Eating Processed Red Meat Every Day Apparently Increases the Risk for Developing Type 2 Diabetes by 51 Percent — according to a Large Study by the Harvard School of Public Health

12 August 2011

This meta-analysis may indicate a valid correlation between aspects of diet and the development of type 2 diabetes

The Harvard School of Public Health reported on a study of type 2 diabetes (the body’s failure to use insulin efficiently):

[An]Pan, senior author Frank Hu . . . and colleagues analyzed questionnaire responses from

37,083 men followed for 20 years in the Health Professionals Follow-Up Study;

79,570 women followed for 28 years in the Nurses’ Health Study I;

and

87,504 women followed for 14 years in the Nurses’ Health Study II.

They also conducted an updated meta-analysis, combining data from their new study with data from existing studies that included a total of 442,101 participants, 28,228 of whom developed type 2 diabetes during the study.

After adjusting for age, body mass index (BMI), and other lifestyle and dietary risk factors, the researchers found that a daily 100-gram serving of unprocessed red meat (about the size of a deck of cards) was associated with a 19% increased risk of type 2 diabetes.

They also found that one daily serving of half that quantity of processed meat—50 grams (for example, one hot dog or sausage or two slices of bacon)—was associated with a 51% increased risk.

Citation

The study authors used “relative risk” terminology to say the same thing and added ideas about what to substitute for red meat:

The results were confirmed by a meta-analysis (442,101 participants and 28,228 diabetes cases): the RRs [relative risks] . . . were 1.19 . . . and 1.51 . . . for 100 g of unprocessed red meat and for 50 g of unprocessed red meat, respectively.

We estimated that substitutions of one serving of nuts, low-fat dairy, and whole grains per day for one serving of red meat per day were associated with a 16–35% lower risk of T2D.

© 2011 An Pan, Qi Sun, Adam M Bernstein, Matthias B Schulze, JoAnn E Manson, Walter C Willett, and Frank B Hu, Red meat consumption and risk of type 2 diabetes: 3 cohorts of US adults and an updated meta-analysis, American Journal of Clinical Nutrition, doi: 10.3945/ajcn.111.018978 (early online publication, 10 August 2011)

Ideally, to put these findings into meaningful context, we need to think about absolute (not relative) risk

Notice that this research (like most medical reporting) simply gives readers the relative risk of something happening, when compared to a group of people that does something slightly different.

For example, take the statement that most readers will extrapolate from this research:

“If I eat 50 grams of processed red meat every day, all else equal, I’ll be 51 percent (1.51 times) more likely to get diabetes type 2.”

Implied is a comparison between (a) people who eat processed red meat daily and (b) those who don’t.

But that increased risk (called “relative risk”) often doesn’t tell the thoughtful reader much that is actually useful. It doesn’t tell us how likely we are to get diabetes in the first place, even if we don’t eat processed red meat.

Measures of relative risk tell us nothing about the disease prevalence that the increased (or decreased) risk is referring to. In other words, would you care if eating red meat elevated your personal near zero risk for getting diabetes (or anything else) by 51 percent?

No — at least not if you’re sane.

So, absolute (not relative) risk is what most people are really concerned with. Absolute risk quantifies the probability of something actually happening to you.

Here is a good example of how measures of relative and absolute risk work together:

Say the absolute risk of developing a disease is 4 in 100 in non-smokers. Say the relative risk of the disease is increased by 50% in smokers.

The 50% relates to the 4 - so the absolute increase in the risk is 50% of 4, which is 2. So, the absolute risk of smokers developing this disease is 6 in 100.

Absolute risk and disease prevalence are related

The above example about smoking implies something that most people miss, until they think about it. If the disease in question is rare, then who cares about even large relative risks of getting ill?

Example

Let’s say I have an aluminum foil hat that I’m fond of wearing because its reflectivity allows my wife to keep track of me in our farm fields. Being metal, it increases the risk of me being hit by lightning.

But if lightning has hit only one human being over the last one million years in the region where I live, will I be willing to displease my wife by getting rid of my metal foil hat?

Probably not. A 51 percent increase in risk (as we have in the diabetes example), relative to a near zero absolute risk, remains a near zero absolute risk.

On the other hand, if she and I live on the top a mountain where lightning strikes are common (say 10 strikes per day), do you think I’ll increase the risk of being crisp-i-fied by wearing my metal foil hat?

A 51 percent increase (as we have in the diabetes example) in an already strong probability of being zapped is worrisome. And who needs that?

The prevalence of diabetes is relatively high in the United States, so increasing one’s risk of getting it by 51 percent is arguably worth paying attention to

According to the Harvard School of Public Health press release:

In the U.S. alone, more than 11% of adults over age 20—25.6 million people—have the disease, according to the Centers for Disease Control and Prevention.

Most have type 2 diabetes, which is primarily linked to obesity, physical inactivity, and an unhealthy diet.

“But wait, didn’t the Harvard meta-analysis provide a much lower diabetes rate?”

Clever of you to catch that.

The researchers wrote, “The results were confirmed by a meta-analysis (442,101 participants and 28,228 diabetes cases) . . . .”

If we do the math, we see that only 6.4 percent of the 442,101 subjects developed Type 2 diabetes over the span of the meta-analysis.

“Isn’t that way less than the 11 percent plus the press release told us about?”

Maybe, maybe not.

The problem lies with our definitions. “Prevalence” counts the proportion of people who have the disease at one time. That is what the press release was presumably reporting from the Centers for Disease Control.

The researchers, on the other hand, simply reported how many people developed the disease over the multi-year period that their meta-analysis covered. Conceptually, we can see that this is a different measure. It includes only the number of new cases over time. It does not count the background number of already existing cases.

And you may also have caught the fact that the Harvard press release statement about diabetes prevalence included type 1 cases.

Note

Type 1 diabetes refers to the body’s failure to produce insulin. It is often called “juvenile diabetes,” even though the condition can develop at any age.

The press writer dismissed the inclusion of type 1 in the 11 percent total estimate as being insignificant, by saying that “most” diabetes over age 20 is type 2 (the body’s failure to use insulin efficiently). Of course, the word “most” (in this context) covers everything from 50.01 to 99.99 percent.

The only number I could find (at the Juvenile Diabetes Research Foundation International) estimated the total of American type 1 diabetes cases at 3 million.

Three million people is a lot. At the time the above estimate was probably made, the United States had about 300 million people. If 3 million of them had type 1 diabetes, that’s equivalent to 1 percent of the population — or roughly 9 percent of the nation’s 11 percent (over age 20) diabetes total.

So, we can approximate the total number of type 2 diabetes case nationally at roughly 10 percent of the population.

That’s a pretty prevalent disease, don’t you think?

However, disease prevalence among a general population does not itself necessarily say anything about our personal absolute risk of getting sick

This caveat is important, especially with diseases that seem to have quite a few risk factors that vary widely in severity among the general population.

For example, say I’m fat, inactive, and eat lots of sugar and carbohydrates each day. Won’t my absolute risk for developing diabetes be much higher than the one for my skinny, physically active, health food-eating neighbors?

All else being equal, yes.

Based on the Harvard School of Public Health press release, we don’t have enough information to calculate each individual reader’s absolute risk of getting diabetes — regardless of their red meat consumption.

“But if so many people have diabetes, isn’t the study worth paying attention to?”

Probably. (And I say that even without having access to the full study.)

That’s the power that a prevalent disease has in attracting our attention. If lots of people have it — and the distinctions in risk factors among them are not ironclad — then we can assume that our personal risk of getting the illness is not so miniscule that we can complacently ignore it.

For example, how fat is fat? When does a significantly elevated risk kick in? How inactive do I have to be, given my own physiology, before I’m inactive enough to be inviting diabetes through the front door? And so on.

That said, you might ask whether I would pay attention to this particular study — given my “wait and see” attitude (based on scientific grounds) toward the majority of other medical studies.

Yes. Even though I’m not fat, not inactive, and watch my carb intake. And if I were fat, inactive, and ate badly, I would pay even more attention.

The study’s:

(i) power of numbers (assuming acceptable statistical methods were used),

(ii) scientific reasonableness regarding the implied connection between highly-processed red meat and health,

and

(iii) probable absence of self-interested data distortion —

give its analysis a public health persuasiveness that most medical studies lack.

In regard to the last element, it is difficult for me to visualize who would pay for a study that was aimed at discrediting processed red meat in favor of a very generalized blend of already recognized more healthful food categories. The research-distorting greed factor is apparently absent.

“So if we take this research seriously, what else should we notice?”

These were comparatively small portions of processed meat. A pound has 454 grams. The 50 gram serving size associated with processed red meat is only 1.76 ounces, less than one-half of a “quarter-pounder.”

The moral? — For most of us, this “diet and diabetes” meta-analysis might be worth paying attention to

Of course, we rarely can know anything for certain.

But sometimes we know enough to take precautionary measures just because the data implies that, if we don’t, something really unpleasant might happen.

Diabetes is nasty. We should do our best to avoid getting it.

PeteFree.com

Eating Processed Red Meat Every Day Apparently Increases the Risk for Developing Type 2 Diabetes by 51 Percent — according to a Large Study by the Harvard School of Public Health