GLP-1 Evidence

Not medical advice. A record of published research and our assessments of it. The limits

Open questions

What the research has not yet answered, why not, and which studies come closest. A question stays open until evidence capable of answering it exists.

Q01Open

Are the anti-inflammatory effects of GLP-1 receptor agonists independent of weight loss, and do they occur in people without excess adiposity?

Domains: inflammation. Best current evidence is biomarker-level in obesity (SELECT hs-CRP; liraglutide vs diet). No normal-weight data.

What the record shows on GLP-1 drugs and inflammation apart from weight loss, as of Oct 6, 2026 — a written reading of this question.

Closest evidence so far

Q02Open

Is chronic GLP-1 receptor agonist use beneficial, neutral or harmful in metabolically healthy people?

Domains: core_question, metabolic, aging. No trial exists. Harms documented in treated populations (GI, gallbladder, lean mass) would apply.

Closest evidence so far

Q03Open

What is the risk-benefit balance in normal-weight adults aged 55-75?

Domains: core_question, adverse_effects, lean_mass, bone. Unanswerable with current evidence; this observatory tracks whether any direct evidence appears.

Closest evidence so far

Q04Partially answered

Are the cardiovascular effects independent of obesity?

Domains: cardiovascular. Within obesity, benefit is largely independent of weight loss (SELECT analyses). Outside obesity, untested.

Closest evidence so far

Q05Open

Do GLP-1 receptor agonists affect biological aging in humans?

Domains: aging. Mechanistic reviews only; no human trial with an aging endpoint.

Closest evidence so far

Q06Open

Is there clinically meaningful neuroprotection (dementia, Parkinson's) in humans?

Domains: dementia, alzheimers, parkinsons, neuroinflammation. Phase 3 trials in established AD and PD were negative; observational prevention signals remain untested by RCT.

Closest evidence so far

Q07Open

What happens to lean mass and muscle function in normal-weight users?

Domains: lean_mass, muscle, bone. No data in normal-weight users; lean mass falls with weight loss in every studied population.

Closest evidence so far

Q08Open

Is there an effective dose for non-weight indications below weight-loss doses?

Domains: metabolic, addiction, cardiovascular. FLOW used 1.0 mg semaglutide; the AUD trial used 0.5-1.0 mg; CVOTs in T2D used lower doses than obesity trials. No dose-ranging data for non-weight endpoints in non-obese people.

Closest evidence so far

Q09Open

What happens with use measured in decades rather than years?

Domains: adverse_effects, cancer, endocrine, bone. Longest follow-up in RCTs is about 5 years (REWIND). Thyroid cancer and other slow outcomes remain uncertain.

Closest evidence so far

Q10Open

How large are rare irreversible harms (NAION, pancreatitis, obstruction) and are they dose-related?

Domains: ophthalmologic, gastrointestinal, adverse_effects.

Closest evidence so far

Q11Open

Does adding glucagon-receptor agonism (retatrutide, survodutide, mazdutide) change the benefit or harm profile relative to GLP-1 agonism alone, and would any of that carry over to normal-weight adults?

Domains: adverse_effects, cardiovascular, lean_mass, body_composition. Under c001-outcomes v1.1 the glucagon co-agonists form a separate subgroup and are not pooled with GLP-1 RAs by default. Every phase 3 trial published so far required obesity, overweight or type 2 diabetes; TRANSCEND-T2D-1 allowed BMI down to 23 but reported no lower-BMI subgroup. Signals to watch: dysesthesia and hypotension (TRIUMPH-2), the class-wide heart-rate rise, and the fat versus lean mass split. TRIUMPH-1 and TRIUMPH-4 are announced but unpublished and are not evidence here until they are.

Closest evidence so far