Skip to content
This GLP
← Research
Evidence

Serious mental illness: a four-year mortality gap of 1.5 points

GLP-1 drugs beat an active comparator on death rates in a population that dies early. The one-year figure in the same paper is not believable.

Dana Sullivan7 min read
Four-year mortality, serious mental illnesson a GLP-14.91%on an SGLT2 inhibitor6.45%HR 0.76 — a gap of 1.54 percentage points.Both arms were on an effective drug. 195,184 matched pairs.

People with schizophrenia, bipolar disorder and major depression die substantially earlier than everyone else, mostly of cardiovascular disease, and that gap has resisted decades of attention. A large emulation asked whether these drugs narrow it, comparing them against SGLT2 inhibitors [1]. The weight side of the same problem is covered in clozapine weight gain and what came back.

At four years, mortality among participants with serious mental illness was 4.91% on a GLP-1 drug against 6.45% on an SGLT2 inhibitor — a hazard ratio of 0.76 with an absolute risk difference of 1.54 percentage points. Among those who also had type 2 diabetes, semaglutide was associated with lower rates of major cardiovascular events, myocardial infarction, stroke and heart failure.

The comparator deserves emphasis because it cuts both ways. SGLT2 inhibitors are not inert — they reduce mortality in their own right — so a margin over them is a harder thing to achieve than a margin over nothing. That makes the direction more interesting. It also means both groups were under active treatment, which removes some of the usual objection that the treated group is simply the group receiving care.

Read the absolute numbers alongside the ratios, as always. A hazard ratio of 0.76 on a four-year mortality of around 6% is a gap of about one and a half people per hundred. That is a real and meaningful difference in a population with excess mortality, and it is a much smaller thing than 0.76 sounds — the same arithmetic set out in the semaglutide and tirzepatide comparison and in the intracranial pressure meta-analysis.

The authors ask for randomized trials, which is correct and also unlikely to happen soon in this population. In the meantime the useful conclusion is narrow: there is no signal here that these drugs are dangerous in serious mental illness, which matters because psychiatric safety concerns have circulated widely around them, as the effort and reward trial approaches from another direction.

Frequently asked

Do GLP-1 drugs reduce death in serious mental illness?
This emulation found lower four-year mortality against SGLT2 inhibitors — 4.91% versus 6.45% — but it is observational, and the authors call for randomized trials.
Why distrust the one-year result?
It reports a 48% mortality reduction within twelve months of the first prescription, larger than any randomized trial of this class has produced over much longer periods. That pattern suggests the treated group was healthier when treatment started.
What was the comparison drug?
SGLT2 inhibitors, which reduce mortality themselves. This is not a comparison against placebo or against no treatment.

Sources

  1. [1] McIntyre RS, Zhang-James Y, Kwan ATH (2026). Glucagon-Like Peptide 1 Receptor Agonists, Mortality, and Cardiovascular Outcomes in Serious Mental Illness JAMA Psychiatry. PMID 42647034

Where to get it

Best GLP-1 injections

Every injectable seller we can verify, with the price each one publishes and an honest read of what the trials measured.

Compare providers →

More in Evidence