Ozempic, Muscle Loss, and Misleading Math
by Ola Hansson
In response to Crémieux’s article Ozempic and Muscle Mass”
(The substack restricts comments to paid subscribers, so I'm publishing this note separately.)
The Problem with Simplistic Body Composition Claims
While I'm not taking a position on whether Ozempic causes more or less muscle loss than natural weight loss, I do see issues with the reasoning about Claim #1:
In each trial, the opposite is true: people lose more fat than muscle, improving their body composition as they lose weight.
Losing more fat than muscle does not necessarily improve body composition. Whether it does depends on your starting point and on what you mean by better composition – lower fat percentage or higher muscle percentage. Those metrics are linked, but they don’t always move in lockstep, or even in the same directions:
- Usually, you raise muscle percentage by lowering fat percentage (good)
- But you can also lower muscle percentage while lowering fat percentage (maybe good for health, definitely bad for relative strength and athleticism)
- And you can even raise muscle percentage while raising fat percentage (maybe bad for health, good for relative strength and athleticism).
The Critical Numbers from Semaglutide Research
In the major semaglutide trial, the average loss was 8.36 kg of fat and 5.26 kg of lean mass. That is roughly 61% fat vs. 39% lean mass.
(Lean body mass, in general, is not the same as muscle. Much of it consists of a constant mass of tissue, organs, and bone. But, here, in the context of losing weight, we are concerned about the variable part. And the variable part is mostly muscle.)
Fat Percentage Changes: Not Always Straightforward
So what happens when your weight loss consists of 61% fat?
- A person who starts with less than 61% fat will see that percentage drop.
- But an extremely obese person with more than 61% fat will actually see fat percentage rise, because he is losing a “better-than-average” slice of his body.
Most people will of course drop fat percentage sharply at that ratio. But the point is a logical one: loss ratio alone doesn’t guarantee success. It must be compared to the original ratio.
Muscle Percentage: Where the Real Concern Lies
And when it comes to muscle percentage, this is not just logical pedantry – it’s a crucial point. A loss ratio of 39% muscle is not obviously beneficial for a majority of people.
Look at it as the mirror image of the fat percentage case:
- Anyone who starts with more than 39% muscle raises that percentage.
- But people below 39% – and that includes most old men and practically all women – will see muscle percentage drop. These people, just like the super-obese case, will lose a better-than-average slice of their bodies. They will thus lose not only absolute strength but also relative strength and the general movement skills and speed that depend on it.
This is bad news for the groups that can least afford to lose muscle.
The Study's Limitations
In the original study, they say:
Although total lean body mass decreased in absolute terms (kg), the proportion of lean body mass relative to total body mass increased with semaglutide.
But this is also not sufficient to draw conclusions about changes in muscle percentage and relative strength. Lean body mass ≠ muscle mass and the study never separates the two.
To calculate the baseline muscle percentage of the participants, we would need the baseline muscle mass, but the study never measured that. We can only be sure that the muscle mass is a LOT less than the 52.4 kg listed under “total lean body mass”, which includes the constant mass of tissue, organs, and bone already mentioned.
(73% of the particpants were women and the average muscle mass of an adult US women is about 17–23 kg).
If the constant part of lean mass is large enough, “lean body mass relative to total body mass” will inevitably increase with any fat decrease – however tiny! It says nothing about muscle mass. Without a baseline muscle percentage, it is still impossible to gauge how the 61/39 ratio affects muscle percentage.
The situation gets even muddier as DXA-scanners probably underdetect fat that’s imbedded inside muscle and other tissue. Furthermore, the underdetection might vary in proportion to fat percentage (measured inaccurately by DXA to begin with). Basic tests of relative strength begin to seem like the only sensible way to determine whether muscle percentage has actually improved or not!
The Broader Perspective
You could argue that a lower fat percentage should be top priority for systemic health, and I wouldn’t disagree. I just want to point out that this pattern of weight loss has drawbacks for performance and quality of life, whatever drug or diet delivers it.
Losing “more fat than muscle” isn’t always an upgrade.
