GLP-1 and lean mass loss: what the clinical data actually shows — and why the receptor mechanism matters

Every major GLP-1 receptor agonist trial uses total body weight as the primary endpoint. It is the clearest number and the most cited. It is also incomplete.

Body weight is a composite of fat mass, lean mass (skeletal muscle, bone mineral, body water, and glycogen), and organ tissue. When subjects lose 15–25% of body weight over 24–72 weeks, what matters for long-term metabolic health is what was lost, not just how much.

The clinical stakes:

• Basal metabolic rate: Skeletal muscle is the primary contributor to resting energy expenditure after organ tissue. Each kilogram of lean mass lost suppresses BMR by approximately 13–15 kcal per day — a compounding deficit that makes long-term weight maintenance progressively harder. A subject losing 15kg with a 40% lean fraction loses roughly 6kg of muscle and approximately 80–90 fewer kcal burned per day at rest.
• Sarcopenia risk: Adults lose 3–8% of muscle mass per decade from the fourth decade onward. Rapid weight loss without structured resistance training accelerates this trajectory. Sarcopenic obesity — low muscle mass with preserved or elevated adiposity — is now actively flagged in GLP-1 follow-up literature as an adverse body composition outcome distinct from weight regain.
• Weight regain composition: STEP-4 (semaglutide discontinuation at 68 weeks) showed approximately two-thirds of lost weight regained within one year of stopping treatment. Regained weight skews toward fat rather than lean mass, creating progressive deterioration in body composition across treatment cycles for subjects who restart.
• Functional outcomes: Grip strength, gait speed, and fall risk correlate with appendicular skeletal muscle mass — outcomes that carry clinical weight in older research cohorts independently of scale weight.

Quantifying lean mass loss per compound, and understanding what receptor mechanism data does and does not support about differences between them, is therefore a primary research question — not a secondary one.

What DXA measures — and what it misses Dual-energy X-ray absorptiometry classifies all body tissue as fat mass or fat-free mass (FFM). FFM includes skeletal muscle, bone mineral, body water, glycogen, and organ tissue — not just muscle protein. When a trial reports lean mass loss, some of that change reflects glycogen depletion and associated water loss (1g glycogen binds ~3g water), particularly during early titration phases. DXA cannot distinguish myofibrillar protein loss from water and glycogen loss. Longer-duration measurements capture more true muscle catabolism relative to early glycogen effects — one reason body composition comparisons across trials of different durations are methodologically imperfect.

The STEP-1 trial (Wilding et al., NEJM 2021) enrolled 1,961 adults with obesity (BMI ≥30, or ≥27 with a weight-related comorbidity) without type 2 diabetes. Participants received semaglutide 2.4mg subcutaneous weekly or placebo over 68 weeks, alongside a structured lifestyle intervention programme.

Primary endpoint: mean change in body weight from baseline. Semaglutide: −14.9% vs −2.4% placebo. That figure is universally cited. The body composition data tells a more complete story.

Body composition — DXA substudy

STEP-1 included DXA body composition measurements in a substudy subset of participants. Key finding: approximately 38–40% of total weight lost under semaglutide was fat-free mass — lean mass rather than adipose tissue.

Measure	Semaglutide 2.4mg	Placebo
Mean total weight change	−14.9%	−2.4%
Approx. fat mass fraction of weight lost	~60–62%	—
Approx. fat-free mass fraction of weight lost	~38–40%	—
Trial duration	68 weeks	68 weeks

Stated concretely: for a research subject starting at 100kg and losing 14.9kg under semaglutide, approximately 5.7–6kg of that loss is lean mass rather than adipose tissue.

Context is critical. Standard caloric restriction in the absence of structured resistance training typically produces lean mass loss in the range of 20–30% of total weight lost. The STEP-1 semaglutide figure appears modestly higher than dietary restriction alone — though direct cross-study comparison is confounded by differences in exercise co-intervention, dietary protein standardisation, and protocol duration.

The mechanism is straightforward. GLP-1 receptor agonism suppresses appetite and reduces total energy intake, but that suppression is not selective for fat substrates. A caloric deficit driven by appetite suppression — without a deliberate increase in dietary protein — draws on all fuel sources, including muscle protein catabolism. Faster, deeper deficits produce proportionally larger lean mass contributions to total weight lost.

STEP-4's discontinuation data adds an important downstream dimension: participants who stopped semaglutide after 20 weeks of treatment regained approximately 6.9% of body weight in the following 48 weeks. Regained weight skewed toward fat rather than lean mass — meaning body composition trajectory worsens across treatment cycles for subjects who stop and restart.

Early-phase lean mass loss — the glycogen caveat Semaglutide produces significant food intake reduction during titration (weeks 1–16). Reduced carbohydrate intake depletes hepatic and muscular glycogen rapidly. DXA measurements in the first 8–16 weeks capture this glycogen-water component as fat-free mass loss — potentially inflating the apparent lean mass fraction of early weight loss relative to longer steady-state conditions. The 68-week STEP-1 timepoint captures a more stable picture, but the early-phase effect remains relevant for short-protocol research designs.

SURMOUNT-1 (Jastreboff et al., NEJM 2022) enrolled 2,539 adults with obesity (BMI ≥30) or overweight (BMI ≥27 with comorbidity) without type 2 diabetes. Participants received tirzepatide at 5mg, 10mg, or 15mg weekly, or placebo, over 72 weeks.

The headline weight loss at the maximum dose (15mg): −20.9% — substantially greater absolute loss than semaglutide in STEP-1. The body composition question is whether the higher total loss carried a proportionally lower lean mass fraction, or a similar one.

Body composition — DXA substudy

SURMOUNT-1 DXA data showed that across dose groups, approximately 33–39% of total weight lost was fat-free mass. This range is broadly consistent with STEP-1 semaglutide data — there is no clear evidence from available data that GIP receptor addition produces a substantially more fat-selective outcome on a proportional basis.

Measure	Tirzepatide 5mg	Tirzepatide 10mg	Tirzepatide 15mg
Mean total weight change	−15.0%	−19.5%	−20.9%
Approx. lean mass fraction of weight lost	~33%	~36%	~39%
Trial duration	72 weeks

The absolute lean mass cost at higher tirzepatide doses is therefore greater in kilogram terms than with semaglutide — not because the fraction is higher, but because the total weight lost is greater. A subject losing 21% of a 100kg starting weight (21kg total) with 39% as lean mass loses approximately 8.2kg of lean mass. A semaglutide subject losing 14.9kg with a 40% lean fraction loses approximately 6kg. The absolute muscle cost is meaningfully higher with tirzepatide at maximum doses, despite similar proportional fractions.

The GIP receptor question

Tirzepatide adds GIP receptor (GIPR) agonism to GLP-1 receptor agonism. The role of GIPR in adipose tissue biology is complex: GIP promotes insulin-dependent fat storage under caloric excess conditions, but GIPR agonism in sustained caloric deficit appears to enhance fat mobilisation — consistent with the greater total weight loss relative to GLP-1 monotherapy.

What SURMOUNT-1 does not support is a strong claim that dual GLP-1/GIP agonism substantially spares lean mass on a per-kilogram-lost basis relative to GLP-1 monotherapy. The proportional data is similar. The greater absolute lean mass loss at maximum tirzepatide doses is a direct consequence of greater total weight loss, not a worse lean-mass ratio.

See the Tirzepatide Research Guide for full mechanism and SURMOUNT programme data.

Retatrutide is a triple agonist: GLP-1 receptor, GIP receptor, and glucagon receptor (GcgR). The third receptor is the key pharmacological distinction from semaglutide and tirzepatide — and the basis for the hypothesis that retatrutide may produce more fat-selective weight loss.

What glucagon receptor agonism adds

Glucagon acts on GcgR to increase hepatic fat oxidation (beta-oxidation), promote thermogenesis via brown adipose tissue activation, raise basal metabolic rate through increased substrate oxidation rates, and increase energy expenditure independent of lean mass changes. These effects are additive to GLP-1 and GIP receptor signalling rather than redundant with them.

The muscle-sparing hypothesis follows: if a portion of the caloric deficit under retatrutide is met by elevated fat oxidation — thermogenic and hepatic — rather than increased muscle protein catabolism, the lean mass fraction of total weight loss should be lower than with GLP-1 monotherapy or GLP-1/GIP dual agonism. The hypothesis is mechanistically coherent. Whether Phase 2 data supports it requires examining what the trial actually measured.

Phase 2 data: Jastreboff 2023 (NEJM)

The retatrutide Phase 2 trial (Jastreboff et al., NEJM 2023) enrolled 338 adults with obesity over 24 weeks across multiple dose arms (1mg, 4mg, 8mg, 12mg weekly). At 24 weeks, the 12mg dose group achieved mean total weight loss of approximately 17.5% — the highest 24-week result reported in a GLP-1 class Phase 2 trial at publication.

Body composition data by DXA showed a high proportion of fat mass contributing to total weight loss across dose groups, consistent with the glucagon receptor thermogenic mechanism. However, the trial was designed and powered as a Phase 2 efficacy and safety study, not as a body composition comparison against tirzepatide or semaglutide.

Parameter	Retatrutide 12mg (Phase 2)	Context
Mean total weight loss at 24 weeks	~17.5%	STEP-1: 14.9% at 68 wks · SURMOUNT-1: 20.9% at 72 wks
Trial phase	Phase 2	STEP-1 and SURMOUNT-1: Phase 3
Total N enrolled	~338	SURMOUNT-1 N: 2,539
Head-to-head body composition vs tirzepatide	Not conducted — no direct comparison exists

What the Phase 2 data does and does not establish

• Duration: 24 weeks is shorter than STEP-1 (68 weeks) or SURMOUNT-1 (72 weeks). Body composition ratios often shift over longer protocols as early glycogen-water effects normalise and sustained muscle catabolism becomes a larger fraction of total loss.
• No head-to-head body composition comparison: There is no randomised trial directly comparing retatrutide to tirzepatide on lean mass endpoints. All cross-compound comparisons are across separate trials with different populations, durations, and co-interventions.
• Phase 2 sample size: Phase 2 body composition substudies are powered for feasibility signal, not definitive lean mass endpoint comparison.
• Phase 3 pending: Retatrutide Phase 3 trials (TRIUMPH programme) are ongoing. Definitive body composition data at scale and longer duration has not been published.

Mechanism is signal, not confirmation The glucagon receptor adds thermogenic fat oxidation that GLP-1 and GIP agonism alone do not provide at equivalent doses. Phase 2 data is consistent with a muscle-sparing benefit. Phase 3 data will determine whether that benefit is statistically and clinically meaningful at scale, and whether it persists past 52 weeks. The question is genuinely open — which is itself a meaningful research signal worth tracking.

For full mechanism, dosing, and Phase 2 outcome data, see the Retatrutide Research Guide.

Two significant pieces of evidence from early 2026 update this picture — one consolidating the research framework, one directly challenging the glucagon receptor hypothesis.

PMC12444289: 2026 systematic review

A 2026 systematic review (PMC12444289) consolidates body composition outcomes across published GLP-1 class Phase 2 and 3 datasets. The review provides a structured comparison of lean mass loss as a fraction of total weight lost across available trials, noting that the 30–40% fat-free mass range appears broadly consistent across the GLP-1 class. Critically, it identifies exercise co-intervention as the most evidence-supported modifier of lean mass outcomes during GLP-1 treatment — more robustly supported than compound selection at the current stage of the literature.

Nature International Journal of Obesity, February 2026: the MC4R mouse finding

A February 2026 paper published in Nature International Journal of Obesity examined body composition outcomes across all three major GLP-1 class analogs — semaglutide, tirzepatide, and retatrutide — in a melanocortin-4 receptor (MC4R) knockout mouse model of obesity. MC4R knockout mice are a validated preclinical model used to study obesity pharmacology and GLP-1 class mechanisms.

The finding: all three compounds reduced lean mass in this model, with no statistically significant difference between them in the lean mass fraction of total weight lost.

This is an important and honest result. It does not invalidate the glucagon receptor hypothesis — murine models carry well-documented limitations in translating to human pharmacological outcomes, and the MC4R knockout model may not fully recapitulate the human GcgR thermogenic pathway. But it is the most recent controlled, same-model comparison across all three compounds on a direct body composition endpoint, and it found no lean mass advantage for retatrutide in that system.

The honest synthesis of the current evidence base:

• The mechanistic case for retatrutide producing fat-preferential loss via GcgR is coherent and consistent with known glucagon receptor biology.
• The Jastreboff 2023 Phase 2 human data is consistent with this hypothesis but was not designed or powered to confirm it definitively.
• The 2026 Nature IJO preclinical data (MC4R model) found no significant lean mass advantage for retatrutide over semaglutide or tirzepatide.
• Phase 3 clinical body composition data at scale and long duration has not been published.

The signal is real. The confirmation is pending. Researchers should hold that uncertainty explicitly rather than resolve it prematurely in either direction.

The strongest evidence-supported modifier of lean mass outcomes during GLP-1 protocols is not pharmacological — it is exogenous co-intervention. Three interventions have the clearest evidence base:

1. Progressive resistance training

Meta-analyses consistently show structured resistance training during caloric restriction attenuates lean mass loss by 40–60% relative to no-exercise conditions. The mechanism is mechanical loading-driven muscle protein synthesis via mTORC1 pathway activation, which counteracts the catabolic signalling that accompanies sustained energy deficit.

Minimum protocol parameters for lean mass preservation during GLP-1 research:

• Frequency: 3 sessions per week minimum, full-body or upper/lower split
• Load: Progressive overload targeting 6–12 repetitions per set (hypertrophic range)
• Volume: 10–20 working sets per muscle group per week across sessions
• Timing: Begin prior to or concurrent with compound initiation — do not delay until after significant weight loss has already occurred

2. Protein intake

Dietary protein is the primary substrate for muscle protein synthesis and the most evidence-supported nutritional intervention for lean mass preservation in caloric deficit. Current evidence-based targets for weight loss contexts:

• Minimum: 1.6g per kg body weight per day
• Target for active resistance training protocols: 2.0–2.4g per kg body weight per day
• Distribution: Across 3–4 meals containing ≥30–40g protein each (above the leucine threshold for maximal acute muscle protein synthesis stimulation per meal)

GLP-1 agonists significantly reduce total caloric intake. Without deliberate protein prioritisation, this reduction proportionally cuts protein consumption during the period when lean mass preservation demands are highest — a compounding problem during aggressive deficit phases.

3. Creatine monohydrate

Creatine has a robust evidence base for lean mass preservation during caloric restriction and for attenuating strength loss during energy deficit. At 3–5g per day, creatine increases phosphocreatine stores, supports resistance training performance, and appears to have direct anti-catabolic effects on skeletal muscle. A 2021 meta-analysis (Lanhers et al.) found significant preservation of lean mass with creatine supplementation during resistance training in hypocaloric conditions. Creatine monohydrate is among the most studied and lowest-risk nutritional interventions in exercise science — a logical co-intervention for any GLP-1 research protocol prioritising lean mass outcomes.

4. BPC-157 for recovery

Body protective compound-157 (BPC-157) has been investigated in preclinical models for effects on tendon and muscle repair, angiogenesis, and recovery from mechanical overload. For researchers running concurrent resistance training protocols alongside GLP-1 compounds — where training volume and tissue repair demands are both elevated — BPC-157 represents an active area of preclinical interest for recovery support.

See the RetaLABS BPC-157 product page for sourcing and purity documentation, and the BPC-157 Research Guide for mechanism and study data.

The following table consolidates available body composition data across the three primary GLP-1 class compounds currently available for Australian research. All figures are from published trial data. Body composition percentages are approximate from DXA substudies. No head-to-head body composition trial exists across all three compounds.

Compound	Trial	Phase	Duration	Total weight loss (top dose)	Approx. lean mass fraction	Receptors
Semaglutide	STEP-1, NEJM 2021	3	68 weeks	14.9%	~38–40% of weight lost as FFM	GLP-1R
Tirzepatide	SURMOUNT-1, NEJM 2022	3	72 weeks	20.9% (15mg)	~33–39% of weight lost as FFM	GLP-1R + GIPR
Retatrutide	Jastreboff 2023, NEJM	2	24 weeks	~17.5% (12mg)	Phase 2 DXA: high fat fraction; Phase 3 pending	GLP-1R + GIPR + GcgR

FFM = fat-free mass (lean mass by DXA). Lean mass fractions are approximate from DXA substudies and should be treated as directional rather than precise. Retatrutide Phase 3 body composition data not yet published. No direct head-to-head body composition trial across all three compounds exists.

Evidence summary

• All three GLP-1 class compounds produce meaningful lean mass loss. None eliminates it.
• Lean mass as a fraction of total weight lost is broadly similar across semaglutide (~38–40%) and tirzepatide (~33–39%). There is no clear proportional advantage for tirzepatide's GIP receptor addition on this endpoint.
• Higher total weight loss at maximum tirzepatide doses means more absolute lean mass lost, even where the percentage fraction is similar to semaglutide.
• The glucagon receptor mechanism in retatrutide provides a credible pathway for fat-preferential loss. Phase 2 human data is consistent with this hypothesis. Phase 3 confirmation is pending.
• A 2026 preclinical study (MC4R mouse model, Nature IJO) found no significant lean mass difference across all three compounds — an important counterpoint that Phase 3 data must address.
• Progressive resistance training and adequate protein intake (≥1.6g/kg/day) remain the most evidence-supported lean mass preservation interventions during any GLP-1 protocol — more robustly supported than compound selection at the current stage of the literature.

Further reading: Retatrutide Research Guide · Tirzepatide Research Guide · Semaglutide Research Guide