Exploring Diabetes vs. Obesity

Teach the controversy?

ZOMG DIABESITY

It seems somewhat accepted in (some?) medical circles that diabetes and obesity are, if not the same thing, closely related. Many seem to even imply that obesity CAUSES diabetes, which I think is flat-out wrong.

My personal thought is that both are caused by a similar defect causally upstream, but it depends on a bunch of factors if just one or both materialize.

So let’s compare diabetes and obesity levels across the world, shall we?

This whole idea of looking into the correlation came from a discussion with a friend who’s lived in China, Japan, and many other Asian countries. He related the extreme diabetes rate the Chinese have, especially given that they’re much, much less obese than the U.S.

Overview

Note: in all of the following diagrams, the Y axis is percentage of a country that has diabetes, and the X axis is percentage of obesity. So up and to the right is bad.

The line is a linear regression, showing how the two values correlate over all data points.

Source for both data sets: https://worldpopulationreview.com.

In the beginning, the font size is super small. That’s just because it’s hard to get 200-300 country names on a diagram and keep it readable. I also deleted a lot of the country names just because they were all writing over each other.

Later, for per-continent graphs, the font size will be bigger and you’ll be able to read all the country names for a given continent.

All countries

At first glance, the regression clearly goes up and to the right. That means diabetes and obesity are pretty strongly related.

Of note: the highest obesity levels are MUCH higher than the highest diabetes rates. Nauru, the most obese country, has a staggering 61% obesity rate, whereas Kiribati, the most diabetic country, “only” has 25% diabetes.

You can also see that most countries are clustered in the bottom left, with a few outliers extending the scales in both dimensions by nearly 50%. We’ll see more in the detailed continent views.

Boxes for continents

But first, let’s draw colorful boxes. These help us visualize where individual regions lie on the scales. Is Oceania a continent? A region? Who knows. Pretty sure that “Middle East” is NOT a continent, but most of the others are. Oh well, that’s the way the data set came.

• Asia and Africa are doing super well on both obesity and diabetes.

• Europe and South America are not quite as good on obesity, with their least-obese countries being about as obese as the most obese Asian/African countries. Interestingly, they do just as well on diabetes, if not better. More on that later.

• North America and the Middle East are quite bad, both significantly more obese and diabetic than Asia, Africa, Europe, and South America.

• The Middle East is still quite a bit worse, especially since the box shape of North America is mostly driven by a few small outliers, as we’ll see later (LOOKING AT YOU, U.S. OF OBESITY!)

• Oceania just blows the top off both in terms of obesity and diabetes. A whole 33% of the obesity axis is JUST for Oceania, and about 20% of the diabetes axis.

Africa

I won’t pretend to know much about Africa. The correlation between obesity and diabetes is positive here.

I do know that Morocco, Tunisia, and Algeria are in North Africa, and South Africa is (I think?) the richest country on the continent. So it could be that the richer countries are more obese.

Not sure about Libya being both the most obese and 2nd most diabetic. Haven’t they had wars/civil wars forever? I’d kind of expect them to be.. not obese and diabetic from that. For reference, the Wikipedia paragraph on Libya post-civil war begins with “In December 2021.”

Asia

The first interesting fact is that the correlation between obesity and diabetes is actually negative here, and that’s probably still influenced by the somewhat-outlier Malaysia, by far the most diabetic Asian country.

On second glance, there seem to be two completely different Asias.

Starting at an incredible 2% obesity rate (!) with Vietnam, there is a cluster of East Asian, South Asian, and Southeast Asian countries on the left. These countries have a obesity range of 2-10%, from Vietnam to Thailand. (Malaysia is an exception, also being a Southeast Asian country. I’m not sure where to put Brunei, I guess it’s SEA too?)

On the right, we have the Central Asian countries, ranging in obesity from 14% (Tajikistan) to 22% (Georgia, not the peachy one). So with the exceptions of Malaysia and Brunei, there’s a 5% gap in obesity between these 2 clusters! That’s quite extraordinary.

If we looked at all of Asia, we’d think the obesity range goes from 2% all the way up to 22%. But really, both clusters have a range of about 8%, and there’s no overlap between them.

Interestingly, the more obese Central Asian cluster is less diabetic than the slim cluster. This is the reason for the negative correlation.

In fact, let’s remove the outliers Malaysia and Brunei for a second:

Now the 2 clusters are completely distinct. The correlation is clearly negative. There is a nearly 5% gap in obesity between the two clusters. Intriguingly, there is both a Central Asian low-diabetes cluster and a Southeast Asian one. There aren’t many mildly diabetic countries, and then there are a lot of very diabetic countries in East Asia and South Asia.

Let’s take a look at the clusters separately. Here’s Peripheral Asia:

Correlation between obesity and diabetes is quite clear.

And here’s Central Asia:

Huh, a negative correlation.

If we’ve learned one thing, it’s that “Asia” isn’t a very useful category for this, and we need to look closer. Asia’s a big place.

Europe

Ah, Europe! Home of pretty churches and smelly cheeses.

But what’s that? The correlation between obesity and diabetes is practically non-existent! The place looks like a random scatter plot. (Also true for actual Europe!)

You’d kind of expect to see the Swiss and the Scandinavians be among the least obese, I guess. And the Italians and the Netherlanders. But if you’d told me that Malta is the most obese, while Italy is among the leanest… Malta’s literally an island off an island off of Italy.

Also how come the Ukrainians have amazing blood sugar control, whereas the Russians are the worst at it?

And how come Italy, France, Albania, Greece, Croatia, and Spain are all over the place? Aren’t they all eating the Mediterranean Diet (tm)? (In fact, maybe the Algerians and Moroccans and Tunisians from our outliers in Africa are eating the Mediterranean Diet, and that’s why they’re among the most obese Africans?)

How come Spain is way more obese than Portugal, but Portugal is way more diabetic?

It all seems pretty random to me.

One interesting fact: Europe’s diabetes range goes from 2-10%, but its obesity range goes from 17-29%. So even the leanest Europeans are as obese as most of the more-obese Central Asians we just saw!

Middle East

Oh boy. Up and to the right.

Interestingly, it actually seems very clustered. I’m pretty sure Yemen’s been at war for decades.

Cyprus?! Shouldn’t that be in Europe? Who makes these rules?!

But if we take those 2 outliers out, the regression becomes even crazier.

Jordan also seems kind of out there by itself. And up top, that looks like another cluster of mostly Emirates and Lebanon and Egypt.

The Middle East is pretty diabetic, with Israel, the lowest-diabetes country, at 7% all the way up to over 20% for the Emirate cluster.

If we take out Yemen (war) and Cyprus (Europe), the obesity range isn’t that wide, but it’s pretty high: 25-35%.

North America

#1! Home of the Brave!

Oh wait, I’m thinking of that other America.

Crap, are we really all the way there to the right? Jeeeez. Yes, we’re a stunning 5% more obese than the next-obese North American country. That’s a gap as big as between Central Asia and Peripheral Asia, remember? We’re about as obese as Kuwait, the most obese Middle Eastern country.

Luckily, we’re not nearly as diabetic. So we’ve got that going for us. We’re not even that diabetic for North America.

And in fact, the correlation between obesity and diabetes isn’t particularly strong in North America.

Fascinating: Mexico is significantly less obese than the U.S., but nearly twice as diabetic. Crazy.

Also interesting is that we have nearly the same diabetes floor as the Middle East. There’s no country in North America with less than 6.5% diabetes.

South America

South America: land of confusion! Another negative correlation, although it’s mostly driven by the extreme outlier country of Guyana. Not sure if they miscounted their diabetes or what, but they have a 17% rate whereas the next highest countries only have 10%. That’s a pretty extreme difference.

I’m generally confused by this graph, maybe because I don’t know that much about South America.

Why is Argentina the most obese? Don’t they constantly crash their economy? Venezuela is embargoed and apparently has hyperinflation, why are they so much more obese than Colombia and Brazil, which are generally doing quite well?

Another 5% floor to diabetes on this continent, though. Interesting how Asia, Africa, and Europe have some countries with much lower diabetes rates. Maybe sampling/measuring differences?

Oceania

Last but not… holy banana. The obesity range in Oceania goes from 20-62%. That’s just stunning.

Again, I don’t know anything about almost any of these countries. It seems that they are very distinct from what I can see, though. E.g. Australia and New Zealand are pretty similar to America culturally and economically, I’d say, whereas some of the others seem to be small island states. So I guess the Western Diet doesn’t lose for once? These 2 seem to be in a similar spot to Canada.

Equally high is the diabetes rate, topped by Kiribati’s 25%. Imagine that, that’s over twice as bad as the U.S. Although even Australia, the least diabetic country, has over 6%.

Let’s cheat a bit: let’s say that New Zealand and Australia are weird outliers in Oceania, because they’re too American. And then, for good measure, let’s remove Papua New Guinea and Fiji. Not for a good reason, but just because I want to show you how weak the correlation becomes:

I think NZ and AUS are a good cluster to remove and make sense. The other 2, honestly no idea. But if we do delete these, the line is pretty flat.

The difference between Samoa and Kiribati is very interesting. They have nearly the same, extreme obesity rate of 46-47%, but samoa only has 8% diabetes (2% less than the U.S.) whereas Kiribati has a mind-blowing 25%.

I do seem to recall anecdotally that some island people seem to have an uncanny ability to metabolize extreme glucose loads. Maybe some specialized adaptation? Or maybe they have some crazy hack in their diet/culture, like taking long walks on the beach after meals, or an ancient tradition of serving metformin after dinner?

Interesting Comparisons

Here I’ll just highlight a few comparisons between continents that I found interesting.

South America vs. Europe

Slightly hard to differentiate, but Europe is in the red jersey whereas South America is in the purple.. never mind, this is not the World Cup.

Essentially, Europe and South America have the exact same obesity distribution if you discount Europe’s Bosnia and Herzegovina.

South America is more diabetic though, even without its extreme outlier, Guyana. The South American countries basically inhabit the upper half of Europe’s diabetes scale.

Asia vs. Europe

I thought this one was interesting because Asia and Europe have almost no overlap in terms of obesity, but have almost the same range of diabetes.

As you can see in this one, the Central Asian countries are closer to Europe in terms of obesity, with the most obese ones being among the slimmer European countries, whereas the East/South/Southeast Asian countries are vastly slimmer.

On the other hand, the Europeans are a lot less diabetic. While there are some very low-diabetes countries in Asia, there’s also a huge cluster up top near the 10%, where only the European outliers sit. And that cluster includes India and China, which add up to over 3 billion people.

This graph is the one that first caught my attention about this whole diabetes vs. obesity thing. It looks almost as if Europe was Asia shifted right by 20% of obesity. You could say that the Europeans are 5x as obese as the Non-Central Asians, yet they are less diabetic? What’s happening here?

Asia vs. North America

Lastly, let’s compare North America and Asia. There are quite a lot of countries here, so I’ve removed some of the names for ease of reading.

I mainly want to focus on the big 3 Northern American countries of Canada, the U.S., and Mexico (in order of maple syrup consumption) and compare them to the East Asian countries.

Canada and Mexico sit at 30% obesity, whereas the U.S. is at 36%. Meanwhile, India, Japan, South Korea, and China all sit around 4-6% obesity.

On the other hand, India, China, Singapore, and North Korea (???), are as diabetic or more than the U.S.! Even Japan and South Korea aren’t far behind at 7.5% vs. the U.S.’s 9% diabetes. Mexico, on the other hand, is vastly more diabetic than any Asian country.

This is even more extreme than the comparison with Europe.

China has 6% obesity, we have 36%. Yet we have the same diabetes rate as them. How is that possible if diabetes and obesity are the same thing, or caused by the same thing, or one is caused by the other?

Where are obesity and diabetes correlated?

It seems that there are some regions/clusters where obesity and diabetes are clearly correlated, and some where they’re not.

Correlated:

• Globally

• Africa (I’ll admit I don’t know enough about Africa to analyze this well, maybe it’s sub-Saharan Africa vs. Westernized Africa + Mediterranean Africa..?)

• Middle East

• Oceania (if we don’t remove any outliers)

• Within East/South/Southeast Asia

Not correlated:

• Europe

• Central Asia vs. Peripheral Asia

• Within Central Asia

• South America

• Basically North America (very weak correlation)

• Basically Oceania if we remove New Zealand, Australia, Fiji, and Papua New Guinea (which drive almost the entire correlation)

Genetics? Diet? Culture?

Who knows.

There are some places where it intuitively makes sense to me. E.g. to my knowledge, countries like India, China, Japan, Korea, and Thailand eat a lot of rice. Maybe rice doesn’t make you obese, but it still gives you diabetes?

And from what I hear, the Central Asian plains cultures eat way more meat and dairy. Maybe those are more obesogenic than a mostly rice-based diet, but protective against diabetes? Or, at least, they’re not more diabetes-inducing than the rice diet.

But don’t many South American countries also eat a lot of rice? Why don’t they see this correlation?

Why is the U.S. super obese, but not particularly diabetic? No clue. Maybe just earlier prevention through the medical system? In general, I could see that diabetes is detected and treated earlier in richer countries. We seem pretty decent at diabetes management, at least compared to how we treat obesity.

Why are the European countries drastically more obese than the South/East/Southeast Asian ones, but not more diabetic, or even slightly less?

And what is it about many of the Oceanian countries that makes them such extreme outliers in both obesity and diabetes? Yet somehow, the Samoans are nearly immune to diabetes?

One theory I’ve heard is that many Asian people are genetically less likely to get obese from the same diet, but they still get diabetes just as much. This data would seem to support that theory (although, of course, it could also be diet/culture).

There’s even speculation that obesity is protective of diabetes, because the body can shuttle excess glucose into fat cells. Or maybe it’s just that they don’t notice it visually, and thus it’s not detected as early?

300lbs and perfect blood sugar on keto

Personally, since I’ve been doing keto for over 7 years now, my blood sugar has been great as long as I can remember. I’ve never tested a fasting blood sugar in the prediabetic (>100mg/dL) range, and am usually way below, e.g. in the 80mg/dL range.

This despite being obese to extremely obese most of my life. Many of these tests were done when I was 250-300lbs. Doctors would regularly be surprised, because they assumed I must be deeply diabetic at that weight.

So at the very least, it seems that if you do pretty strict keto, you won’t get diabetes even if you’re obese. It’s very possible (likely even, I think) that I would’ve gotten diabetic if I’d become obese on a carb-heavy Standard American Diet.

My pet theory is that something makes you obese by overeating, and if you overeat carbs, you’ll become diabetic to boot. Since I obesed myself mostly on protein and fat, I didn’t get diabetes.

Yay me!

I guess one thing to take away from this: even if you can’t lose weight on keto, it’ll almost certainly fix your blood glucose as a prediabetic or non-late-stage Type 2 diabetic.