Weight Gain from Psychiatric Medication, and Where Drugs Like Ozempic Fit

If your weight rose after starting a psychiatric medicine, it is worth being precise about what happened, because the explanation people reach for first is usually the wrong one.

Several psychiatric medicines act directly on the brain systems that decide when you feel hungry, when you feel full, and how rewarding food is. Some also change how the body handles insulin and how much energy it burns at rest. The practical result is that a person can eat what feels like a reasonable amount and still gain weight, or can feel a hunger that does not settle after a meal.

This is a known effect of the treatment. It is not evidence of weak willpower, and it is not something to feel awkward raising in a consultation. Saying so plainly matters, because shame keeps people quiet, and people who stay quiet tend to stop the medicine on their own instead of bringing up the problem. Weight gain is one of the most frequent reasons an effective treatment gets dropped.

At the same time, the medicine is rarely the whole story. Sleep, the severity of the illness itself, physical activity, diet, other medical conditions, family history and ordinary life circumstances all contribute. The honest position is that the medicine is often a major cause and almost never the only one.

Psychiatric medicines do not carry equal risk. Among antipsychotics, olanzapine and clozapine are consistently associated with the greatest weight gain and the largest effect on blood sugar and lipids. Aripiprazole, lurasidone and ziprasidone have more favourable average profiles. Valproate, lithium and mirtazapine can also cause substantial gain.

Averages are not predictions. Some people gain very little on olanzapine, and some gain a great deal on a medicine with a low average risk.

The important point is that choosing a psychiatric medicine on weight alone is a mistake. Clozapine is prescribed when other antipsychotics have failed, and stopping it to protect weight can cost someone the only treatment that has ever controlled their illness. The question is not which medicine causes the least weight gain. It is how to protect physical health without giving up psychiatric stability.

Before anything newer is considered, there is a well-evidenced step that often gets skipped. Metformin, an old and inexpensive diabetes tablet, has better evidence for preventing antipsychotic weight gain than any other medicine. An international guideline published in Schizophrenia Bulletin in 2025 found that starting metformin at the same time as the antipsychotic reduced later weight gain by about 4 kg on average compared with control, and concluded that metformin is the only drug with demonstrated efficacy for preventing this problem. It is worth asking about, particularly when olanzapine or clozapine is being started.

Ordinary monitoring belongs here too. Weight, waist circumference, blood pressure, blood sugar and lipids, measured at the start and then at intervals, so that a problem is caught while it is still small rather than discussed years later.

This is no longer a matter of anecdote. Several randomised trials have tested these drugs specifically in people taking antipsychotics.

In a 2017 trial published in JAMA Psychiatry, 103 people with a schizophrenia-spectrum disorder who were taking clozapine or olanzapine and had prediabetes received liraglutide or placebo. Over 16 weeks the liraglutide group lost 5.3 kg more than placebo, and glucose tolerance returned to normal in 64 per cent of them compared with 16 per cent on placebo.

The COaST trial, published in Lancet Psychiatry in 2025, tested semaglutide in people with schizophrenia taking clozapine who were living with obesity. At 36 weeks the semaglutide group had lost about 13.9 per cent of body weight against 0.4 per cent on placebo. Psychotic symptoms did not worsen, and clozapine blood levels were unchanged. The important limitation is size: the trial was stopped early when the study drug became unavailable, so only 31 people took part.

The HISTORI trial, published in JAMA Psychiatry in 2025, was larger. It gave semaglutide or placebo to 154 people on antipsychotics who had prediabetes and were overweight or obese. At 30 weeks the semaglutide group had lost 9.21 kg more than placebo, HbA1c fell by 0.46 percentage points, and physical quality of life improved. Psychotic symptoms did not worsen and serious adverse events were no more common than on placebo.

A 2025 review pooling trials across psychiatric populations found an average weight reduction of about 5 kg. Nausea, vomiting and constipation were more frequent on the drug, but people did not stop treatment more often because of them.

Two sets of clinical guidelines now list these medicines as an option for antipsychotic-related weight gain: the international INTEGRATE schizophrenia guidelines published in 2025, and Australian and New Zealand guidelines published in 2026. The second describes it as a weak recommendation based on evidence of very low certainty, which is a fair summary of where things stand. This is a recognised option, not a fringe idea, and not yet a settled one.

One gap deserves emphasis. Almost all of this evidence comes from people taking antipsychotics, mostly clozapine and olanzapine. There is no randomised trial testing these drugs for weight gain caused specifically by lithium, valproate or mirtazapine. That does not mean they cannot help. It means the question has not been answered, and anyone who tells you the evidence covers those medicines is going beyond what has been studied.

For roughly two years these medicines carried a warning about suicidal thoughts, and it caused a great deal of worry. That warning was never based on evidence from the drugs themselves. It was carried across from older weight-loss medicines.

On 13 January 2026 the US Food and Drug Administration asked for the warning to be removed from the three weight-loss products that carried it. The decision followed a pooled analysis of 91 placebo-controlled trials involving 107,910 participants, and a separate study of more than 2.2 million people. Neither found an increased risk of suicidal thoughts or behaviour, and neither found an increase in depression, anxiety, irritability or psychosis. Europe's drug safety committee had reached the same conclusion in April 2024.

A 2025 analysis in JAMA Psychiatry pooling 80 placebo-controlled trials points the same way, finding no difference in psychiatric side effects or in depressive symptoms between the drug and placebo.

What that does not mean is that no individual can experience a change in mood while taking one. Large studies describe what happens on average. If mood worsens after starting any medicine, it deserves proper assessment rather than being automatically blamed on the drug, or automatically dismissed as the underlying illness.

This is where a psychiatric assessment adds the most, and it is subtler than asking whether someone has an eating disorder.

The reassuring part first. Binge eating tends to improve. A 2025 review of five studies found reductions in binge eating scores, and a larger 2026 review of 25 trials found reduced binge eating, reduced loss-of-control eating and reduced emotional eating. A history of binge eating is therefore not, by itself, a reason to rule these medicines out.

The same 2026 review found something else. Dietary restraint went up. The authors were careful to say it is not yet clear whether that represents healthy self-regulation or the beginning of rigid restriction. From the outside, in the early stages, the two can look identical. Praise from other people looks identical too.

Distinguishing between them is exactly what a psychiatric assessment is for. Questions worth answering honestly, before starting and again during treatment:

The National Eating Disorders Association puts the position fairly: very little research has been done on these drugs in people with eating disorders, and they can be harmful when used without eating disorder expertise, when inadequately monitored, or when the motivation for weight loss is driven by weight stigma rather than health.

Some people report that they simply stop wanting alcohol, and early evidence does support the observation. In a 2025 randomised trial in JAMA Psychiatry, 48 adults with alcohol use disorder received low-dose semaglutide or placebo for nine weeks. Semaglutide reduced craving, reduced how much people drank on the days they drank, and reduced heavy drinking over time. It did not reduce the number of days on which people drank at all.

That was 48 people over nine weeks, and none of them were seeking treatment for alcohol. A larger trial has since reported. In 2026, The Lancet published a randomised trial of 108 people with alcohol use disorder and obesity given semaglutide or placebo alongside cognitive behavioural therapy for 26 weeks, in which heavy drinking days fell significantly further on semaglutide.

This is now a real finding rather than a hint, and larger trials are underway. It still does not make these medicines a licensed treatment for alcohol dependence, and the established treatments remain the established treatments. It is also worth saying that a reduced urge to drink is not the same as a resolved drinking problem. The wider mental health picture is covered here.

Lithium. There are now several published reports of lithium levels rising after semaglutide was started, in some cases to the point of toxicity, without any change in kidney function or in other medicines. The likely explanation is straightforward: vomiting, reduced fluid intake and eating less all change how the body handles lithium. This amounts to a handful of case reports and no formal study has been done, so the true size of the risk is unknown. The practical response is not alarm but attention. If you take lithium and start one of these drugs, your lithium level should be checked more closely than usual, and vomiting, diarrhoea, tremor, unsteadiness, drowsiness or confusion should be reported promptly rather than waited out.

Clozapine and the bowel. Clozapine slows the gut in most people who take it, and constipation is common. Rarely it progresses to a blockage, which is serious. In the largest study with a proper denominator, roughly 37 in every 10,000 people taking clozapine developed a severe bowel complication, and about 7 in 10,000 died from one. That is a small risk for any individual, but it is a higher number than the blood-count problem clozapine is far better known for. Since these weight-loss drugs also slow the gut and commonly cause constipation, bowel habit stops being a minor detail on a side-effect list. New or worsening constipation while taking both needs treating early, not enduring.

Beyond those two, most psychiatric medicines do not need a dose change. The COaST trial specifically measured clozapine levels and found them unchanged, although in only 31 people. What should be said honestly is that almost no interaction studies have been done between these drugs and psychiatric medicines. Not having found a problem and not having looked are different things.

There is also a simpler mechanism worth naming. Nausea and vomiting cause tablets to be missed. A psychiatric relapse that follows has nothing to do with any pharmacological interaction and everything to do with a week of not keeping medication down.

This belongs in the conversation before starting rather than after.

In the extension of a large semaglutide trial, people who took the drug for 68 weeks and then stopped regained about two thirds of their weight loss over the following year, and the improvements in blood pressure, blood sugar and lipids drifted back towards where they had started.

A systematic review published in the BMJ in January 2026, covering 37 studies and 9,341 people, found that weight came back at an average of about 0.4 kg a month after stopping weight-management medication, and faster for semaglutide and tirzepatide at about 0.8 kg a month. On average, weight returned to its starting point within about 1.7 years.

Individual paths vary and averages do not settle any one person's case. But the practical implications are real. Treatment is likely to be long term, cost becomes a recurring rather than a one-off question, and stopping needs a plan rather than simply running out.

There is a psychological implication as well. If someone has come to feel that their confidence, their standing at work or their worth depends on the new body, weight returning is not experienced as a change in a number. That is much better anticipated in advance than discovered afterwards.

Four questions cover most of it.

Is the current psychiatric medicine essential and working? If it is, protecting it usually matters more than protecting the number on the scale, and the metabolic problem should be treated in its own right.

Could a psychiatric medicine with a lower metabolic risk control the illness just as well? Sometimes yes. That decision has to weigh previous response, how severe relapses have been, and how difficult the illness has been to treat, not weight alone.

Have the established steps been offered? Metabolic monitoring, sleep, activity, diet, and metformin where it applies. These should not quietly disappear from the plan because something newer exists.

Is there a medically appropriate reason for this treatment in its own right? The relevant question is the person's overall metabolic health, assessed by a doctor qualified to prescribe and monitor these medicines. Dissatisfaction with a number on the weighing scale is not the same thing.

Two final points. Weight regain after stopping is not proof of laziness or of psychiatric failure. It is a well-documented biological response to withdrawing a medicine that was regulating appetite. And no injection resolves weight stigma, emotional eating, or a lifetime of being spoken to a particular way about one's body.

These medicines change appetite and weight considerably more reliably than they change a person's relationship with their own body. The second remains psychological work.