44 findings across 16 drug records. Domain pages are descriptive indexes, not advice.
Most represented records
GLP-1 RA class (10)
Semaglutide (6)
Retatrutide (5)
Tirzepatide|semaglutide (head-to-head) (3)
Multiple (review) (3)
Evidence spread
High evidence 2Moderate evidence 21Low evidence 12Very low evidence 9
ApitegromabMARKETED
Body compositionMixed
Very low evidence
Scholar Rock phase-2 EMBRAZE trial (NCT06445075): apitegromab added to tirzepatide preserved additional lean mass versus tirzepatide alone over 24 weeks. The only...
SourceScholar Rock, EMBRAZE phase-2 topline press release, 18 Jun 2025; NCT06445075Source
Full findingScholar Rock phase-2 EMBRAZE trial (NCT06445075): apitegromab added to tirzepatide preserved additional lean mass versus tirzepatide alone over 24 weeks. The only major muscle-preservation combo studied on TIRZEPATIDE (a GIP/GLP-1 dual agonist) rather than semaglutide. NON-GRADUATING: company press topline only.
Fundingindustry - Eli Lilly (trial sponsor; inferred from registration trial)
Scope limitssmall sample (N~102)
Comparatorsplacebo + tirzepatide
BimagrumabMARKETED
Body compositionMixed
Moderate evidence
Foundational human readout for muscle preservation: in adults with T2D and overweight/obesity, bimagrumab MONOTHERAPY (no GLP-1RA) produced large fat-mass loss with a...
SourceHeymsfield SB et al., JAMA Netw Open, 2021Source
Full findingFoundational human readout for muscle preservation: in adults with T2D and overweight/obesity, bimagrumab MONOTHERAPY (no GLP-1RA) produced large fat-mass loss with a simultaneous GAIN in lean mass over 48 weeks, establishing that ActRII blockade is anabolic for muscle while reducing adiposity. This is lean-mass GAIN, distinct from the lean-PRESERVATION seen when such agents are added to incretins.
PopulationPhase-2 double-masked placebo-controlled RCT, N=75 randomised (58 completers), 48 weeks, IV q4w, adults with T2D, BMI 28-40; 9 US/UK sites.
BELIEVE phase-2b (Eli Lilly/Versanis) tested bimagrumab alone or added to semaglutide. The combination drove fat-selective weight loss with lean-mass preservation;...
SourceHeymsfield et al., Nat Med 2026; presented ADA 2025Source
Full findingBELIEVE phase-2b (Eli Lilly/Versanis) tested bimagrumab alone or added to semaglutide. The combination drove fat-selective weight loss with lean-mass preservation; bimagrumab monotherapy GAINED lean while semaglutide alone lost it. Now PEER-REVIEWED (Heymsfield et al., Nat Med 2026).
PopulationBELIEVE phase-2b RCT, N=507, 48-week treatment (results to wk72), randomised double-blind placebo-controlled, 9 arms; bimagrumab IV q12w +/- semaglutide sc weekly; adults overweight/obese with >=1 comorbidity.
Fundingindustry - Eli Lilly (trial sponsor)
Scope limitsDEEPENED (D2): upgraded from landscape note to the peer-reviewed BELIEVE result. VALIDATOR: PMID 41772149 RESOLVED (Heymsfield et al., Nat Med 2026;32(3):869-882, NCT05616013) - status verified, no fallback needed. The per-arm lean DXA percentages (mono +2.5%, sema -7.4%, 92.8% fat fraction, fat -45.7%) are SECONDARY results not in the Nat Med abstract; design-consistent but flagged as not-abstract-verified. Distinguishes lean GAIN (bimagrumab mono) from lean PRESERVATION (combo). Surrogate DXA/BIA, phase-2b, 9 arms dilute per-arm N -> moderate; Lilly-sponsored.
Muscle-preserving combination landscape: bimagrumab (ActRII blockade) is reported to add lean mass while reducing fat, and is being combined with semaglutide to...
SourceNunn E et al., Mol Metab, 2024 (preclinical); Kaiser M et al., Cardiol Rev, 2025 (review)Source
Full findingMuscle-preserving combination landscape: bimagrumab (ActRII blockade) is reported to add lean mass while reducing fat, and is being combined with semaglutide to counter GLP-1-associated lean-mass loss; preclinical work shows combination preserves muscle while enhancing fat loss vs semaglutide alone.
PopulationPreclinical (diet-induced obese mice) plus clinical-development narrative/review; combination under clinical study
Fundingacademic/independent analysis (manufacturer not study sponsor; per-study COI not individually audited)
Scope limitsanimal data; human relevance uncertain
Comparatorssemaglutide
CagriSemaINVESTIGATIONAL
Body compositionDecrease
Low evidence
REDEFINE-1 DXA subgroup (Week 68): CagriSema produced the greatest weight reduction; of weight lost, 66.9% was fat mass and 33.1% lean soft tissue. In participants...
SourceRavussin E et al., REDEFINE-1 body composition, ObesityWeek 2025 / ECO 2026 presentations...Source
Full findingREDEFINE-1 DXA subgroup (Week 68): CagriSema produced the greatest weight reduction; of weight lost, 66.9% was fat mass and 33.1% lean soft tissue. In participants achieving >=30% weight loss, fat-mass proportion fell from 46.3% to 33.2% while lean-soft-tissue proportion rose from 51.3% to 63.2%.
PopulationDXA subgroup of REDEFINE-1 RCT in overweight/obesity; CagriSema 2.4/2.4 mg vs semaglutide 2.4 mg vs cagrilintide 2.4 mg vs placebo; 68 weeks
Fundingacademic/independent analysis (manufacturer not study sponsor; per-study COI not individually audited)
Scope limitsconference/abstract-level
Comparatorssemaglutide; cagrilintide; placebo
Tirzepatide|semaglutide (head-to-head)MARKETED
Body compositionDecrease
Moderate evidence
Systematic review (Annals of Internal Medicine): across 35 RCTs, incretin therapies consistently reduced fat mass and visceral adiposity; muscle-based-index loss was...
SourceBatsis JA et al., Ann Intern Med, 2026Source
Full findingSystematic review (Annals of Internal Medicine): across 35 RCTs, incretin therapies consistently reduced fat mass and visceral adiposity; muscle-based-index loss was highly heterogeneous, with median 28.3% of total weight loss attributable to muscle-based indices and ~two-thirds of interventions exceeding pre-specified muscle-loss benchmarks. No study reported objective physical-function outcomes.
Population35 RCTs (median duration 26 weeks, median N 78) of liraglutide, semaglutide, tirzepatide or dulaglutide in adults with obesity; BIA/DXA/CT/MRI; Jan 2003-Feb 2026
Scope limitsClass-level synthesis; heterogeneity precluded meta-analysis. Note the FFM/muscle benchmark is measurement-method-dependent (~25% for BIA/DXA, ~15% for CT/MRI). Anchors the comparison matrix. The 'no objective physical function outcomes' gap is a recurring caveat.
Comparatorsplacebo; lifestyle
GLP-1 RA classMARKETED
Body compositionMixed
Low evidence
Narrative review of human studies on GLP-1 therapies and sarcopenia: results are mixed - some studies emphasise excessive skeletal-muscle-mass loss while others argue...
Full findingNarrative review of human studies on GLP-1 therapies and sarcopenia: results are mixed - some studies emphasise excessive skeletal-muscle-mass loss while others argue for a protective effect via improved muscle quality (less myosteatosis); no definitive conclusion on net detriment or benefit to skeletal muscle.
PopulationHuman studies of GLP-1-based therapies assessing skeletal-muscle mass, strength/function and structure/quality
Fundingacademic/independent analysis (manufacturer not study sponsor; per-study COI not individually audited)
Scope limitsExplicitly preserves BOTH poles of the sarcopenia question (harm vs muscle-quality benefit). Good neutrality anchor for the muscle-quality column. Echoes the SCWD/FDA workshop (PMID 41362110, DOI 10.1002/jcsm.70147) on endpoint challenges.
Tirzepatide|semaglutide (head-to-head)MARKETED
Body compositionDecrease
Moderate evidence
Systematic review and meta-analysis (20 RCTs, 15,782 adults, DXA/MRI) with absolute lean-mass change (kg) and lean fraction of weight lost as CO-PRIMARY outcomes. Lean...
SourceEisa N, Barood O et al., Diabetes Obes Metab, 2026Source
Full findingSystematic review and meta-analysis (20 RCTs, 15,782 adults, DXA/MRI) with absolute lean-mass change (kg) and lean fraction of weight lost as CO-PRIMARY outcomes. Lean mass was ~25-39% of total weight lost across incretin agents and broadly comparable to lifestyle intervention (no significant difference, p=0.42). The only modality that materially lowered the lean fraction was lifestyle PLUS resistance training (17.5%). Retatrutide is NOT included.
Population20 RCTs, 15,782 adults with overweight/obesity; semaglutide, tirzepatide, liraglutide or lifestyle; DXA or MRI; searched to 20 Jan 2026; random-effects, RoB 2, GRADE.
Fundingacademic / investigator-initiated; no industry sponsor found in indexed metadata (Europe PMC fundingList empty)
Scope limitsNEW source. Most direct head-to-class lean-fraction comparison and the closest thing to an absolute-lean-kg anchor. Decision-relevant: incretin lean fraction is no worse than diet-induced loss of similar magnitude (p=0.42); resistance training is the lever, not the drug. CRUCIAL GAP: retatrutide is ABSENT, so the class benchmark exists but no reta point sits in it. Meta-analysis graded moderate by default. DXA/MRI lean is not muscle function. VALIDATOR: confirm PMID 41877354 resolves to this DOM 2026 meta and the per-drug fractions (sema 35.2%, tirz 25.4%). Cross-ref T5-014, T5-010.
Comparatorslifestyle intervention; lifestyle plus resistance training
GLP-1 RA classMARKETED
Body compositionMixed
Moderate evidence
Network meta-analysis (22 RCTs, 2258 participants): GLP-1RAs reduced absolute lean mass by ~0.86 kg (~25% of weight lost), but the RELATIVE proportion of body weight...
Full findingNetwork meta-analysis (22 RCTs, 2258 participants): GLP-1RAs reduced absolute lean mass by ~0.86 kg (~25% of weight lost), but the RELATIVE proportion of body weight that is lean was unaffected. The same dataset supports both poles of the debate: 'relative lean unchanged' feeds the adaptive reading; the absolute -0.86 kg and the tirz/sema-least-preserving ordering feed the caution reading.
Fundingacademic / independent - no specific grant funding
Scope limitsNEW. The numeric backbone both poles cite; meta-analysis graded moderate by default. Entirely a MASS meta-analysis: no strength/gait/SPPB pooled because the component trials did not measure them. Consistent with the Annals review (C-CLASS-BODYCOMP-01, ~25-28% lean fraction). COROLLARY (Council Methodologist): 'relative lean unchanged' is a COROLLARY of the fixed lean-fraction-of-weight-lost property (cf. SURMOUNT-1 placebo parity), NOT an independent adaptive signal, and does not address absolute lean-kg or function. Cross-ref C-CLASS-BODYCOMP-01.
ADAPTIVE-POLE anchor (both poles preserved): reported lean-mass loss as a proportion of weight lost is heterogeneous (40-60% in some studies, <=15% in others); on...
SourceNeeland IJ et al., Diabetes Obes Metab, 2024Source
Full findingADAPTIVE-POLE anchor (both poles preserved): reported lean-mass loss as a proportion of weight lost is heterogeneous (40-60% in some studies, <=15% in others); on contemporary evidence including MRI muscle-volume work the skeletal-muscle changes appear largely ADAPTIVE (commensurate with the weight loss achieved, with improved insulin sensitivity and reduced muscle fat infiltration suggesting improved muscle QUALITY), rather than pathological wasting. Older age / disease severity flagged as risk modifiers.
PopulationNarrative review of trial + MRI-based evidence on GLP-1RA and dual GLP-1/GIP effects on lean body mass and muscle health in overweight/obesity.
Fundingacademic/independent analysis (manufacturer not study sponsor; per-study COI not individually audited)
Scope limitsADAPTIVE pole. KEY caveat the authors themselves make: DXA 'lean mass' includes organ, bone, fluid, so 'lean preserved' is NOT 'muscle/function preserved'; the adaptive reading rests on MRI muscle-volume + quality inference, NOT strength/function endpoints (which are largely absent, see C-CLASS-BODYCOMP-08). This is the thread-5 T5-014 cross-ref. Counter-pole: C-CLASS-BODYCOMP-06. Companion paper Linge/Birkenfeld/Neeland Circulation 2024 (PMID 39401279) makes the same case; folded here to avoid a duplicate row. COI (validation): co-author J. Linge is employed by AMRA Medical (a commercial MRI muscle-composition analytics firm), directly relevant to the MRI-based 'muscle quality preserved / adaptive' inference. CONFIDENCE (Council Evidence-grader + Red-team): this adaptive/muscle-quality inference is held at NO HIGHER confidence than the caution pole (C-CLASS-BODYCOMP-06); it is down-weighted because the MRI-composition methodology underpinning the 'quality preserved' claim is the AMRA-employed co-author's commercial product. Improved muscle quality / reduced IMAT has NOT been shown to offset quantity loss in any function endpoint (none exist), and IMAT reduction may track total fat loss rather than a muscle-specific effect.
SourceLocatelli JC et al., Diabetes Care 2024Source
Full findingCAUTION-POLE anchor (both poles preserved): narrative review argues incretin agents (liraglutide, semaglutide, tirzepatide, retatrutide named) inducing ~15-24% weight loss also cause rapid lean-mass loss of ~10% / ~6 kg, threatening muscle mass/function and raising sarcopenia/frailty risk, motivating resistance exercise and muscle-preserving adjuncts. The same DXA numbers the adaptive camp reads as benign are read here as a threat: the disagreement is INTERPRETIVE, not over the raw number.
PopulationNarrative review; adults with overweight/obesity on incretin therapy; older-adult sarcopenia/frailty framing.
Fundingacademic/independent analysis (manufacturer not study sponsor; per-study COI not individually audited)
Scope limitsCAUTION pole. Names retatrutide in the lean-loss list but cites NO reta-specific function data (see C-RETATRUTIDE-BODYCOMP-03 gap). Still a MASS argument, not strength data. The ~10%/~6 kg figure is a CLASS generalisation; NO reta-specific lean-kg supports reta's inclusion (the review names reta but does not measure it; Council Methodologist). The 'decade of ageing' is a framing analogy. Two further caution reviews fold here to avoid duplicates: Stefanakis/Mantzoros Metabolism 2024 (PMID 39481534, '>25% of loss is fat-free mass') and the JAMA viewpoint Conte/Hall/Klein 2024 (PMID 38829659, which frames the harm-vs-adaptive question as explicitly UNRESOLVED). Counter-pole: C-CLASS-BODYCOMP-05.
Multiple (review)MARKETED
Body compositionNot directional
Low evidence
DEFINITIONAL YARDSTICK: the ESPEN/EASO consensus defines sarcopenic obesity as excess adiposity PLUS low muscle MASS PLUS low muscle FUNCTION, with a two-step...
SourceDonini LM, Busetto L, Bischoff SC, Barazzoni R, Prado CM et al. (ESPEN/EASO Consensus), O...Source
Full findingDEFINITIONAL YARDSTICK: the ESPEN/EASO consensus defines sarcopenic obesity as excess adiposity PLUS low muscle MASS PLUS low muscle FUNCTION, with a two-step diagnostic pathway that confirms via a muscle-FUNCTION test (e.g. strength) before a mass measure. By this standard, calling incretin lean-mass loss 'sarcopenia' REQUIRES a demonstrated FUNCTION deficit, which the trials have not measured.
PopulationESPEN + EASO international expert consensus statement.
Fundingacademic/independent analysis (manufacturer not study sponsor; per-study COI not individually audited)
Scope limitsThe NEUTRAL yardstick for the harm-vs-adaptive debate: function-first diagnosis means DXA mass loss alone does not equal sarcopenia. Co-authored by Prado and Barazzoni (sarcopenic-obesity authorities). Predates the GLP-1 surge; applied here as the standard, not as a drug study. Graded LOW (expert consensus/guidance, no pooled estimate; Council Evidence-grader). BIDIRECTIONAL (Council Red-team): by ESPEN/EASO criteria, DXA mass loss alone CANNOT be called sarcopenia without a demonstrated function deficit, AND the absence of measured function CANNOT be called muscle-safe either - the class is currently undiagnosable in both directions. This yardstick governs adjudication of the harm-vs-adaptive poles.
GLP-1 RA classMARKETED
Body compositionNot directional
Low evidence
CLASS GAP (the absence is the finding): direct measurement of muscle FUNCTION (grip strength, gait speed, SPPB, chair-stand) is essentially ABSENT from the pivotal...
Full findingCLASS GAP (the absence is the finding): direct measurement of muscle FUNCTION (grip strength, gait speed, SPPB, chair-stand) is essentially ABSENT from the pivotal incretin weight-loss trials; body composition is captured almost entirely as DXA/MRI MASS. Multiple reviews (Neeland 2024; Dubin/Heymsfield 2024; Conte/Hall/Klein; Drucker 2024) converge that strength/function/mobility were not assessed and call for them. Authoritative safety reviews (Drucker, Diabetes Care 2024, PMID 38843460) list muscle strength as an under-characterised SAFETY domain for the class, including retatrutide, survodutide and MariTide.
PopulationClass-wide gap across STEP / SURMOUNT / SUMMIT / SUSTAIN body-composition substudies; synthesised in reviews + one protocol-stage older-adult RCT.
Fundingacademic/independent analysis (manufacturer not study sponsor; per-study COI not individually audited)
Scope limitsTHE ABSENCE IS THE FINDING. Dubin reviewed 28 GLP-1RA trials, all DXA mass, no functional testing, and call for MRI + functional testing. The two HF-PRO signals (6MWD, KCCQ) are folded into the magnitude with an explicit NOT-muscle-strength tag, so they are captured without miscategorising HF symptom relief as muscle function. Drucker PMID 38843460 folded here. Class-level counterpart to the reta-specific gap (C-RETATRUTIDE-BODYCOMP-03). The folded 6MWD/KCCQ are heart-failure-STATUS PROs confounded by dyspnoea/fluid/symptom relief - NOT gait-speed or strength - and do NOT narrow the muscle-function gap (Council Evidence-grader).
GLP-1 RA classMARKETED
Body compositionDecrease
Moderate evidence
FOLK-CLAIM VERDICT ('I'm losing MUSCLE not just fat; eat protein and lift'): MIXED, leaning TRUE-but-reframed. Lean-mass loss is real (~25-40% of weight lost is...
SourceBatsis JA et al., Ann Intern Med, 2026Source
Full findingFOLK-CLAIM VERDICT ('I'm losing MUSCLE not just fat; eat protein and lift'): MIXED, leaning TRUE-but-reframed. Lean-mass loss is real (~25-40% of weight lost is fat-free mass), but the 'losing muscle' framing overstates it - most is obligatory for any large weight loss, the split mirrors placebo/lifestyle, and the LEAN FRACTION is NOT worse with the more potent agents (incl. retatrutide), only the absolute kg. The protein+resistance-training rule is mechanistically sound but its GLP-1-specific trial evidence is MIXED. The genuine concern is functional (sarcopenia in older/sarcopenic-obese patients) and is UNDER-MEASURED (no trial assessed physical function).
PopulationSystematic review (35 RCTs) + DXA substudies of SURMOUNT-1 (tirzepatide) and the reta phase-2 T2D trial.
Scope limitsGRADE moderate (systematic review of RCT DXA data). MATCHED to the D2 lean-mass thread; this row adds the folk-verdict + the reframe (fraction not worse with potency; function under-measured) + the new systematic-review anchor. Belief-vs-reality: the scale drop is ~25-40% fat-free mass (expected, similar to placebo/lifestyle), mostly benign with composition improving - but the older/sarcopenic functional risk is real and untested, and protein+lifting is the right hedge. Reta loses most weight -> most ABSOLUTE lean kg, but not a worse fraction. Cross-ref the D2 body-composition rows.
Comparatorsplacebo; lifestyle weight loss
GLP-1 RA classMARKETED
Body compositionNot directional
Low evidence
FOLK-CLAIM VERDICT ('Ozempic face'): MIXED - the gaunt/hollow/aged/sagging face is REAL (TRUE), but the attribution to the DRUG is FALSE: it is a WEIGHT-LOSS-mediated...
SourceBarone M et al. GLP-1 receptor agonists and skin quality: a systematic review (incl. faci...Source
Full findingFOLK-CLAIM VERDICT ('Ozempic face'): MIXED - the gaunt/hollow/aged/sagging face is REAL (TRUE), but the attribution to the DRUG is FALSE: it is a WEIGHT-LOSS-mediated effect (non-selective loss of facial fat pads + dermal laxity that fails to retract, worse with age), not a drug-specific facial toxicity. The same occurs with any rapid/large weight loss including bariatric surgery and diet. NOT semaglutide-specific (named for first-mover fame); class-wide - and tirzepatide/retatrutide likely produce MORE facial change purely via GREATER weight loss. Folk preventatives (slower loss, more protein, hydration, fillers/microneedling) are mechanistically coherent with the volume-loss explanation.
PopulationDermatology + plastic-surgery systematic reviews of case reports/cohorts + authoritative clinical commentary (Cleveland Clinic); bariatric-surgery facial-change cohorts as the weight-loss-generic comparator.
Fundingacademic/independent analysis (manufacturer not study sponsor; per-study COI not individually audited)
Scope limitsGRADE low: dermatology/plastic-surgery review of mostly case reports + commentary; facial appearance has NEVER been a controlled GLP-1 trial endpoint, and no head-to-head drug-vs-equivalent-diet-weight-loss facial comparison exists. The directional verdict (weight-loss-mediated, not drug-specific) is consistent and well-supported; a minority direct-drug hypothesis (Kruglikov 2025, dermal white-adipose effect) is unproven. BELIEF-VS-REALITY: people believe 'the drug gives you the face'; reality - losing a lot of weight fast gives you the face, and the drug is just an efficient way to do that. Folk-claim frequency: EXTREMELY COMMON as a cultural term.
Comparatorsbariatric-surgery weight loss; diet weight loss
GLP-1 RA classMARKETED
Body compositionDecrease
Low evidence
In an older mostly-female cohort, semaglutide/tirzepatide-associated bone loss tracked weight loss and was greater in non-diabetic patients.
Scope limitsWEIGHT-LOSS-MAGNITUDE-MEDIATED, DM-status-modified; implies the T2D meta understates loss in obesity/non-diabetic use. Retrospective, single-centre.
LiraglutideMARKETED
Body compositionDecrease
High evidence
In a randomised DXA study, liraglutide reduced bone density more than exercise at the same weight loss, and resistance exercise prevented the loss.
SourceLiraglutide vs exercise on bone density (RCT DXA secondary analysis), JAMA Netw Open 2024Source
Fundingacademic / non-commercial (Novo Nordisk Foundation + Danish academic bodies; drug supplied in-kind; disclosed)
Scope limitsTHE PARTLY-DRUG-INTRINSIC bone row (the one not purely weight-mediated - BMD fell beyond what matched weight loss explains). Non-elderly (mean 43y); resistance exercise is protective. Must NOT be mislabelled weight-mediated.
GLP-1 RA classMARKETED
Body compositionNo change
Moderate evidence
In T2D-only RCTs, GLP-1 RAs showed no fracture excess (RR 0.80, NS) and no detectable BMD loss (BMD point estimates modestly higher than control) - a...
SourceGLP-1 RA bone density and turnover (25-RCT meta-analysis), Acta Diabetol 2025Source
Full findingIn T2D-only RCTs, GLP-1 RAs showed no fracture excess (RR 0.80, NS) and no detectable BMD loss (BMD point estimates modestly higher than control) - a reassurance-of-absence in diabetic bone, NOT a demonstrated bone-building effect. This does NOT generalise to obesity/weight-loss populations, where the direct double-blind Hansen RCT and DXA cohorts show resorption-dominant BMD loss.
Fundingacademic / investigator-initiated; no industry sponsor found in indexed metadata (Europe PMC fundingList empty)
Scope limitsREMEDIATED (WI-BMD-01): re-scoped from a near-anabolic class reading to a CONTESTED, T2D-meta-analytic-specific pole. T2D bone tends high-density/low-quality, so a meta of T2D RCTs can report 'no loss / modestly higher BMD' while the direct double-blind weight-loss RCT (Hansen, C-CLASS-BONE-MECH-RCT-01: P1NP null, CTX up, lumbar/hip aBMD down) and obesity cohorts (C-CLASS-BONE-BMD-COHORT-01) show loss. The discordance is explained by population (T2D vs obesity) and weight-loss magnitude; the higher-BMD reading must NOT be transferred to the rapid-weight-loss obesity setting. The fracture RR 0.80 NS is real and retained.
GLP-1 RA classMARKETED
Body compositionDecrease
Moderate evidence
In a 52-week double-blind RCT in adults with increased fracture risk (mostly postmenopausal), semaglutide 1.0 mg did NOT raise bone formation (P-PINP null) but RAISED...
SourceHansen MS, Wolfel EM, Jeromdesella S, et al. Once-weekly semaglutide versus placebo in ad...Source
Full findingIn a 52-week double-blind RCT in adults with increased fracture risk (mostly postmenopausal), semaglutide 1.0 mg did NOT raise bone formation (P-PINP null) but RAISED bone resorption (CTX) and LOWERED lumbar-spine and total-hip BMD versus placebo - a resorption-dominant pattern, not bone-building.
PopulationSemaglutide bone phase-2 double-blind RCT (Hansen et al., EClinicalMedicine 2024, NCT04702516); 64 adults at increased fracture risk (T-score < -1.0 and/or recent low-energy fracture), two Danish hospitals.
Fundingacademic/investigator-initiated (Region of Southern Denmark + Novo Nordisk Foundation + Gangsted Foundation; Novo Nordisk supplied study drug only, not sponsor)
Scope limitsTHE direct double-blind DXA/turnover RCT counterweight to the T2D meta (C-CLASS-BONE-BMD-META-01 / -TURNOVER-01): shows resorption-dominant loss, NOT the meta's favourable-turnover/higher-BMD reading. Graded moderate (phase-2), but it is a direct double-blind randomised DXA + bone-turnover measurement, methodologically stronger than the meta's pooled T2D RCTs for the obesity/weight-loss question. UNSETTLED whether drug-intrinsic or weight-loss-mediated - NO diet-only comparator arm, and the authors attribute the CTX rise to the accompanying weight loss. Postmenopausal/modest n; not an obesity-dose (1.0 mg) trial. Corrects and re-scopes the contested anabolic pole.
Comparatorsplacebo
GLP-1 RA classMARKETED
Body compositionNo change
Low evidence
In a 20-week pilot RCT in older adults (n=20), whole-body and regional BMD and bone-turnover markers did not change significantly, but a post-hoc correlation linked...
SourceDinkla L, Beavers KM, Robbins R, ... Trejo J, Stepanenko A, Cortes TM, et al. Bone minera...Source
Full findingIn a 20-week pilot RCT in older adults (n=20), whole-body and regional BMD and bone-turnover markers did not change significantly, but a post-hoc correlation linked greater weight loss to greater pelvis bone loss (r=0.62) - hypothesis-generating support for weight-loss-mediated loss, not proof.
PopulationOlder-adult GLP-1 bone pilot (Trejo/Cortes et al., Front Aging 2025, NCT05786521); 20 community-dwelling adults aged 65+ with prediabetes/T2D + overweight/obesity, UT Health San Antonio.
Fundingacademic/government (NIH/NIA + San Antonio VA GRECC + Colorado CCTSI; no industry sponsor)
Scope limitsVERY LOW/LOW grade (audit-downgraded from the open-label-RCT rule grade): n=20, pilot, unblinded assessors, whole-body BMD NS, the directional bone-loss trend did not reach significance, and the supportive r=0.62 is a single post-hoc regional correlation. Supports the weight-mediated reading WITHOUT establishing it; do not over-read the r=0.62. Note the paper itself REPEATS the unverified '~5-fold hip/pelvic fracture in 75+ (SELECT)' figure in its discussion - that magnitude is NOT encoded here (open question OQ-BMD-B).
Comparatorslifestyle counselling alone
GLP-1 RA classMARKETED
Body compositionMixed
Moderate evidence
In T2D, the meta reports a favourable-looking turnover shift (beta-CTX down, formation markers up), but this is CONTESTED as a generalisable bone-building effect -...
SourceGLP-1 RA bone density and turnover (25-RCT meta-analysis), Acta Diabetol 2025Source
Full findingIn T2D, the meta reports a favourable-looking turnover shift (beta-CTX down, formation markers up), but this is CONTESTED as a generalisable bone-building effect - likely low-turnover diabetic-bone normalisation. It is directly contradicted by the Hansen double-blind RCT in weight-losing adults (CTX UP, P1NP null) and so must NOT be read as a class anabolic effect.
Fundingacademic / investigator-initiated; no industry sponsor found in indexed metadata (Europe PMC fundingList empty)
Scope limitsREMEDIATED (WI-BMD-01): the favourable-turnover reading is re-scoped to a CONTESTED T2D-specific pole and explicitly REFUTED as a generalisable anabolic effect by the WI-BMD-01 deep-research pass. Most plausibly normalising low-turnover diabetic bone, NOT bone-building; directly OPPOSED by the direct double-blind weight-loss RCT (Hansen, C-CLASS-BONE-MECH-RCT-01: resorption marker CTX UP, formation marker P1NP null, BMD down) and the obesity BMD-loss cohorts. Do NOT transfer to the rapid-weight-loss obesity setting.
EnobosarmMARKETED
Body compositionMixed
Very low evidence
Veru phase-2b QUALITY trial (NCT06282458): enobosarm added to semaglutide in older adults (>=60 y) reported large reductions in lean-mass loss and a physical-function...
Full findingVeru phase-2b QUALITY trial (NCT06282458): enobosarm added to semaglutide in older adults (>=60 y) reported large reductions in lean-mass loss and a physical-function signal versus placebo+semaglutide. This is muscle PRESERVATION (anti-catabolic). NON-GRADUATING: the figures are a company press topline only, with no peer-reviewed publication.
PopulationQUALITY phase-2b dose-finding RCT, N=168, double-blind placebo-controlled, 16 weeks, adults >=60 y on semaglutide; enobosarm 3/6 mg vs placebo.
Fundingindustry - Novo Nordisk (trial sponsor; inferred from registration trial)
Scope limitssmall sample (N~168)
Comparatorsplacebo + semaglutide
LiraglutideMARKETED
Body compositionDecrease
High evidence
Phase 4 MRI RCT (high-CV-risk overweight/obesity, no diabetes): liraglutide 3.0 mg significantly lowered visceral adipose tissue over 40 weeks vs placebo (primary...
SourceNeeland IJ et al., Lancet Diabetes Endocrinol, 2021Source
Full findingPhase 4 MRI RCT (high-CV-risk overweight/obesity, no diabetes): liraglutide 3.0 mg significantly lowered visceral adipose tissue over 40 weeks vs placebo (primary endpoint VAT% by MRI).
Fundingacademic-led (University of Texas Southwestern) with Novo Nordisk collaborator
Scope limitssmall sample (N~185)
Comparatorsplacebo
LiraglutideMARKETED
Body compositionDecrease
Moderate evidence
Pre-specified MRI secondary analysis: liraglutide 3.0 mg reduced thigh-muscle fat (myosteatosis) and adverse muscle composition vs placebo - a muscle-QUALITY...
SourcePandey A et al., J Cachexia Sarcopenia Muscle, 2024Source
Full findingPre-specified MRI secondary analysis: liraglutide 3.0 mg reduced thigh-muscle fat (myosteatosis) and adverse muscle composition vs placebo - a muscle-QUALITY improvement distinct from muscle quantity.
Populationn=128 with follow-up MRI (liraglutide 73, placebo 55); 92.2% women, 36.7% Black; median 36 wk; same trial as NCT03038620
Fundingindustry-Novo Nordisk
Scope limitssmall sample (N~128)
Comparatorsplacebo
MazdutideINVESTIGATIONAL
Body compositionDecrease
Very low evidence
GLORY-1/GLORY-2 phase 3 (Chinese adults with obesity): mazdutide met key secondary endpoints including reduced waist circumference and liver-fat content; dual...
SourceInnovent / Eli Lilly, GLORY-1 and GLORY-2 phase 3 (mazdutide)Source
Full findingGLORY-1/GLORY-2 phase 3 (Chinese adults with obesity): mazdutide met key secondary endpoints including reduced waist circumference and liver-fat content; dual GCG/GLP-1 mechanism described as potentially yielding more favourable fat-vs-muscle partition, but no quantified DXA fat/lean split located.
PopulationGLORY-1 (n=610) and GLORY-2 phase 3 RCTs in Chinese adults with obesity; 48-60 weeks
Fundingindustry - Innovent / Eli Lilly (trial sponsor; inferred from registration trial)
Scope limitsGAP: quantified DXA fat/lean partition for mazdutide not found in this sweep; waist circumference and liver fat are the available body-composition-adjacent readouts. Flagged for follow-up against the published GLORY papers.
Comparatorsplacebo
Multiple (review)MARKETED
Body compositionNot directional
Low evidence
Mechanistic / landscape anchor: incretin weight loss is accompanied by fat-free-mass loss (heterogeneously ~15% to 40-60% of weight lost), motivating adjuncts that...
SourceAimelet V & Holst JJ, Diabetes Obes Metab, 2025Source
Full findingMechanistic / landscape anchor: incretin weight loss is accompanied by fat-free-mass loss (heterogeneously ~15% to 40-60% of weight lost), motivating adjuncts that preserve or add lean mass. The myostatin/activin-ActRII axis is a negative regulator of muscle growth; blocking it (at the receptor, bimagrumab; at ligands, trevogrumab/anti-GDF8, garetosmab/anti-activin A, apitegromab/anti-latent-myostatin) repartitions loss toward fat and can add lean mass. SARMs (enobosarm) anabolise via androgen-receptor selectivity. Bimagrumab and enobosarm have the most human data; most agents remain early-phase with limited long-term safety.
PopulationNarrative reviews of preclinical + phase-2/3 human studies of lean-mass-preserving pharmacotherapy during GLP-1RA / dual-agonist weight loss.
Fundingacademic (Aimelet & Holst, Novo Nordisk Foundation Center for Basic Metabolic Research, University of Copenhagen; no external funding statement; author COIs disclosed)
Scope limitsanimal data; human relevance uncertain
PemvidutideINVESTIGATIONAL
Body compositionDecrease
Very low evidence
MOMENTUM phase 2 MRI-based body-composition substudy: pemvidutide reported ~78.1% of weight lost as fat and ~21.9% as lean mass, framed by the sponsor as...
Full findingMOMENTUM phase 2 MRI-based body-composition substudy: pemvidutide reported ~78.1% of weight lost as fat and ~21.9% as lean mass, framed by the sponsor as 'class-leading' lean-mass preservation.
Fundingindustry - Altimmune (trial sponsor; inferred from registration trial)
Scope limitsconference/abstract-level; small sample (N~50)
Comparatorsplacebo
RetatrutideINVESTIGATIONAL
Body compositionDecrease
Moderate evidence
Retatrutide phase-2 T2D DXA substudy: fat mass fell dose-dependently; the lean-mass conclusion is a SECONDARY, proportional-only claim ('proportion of lean-mass loss...
SourceCoskun T et al. (Eli Lilly), Lancet Diabetes Endocrinol, 2025Source
Full findingRetatrutide phase-2 T2D DXA substudy: fat mass fell dose-dependently; the lean-mass conclusion is a SECONDARY, proportional-only claim ('proportion of lean-mass loss similar to other obesity treatments'), read by the sponsor as reassurance. CRITICAL: no ABSOLUTE lean-kg is reported, and proportional parity at a larger total loss is fully compatible with the LARGEST absolute lean-kg loss of any agent. Completer-only, T2D, fat-primary endpoint.
PopulationSubstudy of the retatrutide phase-2 T2D RCT (NCT04867785), double-blind placebo- and dulaglutide-controlled; 189 enrolled to substudy, 103 with paired baseline+week-36 DXA (completer-only); DXA at week 36.
Fundingindustry-Eli Lilly
Scope limitsDEEPENED (D2): foregrounds the decision-relevant gap - no absolute reta lean-kg anywhere, and the favourable lean conclusion is the sponsor's secondary proportional reassurance, NOT a powered equivalence result (thread-5 C9, OQ-T5-C). PMID DISAMBIGUATION (independent validation): the canonical reta Coskun substudy is PMID 40609566 (DOI 10.1016/S2213-8587(25)00092-0). A prior corpus note had carried PMID 40318682 for this paper - that is WRONG: 40318682 is a DIFFERENT paper (Sattar et al., SURPASS-3 MRI tirzepatide muscle composition, DOI 10.1016/S2213-8587(25)00027-0), NOT a duplicate. ATTRIBUTION (Council Methodologist): this is a WHOLE-DRUG DXA readout; it neither isolates nor implicates the GCGR arm - the GCGR-causes-net-lean-loss hypothesis is unresolved (thread-5) and cannot be inferred from a triple-agonist composite. Read WITH C-RETATRUTIDE-BODYCOMP-03 (the double absence), never alone. Cross-ref T5-001.
SourceSanyal AJ et al., triple agonist retatrutide for MASLD, Nat Med, 2024Source
Full findingPhase 2a MASLD substudy used MRI-PDFF to quantify hepatic fat alongside body-composition assessment; retatrutide markedly reduced liver fat content (ectopic-fat depot adjacent to body-composition remit).
PopulationPre-specified substudy of phase 2 obesity trial, participants with MASLD; MRI-PDFF
Fundingindustry - Eli Lilly (disclosed; retatrutide MASLD substudy)
Scope limitssmall sample (N~98)
Comparatorsplacebo
RetatrutideINVESTIGATIONAL
Body compositionNot directional
Moderate evidence
RETATRUTIDE GAP (double absence): (1) the phase-2 OBESITY trial reported NO DXA/body-composition, so there is NO obesity-population reta lean-mass data at the 12 mg...
SourceGap row derived from absence across the retatrutide phase-2 reports (PMID 37366315 weight...Source
Full findingRETATRUTIDE GAP (double absence): (1) the phase-2 OBESITY trial reported NO DXA/body-composition, so there is NO obesity-population reta lean-mass data at the 12 mg dose; the only reta DXA is the lower-weight-loss T2D substudy. (2) NO physical-function or muscle-strength data exist for retatrutide at all. CONSEQUENCE (Council Red-team): because retatrutide loses the MOST total weight of any agent (~24%) at an apparently constant ~75/25 fat:lean split, the arithmetic implies the LARGEST absolute lean-kg loss of any agent (thread-5 OQ-T5-C), yet it is unmeasured; and reta UNIQUELY adds a glucagon (GCGR) arm whose catabolic amino-acid/ureagenesis axis (thread-5) is a mechanistic reason its absolute lean loss could EXCEED dual/mono agents. Reta is therefore plausibly highest-absolute-lean-loss AND least-evidenced, NOT neutral-because-silent. Whether reta's lean-mass loss is harmful (true sarcopenia) or adaptive is UNTESTED for both obesity-magnitude mass and muscle function.
PopulationRetatrutide programme to date: phase-2 obesity (Jastreboff, NEJM 2023, PMID 37366315, weight only) and phase-2 T2D DXA substudy (Coskun, Lancet D&E 2025, PMID 40609566, fat-mass primary); no function endpoint in any reported reta trial.
Fundingacademic - Canadian Institutes for Health Research (CIHR grant 154321), Drucker review
Scope limitsExplicit reta gap (the absence is the finding); anchored to the Drucker safety review (a real source naming the gap) rather than left sourceless. Mirrors thread-5 OQ-T5-B (T2D reassurance transported onto obesity loss), OQ-T5-C (absolute lean-kg unquantified) and gap #4 (MRI muscle volume + grip/strength absent). The proportional reassurance in C-RETATRUTIDE-BODYCOMP-01 is DXA mass, not function. Cross-ref C9.
Comparatorsplacebo; dulaglutide
RetatrutideINVESTIGATIONAL
Body compositionNot directional
Moderate evidence
EXPLICIT ABSENCE: the retatrutide phase-2a MASLD trial (Sanyal 2024) did not report any DXA/BMD, bone-turnover or fracture endpoint.
SourceSanyal AJ et al. Triple hormone receptor agonist retatrutide for metabolic dysfunction-as...Source
PopulationRetatrutide phase-2a MASLD substudy (Sanyal et al., NCT04881760); reviewed for any DXA/BMD/bone-turnover/fracture endpoint - none present.
Fundingindustry-Eli Lilly
Scope limitssmall sample (N~98); prespecified/exploratory secondary of RCT (substudy)
Comparatorsplacebo
RetatrutideINVESTIGATIONAL
Body compositionNot directional
Moderate evidence
EXPLICIT ABSENCE: the retatrutide phase-2 obesity trial (Jastreboff 2023) did not report any DXA/BMD, bone-turnover or fracture endpoint - no retatrutide bone effect...
SourceJastreboff AM et al. (retatrutide phase 2), NEJM, 2023;389:514-526Source
Full findingEXPLICIT ABSENCE: the retatrutide phase-2 obesity trial (Jastreboff 2023) did not report any DXA/BMD, bone-turnover or fracture endpoint - no retatrutide bone effect can be asserted or excluded; all retatrutide bone inference is indirect class-extrapolation.
PopulationRetatrutide phase-2 obesity trial (Jastreboff et al., NCT04881760); reviewed for any DXA/BMD/bone-turnover/fracture endpoint - none present in the trial report.
SEMALEAN prospective study: semaglutide 2.4 mg reduced total fat mass with an initial lean-mass decline that stabilised after 7 months; handgrip strength improved and...
SourceAlissou M et al. (SEMALEAN), Diabetes Obes Metab, 2025Source
Full findingSEMALEAN prospective study: semaglutide 2.4 mg reduced total fat mass with an initial lean-mass decline that stabilised after 7 months; handgrip strength improved and prevalence of sarcopenic obesity fell from 49% to 33%; REE normalised to lean mass rose from M7 to M12.
Populationn=115 enrolled / 106 completers with obesity (68.9% female, mean BMI 46.3); DXA + handgrip + REE at M0, M7, M12; prospective single-arm
Fundingacademic/investigator - no external funding declared (NOT industry/Novo Nordisk)
Scope limitssmall sample (N~115)
SemaglutideMARKETED
Body compositionDecrease
Moderate evidence
SUSTAIN 8 DXA substudy: semaglutide 1.0 mg vs canagliflozin 300 mg in T2D reduced total fat mass, lean mass and visceral fat; proportion of lean mass rose ~1.2...
SourceMcCrimmon RJ et al., Diabetologia, 2020Source
Full findingSUSTAIN 8 DXA substudy: semaglutide 1.0 mg vs canagliflozin 300 mg in T2D reduced total fat mass, lean mass and visceral fat; proportion of lean mass rose ~1.2 percentage points; changes in visceral fat and fat-to-lean ratio comparable between arms.
Populationn=178 randomised to DXA substudy (sema n=88, cana n=90); 114 with end-of-treatment data; T2D on metformin; 52 weeks; RCT (NCT03136484)
Fundingindustry - Novo Nordisk
Scope limitssmall sample (N~178)
Comparatorscanagliflozin
SemaglutideMARKETED
Body compositionDecrease
Moderate evidence
STOP randomised trial (T2D): semaglutide reduced epicardial adipose tissue vs comparator over the trial period (epicardial-fat-specific RCT readout).
SourceManubolu VS et al. (STOP trial), J Am Coll Cardiol, 2024Source
PopulationSTOP (Semaglutide Treatment effect On coronary atherosclerosis Progression) randomised trial in T2D; CT-based epicardial adipose tissue
Full findingSYNCHRONIZE (phase 3) prespecified MRI substudy: survodutide reduced visceral adipose tissue ~34%, subcutaneous adipose tissue ~28% and lean body volume ~9.8%; fat mass ~78% / lean mass ~22% of total weight lost reported; lean tissue <=10.8% of total tissue-mass change at the highest dose.
PopulationPhase 2 obesity MRI substudy (survodutide / BI 456906); dose range to ~4.8 mg
Fundingindustry - Boehringer Ingelheim / Zealand
Scope limitsconference/abstract-level; no outcome data yet (ongoing)
Comparatorsplacebo
TirzepatideMARKETED
Body compositionDecrease
Low evidence
EHR-linked body-composition digital-phenotyping study (LLM extraction) of routine-care users found tirzepatide associated with GREATER relative lean-body-mass loss...
SourceGreater lean-body-mass decline with tirzepatide than semaglutide in routine care (digital...Source
Full findingEHR-linked body-composition digital-phenotyping study (LLM extraction) of routine-care users found tirzepatide associated with GREATER relative lean-body-mass loss than semaglutide at every timepoint; a 'Depletive GLP-1 metabotype' (>20% TBW loss with >5% LBM loss) was more frequent with tirzepatide.
Population670,422 first-episode GLP-1RA users (semaglutide 456,742; tirzepatide 213,680); 7,965 with paired pre/post body-composition over 12 months; retrospective real-world EHR cohort
Fundingacademic/independent analysis (manufacturer not study sponsor; per-study COI not individually audited)
Scope limitsconference/abstract-level; possible confounding by indication
Comparatorssemaglutide
TirzepatideMARKETED
Body compositionDecrease
Moderate evidence
SURMOUNT-1 DXA substudy: tirzepatide's large total loss was accompanied by a PROPORTIONALLY matched lean loss (~25% of weight lost), and crucially the same ~75/25...
SourceLook M et al., Diabetes Obes Metab, 2025Source
Full findingSURMOUNT-1 DXA substudy: tirzepatide's large total loss was accompanied by a PROPORTIONALLY matched lean loss (~25% of weight lost), and crucially the same ~75/25 split held for PLACEBO too, so the lean fraction is a property of weight loss at this magnitude, not specifically of the drug. The decision-relevant corollary: a much larger absolute loss at the same proportion carries a larger ABSOLUTE lean-kg loss.
PopulationSURMOUNT-1 DXA substudy: n=160 of 2539 (pooled tirzepatide n=124, placebo n=36), adults with obesity/overweight, mean weight 102.5 kg; baseline-to-Week-72 DXA; post-hoc substudy of a phase-3 double-blind placebo-controlled RCT.
Fundingindustry-Eli Lilly
Scope limitspost-hoc (not prespecified); small sample (N~160)
Comparatorsplacebo
TirzepatideMARKETED
Body compositionDecrease
Moderate evidence
SUMMIT CMR substudy (obesity-related HFpEF): tirzepatide reduced LV mass and paracardiac (epicardial + pericardial) adipose tissue vs placebo at 52 weeks; the...
SourceKramer CM et al., J Am Coll Cardiol, 2024Source
Full findingSUMMIT CMR substudy (obesity-related HFpEF): tirzepatide reduced LV mass and paracardiac (epicardial + pericardial) adipose tissue vs placebo at 52 weeks; the paracardiac reduction was driven by the pericardial component, with the LV-mass change paralleling weight loss.
Scope limitsEpicardial/pericardial depot. Coverage notes the volume drop was driven by PERICARDIAL not epicardial fat - a depot-specificity caveat worth preserving. Cross-ref heart-failure domain (SUMMIT).
Comparatorsplacebo
SemaglutideMARKETED
Body compositionMixed
Low evidence
COURAGE phase-2 (Regeneron, NCT06299098) tests trevogrumab (anti-myostatin) +/- garetosmab (anti-activin A) added to semaglutide. Interim results show dose-ordered...
Full findingCOURAGE phase-2 (Regeneron, NCT06299098) tests trevogrumab (anti-myostatin) +/- garetosmab (anti-activin A) added to semaglutide. Interim results show dose-ordered lean preservation, with the triplet most lean-preserving but least tolerated. Final readout pending.
The Echo Compendium