GLP-1 Drugs Were Linked to Lower Mortality in Serious Mental Illness. This Is Not Yet Proof.
A 1.5-million-person study linked GLP-1 initiation to lower mortality and cardiovascular events in serious mental illness. The signal is strong, but it is still observational.
People with schizophrenia, bipolar disorder, or major depression are often told to watch their weight and blood sugar. That advice can sound almost trivial beside the actual risk: serious mental illness is tied to a large, persistent gap in life expectancy, and cardiovascular disease accounts for much of it.
A new study in JAMA Psychiatry asks whether GLP-1 drugs might help close that gap. The result is hard to ignore. Among adults with serious mental illness, those who started a GLP-1 receptor agonist had lower mortality than matched patients who started an SGLT2 inhibitor. The difference appeared within one year and persisted in longer follow-up.
The catch is just as important as the headline. This was a huge health-record analysis, not a randomized trial. It found an association. It cannot tell us that prescribing a GLP-1 drug caused the survival difference.
What the new study found
Researchers at the University of Toronto and SUNY Upstate Medical University used the TriNetX Analytics Network, a multinational collection of electronic health records. They built a target-trial emulation, an observational design that tries to reproduce some features of a randomized trial when randomization is not available.
The comparison was deliberately tougher than GLP-1 treatment versus no treatment. The researchers compared new users of GLP-1 receptor agonists with new users of SGLT2 inhibitors. Both drug classes are used in metabolic care, and both have cardiovascular or kidney benefits in appropriate patients. Using an active comparator helps reduce the chance that the result simply reflects the difference between receiving modern medical care and receiving none.
For the main four-year analysis, the team propensity-matched 1,528,230 adults into 764,115 pairs. That included 195,184 matched pairs with serious mental illness and 568,931 pairs without it. The psychiatric cohort included people with major depressive disorder, bipolar disorder, or schizophrenia.
Among the matched participants with serious mental illness, 4.91% of GLP-1 initiators died during the four-year analysis, compared with 6.45% of SGLT2 inhibitor initiators. The hazard ratio was 0.76, with a 95% confidence interval of 0.74 to 0.78. The absolute risk difference was 1.54 percentage points.
The one-year comparison was also striking. Mortality was 1.46% among GLP-1 initiators and 2.84% among SGLT2 inhibitor initiators. That produced a risk ratio of 0.52 and an absolute difference of 1.38 percentage points.
Among participants who had both serious mental illness and type 2 diabetes, semaglutide initiation was associated with a lower risk of three-point major adverse cardiovascular events than SGLT2 inhibitor initiation. The hazard ratio was 0.77. Other outcomes, including heart attack, stroke, heart failure, and coronary bypass surgery, also favored semaglutide in the reported analysis.
Exploratory analyses extending to 10 years found lower mortality associated with semaglutide initiation in major depression, bipolar disorder, and schizophrenia. Those estimates are interesting, but exploratory results sit lower in the evidence hierarchy than a prespecified primary endpoint.
Why this population needs its own evidence
The cardiovascular burden in serious mental illness is not a side issue. An American College of Cardiology summary of a 2024 state-of-the-art review reports that people with schizophrenia or bipolar disorder die, on average, 15 to 20 years earlier than the general population. Cardiovascular disease is often the cause.
No single mechanism explains that gap. Smoking, physical inactivity, diet, poverty, chronic stress, inflammation, and fragmented access to care all contribute. Some antipsychotics can increase appetite, weight, insulin resistance, and lipid abnormalities. People with serious mental illness are also less likely to receive standard cardiovascular procedures and prevention medications after heart disease develops.
That context changes how the new data should be read. A drug that improves weight, glucose control, and cardiovascular risk may deliver more absolute benefit in a group starting from a higher baseline risk. In the JAMA Psychiatry study, the authors reported greater absolute reductions among people with serious mental illness than among those without it.
It also explains why an obesity-only framing misses the point. For someone taking clozapine or olanzapine, metabolic treatment may be part of protecting continued psychiatric stability. Stopping an effective antipsychotic because of weight gain can carry its own serious consequences. The goal is coordinated care, not trading psychiatric health for metabolic health or vice versa.
This answers a question Oria flagged in June
Earlier randomized trials had already shown that GLP-1 treatment could improve weight and glycemic outcomes in people taking antipsychotics. The HISTORI trial, for example, randomized adults with schizophrenia, prediabetes, and obesity to semaglutide or placebo for 30 weeks. Semaglutide improved HbA1c and body weight without worsening mental health measures.
Those trials were important, but they were not mortality trials. Oria's June review of semaglutide in schizophrenia graded the metabolic evidence highly while deducting for the absence of long-term cardiovascular outcome data in psychiatric populations.
The new TriNetX analysis fills part of that hole. It looks at deaths and cardiovascular events across a very large population. What it does not provide is randomized proof. The right update is not "GLP-1 drugs save lives in psychiatry." It is "the first large comparative outcome signal is now here, and it is strong enough to justify a prospective trial."
Why the comparator matters
SGLT2 inhibitors are not an inert control. They can reduce heart-failure and kidney events in patients with relevant conditions. A lower mortality association versus this class is more informative than a comparison against people who received no modern metabolic drug.
It also makes interpretation harder. Clinicians choose between these classes for reasons that may not appear cleanly in an electronic record. Kidney function, heart failure, obesity severity, previous side effects, insurance coverage, patient preference, and the ability to manage an injection can all influence the prescription. If those factors also affect mortality, some of the observed difference may come from treatment selection rather than the drug itself.
Propensity matching balances measured variables. It cannot balance what the database did not measure well. Prescription records also do not prove that a patient took the medicine consistently for years. Diagnoses can be miscoded, care received outside the network may be missed, and crossover between drug classes complicates long follow-up.
The sample size gives the estimates impressive precision. It does not erase those limitations. A narrow confidence interval around an observational estimate can still describe a biased estimate very precisely.
How this fits with randomized cardiovascular evidence
The broader GLP-1 cardiovascular case is already supported by randomized trials. SELECT enrolled 17,604 adults with overweight or obesity and established cardiovascular disease but without diabetes. Over a mean follow-up of 40 months, major cardiovascular events occurred in 6.5% of participants assigned semaglutide and 8.0% assigned placebo. The hazard ratio was 0.80.
SELECT shows that semaglutide can reduce cardiovascular events in a high-risk population. It does not prove the same effect size in schizophrenia, bipolar disorder, or major depression. People with serious mental illness have different medication burdens, barriers to care, and patterns of baseline risk. That is why the new study is useful and why a dedicated randomized outcomes trial is still needed.
Safety needs psychiatric context too
GLP-1 drugs are not interchangeable wellness products. The current FDA Wegovy label carries a boxed warning about thyroid C-cell tumors observed in rodents and contraindicates semaglutide in people with a personal or family history of medullary thyroid carcinoma or MEN2. Warnings also cover pancreatitis, gallbladder disease, kidney injury from volume depletion, severe gastrointestinal reactions, diabetic retinopathy complications, hypersensitivity, increased heart rate, and aspiration risk around anesthesia or deep sedation.
Several issues deserve extra attention in psychiatric care. Nausea and constipation can compound the gastrointestinal effects of psychotropic drugs. Clozapine already requires careful monitoring, and severe constipation can be dangerous. Semaglutide delays gastric emptying, so clinicians may need to monitor oral medicines that have a narrow therapeutic index or require close clinical observation.
Treatment decisions should involve the prescriber managing the psychiatric condition and the clinician managing metabolic or cardiovascular risk. Rapid weight change, reduced food intake, dehydration, and changes in medication absorption are easier to catch when those teams communicate.
The Oria take
This study moves the conversation beyond weight loss. In a large matched analysis, GLP-1 initiation was associated with lower mortality and fewer cardiovascular events than SGLT2 inhibitor initiation among adults with serious mental illness. The absolute mortality difference was clinically meaningful, and the association appeared across several psychiatric diagnoses.
But the result should change the research agenda before it changes prescribing rules. A pragmatic randomized trial could test whether the association survives randomization, identify which patients benefit most, and measure psychiatric stability, medication adherence, gastrointestinal complications, quality of life, and cardiovascular events together.
For now, the practical message is narrower. Cardiovascular risk in serious mental illness deserves active treatment, not a note buried beneath psychiatric symptoms. GLP-1 therapy may become part of that treatment for eligible patients. The new evidence makes the case more credible, while leaving causation unsettled.
Evidence Grade: B (Promising comparative outcomes evidence). The study used a very large matched cohort, an active comparator, multiple follow-up windows, and sensitivity analyses. The mortality and cardiovascular associations were precise and consistent. The grade remains B because the evidence is observational, exposure came from prescription records, residual confounding is possible, and no randomized cardiovascular outcomes trial has yet confirmed the result in serious mental illness.
Sources
Medical Disclaimer: This article is for informational and educational purposes only. It does not provide medical advice, diagnosis, or treatment. GLP-1 receptor agonists are prescription drugs with contraindications, interactions, and potentially serious adverse effects. People taking psychiatric medication should not start, stop, or change any treatment without guidance from the clinicians managing their psychiatric and medical care. Oria exists to help readers understand peptide and metabolic research, not to replace individualized care.
Related Compounds
Explore Reconstitution Math & Research Reports
Use our interactive U-100 Reconstitution Calculator or browse evidence-graded literature summaries.