Before the big study, nobody was paying attention

The story actually started way earlier. Sytheon, the company that manufactures bakuchiol, first published a paper on the “retinol-like” qualities of bakuchiol in 2014.

No one paid much attention to it, and for good reason. It was published in a journal with a generally low reputation, and the lead author was from Sytheon. There was no evidence that bakuchiol acted on the same pathways as retinol. In fact, it was abundantly clear that bakuchiol did not activate the classic nuclear receptors that retinoids activate. The clinical “data” was flimsy. The study wasn’t registered, had no control group, wasn’t blinded, and didn’t have IRB approval.

So things got more serious in 2018, when the British Journal of Dermatology accepted a study that claimed to be the real thing. This time its authors came from notable academic institutions, including the University of California, Davis in Sacramento and the University of Michigan.

Subscribe now

This study “made” bakuchiol. It’s no surprise that shortly after it came out, Dr. Sam Bunting’s blog cited it and shared the results uncritically. Others quickly piled in, and Dr. Anil Rajani even called bakuchiol a “retinol killer.” But he clearly hadn’t read the original Sytheon paper, because he says bakuchiol works through the same receptors as retinol, which it doesn’t.

Dermatologists and influencers were saying:

“It helps prevent the fine lines and wrinkles, and with pigmentation, elasticity and firmness. You know what retinols do. They’re time-proven, tested, people are used to them. But what I like about it is it works through the same receptors that retinol uses.”

“Does it do the same thing as retinol? It’s thought to have a lot of the same clinical benefits. It’s thought to work in a different way, but have kind of a similar end point, both on a cellular level and then clinically as well. There’s now some decent quality data supporting its actions in the skin in real people. The paper of note that has struck me is the one published in the British Journal of Dermatology, which is a respected journal.”

That seems like it makes sense, right? The British Journal of Dermatology is a reputable journal, so if a clinical study was published there, especially by seemingly independent academic researchers from the University of Michigan and UC Davis, you’d believe it.

Well, in my view, this “bedrock” study is worthless, and actually fake, and the clinicians who ran it deceived not only the reviewers at the British Journal of Dermatology, but the public at large.

How do you cheat in a clinical study?

It actually isn’t hard to cheat, even in a clinical study. When the referee isn’t watching, you can cheat.

The most common form of cheating in clinical studies is to change the definition of success after you’ve seen the results. And this is what the bakuchiol researchers did.

The first protocol they submitted to ClinicalTrials.gov is dated April 7, 2017, by which point the study had already started. The primary endpoint was poorly defined: they said they were going to assess wrinkle appearance and pigmentation, but they weren’t very clear about what they meant.

They also made one very clear mistake. The protocol said they would photograph the subjects before and after to measure wrinkle appearance and pigmentation, and that board-certified dermatologists would grade those photographs on a Likert scale.

But they didn’t do that.

The study was completed in November 2017. Then, on May 5, 2018 (after the study was over), a new protocol was submitted that changed the endpoint. This time the photographs wouldn’t be graded by dermatologists; instead, they would use imaging-based analysis.

See it to believe it. This is the record, and the history of changes.

What makes this case particularly egregious is that they didn’t even specify how the imaging would be used. Basically, what happened is they said, “We’re going to change the rules, and we’ll tell you exactly how later.”

They don’t explain why they made that change or why they made it after the study was over. So who made this decision, and why?

The 100-meter race that ends at 23 meters

Let’s look at an equivalent. You organize a 100-meter race and tell the runners the distance that they’re racing (100 meters). The runner you wanted to win doesn’t win. So you go back and look through the footage. You see that at 23 meters, your runner was ahead. So you say, “Now we’re going to pick the winner based on who was ahead at 23 meters.” Your runner won!

That’s what happened here. The definition of success was changed in a very deceptive manner.

Subscribe now

Cherry-picking after moving the goal line

After moving the goal line, they also cherry-picked, but that only became visible when the manuscript was published.

The imaging software they used can produce 33 different measures of wrinkle appearance and pigmentation. But the researchers reported only one measure of wrinkles and two measures of pigmentation. And they didn’t explain why those three represent the “efficacy” of bakuchiol and retinol and the other 30 don’t. They didn’t even report the results on the other 30. Have a look at the software for yourself.

Logically, if you take 33 measures and get only three “positives,” you have clearly failed.

So what happened? The study was completed. The photographs were taken, and the dermatologists graded them. Someone looked at those results and said, “These aren’t that great. Let’s measure something else.” So they ran the imaging-based analysis and picked the three that showed equivalence between retinol and bakuchiol. What is even more laughable is that even that equivalence is debatable.

Now do you see why this can truly only be viewed as cheating?

One last important detail: the study calls itself a prospective, blinded clinical trial. “Prospective” means the measure is set in advance. But that label was used without explaining how the endpoint was set prospectively, and without explaining how the blinding was done, or whether any blinding was done at all. This, in my view, is completely misleading.

The missing statistics

Oh, one more thing. The study doesn’t actually report the statistics for its primary endpoint.

I have actually never seen a clinical trial where the statistical evaluation of the primary endpoint goes unreported. This is unprecedented. I couldn’t believe it.

This is how the cosmetics industry works

Unfortunately, this is typical of studies we see in the cosmetics industry. Companies commonly measure 20 endpoints and then, after the fact, pick two or three and say that they are the ones they meant when they started the study.

Take a look at the so-called clinicals of any company that has produced a clinical study, and you’ll see two things.

First, you will see that they didn’t select a primary endpoint in advance. In my experience, almost no one picks one thing they are going to measure and define as success.

Second, if they do select a primary endpoint in advance, they alter how that endpoint is measured, in a way that’s favorable to them. And their statistics are bent.

What happened when I asked questions

I wrote to the authors and asked them to explain what looked to me like blatant cheating. They didn’t respond. I wrote again, and I called them. They still didn’t respond.

So on July 23, 2024, I wrote to the British Journal of Dermatology to see how they would respond. After months of prodding, I got this reply by email:

“We conducted an in-depth investigation and felt that the raised concerns do not warrant a retraction according to the COPE guidelines and that the issues have more to do with suboptimal reporting than data fabrication. However, you may wish to submit a formal Letter to Editor for possible publication, which the original authors can then respond to.”

In other words, they didn’t find evidence of plagiarism or fabrication, which are the only grounds on which they would retract a study.

To me, it’s fraud.

So I took them up on their invitation and wrote a letter to the editor, with references, formatting and everything in the format that they require. Within hours of submitting it, the letter was rejected, which is stunning speed for the British Journal of Dermatology. Here’s part of the response from the editor-in-chief, Prof. John McGrath:

“The trial you comment on was accepted by the BJD in Sept 2018 and published in 2019. I regard the time interval too long for the topic still to be of interest to the BJD readership. Your letter raises a number of points in the style of a reviewer reviewing the article. As such, I think you should contact the original authors directly and have a private communication about the points you have raised. Having discussed your submission with other current BJD editors, I’m afraid the decision is not to publish your letter. I would also like to close correspondence on the matter.”

It wasn’t that anything I said was untrue. It was that too much time had gone by and their readers wouldn’t be interested. I had already tried to contact the authors separately, several times.

It’s remarkable that these are the standards of the British Journal of Dermatology. Shame on you.

Who benefited

They should care, because this study was an incredible win for Sytheon. After the study, Sytheon started supplying hundreds of companies with bakuchiol to make new “natural alternatives to retinol.” Sytheon even publishes a chart of the number of bakuchiol brands launched each year, and attributes that success to the BJD study.

I don’t have as strong evidence for this as I do for the endpoint switching, but I believe the researchers misled us about why the study was done.

Why do such a poor study? My only idea is that it was to make bakuchiol look good. And for that to be the purpose, I’d expect there to be some financial incentive behind it.

The paper doesn’t explain where the funding came from. It makes it seem as though it came from UC Davis in Sacramento: that scientists there were simply curious whether bakuchiol was really a natural alternative to retinol, and put up the roughly $100,000 it would cost to run the study. I don’t think that’s what happened. But they don’t tell us, and when I wrote to them and asked them to explain the funding, they didn’t write back.

There’s one more gap in the manuscript. Normally, a study reports where the test agent came from. Did the researchers buy it from the company, which probably sponsored the study, or was it gifted? This paper doesn’t say. That leaves the door open to the possibility that the company supported the study and that this support went undisclosed.

Let’s hold them accountable

It’s unfair that the British Journal of Dermatology rejected my letter to the editor, so I’m asking for your help.

You can reach the British Journal of Dermatology at bjd@bad.org.uk. Please write to them and simply say, politely: We hear that you rejected the letter to the editor regarding the bakuchiol study, and we would like that letter to appear as quickly as possible.

Let’s hold these people accountable. Let’s stop accepting this as real clinical science.

Had you heard that bakuchiol was “the natural retinol”? Has it worked for you, or not? Tell me in the comments.

Leave a comment

I have been working with aestheticians around the world, and they keep saying the same thing: clients who get Botox usually have weak, brittle skin that is difficult to work with.

So what is Botox doing to the skin? I think the answer is clear once you look at how Botox works. Keep reading for my best attempt to explain this, with full disclosure on where the evidence is solid, where it is missing, and where I am speculating.

What Botox does

Botox paralyzes muscles by inhibiting contraction. A muscle that cannot contract cannot crease the skin above it.

But muscle contraction also has an important job with for the lymphatic system on the face. The lymphatic system is essential to facial skin health and relies on muscle contraction to clear lymph fluid. When contraction is reduced, clearance is reduced.

My hypothesis

Here is the path I think runs from a Botox injection to weak skin:

1. Botox inhibits muscle contraction.

2. Less contraction means less lymph clearance.

3. Free fatty acids, a major component of lymph fluid, are not cleared. Fat cells take them up and become bloated.

4. Bloated fat cells signal the fibroblasts in the skin to stop making collagen and elastin, and to start making enzymes that degrade collagen.

If you have followed my work, you know the last step: bloated fat cells lead to weak skin. The new part is the first three steps.

The study nobody has done

I prefer to base what I say on published, peer-reviewed literature. So ideally, I would point you to a paper measuring how much Botox injections in the face reduce lymph clearance.

The issue is that this study has not been done. Although it could be.

I am not the only one who has noticed this gap. A 2024 paper from a group in Zurich, looking at Botox and lymphatic cells, put it this way: “it has been striking that while a few reports examined the effects of BoTA on the blood vascular system, such studies have not been performed for the lymphatic vasculature” (Vasella 2024).

So I looked for indirect evidence.

A case you can see

I found a paper that shows Botox inhibiting lymph flow and causing swelling: “Nonallergic eyelid edema after botulinum toxin type A injection: case report and review of literature” (Chang et al., Medicine, 2015).

The patient received Botox injections for crow’s feet. In the before photos, she is smiling with her eyes. Three days after the injection, there is darkening and swelling, most prominent in the upper eyelids. Two weeks after, significant facial swelling remains.

Her doctors treated her successfully by having her blink repeatedly, along with warm compresses. The blinking was the key: it forced the muscles to contract through the Botox, and that contraction cleared the lymph.

This is the mechanism in a single patient. When the muscle stops working, the fluid accumulates. When the muscle works again, the fluid clears.

How common is it?

The paper also asks how representative this case is.

The authors cite a meta-analysis by Brin et al. of 1,678 participants treated only for glabellar lines and crow’s feet. In that population, about 1.4% developed clinically relevant eyelid edema, meaning swelling visible enough for the clinician to note it.

In Asian subjects, the rate was more than double, at 3.1%. Asian eyelids are structured differently from Caucasian eyelids, and I suspect a blink produces less contraction and therefore less force to clear the lymph. If that is correct, it would make them more susceptible to this effect, which fits the idea that loss of contraction is what matters.

The tip of the iceberg

Is 1.4% a lot?

I think these are the most visible cases, the ones obvious enough that people return to the doctor and say, “My eye is swollen. Please do something.”

People without visible swelling, I would argue, are still experiencing reduced lymph clearance. Based on what I see, I would estimate a reduction of 30 to 50%, but this has not been measured. If I am right, it happens every time they are injected.

As I have covered before, a difference in fat cell size of less than 10% is associated with meaningful increases in inflammation (Henninger 2014). Inflammation, in turn, weakens skin structure.

That would explain what aestheticians are seeing in clients who get Botox repeatedly.

What we don’t know

I want to be clear about where this is solid and where this isn’t.

What is solid: Botox stops muscle contraction. In at least one well-documented case, this led to lymph accumulating in the face, and making the muscle work again cleared it. Visible eyelid swelling occurs in a measurable share of people treated for frown lines and crow’s feet.

What is missing: nobody has measured lymph clearance in the face after Botox. Nobody has measured fat cell size in people who get it regularly. Nobody has followed the skin of repeat Botox users to see whether collagen and elastin decline. The 30 to 50% figure is my estimate, not a measurement. The difference in blink force in Asian eyelids is my interpretation, not a published finding. The brittle skin is what aestheticians report, not a published study.

So this is a hypothesis built from a case report, a meta-analysis, what we know about bloated fat cells, and what people who work with skin every day are observing. I think it is a strong one. It needs testing.

Subscribe now

If you work on faces

Some of you are probably thinking, “Ivan, of course.”

Those of you who practice facial massage and gua sha already know that stagnant lymph means weak skin. You have felt it under your hands.

What I do not think has been discussed is the link to Botox. If this hypothesis is correct, the weak, difficult skin you see in regular Botox clients is not a coincidence. It is what happens when the muscles that move lymph are switched off, again and again.

Tell me if I’m wrong

This is new ground, so I would like to hear from you. If you are an aesthetician, what are you seeing in your Botox clients? If you get Botox yourself, have you noticed changes in your skin? And if you think I have the science wrong, I want to hear that too. Leave a comment below or reply to this email. I read every message.

Thank you for all of your support.

Leave a comment

Sources

1. Vasella M, Wolf S, Grünherz L, Kim BS, Lindenblatt N, Giovanoli P, Gousopoulos E. Evaluation of the effect of botulinum toxin A on the lymphatic endothelial cells. Aesthetic Plast Surg. 2024;48(21):4513–4522. doi:10.1007/s00266-024-04061-7.

2. Chang YS, Chang CC, Shen JH, Chen YT, Chan KKW. Nonallergic eyelid edema after botulinum toxin type A injection: case report and review of literature. Medicine (Baltimore). 2015;94(38). PMID: 26402825.

3. Brin MF, Boodhoo TI, Pogoda JM, James LM, Demos G, Terashima Y, Gu J, Eadie N, Bowen BL. Safety and tolerability of onabotulinumtoxinA in the treatment of facial lines: a meta-analysis of individual patient data from global clinical registration studies in 1678 participants. J Am Acad Dermatol. 2009;61(6):961–970.e11. doi:10.1016/j.jaad.2009.06.040.

4. Henninger AM, Eliasson B, Jenndahl LE, Hammarstedt A. Adipocyte hypertrophy, inflammation and fibrosis characterize subcutaneous adipose tissue of healthy, non-obese subjects predisposed to type 2 diabetes. PLoS One. 2014;9(8):e105262. doi:10.1371/journal.pone.0105262. PMID: 25148116.

The pattern she described is familiar. Someone in their late thirties or forties takes up running, cycling, rowing or a hard martial art. While the body gets leaner and looks younger, the face goes the other way: hollow cheeks, drawn features, a look friends describe as tired or older. The doctors she interviewed saw it most in endurance athletes, and mostly in the cheeks.

Exercise is great for the whole body, and it’s great for the skin too. But for the skin, it comes with an important side effect: the loss of superficial fat. That’s the layer just under the skin that gives the face its volume.

I’ve been thinking about gym face for years. I think it’s real. I also think the standard explanation is incomplete. What follows is my best attempt to explain how it happens, with an honest account of where the evidence is strong, where it’s thin, and where I’m guessing.

Subscribe now

The easy explanation

The usual answer is body fat. Train hard, get lean, and the face loses fat along with everything else. One doctor in the Times piece, Tracy Mountford, described it as an unfortunate trade-off of keeping body fat low.

That’s partly right. If you lose a lot of weight, some of it comes off your face. But it leaves two things unexplained.

First, why does the face look worse than the body? A lean, fit body reads as young. A lean face often reads as old. If this were only about total fat, you’d expect the two to move together.

Second, why doesn’t the face come back? One surgeon in the article noted that when people regain weight, the face is often the last place to show it. A man who cut back on jujitsu after an injury stayed gaunt anyway. If gym face were just empty fat cells, refilling them should be easy. It isn’t.

So here’s my hypothesis. Gym face has two halves. Training empties fat cells all over the body, and the face is unlikely to be the exception. And several things that come with intense training slow down how fast those cells get replaced, and some of them hit the face specifically. The first half explains the loss. The second explains why it doesn’t come back, and maybe why the face is hit harder than the body.

Suspect 1: the burn

I want to give the easy explanation its due, because it’s stronger than people make it sound. Exercise doesn’t just make you leaner in general. It actively empties fat cells. During and after a hard session, your fat tissue releases stored fat for your muscles to burn.

Part of the signal comes from the muscles themselves. Working muscle releases IL-6, and IL-6 helps drive fat release. In a 2019 trial in Copenhagen, adults with abdominal obesity cycled for 12 weeks. Exercise cut their deep belly fat by about 8%. In the group given a drug that blocks IL-6, that loss disappeared (Wedell-Neergaard 2019). Fat just under the skin joins in too. After an hour on a bike, the fat under the skin of the abdomen ramps up its own IL-6 output, in step with the fat it’s releasing (Lyngsø 2002).

So a hard trainer’s facial fat cells are very likely being emptied too, session after session. As I’ll get to, facial fat may give up its fat less readily than belly fat. And I can’t tell you how fast it happens in the face. Nobody has measured it there, and the trial was in belly fat, not faces.

But it can’t be the whole story. Emptied fat cells can refill. That’s what weight regain is. If gym face were only the burn, the face would fill back in when training eased. It often doesn’t. The rest of this post is about why.

Why the face may be different

Here’s the part I think explains why the face goes first. It’s also the most speculative part of this post.

In deep body fat, the number of fat cells is remarkably stable in adults. Radiocarbon dating shows about 10% of them are replaced each year. When people lose a lot of weight, their fat cells shrink, but there aren’t fewer of them (Spalding 2008). So when the weight comes back, the same cells refill. Body fat runs mostly on size.

Facial and dermal fat may run more on renewal. In mice, the fat layer in the skin turns out new, mature fat cells within 11 to 14 days under normal conditions, in step with the hair growth cycle. Deeper fat takes up to eight weeks, and only when the animal is overfed. The authors suggest skin fat is built for constant turnover, while the deeper depot is comparatively dormant (Rivera-Gonzalez 2025). In human cells, more precursors from the face turned into fat cells than precursors from the belly, and they kept that ability through 11 rounds of growth versus 5. The facial fat cells also released less fat when hit with an adrenaline-like drug, and had 60% less of the receptor that senses it (Chon 2015).

If that holds up, the face depends more on making new fat cells than on refilling old ones. Anything that slows regeneration would hit it harder than the belly. That would answer both questions from the top: why the face looks worse than the body, and why it doesn’t come back when the weight does.

I want to be clear about how thin this is. Nobody has measured fat cell turnover in human facial or dermal fat. The fast-renewal data are from mice. The facial cell data are from a dish. It’s a plausible idea that fits the pattern. It isn’t established.

Suspect 2: the bounce

Fat cells can feel mechanical force, and the kind of force matters. The best summary I know is a 2012 review by Naama Shoham and Amit Gefen at Tel Aviv University.

The consistent finding in cell culture: repeated, rhythmic loading suppresses the formation of new fat cells. Cyclic stretching blocked it in mouse preadipocytes (Tanabe 2004), in human umbilical cord progenitor cells (Turner 2008) and in mouse stem cells (Sen 2008). Vibration did the same in human adipose stromal cells (Tirkkonen 2011). Different labs found different signaling pathways doing the work. The direction was the same. And it held up in live animals: in mice, brief daily vibration, at less force than walking, cut new fat formation by about a quarter (Rubin 2007).

Two details stood out to me. In one study, vibration at strains below 0.001% still suppressed fat cell formation, and the number of loading sessions mattered more than their intensity (Sen 2011). And static stretch, a steady pull, seems to do the opposite and promote new fat cells (Hara 2011 in cells, Kato 2010 in mice). Fat cells appear to read the type of signal, not just whether there is one.

Now think about running. Every foot strike sends a jolt up through the body, and the soft tissue of the face moves with it. That’s thousands of cycles an hour, several times a week, for years. It’s the same idea behind what I call rockstar mouth: the areas of the face that move the most lose volume first.

Here’s what I can’t tell you. None of these studies used facial or dermal fat cells. Nobody has measured how much facial tissue actually strains during a run. And the review’s authors caution that fat loss in whole-body vibration studies in animals and people may come partly from muscles burning glucose, not from the fat cells sensing the vibration directly. So the cell and mouse data support the mechanism. They don’t prove it’s happening in your cheeks.

Suspect 3: heavy-session inflammation

Over months and years, exercise lowers inflammation. That’s well established, and it’s one reason exercise is so good for you. But a hard enough session can raise it briefly first.

The signal I care about is TNF-α. In the lab, it’s one of the most reliable blockers of new fat cell formation. And when I went looking at how exercise affects it, I found something I didn’t expect. It has a threshold.

After moderate cardio or short interval sessions, TNF-α in the blood doesn’t measurably move (Windsor 2018). After heavy lifting to failure, it rises about 30% and is back to baseline within a day (Townsend 2015). After a marathon, studies range from no change to more than double (Ostrowski 1999; de Sousa 2021; Ramirez 2025). After ultramarathons, some studies find rises of 20 to 70%, and others find nothing (Waśkiewicz 2025).

So it isn’t that every workout inflames you. Heavy, damaging or very long sessions cross a line that ordinary exercise doesn’t. If you do those most days, your fat cells spend more of the week exposed to a signal that tells them not to regenerate.

The gaps are real. These are blood levels, not levels in facial skin. The rise is modest and short. Nobody has measured fat cell regeneration in the skin after a heavy session. And it’s body-wide, so on its own it doesn’t explain why the face goes first. But if facial fat depends more on renewal, a body-wide brake on regeneration would show there first.

Suspect 4: the sun

The Times piece barely mentions it, but most of the athletes it describes train outdoors. A woman who quit running put it well: hours outside, pounding or pedalling, won’t leave you looking fresh-faced.

UV does reach the fat under the skin. In human skin, sun exposure draws immune cells into that fat and slows how much new fat it makes (Kim 2018). So for outdoor athletes, it’s a real contributor, and the most fixable one.

But I don’t think it’s the main driver. The jujitsu practitioner in the Times piece trained indoors, and plenty of people with gym face do their training under fluorescent lights. I’d treat the sun as something that makes gym face worse, not something that causes it.

What doesn’t fit, and what we don’t know

If bounce were the whole story, runners would get gym face and cyclists wouldn’t. But the doctors in the Times piece named cyclists and rowers too, and the jujitsu practitioner trained indoors. So bounce can’t be the only cause. Cyclists and rowers still burn fat the same way, and long rides and races cross the inflammation threshold. My guess is that the factors add up, and different athletes get there by different mixes.

Age is the other piece. The people in the article were mostly in their late thirties and up. By then, new fat cell formation in the skin has already slowed. That’s been measured directly in human cells (Caso 2013; Liu 2024). An older face has less capacity to absorb an extra hit.

What we don’t know is a long list. No one has studied gym face directly. No one has compared facial fat in runners, cyclists and swimmers at matched body fat. No one has measured fat cell turnover in the face of a hard-training athlete. Everything above is a hypothesis built from human trials, cell studies, animal studies and clinical observation. I think it’s a good one. It needs testing.

What you can do about it

First, keep exercising. Nothing here is an argument for doing less of something that’s so good for your health, and for your skin. It’s an argument for paying attention to how you do it.

If the hypothesis is right, these should help:

• If you train outdoors, protect your face. Train early or late, wear a brimmed hat or visor, use sunscreen. It’s the easiest change, even if it isn’t the whole story.

• Give your body gaps. Build in real recovery days, and don’t stack heavy sessions to failure back to back.

• Mix in lower-impact work. If you run a lot, some of your volume could come from cycling, swimming or the rowing machine.

• Eat enough protein. I recommend about 0.7 grams per pound of body weight. Regeneration needs raw material.

The Times piece was mostly about fillers and Sculptra. Those are real options, and for some people they’re the right ones. They add volume from outside rather than supporting the skin’s own fat cells.

If facial fat really does run on renewal, that’s the process worth supporting. It’s why we created Adipeau: to boost the skin’s natural capacity to regenerate fat cells. In our 3D clinical studies, subjects gained 1.4 cc of volume per side of the face over 6 months. The gains were highly selective to the areas of the face that were depleted.

In our registered clinical study (NCT04769193, published with the Karolinska Institute in 2026), facial volume measured by 3D imaging kept increasing between month 3 and month 6 of using Volume Cream. That study was not done in athletes. I don’t know how people with gym face specifically would respond, and I won’t pretend to.

What I can say is that facial fat cells keep regenerating your whole life. The process can be slowed down. It can also be supported. Gym face can get better, with our products and without them.

You can check out the Strength Gel here.

Tell me if I’m wrong

If you train hard and you’ve watched your face change, I want to hear from you. What do you do, how much, indoors or out? Did anything help? And if you think I’ve got the science wrong, I want to hear that too. Leave a comment below or reply to this email. I read every message.

Leave a comment

Sources

1. Bee P. Are you suffering from “gym face”? The Times, 2 December 2014.

2. Wedell-Neergaard AS, Lang Lehrskov L, Christensen RH, et al. Exercise-induced changes in visceral adipose tissue mass are regulated by IL-6 signaling: a randomized controlled trial. Cell Metab. 2019;29(4):844–855.e3.

3. Lyngsø D, Simonsen L, Bülow J. Interleukin-6 production in human subcutaneous abdominal adipose tissue: the effect of exercise. J Physiol. 2002;543(Pt 1):373–378.

4. Spalding KL, Arner E, Westermark PO, et al. Dynamics of fat cell turnover in humans. Nature. 2008;453(7196):783–787.

5. Rivera-Gonzalez GC, Butka EG, Gonzalez CE, et al. Comparative single-cell lineage tracing identifies distinct adipocyte precursor dynamics in skin and inguinal fat. Cell Stem Cell. 2025;32(8):1267–1284.e8.

6. Chon SH, Pappas A. Differentiation and characterization of human facial subcutaneous adipocytes. Adipocyte. 2015;4(1):13–21.

7. Shoham N, Gefen A. Mechanotransduction in adipocytes. J Biomech. 2012;45:1–8.

8. Tanabe Y, et al. J Cell Sci. 2004;117:3605–3614.

9. Turner NJ, et al. Biochem Biophys Res Commun. 2008;377:1147–1151.

10. Sen B, et al. Endocrinology. 2008;149:6065–6075.

11. Tirkkonen L, et al. J R Soc Interface. 2011;8:1736–1747.

12. Rubin CT, Capilla E, Luu YK, et al. Adipogenesis is inhibited by brief, daily exposure to high-frequency, extremely low-magnitude mechanical signals. Proc Natl Acad Sci USA. 2007;104(45):17879–17884.

13. Sen B, et al. J Biomech. 2011;44:593–599.

14. Hara Y, et al. Sci Signal. 2011;4:1–11.

15. Kato H, et al. Tissue Eng Part A. 2010;16:2029–2040.

16. Windsor MT, Bailey TG, Perissiou M, Meital L, Golledge J, Russell FD, Askew CD. Cytokine responses to acute exercise in healthy older adults: the effect of cardiorespiratory fitness. Front Physiol. 2018;9:203.

17. Townsend JR, Hoffman JR, Fragala MS, et al. TNF-α and TNFR1 responses to recovery therapies following acute resistance exercise. Front Physiol. 2015;6:48.

18. Ostrowski K, Rohde T, Asp S, Schjerling P, Pedersen BK. Pro- and anti-inflammatory cytokine balance in strenuous exercise in humans. J Physiol. 1999;515(Pt 1):287–291.

19. de Sousa CAZ, Sierra APR, Martínez Galán BS, et al. Time course and role of exercise-induced cytokines in muscle damage and repair after a marathon race. Front Physiol. 2021;12:752144.

20. Ramirez D, et al. Effects of marathon running on inflammatory cytokine levels: acute increases and 48-hour resolution. Abstract, National Conference on Undergraduate Research (NCUR), 2025.

21. Waśkiewicz Z, Mukhambet Z, Azerbayev D, Bondarev S. Inflammatory response to ultramarathon running: a review of IL-6, CRP, and TNF-α. Int J Mol Sci. 2025;26(13):6317.

22. Kim EJ, Kim YK, Kim S, Kim JE, Tian YD, Doh EJ, Lee DH, Chung JH. Adipochemokines induced by ultraviolet irradiation contribute to impaired fat metabolism in subcutaneous fat cells. Br J Dermatol. 2018. (Add volume and pages from PubMed.)

23. Caso G, et al. 2013. PMID 22999012.

24. Liu et al. 2024. PMID 38914551.

25. Taebnia et al. Comput Struct Biotechnol J. 2026. (NCT04769193)

It’s easy to get fooled by clinical studies. It just happened to me.

Recently, a social media persona posted a reel discussing a clinical study about the effect of kiwis on skin density.

Yes, kiwis. The headline was that a study in the Journal of Investigative Dermatology found that eating 2 kiwis a day increased skin density by 48% in 8 weeks. Yes, eating kiwis.

When I saw the Kiwi story pop up on my Instagram, I was primed. I had previously done a deep dive into how Vitamin C affects skin structure and function. Vitamin C was one of those major skincare ingredients that I – as a newcomer to skincare in 2017 – had to understand for myself in detail. While doing my research, I came across a review that casually mentioned a few things that surprised me and stuck in my head.

Section 3.1.2: “Dietary supplementation is therefore only expected to be effective in elevating skin vitamin C in individuals who have below-saturation plasma levels prior to intervention.”

Section 3.1.3: “if plasma levels are saturated, then it appears that topical application does not increase skin vitamin C content [51].”

This made sense to me. We all hear that our bodies excrete any excess vitamins that they don’t need. To be crude: “You just piss it out.”

Subscribe now

So when I saw the kiwi story on Instagram, I pulled up the study in the Journal of Investigative Dermatology. The authors were some of the same people who wrote the review paper. They would certainly discriminate subjects per baseline vitamin C status. They did. The extra kiwis did not increase vitamin C levels in those subjects who already had sufficient levels of vitamin C.

“Ha!” I said to myself. “Gotcha.” And without thinking twice, I commented on the social media personality’s post. “Skin density only increased in subjects who were deficient in vitamin c. In subjects who had normal vitamin c levels skin density did not increase.”

I was so pleased with myself. My comment got a bunch of likes.

It turns out I was completely wrong.

I got fooled by this sentence:

“Increased plasma ascorbate was associated with significant increases in skin density, from 0.154 ± 0.027 scanner units at baseline to 0.228 ± 0.035 after the intervention (P < .0001).”

So incredibly freaking sneaky.

This sentence implies that increases in vitamin C cause increases in skin density. They don’t. Skin density – measured by arbitrary ultrasound units increased in all subjects. And numerically the increase was greater in subjects who did not show an increase in plasma vitamin C.

It turns out that this study was just a poor study. No pre-specified endpoint. Arbitrary ultrasound units interpreted as increases in skin density went up. Collagen did not change. Elastin went down. Really, a nothing study.

So what can we learn from my being fooled.

You can’t go by the headlines. “Improved Human Skin Vitamin C Levels and Skin Function after Dietary Intake of Kiwifruit: A High-Vitamin-C Food.” False on its face.

You can’t go by the journal. The Journal of Investigative Dermatology is a serious publication. And they let this paper in, albeit as a letter to the editor. Watch out for those. They are the refuge of the unscrupulous.

You can’t go by the authors. The senior author of this paper is a very well respected scientist. Shame on you for perpetuating this fraud.

You have to subject all studies to strict scrutiny. First, ask whether the study was registered and set a primary endpoint. If it did not, you should be very careful about accepting any conclusions from the study.

This study was only partially registered and did not set a primary endpoint. Without the anchor of a primary endpoint, deception is possible. And given the sponsorship of the study, it seems that deception was always the point.

Leave a comment

Some people commented that they though we were inventing a new marketing term. “No such thing,” they said.

That’s when I knew that I had to create a place where we could really get down to it.

Present ideas. Get feedback, the more critical the better. And build.

I had once thought that YouTube or Instagram could house such discussions, but they don’t work.

Visitors to those platforms post drive by comments, with the expectation that no conversation will ensue.

Subscribe now

And very few people join in.

It feels somehow uncomfortable to comment on a comment.

Here I hope to bring out the scientist that lies within all of us.

To provide an outlet for your natural curiosity, even if you haven’t acquired the scientific vocabulary yet.

I was in your shoes 20+ years ago.

I had no formal scientific training. Not a whit. No biology. No chemistry.

I was a history major with an MBA working in business development for sanofi-aventis when they moved my job to Paris.

Graciously, unbelievably, they offered me a PHd level job in inflammation R&D.

I was encouraged and laughed it. I didn’t really mind the laughter. I preferred not being coddled.

I grew.

After a few years, i would present to the head of R&D, Elias Zerhouni, the former head of the NIH. It was the ultimate “fake it till you make it.” I made it.

After sanofi, I started a biotech company treating an ultra rare genetic disease. We couldn’t afford traditional regulatory advisors, so I had to oversee our pre-IND strategy myself. It worked.

Now as the founder of Adipeau, I publish in peer reviewed science journals and design protocols for clinical studies, learning by doing, as the Germans say .

I want this place to be a source of inspiration and support for you.

We will talk about dermal fat and our work at Adipeau. But we will also talk broadly about the science of skincare ingredients, and the rigor (or lack their of) of skincare science and clinical development.

You don’t need science vocabulary to participate. You can express yourself in plain English. I will respond in kind but in the process will introduce elements of scientific vocabulary to be more precise.

Science is essentiallly democratic. Sometimes the language gets in the way. We are going to work our way through it.

Joyfully,

Ivan