Sugar Sag: Is Your Diet Aging You?

[Regina] (0:00 - 0:08)
Absolutely. If I could just put a cream on my face and still have chocolate cake, right, breakfast, lunch, and dinner, then yeah, absolutely.

[Kristin] (0:09 - 0:09)
That's the dream.

[Regina] (0:09 - 0:10)
Yes. That is the dream. Yes.

[Kristin] (0:15 - 0:38)
Welcome to Normal Curves. This is a podcast for anyone who wants to learn about scientific studies and the statistics behind them. It's like a journal club, except we pick topics that are fun, relevant, and sometimes a little spicy.

We evaluate the evidence, and we also give you the tools that you need to evaluate scientific studies on your own. I'm Kristin Sainani. I'm a professor at Stanford University.

[Regina] (0:39 - 0:44)
And I'm Regina Nuzzo. I'm a professor at Gallaudet University and part-time lecturer at Stanford.

[Kristin] (0:44 - 0:50)
We are not medical doctors. We are PhDs, so nothing in this podcast should be construed as medical advice.

[Regina] (0:50 - 0:55)
Also, this podcast is separate from our day jobs at Stanford and Gallaudet University.

[Kristin] (0:55 - 1:00)
Regina, today we're going to talk about wrinkles, sags, and aged skin.

[Regina] (1:00 - 1:02)
Ooh, this is a scary one.

[Kristin] (1:02 - 1:47)
This is a little scary, yes. We're going to examine the claim that eating a lot of sugar and processed foods will make you look older.

[Regina]
Oh, I'm not going to like this one bit.

[Kristin]
No, this is one of these ones that we might not want to know. There's actually a number of things that we know cause wrinkles and skin aging.

[Regina]
Like aging.

[Kristin]
Yes.

[Regina]
Getting old. Is that big on the list?

[Kristin]
Right, yes. Getting old. Getting old is the main one.

Yes. But we also know that sun exposure and smoking, those two things definitely cause premature aging. It's more controversial, though, Regina, as to whether your diet, what you eat, can influence how old you look.

There's a concept in the literature coined sugar sag, and it's just what it sounds like.

[Regina] (1:48 - 1:52)
That is brilliant. Meaning I eat sugar and then my face sags.

[Kristin] (1:52 - 2:04)
That's the claim. It's not just sugar, though. It's other foods related to sugar that we are going to eventually get to in this podcast.

So don't think that if you don't have a sweet tooth, you get off scot-free here.

[Regina] (2:04 - 2:06)
So eating is also a risk factor.

[Kristin] (2:07 - 2:29)
Yes. Eating, yes. Although it's more specific than that.

We're going to talk about some specific foods that people think might contribute to an aged appearance. And along the way, we're going to talk about some statistical and methodologic topics, something I like to call references to nowhere, as well as we'll talk more about statistical adjustment and things like training and test sets.

[Regina] (2:29 - 2:31)
Oh, that sounds exciting.

[Kristin] (2:32 - 2:35)
It is exciting. Regina, I'm going to start by giving a little biochemistry background.

[Regina] (2:35 - 2:40)
I have not had a biology or chemistry in, like, 30 years.

[Kristin] (2:41 - 2:42)
OK. So you're going to need this refresher.

[Regina] (2:43 - 2:43)
We're going to need the refresher.

[Kristin] (2:44 - 3:00)
All right. Let's talk about proteins and sugars. Proteins.

Proteins are the workhorses of your body. They provide structure. They make up your bones, your skin, and your muscles.

And they also drive all the reactions in your body. And do you remember the structure of proteins at all, Regina?

[Regina] (3:00 - 3:07)
Only because when we talk about proteins that we eat, we talk about amino acids, essential amino acids. So I'm going to guess that.

[Kristin] (3:07 - 4:14)
They are made up of amino acids. They are chains of amino acids, and those chains fold up into specific structures that help to give the protein its function. Now let's talk about sugars.

So do you remember anything about sugars?

[Regina]
They're yummy.

[Kristin]
OK. Right. So carbohydrates are made up of chains of simple sugars, and we're going to be talking about those simple sugars today, things like glucose. And the interesting thing is simple sugars like glucose, they can actually bond with proteins.

I like to think of these sugars put on the surface of proteins as, like, decorations. They're pretty.

[Regina]
Like a scarf?

[Kristin]
Well, they're bonded through covalent bonds, so it's a little tighter bond than that.

[Regina]
Earrings.

Pierced earrings.

[Kristin]
Yeah, like something pinned to you. Yeah.

[Regina]
There you go.

[Kristin]
And these sugars are often added to the protein purposefully after the protein folds, and it helps the protein do its job. One of my favorite examples, HIV, that's the virus that causes AIDS, it actually covers itself with sugar molecules to form kind of a shield, and that helps it evade the immune system.

It's really good at evading the immune system.

[Regina] (4:14 - 4:21)
Like a disguise?

[Kristin]
Oh, like a disguise, yeah.

[Regina]
Like I put on fake hair and a fake mustache, and then no one notices me.

[Kristin] (4:21 - 4:54)
Exactly, yeah. So sugars can be bound purposefully to proteins, but they can also bond to proteins just through random spontaneous reactions. And this is because simple sugars like glucose, they're usually in a ring structure.

But once in a while, that ring structure can break open, and that can expose a reactive chemical group. And if that group happens to bump into the right amino acid on a protein at the right time, then they will just bond spontaneously. It's like you're getting random flyby decorations.

[Regina] (4:55 - 5:00)
Kind of like at the mall, how they would spritz me with perfume, whether I want it or not.

[Kristin] (5:01 - 5:07)
Yes, exactly. Yeah. But when this happens, this spontaneous bonding of the sugar on the proteins, that is called glycation.

[Regina] (5:07 - 5:11)
Glycation. Okay. Only one jargon term so far.

I can handle it.

[Kristin] (5:12 - 5:16)
Regina, let's take an example protein, hemoglobin. Do you remember what that is?

[Regina] (5:18 - 5:19)
Oxygen globules.

[Kristin] (5:19 - 5:19)
Yeah, exactly.

[Regina] (5:20 - 5:20)
That's right?

[Kristin] (5:20 - 5:41)
Yes. It's the protein that carries oxygen around to your tissues. And it floats around in your blood, and sometimes it runs into one of these open sugar rings and becomes glycated, decorated.

And the higher your blood sugar, the more sugars you have running around, the more likely this is to happen. And this glycated hemoglobin, it's something you might have heard of. It's called HbA1c.

[Regina] (5:43 - 5:51)
Oh, I have heard of that. That was just in my labs.

That I just got. In my blood work.

Used to see if you were on the road to diabetes, right?

[Kristin] (5:52 - 6:14)
Exactly. It tells you what percent of your hemoglobin has sugar attached. People with uncontrolled diabetes, their HbA1c is usually between 10 to 15%, meaning that about 10 to 15% of their hemoglobin has been decorated with sugar molecules.

In people without diabetes, it's only around 5%.

[Regina] (6:14 - 6:26)
Oh, is that what the number means that I see on my lab report?

[Kristin]
Yes.

[Regina]
Oh, that makes sense.

That's pretty cool, actually. Yeah. I also see it not just on my labs, but when I'm reading journal articles.

[Kristin] (6:27 - 6:37)
Right. It's a very common outcome in clinical trials because it's a really accurate way to monitor how well people with diabetes are controlling their blood sugar.

[Regina] (6:38 - 6:42)
So why is it so good? Why do we use that instead of just fasting glucose?

[Kristin] (6:42 - 7:11)
Well, hemoglobin sticks around in your blood for about three months. The turnover is about every three months. So it has time to pick up these sugar molecules, and it's actually a really good record of how much blood sugar you had running around for the entire three months.

So if you get a glucose test, that's easier to cheat. I could just be really good a few days before, but the glycated hemoglobin reflects what you've been doing for months.

[Regina] (7:11 - 7:16)
That is clever.

So is it bad to have this glycated hemoglobin? Does it like work less well?

[Kristin] (7:16 - 7:43)
It doesn't seem to affect your oxygen transport. So the glycated hemoglobin itself doesn't seem to be a bad player in the body.

[Regina]
Just a snitch.

[Kristin]
It's a snitch, yes. Yeah, it's telling on you if you ate too much chocolate cake in the last three months. It also tells us about other proteins in the body, because if your hemoglobin has a lot of sugar decorations on it, it's been highly glycated, this means that other proteins in your body are getting glycated at higher rates, too.

[Regina] (7:44 - 7:48)
Hmm. So what happens when you have the sugar sticking to these proteins? Is it harmful?

[Kristin] (7:49 - 8:17)
It isn't necessarily bad, but these glycated proteins are a precursor to something bad. They can undergo additional reactions like chemical rearrangements or bonding to other proteins or fats, and they can become something called an advanced glycation end product. These are called end products because they are not chemically reversible.

It's a whole class of molecules, and we've characterized about 40 different ones at this point.

[Regina] (8:17 - 8:25)
So the glycated protein itself isn't necessarily bad, what you're saying, it's just a precursor to these bad end products.

[Kristin] (8:25 - 8:25)
Exactly.

[Regina] (8:26 - 8:38)
So like a first date.

[Kristin]
Right. Yeah, it's reversible.

You just don't call them back. Exactly, yes.

[Regina]
But the further you go in the process, the harder it is…

[Kristin] (8:38 - 8:47)

to extricate yourself.

[Regina]
Yes, to get out of it.

[Kristin]
Yes. It eventually ends up in marriage, which is the equivalent of the advanced glycation end product.

Oh. Yeah, not reversible. They're not easily reversible.

[Regina] (8:47 - 8:50)
Yeah, that's kind of a pessimistic view of marriage.

[Kristin] (8:50 - 9:11)
Yeah, don't get me started on my thoughts. I think if you're a professional woman making as much as your partner or more, don't get married. That's my advice.

You might have noticed this, Regina, but advanced glycation end product has a nice abbreviation. A for advanced, G for glycation, and E for end product.

[Regina] (9:11 - 9:16)
Oh my goodness. A, G, E, age. Did they do this on purpose?

[Kristin] (9:16 - 9:33)
I don't know if they did that on purpose, but it's one of the few abbreviations in science that I actually enjoy, because A, G, E, of course, spells age, and these glycation end products have been linked to aging. They are not reversible, so they accumulate. They build up in your body as you age.

[Regina] (9:34 - 9:38)
Oh, and they also accumulate in your skin. This is the link to skin aging.

[Kristin] (9:38 - 9:47)
Exactly. One of the proteins that can become decorated with sugar and go on to form these end products, these A, G, E's, is the protein collagen.

[Regina] (9:47 - 9:52)
Collagen. Collagen. I do not have enough of it, apparently, in my skin, in my face.

[Kristin] (9:52 - 9:56)
So you probably know that collagen has something to do with skin aging.

[Regina] (9:57 - 10:03)
Elasticity, right? It affects how elastic and smooth and firm and beautiful your skin is, right?

[Kristin] (10:03 - 10:36)
Right. So let me describe collagen. It's made of these long chains of amino acids, and they're coiled, kind of like springs, and they're laid next to each other, and they work together to give tissues their shape and strength and their elasticity, right, this bounciness.

And collagen can become glycated, decorated with these sugar molecules, just like hemoglobin, but it's more of a worry with collagen because collagen, unlike hemoglobin, it doesn't turn over quickly. There was one study that found that certain types of collagen in your skin, their half-life is 15 years.

[Regina] (10:36 - 10:44)
Oh, my goodness. So it sticks around a long time. Oh, no.

So this is worse than a snitch. So I am running around with the consequences of choices I made 15 years ago.

[Kristin] (10:44 - 11:38)
These might still be influencing your skin potentially.

[Regina]
Oh, I'm screwed.

[Kristin]
Yeah.

You're all screwed. It's a problem because the glycated collagen sits around for years, so it has a lot more time to form AGEs, and one of the AGEs it forms is something called a crosslink. So if you have sugar sticking out from one of these strands of collagen, that sugar might bump into the strand of collagen next to it, the spring next to it, right, and it might spontaneously bond to that collagen strand.

[Regina]
Oh, cool.

[Kristin]
And when that happens, it forms kind of a sticky bridge or crosslink. And I like to think of it, we've got these nice coiled springs, all fresh and bouncy, and then you put some super glue between some of the springs in there, and they adhere to each other.

They stick to each other, and what it does is it makes it less bouncy. It makes your collagen stiff and tough.

[Regina] (11:38 - 11:41)
That doesn't sound pretty. It sounds like my face right now, actually.

[Kristin] (11:41 - 11:49)
Right. We want nice bouncy skin, and as we age, unfortunately, these crosslinks accumulate and our collagen stiffens.

[Regina] (11:49 - 11:52)
So it's not just about the beauty, though. Right.

[Kristin] (11:52 - 12:11)
Collagen doesn't just affect skin aging. It's also found in your connective tissues, like bones and tendons and blood vessel walls. So imagine what happens when those get stiff.

That could be bad. And so now that brings us to, Regina, I want to talk about this fascinating study that I found from 1975. Ooh, old school.

Sometimes classic studies are the best.

[Regina] (12:11 - 12:12)
I love that. Yes.

[Kristin] (12:13 - 12:23)
This, I think, was actually done for forensic purposes. So if, like, a dead body floated up in the river without an ID, they wanted to be able to find a way to say how old they were.

[Regina] (12:24 - 12:26)
Kind of like carbon dating, but for people.

[Kristin] (12:27 - 13:12)
Yeah, exactly. Yes. They recognized that collagen in the body, we just talked about how collagen can accumulate these AGEs, and collagen becomes stiffer as you age.

So collagen in the body records your age, kind of like tree rings, especially internal collagen that hasn't been exposed to things like the sun. So in this study, they took cadavers and they took out the tendon in their diaphragm that's made of collagen, and they put that with an enzyme that digests collagen.

[Regina]
Okay. Why?

[Kristin]
Because the longer it takes to digest that collagen, it's reflecting how stiff and tough and leathery that collagen is. And there is a super-tight correlation between the time it takes to digest that collagen and your actual chronological age.

[Regina] (13:12 - 13:25)
So if your collagen's really chewy, that means you're older and tougher. So I could use this to double-check the age that the guys put down on the dating profiles. Just ask for a sample of their diaphragm collagen.

[Kristin] (13:25 - 13:34)
Right, if you can get their collagen from their diaphragm, which I think they need to be dead. They're not a very good date. Maybe not.

[Regina] (13:34 - 13:35)
Maybe not. I need something else.

[Kristin] (13:35 - 13:47)
We're going to get another way to measure these kinds of things, but it's not as good. But hold that thought. And one of the things I love about this study, Regina, is they had a training set and a test set.

[Regina] (13:47 - 13:50)
Oh, my geeky heart. Oh, I just love it.

[Kristin] (13:50 - 14:12)
I know, right? This makes us so happy. So they took 25 cadavers where they knew the ages.

They ran from about 20 to 90, and they had all died of a trauma, like a car accident or an acute illness. So it wasn't people who had a chronic illness. And they measured the toughness of the collagen, the chewiness of the collagen, and they were able to show that there was a super tight relationship between the toughness of the collagen and their age.

[Regina] (14:13 - 14:14)
Oh, that's fascinating.

[Kristin] (14:14 - 14:22)
And that was their training set. So they used that training set to build a model. And the model turned out to be just a nice straight line relationship, beautiful straight line relationship.

[Regina] (14:23 - 14:23)
Linear relationship.

[Kristin] (14:23 - 14:50)
Linear relationships, and the dots are all really tightly clustered around that line. So they built this nice equation, and then they wanted to test if we apply it to other people, to new dead bodies, can we predict their ages correctly? Because of course, we want to test how good our model is.

Right. But you don't want to test the model on the same people you use to fit the model. You need to test how good the model is on a new set of observations.

[Regina] (14:50 - 15:08)
Otherwise, it's double dipping. The model might be latching on to random noise. I always think of it as seeing shapes in clouds, and they're not really there, but you think you see them.

So we want to let the test set check to see if that model actually learned something real.

[Kristin] (15:08 - 15:16)
Oh, that's a great analogy, Regina. Yeah. You want to know that the model is going to do well when a new body floats up in the river and you have to figure out that person's age.

[Regina] (15:17 - 15:32)
And this is important in modern statistics today, by the way, when you're dealing with big data. Really important to always have a training set that you use to build the model, but then a different test set to test the model.

[Kristin] (15:33 - 16:03)
Their test set was 17 new people, and some of them had actually died of chronic illnesses, so not just acute or trauma. They were aged 22 to 65, and they used this method to predict their ages. And it turns out for 12 of them, they were off by an average of just two years.

So it's really good carbon dating. That's impressive. But another thing I love about this paper is they gave all of the data.

It's not in an Excel spreadsheet. It's just in a table in the paper. But they said, you know, what the person died of and their predicted age and their actual age.

[Regina] (16:03 - 16:05)
I love old papers because they do this.

[Kristin] (16:06 - 16:16)
Yeah. So I can look at that and learn something else that they didn't even talk about. There were two people that they were predicted to be about a decade older than they actually were, and they both died of alcoholism.

[Regina] (16:16 - 16:20)
Well, that makes sense. That would kind of track with what we might expect.

[Kristin] (16:20 - 16:38)
Yeah. Alcoholism ages you. There were also three type one diabetics in the study, and these are the interesting ones. They probably had uncontrolled diabetes, two died of heart attacks, one of kidney failure, and they were in their 30s or early 40s.

Their predicted ages were 84 and 106 and 106.

[Regina] (16:38 - 16:53)
Whoa. Way off.

[Kristin]
Way off.

[Regina]
So their collagen was chewy, more like collagen from a really old person. And this was in diabetics, right?

[Kristin]
Yes.

[Regina]
And it's because of this glycated collagen that we're talking about.

[Kristin] (16:53 - 17:06)
Exactly. These three diabetics likely had a lot of sugar running around. And we believe that this led their collagen to form a lot of crosslinks and other AGEs, and this would cause their collagen then to be super leathery.

[Regina] (17:07 - 17:19)
So this brings us back to sugar sag. If their internal collagen is stiff and leathery, then the collagen in their skin might be too.

[Kristin]
That's right.

[Regina]
Did they look like they were 106 years old?

[Kristin] (17:20 - 17:54)
No. And the authors, again, a nice thing about this paper, the authors talked about this. They said specifically that they did not look like they were elderly centenarians. They noted that in the paper.

So this study tells us that sugar can dramatically affect collagen, but we haven't yet tied that to an aged appearance.

[Regina]
Fascinating.

[Kristin]
Fascinating study, right?

Regina, I want to make a little statistical side note here. Because in this case, we're talking about three diabetics, which is a very small sample.

[Regina] (17:54 - 17:54)
Barely even a sample.

[Kristin] (17:54 - 18:05)
Right. But it's compelling. So I just think it's important since we're going to talk in this podcast a lot about why we hate small samples. I think it's important that we point out why we find this small sample compelling.

[Regina] (18:05 - 18:10)
We are talking about huge effect here in a well-designed study.

[Kristin] (18:10 - 18:12)
Yes, exactly. This is not subtle.

[Regina] (18:12 - 18:13)
No, no.

[Kristin] (18:13 - 18:16)
This is 60 years between the predicted age and the actual age.

[Regina] (18:16 - 18:27)
Right. You almost don't need statistics. I saw one researcher describe it as the intraocular test, meaning it's so big it hits you between the eyes and you don't even need a statistical test.

[Kristin] (18:28 - 18:48)
Exactly. So, Regina, diabetics are this great model system of the effects of sugar in the body. And a lot of studies have been done in both type 1 and type 2 diabetes since this 1975 study. But it's not just people with diabetes we are worried about, right?

People without diabetes also accumulate AGEs as they age, just at a slower rate.

[Regina] (18:49 - 19:00)
So this leads to the idea that if I eat more sugar, even though I am not diabetic, this could still lead to chewier collagen and my skin just falling apart faster.

[Kristin] (19:01 - 19:09)
Exactly. That's the idea of sugar sag. So fast forward a little bit.

There are now a bunch of ways that scientists can measure AGEs that are much easier than cutting people up.

[Regina] (19:10 - 19:12)
No dead bodies.

[Kristin] (19:13 - 19:36)
No dead bodies. Exactly. One of the first ways they discovered for measuring these is kind of interesting. So some of the AGEs have a brown or yellow color.

[Regina]
Ew.

[Kristin]
Yeah. Did you know, actually, AGEs were originally discovered in the beginning of the 20th century in cooking.

[Regina]
Oh, really?

[Kristin]
So when you brown chicken or you make toast that brown color, you are creating AGEs.

[Regina] (19:37 - 19:37)
No.

[Kristin] (19:37 - 19:43)
Yep.

[Regina]
The yummy part.

[Kristin]
Yep.

We eat them all the time.

[Regina]
Oh, no.

[Kristin]
Hold that thought, because we're going to get back to that in this episode.

[Regina] (19:43 - 19:46)
So good foreshadowing. I'm going to stick around for that one. Thank you.

[Kristin] (19:47 - 20:03)
Right. So in addition to this brown or yellow color, they also emit fluorescence. And so you can measure them in the skin non-invasively by using this special tool that measures fluorescence coming out of the skin.

It's called a fluorescence spectroscope.

[Regina]
So they're glowy.

[Kristin]
Yes.

[Regina] (20:03 - 20:18)
So can I get this special scope, fluoroscope, or whatever, and use this on my dates? It sounds a lot easier than taking out their diaphragm. But I want to make sure they're not lying about their age, because sometimes I got to say I've got my suspicions.

[Kristin] (20:19 - 20:41)
Regina, are you looking for dating carbon dating?

[Regina]
Carbon dating, dating, yes.

[Kristin]
So Regina, you're going to love this.

There actually is a device called an AGE reader. It's about the size of a shoebox, and all you have to do is rest your forearm on it for 12 seconds, and it shines a light on your skin to measure fluorescence, and it gives an AGE reading.

[Regina] (20:41 - 20:43)
Can I buy this on Amazon?

[Kristin] (20:44 - 20:58)
Well, it's used in research, so you could buy it, but it's probably pretty pricey. And the relationship between this measurement and your chronological age, it's not a tight relationship the same way the chewiness of the collagen was. So it's probably not going to be super accurate.

[Regina] (20:59 - 21:06)
That's okay. I'm fine with it. I got my credit card out.

All right. All right. So Kristin, scientists know how to measure these things.

[Kristin] (21:06 - 21:36)
Right, exactly.

And so now lots of studies have been done, because we don't have to cut up dead bodies, and they have linked high AGEs with bad things. So for people with diabetes, if they have high AGEs, that's been correlated with having more diabetic complications, like more heart disease, higher blood pressure. And Regina, you're going to love this one.

I am going to get the sex into this episode, because higher AGEs have also been linked to more erectile dysfunction.

[Regina] (21:38 - 22:00)
Great job bringing the sex in there. And Kristin, I'm so proud of you for saying the word erectile dysfunction.

Okay. This does make sense, because you want penises to be spongy. You want penises supple and juicy, right?

Did they measure the AGEs in the arm or the penis?

[Kristin] (22:01 - 22:06)
Oh, it was just on the arm. So there were multiple studies, and those studies actually all used the AGE reader.

[Regina] (22:07 - 22:20)
Oh, I want one. But forget about using it for aging. Now that you've told me this secret, can I use it on a date to get a preview of what to expect later on?

[Kristin] (22:20 - 22:32)
Well, yeah, in theory. I mean, I don't know how tight the correlation is, plus the AGE reader is kind of bulky, so I'm not quite sure how you're going to get the men to put their arm in the thing. How are you going to explain that to them?

[Regina] (22:32 - 22:39)
That's okay. We can do some product development on this, because I think we can come up with a pocket spectrometer, like a wand.

[Kristin] (22:39 - 22:46)
Oh, like a wand that you just whip out of your purse and run it over their arm really quickly to figure out what the night's going to look like.

[Regina] (22:46 - 22:58)
Oh, we can make a lot of money. Especially women of a certain age, right? Dating men of a certain age.

[Kristin]
Oh, absolutely.

[Regina]
Did we just copyright it by describing it? I think so.

[Kristin] (22:59 - 23:00)
We did, Regina. It's ours now.

[Regina] (23:00 - 23:00)
All right.

[Kristin] (23:02 - 23:48)
All right, back to sugar sag, away from penises. Right. So AGEs have been linked to bad things in people without diabetes, too.

So elderly people who have more AGEs in their blood, they also tend to have more chronic diseases. Makes sense. Right.

There's a lot of biological reasons to think that these AGEs might actually be doing damage in the body. It's not a stretch to think that stiff collagen might be bad. Your arteries and your blood vessels, if they are stiffer, that could lead to more heart disease.

And also, there's tons of other proteins in the body that can form these AGEs. Another protein that helps keep skin supple is the protein elastin. If sugar damage is elastin, this could also age your skin.

It's biologically plausible.

[Regina] (23:49 - 23:55)
It is plausible, unlike some studies that we talk about where the biology is not quite worked out so well.

[Kristin] (23:55 - 24:08)
But as we also talk about, Regina, correlation is not causation, and these are all correlation studies. AGEs definitely accumulate with age. They correlate with bad things, but they might just be an innocent bystander.

[Regina]
Could be a marker.

[Kristin]
A marker, yes.

[Regina] (24:08 - 24:09)
Not actually doing damage.

[Kristin] (24:09 - 24:38)
Yes. And remember, in that cadaver study, internally, that collagen looked 60 years above their actual AGEs, but they didn't look on the outside like they were 106.

[Regina]
But that's what we're really interested in today.

[Kristin]
That's what we're interested in, is how did they look, because that's most important, right? So, yeah. We now are going to narrow in on the actual claim for today.

This is the background biology, but does it make you look older?

[Regina]
And we do have, you have studies to talk about?

[Kristin]
Right. You need studies that actually measure how old do you look. So we have to talk about those next.

[Regina] (24:38 - 24:57)
Oh, great. So, Cliffhanger, let's do a short break first.

Kristin, we've talked about your course, Writing in the Sciences, on Coursera.

Maybe you could tell people a little bit more about it.

[Kristin] (24:57 - 25:13)
It's a self-paced course for anyone who needs to write scientific papers, and I give a lot of practical demonstrations for how to improve your writing to make it much more clear and concise. And you can earn a certificate from Coursera. You can find a link to that course on our website, NormalCurves.com.

[Regina] (25:18 - 25:32)
Welcome back to Normal Curves. Today we're looking at the claim that eating a lot of sugar in processed food makes you look older. So, Kristin, where are we?

[Kristin] (25:32 - 26:04)
We've learned from studies in diabetes that when you have high sugar levels in your blood, your body produces more of these advanced glycation end products, or AGEs for short. These AGEs also build up in your body as you age, even if you don't have diabetes, and they impact collagen, making it stiffer. They also affect other proteins that help keep skin supple and smooth.

So, it's not too far-fetched to think that if you're eating a high-sugar diet, even if you don't have diabetes, you might be raising your AGE levels, which could lead your skin to look aged.

[Regina]
So, this is the sugar sag.

[Kristin] (26:04 - 26:11)
Yes.

[Regina]
This is what we're talking about. So, it's plausible.

Yes. But what happens, like, is there evidence, direct evidence?

[Kristin] (26:11 - 26:33)
Great question, Regina. As it turns out, there aren't many studies with that direct link. We're going to discuss two studies in just a bit, but I'm going to warn you right now, they aren't exactly earth-shattering.

[Regina]
Oh, no.

[Kristin]
I want to talk for a minute, though, about the review articles that I came across. Because when I tried to get into this topic, the first thing I did was pull a few recent review articles, hoping they would point me to that direct evidence. And I was a little disappointed.

[Regina] (26:33 - 26:34)
Oh, no. They're not good?

[Kristin] (26:35 - 26:52)
They were very thin on any actual evidence that if I eat too much sugar, my skin is going to be affected. As we're going to see, a lot of what I found was, like, a review article citing a previous review article, which in turn cited another review article. So, it was very heavy on narrative and short on hard evidence.

[Regina] (26:53 - 27:05)
I'm picturing, like, those gag birthday gifts where you get this big box and it's wrapped and you unwrap it and then there's a smaller box and you unwrap it and you keep going to the end. And then it's an empty box.

[Kristin] (27:06 - 27:30)
That is a perfect analogy, as we're going to see, Regina. I ended up with an empty box, pretty much, I think. And, Regina, I think it's worth talking, actually, about two specific issues that I noticed in the review articles.

First off, I think they're doing a bit of marketing. This claim has some pretty great branding. I mean, using A-G-E for Advanced Glycation End Product?

Age! I mean, that just makes you think, yes, of course they age you.

[Regina] (27:30 - 27:34)
It's kind of brilliant. It's almost like some marketing people came up with this directly. Sugar Sag.

[Kristin] (27:34 - 27:52)
That is so clever and catchy, right? I know. And Sugar Sag has some serious profit potential, whether it's real or not.

If people believe it's real, it creates a market because skin care companies can say, hey, I've got something to fight this glycation. And it sounds really scientific. And that's going to get people to buy their products.

[Regina] (27:53 - 28:01)
Absolutely. If I could just put a cream on my face and still have chocolate cake, right, breakfast, lunch, and dinner, then yeah, absolutely.

[Kristin] (28:01 - 28:02)
That's the dream. Yes.

[Regina] (28:02 - 28:03)
That is the dream. Yes.

[Kristin] (28:04 - 28:21)
So there's some worry that maybe these review articles are a little bit more about marketing than about science. Just to take an example of one of the review articles that has a single author. Just for fun, I'm going to read you some of her conflicts of interest.

She has about 80 total. I'm just going to read you the A's and the B's.

[Regina]
80? Did you say eight zero?

[Kristin]
Yes. 80.

[Regina] (28:21 - 28:22)
Oh my goodness.

[Kristin] (28:22 - 28:30)
Aclaris Therapeutics, Aesthetics Biomedical, Allergan, Almirall, Amderma, Amgen, AO Biome, Arcutis Biotherapeutics.

[Regina] (28:30 - 28:31)
Wait a minute. She's funded by all of these?

[Kristin] (28:31 - 29:01)
She's getting money somehow. Maybe she's just getting paid for a random talk.

We don't know. But she is getting some money from all of these.

[Regina]
And this is one person?

[Kristin]
One author on this review article. Okay, let's get to the B's. Okay.

Bausch Health, Bayer, Biopelle, Boehringer Ingelheim Pharmaceuticals, Boots, Botanics Pharmaceuticals, Brickell Biotech, Burt's Bees. That's just the A's and the B's. And let me highlight a few others you might recognize.

Estée Lauder, L'Oreal, General Mills, GlaxoSmithKline, NuSkin, Perricone, Pfizer, Procter & Gamble. It goes on and on.

[Regina] (29:01 - 29:03)
We do not have enough conflicts of interest.

[Kristin] (29:04 - 29:07)
We clearly do not have enough conflicts of interest.

We are missing out here, Regina. We're doing something wrong.

[Regina] (29:07 - 29:12)
We are missing out.

Yeah. How did we get that?

[Kristin] (29:12 - 29:19)
I know. Not only that, but she received an honorarium just for writing the manuscript. I want to get paid to write my scientific manuscripts, you know?

[Regina] (29:19 - 29:21)
Wow. Yeah.

[Kristin] (29:21 - 29:33)
And she is supported by a grant from SkinCeuticals. And, Regina, here's the kicker. SkinCeuticals actually makes a product that they claim blocks glycation, blocks the formation of these AGEs.

[Regina] (29:33 - 29:44)
I think I've seen that one, actually. It's like an age interrupter or something like that.

[Kristin]
It's called Age Interrupter Advanced.

[Regina]
Yes. Oh, maybe I should try this. Maybe they'll send us a free sample.

[Kristin] (29:45 - 30:04)
Well, it might depend on how this episode wraps up, whether they may not like our take on it. But, Regina, it costs $185 a bottle. I was kind of blown away by that.

And the author of that review article that we were just talking about, she was heavily involved in the development of this skin cream.

[Regina] (30:04 - 30:11)
Did she do some of the so-called clinical trials where they have like nine people and no randomization? No control.

[Kristin] (30:12 - 30:26)
Twenty. Yeah. And she's the one rating whether or not their wrinkles improved or not. Yes, exactly.

[Regina]
We should do a future episode about these tiny clinical trials.

[Kristin]
These kinds of clinical trials that skin care companies sometimes run, that would be a really interesting episode.

[Regina] (30:26 - 30:26)
Yeah.

[Kristin] (30:27 - 30:27)
Let's save that.

[Regina] (30:27 - 30:32)
So, this author then is potentially making money off of this whole sugar sag narrative.

[Kristin] (30:33 - 30:42)
Oh, she is definitely making money off of the sugar sag narrative. Yes. So, there may be some incentive for her to make the narrative a little tighter than it actually is.

Yeah.

[Regina] (30:43 - 30:50)
So, we need to keep all of this in the back of our minds when we're reading these types of review papers. There might be something to be gained here by some of the authors.

[Kristin] (30:51 - 31:05)
The second thing I want to talk about with regards to these review articles, though, Regina, is something that drives me absolutely bonkers. And if you'll permit me, I want to go on a little statistical detour, but it's not about how we calculate statistics. It's about how we cite them.

Oh. So, let’s call it statistics adjacent.

[Regina] (31:05 - 31:10)
Uh-huh. So, is it a little bit of sleuthing?

[Kristin]
It's citation sleuthing.

[Regina]
Oh, citation sleuthing. Yes. I love that.

[Kristin] (31:10 - 31:15)
It's something I call references to nowhere, which I rant about in my Coursera course, Writing in the Sciences.

[Regina] (31:16 - 31:17)
Which is a very good course, by the way.

[Kristin] (31:17 - 31:18)
Oh, thank you, Regina.

[Regina] (31:18 - 31:19)
Put in a word for that.

[Kristin] (31:19 - 31:47)
All right. I was reading a review article from 2023. This was in the International Journal of Molecular Science, different author.

I came across a sentence that talked about an intervention study where they claimed to have reduced glycated collagen in people by 25 percent.

[Regina]
Ooh. That sounds like it could be good.

[Kristin]
Right. I wanted to find the original study because it's the kind of evidence that I was looking for. But, trying to find this study, Regina, it sent me down a rabbit hole of citations.

[Regina] (31:48 - 31:50)
Are you going to take us down that rabbit hole with you?

[Kristin] (31:52 - 32:06)
Oh, yes. All right. I'm now going to read you verbatim the sentence from the 2023 review article. Just a warning, they use the term glycosylated rather than glycated, and I'm going to read the second half of the sentence, even though it's not related to the intervention study.

You're going to see why in a minute.

[Regina] (32:06 - 32:07)
Okay.

[Kristin] (32:07 - 33:52)
Listen carefully to the words. Here it is. Interestingly, strict control of blood sugar for four months reduced the production of glycosylated collagen by 25 percent, and low-sugar food prepared by boiling could also reduce the production of AGEs.

They cite for that sentence a 2020 article, so of course I went and pulled that 2020 article hoping to find the study behind that 25 percent statistic. But it was just another review paper.

[Regina]
Oh, no.

[Kristin]
It wasn't original research. Okay. And let me read you the sentence I found in that 2020 study.

See if you notice anything familiar, Regina. However, strict control of blood sugar for four months can reduce the production of glycosylated collagen by 25 percent, and low-sugar food prepared by boiling can also reduce the production of AGEs.

[Regina]
It's like the same sentence, right?

[Kristin]
It's exactly the same sentence. I think they changed however to interestingly. That is not a plagiarism get out a free card.

[Regina]
One adverb does not originality make.

[Kristin]
Yes, exactly.

Lifting the entire sentence and just changing one or two words is plagiarism, and it's not a gray area. The 2023 article plagiarized the 2020 article. But now, Regina, continuing along with my sleuthing, the 2020 article gave three citations for this 25 percent statistic.

[Regina]
Sounds promising.

[Kristin]
Sounds promising, right? But guess what?

They were all review articles, so I still haven't found the study. They were from 2015, 2013, and 2010. All three review articles had the same sentence: strict glycemic control can reduce glycated collagen by 25 percent in four months.

So, in other words, the 2015 and 2013 articles plagiarized from the 2010 article. That 2015 article does give two citations, but those turn out to be just the 2013 and 2010 review articles.

[Regina] (33:53 - 33:56)
So, you're down to like the fourth empty box now, gift wrap box.

[Kristin] (33:56 - 34:11)
Yes, exactly. It's fun unwrapping all these presents, I guess, but yeah, I'm coming up empty. And then, guess what, Regina, the 2013 and 2010 articles give no citation at all.

The sentence just appears in both of those articles with no citation.

[Regina] (34:11 - 34:17)
No citation?

How could they do that? You can't do that.

[Kristin]
It just states this as fact.

[Regina]
So, this is the empty box. You are down now to the empty box.

[Kristin] (34:17 - 34:27)
I am down to the empty box, yes, exactly. So this is why I call these references to nowhere.

[Regina]
To nowhere.

[Kristin]
You know, I'm at a dead end, so I thought I would go to Google for help.

[Regina] (34:27 - 34:27)
Of course.

[Kristin] (34:27 - 34:31)
I was able to Google around until I actually found the underlying paper.

[Regina] (34:31 - 34:32)
Wow. That's impressive.

[Kristin] (34:33 - 34:58)
I know. It was a paper from 1991. Quite old.

Yeah. It was a study in 14 people. These were all diabetics with poor sugar control, and they were put on a strict diet.

There was no control group or anything, but they did get their blood sugar down quite a bit, and they reduced the amount of glycated collagen in their skin.

[Regina]
That's pretty cool.

[Kristin]
It's a good study, but that 25% statistic that people were citing, the citing authors got it wrong.

[Regina] (34:59 - 34:59)
Oh, no.

[Kristin] (35:00 - 35:04)
They made it sound like there was a 25% reduction in the total amount of glycated collagen in the skin.

[Regina] (35:05 - 35:06)
But that's not what happened?

[Kristin] (35:06 - 35:19)
No. In the original paper, the authors reported a 25% reduction toward the normal mean. So, the diabetics start with higher levels than normal, they reduce the extra amount by 25%, not the total amount.

[Regina] (35:19 - 35:23)
Hmm. So, they just reduced the icing on the cake, not the cake itself.

[Kristin] (35:24 - 35:40)
That's a really good analogy. Yeah. 25% of the icing went, not 25% of the total cake.

Yes. So, that 25% is actually an exaggeration, but this got lost in translation because the authors were citing review articles and not going back to that original source to see what that actual statistic was.

[Regina] (35:40 - 35:45)
It sounds like a game of telephone. Things get all dropped and garbled along the way.

[Kristin] (35:45 - 36:24)
That's a perfect analogy. Yes. And that's what happened here.

We lost some of the original message. And sometimes, Regina, it's even worse than this, the message that comes out at the end of the chain has no resemblance whatsoever to what started, what was in the original source. So, the take-home message here, one, you should never plagiarize.

That's just bad for many reasons. And two, always go back to the original source. You should never trust that someone else's summary of an underlying study is correct.

That is a bad bet. I would venture to say that at least half of the citations in the scientific literature are wrong in some way. So, if you just copy someone else's summary, you are probably going to get it wrong, too.

[Regina] (36:25 - 36:29)
I think this is probably surprising to a lot of people. It's because there's no fact-checking.

[Kristin] (36:29 - 36:35)
It's not like when we write for newspapers and magazines, Regina, that gets fact-checked very, very carefully.

[Regina] (36:35 - 36:46)
Very carefully. Every tiny little bit, they look into it. But strangely enough, scientific articles in the published literature, not fact-checked. So weird.

[Kristin] (36:47 - 36:56)
Peer reviewers don't have time to go through and fact-check. There's just too many references.

When I'm doing peer review, I don't have time to check everybody's citations. No, I'm checking their statistics. Did they do the study right?

[Regina] (36:56 - 37:05)
You know, journals should offer a bounty. You know how sometimes they offer like a bounty for finding flaws in code or whatever, like a hacking. They should offer a little.

[Kristin] (37:05 - 37:14)
I think the literature is already too polluted. It's like global warming. We can't really do much to fix the ones that are already there, but maybe everybody going forward can do better.

[Regina] (37:14 - 37:14)
Going forward.

[Kristin] (37:15 - 37:40)
Going forward. Regina, that was a little aside about how to do good papers, but let's bring it back now to the actual evidence on sugar and aged skin.

[Regina]
This is the direct evidence.

[Kristin]
Yeah, the two studies that I found that gave some evidence of this link. First study, it's in an MDPI journal.

[Regina]
Oh no.

[Kristin]
Yeah. It's not that every article published in an MDPI journal is bad, but there's a lot of variability.

[Regina] (37:41 - 37:42)
Very high variability.

[Kristin] (37:43 - 38:04)
This was a 2023 study from Brazil. They looked at 28 women aged 39 to 55, and there were 14 people without diabetes and 14 people with diabetes. So we're still using our model system of diabetes because it's just the best chance we have of seeing.

Now they did some cool measurements of their skin. Not that I want these measurements done on me. I do not.

[Regina] (38:04 - 38:07)
No? Oh, like measuring wrinkles?

[Kristin]
Oh yes.

[Regina]
Oh my.

[Kristin] (38:07 - 38:08)
Yes.

[Regina] (38:08 - 38:10)
How do you measure it? What are they doing?

[Kristin] (38:10 - 38:20)
Okay, so it's a high resolution photograph, and then it goes through an actual program that's been written to calculate the depth, area, and volume of your wrinkles. Don't you want that done on you, Regina?

[Regina] (38:20 - 38:25)
Oh my goodness. I do not. I mean, part of me wants the quantification, but- I do not want to know.

[Kristin] (38:25 - 39:06)
Right? No. So they did that.

They also measured the elasticity in their skin. It's a special instrument. You take a little skin biopsy.

You can measure how bouncy the skin is. And what they found is on a number of these metrics, not all, but on many of them, the diabetic group had more aged appearance than the non-diabetic group. Like more wrinkles around the eyes and around the mouth, which sounds like, hey, there's the evidence.

Because again, our diabetics, that's our model system of high sugar. Here's the problem, Regina.

[Regina]
What?

[Kristin]
They never tell us the ages of the diabetics versus the non-diabetics. Maybe they were the same age. But there's no information in the paper on the actual ages.

And 39 to 55 is a rather wide age range.

[Regina] (39:06 - 39:08)
I looked much better when I was 39.

[Kristin] (39:08 - 39:16)
I want to look like I looked when I was 39, yes. So we don't know. Maybe all the diabetics were age 53 on average, and the non-diabetic group was 43 on average.

[Regina] (39:16 - 39:21)
The biggest thing, if you're talking about them looking older, first of all, you have to know if they are older.

[Kristin] (39:21 - 40:16)
Yes, exactly. Exactly, Regina. So people cite this as evidence.

Unfortunately, it's just a big question mark for me. It's not proof of anything. And, you know, there are a lot of other red flags in the paper that indicate that it's not of the highest quality.

I'll put some of them in the show notes, but I'm just going to mention one thing because we've been talking today about references, and I noticed a bad reference. I came across this sentence. This type of alteration has already been reported in the diabetic population where altered skin elasticity was detected.

So this sounded to me like a study about skin elasticity in people with diabetes.

[Regina]
Right. The direct evidence that we're looking for.

[Kristin]
Exactly. So I pulled the paper expecting something in humans. It was a cell culture study with no mention of people with diabetes.

[Regina]
No way. Come on.

[Kristin]
So this is another reference to nowhere because the statement in the paper is totally divorced from what is actually in the cited source.

[Regina] (40:17 - 40:22)
Okay. So that's study number one. Not much there, to say the least.

What about study two?

[Kristin] (40:22 - 40:34)
All right. Second study from the Netherlands, 2013. It's about 600 people, both women and men.

This is mostly in non-diabetics, which is the group we're interested in. They did a one-time non-fasted glucose measurement.

[Regina] (40:34 - 40:39)
Wait, non-fasted?

[Kristin]
Non-fasted glucose one time.

[Regina]
So if I had some cookies before I showed up?

[Kristin] (40:39 - 40:49)
So, it's really not a great measure of sugar. That's their metric of sugar. Not a great one.

They also want to link that to perceived age. So they measured perceived age, which is kind of a fascinating variable.

[Regina] (40:49 - 40:52)
It is. It says who? Perceived by whom?

[Kristin] (40:52 - 41:14)
Oh, great question.

So here's how this measurement works. The subjects come into the study, no makeup, no hairstyling. They pull their hair back out of their face.

This is a little scary. There's a high-resolution photo taken right up close and personal. I don't want this.

And then that photo is shown to 60 independent assessors, and each of those assessors looks at the photo and they guess the age.

[Regina] (41:14 - 41:24)
So they asked other people to look at this microscopic close-up of your face and say how old do you look? Ooh. Yeah.

Again, online dating kind of reminds me of that.

[Kristin] (41:24 - 42:13)
It's a little like online dating, I guess. It's 60 different raters, and then they take the average of all the raters. It's kind of crowdsourcing.

Participants were in their early 60s. They had almost 600 non-diabetics and just 33 diabetics. They divided the non-diabetics into three groups according to their blood sugar, low, medium, and high.

The low-glucose group was a little younger than the high-glucose group, and they were a little younger than the diabetic group. But thankfully, the authors actually did account for age in this study on a logical age.

[Regina]
Okay. Good.

[Kristin]
Right. They get bonus points, although the bar is low here.

[Regina]
The bar is very low. Yes. So the low-glucose group appeared on average 3.8 years younger than their actual age, which sounds good, right?

Except the high-glucose group appeared 2.6 years younger than their actual age. So it wasn't quite as good, but it wasn't that different.

[Regina] (42:13 - 42:16)
Everyone looks younger than what they are. Who are these nice raters?

[Kristin] (42:16 - 43:26)
I want these raters, you know, rating my age, I guess. But here's the kicker, Regina. The people with diabetes, they appeared 4.6 years younger on average than their actual age.

[Regina]
Whoa. They appeared younger?

[Kristin]
Yeah.

They appeared more young for their actual age than the other groups. So you might look at that and say, wait, what? This is getting cited as evidence of sugar sag, but they look younger?

[Regina]
Hmm. Did they look at weight?

[Kristin]
Okay, good, Regina. Yes, you've got it. Yes, weight is really critical here.

The diabetic group was much heavier than the other groups. And the high-glucose group was also a little heavier than the low-glucose group. So they recognized that they needed to adjust for weight, because a little aside here, Regina, even though everybody talks about how being heavy is bad for you, in fact, if you are over 40, for women, if you are carrying more weight, you actually look younger.

[Regina]
Oh, yeah. It fills out all the wrinkles and the sags.

[Kristin]
Right.

You're filling all that sagging in with volume from fat. Interestingly, before the age of 40, thinner women actually look younger than heavier women. It's something about their cheekbones sticking out more.

[Regina] (43:27 - 43:37)
Hmm, I can see that. But Kristin, you and I are out of that age zone where it's good to be thin. So pardon me, be right back, going to gain some weight right now.

[Kristin] (43:38 - 44:29)
We should both be eating more chocolate cake, right? Fill out the volume. So the diabetics actually look younger for their age than the other groups, but they're much heavier.

So the authors needed to account for weight. And what do you think they did to try to account for that, Regina?

[Regina]
Statistical adjustment, maybe?

[Kristin]
They did statistical adjustment. They adjusted for BMI and chronological age and a few other things. So what they're doing actually is now answering a very narrow question.

The question is, do people with higher blood sugar look older for their age and weight? The model essentially tries to create an imaginary world where all four groups have the same average chronological age and weight. It's estimating how old the diabetic group would look if they were a little younger and thinner and how old the low-glucose group would look if they were a little older and heavier.

[Regina] (44:30 - 44:38)
So it reminds me of filters, you know, in Snapchat, where you can see what you would look like as an old lady or a baby, and it takes all the wrinkles out or puts them in.

[Kristin] (44:39 - 44:55)
Yeah, and you can probably program in something to see what you would look like if you were heavier or you were thinner, right?

[Regina]
Oh, absolutely. They've got apps for everything these days.

[Kristin]
Exactly. It's like applying one of those to make all the groups match on chronological age and weight and then compare them.

[Regina] (44:56 - 45:03)
So that would make all the effects of sugar really stand out because you're holding these other things constant.
[Kristin] (45:03 - 45:21)
That's the idea, exactly. We want to isolate that effect of sugar. And when we apply this model, they project that if you apply all these filters that the low-glucose group would look about 59.7 years on average, the high-glucose group 60.4, and the diabetic group 61.1, and this was a statistically significant difference.

[Regina] (45:22 - 45:36)
Not a very big effect, though. Not a big effect, especially in the non-diabetic group. We're talking, what, like less than a year?

Kristin, do I look that much different than eight months ago? Look at me.

[Kristin] (45:37 - 45:49)
Regina, you do not.

So yeah, this is not a big effect, and it also relies on a lot of assumptions. So for example, the model assumes that there is a linear relationship between perceived age and chronological age, and that one's almost certainly wrong.

[Regina] (45:49 - 45:52)
Oh, yeah. You go downhill faster when you get older.

[Kristin] (45:53 - 46:19)
So in a way, you're kind of penalizing the diabetic group because they're older. They're declining faster. So I'm not blown away by these results, Regina.

So first, remember for the non-diabetics, they're using that non-fasted glucose, not a great measure of your chronic sugar exposure. And then we see the diabetics actually look young for their age, and it's only when you account for weight that they look worse for their age. And the model's making a lot of assumptions, and then finally, even if this is a real effect, it's really quite tiny.

[Regina] (46:20 - 46:28)
Hmm. Yeah, I don't know this is one I'm going to worry about.

I think the weight might be giving me a better boost in my skin.

[Kristin] (46:28 - 46:37)
Right. So if you ate a lot of chocolate cake for a long time, you would boost the volume in your face, and that might counteract the AGEs that you accumulate. I don't know.

[Regina] (46:38 - 46:45)
I'd be happier. I'd have a big smile on my face, and that makes me younger. Yes, we're all happier if we get to eat chocolate cake.

[Kristin] (46:46 - 47:12)
And Regina, that's pretty much it. There's those two studies. There's a handful of studies where they didn't find any correlation, actually.

One of them was in women who were aged 28 years old. So it might not be surprising that they didn't find anything because, you know, it's hard to pick up subtle effects if you're young. All right, up until now, Regina, we've been talking about how your body makes AGEs.

These are produced in your body, right, because you have high levels of sugar, either because you eat a lot of sugar or maybe there's a problem with how you process sugar in the body, right?

[Regina] (47:12 - 47:12)
Mm-hmm.

[Kristin] (47:12 - 47:16)
But there's another way that we get AGEs. We actually eat them in our diets.

[Regina] (47:16 - 47:44)
Oh, no. Yeah, this is the part that I do not want to hear about, but I do. So we're going to take a break now so I can prepare myself mentally.

[Kristin]
Prepare yourself.

[Regina]
And then we'll come back up.

Kristin, we've talked about your medical statistics program.

It's just a fabulous program available on Stanford Online. Maybe you can tell listeners a little bit more about it. It's a three-course sequence.

[Kristin] (47:44 - 47:57)
If you really want that deeper dive into statistics, I teach data analysis in R or SAS, probability, and statistical tests, including regression. You can get a Stanford Professional Certificate as well as CME credit.

[Regina] (47:58 - 48:17)
You can find a link to this program on our website, normalcurves.com.

Welcome back to Normal Curve. The claim we're looking at today is that eating a lot of sugar in processed foods makes you look older.

[Kristin] (48:17 - 48:26)
I am going to put a shock into your system now, Regina, because I have a sweet tooth. I do miss sugar. You don't have a sweet tooth, but you love French fries.

[Regina] (48:26 - 48:27)
I do love French fries. Don't take away my French fries.

[Kristin] (48:27 - 48:43)
So I have bad news for you.

[Regina]
No, don't take them away.

[Kristin]
So it's not just about eating sugar, potentially, because the place that these AGEs were first discovered was in cooking, in the context of browning food. Remember I said they have a brown color?

[Regina] (48:43 - 48:43)
Oh, I do remember that.

[Kristin] (48:43 - 48:51)
When you brown food, the reason the food is brown is because of AGEs. So take, like, chicken. If you brown chicken in a frying pan, okay?

[Regina] (48:51 - 48:52)
So AGEs taste yummy?

[Kristin] (48:52 - 49:28)
They taste delicious. They're super yummy. Oh, yes.

For some people, it might even be more yummy than sugar. So sugar itself doesn't have AGEs in it, but your body makes AGEs when there's sugar in the body. But you can also create AGEs in things that you ingest.

[Regina]
Oh, my.

[Kristin]
I know. Isn't this awful?

[Regina]
Oh, my French fries.

[Kristin]
So if you eat AGEs, you do absorb some of them. So AGEs are produced whenever you have proteins and even a little bit of sugar hanging around and you heat things to a high temperature, especially dry cooking.

So things like grilling or barbecuing.

[Regina] (49:29 - 49:32)
But I love roasting grilled vegetables.

[Kristin] (49:32 - 49:59)
Frying. All of these are bad.

[Regina]
Like caramelizing things?

[Kristin]
Caramelizing, bad.

Yes. And processed foods. So when they're manufactured, like cookies and muffins that you manufacture in a plant, it's dry cooking at high temperature.

So they also have a lot of AGEs in them. And Regina, believe it or not, they sometimes actually deliberately add AGEs to processed foods for the flavor and coloring, because they do taste good.

[Regina] (49:59 - 50:07)
So like brown cheese, when it gets all nice and bubbly and gooey?

[Kristin] (50:07 - 50:08)
Brown cheese is bad. Pizza is bad for these too. Toast.

[Regina] (50:08 - 50:09)
Toast? No. All the crispy things?

[Kristin] (50:09 - 50:18)
Exactly. And you mentioned grilled vegetables.

You think vegetables are good for you, but when they're grilled, they have a lot of AGEs.

[Regina] (50:18 - 50:21)
I love grilled broccoli, though. I can't believe you're doing this to me.

[Kristin] (50:21 - 50:23)
Yeah. You want to steam your broccoli.

[Regina] (50:23 - 50:25)
No. I do not.

[Kristin] (50:26 - 50:51)
All right, let's do a little cooking chemistry lesson here, okay? There are a couple of different factors that determine how many AGEs are made in your food. So first one is temperature.

High temperatures drive these reactions. Second thing, pH. So you want to marinate your meat or fish in vinegar or lemon, because lower pH, more acidity helps to prevent the creation of AGEs.

Third thing, humidity. Dry cooking is bad. Wet cooking is good.

[Regina] (50:52 - 50:53)
No. No to all three of these.

[Kristin] (50:54 - 51:07)
But you like lemon on your food, right?

[Regina]
And grilled.

[Kristin]
Well, by adding the lemon, you counteract some of the grilling.

But if you boil things in water and liquid or stews, the temperature doesn't get that high.

[Regina] (51:08 - 51:16)
And it tastes gross.

[Kristin]
So you have to steam or boil everything.

[Regina]
No, I'm telling you.

No. So you're telling me my air fryer—

[Kristin] (51:16 - 51:33)
Air fryers are bad. Yes.

[Regina]
No.

[Kristin]
Yes, you're going to have to return it.

[Regina]
How can they be bad?

There's no oil.

[Kristin]
Because you're cooking things at a dry heat and high temperatures. And another thing that makes more AGEs, Regina, is if you add sugar.

So like barbecued meat, because barbecue sauce has a lot of sugar, or honey roasted nuts.

[Regina] (51:34 - 51:35)
I love roasted nuts. Roasted nuts are like one of the best things, right next to french fries.

[Kristin] (51:35 - 51:48)
Yeah, but AGEs. Sorry.

The good news, though, Regina, is that grilled meat has a lot more AGEs than grilled vegetables. So your grilled broccoli isn't the worst thing in the world.

[Regina] (51:49 - 51:52)
So I can tell that eating dinner at your house is going to be a lot less fun now.

[Kristin] (51:52 - 51:58)
I know. Now we've got the green juice, the carrot juice, which you call the dirt juice.

[Regina] (52:00 - 52:05)
And now we're steaming and boiling things instead of nicely browning and grilling them.

[Kristin] (52:05 - 52:57)
Yeah, sorry. Yeah, I've discovered you can boil chicken in water and then you drain the water and add the sauce.

[Regina]
Yummy?

[Kristin]
Yeah, I mean— I'm a terrible cook anyway, and I hate to cook, so it doesn't really matter.

[Regina]
Let's go out to eat while I'm here.

[Kristin]
All right. So this is where the AGEs come from in the diet. Studies show we absorb somewhere about 10 to 30 percent of these.

But we don't really know, are they actually doing damage in the body? That's controversial. I think this is a fun little side note here, actually.

There was a special issue of the Journal of Molecular Nutrition and Food Research in 2007. They dedicated an entire issue to this controversy. Are dietary AGEs, the ones you eat, not the ones your body creates, are they really harmful to humans?

And that whole issue came out of a workshop that was held. And do you want to guess what company sponsored the workshop, Regina?

[Regina] (52:58 - 53:11)
Either someone who wants AGEs to be bad, like a face cream or pseudo-pharma company, or someone who wants AGEs to be just fine, which would be the grilled cheese lobby.

[Kristin] (53:12 - 53:31)
Yes, it was Nestle. So I think what happened is this narrative was coming up in the literature about these dietary AGEs. So Nestle sponsored a workshop.

They wanted to make sure to say, hey, no, it's a debate. We don't have enough robust evidence. So there's a whole set of papers in this journal going back and forth about whether these are actually harmful.

[Regina] (53:31 - 53:43)
OK. Good to keep in mind. I'm going to use that. It's debatable, as justification to keep my French phrase. So have there been studies on dietary AGEs?

[Kristin] (53:43 - 54:09)
There have been some short-term intervention studies. They put people on low AGE diets, like they tell them to boil their food rather than grill it. And it does reduce the AGEs in their blood.

And they've tried to link these low AGE diets to improvements in health outcomes, things like vascular health or insulin resistance. The evidence is really mixed. There's actually a lot of negative studies.

But you know, short-term studies, are people really following the diet? Yeah.

[Regina] (54:09 - 54:10)
Yeah.

[Kristin] (54:10 - 54:17)
And it's hard to do a short-term study for appearance, Regina, right? Changing our diet for four weeks is probably not going to affect how old we look.

[Regina] (54:18 - 54:19)
Yeah. Not enough time.

[Kristin] (54:20 - 55:06)
Not enough time. Yeah. So the best we can really do for evidence here is looking for some observational diet studies.

So I went out and tried to find, has anybody like measured the AGEs in the diet and tried to link that to facial appearance? And I couldn't find any studies. But I did find a few studies that looked at people's overall diets and the relationship to wrinkles.

And that's as close as we're going to get. So we're going to talk about two of those studies. There was a 2007 study that used a big representative sample from the U.S. These were women aged 40 to 74. And they were given what's called a 24-hour dietary recall. So they had to write down everything they ate in the last 24 hours to get a snapshot of their diet. And then dermatologists rated their skin.

[Regina] (55:07 - 55:10)
Experts. Experts looked at their skin, not just random people on the internet.

[Kristin] (55:10 - 55:22)
Yeah. Exactly. Dermatologists rated their wrinkles and they tried to see if that related to their dietary patterns.

And the paper concluded that if you eat more carbs, you have more wrinkles, and also if you eat more fats, you have more wrinkles.

[Regina] (55:23 - 55:35)
Mmm. So this is not specifically about AGEs, but if you're having a lot of carbs and fats, that's likely to be this Western diet with a lot of processed foods that are high in AGEs.

[Kristin] (55:36 - 55:52)
That's the link.

Exactly.

Regina, the paper is kind of bizarre, though.

[Regina]
Okay.

[Kristin]
Because of the way they quantified wrinkles, they divided people into two groups.

You have a wrinkled appearance or you don't have a wrinkled appearance. They turned it into a binary variable.

[Regina] (55:52 - 56:08)
That is weird. Wrinkles are like on a spectrum. How do you turn this into a binary variable?

What are you completely smooth and then you're like one big wrinkle?

[Kristin]
Regina, am I wrinkled or not wrinkled, I want to ask you?

[Regina] (56:08 - 56:19)
I'm not going to answer that one. Nope. Nope. Nope.

[Kristin]
Well, I actually really want to know, though, Regina, if we would make the cutoff in this paper for being in the non-wrinkled group.

[Regina] (56:19 - 56:22)
I feel like you have to be a baby in order to be not wrinkled.

[Kristin] (56:22 - 57:21)
Well, interestingly, this was a paper on 40 to 74 year old women and only about 20% were classified as having a wrinkled appearance. So you and I, we might actually have made it into the non-wrinkled appearance group. It seems that you had to have a really wrinkled appearance in order to make it into the wrinkled group.

[Regina]
This is just weird.

[Kristin]
I mean, why would you turn wrinkles into a binary variable? It's just bizarre.

Another bizarre thing, remember, they're claiming that a high carb diet is linked to wrinkles and a high fat diet is linked to wrinkles. But in one of their tables, they compare the carb intake and the fat intake between the wrinkled group and the non-wrinkled group. And guess what?

There is no difference.
[Regina]
Really?

[Kristin]
The wrinkled group averaged 160 grams of carbohydrates per day versus 158 grams in the non-wrinkled group and the fat intakes were almost identical.

And these means have already been adjusted to account for age, BMI, smoking, sun exposure, all the important things.

[Regina] (57:21 - 57:29)
But you said if you eat more carbs or more fat, then you have more wrinkles. So how did we get there? What happened?

[Kristin] (57:29 - 57:49)
Well, a little statistical shenanigans, actually, Regina. They only found an effect because they ran a regression model that put fats, carbs, thiamine and vitamin C all into the same regression. And when they put them all together, then suddenly they found significant differences on all these nutrients between the groups.

[Regina] (57:50 - 58:03)
That is just weird, though. So if I'm understanding correctly, carbs were not different between the groups until after you account for, what did you say, fats, vitamin C, thiamine. How do you even interpret that?

[Kristin] (58:03 - 58:16)
Yeah, at this point, you are answering a very, very narrow question. If two people have the same fats, thiamine and vitamin C levels in their diet, then what is the effect of carbs? Well, what does that even mean?

[Regina] (58:16 - 58:25)
Diets do not work that way. It's not like a lever, you know, where you just move them independently around.

[Kristin] (58:25 - 58:38)
Exactly. As we discussed earlier, statistical adjustment is like applying filters to create an imaginary world. Sometimes this makes sense. Like you can predict what someone might look like if they were younger and thinner.

But what happens when we apply all of these diet filters?

[Regina] (58:38 - 58:53)
Right. Like if we want to look at the effect of carbs on wrinkles, then we have to imagine this weird world where somehow people eat the exact same amount of thiamine, fat, vitamin C, but they differ only on carbs. No, no way.

[Kristin] (58:53 - 59:26)
It would be a very bizarre and unrealistic world. Like maybe we have the fruity pie diet. You eat tons of fruity cakes and pies like mango cake and cherry pie.

So it's high in fat, high in vitamins, and high in carbs. And we're comparing that to, say, the fatty vegan diets, like a vegan diet with tons of nuts and avocados. So also high in fats and vitamins, but lower in carbs.

And then we ask, does the fruity pie group have more wrinkles than the fatty vegan group? It's weird.

[Regina] (59:27 - 59:58)
Yeah, the model at this point is just chasing statistical noise. It's fitting nothing. It's like clouds in the sky.

Kristin, I love how you brought this up because this is something that we should all be looking out for in papers is this fitting statistical noise. And how we can spot it is when a bunch of variables in isolation have no relationship to the outcome, but then you shove them together in the same model and all of a sudden they are significant.

[Kristin] (59:58 - 1:02:57)
Regina, this is where we need a training set and a test set like we talked about earlier in this episode. That's the way you can figure out if you're fitting noise. If we were to take the model here and apply it to a new group of people to predict based on their diets who was going to be wrinkled and who was going to be non-wrinkled, the model would almost certainly fail to accurately classify people.

It's likely to completely fall apart and have no predictive ability.

[Regina]
So are there any decent studies?

[Kristin]
There was.

There was a study that I found, a 2018 paper in the Journal of the American Academy of Dermatology. They looked at 2,700 elderly patients and they took digital photos and then put it through this software that quantifies your wrinkles.

[Regina]
Wow.

[Kristin]
This was a good study, though, because they said ahead of time, they pre-specified what aspect of diet are we going to look at? So rather than like, oh, the carbs after the vitamin, after the vitamin C, they pre-specified we're going to look at one dietary pattern. And that's the primary hypothesis.

So that keeps them from cherry picking later. So this was a study in the Netherlands and they looked at something called the Dutch Healthy Diet Index. It's a score from zero to 90.

And you get a high score if you eat a lot of fish, fruits, vegetables, fiber, things that are considered unprocessed healthy foods. You get a low score if you eat a lot of saturated fats, trans fatty acids, acidic drinks, sodium, alcohol, things that would be considered more of a Western diet, a lot of processed foods, right? So again, it's not comparing a high AGE diet to a low AGE diet, but the healthy diet is likely to be lower in AGEs than the unhealthy diet.

[Regina]
What'd they find?

[Kristin]
For women, but not men, they did find an association, not huge, but every 10 points higher you were on this Dutch healthy diet scale, you had 5% less wrinkling for your age. So, you know, if you went from zero to 90, you know, you might be able to see that.

[Regina]
That would be a big jump.

[Kristin]
A big jump. Yes.

And this was after accounting for all of the usual players, like they adjusted for age, BMI, smoking, sun exposure, and so on. So maybe there's something there. They have this great sentence, though, in their abstract, Regina, which I kind of love.

They say, other health conscious behaviors of study participants could have influenced the results.

[Regina]
Oh, you think? Maybe?

[Kristin]
Right. So, of course, if you are eating a healthy diet, what else are you doing that might influence how old you looked, right? Maybe you have a lot more money and you're going to the doctor and getting cosmetic surgery or Botox, right?

So it looks like, yeah, maybe having a healthy diet, you're going to look a little younger, but could it also just be that you're generally a health conscious person, right? They did another fun thing, though, after they looked at that pre-specified healthy diet index, they wanted to look at other dietary patterns, too. But again, rather than like looking at the carbs separate from the fat, separate from the vitamin C, which gets really statistically tricky, they did something cool called a principal components analysis.

[Regina] (1:02:58 - 1:02:59)
PCA. I love PCA.

[Kristin] (1:03:00 - 1:04:01)
Oh, yeah. We are not going to go into the details of that today. I'm going to save that for another episode.

Good. It'll be a fun episode, I promise. But basically what it's doing is it's picking out patterns of diet.

So like people who eat French fries, not you, Regina, but many people who eat French fries also eat a lot of hamburgers and drink a lot of Coke. It's looking for clustering, clusters of food types that go together. So they picked out three patterns, one which they called a healthy pattern, one which they called an unhealthy pattern and an intermediate.

The healthy pattern were people who were eating a lot of healthy things like vegetables and fish and it also included a little bit of wine. The unhealthy pattern was a lot of meats, snacks, soft drinks, coffee and non-wine alcoholic beverages. So again, not totally tracking with AGEs, but probably the healthy diets likely to have less AGEs.

They found that the unhealthy dietary pattern was significantly linked to wrinkles. But again, it was a small effect. If you went up a standard deviation in the unhealthy dietary pattern, that's a lot.

Your wrinkling went up like three percent.

[Regina] (1:04:02 - 1:04:14)
Oh, yeah. That's small. I personally want to see what the trade-off is, right?

Like to get those 10 points higher on this healthy diet, how much do I need to give up and is it really worth it to get three percent fewer wrinkles?

[Kristin] (1:04:14 - 1:04:29)
Or maybe you could just eat more vegetables and more fish, boiled and steamed, of course.

[Regina]
No, no, no. No thanks.

[Kristin]
All right. So, Regina, that's pretty much the evidence. There just aren't that many studies that looked at skin aging or appearance.

[Regina] (1:04:29 - 1:04:36)
And that's probably for a good reason. Yeah. Cancer, other things, diseases maybe take precedent.

[Kristin] (1:04:36 - 1:05:02)
Yeah. I support having more cancer studies than more wrinkle studies. Absolutely. Though I have to say, though, Regina, I think there is a reason to study appearance and wrinkles if you want to motivate behavioral change, which is a lot of public health, right?

I used to write for Allure, for example, about sun exposure. And if you tell people, you know, too much sun exposure, you might get cancer. I do not think that's as motivating as if you tell people too much sun exposure, you might get wrinkles.

[Regina] (1:05:03 - 1:05:10)
So, vanity. Forget about health benefits, blah, blah, blah, living longer, better heart health. No, you'll look better.

[Kristin] (1:05:11 - 1:05:19)
Well, I think it's human nature, right? Short-term, vain outcomes actually might be more motivating than long-term health outcomes. It's just how human nature works.

[Regina] (1:05:20 - 1:05:36)
I think you can get some grant money for this, actually.

[Kristin]
Oh, our conflicts of interest, Regina.

[Kristin]
Finally!

[Regina]
Kristin, are we ready to wrap up this episode now? I think so. This is where we rate the strength of evidence for the claim.

Could you repeat the claim, please, Kristin?

[Kristin] (1:05:36 - 1:05:40)
The claim today is that eating more sugar and processed foods will make you look older.

[Regina] (1:05:41 - 1:05:57)
Got it. And we rate this with our smooch scale, one to five, like an Amazon scale. One smooch means little to no evidence supporting that claim.

Five means really strong evidence in favor of it, so kiss it or diss it, Kristin, what do you say?

[Kristin] (1:05:57 - 1:06:55)
Regina, I'm going to go with three kisses on this one. And I need to say right up front, I might be a little biased. I have changed my diet because of the cancer.

I've cut down on a lot of sugars, trying to eat really healthy, so I probably just really want that to do something good for me, so I might be biased. To me, the most compelling evidence actually was that cadaver study. That idea that the cross-linked collagen makes the collagen that much tougher and older, it does make it seem biologically plausible to me that sugar is doing something in your body.

But it's a leap of faith to go from that to that it actually affects your appearance. And I have to say, if there is some effect on appearance, I think it's quite small, and it's probably dwarfed by a lot of other things. So if you, Regina, are looking at, do I boil my broccoli or do I use sunscreen, absolutely go for the sunscreen first, because that's going to have a much bigger effect.

Any effect of the sugar in the processed foods is probably much smaller.

[Regina] (1:06:56 - 1:07:30)
Yeah, going to be small. Okay, so you're giving it three smooches?

[Kristin]
How about you?

[Regina]
I'm going to give it 2.5, not because I disbelieve it, but because there's not a lot of hard evidence connecting all the dots. It does make sense that eating a lot of junk food, sugar, processed food, probably bad for you. Right.

But I'm wondering whether it's enough to overcome genetics and sun exposure and all these other things. Stress. When I am stressed, I look much older.

People can tell right away, and they comment on it. Thanks, guys. Very much.

Okay. Methodological morals. Do you have yours, Kristin?

[Kristin] (1:07:31 - 1:07:36)
Regina, mine's going to be, when you plagiarize, you steal the errors, too.

[Regina] (1:07:36 - 1:07:42)
Oh, I love it. Oh, I want to put that in a coffee cup. That one is fantastic.

[Kristin] (1:07:42 - 1:07:44)
Okay. And Regina, what's your methodologic moral?

[Regina] (1:07:45 - 1:07:52)
How about this? Overdone statistical adjustment is like overdone photo filters. At a certain point, it's just laughable.

[Kristin] (1:07:53 - 1:07:56)
Yeah. I love that. Yeah, that statistical model we talked about was laughable.

[Regina] (1:07:57 - 1:08:15)
Thank you, Kristin. This has been fascinating. We're still going to eat out anytime I come visit you next.

No mushy broccoli for me.

[Kristin]
Well, that's okay, because I hate to cook anyway, so sure.

[Regina]
But you really have taught me a lot, and I think the audience listening in, too.

Thank you, Kristin, and thank you, everyone, for tuning in.

[Kristin] (1:08:16 - 1:08:17)
Thanks, everyone, for tuning in. Thanks, Regina.