A patient returns at week 52 of semaglutide therapy having lost 18% of baseline body weight. The scale result reads as an unambiguous success. The DEXA scan ordered alongside it tells a more complicated story: a meaningful share of that loss came from lean tissue, not fat. This is not a hypothetical — it is close to what the body composition substudy nested inside the STEP 1 trial actually found, and it is the reason muscle loss on GLP-1 therapy has moved from an afterthought to an active research question.
Muscle loss on GLP-1 therapy, as measured by DEXA substudies, is a real and quantifiable phenomenon — but the data are narrower than the headlines built on top of them. Only a fraction of each pivotal trial's enrollment underwent dual-energy X-ray absorptiometry scanning, the numbers describe total lean mass rather than isolated skeletal muscle, and the comparison point — diet-induced weight loss without medication — also comes with substantial lean tissue loss. Understanding what these substudies measured, and what they didn't, is necessary before translating a percentage into clinical concern.
What DEXA Substudies Actually Measure
Dual-energy X-ray absorptiometry partitions body mass into three compartments — fat mass, lean mass, and bone mineral content — using differential attenuation of two X-ray energy levels passed through tissue. It is considered a reference-standard method in body composition research, more precise than bioelectrical impedance analysis (BIA) and far more informative than scale weight or BMI alone.
Critically, the 'lean mass' DEXA reports is not synonymous with skeletal muscle. It includes muscle tissue, but also organ mass, connective tissue, and total body water, the last of which can shift substantially during rapid weight loss independent of actual muscle fiber loss. A drop in lean mass on a DEXA readout can reflect some combination of fluid shifts, glycogen-bound water depletion, and genuine myofibrillar loss — the scan alone cannot separate these cleanly.
This matters for interpreting trial substudies. When a DEXA substudy reports that lean mass accounted for a given percentage of total weight lost, that figure is a reasonable proxy for tissue composition change, but it overstates precision if read as a direct measurement of muscle fiber atrophy. More granular methods — MRI-based muscle volumetrics or D3-creatine dilution for muscle mass specifically — exist in the research literature but have not been widely deployed inside GLP-1 pivotal trials to date.
The mechanistic backdrop for why any lean loss occurs is tied to the drug class's core pharmacology — reduced caloric intake via delayed gastric emptying and central appetite suppression. The mechanism behind that appetite effect is covered in detail in a related analysis of GLP-1 receptor agonists and gastric emptying, which is also the mechanistic root of the tolerability profile associated with these drugs.
The STEP 1 Body Composition Substudy: Semaglutide DEXA Findings
The most frequently cited DEXA data for semaglutide 2.4 mg comes from a body composition substudy nested within the STEP 1 trial (Wilding et al., NEJM 2021, PMID 33567185), published separately by Wilding and colleagues in Diabetes Obesity and Metabolism (2022). STEP 1 itself enrolled adults with overweight or obesity without diabetes and ran 68 weeks; the parent trial reported mean weight loss of approximately 14.9% with semaglutide versus 2.4% with placebo.
The DEXA substudy involved a nested subset of participants, not the full STEP 1 cohort, who underwent scanning at baseline and at 68 weeks. The commonly cited finding from that analysis is that lean mass accounted for approximately 39.7% of total weight lost in the semaglutide arm. Fat mass made up the majority of the remainder, consistent with the drug's intended effect, but the lean mass contribution was large enough to be flagged as clinically relevant by the study authors themselves.
What the substudy did not report is equally important: no paired functional testing — grip strength, chair-stand performance, gait speed — accompanied the imaging, so whether that lean mass reduction translated into measurable weakness is [CITATION NEEDED: functional outcome data paired with STEP 1 DEXA substudy]. The substudy also pre-dates longer-term data on whether lean mass is regained, partially regained, or permanently reduced after the main trial period, a question addressed separately in research on GLP-1 discontinuation and weight regain from STEP 4 follow-up.
Benchmarking Against Diet-Induced Weight Loss
A lean mass figure in isolation is hard to judge without a comparison point. The relevant benchmark is diet-induced weight loss without pharmacologic intervention, where fat-free mass loss has long been documented as a predictable companion to caloric restriction. Weinheimer, Sands, and Campbell's 2010 systematic review (Nutrition Reviews, PMID 20591112) pooled data across energy-restriction studies in middle-aged and older adults and estimated that fat-free mass accounted for roughly 20-30% of total weight lost through diet alone, trending higher with more aggressive restriction and lower baseline muscle mass.
Set against that range, the 39.7% lean mass contribution reported in the STEP 1 substudy sits above the typical diet-only benchmark, though the comparison carries caveats: different populations, different measurement windows, and different rates of total weight loss make a clean apples-to-apples read difficult. Semaglutide's 68-week, ~15% weight loss is also larger in magnitude than many of the diet studies pooled in the 2010 review, and larger total weight loss is independently associated with a larger absolute (though not necessarily proportional) lean mass loss.
The practical interpretation favored in the literature so far is that GLP-1-associated lean mass loss looks proportionally similar to, or modestly elevated above, what caloric restriction alone would be expected to produce at a comparable magnitude of weight loss — not a categorically distinct phenomenon unique to the drug class. Whether that holds at even larger weight-loss magnitudes, as seen with newer triple agonists, remains an open question worth tracking as more substudies report out.
Tirzepatide and Dual Agonism: Body Composition Signals
Tirzepatide, a dual GIP/GLP-1 receptor agonist, produces larger mean weight loss than semaglutide in head-to-head data — the SURMOUNT-5 trial comparison is detailed in a dedicated review of tirzepatide versus semaglutide head-to-head trial data. Larger total weight loss raises the obvious question of whether lean mass contribution scales proportionally, stays flat, or improves with GIP co-agonism, which has been proposed in preclinical work to have a more favorable body composition signature.
Published DEXA substudy data specific to tirzepatide's lean-to-fat loss ratio at the level of detail available for semaglutide's STEP 1 substudy is less consistently reported in the literature reviewed here, and specific percentage figures should be treated as [CITATION NEEDED: tirzepatide DEXA substudy lean mass percentage of total weight loss, by trial and timepoint] until a primary source is confirmed. What is established is that tirzepatide's long-term maintenance dosing data, including findings from SURMOUNT-4 on withdrawal and regain, shows that weight regain after treatment cessation is substantial, which carries its own body composition implications covered in the dedicated analysis of SURMOUNT-4 withdrawal and regain findings.
Until head-to-head DEXA substudy data directly comparing tirzepatide and semaglutide lean mass contributions are published with matched methodology, claims that one drug class 'spares' muscle better than another should be treated as preliminary or extrapolated from indirect comparisons rather than settled findings.
Mechanistic Drivers of Lean Mass Loss
Three overlapping mechanisms plausibly explain lean mass loss during GLP-1 therapy, none of which point to a direct catabolic action of the drug on muscle tissue itself. The first and most direct is caloric deficit: GLP-1 receptor agonism suppresses appetite centrally and slows gastric emptying peripherally, producing a substantial spontaneous reduction in intake. Any sufficiently large caloric deficit, regardless of cause, produces some fat-free mass loss as an expected physiological consequence.
The second driver is reduced physical activity accompanying rapid weight loss, particularly in patients who do not maintain or increase resistance training during treatment. Lower mechanical loading on skeletal muscle removes one of the primary stimuli that would otherwise counteract lean tissue breakdown during a caloric deficit.
The third, less studied, driver is the rate of weight loss itself. Faster weight loss trajectories are generally associated with a higher proportional lean mass contribution compared to slower, more gradual loss at an equivalent caloric deficit, based on broader body composition research outside the GLP-1 literature specifically. Given that GLP-1 therapies can produce weight loss rates exceeding what most unmedicated caloric restriction protocols achieve, this rate-dependent effect is a plausible contributor worth isolating in future substudy designs, though it has not been directly tested within a GLP-1 DEXA substudy to date.
Who Is at Highest Risk for Clinically Meaningful Lean Loss
Baseline characteristics appear to modify lean mass loss risk, though dedicated subgroup DEXA analyses remain sparse. Older adults are a recognized higher-risk group because age-related sarcopenia already reduces baseline muscle reserve, meaning the same proportional lean mass loss represents a larger relative hit to functional capacity than it would in a younger patient.
Patients with lower baseline muscle mass relative to total body mass — which can include some patients with sarcopenic obesity — are another plausible higher-risk group, since they have less lean tissue buffer before functional thresholds like grip strength or mobility are affected. Rate of weight loss is a third risk marker: patients losing weight unusually quickly, whether due to high responsiveness to the medication or aggressive concurrent caloric restriction, appear more likely to show a higher lean-to-fat loss ratio, consistent with the broader body composition literature on rapid weight loss generally.
None of the pivotal GLP-1 DEXA substudies reviewed here have published a dedicated age-stratified or baseline-muscle-stratified breakdown of lean mass loss at the level of detail needed to quantify risk precisely by subgroup [CITATION NEEDED: age-stratified DEXA lean mass outcomes from a GLP-1 pivotal trial substudy]. This is one of the clearer gaps in the current literature and a reasonable target for clinicians counseling older or lower-muscle-mass patients before initiating therapy.
Monitoring Protocols: DEXA, BIA, and Functional Testing
Trial-grade DEXA scanning is not standard practice in routine clinical monitoring of patients on GLP-1 therapy, largely due to cost, access, and the absence of an established interval protocol outside research settings. Where body composition monitoring does occur clinically, BIA is more commonly used because it is inexpensive and available in many clinic settings, despite being less precise than DEXA and sensitive to hydration status, recent exercise, and time of day.
A monitoring approach grounded in the substudy literature would reasonably include a baseline body composition assessment before initiating therapy in higher-risk patients, periodic reassessment at intervals matching major dose titration steps (discussed in detail in a review of semaglutide dose titration schedules from pivotal trials), and functional testing — grip strength dynamometry or a simple chair-stand test — that no pivotal trial DEXA substudy has yet paired with imaging data but that would close an obvious gap between a composition number and real-world strength or mobility.
Laboratory monitoring overlapping this picture, such as periodic assessment of markers relevant to muscle and metabolic status, should be guided by the prescribing clinician rather than a fixed universal schedule, since optimal intervals have not been established in a dedicated trial.
Mitigation Strategies in the Literature
Resistance training and adequate protein intake are the two interventions with the strongest independent evidence base for preserving lean mass during caloric deficit, established well before GLP-1 therapies existed. Extrapolating from that general body composition literature, a reasonable clinical hypothesis is that structured resistance training combined with protein intake sufficient to support muscle protein synthesis would attenuate the lean mass loss observed in DEXA substudies.
That hypothesis has not yet been confirmed by a randomized trial that pairs resistance training with GLP-1 therapy and DEXA follow-up at the scale of STEP 1's substudy [CITATION NEEDED: RCT testing resistance training as a lean-mass-preservation intervention during GLP-1 therapy with DEXA confirmation]. Several smaller studies and ongoing trials are reportedly investigating this question, but a definitive, adequately powered substudy result is not yet available in the published literature reviewed here.
Related tissue-preservation strategies, including the physiological rationale for why lean mass protection matters during rapid pharmacologic weight loss, are covered in more depth in a dedicated article on GLP-1 therapy and lean muscle mass preservation, which addresses practical considerations clinicians raise when counseling patients before and during treatment.
Research Gaps and What Remains Unclear
Several questions remain open even after accounting for the STEP 1 substudy and related data. First, no pivotal trial DEXA substudy has published matched functional outcome data, so the clinical significance of a given percentage lean mass loss — whether it meaningfully affects strength, fall risk, or metabolic rate — is inferred rather than directly demonstrated.
Second, long-term trajectories after the initial 52-to-68-week trial windows are not well characterized by DEXA; whether lean mass lost during active treatment is regained, stays suppressed, or continues declining with extended use is an open question tied to the broader discontinuation and maintenance literature. Third, comparative DEXA data across drug classes — semaglutide, tirzepatide, and emerging triple agonists — using matched methodology and timepoints does not yet exist in a form that supports confident head-to-head claims about which agent is more muscle-sparing.
Closing these gaps will likely require purpose-built substudies designed from the outset to pair DEXA or more granular imaging with functional testing across age and baseline-muscle-mass strata, rather than retrofitting body composition analysis onto trials designed primarily around weight and glycemic endpoints.
Clinical Takeaway
The DEXA substudy evidence available to date supports a specific, bounded conclusion: lean mass loss during GLP-1 therapy is real, proportionally comparable to or modestly above what diet-induced weight loss alone would be expected to produce, and measured in only a fraction of each trial's total participants using total lean mass rather than isolated skeletal muscle. It does not support either extreme — neither dismissing muscle loss as a non-issue nor treating it as a unique hazard specific to this drug class.
For clinicians managing patients on GLP-1 therapy, the practical next step is a baseline risk assessment — age, baseline muscle mass, rate of expected weight loss — followed by a monitoring plan appropriate to that risk level, developed with the prescribing clinician rather than derived from a single substudy percentage.
This article summarizes research and does not constitute medical advice. Consult a licensed clinician for diagnosis, treatment, or any decisions about medications or supplements.