Muscle Loss on GLP-1 Medications: Why It Happens and What to Do About It

Medical Disclaimer: This article is for informational and educational purposes only and does not constitute medical advice. Consult a licensed healthcare provider before making any changes to your treatment. All pharmacological approaches mentioned beyond resistance exercise are either investigational or approved only for specific indications not covered here.

Semaglutide and tirzepatide produce weight loss results that were unthinkable a decade ago — 15 to 22 percent of body weight in the major clinical trials. For millions of people with obesity and its metabolic consequences, these results are genuinely transformative.

But a number that doesn’t appear in most headlines deserves more attention: roughly 25 to 40 percent of the weight lost on GLP-1 medications comes from lean mass — the muscle, bone mineral density, and organ tissue that is not body fat.

For someone losing 50 pounds on semaglutide, that could mean 12 to 20 pounds of muscle lost alongside the fat. Whether that matters — and what to do about it — is one of the most important practical questions facing the tens of millions of people now taking these medications.

First: Why Weight Loss Always Involves Some Muscle Loss

Muscle loss during weight loss is not unique to GLP-1 medications. Any form of caloric restriction causes some lean tissue loss alongside fat reduction. This is a fundamental feature of human metabolism, not a drug-specific side effect.

The underlying biology: skeletal muscle is in a continuous state of protein turnover — muscle protein synthesis (MPS) and muscle protein breakdown (MPB) happen simultaneously, and their relative rates determine whether you gain, maintain, or lose muscle. In the fed state, insulin and amino acid availability activate the mTOR/PI3K-Akt/S6K1 pathway, driving MPS. In the fasted or calorie-restricted state, mTOR activity drops, IGF-1 falls, and the body shifts toward net protein catabolism — breaking down muscle protein to provide amino acids for gluconeogenesis and other metabolic demands.

During prolonged caloric restriction, this process is compounded: reduced protein intake (because appetite is suppressed), lower circulating insulin (because energy intake is down), and reduced IGF-1 (which tracks with caloric state) all contribute to net muscle breakdown. The rate of lean mass loss during weight loss depends on:

Rate of weight loss: Faster caloric deficits give the body less time to adapt and tend to produce higher proportions of lean mass loss
Protein intake: Higher protein intake during caloric restriction preserves lean mass by maintaining amino acid availability for MPS
Resistance exercise: Mechanical loading of muscle fibers activates mTOR through mechanosensors (integrins, FAK/Src kinase pathway) — a nutrient-independent MPS signal that competes with the catabolic environment
Hormonal environment: Testosterone, estrogen, IGF-1, and growth hormone all influence muscle protein balance; deficiencies in any of these amplify lean mass loss during weight reduction

The Molecular Atrophy Pathway: Why Muscle Protein Breaks Down

Understanding the precise molecular mechanism driving muscle breakdown during caloric restriction explains why the interventions that work — protein intake and resistance exercise — are effective, and why they are approximately additive in their protection.

The FOXO-atrophy program: The upstream controller of skeletal muscle protein breakdown is the FOXO family of transcription factors — specifically FOXO1 and FOXO3 in muscle. Under normal fed conditions, insulin and IGF-1 signaling maintains elevated PI3K → Akt (protein kinase B) activity. Active Akt phosphorylates FOXO at specific serine and threonine residues, which sequesters FOXO proteins in the cytoplasm where they cannot reach their nuclear target genes. The muscle atrophy program is held off as long as insulin and IGF-1 signaling remains active.

During GLP-1-induced caloric restriction, insulin and IGF-1 fall substantially. Akt activity drops. Unphosphorylated FOXO1 and FOXO3 translocate into the nucleus and activate two muscle-specific E3 ubiquitin ligases:

MuRF1 (Muscle RING Finger 1): Targets structural myofibrillar proteins — particularly myosin heavy chains — for ubiquitin tagging. Polyubiquitinated myosin is then directed to the 26S proteasome for degradation. MuRF1 is the primary executioner of contractile protein breakdown during atrophy.
Atrogin-1 (MAFbx): Targets muscle regulatory transcription factors (MyoD) and components of the protein synthesis machinery (eIF3f), suppressing the muscle’s anabolic capacity at the same time that MuRF1 is degrading structural proteins. Atrogin-1 acts as a meta-regulator: it compounds the atrophy by reducing the muscle’s ability to resist it.

Together, FOXO-driven MuRF1 and Atrogin-1 upregulation produces the accelerated muscle protein breakdown of caloric restriction — not through nonspecific proteolysis, but through targeted degradation of thick filaments and anabolic machinery.

Why protein and resistance exercise target this pathway at the root:

Dietary leucine activates mTOR through the Sestrin2-GATOR2 amino acid sensing complex — a pathway that requires mechanistic detail to understand why leucine specifically (and not all amino acids equally) suppresses the atrophy program. Leucine enters myocytes primarily via the LAT1/SLC7A5 large neutral amino acid transporter. Once intracellular, leucine directly binds Sestrin2 at a hydrophobic leucine-binding pocket in Sestrin2’s C-terminal domain; in the leucine-free state, Sestrin2 maintains constitutive inhibitory interaction with the GATOR2 complex (a five-subunit scaffold: MIOS, WDR24, WDR59, SEH1L, SEC13). When leucine occupies Sestrin2, the Sestrin2–GATOR2 interaction is disrupted → Sestrin2 releases GATOR2 → GATOR2 is now free to inhibit GATOR1 (a four-subunit GTPase-activating protein: DEPDC5, NPRL2, NPRL3, and KICSTOR scaffold). GATOR1 would otherwise act as a GAP (GTPase-activating protein) for RagA/B, keeping RagA/B in the inactive GDP-loaded state. With GATOR1 inhibited by GATOR2, RagA/B remains in the GTP-loaded active state. GTP-RagA/B forms an obligate heterodimer with GDP-RagC/D → this Rag heterodimer is anchored at the lysosomal surface by the Ragulator pentamer complex (LAMTOR1–5) → GTP-RagA–RagC heterodimer binds mTORC1 through the Raptor subunit → mTORC1 is recruited to the lysosomal membrane where it encounters Rheb-GTP (the GTPase that directly activates mTORC1 kinase activity) → mTORC1 is activated → phosphorylates S6K1 (at Thr389) and 4E-BP1 (at Thr37/46 then Ser65/Thr70) → ribosome biogenesis and cap-dependent mRNA translation → muscle protein synthesis (PMID: 21036374). Separately, mTORC1 activity promotes Akt phosphorylation through a positive feedback loop, which then phosphorylates FOXO back into cytoplasmic sequestration — suppressing MuRF1 and Atrogin-1 gene transcription. Adequate leucine at each meal is therefore not merely substrate provision; it is upstream atrophy pathway inhibition via a specific amino acid sensor (Sestrin2) that feeds directly into the mTOR/FOXO axis.
Resistance exercise activates mTOR through integrin-mediated mechanosignaling (integrins → FAK/Src kinase → PI3K → Akt → mTOR) — a nutrient-independent route that independently suppresses FOXO nuclear translocation. The mechanical MPS signal competes with the FOXO atrophy program even in a caloric deficit.

Because protein and resistance exercise suppress the FOXO atrophy program through different upstream signals (leucine-nutrient vs. mechanical-integrin), their effects are approximately additive — and the reason the combination is additive rather than redundant is specific to the two-step architecture of mTORC1 activation. Leucine’s Sestrin2/GATOR2/RagA-B-GTP pathway controls mTORC1 recruitment to the lysosomal surface via the Ragulator pentamer complex (LAMTOR1–5); this step delivers the mTORC1 complex to the lysosomal membrane but does not activate its kinase domain. Mechanical/integrin/FAK/PI3K/Akt/TSC2 signaling controls mTORC1 activation at the lysosomal surface by maintaining Rheb in the GTP-loaded state — Rheb-GTP directly contacts the FRB domain of mTOR and triggers catalytic activity; this step activates the mTORC1 kinase but only when mTORC1 is already membrane-recruited via the leucine/RagA-B/Ragulator arm. The two steps are sequentially required and non-redundant: without leucine → Sestrin2 remains in GATOR2-inhibiting conformation → GATOR1 remains active → RagA-B stays GDP-loaded → Ragulator cannot dock mTORC1 at the lysosomal surface → mTORC1 cannot encounter Rheb regardless of Akt activity and Rheb-GTP abundance; without mechanical/Akt/TSC2 suppression → Rheb remains GDP-loaded → mTORC1 recruited to the membrane is not activated regardless of leucine availability. Because each pathway unlocks a distinct rate-limiting bottleneck in the same sequential activation architecture — leucine-driven membrane recruitment and Akt-driven kinase activation — combining them produces maximal mTORC1 output where either alone produces partial activation at best. The effects are additive (each removes a separate bottleneck) rather than redundant (which would imply they compete for the same regulatory node) or supra-additive (which would imply each amplifies the other’s signal amplitude). This is the molecular basis for the consistent finding that combining resistance exercise with protein optimization preserves substantially more lean mass than either alone.

Why GLP-1 Medications May Be Particularly Challenging for Lean Mass

The appetite suppression mechanism of GLP-1 medications creates a specific pattern of reduced intake that may compound lean mass loss beyond what the caloric deficit alone would predict.

GLP-1 receptors in the hypothalamus (arcuate nucleus) and brainstem (nucleus tractus solitarius) signal satiety — both from gut-derived endogenous GLP-1 and from pharmacological doses. This suppression is non-specific: patients eat less of everything, including protein. People on GLP-1 medications commonly report difficulty meeting protein targets not because they’re making poor food choices, but because they’re simply not hungry enough to eat adequate volume.

The result is a doubly compounded problem: caloric restriction reduces mTOR signaling, and insufficient dietary protein reduces the substrate available for muscle protein synthesis. This is why the 1.2–1.6 g/kg/day protein target is so important and so often unmet in GLP-1 patients — it requires deliberate, structured protein intake at every meal, often prioritizing protein before other foods.

Additionally, GLP-1 therapy-induced rapid weight loss in patients with pre-existing low muscle mass (common in older adults with obesity) starts from a worse baseline. The proportion of lean mass lost is similar, but the absolute amount of muscle loss is clinically more significant when muscle reserve is already reduced.

Why the Muscle Loss Question Has Become Urgent

A January 2026 review in the Journal of Clinical Medicine by Arora, Conde, and Desouza (PMID: 41598480) at the University of Nebraska Medical Center is among the most systematic assessments of this problem to date. The authors frame it clearly: the massive and growing population using GLP-1 medications, combined with the scale of weight loss these drugs produce, has made muscle preservation a first-tier clinical priority.

Why lean mass preservation matters:

Metabolic function: Skeletal muscle is the primary site of glucose disposal — accounting for approximately 70–80% of insulin-mediated glucose uptake. The mechanism operates through GLUT4 vesicle translocation: when insulin binds the insulin receptor (IR) on the myofiber surface, IR autophosphorylates its β-subunit tyrosines and recruits IRS-1, which activates PI3K to generate PIP3 at the inner plasma membrane leaflet. PIP3 recruits PDK1 and mTORC2, which phosphorylate Akt at Thr308 and Ser473, respectively. Activated Akt then phosphorylates AS160 (TBC1D4) at Thr642 and Ser588, suppressing its Rab-GAP activity. With AS160 inhibited, Rab8A and Rab14 on GLUT4 storage vesicles (GSVs) remain GTP-loaded, enabling KIF5B kinesin-mediated translocation of the GSV pool from its perinuclear depot to the sarcolemma. There, SNARE complex assembly — VAMP2 on the vesicle pairing with syntaxin-4 and SNAP23 on the plasma membrane — drives membrane fusion and GLUT4 insertion. Each GLUT4 molecule inserted increases the sarcolemmal glucose-transport capacity, and an insulin-stimulated myofiber can increase its surface GLUT4 density fourfold or more. Loss of muscle mass therefore directly reduces the total number of GLUT4-expressing myofibers available per insulin stimulus; fewer fibers mean fewer translocation events and a proportionally reduced whole-body glucose disposal capacity, manifesting clinically as systemic insulin resistance. This creates a tension: you lose weight for metabolic health, but the way you lose it may partly undermine the metabolic goal.

Long-term weight maintenance: Muscle mass is metabolically active tissue. Reduced muscle from weight loss lowers basal metabolic rate, which is a known contributor to weight regain. The cellular mechanism is more specific than it sounds: skeletal muscle at rest expends energy primarily through three processes. First, sarcolemmal Na⁺/K⁺-ATPase (the sodium-potassium pump) is the dominant consumer — each pump hydrolyzes one ATP per cycle to export 3 Na⁺ and import 2 K⁺, maintaining the electrochemical gradient across the myofiber membrane, and this basal ion-gradient maintenance accounts for roughly 20–25% of resting skeletal muscle energy expenditure (PMID: 10938009). Second, resting myosin ATPase activity — even non-contracting muscle maintains sarcomere structural integrity through low-level actin-myosin cross-bridge cycling (rigor-to-relaxed transitions), which requires continuous ATP hydrolysis. Third, the muscle protein turnover cycle (MPS − MPB) is energetically costly even when net balance is zero: both ubiquitin-proteasome degradation and ribosomal translation require ATP. When GLP-1-driven weight loss eliminates myofiber mass, every lost kilogram of skeletal muscle removes a proportional quantity of sarcolemmal Na⁺/K⁺-ATPase pumps, sarcomere contractile units, and ribosomal/proteasomal machinery. The aggregate result is a lower total resting energy expenditure — the same caloric intake that previously maintained body weight now generates a positive energy balance, producing persistent weight-regain pressure that is independent of appetite or behavioral relapse. Patients who regain weight after stopping GLP-1 medications — and many do; the SURMOUNT-4 discontinuation trial showed average regain of two-thirds of lost weight over 48 weeks — tend to regain primarily fat rather than muscle, resulting in a worse body composition than they started with. The mechanistic asymmetry driving this fat-dominant regain is kinetic and pathway-specific. When caloric intake resumes after GLP-1 discontinuation, insulin elevation activates SREBP-1c (sterol regulatory element-binding protein 1c) in hepatocytes and adipocytes within hours — SREBP-1c is a membrane-bound transcription factor cleaved by S1P/S2P proteases when the ER senses insulin-driven PI3K/Akt signaling, releasing its N-terminal domain to the nucleus where it transactivates FASN, ACC1, and SCD1, upregulating de novo lipogenesis. In adipocytes, the same insulin/Akt signal phosphorylates PDE3B (phosphodiesterase 3B), rapidly depleting cAMP → PKA activity falls → HSL (hormone-sensitive lipase) loses its activating phosphorylation at Ser563 and Ser660, and perilipin-A loses its PKA-mediated phosphorylation that normally allows cytoplasmic lipase access to the lipid droplet surface. With HSL dephosphorylated and perilipin-A stabilizing the droplet coat, lipolysis is arrested; simultaneously, DGAT1 (diacylglycerol acyltransferase 1) esterifies fatty acids into triglycerides for storage. The net result: fat re-accumulation is initiated within hours of resumed eating because it requires only substrate (dietary fat and glucose) and insulin — both immediately available. Muscle re-accretion operates on a fundamentally different regulatory logic: it requires integrin/FAK/mTORC1 mechanical signaling (generated only by resistance exercise-induced contractile strain, not by caloric surplus) AND per-meal leucine threshold activation of the Sestrin2/GATOR2/RagA/mTORC1 pathway (requiring ≥2–3g leucine per meal to fully activate mTORC1 at the lysosomal surface, regardless of total caloric intake). Returning appetite and weight regain provide neither input. The result is weeks-to-months of consistent resistance training plus adequate per-meal protein intake required before meaningful myofibrillar protein accretion occurs — while fat storage reactivates the same day eating resumes. This phenomenon, sometimes called “weight cycling with body composition disadvantage,” makes each subsequent weight loss attempt more difficult metabolically.

Physical function and falls: Sarcopenia — age-related muscle loss — is associated with increased fall risk, reduced independence, and higher healthcare utilization in older adults. Accelerating lean mass loss in a patient population that skews toward middle age and older creates a meaningful functional risk. The mechanism linking muscle loss to falls is more specific than general “weakness”: sarcopenia preferentially depletes type II (fast-twitch) muscle fibers, and the reason this is catastrophic for fall prevention is that postural recovery from a balance perturbation depends on peak force production velocity, not steady-state strength.

The fiber-type selectivity of disuse atrophy operates through motor unit physiology. Skeletal muscle fibers are innervated in threshold-ordered motor unit pools: type I (slow-twitch) fibers are recruited by low-threshold motor neurons (small soma diameter, lower rheobase) that fire continuously during low-intensity activities — standing, walking, sustained posture. Type II (fast-twitch) fibers are recruited only by high-threshold motor neurons that fire during high-force, high-velocity tasks — rapid direction changes, stair negotiation, reactive stepping. In sedentary individuals and in aging, high-threshold motor units are selectively underrecruited: the type II motor neuron is rarely brought above its activation threshold during routine daily activity, resulting in chronic electrical silence of type II fibers. Chronically silenced type II fibers reduce protein synthesis because mTORC1 activation in skeletal muscle is mechanoresponsive — it requires the upstream activator mechanosensitive kinase FAK (focal adhesion kinase) to phosphorylate the integrin-linked kinase (ILK)/mTORC1 activation scaffold, and FAK is only activated by cyclic mechanical strain generated during active fiber contraction. A type II fiber that is never recruited generates no contractile strain → no FAK phosphorylation → reduced mTORC1 → reduced muscle protein synthesis → net negative protein balance → type II fiber cross-sectional area decreases preferentially, even while type I fiber area is maintained by their ongoing low-threshold tonic recruitment. GLP-1-driven caloric restriction deepens this asymmetry: it reduces circulating IGF-1 (via hepatic GH signaling suppression under caloric deficit) and reduces the amino acid substrate availability for MPS, further suppressing the already-silenced type II fiber’s protein synthesis to near zero while the tonic-firing type I fibers maintain partial protein turnover through continued mechanical stimulation.

The fall risk consequence follows from this fiber-type geometry. When a perturbation — a foot catching an uneven surface, a sudden push, an unexpected weight shift — displaces the center of mass toward the edge of the base of support, the central nervous system has approximately 100–200 milliseconds to initiate a compensatory stepping response before the center of mass crosses the base boundary and a fall becomes mechanically inevitable. This reactive stepping response depends on the ankle dorsiflexors (tibialis anterior), hip abductors (gluteus medius/minimus), and knee extensors (quadriceps) generating rapid, high-force contractions in sequence to reposition the foot, arrest the toppling moment, and restore equilibrium. These are high-threshold type II motor unit tasks: the required contraction velocity and force amplitude — sufficient to accelerate the limb in 80–120 ms — falls entirely in the type II fiber’s performance domain. The neuromuscular pathway is: vestibular/proprioceptive perturbation → brainstem vestibular nuclei → corticospinal tract + lateral reticulospinal tract → high-threshold α-motor neurons → type II fiber contraction — a chain that requires type II fibers to be present, innervated, and of adequate cross-sectional area to generate the required force impulse within the available time window. When GLP-1-accelerated lean mass loss has already depleted type II fiber area on top of pre-existing sarcopenic loss, this force-production capacity is below threshold: the reactive stepping response initiates correctly at the neural level but the muscle cannot generate adequate ankle dorsiflexion and hip abduction velocity, the compensatory step undershoots, and the center of mass clears the base of support — producing a fall that a younger or better-preserved neuromusculoskeletal system would have arrested. This is the specific mechanistic pathway through which GLP-1-accelerated lean mass loss translates to elevated fall risk in older patients — not general weakness, but selective type II fiber depletion eroding the peak force velocity reserve that reactive postural control depends on.

Bone density: Lean mass loss is associated with reductions in bone mineral density. GLP-1 medications appear to also reduce bone formation markers. Falls from muscle weakness compound bone fracture risk in a population where osteoporosis is already prevalent.

The bone loss mechanism during GLP-1-driven weight loss operates through three converging arms. Mechanical unloading is primary: skeletal loading is sensed by osteocytes through two parallel mechanosensory pathways operating in the lacuno-canalicular network — a system of fluid-filled microscopic channels connecting osteocytes through cortical bone. In the primary cilia pathway, each osteocyte extends one non-motile primary cilium into the canalicular fluid space; the ciliary membrane houses polycystin-1 (PC1/PKD1) and polycystin-2 (PC2/PKD2) — the PC1/PC2 heteromeric complex functions as a mechanically gated Ca²⁺-permeable channel; cyclic mechanical strain in bone generates pulsatile fluid flow through the lacuno-canalicular network → fluid shear deflects the primary cilium → PC2 opens → Ca²⁺ entry into the osteocyte cytoplasm; simultaneously, the cilia-localized isoform adenylyl cyclase III (AC3) is activated by the local Ca²⁺ rise → cAMP accumulation → PKA → CREB Ser133 phosphorylation in osteocytes → CREB occupies repressive response elements in the SOST (sclerostin) gene promoter, suppressing sclerostin transcription. In the Piezo1 pathway, deformation of the bone matrix creates membrane tension directly in the osteocyte plasma membrane → Piezo1 stretch-activated ion channels open → Ca²⁺ influx → calcineurin → NFAT nuclear translocation in osteocytes → NFAT drives upregulation of Wnt ligands and further SOST repression. Both Ca²⁺ streams converge: active mechanical loading holds the SOST promoter repressed (via CREB through the cilia/AC3 arm and NFAT through the Piezo1 arm), keeping sclerostin output low and Wnt/β-catenin signaling intact in adjacent osteoblast precursors. When body weight falls during GLP-1-driven weight loss, tibial and femoral osteocytes register reduced cyclic strain → reduced lacuno-canalicular fluid flow → less PC2/Ca²⁺/AC3/cAMP/CREB and less Piezo1/Ca²⁺/NFAT signaling → the SOST promoter is de-repressed → sclerostin output rises. Low strain-driven upregulation of sclerostin (SOST), an osteocyte-secreted Wnt antagonist. Sclerostin binds the LRP5/LRP6 co-receptors on osteoblast precursors, preventing Wnt ligands from activating LRP5/6 and Frizzled — this keeps the cytoplasmic destruction complex (APC/Axin/GSK-3β/CK1) constitutively active, continuously phosphorylating β-catenin at Thr41/Ser37/Ser33 → β-catenin ubiquitinated by β-TrCP E3 ligase → proteasomal degradation → less nuclear β-catenin/LEF/TCF transcriptional activity → reduced expression of RUNX2, SP7/Osterix, and downstream targets COL1A1 and osteocalcin → fewer committed osteoblasts → depressed bone formation. P1NP (procollagen type I N-terminal propeptide) and osteocalcin, the clinical formation markers, fall within 8–12 weeks of significant weight loss onset (PMID:34550405).

The second arm is IGF-1 reduction: caloric restriction lowers hepatic IGF-1 output, and osteoblasts express IGF-1R abundantly. Osteoblast IGF-1R/IRS-1/PI3K/Akt/mTORC1 signaling drives osteoblast proliferation and COL1A1 synthesis; when circulating IGF-1 falls proportionally to the degree of caloric deficit, osteoblast mTORC1 activity declines, amplifying the SOST/Wnt-driven suppression of bone formation. A third, GLP-1-specific contribution involves slowed gastric emptying: delayed chyme transit to the duodenum/jejunum reduces the postprandial calcium absorption spike because active transcellular Ca²⁺ transport via TRPV6 and calbindin-D9k in duodenal enterocytes is concentration- and transit-time-dependent — slower delivery means less peak luminal calcium and less absorptive surface contact per meal, reducing mineralization substrate availability.

GLP-1 receptors on osteoclasts partially offset these formation losses. The mechanism by which GLP-1R/Gs/cAMP/PKA signaling reduces osteoclast activity is worth understanding because it clarifies both why the protection exists and why it proves insufficient. GLP-1R activation in osteoclast precursors elevates cAMP, which activates two parallel effectors: PKA and Epac1 (exchange protein directly activated by cAMP). PKA phosphorylates components of the RANKL/RANK/TRAF6 signaling complex that drives osteoclast differentiation — specifically, PKA-dependent phosphorylation of TRAF6-associated kinases reduces their ability to activate the JNK/AP-1 and NF-κB arms of the RANK pathway. Epac1 operates through a distinct route: Epac1 is a guanine nucleotide exchange factor for Rap1 GTPase; Epac1/Rap1 signaling inhibits c-Fos induction downstream of RANKL → reduced AP-1 transcriptional activity at the NFATc1 promoter. NFATc1 is the master transcription factor for osteoclast terminal differentiation: it drives expression of cathepsin K (the cysteine protease that cleaves type I collagen at the bone surface), TRAP (tartrate-resistant acid phosphatase), MMP-9 (matrix metalloproteinase for collagen degradation), and integrin αvβ3 (which anchors the mature osteoclast sealing zone to hydroxyapatite). Reduced NFATc1 nuclear translocation means fewer mature osteoclasts are generated and those that differentiate express lower levels of the resorptive machinery — directly explaining why CTX (a collagen cross-link fragment released by cathepsin K) and NTX fall in GLP-1 agonist users.

The paradox is that this cAMP/PKA/Epac1/Rap1/NFATc1-mediated resorption suppression, while real and measurable, is numerically overwhelmed by the formation-side suppression from weight-loss-driven SOST upregulation and IGF-1 reduction. The ratio of formation-to-resorption suppression — where formation falls more than resorption during active caloric deficit — is the mechanistic reason net bone mineral density declines over 12–24 months despite the GLP-1R osteoclast-protective effect being present throughout. Resistance exercise provides a targeted countermeasure: muscle force transmitted through tendon insertions to cortical bone generates local osteocyte strain precisely at enthesis sites, suppressing SOST locally → restoring Wnt/β-catenin → maintaining RUNX2/osteocalcin at insertion zones — a spatially targeted rescue of the formation program that no pharmacological approach fully replicates. This represents a second distinct benefit of resistance training during GLP-1 therapy, independent of its lean mass preservation effects.

Sex Differences in GLP-1 Lean Mass Loss

Women may face higher lean mass loss risk during GLP-1 therapy than men, for several biological reasons:

Hormonal baseline: Estrogen has anabolic effects on muscle — it augments mTOR signaling and muscle protein synthesis through estrogen receptor alpha (ERα) activation in myocytes. Postmenopausal women who have lost this hormonal support have a reduced baseline MPS capacity. Adding the catabolic pressure of caloric restriction on top of an already-reduced anabolic environment accelerates lean mass loss disproportionately.

The ERα mechanism in skeletal muscle operates through two distinct arms. In the non-genomic arm (acute, minutes): estradiol binds membrane-associated ERα on the sarcolemma → ERα directly associates with the p85 regulatory subunit of PI3K (the same PI3K recruited downstream of IGF-1R) → PI3K activation → PIP₂→PIP₃ → PDK1 → Akt Thr308/Ser473 phosphorylation → mTORC1 → S6K1/4E-BP1 → muscle protein synthesis. This rapid non-genomic ERα-PI3K coupling means estradiol amplifies the same Akt-mTOR signaling that leucine and IGF-1 activate — effectively lowering the anabolic threshold. In the genomic arm (hours to days): nuclear ERα binds estrogen response elements (EREs) in the promoters of IGF1R (IGF-1 receptor) and IRS1 (insulin receptor substrate-1) genes → increased transcription of IGF-1 receptor and IRS-1 protein → greater signal amplification at each step of the IGF-1 → PI3K → Akt cascade per unit of available IGF-1 (PMID:25526573). The combined effect is that estrogen both directly activates mTOR (non-genomic) and sensitizes the entire insulin/IGF-1 signaling axis (genomic) — a dual mechanism that sets a higher baseline MPS capacity in premenopausal women. After menopause, both arms are lost simultaneously, reducing mTOR responsiveness to leucine and IGF-1 signals and contributing to the anabolic resistance observed in postmenopausal women even before any caloric restriction begins.

Starting muscle mass: Women have lower absolute muscle mass than men on average, even when controlling for body weight. A 30% lean mass loss from caloric restriction represents a higher fraction of total muscle reserve in a woman who started with less. The underlying biology is androgenic: testosterone acts through androgen receptor (AR) in muscle satellite cells — the resident stem cells that donate new myonuclei to muscle fibers after damage or growth stimulus — to drive satellite cell proliferation and the addition of new myonuclei. More myonuclei expand the myonuclear domain (the volume of cytoplasm each nucleus can support) and increase the transcriptional capacity of the fiber for contractile protein synthesis. Men accumulate a substantially larger myonuclear complement during puberty and early adulthood than women (testosterone-driven AR signaling during this window is responsible for the higher fiber cross-sectional area and total myonuclear number that defines male lean mass advantage). Women entering GLP-1 therapy therefore start with a smaller muscle reserve and a lower myonuclear ceiling — meaning the same absolute 30% lean mass loss removes a proportionally greater share of contractile capacity and leaves less reserve to defend against further atrophy.

Protein intake challenges: Women who begin GLP-1 therapy with protein intake already near or below optimal for lean mass maintenance face a specific mechanistic vulnerability. GLP-1-mediated appetite suppression reduces total food volume consumed at each meal. The minimum per-meal leucine dose required to activate the Sestrin2/GATOR2/mTORC1 signaling complex — the leucine sensor controlling muscle protein synthesis (MPS) — is approximately 2–3 g leucine, equivalent to roughly 25–30 g of high-quality complete protein. Below this threshold, even if total daily protein appears adequate on paper, individual meals fail to generate a robust mTORC1/S6K1/4E-BP1 signaling response, and MPS is not meaningfully stimulated. A woman who was previously eating 25 g protein per meal (just at threshold) will, after GLP-1-induced meal size compression, consume 15–18 g per meal — reliably below the Sestrin2 activation threshold at every eating occasion. Total daily protein may drop from 80 g to 50 g, but the critical failure is that no single meal achieves leucine threshold. The mathematical consequence: net protein balance (MPS minus muscle protein breakdown) remains negative at every meal, and lean mass erodes continuously even when reported dietary intake looks “reasonable” on a daily gram total. This per-meal threshold failure is more likely in women because they start closer to threshold; men with higher baseline intake have more margin before GLP-1 meal size reduction breaches the leucine sensor floor.

The practical implication: postmenopausal women on GLP-1 therapy need particularly aggressive lean mass monitoring and particularly disciplined protein/resistance exercise protocols. DEXA scanning is especially warranted in this population.

What the Clinical Trial Data Shows

The landmark GLP-1 trials measured total body weight but not all provided detailed body composition analysis. The data that exists is concerning but also reveals important nuance.

In the STEP trials (semaglutide 2.4mg), participants lost an average of 15.3% body weight over 68 weeks. Studies using DEXA scanning to examine body composition found that approximately 30% of the lost weight came from lean mass.

In the SURMOUNT trials (tirzepatide), total weight loss was larger (up to 22.5% at the highest dose), and body composition data showed similar proportions of lean to fat loss. This outcome is mechanistically informative: tirzepatide adds GIP receptor (GIPR) agonism to GLP-1R agonism, yet GIPR co-activation does not improve the lean-to-fat loss ratio, and the reason lies in what GIPR signaling actually does in skeletal muscle. GIPR is a Gs-coupled GPCR; in skeletal myocytes, its activation raises cAMP → PKA → phosphorylation of pyruvate dehydrogenase kinase 4 (PDK4) and hormone-sensitive lipase (HSL), shifting intramyocellular substrate oxidation toward glucose and fatty acids (improved energy substrate utilization). The PDK4 mechanism deserves unpacking: PDK4 is a mitochondrial matrix kinase that normally phosphorylates the E1α subunit of the pyruvate dehydrogenase complex (PDC) at Ser293, Ser300, and Ser232 → phospho-E1α cannot bind its thiamine pyrophosphate (TPP) cofactor in the correct orientation → PDC is catalytically inactivated → pyruvate cannot be decarboxylated to acetyl-CoA → less carbon enters the TCA cycle → glucose oxidation is constrained. When GIPR/Gs/cAMP elevates PKA activity in myocytes, PKA phosphorylates PDK4 itself at regulatory Ser32 and Ser88 → this phosphorylation reduces PDK4’s affinity for PDC E1α (negative autoregulatory loop) → PDK4 kinase activity toward its PDH substrate falls → E1α Ser293/Ser300 phosphorylation is reversed by the constitutively active pyruvate dehydrogenase phosphatase (PDP1/PDP2) → PDC returns to its dephosphorylated, active state → pyruvate flux into acetyl-CoA resumes → glucose is oxidized efficiently through the TCA cycle. Simultaneously, PKA-mediated HSL phosphorylation at Ser563/Ser659/Ser660 drives intramyocellular lipid droplet triglyceride hydrolysis → long-chain fatty acids released for mitochondrial β-oxidation, feeding additional acetyl-CoA into the TCA cycle. The two arms together — restored glucose oxidation through PDK4 disinhibition of PDC, and activated fatty acid oxidation through HSL — improve myocellular energetics and insulin sensitivity. This is metabolically beneficial for insulin sensitivity, but it is not an anabolic signal — GIPR does not couple to IGF-1R/IRS-1/PI3K/PIP₃/PDK1/Akt-Thr308, so it does not phosphorylate and cytoplasmic-sequester FOXO1/FOXO3 via 14-3-3 proteins, meaning it exerts no suppression of the MuRF1 (TRIM63) and Atrogin-1/MAFbx (FBXO32) E3 ubiquitin ligase program that drives myofibrillar protein degradation. Simultaneously, tirzepatide’s superior efficacy produces a larger total caloric deficit, and this larger deficit drives a proportionally greater VMH glucokinase/K-ATP → CRH/PVN → ACTH → StAR/cortisol HPA axis response — more cortisol → more GR/GRE-driven FOXO1 and REDD1 transcription → more mTORC1 suppression and ubiquitin-proteasome activation in myocytes. The net effect: GIPR’s marginal glucose oxidation advantage in muscle is offset by the greater catabolic hormonal environment generated by a larger caloric deficit, producing lean/fat loss ratios that converge with semaglutide despite the mechanistically distinct receptor pharmacology.

Importantly, participants in these trials did not receive structured resistance exercise protocols or high-protein dietary counseling aimed specifically at lean mass preservation — these were drug efficacy trials, not optimization trials. This is significant: the 25–40% lean mass loss figure likely represents a near-worst-case scenario relative to what could be achieved with concurrent muscle-protective interventions. When resistance exercise and protein-focused dietary counseling are added, early data suggests lean mass preservation substantially improves.

Rate of Loss Matters: Slower Can Preserve More Muscle

An underappreciated variable in lean mass preservation is the rate at which weight is lost. Rapid caloric restriction (>1.5% body weight per week) is consistently associated with higher lean mass loss percentages than more gradual weight reduction in the general obesity literature.

The molecular mechanism explains why rate — not just magnitude — matters. Rapid caloric restriction triggers a hypothalamic-pituitary-adrenal (HPA) axis stress response, elevating cortisol in proportion to the severity of the energy deficit. The transduction pathway from energy deficit to elevated cortisol operates through a chain of neuroendocrine events. Hypothalamic glucose sensing: The ventromedial hypothalamus (VMH) and arcuate nucleus (ARC) harbor specialized glucose-sensing neurons that express glucokinase (hexokinase-IV), an enzyme with a high glucose Km (~8 mM) that functions as a kinetic glucose sensor. At euglycemia, glucose flux through glucokinase generates sufficient ATP → elevated intracellular ATP:ADP ratio → closure of K-ATP channels (Kir6.2/SUR1) → membrane depolarization → sustained neuronal firing. This tonic activity provides inhibitory input to CRH-secreting neurons, keeping the stress axis restrained during normal metabolic conditions. During rapid caloric restriction, blood glucose falls and glucose delivery to the VMH/ARC decreases → glucokinase activity falls disproportionately (low-affinity kinetics amplify the glucose signal) → less glycolytic ATP → K-ATP channels open → K⁺ efflux → membrane hyperpolarization → glucose-excited neurons become quiet → their tonic inhibition of CRH (corticotropin-releasing hormone) neurons in the paraventricular nucleus (PVN) is withdrawn. PVN → pituitary: Disinhibited CRH neurons fire and release CRH into the hypothalamo-pituitary portal vessels → CRH reaches anterior pituitary corticotrophs → binds CRH-R1 (Gs-coupled GPCR) → adenylyl cyclase → cAMP → PKA → immediate exocytosis of pre-formed ACTH granules + CREB-driven transcription of the POMC gene (processed by prohormone convertase 1/3 to yield ACTH). ACTH enters systemic circulation. Adrenal cortisol synthesis: ACTH binds MC2R (melanocortin-2 receptor) on adrenal zona fasciculata cells — a Gs-coupled receptor → cAMP → PKA acting on two timescales: (1) Acute — PKA phosphorylates StAR (steroidogenic acute regulatory protein) at Ser194 → StAR shuttles cholesterol from the outer to the inner mitochondrial membrane (the rate-limiting step in steroidogenesis) → CYP11A1 (P450scc) cleaves the side-chain to form pregnenolone → ER-localized HSD3B2 → CYP17A1 → CYP21A2 → CYP11B1 (11β-hydroxylase) complete the conversion to cortisol, which is released into circulation within minutes; (2) Chronic — PKA/CREB drives transcriptional upregulation of StAR, CYP11A1, CYP17A1, and CYP11B1 genes over hours to days, expanding steroidogenic capacity in proportion to sustained ACTH drive. The resulting plasma cortisol elevation scales with the magnitude of the energy deficit — explaining why faster, more severe caloric restriction produces more cortisol than a gradual deficit of equal cumulative magnitude. Cortisol binds the glucocorticoid receptor (GR) in myocytes; the ligand-bound GR translocates to the nucleus and binds glucocorticoid response elements (GREs) in the FOXO1 gene promoter, directly upregulating FOXO1 transcription. This is an active catabolic signal — not merely the passive loss of insulin/IGF-1 anabolic drive — which is why rapid restriction produces disproportionately higher lean mass loss than the caloric deficit alone would predict from the FOXO-atrophy model above. The GR simultaneously drives transcription of REDD1 (regulated in development and DNA damage response 1), an mTORC1 inhibitor: REDD1 releases TSC2 from its 14-3-3 sequestration → free TSC2 acts as a GAP for Rheb → Rheb-GTP is hydrolyzed to Rheb-GDP → mTORC1 is no longer activated by Rheb → S6K1/4E-BP1 are dephosphorylated → muscle protein synthesis is further suppressed. The result: at the same total caloric deficit endpoint, a faster trajectory generates more cumulative cortisol exposure → more GR/FOXO1 and GR/REDD1 signaling → both MPS suppression and accelerated MPB, amplifying lean mass loss beyond what slower, less stressful restriction produces.

GLP-1 medications, particularly at maximum doses, can produce weight loss rates of 1–2% of body weight per week in the early months. This is faster than many conventional dietary approaches, which may partly explain why lean mass loss on these drugs is a more salient concern than on slower interventions. The mechanism by which dose titration rate specifically affects lean mass — as opposed to total caloric deficit magnitude — is worth making explicit, because it explains why the recommendation to “titrate more slowly” is not merely cautious but has a defined molecular basis.

GLP-1R signaling in the hypothalamic arcuate nucleus suppresses NPY/AgRP orexigenic neurons in a dose-responsive manner: higher doses produce more complete arcuate satiety signaling → larger acute caloric deficit per meal than lower doses. When titration jumps from a lower to a maximum dose in a compressed timeframe (e.g., 4-week intervals), the step change in daily caloric intake is larger than if the same final dose is reached over 12–16 weeks. Each discrete dose step creates a corresponding acute energy deficit that the VMH glucokinase sensors register as a rate-of-change signal: the faster the deficit builds per step, the more pronounced the VMH glucose-excited neuron silencing → the larger the CRH disinhibition burst per step → the larger the ACTH surge → the more StAR Ser194 phosphorylation events per unit time → the higher the peak cortisol concentration achieved. This peak cortisol amplitude — not only the cumulative cortisol exposure — is the critical variable for myocyte GR nuclear translocation. The glucocorticoid receptor binds cortisol with a Kd of approximately 5–20 nM; at moderate cortisol concentrations (normal range 30–200 nM), GR occupancy is partial and GR nuclear translocation is proportional. During a rapid titration step generating a cortisol surge into the 400–600 nM range, GR occupancy approaches saturation → the nuclear-to-cytoplasmic GR ratio reaches its maximum → FOXO1 gene transcription and REDD1 gene transcription are driven at maximum amplitude simultaneously → the most severe convergent suppression of mTORC1 (via REDD1/TSC2/Rheb) and the most aggressive activation of MuRF1/Atrogin-1 (via FOXO1 nuclear accumulation) occurs per dose step.

Slow titration prevents these peak cortisol saturation events. With extended intervals between dose steps, the magnitude of energy deficit at each transition is smaller → glucokinase-sensing is graded rather than threshold-crossed → CRH/ACTH/cortisol responses per step are proportionally attenuated → GR remains below saturation → nuclear FOXO1 and REDD1 induction per step is submaximal. Additionally, the stable periods between slow titration steps allow partial recovery of the muscle protein synthesis machinery: during plateau periods at a given dose, dietary leucine availability (Sestrin2/GATOR2/mTORC1) and mechanical signaling from exercise (integrin/FAK/Akt) continue operating at their reduced but non-zero level. Each subsequent dose step therefore begins from a partially recovered mTORC1/FOXO3a baseline rather than a continuously maximally suppressed one. The cumulative lean mass loss over the full titration trajectory may be similar in total caloric deficit achieved, but the plateau-and-step pattern of slow titration keeps the peak amplitude of the catabolic signaling below the level that fully saturates GR and eliminates residual mTORC1 activity — producing a meaningfully better lean-to-fat composition of loss along the way.

This doesn’t mean lower GLP-1 doses are always preferable — the metabolic benefits are dose-dependent. But it does mean that for patients with significant lean mass concerns (older adults, those with pre-existing sarcopenia, competitive athletes), discussion with their prescriber about dose titration rate — going up more slowly — may be worth having. The cortisol peak amplitude argument provides a molecular rationale for why a slightly slower trajectory of weight loss may preserve more lean tissue even if the total weight loss endpoint is the same.

What You Can Do Right Now: The Evidence

1. Resistance Exercise (High Priority, Strong Evidence)

This is the most important practical recommendation and the one with the strongest evidence base for anyone losing weight — on GLP-1 medications or otherwise.

Resistance exercise (weightlifting, resistance bands, bodyweight exercises, machines) activates mTOR through mechanical mechanosensors in muscle fiber membranes (integrins and FAK/Src kinase), independent of nutritional status. This mechanically-driven MPS signal competes with the catabolic environment of caloric restriction. The molecular cascade that drives this is now well-characterized: α7β1 integrin (the dominant mechanosensory integrin in skeletal muscle, linking intracellular costamere scaffolds to laminin-211 in the endomysium) senses tensile force during contraction → cytoplasmic vinculin/talin domains undergo force-induced conformational unfolding → expose cryptic binding sites for scaffolding proteins at focal adhesion plaques → FAK (focal adhesion kinase) Tyr397 autophosphorylation in trans within focal adhesion clusters → Src kinase SH2 domain binds pFAK Tyr397 → Src phosphorylates FAK at Tyr925 → this pTyr925 site recruits the p85 regulatory subunit of PI3K → PI3K converts PIP₂ to PIP₃ at the inner sarcolemmal leaflet → PIP₃ recruits PDK1 and activates mTORC2 → Akt is phosphorylated at Thr308 (PDK1) and Ser473 (mTORC2) → dual-phosphorylated Akt phosphorylates TSC2 at Thr1462 → TSC2 GTPase-activating function toward Rheb is inactivated → Rheb-GTP accumulates at the lysosomal membrane → Rheb-GTP directly activates mTORC1 kinase (through the FRB domain of mTOR) → mTORC1 phosphorylates S6K1 at Thr389 and 4E-BP1 at Thr37/46 → ribosomal S6 protein and cap-dependent mRNA translation → net muscle protein synthesis. Simultaneously, the same active Akt phosphorylates FOXO3a at Ser253 and Ser315 → 14-3-3ζ binds phospho-FOXO3a → FOXO3a is sequestered in the cytoplasm → cannot enter the nucleus to transcribe MuRF1 and Atrogin-1 → contractile protein ubiquitination is actively suppressed during the post-exercise anabolic window. A parallel mechanosensory arm operates through Piezo1 stretch-activated ion channels on the sarcolemma: mechanical deformation during contraction gates Piezo1 → transient Ca²⁺ entry → intracellular Ca²⁺ spike activates CaMKII (Ca²⁺/calmodulin-dependent protein kinase II) at Thr286 (autophosphorylation) → CaMKII phosphorylates and inhibits eEF2K (elongation factor 2 kinase) → eEF2K can no longer phosphorylate eEF2 at Thr56 → eEF2 remains GTP-competent and ribosome-associated → translational elongation rates are maintained during and after contraction-induced mTORC1 activation. The consequence of this integrin/FAK/PI3K/Akt/mTORC1 + Piezo1/CaMKII/eEF2 dual cascade is that resistance exercise simultaneously activates muscle protein synthesis (mTORC1 → S6K1/4E-BP1) and suppresses muscle protein breakdown (Akt → FOXO3a sequestration → MuRF1/Atrogin-1 repressed) through mechanosensory signals that operate entirely independently of leucine, insulin, and IGF-1. This is why the mechanical MPS signal from resistance exercise remains active even in the calorie-restricted, low-insulin, low-IGF-1 environment created by GLP-1 therapy — it bypasses the very pathway that caloric restriction impairs.

In the context of caloric restriction:

Multiple RCTs demonstrate that resistance exercise during weight loss preserves lean mass more effectively than aerobic exercise alone
Even modest resistance training — 2–3 sessions per week — produces meaningful lean mass protection
Volume and progressive overload matter: muscles need to experience increasing challenge over time to maintain the MPS signal

Practical guidance: Progressive resistance exercise targeting major muscle groups (lower body, upper body, core) 2–3 times per week is the current evidence-based standard. If you’re new to resistance training, a physical therapist or certified trainer can establish a safe starting protocol.

The review by Arora et al. explicitly identifies resistance exercise as the most evidence-backed currently available intervention for lean mass preservation during GLP-1 therapy — ahead of any pharmacological option.

2. Protein Intake (High Priority, Strong Evidence)

Adequate protein intake during caloric restriction is essential for minimizing lean mass loss. Muscle protein synthesis requires amino acid availability; when dietary intake is insufficient, the body supplements from muscle protein breakdown.

Current guidance: Most sports nutrition and clinical nutrition organizations recommend 1.2–1.6 g of protein per kilogram of body weight per day during active weight loss, with some recommending up to 2.0 g/kg for older adults (who have reduced leucine sensitivity — the amino acid that most potently activates mTOR in muscle) and athletes.

This is often difficult to achieve on the appetite-suppressed diets typical of GLP-1 therapy. Intentional protein prioritization is required: eat protein first at every meal, and consider protein supplementation (whey protein, casein, pea protein) to hit targets when food volume is limited.

The leucine threshold and why protein quality matters: Not all dietary protein is equally effective at stimulating MPS. The key determinant is leucine content — leucine is the amino acid that most potently activates mTOR in muscle cells, acting through the Sestrin2-GATOR2 complex as a direct amino acid sensor. There is a “leucine threshold” concept: a minimum leucine dose per meal (~2.5–3g) is required to produce a maximal MPS response; below this threshold, MPS activation is submaximal regardless of total protein consumed.

This explains why protein source matters during GLP-1 therapy when appetite is suppressed and total food intake is limited:

Whey protein: High leucine content (~10% by amino acid composition), fast digestion kinetics — the gold standard for MPS activation per gram
Casein: Lower leucine than whey, slower digestion — better for sustained amino acid delivery between meals
Animal proteins generally: Complete amino acid profiles with adequate leucine per serving
Plant proteins: Lower leucine per gram than animal sources; require larger quantities or strategic combinations to reach the leucine threshold

Additionally, older adults experience anabolic resistance — reduced sensitivity of muscle to leucine signaling, requiring a higher leucine dose (and thus higher total protein intake) to achieve the same MPS response as younger adults. The molecular basis for this right-shifted leucine dose-response curve involves three converging deficits in aged skeletal muscle. First, Sestrin2 protein expression declines approximately 30–40% in aged rodent and human skeletal muscle — attributed to reduced PGC-1α-driven transcription as mitochondrial biogenesis slows with age. Because Sestrin2 is the primary leucine sensor that, when leucine-bound, dissociates from GATOR2 to relieve its inhibition of Ragulator/RagA-GTP and thus permit mTORC1 lysosomal recruitment, a lower Sestrin2 pool means that a higher ambient leucine concentration is required to displace the same proportion of Sestrin2 from GATOR2 — the mTORC1 activation curve via the RagA/Ragulator/lysosomal pathway is right-shifted. Second, a post-mTORC1 ribosomal biogenesis bottleneck limits translation capacity independently of leucine sensing: TIF-1A (transcription initiation factor 1A, the RNA Polymerase I co-activator) expression declines in aged myocytes, reducing the rate at which rRNA genes are transcribed and new 28S/18S ribosomal RNA is produced; even when upstream mTORC1 activity is pharmacologically equalized between young and old muscle, the aged muscle assembles fewer new ribosomes per unit time, limiting how rapidly mTORC1 activation can be converted into new structural protein. Third, somatopause (the age-related decline in GH pulse amplitude and circulating IGF-1) reduces basal Akt activity in aged myocytes: with lower tonic IGF-1R/IRS-1/PI3K/Akt signaling, there is less constitutive TSC2 Ser939/Thr1462 phosphorylation → less Rheb-GTP accumulation on the lysosomal surface → the mTORC1 complex is less pre-primed to respond when the leucine sensor does fire, attenuating the amplitude of the MPS response even for the fraction of Sestrin2 that is successfully displaced. Together these three deficits — attenuated leucine sensing, downstream ribosomal biogenesis capacity limitation, and reduced basal Akt/Rheb priming — explain why aged muscle requires not merely more protein but specifically more leucine per meal and why leucine-enriched formulas (not merely caloric protein supplements) are the rational intervention. This is the biological rationale for the 1.6–2.0 g/kg/day recommendation in older adults, versus 1.2–1.6 g/kg/day for younger individuals, and why protein supplementation with leucine-rich sources (whey, leucine-enriched formulas) is specifically emphasized for older adults, not just any protein source.

3. Creatine Supplementation (Moderate Evidence, Low Risk)

Creatine monohydrate at 3–5 g/day has one of the best safety and tolerability profiles of any dietary supplement. Its primary mechanism is phosphocreatine replenishment for ATP regeneration during high-intensity exercise — a process worth understanding at the molecular level because it explains both the performance benefit and the lean-mass relevance. Creatine enters myocytes via the SLC6A8 (CrT1) sodium-dependent creatine transporter and, once intracellular, is phosphorylated by creatine kinase (CK) using mitochondrial-derived ATP to form phosphocreatine (PCr). During high-intensity contraction bursts, myosin ATPase hydrolyzes ATP to ADP + Pi at rates that transiently exceed the capacity of mitochondrial oxidative phosphorylation (which requires seconds to upregulate electron transport chain flux). The accumulating ADP activates CK in reverse: PCr + ADP → creatine + ATP — CK immediately regenerates ATP from the phosphocreatine buffer, maintaining the ATP:ADP ratio at the myosin cross-bridge even while oxygen-dependent ATP resynthesis is still ramping up. This buffer capacity sustains near-maximal power output for approximately 8–12 seconds and accelerates ATP resynthesis between explosive efforts (mitochondrial CK at the outer mitochondrial membrane uses oxidative phosphorylation to recharge the PCr pool during inter-set rest periods). Creatine supplementation at 3–5 g/day increases total intramuscular creatine and phosphocreatine stores by approximately 20–30% above unsupplemented baseline — raising the ceiling of the ATP buffer before oxidative phosphorylation must take over, delaying fatigue at any given power output, and shortening the PCr recharge time between efforts. In the context of GLP-1 therapy and caloric restriction, this PCr buffer has a secondary muscle-preserving consequence: by maintaining ATP:ADP ratio at the myosin head for more repetitions per set, it prolongs the mechanical loading signal per training session → greater cumulative activation of the integrin/FAK/PI3K/mTORC1 mechanosensory cascade → more robust suppression of FOXO3a/MuRF1/Atrogin-1 per workout despite the catabolic hormonal environment. This exercise-sensitization pathway is, however, entirely dependent on active muscle contraction to generate the FAK Tyr397 autophosphorylation that initiates the integrin/FAK/PI3K/mTORC1 cascade — in sedentary muscle between training sessions, the PCr buffer is recharging rather than extending a mechanical loading signal, and the FAK mechanosensory cascade is quiescent regardless of intracellular creatine concentration. A second creatine mechanism operates through a mechanistically orthogonal, exercise-independent trigger — explaining why creatine supplementation shows lean mass benefits even in weight-loss populations with limited exercise capacity or compliance, as the meta-analysis literature consistently documents:

Meta-analyses of creatine supplementation during weight loss show modest preservation of lean mass compared to placebo, even without exercise
Creatine promotes cellular hydration in muscle (water retained intramyocellularly), which independently signals muscle protein synthesis through osmotic sensing. The molecular chain: creatine enters myocytes via SLC6A8 (CrT1), the Na⁺-dependent creatine transporter, which co-transports 2 Na⁺ ions per creatine molecule → osmotic gradient draws water inward through AQP4 aquaporin channels on the sarcolemma → myocyte volume increases. This cell swelling is detected by membrane-embedded integrin mechanoreceptors (α7β1 integrins sense changes in sarcolemmal tension) → integrin ligation activates FAK (focal adhesion kinase)/Src kinase → PI3K → PIP₃ → Akt → mTORC1 phosphorylates S6K1 and 4E-BP1 → ribosomal mRNA translation of muscle structural proteins. A parallel osmosensitive arm activates ERK1/2 through Raf-1 (osmotic swelling independently activates Ras/Raf/MEK/ERK). Together, the cell volume expansion from creatine-driven water retention activates two distinct mTORC1/ERK anabolic pathways without any change in nutritional or hormonal status — explaining why creatine’s volumizing effect is not simply cosmetic but mechanistically pro-anabolic in the context of caloric restriction where nutrient-driven mTOR signaling is attenuated
There is no evidence that creatine causes fat gain or negatively interacts with GLP-1 mechanisms

Creatine is widely available over the counter, inexpensive, and has a safety profile established over decades of use. It is one of the few supplements that justifies recommendation alongside the exercise and protein pillars for lean mass preservation.

4. HMB (Beta-Hydroxy Beta-Methylbutyrate) — Emerging Evidence

HMB is a metabolite of leucine (the primary mTOR-activating amino acid) that mediates much of leucine’s anti-catabolic effect at concentrations too low to fully activate mTOR directly. The molecular mechanism of HMB’s anti-catabolic action operates through the PI3K/Akt/FOXO axis: HMB activates PI3K independently of the insulin receptor/IRS-1 pathway — a distinction that matters in the insulin-resistant muscle environment common to obesity and to the caloric restriction imposed by GLP-1 therapy, where IRS-1 Ser307 phosphorylation (from the mTORC1 negative feedback loop) blunts normal insulin signaling. By activating PI3K upstream of IRS-1, HMB reaches the Akt node regardless of IRS-1 status. Akt phosphorylates FOXO1 at Thr24 and FOXO3a at Thr32 → cytoplasmic sequestration by 14-3-3 proteins → FOXO transcription factors cannot enter the nucleus → MuRF1 (muscle RING-finger protein 1) and Atrogin-1/MAFbx are not transcribed → ubiquitin-proteasome tagging of myosin heavy chain, troponin I, and other contractile proteins is reduced (PMID:19364184). In parallel, HMB activates mTORC1 through a Rheb-dependent mechanism, driving S6K1 and 4E-BP1 phosphorylation to modestly increase muscle protein synthesis rate. The net result — simultaneous suppression of protein degradation via FOXO/MuRF1/Atrogin-1 and stimulation of protein synthesis via mTORC1 — explains why HMB exerts both anti-catabolic and modestly pro-anabolic effects, and why its benefit is most pronounced in populations with anabolic resistance (older adults, catabolic illness) where leucine itself cannot generate sufficient Akt activation to suppress FOXO transcription factors.

Clinical trial results with HMB are mixed, but the better-designed studies in older adults and in caloric restriction contexts show meaningful lean mass preservation. The 2026 Arora et al. review includes HMB as a supplementation candidate worth discussing with a prescriber. Standard dosing is 3 g/day in divided doses (1 g with each meal). HMB is available over-the-counter and has a good safety profile.

HMB is likely more beneficial in populations with reduced leucine sensitivity (older adults, those with metabolic disease) than in younger healthy individuals, where protein optimization and resistance exercise produce stronger effects.

5. Track Lean Mass, Not Just Weight (Important but Underused)

Standard scale weight does not distinguish between fat and muscle. A patient could be losing primarily lean mass while maintaining body fat and appear to be “succeeding” by weight metrics alone.

Body composition measurement options:

Method	Accuracy	Cost	Availability
DEXA (dual-energy X-ray absorptiometry)	Gold standard — separate fat, lean, bone	$50–150 per scan	Medical centers, some gyms, radiology clinics
InBody bioimpedance	Good for tracking trends, less absolute accuracy	$10–30 per scan (clinic); home scales cheaper	Many weight loss clinics, performance gyms
Hydrostatic weighing	Near-gold standard	$50–80	Sports science labs
Skinfold calipers	Operator-dependent; useful for trends	Minimal	Primary care, fitness professionals

Recommended: Baseline DEXA before or early in GLP-1 therapy, then repeat at 6 months and 12 months. This allows your prescriber to see whether you’re losing fat vs. lean mass and adjust the intervention accordingly. If lean mass loss is >30–35% of total weight loss, intensifying the resistance exercise and protein protocol — or discussing pharmacological options — is appropriate.

The Pharmacological Pipeline: What’s Being Investigated

For patients with significant concerns about lean mass preservation — particularly older adults, patients with pre-existing sarcopenia, or those anticipating very large weight losses — the 2026 review surveys several pharmacological candidates being studied as adjuncts to GLP-1 therapy.

Tesamorelin (Most Clinically Advanced Peptide Option)

Tesamorelin is a synthetic analogue of GHRH (growth hormone-releasing hormone) — the hypothalamic signal that prompts the pituitary gland to produce growth hormone. Unlike direct GH injection, tesamorelin works through the physiological GH axis, meaning GH release remains subject to normal feedback regulation.

Why it’s relevant here: Tesamorelin is FDA-approved for the treatment of HIV-associated lipodystrophy — a condition where antiretroviral drugs cause excess visceral fat accumulation. The pivotal LIPO-010 trials demonstrated that tesamorelin reduced visceral fat by approximately 15% compared to placebo while preserving or improving lean body mass. This lean-mass-sparing property is precisely what’s needed in the GLP-1 context.

The rationale for combination: GH axis stimulation promotes fat mobilization (lipolysis) and muscle protein preservation through IGF-1 signaling — and the mechanism by which these two arms shift weight loss composition explains why the combination is physiologically rational. Lipolysis arm: GH binds GHR on white adipocytes → JAK2 trans-autophosphorylation → STAT5b Tyr699 phosphorylation → STAT5b drives transcription of ATGL (adipose triglyceride lipase) and HSL (hormone-sensitive lipase), the two rate-limiting triglyceride hydrolases at the lipid droplet surface. ATGL cleaves triglyceride → diacylglycerol; HSL cleaves diacylglycerol → monoacylglycerol; monoacylglycerol lipase completes hydrolysis → glycerol + free fatty acids (FFAs) are released into systemic circulation. Elevated plasma FFAs are taken up by hepatocytes and skeletal muscle for β-oxidation → acetyl-CoA → TCA/OXPHOS → ATP. When FFAs supply sufficient caloric substrate, the liver’s gluconeogenic demand for amino acids (released by muscle via the alanine cycle: muscle transamination of pyruvate → alanine → hepatic deamination → glucose + urea) is reduced — this is GH’s protein-sparing effect: caloric deficit is met by fat mobilization rather than by catabolizing muscle protein for gluconeogenesis. IGF-1 anabolic arm: GH → hepatic GHR/JAK2/STAT5b → IGF-1 secretion → circulating IGF-1 binds IGF-1R on myocytes → IRS-1/PI3K/PIP₃/PDK1/Akt Thr308 + mTORC2/Akt Ser473 → fully active Akt phosphorylates FOXO3a at Thr32/Ser253 → 14-3-3ζ sequesters FOXO3a in the cytoplasm → MuRF1 and Atrogin-1/MAFbx gene transcription is suppressed → ubiquitin-proteasome targeting of myosin heavy chain is attenuated; mTORC1 simultaneously phosphorylates S6K1/4E-BP1 → cap-dependent mRNA translation → muscle protein synthesis maintained at a higher rate than the caloric deficit alone would permit. The convergence of FFA-driven protein sparing and IGF-1-driven FOXO3a sequestration is mechanistically why tesamorelin-treated HIV-lipodystrophy patients lost visceral fat while simultaneously preserving lean mass — and why the same logic applies to the GLP-1 context. During GLP-1-induced caloric restriction, adding GHRH-axis support could theoretically shift the composition of weight loss toward fat and away from muscle through both arms simultaneously. The metabolic interaction requires monitoring: GH increases hepatic insulin resistance through a specific molecular mechanism — GH binds GHR on hepatocytes → JAK2 activation → STAT5b drives transcription of SOCS1 (suppressor of cytokine signaling 1) and SOCS3 as negative-feedback genes → SOCS1/SOCS3 proteins bind JAK2 and IRS-1/IRS-2 (the insulin receptor substrate proteins through which the insulin receptor normally activates PI3K/Akt/FoxO1 in hepatocytes) → SOCS1/SOCS3 recruit the Elongin BC/Cullin5/SOCS-box (ECS) E3 ubiquitin ligase via their SOCS-box domain → IRS-1 and IRS-2 are polyubiquitinated and targeted for proteasomal degradation → reduced IRS-1/IRS-2 protein abundance in hepatocytes → attenuated insulin receptor/IRS-1/PI3K/Akt signaling when portal insulin arrives postprandially → hepatic gluconeogenesis (via FOXO1/G6Pase/PEPCK) is less effectively suppressed → GLP-1’s glucose-lowering effect is partially counteracted at the hepatic level; HbA1c should be checked at 6-week intervals when combining the two.

Current status: No large RCTs have published results for tesamorelin combined with GLP-1 agonists in non-HIV patients. This is an investigational use. Investigator-initiated studies are in early stages as of 2026.

GH Secretagogues (Ipamorelin/CJC-1295, Sermorelin)

Ipamorelin, CJC-1295, and sermorelin stimulate GH production through GHRP and GHRH mechanisms. The lean-mass-preserving rationale parallels tesamorelin: GH axis support during caloric restriction shifts body composition toward fat loss.

The receptor pharmacology differs between the two compound classes and is worth understanding because it explains why they are often combined. Ipamorelin is a growth hormone-releasing peptide (GHRP) — it binds the GHSR-1a (growth hormone secretagogue receptor 1a), a Gq-coupled GPCR expressed on pituitary somatotrophs. GHSR-1a/Gq activation → PLCβ (phospholipase C-β) → hydrolysis of PIP₂ → IP₃ + DAG (diacylglycerol). IP₃ triggers Ca²⁺ release from the endoplasmic reticulum into the cytoplasm; DAG activates PKC (protein kinase C). The combined Ca²⁺/PKC signal drives rapid GH vesicle exocytosis from secretory granules — a fast, spike-shaped GH pulse. Ipamorelin’s selectivity (binding GHSR-1a without activating GHS-R subtypes that release cortisol or prolactin) is its key pharmacological advantage over older GHRPs like GHRP-2 and GHRP-6, which also activate non-GH secretagogue receptors.

CJC-1295 (and sermorelin) are GHRH analogues — they bind the GHRH receptor (GHRH-R), a Gs-coupled GPCR on the same somatotrophs. GHRH-R/Gs → adenylyl cyclase → cAMP elevation → PKA phosphorylation of: (1) voltage-gated L-type Ca²⁺ channels → Ca²⁺ entry → GH vesicle fusion with the plasma membrane; (2) CREB → GH1 gene transcription → increased GH protein synthesis to replenish secretory granules; (3) GHRH-R upregulation itself → sensitization to subsequent GHRH pulses. The Gs/cAMP arm thus operates on a longer timescale than the Gq/Ca²⁺ arm, building GH synthetic capacity over hours rather than just triggering immediate release.

Because the two receptor systems are entirely distinct (Gq vs. Gs, PLCβ vs. adenylyl cyclase), combining a GHRP (ipamorelin) with a GHRH analogue (CJC-1295 or sermorelin) produces supra-additive GH release — the Gq-driven Ca²⁺ spike from ipamorelin is amplified by the pre-existing cAMP/PKA sensitization from CJC-1295/sermorelin, producing larger and more sustained GH pulses than either compound alone at equivalent doses. The supraadditivity is not simply additive receptor coverage — it reflects three distinct molecular sensitization events that GHRH-R/cAMP/PKA priming applies to the exact same Ca²⁺ signaling machinery that ipamorelin/GHSR-1a subsequently activates. First, Cav1.2 L-type Ca²⁺ channel sensitization: PKA phosphorylates the Cav1.2 α1C subunit at Ser1928 on the C-terminal regulatory domain → increased channel open probability and unitary conductance. When ipamorelin then triggers GHSR-1a/Gq/PLCβ/IP₃/DAG, the IP₃-mediated ER Ca²⁺ release is amplified by coincident Cav1.2 Ca²⁺ entry — which PKA has maximized — producing a larger total intracellular Ca²⁺ transient than ipamorelin alone can generate without prior PKA priming. Second, IP₃R type 2 sensitization: PKA phosphorylates the IP₃ receptor type 2 (IP₃R2) at Ser1589, increasing the receptor’s open probability at any given IP₃ concentration. Ipamorelin-triggered PLCβ produces the same IP₃ quantity regardless of cAMP state, but that IP₃ drives larger Ca²⁺ release from the somatotroph ER when IP₃R2 is pre-phosphorylated by PKA — a classic coincidence detector where the magnitude of the Ca²⁺ response is determined by both ligand concentration and receptor phosphorylation state. Third, CREB-driven granule loading: CJC-1295 → cAMP/PKA/CREB → GH1 gene transcription → increased GH protein synthesis and granule-filling over hours to days. When ipamorelin triggers exocytotic fusion of secretory granules, the granule pool available per fusion event is larger — more GH protein per vesicle, more vesicles loaded and docked at the plasma membrane — amplifying the quantity released per Ca²⁺ transient. The consequence of these three simultaneous PKA-driven priming events is that ipamorelin’s Gq/Ca²⁺ exocytotic signal fires into a somatotroph that has more Ca²⁺ available per pulse (Cav1.2), releases more Ca²⁺ from the ER per IP₃ molecule (IP₃R2), and has more GH protein ready to release per granule fusion (CREB/GH1) than it would in the absence of GHRH-R cAMP signaling — explaining why the combination is consistently described as supra-additive rather than simply additive. This is the mechanistic basis for the clinical convention of combining these compound classes. The GH released then acts on hepatocytes (GH-R/JAK2/STAT5b → IGF-1 synthesis) and directly on skeletal muscle and adipocytes to promote fat mobilization and protein anabolism during caloric restriction.

These are investigational uses. Access is through compounding pharmacies with a valid prescription. Current FDA compounding guidance affects the legal landscape for some of these peptides, and patients should confirm they are working with licensed prescribers and PCAB-accredited pharmacies.

Bimagrumab (Most Promising Investigational Antibody)

Bimagrumab is a monoclonal antibody that blocks activin type II receptors — the signaling pathway (via myostatin/activin A) that normally limits muscle growth and promotes muscle breakdown. By blocking this brake on muscle growth, bimagrumab allows muscle protein synthesis to proceed without the inhibitory signal.

The intracellular cascade bimagrumab interrupts is the myostatin/SMAD pathway. Myostatin (GDF-8) and activin A bind activin type IIB receptor (ActRIIB) on the sarcolemma, which recruits and transphosphorylates the type I co-receptor ALK4 or ALK5. The activated ALK4/5 kinase then phosphorylates SMAD2 and SMAD3 at the conserved SSxS motif in their C-terminal MH2 domain. Phospho-SMAD2/3 dissociate from the receptor complex, bind the co-SMAD SMAD4, and the resulting heterotrimer translocates to the nucleus where it binds SMAD-binding elements in the promoters of MuRF1 (muscle RING finger protein-1) and Atrogin-1/MAFbx — the two E3 ubiquitin ligases that tag myofibrillar proteins (myosin heavy chain, titin, and troponin I) for proteasomal degradation. Separately, SMAD2/3 nuclear activity also suppresses Akt signaling, blunting mTORC1-driven protein synthesis. Bimagrumab occupies ActRIIB with high affinity (Kd ~1 nM), preventing myostatin and activin A from initiating this cascade; MuRF1 and Atrogin-1 are not upregulated via this pathway, the Akt suppression signal is absent, and the net balance shifts from catabolism toward anabolic protein retention (Lach-Trifilieff et al., PMID:24623528; Han et al., PMID:23530121).

Phase 2 results: A Phase 2 trial published in JAMA Internal Medicine (Heymsfield et al., PMID:34694348) in patients with type 2 diabetes and obesity showed significant increases in lean mass alongside decreases in fat mass. These results are unusual: most weight loss interventions lose some lean mass; bimagrumab produced a shift toward gaining lean mass simultaneously. The simultaneous fat loss requires a distinct mechanistic explanation from the muscle-gain arm, because ActRIIB/SMAD2/3 signaling has roles in adipose tissue biology that are separate from its muscle atrophy functions.

Why bimagrumab also reduces fat mass. Activin A (a TGF-β superfamily ligand that signals through the same ActRIIB receptor that binds myostatin) is secreted by white adipocytes and adipose stromal cells in proportion to adipose tissue mass — and it acts in an autocrine/paracrine loop that promotes white adipocyte lipid storage and suppresses thermogenic beiging programs. The mechanism: activin A → ActRIIB → ALK4/5 → SMAD2/3 phosphorylation in adipocyte precursors → nuclear SMAD2/3/SMAD4 complex → direct transcriptional repression of PGC-1α (PPARGC1A) and UCP1 (uncoupling protein 1, the mitochondrial proton leak channel responsible for non-shivering thermogenesis in brown and beige adipocytes) → adipocyte precursors that would otherwise adopt a beige/thermogenic identity under β-adrenergic or PPAR-γ agonist stimulation are instead diverted toward white lipid-storing phenotype; simultaneously, SMAD3 promotes FASN (fatty acid synthase) and DGAT1 (diacylglycerol acyltransferase 1) transcription in mature white adipocytes, increasing de novo lipogenesis rate and triglyceride esterification capacity. When bimagrumab blocks ActRIIB, this activin A/SMAD3 adipose program is dismantled: SMAD2/3 nuclear translocation is blocked → PGC-1α and UCP1 transcriptional repression is relieved → adipocyte precursors in the beiging-permissive microenvironment adopt thermogenic identity → beige adipocytes with functional UCP1 insert into the inner mitochondrial membrane → UCP1 dissipates the proton gradient generated by the electron transport chain as heat rather than capturing it as ATP → caloric energy is wasted as thermogenesis without any change in food intake or exercise output. The net metabolic consequence in the whole organism is an increase in resting energy expenditure that shifts the energy balance toward fat oxidation, explaining why bimagrumab-treated subjects lose fat mass in the trial even though the drug produces no GLP-1-mediated appetite suppression and no behavioral caloric restriction. The muscle-gain and fat-loss arms are therefore mechanistically orthogonal: blocking ActRIIB/SMAD3 in skeletal muscle relieves MuRF1/Atrogin-1-mediated myofibrillar catabolism → lean mass accrues; blocking the same receptor in adipose tissue relieves SMAD3-mediated UCP1 suppression → beige thermogenic adipogenesis → fat oxidized as heat. Both effects trace to the same ActRIIB blockade but operate through tissue-specific SMAD3 target gene sets.

Current status: Phase 3 trials are underway. Bimagrumab is not yet approved for any indication and is not accessible outside of clinical trial settings.

Enobosarm (SARM — Further from Approval)

Enobosarm (ostarine) is a selective androgen receptor modulator (SARM) studied for muscle wasting conditions. It is investigational, not approved, and carries a different risk profile from the GH-axis approaches due to androgen receptor activity. The molecular mechanism: The androgen receptor (AR) is a ligand-dependent nuclear transcription factor in the nuclear receptor superfamily. Testosterone or DHT binding to the AR ligand-binding domain (LBD) induces a conformational change in helix-12 (the AF-2 activation helix at the C-terminus of the LBD) → helix-12 repositions over the hydrophobic pocket, creating a coactivator binding surface called the activation function-2 (AF-2) groove → coactivator proteins containing LXXLL nuclear receptor-binding motifs (SRC/p160 family coactivators, CBP/p300) are recruited → AR homodimerizes, translocates to the nucleus, and binds androgen response elements (AREs) in target gene promoters → transcription of AR target genes including IGF1, follistatin, and genes encoding thick-filament structural proteins (myosin heavy chains, actin) → muscle protein synthesis. SARMs like enobosarm bind the same AR LBD but induce a subtly distinct helix-12 geometry compared to natural androgens — one that preferentially recruits muscle-enriched coactivators (SRC-2/GRIP1) while less efficiently recruiting the coactivators required for full androgenic activity in prostate, seminal vesicle, and hepatic tissue. This tissue-selective coactivator recruitment is the mechanistic basis for the “selective” in SARM. Enobosarm additionally suppresses the muscle atrophy program through AR-driven IGF1 transcription → IGF-1/IGF-1R/PI3K/Akt → FOXO1/3 phosphorylation/cytoplasmic sequestration → reduced nuclear FOXO → attenuated MuRF1 and Atrogin-1 transcription → less ubiquitin-proteasome targeting of myosin heavy chain. The risk profile concern is that tissue selectivity is incomplete in current SARMs: partial androgen AR activation in non-target tissues (hepatic lipid metabolism changes via hepatic AR; SHBG displacement increasing free estradiol) and potential testosterone suppression via hypothalamic-pituitary feedback remain unsolved challenges, distinguishing SARMs from the GH-axis and anti-myostatin approaches above. The HPG suppression mechanism is specific: enobosarm → AR activation on arcuate nucleus KNDy (kisspeptin/neurokinin B/dynorphin) interneurons → AR/ARE-driven transcription of the prodynorphin gene (PDYN) → increased dynorphin release from KNDy axon terminals at GnRH pulse generator sites → dynorphin binds κ-opioid receptors (KOR/OPRK1) on GnRH-secreting neurons → KOR/Gαi → adenylyl cyclase inhibition → reduced cAMP → attenuated GnRH pulse amplitude and frequency entering the hypothalamo-pituitary portal vessels → reduced LH secretion from gonadotrophs → reduced LHCGR/Gαs/cAMP/PKA/StAR Ser194 phosphorylation in testicular Leydig cells → less cholesterol shuttled to the inner mitochondrial membrane → CYP11A1/CYP17A1/HSD17B3 throughput falls → reduced testosterone biosynthesis. Simultaneously, enobosarm → gonadotroph AR/ARE → follistatin (FST) transcription → follistatin sequesters circulating activin A (the primary FSH-stimulating signal released from Sertoli cells) → reduced FSH → Sertoli cell spermatogenic support declines, compounding HPG suppression. The clinical consequence: circulating testosterone falls despite enobosarm’s own AR agonism in muscle — removing the endogenous anabolic androgen signal that normally synergizes with AR occupation, and potentially attenuating net myofibrillar protein synthesis below what the SARM-occupied AR alone can sustain. The 2026 review includes enobosarm as part of the pharmacological landscape, but it is further from clinical availability than tesamorelin or bimagrumab.

Special Populations: Where the Risk Is Highest

Older Adults (65+)

Sarcopenia — age-related muscle loss — affects an estimated 10–30% of older adults and is already associated with falls, fractures, hospitalization, and loss of independence. Adding the lean mass loss from GLP-1 therapy to an already-reduced muscle reserve substantially increases functional risk.

For older adults starting GLP-1 medications, the following warrant particular attention:

Baseline DEXA and gait speed or grip strength assessment before starting
Protein targets at the higher end of recommendations (1.6–2.0 g/kg/day)
Resistance exercise prescription from a physical therapist if functional limitations exist
Discussion with prescriber about whether a slower dose titration rate is appropriate

Postmenopausal Women

As detailed above, estrogen loss reduces MPS capacity, making postmenopausal women particularly vulnerable to lean mass loss during caloric restriction. DEXA monitoring, high protein targets, and consistent resistance exercise are especially important in this group.

Patients with Pre-Existing Sarcopenic Obesity

Sarcopenic obesity — having both excess fat and reduced muscle mass — is more common than recognized, particularly in the population most likely to be prescribed GLP-1 medications. These patients may show normal or only mildly elevated BMI while having already lost clinically significant muscle. Standard weight-based monitoring misses this; body composition assessment before and during GLP-1 therapy is essential.

A Practical Framework for GLP-1 Patients

Immediately actionable:

Start or maintain resistance exercise, 2–3 sessions per week targeting major muscle groups
Prioritize protein at every meal; aim for 1.2–1.6 g/kg body weight (higher for older adults)
Add creatine monohydrate 3–5 g/day — safe, inexpensive, and evidence-supported
Request a baseline DEXA scan if you haven’t had one; repeat at 6 months and 12 months

Discuss with your prescriber: 5. Ask specifically about body composition monitoring, not just weight monitoring 6. If significant lean mass loss is documented (>30% of total weight loss from lean mass on DEXA), ask whether dose titration adjustment, specialist referral, or pharmacological adjuncts are appropriate

Watch the pipeline: 7. Bimagrumab Phase 3 results, when published, will be the signal that changes pharmacological options for this indication 8. Tesamorelin combination trial results — if positive — would provide an evidence base for prescribing it alongside GLP-1 therapy

The Bigger Picture

The muscle loss question during GLP-1 therapy illustrates a broader challenge in pharmacology: a drug can be enormously effective for its primary endpoint while creating secondary problems that require their own management. The answer is not to stop taking medications that provide meaningful metabolic benefit — it is to understand the full picture and manage the secondary effects intelligently.

For the vast majority of GLP-1 patients, resistance exercise, adequate protein, and creatine are sufficient to produce a favorable body composition trajectory. Pharmacological adjuncts are investigational and not currently standard of care. But the research pipeline is active, the physiological rationale for GH axis support during weight loss is sound, and the evidence base will evolve over the next several years.

Understanding where the science is now — and what questions are actively being answered — puts patients in a better position to have informed conversations with their prescribers about whether additional monitoring or interventions are appropriate for their specific situation.

References

Arora G, Conde KR, Desouza CV. Pharmacologic Treatments for the Preservation of Lean Body Mass During Weight Loss. J Clin Med. 2026;15(2):541. PMID: 41598480
Wilding JPH, Batterham RL, Calanna S, et al. Once-Weekly Semaglutide in Adults with Overweight or Obesity. N Engl J Med. 2021;384(11):989-1002. PMID: 33567185
Jastreboff AM, Aronne LJ, Ahmad NN, et al. Tirzepatide Once Weekly for the Treatment of Obesity. N Engl J Med. 2022;387(3):205-216. PMID: 35658024
Falutz J, Allas S, Blot K, et al. Metabolic effects of a growth hormone-releasing factor in patients with HIV. N Engl J Med. 2007;357(23):2359-70. PMID: 18057338
Heymsfield SB, Coleman LA, Miller R, et al. Effect of Bimagrumab vs Placebo on Body Fat Mass Among Adults With Type 2 Diabetes and Obesity. JAMA Intern Med. 2021;181(12):1545-1554. PMID: 34694348
Stokes T, Hector AJ, Morton RW, McGlory C, Phillips SM. Recent Perspectives Regarding the Role of Dietary Protein for the Promotion of Muscle Hypertrophy with Resistance Exercise Training. Nutrients. 2018;10(2):180. PMID: 29414942
Lanhers C, Pereira B, Naughton G, et al. Creatine Supplementation and Lower Limb Strength Performance: A Systematic Review and Meta-Analyses. Sports Med. 2015;45(9):1285-1294. PMID: 25946994
Cianflone D, Marenzi G, Calabrese F, et al. Prevention of muscle atrophy through the use of β-hydroxy-β-methylbutyrate (HMB) supplementation. Aging Clin Exp Res. 2022;34(3):489-498. PMID: 34637100

Muscle Loss on GLP-1 Medications: Why It Happens and What to Do About It

First: Why Weight Loss Always Involves Some Muscle Loss

The Molecular Atrophy Pathway: Why Muscle Protein Breaks Down

Why GLP-1 Medications May Be Particularly Challenging for Lean Mass

Why the Muscle Loss Question Has Become Urgent

Sex Differences in GLP-1 Lean Mass Loss

What the Clinical Trial Data Shows

Rate of Loss Matters: Slower Can Preserve More Muscle

What You Can Do Right Now: The Evidence

1. Resistance Exercise (High Priority, Strong Evidence)

2. Protein Intake (High Priority, Strong Evidence)

3. Creatine Supplementation (Moderate Evidence, Low Risk)

4. HMB (Beta-Hydroxy Beta-Methylbutyrate) — Emerging Evidence

5. Track Lean Mass, Not Just Weight (Important but Underused)

The Pharmacological Pipeline: What’s Being Investigated

Tesamorelin (Most Clinically Advanced Peptide Option)

GH Secretagogues (Ipamorelin/CJC-1295, Sermorelin)

Bimagrumab (Most Promising Investigational Antibody)

Enobosarm (SARM — Further from Approval)

Special Populations: Where the Risk Is Highest

Older Adults (65+)

Postmenopausal Women

Patients with Pre-Existing Sarcopenic Obesity

A Practical Framework for GLP-1 Patients

The Bigger Picture

References

Read Next

Anti-Aging Peptide Protocols: A Framework for Evidence-Based Longevity Approaches

Anti-Aging Peptides: What the Research Actually Shows

AOD-9604 and Fragment 176-191: What the Clinical Trial Record Actually Shows

Ready to Apply These Insights?

First: Why Weight Loss Always Involves Some Muscle Loss

The Molecular Atrophy Pathway: Why Muscle Protein Breaks Down

Why GLP-1 Medications May Be Particularly Challenging for Lean Mass

Why the Muscle Loss Question Has Become Urgent

Sex Differences in GLP-1 Lean Mass Loss

What the Clinical Trial Data Shows

Rate of Loss Matters: Slower Can Preserve More Muscle

What You Can Do Right Now: The Evidence

1. Resistance Exercise (High Priority, Strong Evidence)

2. Protein Intake (High Priority, Strong Evidence)

3. Creatine Supplementation (Moderate Evidence, Low Risk)

4. HMB (Beta-Hydroxy Beta-Methylbutyrate) — Emerging Evidence

5. Track Lean Mass, Not Just Weight (Important but Underused)

The Pharmacological Pipeline: What’s Being Investigated

Tesamorelin (Most Clinically Advanced Peptide Option)

GH Secretagogues (Ipamorelin/CJC-1295, Sermorelin)

Bimagrumab (Most Promising Investigational Antibody)

Enobosarm (SARM — Further from Approval)

Special Populations: Where the Risk Is Highest

Older Adults (65+)

Postmenopausal Women

Patients with Pre-Existing Sarcopenic Obesity

A Practical Framework for GLP-1 Patients

The Bigger Picture

References

Stay Informed on Peptide Research

Read Next

Anti-Aging Peptide Protocols: A Framework for Evidence-Based Longevity Approaches

Anti-Aging Peptides: What the Research Actually Shows

AOD-9604 and Fragment 176-191: What the Clinical Trial Record Actually Shows

Ready to Apply These Insights?