Metformin and Antipsychotics: A Review of Their Combined Use

Adding metformin to antipsychotic treatment consistently reduces the metabolic harm these medications cause, with meta-analyses across hundreds of patients showing roughly 3 to 4 kilograms less weight gain compared to placebo. The combination has been studied for over two decades, and while metformin is best known as a first-line diabetes drug, its role alongside antipsychotics addresses a different clinical reality: the substantial weight gain, insulin resistance, and cholesterol disturbances that antipsychotic medications trigger in many patients. The evidence supporting this pairing has grown strong enough that recent clinical guidelines now recommend considering metformin at the very start of antipsychotic therapy, though real-world prescribing still lags well behind the science.

Why Antipsychotics Cause Metabolic Problems

Antipsychotic medications, particularly the newer “atypical” or second-generation class, are effective at managing psychosis, bipolar disorder, and other conditions. But they come with a metabolic cost that goes beyond simple overeating. These drugs interact with a range of receptors in the brain and body that regulate appetite, fat storage, blood sugar, and cholesterol. In the hypothalamus, they disrupt signaling through dopamine, serotonin, histamine, and acetylcholine pathways, which ramps up appetite-promoting signals and increases the body’s sympathetic nervous system output. That cascade raises glucagon levels and pushes the liver to produce more glucose.¹

The two antipsychotics most strongly linked to weight gain, olanzapine and clozapine, happen to bind tightly to serotonin 5-HT2C and histamine H1 receptors, both of which play important roles in appetite regulation. Blocking these receptors drives up food intake. But the metabolic damage isn’t limited to eating more. Blocking muscarinic M3 receptors in the pancreas can directly impair insulin release, contributing to diabetes risk even independently of weight gain. Antipsychotics also cause the body to secrete more leptin without the expected drop in appetite, a sign that the brain’s leptin-sensing machinery has been thrown off.² The result is a cluster of problems: weight gain, rising blood sugar, worsening cholesterol, and growing insulin resistance that together raise the risk of diabetes, heart disease, and early death.

How Metformin Pushes Back

Metformin works through several mechanisms that happen to counter much of what antipsychotics do metabolically. In the liver, it suppresses excess glucose production, which directly opposes the antipsychotic-driven increase in hepatic glucose output.³ But what makes the pairing especially logical is that metformin also acts in the brain, where many of the antipsychotic metabolic effects originate.

In the hypothalamus, metformin suppresses AMPK activity, which sounds counterintuitive since it activates AMPK in the liver. But this tissue-specific difference is key. Suppressing hypothalamic AMPK reduces the production of appetite-promoting signals and decreases food intake. Metformin also restores leptin sensitivity in the hypothalamus, directly countering the leptin resistance that antipsychotics create. Brain imaging studies in people with type 2 diabetes have shown that metformin changes activity in brain regions involved in food-reward processing, which could help explain why patients report reduced cravings.⁴

A third axis of action involves the gut. Antipsychotic treatment has been shown to shift the composition of intestinal bacteria in ways associated with metabolic dysfunction, reducing beneficial species and increasing others linked to obesity. Metformin partially reverses these changes and plays a role in correcting disrupted satiety-related signaling between the gut and brain.⁵ So rather than just treating one downstream consequence, metformin addresses the metabolic disruption at multiple levels simultaneously.

What the Weight Data Actually Show

Several large meta-analyses have now quantified metformin’s effect on antipsychotic-related weight gain. A 2016 meta-analysis pooling 12 studies with 743 patients found that metformin treatment reduced body weight by about 3.3 kg more than placebo, with significant reductions in BMI and insulin resistance as well.⁶ A more recent 2024 meta-analysis focusing specifically on co-commencement found a similar figure of about 3.1 kg less weight gain in the metformin group.⁷

A network meta-analysis published in JAMA Psychiatry compared multiple weight-management drugs used alongside antipsychotics and found that metformin, drawing on 16 trials, produced an average weight reduction of roughly 3.9 kg versus placebo. The same analysis found that semaglutide and liraglutide, both GLP-1 receptor agonists, outperformed metformin on weight loss alone, with semaglutide achieving around 11 kg of weight reduction. Both semaglutide and metformin achieved clinically meaningful weight loss of 5% or greater.⁸

These numbers might seem modest in absolute terms, but context matters. In populations where antipsychotics can cause 7 to 10 kg or more of weight gain in the first year, shaving off 3 to 4 kg represents a meaningful reduction in cardiometabolic risk. And unlike many interventions studied for this purpose, metformin has shown consistent effects across different antipsychotics, different study designs, and different populations.

Beyond Weight: Lipids, Insulin, and Cardiovascular Risk

The metabolic benefits of metformin in this population extend beyond the scale. A pooled analysis of two randomized, placebo-controlled trials tracked lipid changes over 24 weeks and found that metformin produced a substantial drop in LDL cholesterol. The proportion of patients with elevated LDL fell from comparable baseline levels to about 25% in the metformin group versus 65% in the placebo group by week 24. Metformin also significantly lowered total cholesterol and triglycerides while raising HDL (“good”) cholesterol.⁹

An important finding from that research was that the timeline for these improvements differs. Insulin resistance began improving around week 12, but the lipid improvements didn’t become statistically significant until week 24. This suggests that shorter trials may underestimate metformin’s full benefit on cholesterol, and that clinicians should allow adequate time before judging whether the drug is working on the lipid front. Animal studies have also confirmed that metformin reverses olanzapine-induced insulin resistance through mechanisms that include reducing inflammatory activity in fat tissue.¹⁰

Starting Early Makes a Difference

One of the strongest messages from the research is that timing matters. Starting metformin at the same time as the antipsychotic, rather than waiting until weight gain has already taken hold, produces better results. A Cochrane review identified metformin as the only drug with evidence supporting prevention of antipsychotic-induced weight gain, with co-commenced patients gaining about 4 kg less than controls.¹¹ A large naturalistic study of patients prescribed clozapine found that the weight gap between metformin and control groups actually widened over time, from about 3.1 kg at six months to 5.4 kg at twelve months.¹²

Interventions that begin after weight gain is already established rarely help patients return to their original weight. One review concluded that adjunctive medications for weight management appear more effective when started at or near the initial antipsychotic exposure.¹³ A trial in drug-naive, first-episode schizophrenia patients confirmed this: metformin successfully attenuated both weight gain and insulin resistance when added from the beginning of olanzapine treatment.¹⁴ This doesn’t mean late addition is worthless, but it highlights that the earlier the intervention, the better the outcome.

Children and Adolescents

Antipsychotic prescribing has increased substantially in younger populations, making the metabolic side effects a growing concern in pediatric psychiatry. A meta-analysis of randomized controlled trials in children and adolescents found that 12 to 16 weeks of metformin therapy was associated with a significant reduction in weight (about 4.5 pounds more than placebo) and a significant reduction in BMI z-score and insulin resistance.¹⁵

The side effects were predictable: higher rates of nausea, vomiting, and diarrhea in the metformin group, which is consistent with what’s seen in adults. However, these gastrointestinal issues didn’t lead to more kids stopping the medication — the dropout rates were similar between metformin and placebo groups. An earlier randomized trial found that while children on placebo continued gaining about 0.31 kg per week, those on metformin essentially stabilized, and because these were growing children, that stabilization translated into meaningful reductions in weight-for-age z-scores.¹⁶ In a growing child, simply stopping the upward trajectory without causing weight loss can be enough to normalize body composition over time.

Cognitive Effects That Weren’t Expected

One of the more intriguing findings from recent research is that metformin may improve cognitive function in people taking antipsychotics, independently of any improvement in psychiatric symptoms. Cognitive impairment is a core feature of schizophrenia and one of the strongest predictors of functional outcomes, yet few treatments effectively target it.

A 24-week trial found that patients receiving antipsychotics plus metformin showed significantly greater improvement in composite cognitive scores compared to those on antipsychotics alone. The improvements were most pronounced in processing speed, working memory, verbal learning, and visual learning. Crucially, the researchers assessed psychiatric symptoms separately and found no group differences, meaning the cognitive gains were not simply a byproduct of feeling better overall.¹⁷

A meta-analysis across 12 studies found a favorable trend for metformin over placebo on cognitive measures, though it didn’t reach statistical significance overall. However, subgroup analysis showed that studies with 24 weeks of follow-up showed a stronger and near-significant improvement compared to those lasting only 12 weeks, suggesting that the cognitive benefit may need time to develop.¹⁸ Follow-up research has linked these cognitive improvements to changes in brain energy metabolism and restored functional connectivity in the hippocampus.¹⁹

This is still early-stage evidence, and no one is recommending metformin as a cognitive enhancer for schizophrenia based on current data. But it raises an interesting possibility that the metabolic improvements metformin provides (better insulin sensitivity, improved glucose utilization in the brain) may have downstream cognitive benefits, especially since insulin resistance itself is increasingly recognized as a contributor to cognitive decline.

How GLP-1 Receptor Agonists Compare

The emergence of GLP-1 receptor agonists like semaglutide and liraglutide has changed the conversation about managing antipsychotic-related metabolic problems. These newer drugs produce dramatically greater weight loss — the JAMA Psychiatry network meta-analysis found semaglutide outperforming metformin by roughly threefold on weight reduction.⁸

A large real-world cohort study comparing GLP-1 receptor agonists to metformin in antipsychotic-treated patients found striking differences. The incidence of new type 2 diabetes was about 2% in the GLP-1 group versus about 7% in the metformin group, and the GLP-1 group also had lower all-cause mortality. BMI reductions were also significantly greater with GLP-1 receptor agonists.²⁰

So why isn’t metformin simply replaced? Cost and access are the obvious answers. Metformin is generic and extremely inexpensive, while GLP-1 receptor agonists cost many times more and face supply constraints. Metformin also has a much longer safety track record, comes in oral form (most GLP-1 agonists require injections, though oral versions are emerging), and has been studied far more extensively in psychiatric populations. For most patients, metformin remains the practical first choice, with GLP-1 agents reserved for cases where metformin proves insufficient or where the metabolic risk is especially high.

Practical Barriers to Prescribing

Despite the evidence, metformin co-prescription with antipsychotics remains uncommon in practice. A UK primary care study identified several barriers. Clinicians worry about stacking metformin’s gastrointestinal side effects (nausea, abdominal pain, diarrhea) on top of the gut-related effects antipsychotics already cause. Alcohol use complicates matters, since alcohol enhances metformin’s blood-sugar-lowering effect and increases the risk of lactic acidosis. Renal function needs annual monitoring, and vitamin B12 levels should be checked periodically, especially in people at higher risk of deficiency. Patients who are less engaged with routine monitoring may therefore be seen as poor candidates.²¹

There’s also a structural problem. In many healthcare systems, the psychiatrist prescribes the antipsychotic while the general practitioner manages metabolic health. Each side can assume the other is handling cardiometabolic monitoring, leading to missed interventions. Additionally, antipsychotic-induced weight gain is not always listed as an approved indication for metformin in prescribing formularies, creating administrative hesitation even when the clinical rationale is solid. Updated guidelines now explicitly recommend considering metformin at antipsychotic initiation, but translating guidelines into routine practice is notoriously slow in psychiatric care.

Does Metformin Help Patients Stick With Their Antipsychotics

Weight gain is one of the most common reasons patients stop taking their antipsychotics, and non-adherence in schizophrenia and related conditions is linked to relapse, hospitalization, and worse long-term outcomes. This creates a secondary benefit of metformin that doesn’t show up in metabolic data alone.

A study using insurance claims data found that patients who took metformin alongside their antipsychotic had better adherence to and persistence with their antipsychotic medication at both 6 months and 12 months compared to those not taking metformin. This held true whether metformin was started early (at the same time as the antipsychotic) or added later. The implication is that by reducing or preventing the weight gain that drives patients to stop treatment, metformin indirectly supports psychiatric stability.²²

Individual Variation and the Pharmacogenomics Angle

Not everyone responds to metformin equally. Some patients see substantial metabolic improvements, while others see little change. This variability has prompted interest in whether genetic factors can predict who will benefit most. Research in youth with autism spectrum disorder who experienced medication-induced weight gain found evidence that genetic variation, of the kind already known to influence metformin’s effect in diabetes, may also predict how well metformin works for antipsychotic-related weight gain.²³

This is still preliminary work, and genetic testing to guide metformin prescribing in this context is not part of current clinical practice. But it points toward a future where the decision to add metformin could be informed by a patient’s individual metabolic genetics rather than waiting to see whether they gain weight first. Given that early intervention appears more effective than reactive treatment, identifying high-risk patients before the weight gain starts would be genuinely useful.

Differences Across Antipsychotics

The degree of metabolic disruption varies considerably across individual antipsychotics. Olanzapine and clozapine are consistently identified as the worst offenders for weight gain, while aripiprazole and ziprasidone cause substantially less. This raises the question of whether metformin is equally effective across the board.

The available trials have tested metformin alongside several different antipsychotics. The Schizophrenia Bulletin guideline review included studies involving clozapine, olanzapine, risperidone, and sulpiride, and the overall weight-attenuation effect remained consistent across these agents.¹² A trial specifically studying overweight outpatients with schizophrenia or schizoaffective disorder, many of whom were taking various antipsychotics, found that metformin produced about 2 kg more weight loss than placebo along with significant improvements in triglycerides and hemoglobin A1c.²⁴

The practical takeaway is that metformin appears to help across the antipsychotic spectrum, though its absolute benefit is likely largest when paired with the most metabolically disruptive agents. For someone on clozapine or olanzapine, where yearly weight gain can be dramatic, the case for adding metformin is especially strong. For someone on a more weight-neutral antipsychotic who still experiences metabolic changes, metformin remains a reasonable option but the expected gains will be more modest.