A medRxiv preprint posted April 13, 2026 (nference + Mayo Clinic) used EHR-linked digital phenotyping on 670,422 first-episode GLP-1 users to compare body-composition trajectories on tirzepatide versus semaglutide in routine clinical care. The headline: tirzepatide users lose roughly 2% more lean body mass than semaglutide users at 12 months, and a new "Depletive metabotype" — patients losing >20% body weight AND >5% lean mass — was 54% more common on tirzepatide (10.3% vs 6.7%).
Research-context information only. Peptides and pharmaceuticals discussed below are research compounds or prescription drugs. Protocols, doses, and reactions reported come from published research and self-reported community sources. This article reports what has been documented, not what should be done. Consult a licensed physician for personal medical decisions.
This is the first large-scale, real-world body-composition comparison of the two market-leading GLP-1 drugs outside the trial environment — and it tilts the muscle-preservation question more sharply toward semaglutide than the SURMOUNT and STEP DXA substudies suggested.
What the New Data Actually Shows
The analysis tracked 670,422 first-episode GLP-1 receptor agonist users in EHR data: 456,742 on semaglutide and 213,680 on tirzepatide. Of those, 7,965 had paired pre- and post-initiation body-composition measurements within 12 months, which forms the analytic core.
Lean body mass loss on tirzepatide exceeded semaglutide at every measured timepoint:
| Timepoint | Excess lean mass loss on tirzepatide vs semaglutide |
|---|---|
| 3 months | 1.1% more |
| 6 months | 1.5% more |
| 9 months | 1.3% more |
| 12 months | 2.0% more |
The gap widens over the first year and stays statistically significant after adjustment for baseline weight, sex, age, and weight-loss trajectory.
The more striking finding is the Depletive metabotype prevalence. The authors define this as a patient who loses more than 20% total body weight AND more than 5% lean body mass within 12 months — high efficacy with high muscle cost. On tirzepatide, 10.3% of patients hit this profile. On semaglutide, only 6.7% did (p<0.001). Tirzepatide's greater weight-loss efficacy comes with a 54% higher chance of crossing into Depletive territory.
The flip side they call the Prime metabotype: high weight loss with preserved lean mass. Most patients on either drug don't reach either extreme, but the distribution skews differently — tirzepatide pulls the curve toward Depletive, semaglutide skews toward Prime.
Among 3,746 examined EHR phenotypes, baseline musculoskeletal pain was the strongest predictor of lean mass loss. Cervicalgia (neck pain) carriers lost 4.1 percentage points more lean mass on semaglutide and a striking 14.3 percentage points more on tirzepatide. Knee pain carriers lost 4.8 percentage points more on semaglutide and 13.4 percentage points more on tirzepatide. The drug effect is real, but the patient effect — being mobility-limited at baseline — amplifies the tirzepatide gap by roughly 3-4x relative to semaglutide.
Why This Matters More Than Trial DXA Data
The SURMOUNT and STEP DXA substudies that established the "25-30% of weight loss is lean mass" baseline were done in trial populations: motivated patients, structured visits, baseline DXA, often dietitian contact. The new analysis is the opposite. EHR data captures the average patient on the average prescriber's average schedule — no structured resistance training, no protein counseling, no DXA. Whatever the trials measured under ideal conditions, this is what the body-composition curve looks like in the wild.
Two things change in real-world data:
-
Lean mass loss is higher in absolute terms because patients aren't doing the resistance training that trial sub-populations sometimes do. The "natural" lean-mass loss on a GLP-1 without exercise is closer to 30-40% of total weight lost, not the 25% that body-composition substudies cite.
-
The drug-to-drug gap is sharper. Tirzepatide's larger total weight loss compounds with the lack of exercise stimulus. The 2% extra lean-mass loss at 12 months sounds small until you scale it: on a 100 kg starting weight with 25% body fat (25 kg fat, 75 kg lean), an extra 2% lean mass equals 1.5 kg of additional muscle attrition over a year — roughly one full year of natural aging muscle loss, on top of whatever the drug class is already costing you.
The patient-effect amplifier is the part that's new. SURMOUNT and STEP didn't enroll for baseline musculoskeletal pain or stratify by mobility status. This data does — and the message is that the patients who probably need GLP-1s most (obese, knee-pain, can't easily exercise) are exactly the ones losing the most lean mass on the more potent option.
What Tirzepatide Buyers Should Do Now
The practical implications, ordered by leverage:
1. If you're starting tirzepatide and you have baseline knee, hip, back, or neck pain, the conversation about semaglutide vs tirzepatide is now more nuanced than "tirzepatide loses more weight." The musculoskeletal-pain phenotype amplifies lean-mass loss roughly 3x on tirzepatide vs semaglutide. For a mobility-limited patient, the absolute weight-loss advantage of tirzepatide may not be worth the body-composition trade-off.
2. Resistance training 2-3 days per week is non-negotiable, especially on tirzepatide. The trial DXA data suggests structured resistance training partially blocks GLP-1 lean-mass loss. The new EHR data suggests its absence is one of the biggest reasons routine-care patients lose more muscle than trial subjects. Walking is not enough. Bodyweight squats, push-ups, rows — or a gym program — at minimum.
3. Protein floor: 1.2-1.6 g per kg body weight per day, divided across 3-4 meals. GLP-1s suppress appetite, which suppresses protein intake by default. Whey or casein shakes, Greek yogurt, eggs, and protein-fortified meals are easier to hit than chicken breast when you're not hungry.
4. Consider a muscle-preserving peptide stack from day one, not after you notice the loss. Lean mass is hard to add back at any age — much harder after 50 — so prevention beats remediation. The most-used categories:
- Cagrilintide — amylin analogue, appetite suppression via a non-GLP-1 mechanism. The cagrilintide-semaglutide combo (CagriSema) shows better fat-to-lean loss ratios than GLP-1 monotherapy in REDEFINE-1 body-composition data. The most evidence-backed add-on.
- Tesamorelin + Ipamorelin — restores GH/IGF-1 axis. Most-discussed adjunctive stack for "GLP-1 plus muscle preservation" in community protocols.
- MK-677 (ibutamoren) — oral GH secretagogue. Cheap and easy, but elevates IGF-1 long-term and may worsen insulin resistance. Less appropriate for diabetic patients.
5. Monitor body composition, not just the scale. A DXA scan every 6 months is the gold standard. Cheaper: bioimpedance plus grip strength (squeeze a hand dynamometer at baseline and every 3 months). If grip strength falls more than 5 kg from baseline, that's a body-composition red flag — talk to your prescriber about adjustments.
For vendors with the relevant compounds:
- Best cagrilintide vendors — current pricing and stock
- Deep dive: Best Tesamorelin + Ipamorelin Vendors — pre-mixed GHRH+GHRP blend
- Best MK-677 vendors — oral GH secretagogue
- Best tirzepatide vendors — current pricing if you're staying on tirz
- Best semaglutide vendors — current pricing if switching
- All vendor discount codes — current coupon round
