GLP-1 Weight Loss and Muscle: What the Evidence Really Shows

The scanner table is narrow and flat, and the body on it does nothing. No effort, no performance, no fatigue to measure. Dual-energy X-ray absorptiometry passes over what lies there and divides it into three compartments: bone mineral, fat, and lean soft tissue. The third compartment is where the argument lives. It counts contractile muscle, and it also counts body water, glycogen, organs, and connective tissue. When the SURMOUNT-1 substudy reported that tirzepatide reduced lean mass by 10.9% over 72 weeks, this is the instrument that said so — and its silence covers everything the instrument cannot say.

What the scanner cannot register is the thing people actually fear losing: strength, function, the capacity to rise and carry and climb in midlife and later life, when reserve may already be shrinking. Nor can it weigh what excess adiposity itself costs — restricted movement, insulin resistance, cardiovascular disease, sleep apnea, painful joints. The useful question is not whether the scale moved. It is what tissue changed, whether strength and physical function changed with it, and whether a particular person became more or less capable. On the first part of that question the evidence is solid; on the second, it is only beginning.

The 25% number needs a label

A 2025 network meta-analysis of 22 randomized trials found that GLP-1 and related agents reduced body weight, fat mass and lean mass. On average, lean mass accounted for about one quarter of the weight lost. The DXA substudy of SURMOUNT-1 produced a similar ratio. Among 160 participants, tirzepatide reduced body weight by 21.3%, fat mass by 33.9% and lean mass by 10.9% at 72 weeks; roughly 75% of lost weight was fat and 25% was lean. The placebo group had a similar fat-to-lean proportion, although it lost much less weight overall.

That result supports concern without supporting panic. “Twenty-five percent of weight loss was lean mass” is not the same statement as “twenty-five percent was contractile skeletal muscle,” and neither statement tells us that strength fell by 25%. Dual-energy X-ray absorptiometry, or DXA, divides the body into bone mineral, fat and lean soft tissue. Its lean compartment includes body water, glycogen, organs, connective tissue and muscle. Hydration and glycogen can change during a large energy deficit. Some fat-free tissue also disappears as a smaller body requires less supporting tissue.

Percentages can further confuse the picture. A person may lose lean tissue in absolute kilograms while the percentage of the smaller body composed of lean tissue rises, because fat fell faster. A 2026 meta-analysis of seven obesity-dose trials found exactly that pattern: absolute lean mass decreased while lean mass as a proportion of body weight improved. Neither metric should be allowed to cancel the other.

Muscle quantity is not muscle quality

Sarcopenia is not diagnosed from one whole-body lean-mass number. Muscle quantity, strength and physical performance are related but distinct. A recent physiological review concluded that comparative evidence on GLP-1 therapy and muscle quantity remains highly heterogeneous and imprecise. Muscle-quality results could not even be pooled because studies used incompatible methods and often omitted needed data. The reviewers judged the evidence low to very low certainty and cautioned both against equating lean-mass decline with sarcopenia and against declaring vulnerable patients safe.

The SEMALEAN study shows why function changes the story. In 106 patients with severe obesity treated with semaglutide, DXA lean mass fell by about 3 kilograms during the first seven months and then stabilized, while handgrip strength improved by 4.5 kilograms at 12 months. That is encouraging, but it was a prospective cohort without a randomized control group. Participants who completed follow-up may differ from those who did not, and handgrip is only one window into function. The study shows that lean tissue and strength can move in different directions; it does not prove that no clinically important muscle was lost.

Exercise adds an outcome the injection cannot

The strongest direct exercise evidence comes from liraglutide, not today’s more potent semaglutide or tirzepatide regimens. After an eight-week low-calorie diet, 195 adults with obesity were randomized to a year of exercise, liraglutide, both or placebo. The combination produced the largest sustained weight loss and roughly twice the improvement in body-fat percentage seen with either active treatment alone.

A later analysis examined what people could do. The exercise program mixed interval cycling with circuit training; it was not a pure resistance-training trial. Compared with liraglutide alone, exercise plus liraglutide improved stair-climbing time and cardiorespiratory fitness. Absolute knee-extensor strength was preserved across the active groups. Relative strength—strength divided by body weight—improved partly because the denominator became smaller. This is valuable evidence that structured movement supplies fitness gains the drug alone did not, but it does not quantify how much lean mass resistance training preserves during semaglutide or tirzepatide treatment.

For older adults, much of the practical case still comes from non-drug weight-loss studies. In a randomized trial of 160 adults with obesity, resistance training or combined aerobic and resistance training limited lean-mass loss more than aerobic exercise alone while producing larger strength gains. The combined program improved overall physical performance most. Participants were not taking GLP-1 drugs, so this is biologically relevant extrapolation rather than direct proof. Exercise-related injuries also occurred, a reminder that “just lift” is not a complete plan for someone with frailty, arthritis, balance problems or cardiovascular limitations.

Protein is a foundation, not a force field

Appetite suppression can make adequate food intake difficult, so protein deserves attention. A meta-analysis of 20 randomized trials in adults older than 50 found that higher-protein diets during energy restriction retained more lean mass and lost more fat mass than lower-protein diets. Yet another meta-analysis, restricted to nonfrail community-dwelling older adults whose habitual intake was often already sufficient, found no significant added benefit from protein supplements for lean mass, strength, gait speed or chair-rise performance.

Those results are not contradictory. Correcting a shortfall during rapid weight loss is different from adding a powder to an already adequate diet. Protein targets also cannot be safely universalized from social posts. Total energy intake, body size, kidney function, illness, food tolerance and training all change the calculation. A 2026 review found only seven heterogeneous studies in adults over 50 with multiple long-term conditions, and none directly analyzed protein’s effect in relation to multimorbidity. More protein is not automatically better, and a supplement cannot provide the mechanical stimulus or motor learning of resistance exercise.

The trial that reveals the gap

The clearest sign that the answer is unfinished is LEAN-PREP. This 2026 publication is a protocol, not a result. Investigators plan to randomize 232 adults starting semaglutide or tirzepatide to usual care, resistance exercise, protein support or both. They will measure quadriceps area by MRI, DXA body composition, intramuscular fat, strength and physical function. It is almost exactly the experiment social media implies has already been completed.

Until those results arrive, clinicians and patients have to combine direct but incomplete GLP-1 evidence with stronger general evidence from intentional weight loss and aging. For a robust younger adult, a modest DXA change with stable or improving function may mean something very different than the same change in an older adult with recent falls, low baseline strength, poor intake or unintentional weight loss.

A better scoreboard than weight alone

A muscle-aware treatment discussion can track the pace and quality of weight loss alongside symptoms, food intake and what a person can actually do. A scan may add context, but repeated measures should use a consistent method and should not outrank walking speed, chair rises, grip or task-specific strength. New fatigue or weakness deserves evaluation rather than an automatic instruction to eat more protein or train harder.

The honest verdict is balanced. GLP-1 and GIP/GLP-1 medicines can produce clinically important fat and weight loss, and absolute lean mass often falls with it. Existing studies do not show a uniform collapse in muscle function, but they are too inconsistent to rule out meaningful harm in people with limited reserve. Resistance training and adequate nutrition are evidence-aligned parts of comprehensive obesity care; the exact degree to which they protect muscle during potent incretin treatment is still being tested. The goal is not the smallest body. It is a healthier body that can carry its owner through daily life. Educational, not medical advice.