GLP-1 drugs and exercise for muscle and weight maintenance

GLP-1 drugs can reduce weight, but exercise still has a role. Read the evidence on lean mass, muscle function and weight regain after treatment ends.

日本語で読む（Japanese version）

The number on the scale is falling, and clothes that used to feel tight fit again. Yet the exercise slot in your calendar remains empty. If the medication is already helping you lose weight, does making time to move still matter? A visible result can make that question more pressing, rather than make it disappear.

Studies of GLP-1 receptor agonists and related treatments show reductions in body weight and fat mass, alongside reductions in lean mass.[17] Studies following people after treatment stops also find weight regain.[1] Losing weight during treatment and maintaining your body afterward are connected questions, but the evidence does not give them an identical answer.

This article examines what exercise trials, body-composition studies and post-treatment follow-up actually measured. It is a way to prepare for a more specific conversation with your doctor about movement, muscle and maintenance. Decisions about starting, continuing or stopping medication belong with your doctor; a successful morning weigh-in cannot make those decisions for you.

Weight loss and exercise answer different questions

GLP-1 receptor agonists are medications that use the action of hormones in the body. Incretins are hormones released from the gut in response to nutrients. Treatments based on their action include drugs that target GLP-1 receptors and drugs that also act on other receptors.[6] Everyday discussions often group them together as weight loss drugs. Research distinguishes between them, and those distinctions matter when reading results.

A narrative review, which discusses and interprets a collection of studies, reported weight losses of approximately 15–24% in trials of related medications in adults with overweight or obesity.[6] This range covers different drugs and trials. The review’s summary does not provide a confidence interval for the range as a whole. It should therefore be read as a description of the results discussed, rather than a precise prediction for an individual starting treatment.

A separate meta-analysis combined 22 randomized controlled trials involving 2,258 adults with diabetes and/or overweight or obesity. It found a mean body-weight difference of −3.55 kg relative to comparison treatments, with a 95% confidence interval of −4.81 to −2.29 kg.[17] In these trials, random assignment determined which treatment participants received. The interval around the −3.55 kg estimate expresses uncertainty about the average difference between treatments, rather than the range of weight changes every participant experienced.

The percentage range and the difference in kilograms have different reference points. How much weight participants lost during treatment is a different question from how much their result differed from a comparison group.[6][17] Putting the larger-looking number first does not produce a ranking of drug effectiveness. A bathroom scale has no space for the name of the comparison treatment, but a research result needs one.

Exercise adds another set of questions. Does combining it with medication change body weight, body-fat percentage or maintenance after treatment? A trial can find different answers for those different outcomes.[2] Asking only how many kilograms exercise can remove compared with medication leaves out the composition of the weight change and the period after active treatment ends.

The exercise-focused review proposes individually tailored resistance training alongside incretin-based treatment to help preserve muscle and improve body composition.[6] That proposal accepts that medication can produce substantial weight loss. It also gives exercise a distinct job to consider: supporting the body through treatment and preparing for maintenance, rather than simply competing with the drug on the scale.

Lean mass is not the same measurement as muscle

Lean mass means the body’s components other than fat. It includes muscle, but also organs, bone, fluids and water within fat tissue.[14] Reading “lean mass decreased” as “exactly that much muscle disappeared” changes the thing the study measured. The distinction is especially useful when a body-composition result appears to tell a more alarming story than the weight result alone.

In the meta-analysis of 22 randomized trials, the mean difference in fat mass was −2.95 kg, with a 95% confidence interval of −4.11 to −1.79 kg. The mean difference in lean mass was −0.86 kg, with an interval of −1.30 to −0.42 kg. Lean mass loss was reported to account for approximately 25% of total weight loss.[17] Fat loss and a reduction in tissue other than fat were both part of the findings.

The comparison groups in this analysis included placebo and other active treatments.[17] A placebo provides a control against which a medication can be evaluated; an active comparator is another treatment. The participants included adults with diabetes and/or overweight or obesity. Those details describe the evidence behind the average. They also explain why the average cannot simply be assigned to every person taking a related drug.

A 2024 review found much wider variation across studies. In some reports, lean mass reduction accounted for 40–60% of total weight lost; in others, the proportion was approximately 15% or less.[14] The denominator is the amount of weight lost, and the numerator is the lean mass included in that loss. These are not percentages of all the muscle a participant had before treatment.

The review identifies differences in populations, medications and accompanying illnesses as possible reasons for the variation.[14] It also discusses imaging evidence suggesting that reductions in muscle volume may be an adaptation consistent with the weight loss achieved. Changes in fat within muscle and insulin sensitivity may contribute to improved muscle quality.[14] Amount, composition and function are therefore separate parts of the muscle question.

This interpretation does not mean that muscle deserves no attention. The same review highlights older age and disease severity as reasons for particular care, and calls for better assessments of mobility, strength and function as well as muscle mass.[14] A person who notices a change when climbing stairs or carrying a bag has something relevant to report, even if a body-composition printout is the starting point of the discussion.

Reviews disagree about the size of the muscle concern

The 2024 review focused on exercise cites reports of approximately 10%, or approximately 6 kg, of lean mass loss with incretin-based weight loss drugs. It proposes resistance exercise as a way to help preserve muscle.[6] Resistance exercise means movement against a load; strength training is a familiar example. The review connects a concern about treatment-related body composition with evidence on training that works directly on muscle.

It describes supervised resistance-training interventions lasting more than 10 weeks in men and women. Those studies reported increases of approximately 3 kg in lean mass and approximately 25% in strength.[6] These are findings from supervised training research used to support the proposal. They are not a promise that someone taking a GLP-1 drug will gain those amounts, or that every exercise class will reproduce the interventions reviewed.

A 2026 systematic review presents a less alarming overall picture. It examined 36 studies qualitatively and pooled 24 quantitatively in adults with overweight or obesity, with or without type 2 diabetes. It concluded that fat loss predominated and lean mass reductions were modest. At 6 months, lean mass was described as largely preserved.[18] Reading only the review that raises the strongest concern would omit this finding.

The same systematic review reported weight reductions of approximately 9% at 3 months, 5% at 6 months and 4% at 12 months.[18] These figures came from subgroup analyses by medication type and treatment duration. At 6 months, semaglutide, liraglutide and exenatide had comparable effects; at 12 months, the review described variation between agents, particularly liraglutide.[18]

Those percentages are pooled findings for studies at each duration. They should not be read as a single person’s weight-loss trajectory becoming progressively smaller over time. A table organized by duration can look like a diary, but its entries may summarize different study groups. Keeping the study design attached to the number helps avoid a conclusion the review did not establish.[18]

Meanwhile, the analysis of 22 randomized trials found a significant average reduction in absolute lean mass.[17] Relative preservation in one review and an average reduction in another do not have to be forced into a single verdict. Population, treatment, duration and measurement differ across the evidence, and the 2024 muscle review explicitly describes heterogeneity in results.[14][17][18]

There is a further distinction within the 22-trial analysis itself. It assessed absolute lean mass and relative lean mass, defined as percentage change from baseline. It reported a reduction in the absolute measure but no effect on the relative measure.[17] Asking about kilograms lost and asking about change relative to the starting point can produce different descriptions. When hearing that a measure was “preserved,” it helps to establish which measure was assessed.

The useful interpretation is that fat loss is supported, lean tissue changes are not uniform, and muscle quantity and function both deserve attention.[14][17][18] Nutritional and exercise interventions are proposed, but the reviews do not specify an individual resistance-training frequency or load.[6][18] Discuss the exercise appropriate to your treatment with your doctor.

Numbers: body composition and life after treatment

22 trials and 2,258 participants: the scale of the randomized evidence in the body-composition meta-analysis.[17]

−0.86 kg: the mean lean mass difference relative to comparison treatments, with a 95% confidence interval of −1.30 to −0.42 kg.[17]

40–60% versus approximately 15% or less: proportions of total weight loss attributed to lean mass reduction in different studies reviewed.[14]

−5.1 kg: the difference between previous combination treatment and medication alone in weight change from randomization to the post-treatment assessment.[2]

5.63 kg: average weight regain after discontinuation in participants with obesity, with a 95% confidence interval of 3.52 to 7.73 kg.[12]