Metabolic Medicines

A metabolic medicine, in the sense this brief uses, is a drug that acts on the incretin and energy-balance signalling system — GLP-1, GIP, glucagon, amylin — and produces weight loss large enough to change the course of other diseases. The class matters because of an arithmetic with no precedent in pharmacology: a once-weekly injection that removes a fifth of body weight, where the previous best non-surgical drugs removed five per cent and were withdrawn for cardiac and psychiatric harms.

This brief is a synthesis, and the joint question falls between three others on the map. Longevity Therapies owns the geroscience question and the replication infrastructure that scores it. Precision Medicine owns stratification and the failures of targeting, and supplies the lesson this brief borrows: a biomarker that moves is not an outcome that matters. Human Flourishing owns the measurement of subjective benefit, the axis on which a drug that makes people lighter and less hungry is most often defended and least often measured. The question none of them owns is the one here: when one drug class moves a common upstream variable and thereby touches cardiology, nephrology, hepatology, sleep medicine, orthopaedics and psychiatry at once, how do we work out what it is actually worth, and to whom?

Frontier The gap between what has been measured and what is being sold is now the most important fact about the field. Cardiovascular benefit in established disease is measured. Kidney benefit in diabetic kidney disease is measured. Liver histology in MASH is measured. Addiction, dementia prevention and geroprotection are not measured, are being marketed, and in the one large case that has read out — oral semaglutide in early Alzheimer’s disease — the answer was negative in roughly 3,800 randomised patients.

Established A note on sourcing. This brief was commissioned in September 2026 from the Institute’s research base. Reading-list entries without links are cited from the bibliographic record rather than re-fetched, and claims are dated no later than early 2026 unless carried by a linked source.

Established Weight loss: three anchor trials and one head-to-head. STEP 1 (semaglutide 2.4 mg, 68 weeks, n = 1,961) gave −14.9 per cent against −2.4 on placebo, a difference near 12.4 percentage points (95 per cent CI roughly −13.4 to −11.5), with 86 per cent of the drug arm losing at least 5 per cent against 32 per cent. SURMOUNT-1 (tirzepatide, 72 weeks, n = 2,539) gave −15.0, −19.5 and −20.9 per cent at 5, 10 and 15 mg against −3.1. SURMOUNT-5 randomised the two drugs against each other in 751 adults for 72 weeks: −20.2 against −13.7 per cent, a difference near 6.6 points (95 per cent CI roughly −8.2 to −5.0). The class ranking was established by randomisation, not cross-trial arithmetic.

Established Cardiovascular outcomes: one placebo-controlled trial carries the indication, and its absolute effect is modest. SELECT randomised 17,604 adults with overweight or obesity and established cardiovascular disease but without diabetes. Major adverse cardiovascular events: 6.5 per cent against 8.0 per cent over a median 39.8 months, hazard ratio 0.80 (95 per cent CI 0.72 to 0.90). A fifth in relative terms, 1.5 percentage points in absolute terms, roughly 67 patients treated for three years per event prevented. Both numbers are true; one of them is usually quoted.

Established Kidney and heart-failure outcomes. FLOW randomised 3,533 adults with type 2 diabetes and chronic kidney disease to semaglutide 1.0 mg and stopped early for efficacy: hazard ratio 0.76 on the composite kidney endpoint (95 per cent CI 0.66 to 0.88), the strongest non-cardiac result in the class and in a population defined by albuminuria rather than weight. In heart failure with preserved ejection fraction, STEP-HFpEF (n = 529) improved the Kansas City questionnaire by about 7.8 points (95 per cent CI 4.8 to 10.9), while SUMMIT (n = 731) gave a hazard ratio of 0.62 for cardiovascular death or worsening heart failure (95 per cent CI 0.41 to 0.95) on a few dozen events. Symptom result solid; event result promising and imprecise.

Established Sleep apnoea and knee osteoarthritis: two indications won on mechanism-free endpoints. SURMOUNT-OSA cut the apnoea-hypopnoea index by about 25 events per hour against 5 on placebo without positive-airway-pressure therapy, and about 29 against 5 with it; tirzepatide was approved for moderate-to-severe obstructive sleep apnoea with obesity in December 2024. STEP 9 (n = 407) improved WOMAC knee pain by about 14 points more than placebo (95 per cent CI roughly 20.3 to 7.8).

Established Liver disease: a histological endpoint met, on an accelerated pathway. ESSENCE part 1 assessed semaglutide 2.4 mg at 72 weeks in about 800 patients with MASH and stage 2–3 fibrosis: steatohepatitis resolution without worsening fibrosis in about 63 per cent against 34 per cent, fibrosis improvement without worsening steatohepatitis in about 37 per cent against 22 per cent. Accelerated approval followed in August 2025, so the outcome that matters — cirrhosis, transplant, death — remains conditional on a confirmatory trial. Resmetirom, approved in March 2024 by the same route, gave resolution in about 26–30 per cent against 10 per cent.

Frontier The head-to-head cardiovascular trial did not show superiority. SURPASS-CVOT compared tirzepatide against dulaglutide — an active comparator, not placebo — in roughly 13,000 adults with type 2 diabetes and cardiovascular disease. Non-inferiority was met; the point estimate for major adverse cardiovascular events sat near 0.92 with an upper confidence bound at or just above 1.0, so superiority was not established. The dual agonist’s extra weight loss did not produce a demonstrable extra cardiovascular benefit over an older drug in this population. That negative sits inside a commercially successful programme and is rarely quoted.

Established The dementia trials failed. The evoke and evoke+ programme randomised roughly 3,800 patients with early Alzheimer’s disease to oral semaglutide 14 mg or placebo. In late 2025 the sponsor announced that neither trial met its primary endpoint on the clinical dementia rating sum of boxes. Years of observational and mechanistic argument pointed the other way. The randomised answer is the one that counts, and this is the single most important negative result in the class.

Frontier Addiction: the claim is everywhere, the randomised evidence is two small trials. A 2025 randomised trial of low-dose semaglutide in alcohol use disorder enrolled 48 people for nine weeks and reported reduced drinks per drinking day and reduced cue-elicited craving, without an effect on drinking days overall. An earlier Danish exenatide trial in 127 patients missed its primary endpoint. Against that sit retrospective health-record analyses that cannot separate the drug from the kind of patient who gets it. No adequately powered phase 3 trial in any substance use disorder has reported, and marketing for addiction is running years ahead of an evidence base one n = 48 study wide.

Established Durability: stopping reverses most of the effect, measured twice. In the STEP 1 extension, participants coming off semaglutide regained about two-thirds of the lost weight within a year, ending roughly 5.6 per cent below baseline against 12.4 points of treatment difference at week 68. SURMOUNT-4 randomised patients after a 36-week tirzepatide lead-in: continuing patients lost a further 5.5 per cent over 52 weeks while those switched to placebo regained about 14 per cent, a difference near 19 points. The drugs treat obesity while taken, as antihypertensives treat blood pressure while taken.

Frontier Real-world persistence is far below trial persistence, and the numbers come from interested parties. Claims analyses published in 2024–2025 report roughly half of patients stopping within a year, one finding about 15 per cent still on therapy at two years (payer and pharmacy-benefit-manager analyses, not peer-reviewed, with a financial interest in lower utilisation). The direction is consistent across sources with opposite incentives; the exact fraction is not.

Frontier Lean mass: the loss is real, the proportion is roughly what dieting produces, and nobody has measured what it does to function. Body-composition substudies report a quarter to two-fifths of the weight lost as fat-free mass — about 39 per cent in the STEP 1 substudy, closer to a quarter in SURMOUNT-1. DXA fat-free mass includes water, glycogen and organ mass, and falls in any substantial weight loss including surgery. The trial that matters — grip strength, gait speed, falls, fractures and independence in older adults, randomised, over years — has not been done. The activin-pathway antibodies developed against this worry, one phase 2b combination reporting over 90 per cent of weight lost as fat, have no functional outcome data at all.

Established Access is a policy artefact, not a manufacturing one, and the prices moved fast. United States list prices near 1,000–1,350 dollars a month in 2024 coexisted with cash-pay channels at roughly 350–500 dollars by 2025 and a federal pricing arrangement announced in November 2025 promising lower cash prices and a Medicare pathway from 2026. Semaglutide was selected in the second cycle of Medicare price negotiation, effective 2027. Frontier The 2026 implementation is the part a September 2026 reader should treat as unsettled; published rulemaking and formulary listings would confirm it, announcements do not.

Frontier How much of the benefit is weight and how much is the drug? SELECT’s event curves separated earlier than the weight curves plausibly explain, and the effect did not track weight lost in post-hoc analysis. Direct vascular anti-inflammatory action, blood-pressure and lipid effects, and visceral-fat loss specifically are all plausible and none is established. If the benefit is weight, any route to the same weight loss should do; if it is the molecule, surgery and incretins are not substitutes.

Frontier Does primary prevention work? Every hard-outcome trial in the class enrolled people with established disease — prior myocardial infarction, diabetic kidney disease, heart failure. Nobody has randomised people with obesity and no cardiovascular disease to a decade of incretin therapy against placebo with mortality as the endpoint. That is increasingly the population being treated, and the evidence for treating it is extrapolation.

Frontier Triple agonism: how much more is there? Retatrutide, a GLP-1/GIP/glucagon triple agonist, produced about 24 per cent mean weight reduction at 48 weeks at 12 mg in a phase 2 trial of 338 patients — the largest figure in the class and still on a rising curve at the end of dosing. Speculative Phase 3 readouts were scheduled for 2026; a reader in September 2026 should treat any specific phase 3 number as unconfirmed until the primary publications appear, and should expect the phase 3 effect to be smaller than phase 2, as it was for every predecessor.

Frontier Oral small molecules change the supply problem, not obviously the efficacy ceiling. Orforglipron, a non-peptide oral GLP-1 agonist, reported phase 3 weight reductions around 11–12 per cent at 72 weeks — below injectable tirzepatide, above older oral agents, and manufacturable by conventional synthesis rather than peptide production. If that holds, the class splits into a high-efficacy injectable tier and a scalable oral tier, and the second tier is the one that reaches middle-income health systems.

Frontier Amylin combinations underperformed their own guidance: cagrilintide-semaglutide reported roughly 23 per cent at 68 weeks against about 16 for semaglutide in the same trial, a real improvement received as a disappointment because the sponsor had signalled 25.

Speculative Geroscience remains a watchlist item, not a finding. SELECT reported nominally fewer deaths from any cause (hazard ratio near 0.81, 95 per cent CI roughly 0.71 to 0.93) in a trial neither designed nor alpha-protected for that endpoint. The National Institute on Aging’s Interventions Testing Program — the blinded three-site mouse lifespan facility described in Longevity Therapies — has produced male-biased lifespan extension for acarbose and canagliflozin and no incretin result. Until it does, “GLP-1 drugs slow ageing” is a hypothesis with a marketing budget.

Established Peptide manufacture is the physical constraint, and it is fill-finish more than synthesis. A weekly injectable for a hundred million people is billions of sterile presentations a year. The two dominant manufacturers responded with multi-billion-dollar plant programmes and, in one case, the acquisition of a contract manufacturer largely for its sterile filling capacity. The bottleneck is aseptic lines and the years they take to qualify.

Frontier An oral small molecule removes most of that constraint and creates a different one. Conventional synthesis and tabletting scale far more cheaply than peptides and cold chain, which is what makes the oral tier the plausible route to middle-income countries. The new constraints are active-ingredient supply concentrated in few chemical producers, and bioavailability that currently requires strict fasting conditions.

Established Clinical capacity, not drug supply, limits appropriate use. Dose titration, side-effect management, gallbladder screening, contraception counselling and — if the lean-mass worry is taken seriously — resistance training advice add up to a primary-care workload in a specialty most systems do not have. Telehealth filled the gap at an unmeasured quality.

Frontier There is no validated endpoint between weight and death. Regulators accept percentage weight change for approval and hard events for outcome claims; nothing in between — function, disability, freedom from polypharmacy — is standardised enough to power a trial on. It is the same gap Longevity Therapies identifies for ageing, and it is why each beyond-obesity indication costs a full outcome trial.

Established The class depends on receptor pharmacology worked out over four decades and on nothing exotic. GLP-1 was identified in the 1980s and the long-acting analogue problem solved by fatty-acid acylation. No enabling platform sits upstream, which explains the speed: the science was ready long before the indication was.

Frontier It depends on body-composition measurement unfit for the question being asked of it. DXA-derived lean mass cannot distinguish contractile muscle from water and organ mass. D3-creatine dilution and biopsy methods exist and are not in routine trial use, so resolving the muscle question requires adopting a better instrument first.

Speculative The beyond-obesity claims depend on neurobiology nobody has pinned down. If incretin signalling in the brain modulates reward, the addiction and mood claims have a mechanism; receptor distribution is known, the human causal chain is not. Bioelectric Medicine covers the adjacent case where a plausible neuro-immune mechanism took two decades to produce one randomised result.

Frontier The decisive experiment this brief is waiting on is a randomised trial in adults over 65 with obesity, powered for physical function, falls and fractures rather than weight, comparing incretin therapy alone against incretin therapy plus supervised resistance training and protein targets, over at least three years. It tests the one harm that could invert the risk-benefit calculation in the fastest-growing treated population, and no existing trial can answer it: the outcome trials enrolled younger patients and measured body composition by DXA in substudies of a few hundred. Nobody has scheduled or funded it, and until it reports every statement about incretins and sarcopenia is inference from proportions of fat-free mass.

Frontier Second: a primary-prevention outcome trial. Randomise adults with obesity and no established cardiovascular disease to an incretin or placebo and follow to death and major events. It is the only way to tell health systems whether the largest group they treat gets the benefit measured in a much sicker one.

Frontier Third: a maintenance-strategy trial. After a year at full dose, randomise to continuous full dose, reduced dose, intermittent dosing or structured withdrawal with support, with weight and events at three years. SURMOUNT-4 and the STEP 1 extension show that stopping fails; they do not show that lifetime full-dose therapy is the only alternative, and that difference is the difference between a payable and an unpayable bill.

Frontier Fourth: a properly powered randomised trial in a substance use disorder. Several hundred patients, six months, an accepted endpoint such as percentage of heavy drinking days, in people without obesity so weight is not the mediator. It is cheap by the standards of this class, and that it has not been run while the claim is marketed is this brief’s clearest case of promotion outrunning evidence.

Speculative Fifth, and only after the others: an Interventions Testing Program lifespan run on an incretin. A small fraction of the cost of any of the above, in the one blinded three-site replication facility that exists for lifespan claims, settling whether the geroscience story has an animal foundation before it is sold further.

Established The device, not the molecule, is the manufacturing product. A once-weekly single-use pen is an injection-moulded assembly with a glass cartridge, a needle, a dose-setting mechanism and a cold chain; scaling it to hundreds of millions of units a year is the sterile-manufacturing and logistics problem vaccine programmes know, which is why capacity announcements are measured in billions of dollars and years.

Frontier Oral delivery is an engineering achievement more than a pharmacological one. Peptide oral formulations rely on permeation enhancers and deliver a small fraction of the dose; the non-peptide small molecules avoid this entirely. The engineering consequence is a two-order-of-magnitude difference in cost per patient-year between an injectable peptide and a synthesised tablet, which is the whole of the access question.

Speculative Long-acting depots have no clinical data in obesity. A six-month implant would change the shape of the persistence problem; nothing published shows one works yet.

Established Bariatric surgery is the comparator nobody runs head-to-head against. Sleeve gastrectomy and gastric bypass produce 25–30 per cent total weight loss sustained for years, with observational mortality benefit over decades. Tirzepatide’s 20.9 per cent brought drugs within sight of surgery, and no adequately powered randomised comparison with hard outcomes exists.

Established Cell therapy for diabetes is the adjacent technology with the most convincing recent data. Stem-cell-derived islet replacement produced insulin independence in a proportion of type 1 diabetes recipients in sponsor-reported results, at the cost of immunosuppression — a useful contrast with an indefinite weekly injection. Lab-Grown Organs and Whole Organ Regeneration own that thread.

Frontier The microbiome route to obesity is the cautionary neighbour. A pooled reanalysis found the gut-community-to-obesity association far weaker than the literature implied, with classification near chance; the flagship intervention — a pasteurised Akkermansia muciniphila supplement in 32 volunteers — reported marker changes and about half a kilogram of difference. Microbiome Engineering carries it in detail: the same shape of claim, three orders of magnitude less effect, sold in the same shops.

Frontier Nucleic-acid cardiometabolic drugs converge on the same patients from another direction. Twice-yearly small interfering RNA against PCSK9 and antisense agents against lipoprotein(a) address the lipid half of cardiometabolic risk at dosing intervals incretins cannot match; RNA Medicines owns the platform. The combination as a single regimen has not been tested as a strategy.

Established Obesity is an indication regulators accept and a benefit payers are allowed to refuse. That asymmetry is the institutional core of the field. United States Medicare law excluded drugs used for weight loss; the 2024 proposal to reinterpret it was not finalised, and coverage arrived instead through cardiovascular and sleep-apnoea indications and a negotiated pricing arrangement. A drug approved for a disease a payer is statutorily barred from covering it for is an institutional fact, not a scientific one.

Established Indication creep is how coverage actually expands. Weight, then cardiovascular risk reduction, then sleep apnoea, then kidney disease, then MASH: a sequence of reimbursable diseases, each won with its own trial. It is a rational response to payer rules with a specific distortion — the trials that get run unlock a reimbursement code rather than answering the largest clinical uncertainty.

Frontier National systems are rationing by eligibility criteria and by waiting. Fixed-budget systems have restricted prescribing to narrow body-mass-index and comorbidity bands, with specialist gatekeeping and multi-year phase-ins, producing a two-tier pattern inside single-payer systems while cash-pay and telehealth channels serve those who can pay. Whether that is rationing or triage depends on what the marginal patient gains, which returns to the unmeasured function endpoint.

Frontier Price competition is arriving through patent expiry rather than policy. Semaglutide’s protection lapses on different dates in different jurisdictions, with Canada, Brazil, India and China identified as early markets for generic entry from 2026. Generic injectable semaglutide at a fraction of the originator price would change access faster than any reimbursement negotiation. A September 2026 reader should check actual launches rather than announcements.

Established The evidence infrastructure is entirely sponsor-owned. Every pivotal trial named here was designed, funded and analysed by the manufacturer of the drug it tested — normal, and the reason the questions in the experiments section are unfunded: none would increase sales and two could reduce them.

Established The stigma argument cuts both ways and both directions are evidenced. Effective pharmacotherapy supports the claim that obesity is physiological rather than a failure of will. It also expands the population defined as needing treatment, and the same drugs are used cosmetically at doses and body-mass ranges no trial studied.

Frontier Adolescent use is where the evidence thins fastest. Adolescent trials show weight effects comparable to adults over about a year. Nothing is known about someone who begins hormonal appetite suppression at 14 and continues for fifty years. Prescribing here is a bet on extrapolation and should be labelled as one.

Frontier Eating-disorder risk is plausible, hard to measure and under-monitored. An appetite suppressant is being administered into a population containing undiagnosed restrictive and binge-eating disorders; trials excluded most such patients, telehealth screens inconsistently, and the systems that would detect a population signal are passive reporting systems. Absence of a signal here is weak evidence.

Established Global distribution is inverted relative to need. Obesity and diabetes prevalence rises fastest in middle-income countries while supply has gone to high-income markets with cash-paying patients. International essential-medicines processes have engaged with the class for diabetes with cardiovascular or kidney disease rather than obesity broadly — defensible triage, and an admission that the world cannot afford the broad indication.

Frontier What a life goes better for is the question Human Flourishing owns, and this class makes it concrete. Trials measure weight, events and disease-specific questionnaires. They do not measure whether a person on long-term appetite suppression enjoys eating, socialises, or reports a better life at five years. The instruments exist; the trials did not use them; and the resulting evidence base is unusually silent about the experience of taking the drug.

Frontier If a large fraction of adults in rich countries take an appetite-suppressing drug indefinitely, the first effects are economic rather than medical. Food manufacturers, restaurants, airlines and insurers have published projections of changed consumption (industry analyses with commercial interests in both directions). Measured population-level dietary change attributable to the drugs is thin; the projections are ahead of the data, as they were for every previous consumption shock.

Speculative A cheap oral version would be a public-health instrument with few precedents. The closest analogues are statins and antiretrovirals: molecules whose cost collapsed and whose population effect became a matter of delivery rather than discovery. At generic pricing the binding constraints become diagnosis, prescribing capacity and adherence — the constraints that limit hypertension control today, in systems that manage it badly.

Speculative The demographic argument is weaker than it sounds. Compressing cardiometabolic morbidity shifts health spending later rather than removing it, and what dominates the last years of life without obesity-driven vascular disease is dementia and frailty — where this class has now returned a large negative result.

Handwave The claim that this class ends the obesity epidemic is the handwave. It requires lifetime adherence in a population that stops within a year, universal access in a market that allocates by price, and a supply chain sized for a fifth of the adult population. Each of those is a separate unsolved problem, and the argument usually treats all three as logistics.

These horizons track evidence and access milestones rather than molecules, because the molecules already exist and the uncertainty is what will be proven about them and who will get them.

• 10 yr: Frontier Generic injectable semaglutide in several large middle-income markets at prices an order of magnitude below originator levels; triple-agonist phase 3 results published and, if the pattern holds, smaller than phase 2; a confirmatory hard-outcome trial in MASH reading out; oral agents established as the volume product. Speculative A function-and-fracture trial in older adults, if anyone funds it.
• 25 yr: Speculative Cohorts who began therapy in adolescence report; the maintenance question — continuous, intermittent or withdrawal — is settled empirically rather than by default; incretin-plus-lipid-agent regimens become standard in high-income systems.
• 50 yr: Speculative Obesity is managed like hypertension: cheap, generic, chronically under-treated in exactly the populations with the most disease. Whether that produced a population mortality shift is answerable by then and not now.
• 100 / 250+ yr: Handwave Any claim at this range — permanent pharmacological energy-balance control, or its becoming unnecessary because food environments changed — is a claim about institutions and culture on a timescale where neither is predictable.

• Depends on Less than most briefs in this category: the pharmacology needed no enabling platform, which is why the field moved so fast. What it does depend on is the endpoint problem owned by Longevity Therapies and the stratification lessons of Precision Medicine, whose central negative result is that responder prediction is harder than the marketing of responder prediction. Neither blocks an approval here; both block the beyond-obesity claims.
• Requires (not on this map) A payer rule that funds a drug taken for decades for a condition many systems still class as lifestyle; aseptic filling and device assembly at a scale the injectable market has never needed, measured in years of plant qualification rather than chemistry; an oral active-ingredient supply chain that does not run through the same two manufacturers, since the access case rests entirely on the oral tier; the randomised evidence on strength, falls and fractures in older adults that no sponsor has reason to generate; a regulator-accepted endpoint between percentage weight change and death; and enough post-expiry competition to hold prices down in markets that cannot pay originator prices.
• Enables Cheap and oral, the class enables cardiometabolic risk management as one primary-care protocol rather than five specialties; MASH treated before cirrhosis rather than transplanted after; plausibly a fall in dialysis incidence, the most expensive chronic service most systems run; and population-scale trials of combination regimens currently unaffordable because the incretin costs more than the trial.
• Adjacent Microbiome Engineering supplies the null result this field should keep in view; RNA Medicines the twice-yearly lipid agents that will share these patients; Lab-Grown Organs and Whole Organ Regeneration the replacement logic for organs metabolic disease destroys; Exposomics the environmental-cause account a pharmacological fix does not address; and Human Flourishing the measurement the trials omitted.

Established “These drugs make you lose 20 per cent of your body weight.” That was the mean at the top dose of tirzepatide over 72 weeks, in people who stayed on the drug and were supported by trial staff. Real-world persistence analyses find a large fraction stopping within a year, and the STEP 1 extension shows two-thirds of the loss returning within a year of stopping. The correct sentence: the drugs remove about 15–21 per cent of body weight for as long as they are taken and tolerated.

Established “The cardiovascular benefit is enormous.” It is a 20 per cent relative and a 1.5-percentage-point absolute reduction over about three and a half years in people with established cardiovascular disease — roughly 67 patients treated for that period per event prevented. Quoting only the relative figure to people without established disease misstates their expected benefit by a large factor.

Frontier “GLP-1 drugs treat addiction.” The randomised human evidence in alcohol use disorder is a 48-person nine-week trial with a positive secondary-endpoint pattern and a 127-person trial that missed its primary endpoint. The rest is electronic-health-record epidemiology that cannot control for who receives these drugs. The claim is being advertised by prescribing services. It may yet prove true; today it rests on fewer than two hundred randomised patients.

Established “They protect the brain.” Two randomised trials with roughly 3,800 patients in early Alzheimer’s disease failed their primary endpoint. Observational dementia-incidence associations in diabetes cohorts remain, and they now have to explain why the best-powered randomised test came back negative.

Frontier “They destroy your muscle.” The fraction of weight lost as fat-free mass — a quarter to two-fifths depending on trial and method — is in the range produced by dieting and by bariatric surgery, and DXA lean mass is not contractile muscle. What is genuinely unknown is function in older adults over years, which nobody has measured and which the counter-reassurance also cannot cite.

Speculative “They are anti-ageing drugs.” The basis is a nominal all-cause mortality signal in a trial powered for cardiovascular events, plus mechanism. A blinded three-site mouse lifespan facility exists precisely so such claims can be tested cheaply, and it has published no incretin result. Cellular Rejuvenation and Longevity Therapies set out the standard of evidence for a geroprotector; nothing in this class meets it yet.

Frontier “Compounded and research-grade versions are the same drug.” Compounded semaglutide was legal in the United States only while the product was in shortage, and that window closed in 2025. What replaced it — peptides sold as research chemicals without prescription — has no established identity, purity or dose accuracy. This class’s safety record does not transfer to a vial of unknown provenance.

Handwave “Obesity is solved.” The claim requires that a drug taken by a minority of eligible patients, stopped by half of them within a year, priced above what most health systems can pay, and never tested for primary prevention or long-term function, has resolved a condition with environmental and economic determinants. What happened is that obesity became pharmacologically treatable, a large, real and narrower thing.