Skip to content
This GLP
← Research
Evidence

Crohn's disease: hospitalization halved, mortality disowned

In 9,766 matched pairs, hospitalization at one year was 10.0% against 24.7%. The same study reports a six-fold mortality difference its authors say is probably not real.

Ruth Alvarez6 min read
One year, 9,766 matched pairson a GLP-1not on onehospitalization10%24.7%ED visit18.9%25.3%TNF inhibitor use7.1%10.9%Corticosteroid use at one year: unchanged.

Obesity affects between 15% and 40% of adults with Crohn’s disease, and a 2026 cohort study asked what happens to them on a GLP-1 drug [1]. It drew on 162 organizations, matched 9,766 pairs on more than thirty covariates, and followed them for one year and again for five.

The utilization results are substantial and consistent. Hospitalization at one year was 10.0% against 24.7% (RR 0.40, 95% CI 0.38–0.43), emergency visits 18.9% against 25.3% (RR 0.75, 95% CI 0.71–0.79), and both associations held at five years. These are the outcomes that show up on a bill as well as in a chart, which makes them worth more attention than they usually get beside the weight figures.

The study also reports one-year mortality at 0.7% against 4.2% (RR 0.17, 95% CI 0.13–0.22), and then tells you not to believe it. Its authors write that the mortality difference is unlikely to reflect a true causal effect and should be treated as hypothesis-generating, and that healthy-user bias cannot be excluded. This is the second paper in this library to publish a dramatic mortality number and disown it in the same breath; the first is covered in the authors do not believe their own hazard ratio, and the reasoning there applies here unchanged.

The TNF inhibitor result needs the same discipline. Use was 7.1% against 10.9% (RR 0.65, 95% CI 0.60–0.72), which reads like less need for biologic escalation — and the authors state plainly that it should not be read that way, because their database cannot distinguish starting a TNF inhibitor from continuing one. A lower rate of recorded use is not the same fact as a lower rate of disease progression.

The most informative number is the one that did not move. Corticosteroid use at one year was the same in both arms, RR 1.00 with a confidence interval of 0.96 to 1.03. If a healthy-user effect were driving everything, it should have dragged that outcome too. It did not, which is a point in the findings’ favor, and by five years steroid use was modestly lower at RR 0.91.

None of this makes a GLP-1 a Crohn’s treatment, and the paper does not claim it. What it supports is metabolic safety in a group whose disease is often cited as a reason for caution, with a call for prospective work. That caution is mostly about motility, and the evidence behind it is thinner than the warnings suggest — the gastroparesis signal is worth reading before assuming it settles the question. Anyone with inflammatory bowel disease weighing one of these drugs is in a conversation with a gastroenterologist, not an intake form — the same conclusion reached from a different direction in the ulcerative colitis evidence.

Frequently asked

Do GLP-1 drugs help Crohn's disease?
This study does not show that, and does not claim it. It found lower hospitalization and emergency use among matched patients with Crohn's and obesity, and its authors present the work as supporting metabolic safety rather than treatment effect.
Why dismiss the mortality result?
Because the authors do. They write that the difference is unlikely to be causal, should be treated as hypothesis-generating, and that healthy-user bias cannot be excluded — people well enough to be started on a GLP-1 differ from those who are not.
Did fewer patients need biologics?
That cannot be concluded. Recorded TNF inhibitor use was lower, but the authors state their data cannot distinguish starting one from continuing one, so a lower recorded rate is not evidence of less escalation.

Sources

  1. [1] Ganju N, Ssebambulidde K, Gupta S, Adidam S, et al. (2026). GLP-1 receptor agonists and clinical outcomes in adults with Crohn's disease and obesity: a propensity score-matched real-world cohort study Scientific Reports. PMID 42321339

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