The blind spot is the metric. A weight-loss programme is normally judged on one number — kilograms, or percentage of body weight lost — because that is the endpoint the pivotal trials used, and that number cannot tell the difference between the fat a patient came to lose and the muscle they did not. Someone can hit the stated target in full while a quarter or more of what came off was lean tissue, and a programme that ends there leaves the year afterwards unaddressed. This article covers what the scale hides, what the trial data actually show about the fat-to-lean split, and what closing the gap would require.
Body composition: what body weight is made of — fat mass, and lean tissue including muscle, bone and body water. Two people at the same weight can have very different compositions, and one person can change composition substantially with no movement on the scale at all.
What does a weight-loss programme actually measure?
Body weight, and figures derived from it. That is the endpoint in the pivotal trials, the basis of BMI, and the number written in the notes at every review.
It is not a careless choice. A scale is cheap, immediate and identical in every clinic, and every published result a programme might be measured against was recorded the same way. The criticism is narrower than calling it the wrong instrument: it answers how much precisely and is silent on what.
The imbalance is visible in the trials themselves. SURMOUNT-1 ran its DXA sub-study on 160 people (Look 2025); STEP 1's equivalent covered 140 and was presented as a conference analysis rather than a peer-reviewed paper (Wilding 2021), against 1,961 randomised in the parent trial (Wilding 2022). Weight was measured on everyone and composition on a subset. Clinical practice inherited that ratio honestly, because it is the one the literature handed over.
What does the scale hide?
The split. Every kilogram lost is a mixture of fat and fat-free mass, and the mixture is not fixed.
Roughly three-quarters of the loss in the SURMOUNT-1 sub-study was fat and about a quarter lean mass (Look 2025) — while a review reading STEP 1's supplementary data puts the semaglutide figure nearer 40% of the kilograms lost (Mechanick 2025), reconstructed from a conference abstract's appendix rather than stated in a primary paper, so it is one review's reading rather than a settled number. Ordinary dieting occupies the same range, at about 14% of weight lost as fat-free mass on standard low-calorie diets and roughly 23% on very-low-calorie ones, varying by sex (Chaston 2007).
Getting lean vs losing weight sets those studies out in full and is the better place to read them side by side. What matters here is that no single figure is quotable: the "a quarter of weight lost is lean tissue" rule of thumb has been criticised as an oversimplification that moves with age, inactivity and exercise (Heymsfield 2014). A number that unstable cannot be a programme's only reported outcome — which is the point this article is making, and it survives whichever figure turns out to be closest.
Is the medication responsible for the muscle?
No. It is an easy thing to get backwards, and getting it backwards leads to the wrong response.
The two arms of that sub-study came out level: about a quarter of the weight lost was lean mass on tirzepatide, and about a quarter on placebo, where the placebo arm was diet alone (Look 2025). A large deficit costs lean tissue whatever produced it.
The practical consequence is the whole reason to get the direction right. If the medicine were the cause, avoiding the medicine would be the response. Because the deficit is the cause, the response is to change what surrounds the deficit — training and protein — and that is something a patient can act on rather than opt out of.
What holds muscle during weight loss?
A resistance-training stimulus, and this is one of the better-evidenced findings in the whole area.
A meta-analysis of 114 trials covering 4,184 people with overweight and obesity found lean mass was statistically unchanged when resistance training accompanied caloric restriction (Lopez 2022). In obese older adults specifically, six randomised trials pooled together found that adding resistance training to caloric restriction prevented an estimated 93.5% of the lean-mass loss otherwise seen — and did so without blunting fat loss or total weight loss (Sardeli 2018). That 93.5% figure comes from an older-adult population and should not be assumed to hold at every age.
The consequence of not doing it is not cosmetic. Sarcopenic obesity — excess adiposity together with low muscle mass or function — was associated with roughly 21% higher relative risk of all-cause mortality when pooled across 23 studies and about 51,000 adults (Zhang 2019). That is an association across studies using varying definitions, not a demonstration that one person's weight-loss programme caused it. It is still the reason lean mass is worth treating as an outcome rather than as a rounding error.
What happens when the programme ends?
The weight returns, and the published trials are unambiguous about it.
In the STEP 1 extension, adults who came off semaglutide after 68 weeks regained about two-thirds of what they had lost within the following year, and most of the cardiometabolic improvements moved back towards baseline over the same period (Wilding 2022). Pooling six cessation trials and 3,236 people, about 60% of the weight lost was back within a year, with the curve decelerating rather than continuing at the same rate (Budini 2026).
Here is the part that is genuinely unknown, and it is worth saying plainly rather than filling in. Nobody has measured what regained weight is made of. The largest discontinuation meta-analysis reports body weight, waist circumference and BMI, and stops there (Berg 2025). The idea that it all comes back as fat is an extrapolation from a different literature, not an established GLP-1 finding. The muscle case does not need that claim; it rests on the composition of the weight lost, which is measured.
What follows is not that stopping is a mistake. It is that stopping without a plan for the year afterwards is the second half of the same blind spot as measuring only the scale during treatment. Our article on what happens when you stop goes through the withdrawal trials in detail.
What would closing the blind spot look like?
Four things, in order of how cheap they are.
Waist, measured the same way each time. Sensitive to fat, insensitive to the muscle that is offsetting it on the scale. Across 78 studies in 14 countries, a boundary of about half your height best discriminated cardiometabolic risk (Browning 2010). It costs a tape measure.
Strength in training. The most immediate feedback a session gives, and worth keeping in its own right — but not a read on muscle mass. In a randomised trial, about 7% weight loss by dieting alone reduced lean mass measurably while measured muscle strength did not change at all (Weiss 2017). Separately, grip strength has been studied in its own right: in a cohort of 139,691 people across 17 countries, lower grip strength was associated with higher all-cause and cardiovascular mortality (Leong 2015). Grip strength is a strength measure, not a measure of muscle mass, and that association is observational.
A body-composition scan, repeated on the same device. Across fifteen bioimpedance devices tested against a four-compartment model, only a third met a strict accuracy standard for a single measurement, but most tracked change over 12 to 16 weeks considerably better (Siedler 2022). The practical reading is that consistency beats precision: the same machine, the same conditions, repeated. At the time of writing, published Singapore prices ran roughly $40–45 for an InBody scan and around $300–310 for a DEXA body-composition scan (ATA Medical price list, July 2026); prices change, so check the provider's own page. Our guide to body-composition testing in Singapore compares the options.
A written exit plan, made at the start rather than at the end. What the maintenance phase looks like, what the training and protein floor is, and what triggers a conversation with the prescriber.
GetLean does not ask patients for a body-composition scan or measurement at any point. The programme is telehealth only; what is tracked is weight, reported through the daily WhatsApp check-ins, together with the habits those check-ins are built around.
At GetLean, our philosophy is that GLP-1 medication should act as a catalyst — not something to depend on indefinitely. That commits us to two things the scale alone will not show: protecting lean mass while the weight comes off, and planning the exit from the start. Individual results vary, and clinical-trial figures describe the populations that were studied. If you are considering treatment, check your eligibility and speak to a doctor about whether it is suitable for you.
Common questions
What is the blind spot in doctor-led weight loss?
The metric. Programmes are judged on body weight, and body weight does not distinguish fat from muscle. A patient can hit the stated target in full while a substantial share of what came off was lean tissue — and the same metric says nothing about what happens after the medication stops.
How much of the weight lost on GLP-1 medication is muscle?
It depends on the study, and the range is wide enough that no single figure should be quoted as the answer. The SURMOUNT-1 sub-study of 160 people put about a quarter of the loss as lean mass (Look 2025); a review reading STEP 1's supplementary data puts semaglutide nearer 40%, reconstructed from a conference abstract's appendix rather than a primary paper (Mechanick 2025). It moves with the size of the deficit, sex, age and training (Heymsfield 2014), and getting lean vs losing weight compares the studies in detail.
Does GLP-1 medication cause more muscle loss than dieting?
The evidence does not show that. The drug and placebo arms of the SURMOUNT-1 sub-study came out level on the lean share of weight lost (Look 2025), and ordinary dieting lands in the same range at roughly 14% to 23% depending on how aggressive the diet is (Chaston 2007). What drives the muscle loss is the size of the deficit, not what opened it.
What should a weight-loss programme measure besides weight?
Waist, strength in training, and where it is available a body-composition measurement repeated on the same device. Bioimpedance devices track change over 12 to 16 weeks better than they measure an absolute value at one sitting (Siedler 2022), so using the same scanner each time matters more than which scanner it is.
Why does an exit plan matter if the programme worked?
Because the weight comes back without one. Adults who stopped semaglutide after 68 weeks regained roughly two-thirds of what they had lost within a year (Wilding 2022), and pooled cessation trials put the figure near 60% at one year (Budini 2026). Speak to your prescriber before changing or stopping any medication.