Semaglutide linked to 21% lower psychiatric hospitalization risk in bipolar disorder patients
GLP-1 drugs keep showing up in unexpected places — from bipolar disorder wards to dental offices to Alzheimer's labs. This week's research paints a broader picture of what these metabolic drugs might (and might not) be doing beyond blood sugar and the scale.
🧠 Semaglutide linked to lower psychiatric hospitalization in bipolar disorder — but it's not a class effect
A Swedish nationwide registry study followed 14,694 people with bipolar disorder who also used antidiabetic medications between 2009 and 2024. Here's what stood out:
- Semaglutide was associated with a 21% lower risk of psychiatric hospitalization (adjusted hazard ratio 0.79, 95% CI 0.69–0.91) compared to periods when the same individuals were not on a GLP-1 drug — a within-person comparison design that helps reduce confounding
- Semaglutide was also linked to a 17% lower risk of bipolar relapse specifically (aHR 0.83, 95% CI 0.69–0.99)
- Liraglutide and dulaglutide showed no significant association with reduced hospitalization — suggesting this may not be a property shared by all GLP-1 drugs
Why it matters: This observational study can't confirm causation, and the mechanism is unknown. But the within-person design (each person serves as their own control across time periods) is more rigorous than simple group comparisons. If replicated in a randomized trial, it could point to a meaningful psychiatric benefit of semaglutide specifically — not GLP-1 drugs as a whole.
Key Findings
🫀 GLP-1 drugs may improve heart failure outcomes even without diabetes
A meta-analysis of 11 studies looked at GLP-1 drugs (including tirzepatide) in patients with heart failure with preserved or mildly reduced ejection fraction (HFpEF/HFmrEF — types of heart failure where the heart pumps normally but doesn't relax well):
- Pooled randomized trial data showed a 32% lower risk of heart failure hospitalization (hazard ratio 0.68, 95% CI 0.52–0.88)
- The combined risk of cardiovascular death or heart failure events was also lower (HR 0.76, 95% CI 0.65–0.87)
- Critically, a meta-regression found no evidence that having type 2 diabetes changed the size of the benefit — all interaction p-values were >0.05
🧬 An oral GLP-1 drug activates a brain-protective metabolic loop — in mice
A Cell Metabolism study tested OHP2, an experimental oral GLP-1 drug designed to cross the blood-brain barrier, in mouse models of Alzheimer's disease:
- OHP2 primarily activated GLP-1 receptors on astrocytes (brain support cells), boosting their aerobic glycolysis (energy production from glucose) and lactate release
- Neurons took up that astrocyte-derived lactate, which increased a specific histone modification (H3K9la — a chemical tag on DNA-packaging proteins) that then promoted lipid transfer back to astrocytes
- This bidirectional metabolic loop between astrocytes and neurons was associated with reduced metabolic disturbances in the Alzheimer's mouse model
This is animal research, and whether OHP2 or this mechanism translates to humans is not yet known.
🏋️ Exercise — not liraglutide — drove vascular improvements in a weight-loss trial
A secondary analysis of the S-LiTE trial randomized 130 adults with obesity (after diet-induced weight loss) to exercise and/or liraglutide for 52 weeks:
- Exercise alone or combined with liraglutide reduced carotid intima-media thickness (a marker of artery wall thickening linked to atherosclerosis) and lowered inflammatory cytokines including interleukin-6 and interferon-γ
- The combination of exercise + liraglutide also improved endothelial function markers (sICAM-1, sVCAM-1, tPA — proteins that reflect how well blood vessel linings are working)
- Liraglutide alone showed none of these vascular improvements
⚖️ Bariatric surgery still outpaces GLP-1 drugs for weight loss at 1–3 years
A real-world study of 44,025 patients at two urban health systems compared weight loss across treatments over up to 3 years:
- Sleeve gastrectomy: ~24% total weight loss at 1 year, ~22% at 3 years
- Gastric bypass: ~30% at 1 year, ~28% at 3 years
- Tirzepatide (intention-to-treat): ~9% at 1 year
- Semaglutide (intention-to-treat): ~5% at 1 year, ~7% at 3 years
- Even with 1 year of continuous GLP-1 use (per-protocol), tirzepatide reached ~12% and semaglutide ~9% — still well below surgery
This is a retrospective study, so patient selection differences between groups can't be fully ruled out.
👃 GLP-1 drugs linked to higher risk of smell and taste disturbances
A large retrospective cohort study used electronic health records from 876,948 matched patients with type 2 diabetes (438,474 on GLP-1 drugs, 438,474 on other diabetes medications):
- GLP-1 drug users had a 48% higher overall risk of smell and taste disturbances (hazard ratio 1.48, 95% CI 1.37–1.61)
- Smell disturbances specifically: HR 1.81 (95% CI 1.58–2.07)
- Taste disturbances specifically: HR 1.52 (95% CI 1.35–1.71)
- The signal was consistent across the full 2-year follow-up period
This is an observational study using diagnostic codes — it can identify associations but not confirm that GLP-1 drugs directly cause these effects.
💊 A pill version of a GLP-1 drug produced up to 9.4% weight loss in Chinese adults without diabetes
A phase 2 randomized controlled trial across 29 centers in China tested HRS-7535 — an oral small-molecule GLP-1 drug — in 235 adults with obesity (BMI 28–40) but no diabetes:
- At 26 weeks, the 180 mg/day dose produced a placebo-adjusted weight loss of -6.87% (vs. -2.50% for placebo, p<0.0001)
- The 60 mg dose reached -4.60% placebo-adjusted loss (p=0.0006); the 30 mg dose did not reach statistical significance (-0.49%, p=0.71)
- Gastrointestinal side effects were most common and mostly mild-to-moderate, peaking during dose escalation
This is a phase 2 trial — larger confirmatory studies are needed.
Implications
This week's research collectively suggests GLP-1 drugs are doing something beyond managing blood sugar and body weight — touching psychiatric risk, liver disease, heart failure, and even brain metabolism in animal models. But the pattern is also clear: benefits vary by drug, by patient, and by what's being measured, and exercise still does things these medications apparently can't. As access and eligibility questions intensify (nearly 77% of Medicare patients may qualify under current or near-future indications), the field is shifting toward asking not just 'does it work?' but 'for whom, at what dose, and combined with what?'
Studies in this issue
Primary sources used for this newsletter.
- Use of Glucagon-Like Peptide 1 Receptor Agonists and the Risk of Hospitalization in Bipolar Disorder from 2009 to 2024main storyActa psychiatrica Scandinavica2026-06-22PMID 42324672
- Smell and Taste Problems in People Using Glucagon-Like Peptide-1 Receptor Agonist Medicineskey findingJAMA otolaryngology-- head & neck surgery2026-06-25PMID 42348217
- Real-World Effects of Semaglutide and Tirzepatide Compared to Weight-Loss Surgerykey findingObesity (Silver Spring, Md.)2026-06-25PMID 42345739
- Oral GLP-1 treatment may help brain support cells and neurons share energy and fats, providing brain protectionkey findingCell metabolism2026-06-22PMID 42330959
- Glucagon-Like Peptide-1 Receptor Agonists Improve Heart Health in Heart Failure with Normal Pumping, Regardless of Diabeteskey findingCJC open2026-06-22PMID 42327438
- HRS-7535, an oral drug activating the GLP-1 receptor, tested in Chinese adults with obesity but no diabetes in a controlled clinical trialkey findingNature communications2026-06-22PMID 42331800
- Exercise and liraglutide's effects on blood vessel health and inflammation during weight loss maintenancekey findingNature metabolism2026-06-24PMID 42342869
Continue reading
All GLP-1 Therapies issuesGet the next GLP-1 Therapies issue
Seven papers, once a week. Free.