Tuesday, October 14, 2014

Obesity → Diabetes

A new study adds to the evidence that the prevalence of type 2 diabetes is rapidly increasing in the US, and our national weight problem is largely to blame.

The Centers for Disease Control (CDC) currently estimates that a jaw-dropping 33 percent of US men, and 39 percent of US women, will develop diabetes at some point in their lives (1).  Roughly one out of three people in this country will develop diabetes, and those who don't manage it effectively will suffer debilitating health consequences.  Has the risk of developing diabetes always been so high, and if not, why is it increasing?

In the same issue of the Annals of Internal Medicine as the low-carb vs. low-fat study, appears another study that aims to partially address this question (2).


Menke and colleagues examined records of diabetes prevalence from the Centers for Disease Control's NHANES surveys.  They identified cases of diabetes by either 1) self-report of a diabetes diagnosis, or 2) fasting plasma glucose of 7 mmol/L (126 mg/dL) or more.

Between the 1976-1980 and the 2007-2010 survey periods, the prevalence of diabetes in men increased from 4.7 to 11.2 percent, and among women it increased from 5.7 to 8.7 percent.  In other words, the overall prevalence of diabetes has nearly doubled over the last 30 years.

The investigators then looked at a few physical and demographic characteristics to see how well they explained (mathematically speaking) the increase.  Adjusting for age diminished the trend slightly-- since our mean age has increased over the last 30 years and age is a major risk factor for diabetes.  Adjusting for population race/ethnicity changes had little impact on the trend.

However, the investigators found that adjusting for body mass index (BMI) and/or waist circumference greatly attenuated the trend, suggesting that the increase in diabetes prevalence is intimately linked to our expanding waistlines.

The story becomes a bit more complicated from here.  Increasing BMI accounted for essentially all of the increase in diabetes prevalence in women, but in men it only accounted for some of it.  There are apparently other unidentified factors in men, such as changes in physical activity and/or sedentary time, deteriorating sleep habits, or better diabetes diagnosis, that have played a role in the increase.

This study adds to a large body of evidence linking excess weight with diabetes risk.  Here's a graph illustrating the relationship between BMI and the likelihood of developing diabetes over a 5-year period among US men (3):


People in the highest weight category had a 42 times higher diabetes risk than people in the lowest weight category.  Body fatness is quite obviously a dominant risk factor for developing diabetes, even though lean people do sometimes develop diabetes*.

Obesity and diabetes risk are closely linked to one another, although there are other important risk factors, such as genetics, age, and inactivity.  Even a modest degree of fat loss can greatly reduce the risk of developing diabetes among people who are overweight or obese (4).


* One confounding factor is that uncontrolled diabetes itself causes weight loss.  People are sometimes diabetic for a while before receiving a diagnosis, such that they aren't as overweight by the time they're diagnosed.  Often, that person was previously overweight or obese, and excess weight contributed to the development of their disease.  This confounds the relationship between excess weight and diabetes.  However, some people do develop diabetes without ever having excess body fat.  This can be due to a combination of diet and lifestyle factors, age, and unlucky genetics.

21 comments:

Beth@WeightMaven said...

You say "our national weight problem is largely to blame" ... but does this paper prove causation or just correlation? Couldn't it be that what makes us fat eventually makes us diabetic? Is Unger and Scherer's paper suggesting that obesity is protective at first all that heretical?

Tucker Goodrich said...

"One confounding factor is that uncontrolled diabetes itself causes weight loss."

That's fairly significant. So diabetes may be the body's reaction: an attempt to dispose of unwanted fuel?

What's the mechanism, do we know?

Steve Parker, M.D. said...

U.S. men's lifetime risk of developing diabetes has caught up with women's: four in 10. That's according to CDC-affiliated researchers.

Reference at Lancet:http://www.thelancet.com/journals/landia/article/PIIS2213-8587(14)70161-5/fulltext#aff1

-Steve

valerie said...

I'm with Beth on the question of causation vs correlation. Your title implies causality, but the data you present shows only association.

Also, you mention that weight loss in overweight or obese people reduces their risk of diabetes. Wouldn't that be true for everyone, including people of normal weight?

Finally, in your chart showing risk of diabetes against BMI, you seem to have picked the most dramatic relative risks available in all the literature. When I google "risk ratio BMI diabetes" I get plenty of much more modest results.

Stylooke said...

Hi Tucker,

By refusing to take up glucose through insulin, the body excretes part of the excess blood sugar through the urine, leading to less fat production and more weight loss.

In my opinion insulin resistance is a natural mechanism in reaction to an excess of energy. By reducing the total amount of calories (and maybe glucose (starch) itself, probably not necessarily) insulin resistance is quickly reversed. Also look at geriatric bypass patients which are typically cured of their diabetes type II within DAYS after their operation.

Nils

Stylooke said...

I obviously meant bariatric surgery, oops!

Nils

Unknown said...

Thanks for a most informative article. The mechanism of excessive weight gain has always interested me. A fantastic lecture on the microbiome of the gut http://youtu.be/GAF-AblJpfM ( about 46:40 or so) suggests that the contamination of the food chain by sub-therapeutic doses of antibiotics may be playing a significant roll in the development of obesity and the subsequent disease like sleep apnea and diabetes.
Cheers,
Don S in sunny Vancouver, Canada

Unknown said...

Why is it that type II diabetes is considered to have a significant genetic component? I wrote about type II diabetes last week here: diabetesT2 and frankly felt that type I has a bigger genetic component.
My understanding is that type I has fewer, but more significant associated genes (many on chromosome 6 and the HLA cluster). Type II seems to have a lot of weakly associated genes, which to me says there is no real link. There will be a few people who are more prone to type II diabetes by bad luck, but to have all the associated genes would be virtually impossible. For me causation of type II is all about poor diet and sedentary lifestyle.

Stephan Guyenet said...

Hi Beth and Valerie,

My statement of causality isn't just based on this paper, it's based on a large body of evidence suggesting that fat gain increases T2 diabetes risk, and fat loss decreases it. That's why I wrote "A new study adds to the evidence..."

Valerie, you asked about my selection of studies. The reason this one appears so much more impressive is that the weight categories were divided up much more finely than in most other studies. When you lump together large BMI chunks, it greatly attenuates the apparent differences. Also, this study was prospective, which diminishes that confounder I mentioned in my article that people with diabetes lose weight pre-diagnosis due to the diabetes. The study I cited is consistent with many others showing that excess fat is a dominant risk factor for diabetes, along with age, inactivity, and genetics.

Hi Tucker,

I don't know whether or not glycosuria (glucose dumping in the urine) is a "deliberate" response to excess glucose, or simply the result of a system (glucose re-uptake by the kidney) exceeding its normal tolerances. I'd guess the latter.

Hi Robin,

T2 diabetes has a significant genetic component, but you're right that the genetic contribution is likely from many small additive effects. Most of the gene variants whose function is known relate to pancreatic function.

JWO said...

Blacks and Hispanics have higher rates of obesity and diabetes and the Blacks + Hispanic percent of the population has risen. This complicates the simple picture presented.

http://www.cdc.gov/features/dsobesityadults/
http://minorityhealth.hhs.gov/omh/browse.aspx?lvl=4&lvlID=18

Stephan Guyenet said...

Hi Jim,

Controlling for race/ethnicity had little impact on the results.

Robert Andrew Brown said...

gut http://youtu.be/GAF-AblJpfM

Thanks Donald this is indeed an amazing must watch lecture

RLL said...

Many of us diabetics who now maintain near normal BGs (A1C 5.0, BGs 30 day average of 93) suspect that early diabetes contributed to our weight gain. Very low carb and high fat help us more than other people in maintaining appropriate weight.

Sergio said...

Hi Stephan,
is there an explanation for the fact that men overtook women?

"Between the 1976-1980 and the 2007-2010 survey periods, the prevalence of diabetes in men increased from 4.7 to 11.2 percent, and among women it increased from 5.7 to 8.7 percent"

It's pretty natural to think that obesity is less damaging in women...

marksuave25 said...

Black's percentage of the population has remained constant for decades. Hispanics' percentage of the population has grown substantially over that time.

Don S said...

First a question - are you familiar with any data that has tracked if the rate of T2DM within each BMI grouping has changed over the years? (The graph you present is from 1994. Of note in that article even after adjusting for BMI relatively increased waist circumference still presented a 3.5 RR.)

What I can find is this article - http://ajcn.nutrition.org/content/81/3/555.long - which shows WC far outperforming BMI as a predictor even after controlling "for physical activity, smoking, parental history of diabetes, dietary intakes of trans fat and cereal fiber ..."

What this suggests to me is that while weight, and even fat mass, may be major contributors, that which controls the partitioning of fat mass appears to be an even stronger factor. If those factors (whatever they might be) have changed over time then the rate of T2DM per BMI group (matched for ethnicity and FH) should be increasing. Those factors might, for example, include a greater proportion of calories coming from added sugars, or possibly other items mentioned in your previous series detailing risk factors for T2DM.

Thoughts?

Don S said...

Following up on my own questions ...

Interestingly, WC within BMI groupings has increased (http://onlinelibrary.wiley.com/doi/10.1038/oby.2007.505/full): "Among both men and women, significant increases in mean WC were found in all three BMI categories. Both men and women with a BMI of ≥30 kg/m2 had the largest increase in mean WC"

So while BMI has increased, WC as a proxy for central obesity has increased more, at all levels of BMI.

Also of note in the context that your featured article was unable to analyze by WC and found BMI alone less well explained the increase in T2DM in males - the article I cited documents that between 1999 and 2004 males had a significnat increase in WC per BMI level while women did not.

Of interest the association between BMI and T2DM with increased BMI has persisted through the years but appears to be less large than it was in that 1994 study- http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3976458/ - the RR of those in class 2 and 3 obesity is now 3.6 and 5.1 instead of the 42 it was for the combined group in 1994.

Let us assume that the genetics predisposing to central obesity have not changed - a key question now becomes what has changed in lifestyles or environments such that the nature of the obesity has changed, such that fat is more often partitioned into central stores, and such that the overweight by BMI are now more often obese by WC measures and more commonly diabetic.

Hypotheses?

Mister Roboto said...

My own opinion is that the impact of high fructose corn syrup consumption is the missing link in the diabetes-obesity chain. Not only is consuming a more noxious form of refined sugar likely to make you fatter, but there is also evidence that HFCS promotes leptin-resistance. Leptin is the hormone that signals your physiology that you have had enough to eat. Resistance to it would make people want to eat more, which is bad news if they are eating the Standard American Diet.

Anonymous said...

@Don S.

Regarding the association between obesity and diabetes, you said: "Hypotheses?"

Here is one: The increase in diabetes throughout the past century is due mainly to the increased intake of linoleic acid (LA) from vegetable oils which elevates endocannabinoids thereby inducing hypothyroidism and insulin resistance that leads to diabetes.

Throughout the past century total dietary fat has stayed about the same but vegetable oil consumption has increased from 20% to 60% of total fat intake. LA intake has increased from 2.2% to 7.2% of calories mainly due to increased soybean oil consumption. (PMID 21367944)

An ecological study assessed the association between changes in intake of 273 food commodities and increased obesity during the past century and found a strong association with increased obesity and increased food intake only for those foods high in LA (soybean oil, poultry and shortening). A weak association was found between increased sugar intake and obesity. The same study found that increasing dietary LA acid in rodents elevates LA derived endogenous endocannabinoids and induces obesity, independent of food intake, which is preferentially directed to adipose tissue. (PMID 22334255)

Linoleic acid (LA) is the precursor to arachidonic acid which is the backbone of eicosenoids called endocannabinoids (endogenous cannabinoids) which activate the same receptors activated by marijuana. The endocannabinoid system consists of the endocannabinoids, endocannabinoid receptors CB1 and CB2, and enzymes that degrade the receptors. The breakdown of the endocannabinoids by degrading enzymes reduces the effects of endocannabinoids compared to the effects of marijuana. The most studied endocannabinoid is anandmide which is degraded by fatty acid amide hydrolase (FAAH). FAAH knockout mice are prone to excess energy storage, insulin resistance and adiposity whereas mutations in FAAH are associated with obesity in humans.

Elevated anandamide promotes centrally mediated hypothyroidism which causes reduced energy expenditure that induces obesity and results in hepatic insulin resistance and elevated triglycerides. (PMID 22912404) It has been known since the 1950s that excessive LA intake results in hypothyroidism.

The prevalence of hypothyroidism is quite high in type 2 diabetes patients above 45 years old and more so if their BMI is over 25. (PMID 23565418)

Obesity and a higher intake of linoleic acid are independent risk factors for dysregulation of glucose tolerance. (PMID 8969289)

In summary, there is a strong association between increased LA intake of increased obesity during the past century, and the cause, centrally mediated hypothyroidism that reduces energy expenditure and causes insulin resistance, seems well understood. The association between linoleic acid induced hypothyroidism, obesity, insulin resistance and.type 2 diabetes also seem well established. It seems reasonable to conclude that the increase in dietary LA during the past century may be a significant contributor to the increased incidence of diabetes.

Emily Miller said...

Yes, I agree with you Stephan that over weight is one of the major cause for diabetes.So its very important to do regular exercise and control your diet so that it will help you to control obesity and also follow doctor's instruction for obesity as well as diabetes.
Jennie

Juanita Swindell said...

The numbers certainly are scary. What’s even more scary is the fact that we ignore them most of the time – after all they are just numbers that mean nothing to us. That is, until we find ourselves one of those numbers. That’s what happened to me. I usually just glanced at these statistics and never gave them a second thought. To me, diabetes was just something that happened to other people. Well, not anymore. My doctor told me I was diabetic a few days ago and suddenly my whole perspective has changed on the statistics. My advice to everyone out there is to make those lifestyle changes today so you are not one of those statistics. Diabetic Problems