of structures involved in learning, action and responding to important events. The evidence is more specific than the slogan “too much dopamine in the brain”. It concerns particular processes and locations, with considerable differences between people.

A review of 21 brain-imaging studies, involving 269 patients and 313 controls, found a moderately large average difference in dopamine production and release. The result was clearer in dorsal parts of the striatum than in the subdivision emphasised by some older theories. Importantly, the review excluded studies specifically investigating schizophrenia that had not responded to treatment. [Where the dopamine differences occur](https://doi.org/10.1093/schbul/sbx180)

A larger analysis of 65 studies confirmed increased dopamine production and release on average, but did not find significantly elevated average production in the treatment-resistant groups. Measurements of dopamine receptors and transporters did not show the same consistent pattern. A general statement about all dopamine functions in all people with schizophrenia would therefore go beyond the evidence. [Variation in dopamine findings between groups](https://doi.org/10.1016/j.biopsych.2019.07.008)

One explanation is that altered dopamine signalling can make an ordinary event feel unusually important, personally relevant or threatening. A person may then try to understand why it feels so significant, contributing to a troubling belief. This is a plausible connection between a brain process and experience, not a scan-based explanation of how any particular belief arose. Dopamine findings are most informative about hallucinations and delusions. They explain less about persistent thinking difficulties, low motivation or limited emotional expression. [Dopamine and other mechanisms](https://doi.org/10.1001/jamapsychiatry.2026.1684)

## Other brain messengers and connections

Glutamate helps nerve cells activate other cells; another messenger, called GABA, usually helps restrain activity. Researchers investigate how these processes work together, rather than assuming that more or less of one chemical alone explains schizophrenia.

A leading proposal concerns NMDA receptors, one type of receptor that responds to glutamate. Reduced receptor activity might disturb the balance of activation and restraint in brain circuits, with effects on wider networks and dopamine regulation. Genetic findings involving glutamate receptors support this line of investigation. Drugs that block NMDA receptors can also produce psychotic and thinking difficulties. Even so, the complete sequence remains a working explanation, not a demonstrated pathway shared by everyone. [Review of brain signalling and schizophrenia](https://doi.org/10.1038/s41572-025-00667-6)

Studies of brain tissue after death have found differences in molecules used for inhibitory signalling. It remains uncertain whether these changes helped initiate illness, followed other changes, or partly compensated for them. Likewise, a review of 92 studies measuring glutamate found different patterns according to brain region, illness stage, medication and treatment response. These measurements concern chemical concentrations in tissue; they do not directly reveal receptor activity or communication at each connection. A lower measurement in one place cannot establish a general glutamate shortage. [Glutamate measurement review](https://doi.org/10.1177/00048674241254216)

Brain scans show differences across networks, including the tissue containing nerve-cell bodies, the pathways connecting regions, and patterns of activity. A 2025 review brought together 50 reports of combined imaging studies. Such work supports widespread involvement of brain circuits. Comparing one group early in illness with another group years later, however, does not demonstrate that every individual passes through the same changes. Illness, medication and circumstances may all influence the findings. [The brain-imaging evidence](https://doi.org/10.1038/s44220-025-00493-5)

## Immunity and the development of brain connections

The immune system also helps regulate connections between nerve cells. A particularly influential genetic finding concerns C4, part of an immune system known as complement. Genetic variations linked to greater predicted production of C4A in the brain were also linked to schizophrenia susceptibility. Related experiments showed a role for C4 in removing connections between nerve cells. This offered a specific bridge between genetic evidence and a possible biological process. [The original C4 study](https://doi.org/10.1038/nature16549)

The often-repeated conclusion that schizophrenia is caused by excessive removal of brain connections in adolescence is much less secure. A 2026 review found only 11 eligible studies relating C4 measures to brain imaging. Results differed across brain regions, populations and methods. Excessive removal of connections and inflammatory effects remained possible explanations, rather than an established general human disease process. [C4 and brain imaging: the later review](https://doi.org/10.1016/j.bbi.2026.106873)

Other immune differences have been reported. A review of 69 studies involving 5,710 people found average differences in several inflammatory substances in the fluid surrounding the brain and spinal cord, including substances called interleukins 6 and 8. Yet average routine fluid measurements remained within normal ranges. Inflammation, changes in the barrier separating blood from brain tissue, and changes in fluid movement could all contribute. These results do not show which changes start the illness. [The review of inflammatory fluid measurements](https://doi.org/10.1016/j.schres.2024.02.001)

There is a separate clinical issue: some illnesses in which the immune system attacks the nervous system can first appear as psychosis, occasionally before obvious neurological signs emerge. They need appropriate medical investigation. Their existence does not mean that schizophrenia generally is an autoimmune disease. [Consensus guidance on psychosis of suspected autoimmune origin](https://doi.org/10.1016/S2215-0366%2819%2930290-1)

## How expectations shape perception

Another research approach asks how the brain combines expectations with new information. Expectations help everyone interpret uncertain sights and sounds. Some theories propose that hallucinations or delusions arise when the balance between expectation and incoming evidence changes.

A September 2026 review combined 27 studies involving 904 people with schizophrenia-spectrum conditions and 1,039 controls. Across the perceptual tasks tested, it found essentially no average difference in reliance on expectations. Results varied substantially between studies. This challenges a general explanation that applies across these tasks, while leaving room for particular symptoms, situations or subgroups. It also does not test every version of the broader theory. [The review of expectations and perception](https://doi.org/10.1017/S0033291726105455)

Researchers therefore disagree less about whether many biological and psychological processes are involved than about how they fit together. Dopamine changes might be downstream of developmental, environmental and other brain changes. Alternatively, partly separate pathways may lead to similar experiences. Proposed biological subgroups still need to show that they reliably predict course or treatment response in new populations. There are no established routine categories of “dopamine schizophrenia”, “glutamate schizophrenia” or “inflammatory schizophrenia”.

## Why research findings cannot yet identify one person's cause

Each research method answers some questions better than others. Illness cannot change the genetic sequence a person inherited at conception, but a genetic association does not by itself identify the responsible gene, cell or developmental period. Environmental studies must disentangle overlapping experiences and family influences. Studies following people over time help, but subtle emerging illness may already have changed exposure or behaviour before a diagnosis is made.

Brain imaging and studies of tissue after death often examine people after illness and treatment have begun. A reproducible difference can therefore be a cause, consequence, adaptation or medication effect. Finding that a drug helps by changing a particular system also does not prove that an original deficiency in that system caused the disorder. A successful treatment can act on a useful point in a chain of events without reversing its first cause.

Better answers will require repeated observations beginning before illness, studies that test whether changing a suspected mechanism produces its predicted effects, and proposed subgroups that hold up in independently recruited patients. Theories should specify results that would count against them, rather than accommodating every finding afterwards. Under-represented ancestries and incomplete measurement of social conditions remain important gaps. A review published recently may also rely on studies conducted many years earlier. [Current questions about causes and mechanisms](https://doi.org/10.1136/bmj-2026-635470)

## Treatment should address several different goals

Treatment aims can include reducing frightening experiences, preventing another episode, improving concentration, returning to work or education, and making everyday life more manageable. Success on one measure does not establish success on all of them. The best-supported general approach combines an individually chosen antipsychotic medicine with psychological treatment, family support where suitable, and practical help with daily life. The plan should reflect what matters to the person and be reviewed as needs change. [American Psychiatric Association treatment guideline](https://www.psychiatry.org/psychiatrists/practice/clinical-practice-guidelines/schizophrenia)

Antipsychotics have the strongest evidence for reducing hallucinations, delusions and disorganised thinking during an episode. A large comparison of 32 oral medicines drew on 402 trials involving 53,463 adults with more than one episode. Most medicines improved symptoms, but unwanted effects differed greatly. There was no universally best first choice. [Comparison of oral antipsychotics](https://doi.org/10.1016/S0140-6736%2819%2931135-3)

For example, olanzapine has relatively strong average symptom benefits but can cause substantial weight gain and changes affecting diabetes and heart health. Risperidone and paliperidone can raise the hormone prolactin. Aripiprazole generally has a more favourable weight and metabolic profile but can cause severe inner restlessness. These experiences matter both in their own right and because they affect whether treatment is sustainable.

The 2026 Australian and New Zealand guideline favours considering initial options with relatively favourable effects on weight, blood sugar and heart health, including aripiprazole, brexpiprazole, cariprazine and lurasidone. That is a judgement about balancing benefits and harms, not evidence that one of these medicines works best for everyone. [The Australian and New Zealand guideline](https://doi.org/10.1177/00048674251406058)

## Preventing relapse and reviewing medication over time

Continuing effective medication substantially reduces the chance of relapse in clinical trials. A Cochrane review included 75 trials and 9,145 participants overall. In the trials reporting relapse over seven to twelve months, about 24% of people continuing antipsychotics relapsed, compared with 61% assigned to placebo or discontinuation. That is about 37 fewer relapses per 100 people over that period, or one additional relapse prevented for roughly every three people treated. It is not a lifetime prediction or an individual guarantee. [Antipsychotics for relapse prevention](https://doi.org/10.1002/14651858.CD008016.pub3)

The comparisons also have limits. Some trials selected people who had already responded to medication, and withdrawal can affect outcomes. The results cannot be treated as exact forecasts for every carefully supported gradual reduction. They nevertheless provide strong evidence for continuing treatment as a means of relapse prevention.

Long-acting injections are another way of receiving antipsychotic medicine. They can be useful when someone prefers them or finds taking daily medication difficult. Across randomised trials, injections produced a modest average reduction in hospitalisation or relapse compared with oral treatment: about 12% lower relative risk, with uncertainty around that estimate. Much larger benefits appeared in studies comparing people before and after starting injections; those designs are more vulnerable to other explanations. Reliable delivery can be valuable, but its benefit varies between people. [Injections compared with oral medication](https://doi.org/10.1016/S2215-0366%2821%2900039-0)

For adults whose first psychotic episode has settled, WHO recommends offering at least seven to twelve months of maintenance treatment, taking effectiveness, unwanted effects and preference into account. That minimum is not an automatic stopping date. Decisions about continuing or reducing treatment require a review of the person's course and circumstances. [WHO guidance after a first psychotic episode](https://www.who.int/teams/mental-health-and-substance-use/treatment-care/mental-health-gap-action-programme/evidence-centre/psychosis-and-bipolar-disorders/duration-of-antipsychotic-treatment-in-individuals-with-a-first-psychotic-episode)

The HAMLETT trial illustrates the uncertainty about longer-term trade-offs. It randomly assigned 347 people whose first psychotic episode had settled to early dose reduction or discontinuation, or maintenance treatment. Reduction produced more relapse in the first year and no advantage on the main measure of functioning reported by participants. Researchers' ratings of functioning were better in the reduction group at years three and four. However, medication doses had become similar after the first year, so the later difference cannot simply be attributed to staying on less medication. These findings from first-episode care also cannot automatically be transferred to people with repeated episodes. [The HAMLETT trial](https://doi.org/10.1001/jamapsychiatry.2025.2525)

There is a legitimate clinical discussion about supported reduction for selected people. Current evidence does not support routine early stopping. Changes should be agreed with the treating team; NICE recommends gradual withdrawal when undertaken, with continuing monitoring for relapse. [NICE guidance on continuing treatment and withdrawal](https://www.nice.org.uk/Guidance/CG178/chapter/recommendations)

## When symptoms persist: clozapine and specialist treatment

If psychotic symptoms persist despite two adequate antipsychotic trials, guidelines recommend offering clozapine. First, clinicians need to check the diagnosis, doses, duration and whether enough medication was actually taken or absorbed, alongside substance use, physical illness and other possible reasons for poor response. Repeatedly trying further standard medicines can delay access to the recommended treatment. [NICE guidance on treatment resistance](https://www.nice.org.uk/Guidance/CG178/chapter/recommendations)

The Royal College of Psychiatrists reaffirmed this role in January 2026 and highlighted underprescribing and avoidable delay. Clozapine needs careful monitoring because it can cause dangerously low white blood cells, inflammation of the heart, severely slowed bowel movement, and substantial weight and metabolic problems. Sleepiness and excess saliva can also be burdensome. Blood tests alone do not cover all these risks. [The Royal College's clozapine statement](https://www.rcpsych.ac.uk/docs/default-source/improving-care/better-mh-policy/position-statements/position-statement---ps01-26---clozapine.pdf)

Clozapine has particular evidence concerning suicidal behaviour. In the two-year InterSePT trial, 980 high-risk people with schizophrenia or schizoaffective disorder received clozapine or olanzapine. Clozapine reduced the trial's combined suicidal-behaviour outcome; there were 34 suicide attempts in its group compared with 55 in the olanzapine group. Too few people died by suicide to establish that clozapine reduced suicide deaths in this trial. Preventing attempts and demonstrating lower mortality are different findings. [The InterSePT trial](https://doi.org/10.1001/archpsyc.60.1.82)

The recommendation for clozapine should not conceal uncertainty about the size of its advantage over other newer medicines. A 2025 analysis of 19 trials, involving 1,599 people, did not demonstrate better symptom reduction with clozapine at six to eight weeks. The estimated difference was near zero, but the evidence was judged very uncertain; individual patient data were available for only 1,052 participants, and early discontinuation and differences between trials complicated interpretation. The study neither establishes equal effectiveness nor settles longer-term benefits. [The comparative clozapine analysis](https://doi.org/10.1016/S2215-0366%2825%2900001-X)

Electroconvulsive therapy, or ECT, has a more specialised place. It may be used for *catatonia*, a serious disturbance of movement and responsiveness. Specialist guidance recommends a medicine from the benzodiazepine group, usually lorazepam, and/or ECT for this syndrome. ECT may also be considered alongside clozapine in selected cases that have remained resistant, although that evidence is weaker. It is not routine treatment for everyone with schizophrenia. [Specialist guidance on catatonia](https://doi.org/10.1177/02698811231158232)

## Talking treatments and family support

Cognitive behavioural therapy for psychosis helps people explore interpretations of troubling experiences, develop ways of coping and reduce distress. A 2024 Cochrane review of first-episode and recent-onset psychosis included 28 trials; 26 supplied data from 2,407 people. Adding therapy to standard care produced modest average benefits in symptoms and functioning. Confidence varied by outcome, and the review did not establish a clear reduction in relapse in that population. [CBT added to care early in psychosis](https://doi.org/10.1002/14651858.CD015331.pub2)

Another review found small-to-medium benefits for general and psychotic symptoms at treatment completion, with less consistent evidence at longer follow-up. Therapy quality, the comparison service, illness stage and whether assessors knew the treatment assignment can all affect results. People and therapists cannot usually be kept unaware of which talking treatment is being delivered, which makes these trials harder to interpret. [The broader review of CBT findings](https://doi.org/10.1016/j.eclinm.2023.102392)

Most studies tested therapy alongside usual care, generally including medication. They do not establish that therapy alone provides equivalent protection in established psychotic illness. Recommendations to offer it reflect its complementary role, rather than a claim that it solves every symptom or prevents every recurrence.

Structured family interventions combine explanation, communication skills, problem-solving and planning for possible deterioration or crises. They can reduce the burden on relatives and may reduce relapse. The 2024 Cochrane review included 26 trials and 1,985 patients. Its estimate of about one-third lower relative relapse risk shortly after treatment relied on only four trials with 229 people, and confidence in that estimate was low. Evidence for reduced caregiver burden was more secure. A structured programme should not be equated with occasional information-giving, and family involvement needs to fit the person's circumstances and wishes. [The family-intervention review](https://doi.org/10.1002/14651858.CD013541.pub2)

## Recovery in work, learning and everyday life

Coordinated early intervention brings medication, psychological work, family support, case management, and help with work or education into one service. A review of ten randomised trials with 2,176 participants found better engagement and several clinical and everyday outcomes during treatment. Hospitalisation was about 26% less likely in relative terms, although possible publication bias adds uncertainty. It is less clear how long the advantages persist after specialist support ends. [Coordinated early intervention research](https://doi.org/10.1001/jamapsychiatry.2018.0623)

For people who want paid work, individual placement and support is particularly well supported. It combines assistance finding an ordinary job with ongoing employment and mental health support. Across 32 trials involving 7,665 people, about 49% obtained employment during follow-up with this approach, compared with 28% using comparison services. About 45% of participants had schizophrenia-spectrum conditions; the rest had other mental disorders. These are therefore not schizophrenia-specific predictions, and employment does not by itself establish improvement in every other area of life. [The supported-employment review](https://doi.org/10.1017/S2045796022000300)

*Cognitive remediation* means structured practice and strategies for attention, memory and other thinking skills. Across 130 studies with 8,851 participants, average gains were small: 0.29 standard-deviation units for thinking skills and 0.22 for everyday functioning. Those numbers compare the size of changes with variation between people; 0.29 does not mean a 29% improvement. Programmes tended to do better when trained therapists helped people develop strategies and connect them with wider rehabilitation. The useful question is whether practice helps someone carry out meaningful activities beyond the sessions. [The cognitive-remediation review](https://doi.org/10.1001/jamapsychiatry.2021.0620)

People with repeated admissions or difficulty staying connected to services may need more intensive community support. Access to treatment, practical assistance and continuity of relationships matter alongside which particular intervention is chosen. Recovery goals can include study, relationships, independent living or enjoyable activity as well as paid employment.

## Low motivation, thinking difficulties and physical health

The term *negative symptoms* refers to reductions in things such as motivation, speech and outward emotional expression. It does not mean bad behaviour or a negative attitude. These difficulties can remain even when hallucinations and delusions improve, and treating them is still an important gap.

Some apparently similar difficulties have other contributors: depression, continuing psychosis, medication-related sleepiness or movement problems. Those should be assessed because they may be treatable. European guidance recognises promising evidence for medicines including cariprazine, but does not identify a universally effective treatment for persistent primary negative symptoms. [European guidance on negative symptoms](https://doi.org/10.1192/j.eurpsy.2021.13)

Thinking difficulties may also remain after psychosis settles. Simply increasing antipsychotic doses does not reliably resolve them. Rehabilitation and cognitive remediation can help, alongside reviewing medicines that impair alertness or thinking. The aim is to support the person's ability to act and participate, not merely to lower a symptom score.

Physical health needs active care. Weight, blood pressure, blood sugar, blood fats, movement effects and other unwanted experiences should be monitored and addressed. Help with stopping smoking, physical activity, and treatment of diabetes and heart disease belongs within effective mental health care. The 2026 Australian and New Zealand guideline also recommends considering metformin early for medication-related weight gain, particularly with higher-risk medicines such as olanzapine and clozapine. This requires an individual assessment of benefits, reasons not to use it and possible harms. [Physical health and treatment recommendations](https://doi.org/10.1177/00048674251406058)

## A newer medicine and what remains uncertain

Xanomeline–trospium, marketed as Cobenfy, offers a different way of affecting brain signalling. It targets receptors responding to acetylcholine rather than directly targeting dopamine receptors in the usual way. US regulators approved it for adults with schizophrenia on 26 September 2024. It can cause digestive problems and has important warnings, including difficulty emptying the bladder and increased heart rate. A different mechanism does not mean freedom from unwanted effects. [The FDA approval announcement](https://www.fda.gov/news-events/press-announcements/fda-approves-drug-new-mechanism-action-treatment-schizophrenia)

In the five-week EMERGENT-3 trial, 256 participants received the medicine or placebo. The medicine produced an additional average reduction of 8.4 points on a clinician-rated symptom scale. The trial established short-term benefit compared with placebo, but did not compare it with an established antipsychotic. [The EMERGENT-3 trial](https://doi.org/10.1001/jamapsychiatry.2024.0785)

A 2026 publication reported a year-long study in which 277 of 566 treated participants completed follow-up. Everyone received the medicine, without a comparison group. Continuing improvement among those who remain in such a study cannot settle comparative effectiveness, relapse prevention or long-term recovery for everyone who starts it. [The year-long EMERGENT-5 study](https://doi.org/10.1016/j.schres.2025.12.015)

Its place among other treatments therefore remains under investigation. The supplied November 2026 research recorded NICE's UK appraisal as still in development; availability and recommendations should be checked locally rather than inferred from US approval. Effective treatment through this pathway also does not prove that schizophrenia began with a shortage of acetylcholine. [NICE's appraisal project](https://www.nice.org.uk/guidance/indevelopment/gid-ta11925)

## What better evidence and better care need to address

The strongest treatment conclusions concern reducing acute psychosis and preventing relapse. The best long-term balance for an individual is less settled. Trials lasting a few weeks cannot tell us enough about years of independence, work, physical health and medication burden. Several favourable reviews may also contain many of the same trials, rather than independent replications. When participants leave a study, their later experiences may be missing from the apparent results.

Research particularly needs to identify who can safely reduce medication with support, what reliably improves persistent motivation and thinking difficulties, how much clozapine helps under well-delivered longer-term treatment, and how newer medicines compare with established options over years. Studies should measure relapse and hospitalisation alongside people's own assessments of functioning, quality of life and unwanted effects.

A useful care plan brings these concerns together: a jointly agreed medication approach, accessible psychological and practical support, active attention to physical health, and regular review of progress towards the person's goals. Benefits and burdens deserve discussion together. None of the group averages in this essay identifies the right medicine, dose or stopping schedule for an individual. If someone cannot stay safe or there is an immediate risk of harm, urgent local medical or crisis help is needed.

---

AI-assisted adaptation of supplied research; not independently clinically reviewed. Original essay text: CC0 1.0. Published 7 November 2026.