GLP-1 receptor agonists — semaglutide (Ozempic, Wegovy), tirzepatide (Mounjaro, Zepbound), and related medications — have changed the landscape of weight management. Their efficacy is real and well-documented. But as tens of millions of people worldwide now use these drugs, a body of evidence is building around a problem that is not always flagged in the headlines: significant muscle loss, and a documented decline in physical activity among users. For anyone thinking seriously about fitness and long-term health, this matters.
How much muscle do GLP-1 drugs actually reduce?
The major clinical trials for these drugs have been consistent on this point. In the STEP 1 trial of semaglutide, approximately 39 to 45 percent of weight lost was lean tissue, not fat. The SURMOUNT-1 trial of tirzepatide showed roughly 25 percent of weight lost came from lean tissue. A 2025 prospective study using DXA imaging in real-world semaglutide patients found significant reductions in lean mass at both 7 and 12 months of treatment.
These are not trivial figures. Lean tissue includes muscle, bone, and organs — and muscle in particular is the tissue that drives metabolic rate, protects joints, preserves functional mobility, and becomes progressively harder to rebuild after midlife. Resistance exercise training is the primary tool for preserving lean mass during weight loss, and this applies to GLP-1-induced weight loss as it does to any other caloric deficit. The drug itself does not change that biology.
Users are moving less, not more
A reasonable assumption when someone loses significant weight is that movement becomes easier and physical activity increases. New research presented at ENDO 2026, the Endocrine Society’s annual meeting, directly tested that assumption using objective wearable data — and found the opposite.
The study analysed NIH All of Us Research Program data linking electronic health records with Fitbit activity data from 753 adults with obesity who had started a GLP-1 medication. Average daily steps declined from 5,047 to 4,487 — a drop of 560 steps — and moderate-to-vigorous physical activity fell from 28 minutes per day to 22 minutes. Both changes were statistically significant. The decline was steepest among men and people with musculoskeletal pain. Weight loss from the medication did not lead to increased physical activity. For most participants, it led to less.
The study is preliminary — presented at a conference rather than published in a peer-reviewed journal — and warrants the usual caution about observational data. But it is consistent with clinical observations about appetite suppression-related fatigue and the practical reality that feeling lighter does not automatically translate into wanting to exercise more.
Why this combination is a problem
The convergence of these two findings creates a specific risk. GLP-1 drugs reduce both fat and muscle. Physical activity — particularly resistance training — is the most effective intervention for preserving muscle during caloric restriction. If users are simultaneously losing muscle and reducing the activity that would protect it, the outcome is weight loss that undermines long-term metabolic health, functional strength, and physical resilience, even as the scale moves in the desired direction.
A 2025 case series published in a peer-reviewed journal found that patients who combined GLP-1 therapy with structured resistance training three to five days per week preserved substantially more lean tissue — some actually gained lean mass while losing 13 to 33 percent of their body weight. Protein intakes in that group ranged from 1.6 to 2.3 grams per kilogram of fat-free mass per day. That is a meaningfully higher protein target than most dietary guidance, and it was combined with consistent resistance training — not walking more or general activity advice.
What the emerging guidance recommends
The research base here is still developing. Long-term randomised controlled trials combining GLP-1 therapy with structured exercise protocols are underway at multiple institutions — including at the University of Exeter and Rice University — but results are months to years away. In the meantime, the available evidence points toward a few consistent recommendations.
Resistance training, not just aerobic activity, is the priority. Walking more may support general health but does not produce the anabolic stimulus that protects muscle during a caloric deficit. Protein intake needs to be deliberate, likely in the range of 1.6 grams per kilogram of body weight daily at minimum, distributed across meals. And the timing matters: GLP-1 appetite suppression can make it genuinely difficult to consume adequate protein and total calories. People on these medications often need structured nutritional planning, not simply eating less whenever hunger signals subside.
A 2025 mini-review in Frontiers in Clinical Diabetes and Healthcare argued that exercise professionals and physical therapists should be actively integrated into GLP-1 treatment pathways rather than left as an optional add-on. The framing — that structured exercise is a non-negotiable component of GLP-1 therapy for anyone who cares about the quality of the weight loss, not just the quantity — is gaining traction in clinical circles, even if practice has not yet caught up with the evidence.
An honest caveat
It is worth being clear about what the research does not yet establish. Most muscle loss data comes from clinical trials designed to measure weight loss, not from randomised studies of exercise-plus-GLP-1 versus drug-alone. The case series supporting exercise for muscle preservation is small and non-randomised. The ENDO 2026 activity data is observational and preliminary. What we can say with reasonable confidence is that muscle loss during GLP-1 therapy is real, substantial, and not trivially avoided; that physical activity tends to decline rather than increase after starting these medications; and that resistance training and adequate protein are the most evidence-consistent tools for addressing both problems. The gaps in the evidence are about the precise magnitude of the effect and the optimal protocol — not about whether the problem exists.