Body Recomposition on a GLP-1: When the Scale Stalls but Your Body Doesn't

Body recomposition is the simultaneous loss of fat and gain — or preservation — of lean muscle tissue. It hides on the scale because a bathroom scale measures one thing: the total downward force of everything in your body, including bone, organs, muscle, fat, glycogen, and about 60% water.

Muscle tissue has a density of roughly 1.06 g/ml. Fat tissue sits near 0.9 g/ml. That means a pound of muscle occupies about 18% less space than a pound of fat. Swap two pounds of fat for two pounds of muscle and the scale reports nothing happened, while your waistband reports something quite significant did.

This is not a new phenomenon, but it is unusually relevant on a GLP-1 for one reason: the standard trajectory on these medications is fast early loss followed by a slower phase. In SURMOUNT-1, participants on 15 mg tirzepatide reached 20.9% total body weight loss at 72 weeks, but the steepest portion of that curve was in the first 6-9 months (Jastreboff et al., *NEJM*, 2022). When rapid loss slows, many people interpret the change in slope as a failure — right at the point where the composition of what they are losing has actually improved.

There is a second reason the scale misleads on a GLP-1 specifically. These medications alter fluid balance, sodium handling and glycogen stores. Glycogen binds roughly three grams of water per gram stored. Increase your carbohydrate intake by 100 grams for a few days — say, a holiday weekend — and you can add two to three pounds of pure water weight that has nothing to do with fat.

So the scale is not lying. It is answering a question you did not ask.

You tell them apart by measuring something other than weight. Recomposition and a genuine plateau produce identical scale readings and completely different everything else.

Signs you are recomposing:

• •Waist circumference is dropping — even a quarter inch a month counts
• •Clothes fit differently, especially at the waist and upper arms
• •You are adding reps or weight in the gym
• •Progress photos in the same light and pose look different at 8 weeks
• •Energy is stable and you are eating adequately

Signs it is a genuine stall:

• •Waist, hips and arm measurements are unchanged over 6-8 weeks
• •Strength is flat or declining
• •You feel cold, tired, and are eating well under 1,200 calories
• •Sleep is poor and recovery from workouts is slow
• •You have been at the same dose for many months

That second cluster deserves attention rather than more restriction. Under-eating is one of the most common causes of a true stall on a GLP-1, because losing lean tissue reduces resting energy expenditure and makes the deficit shrink from the other end. Our guide to [why GLP-1 stalls happen and how to break them](/blog/glp1-weight-loss-plateau-why-stalls-happen-and-how-to-break-them) covers dose, sleep, alcohol and other drivers, and [eating enough on a GLP-1 during menopause](/blog/eating-enough-on-glp1-during-menopause-calorie-floor) covers the calorie floor question directly.

A practical rule: give any apparent plateau eight weeks and a tape measure before you change anything. Four weeks is not long enough to distinguish signal from water weight, especially for anyone still cycling hormonally.

Measure four things, on a fixed schedule, and let the scale become one data point among several rather than the verdict.

1. Waist circumference — every two weeks. Measure at the navel, first thing in the morning, after using the bathroom, tape snug but not compressing. Waist circumference correlates with visceral fat — the metabolically active fat around your organs — better than weight does, and visceral fat is the fat that matters most for cardiovascular and metabolic risk. A drop here while the scale holds is unambiguously good news.

2. Strength numbers — every session. Write down the weight and reps. A squat that went from 8 reps at 60 lb to 10 reps at 70 lb over six weeks is proof of preserved or gained muscle, and no scale can tell you that.

3. Progress photos — every four weeks. Same three angles, same lighting, same time of day, same minimal clothing. Most people cannot see change in the mirror day to day because the change is too gradual; a side-by-side at eight weeks is often startling.

4. Body composition — every three to six months, if accessible. A DXA scan (dual-energy X-ray absorptiometry) gives you fat mass, lean mass and bone density in one measurement and typically costs $50-150. Smart scales using bioimpedance are far less accurate in absolute terms but are reasonably good at showing direction over time if you use them under identical conditions.

And if you do weigh: weigh daily, but only look at the 7-day rolling average. Individual days are dominated by sodium, hydration, glycogen and bowel contents. The weekly average strips most of that out.

We cover the full measurement protocol in [tracking progress beyond the scale](/blog/tracking-progress-on-glp1-in-menopause-beyond-the-scale), including a simple template you can keep on your phone.

You can build some, and you can almost certainly preserve most of what you have — which is the more important goal for the majority of people on these medications.

Building meaningful new muscle in a calorie deficit is difficult for anyone. It happens most readily in three groups: people new to resistance training, people returning after a long layoff, and people carrying a substantial amount of body fat, which functions as an internal energy source. Many GLP-1 users fall into at least one of those categories, which is why beginner lifters on these medications sometimes see genuine strength and size gains alongside fat loss.

For everyone else, the realistic and still-valuable goal is preservation. This matters a great deal, because the default without intervention is not neutral. The STEP 1 body-composition substudy found approximately 40% of the weight lost on semaglutide came from lean tissue (Wilding et al., *NEJM*, 2021). The SURMOUNT-1 substudy reported a more favourable ratio for tirzepatide, closer to 25%. Neither is where you want to be if you can influence it.

What shifts the ratio, in order of impact:

Resistance training, 2-3 times a week. Non-negotiable. Mechanical load is the signal that tells the body to keep muscle when energy is scarce. Full-body sessions of 30-40 minutes are sufficient; you do not need a bodybuilding split.

Protein at 1.2-1.5 g per kg of body weight. Distributed across three meals of 25-35 grams each, since muscle protein synthesis responds to per-meal dose rather than daily total alone.

Not under-eating. A very steep deficit overwhelms the protective effect of training and protein.

Sleep. Growth hormone and muscle repair are sleep-dependent. Chronic short sleep measurably increases the proportion of weight lost as lean mass.

Our guide to [preserving muscle with strength training on a GLP-1](/blog/strength-training-on-glp1-how-to-preserve-muscle-while-losing-weight) has a specific beginner programme, and [the sarcopenia double risk in menopause](/blog/glp1-menopause-muscle-loss-the-sarcopenia-double-risk) explains why this is more urgent for women over 45.

Yes — and for most people it should mean abandoning a specific number in favour of a range and a set of physical markers.

A goal weight chosen at 30, or derived from a BMI chart, does not account for how much muscle you carry. Two women at 165 lb and the same height can have a 12-point difference in body fat percentage and look and function entirely differently. If you spend a year on a GLP-1 doing serious resistance training, arriving at 165 lb with more muscle than you have ever had is a categorically better outcome than arriving at 155 lb depleted.

This is more than cosmetic. Lean mass drives resting metabolic rate, the calories you burn at rest, which typically accounts for 60-70% of total daily energy expenditure. It also drives insulin sensitivity, functional strength, fall resistance and bone loading. The STEP 1 extension study found participants regained roughly two-thirds of lost weight within a year of stopping semaglutide (Wilding et al., *Diabetes, Obesity and Metabolism*, 2022) — and someone who preserved lean mass on the way down has a materially easier maintenance phase than someone who did not.

Better targets than a scale number:

• •Waist-to-height ratio below 0.5 — a simple, well-validated marker of cardiometabolic risk
• •A strength standard — for example, carrying half your body weight in each hand for 60 seconds, or a set of 10 unassisted push-ups
• •A body fat range rather than a weight, if you have DXA access
• •Functional goals — a flight of stairs without breathlessness, a hike you have wanted to do

If you are approaching maintenance, this reframing becomes practical rather than philosophical. [Finding your GLP-1 maintenance dose](/blog/glp1-maintenance-dose-how-to-keep-weight-off) covers the medication side; the training and protein side is what determines whether the body you maintain is one you want to keep.

The single most useful sentence for anyone in a flat-scale stretch: the scale measures how much of you there is, not what you are made of. On a GLP-1, what you are made of is the part you can still influence.