Tuesday, October 17, 2006

Warning - Low-Carb Diets Bad for Diabetics - So Say Researchers!

Where do I even start with this one, published in the October 3 issue of Nutrition? Low-carbohydrate and high-fat intake among adult patients with poorly controlled type 2 diabetes mellitus. (abstract)

It's not really as study per se, but the findings of baseline dietary habits of subjects about to start a trial to investigate dietary intervention to control type II diabetes.

What caught my attention was the conclusion in the abstract, "This dietary pattern may represent a popular trend that extends beyond our particular study and, if so, has serious cardiovascular implications in this vulnerable population of T2DM patients."

What, pray tell, is the dietary pattern in question?

Well, if we rely on publication title and what the researchers tell us, it is a low-carbohydrate diet; not only that, it's this low-carbohydrate dietary pattern reported that is leading the way to cardiovascular complications down the road.

Abstracts like this one should have a warning: Beware, you're about to enter the assumption-spin zone!

You know me....never quite content with the abstract alone, I want to see the data!

How about we take a look?


Before I go through the data reported in the study, how about I first review the ADA's definition of what level of carbohydrate intake they consider a low-carbohydrate diet. You'll see in a moment why this definition is important.

From the latest update to the ADA position statement: Nutrition Recommendations and Interventions for Diabetes–2006:

Twice they state the minimum carbohydrate intake levels to consume. First in the energy balance section with "Low-carbohydrate diets (restricting total carbohydrate to less than 130 g/day) are not recommended in the treatment of overweight/obesity;" and again in the nutrition recommendations section with "Low-carbohydrate diets, restricting total carbohydrate to less than 130 g/day, are not recommended in the management of diabetes."

Even before they updated the Standards of Care this summer, their position statement published in January, Standards of Medical Care in Diabetes - 2006, stated "Low-carbohydrate diets (restricting total carbohydrate to less than 130 g/day) are not recommended in the management of diabetes."

So we can see the ADA is firm and clear in the belief that individuals, even those with type II diabetes, must consume at least 130g of carbohydrate each day; otherwise they are consuming a low-carbohydrate diet.

With me so far?

Now in the study above, the researchers specifically cite the same January 2006 position statement as their reference for their statement "Diets that provide low carbohydrate, low fiber, and high saturated fat contribute to disease complications in diabetes and are not recommended [12];" with reference 12 being Standards of Medical Care in Diabetes - 2006. Diabetes Care 2006; 29(suppl 1):S4-42.

I am making a point of this citation because the researchers state their objective in publishing this paper as "...the specific prevalence of low-carbohydrate diet trend in patients with T2DM has not been documented. Thus, the objectives of the present study were to examine baseline dietary, physiologic, and demographic information from adult patients with poorly controlled T2DM in an academic medical center."

So far, so good.

The researchers recruited 163 potential subjects from eight physicians who agreed to participate and help find subjects, along with the use of intranet messages and flyers to find patients willing to participate if the individual's physician cleared them to be screened as a candidate. Then these 163 candidates were included or excluded by telephone interviews, HbA1c screening and inclusion/exclusion criteria. The screening resulted in 40 subjects accepted and enrolled in the Diabetic Education Eating Plan Study.

The exclusion criteria included this curious reason to reject a subject, "Currently adhering to a low-carbohydrate diet such as the Atkins Diet or the South Beach Diet."

The reason given why this would exclude someone from enrollement in the study? From an intervention perspective, this exclusion was because these diets are low in all sources of carbohydrate, and modification of the type of carbohydrate (GI) will have a limited effect on glycemic load (GL) and therefore on HbA1c.

Okay, so those already eating low-carbohydrate diets were excluded because modifying their dietary carbohydrate within the study protocol for GI and GL of carbohydrates won't result in the data showing a benefit to HbA1c.

In fact, this is a very subtle admission that the dietary intervention the researchers are about to subject their participants to is not better for them if they are already following a low-carbohydrate diet.They excluded subjects already doing a low-carbohydrate diet because carbohydrate restricted diets are already low GI and GL; those individuals are already assumed to be benefiting from lower GI and GL, and increasing their carbohydrate actually risks the trial data in the future!

Basically, from what I can see, the researchers excluded these individuals because there is adequate data available that tells them if they modify the diet to be higher in carbohydrate, even though intake will be within the protocol definition of lower GI and GL, these folks will at risk to be negatively affected; and the potential for such would negatively affect the data outcome seeking to show benefit from dietary intervention in the trial.

So then, how did the researchers - after painstakingly excluding those following a low-carbohydrate diet - conclude their study subjects were eating a low-carbohydrate diet?

In their conclusions they state, straight out "All participants stated that they were engaged in dietary management of their diabetes, although they specifically stated they were not currently following on of the popular low-carbohydrate diets and therefore met study criteria"

They reported average intake was 159g carbohydrate per day, it did not fall below the ADA minimum intake of 130g, nor meet the ADA definition of a low-carbohydrate diet, yet these inconvenient facts do not stop the researchers from concluding this was a low-carbohydrate diet.

"Low-carbohydrate and high-fat intakes were observed at baseline among most participants with poorly controlled T2DM at our primary care clinic, despite the exclusion of patients following low-carbohydrate dietary programs such as the Atkins and South Beach diets."

If that wasn't enough, they continued, "Our obervations may be representative of many other patients with T2DM, and perhaps of a trend in the wake of the low-carbohydrate diets. Such a diet likely has cardiovascular implications for patients with T2DM, obesity, hypertension, and hyperlipidemia. Although many researchers are advocating this low-carbohydrate approach to diabetic management, more research is needed to determine the effect of this dietary recommendation on other macronutriets such as saturatred fat and fiber."

Oh, and it keeps getting better - "We found saturated fat intake to be more than twice that of the American Heart Association recommendation. We speculate that, when reducing carbohydrate intake to control weight and hyperglycemia, participants appeared to have replaced the energy they previously got from carbohydrate with energy from fat. Further, participants did not choose to replace carbohydrates with the monnounsaturated or polyunsaturated fats that have been shown to be cardioprotective. The ADA recommends a diet with less than 7% saturated fat content for people with diabetes. It is well understood that saturated fat is one of the main factors contributing to elevation of low-density lipoprotein cholesterol, which can increase risk of cardiovascular disease and overall inflammation."

On the one hand we have a population of subjects that specifically stated they're not following a low-carbohydrate diet and their reported intake of carbohydrate confirms they are eating more than 130g of carbohydrate a day as insisted upon by the ADA; then, on the other we have a team of researchers saying this dietary patten is a low-carbohydrate diet; who then go on to warn of dire consequences, even though their references and citations confirm their subjects are not eating a low-carbohydrate diet.

And this made it through peer-review?

This was accepted and published?

What the baseline data clearly shows is their study population is consuming a crappy diet.

But hey, don't let inconvenient facts deter a good assumption!

The fact is, one simply does not consume an average glycemic index [GI] of 80.7 (based on referent white bread = 100), and glycemic load [GL] of 133.62, if they're following a low-carbohydrate diet. No can do.

Add to that, a carbohydrate intake of an average 159g daily exceeds the ADA minimum of 130g per day!

But, the sad fact is, there is a lot of mileage to be gained if you can find a way to include, talk about and then discredit low-carbohydrate or carbohydrate restricted diets as a therapy for type II diabetes.

Just call whatever it is you find "low-carbohydrate," point out all the negative potential problems you can think of, cite and promote the ADA diet as ideal, and don't forget to ignore all data from the studies available that resulted in statistically significant findings of benefit - and, guess what? You've found the recipe for successful publication these days!

The researchers did not include any acknowledgement of the glaring, fatal flaw in their conclusion that the dietary pattern of their subjects was the cause their elevated HbA1c and concurrent complications. They assumed that it was because of the "low carbohydrate" diet, providing an average 159g of carbohydrate each day.

Why is this a fatal flaw?

The researchers did not investigate their subject populations progression of diabetes since diagnosis. That is, they did not seek to answer the question - are our subjects better or worse today than when they were diagnosed? They didn't seek to know if their subjects diabetes had progressed or improved since diagnosis - instead they assumed the HbA1c at 8.3%, hyperlipidemia, hypertension and other associated complications observed was a consequence of of the reported dietary patten instead of investigating their previous measures of HbA1c, cholesterol, blood pressure, etc.

They simply do not know if their subjects diabetes is worse or better than it was at diagnosis, or at any point previously in their history, because they didn't bother to look.

It was just so much easier to assume the observed complications and poor control were such because the of their dietary pattern.

Intellectual honesty demands we insist researchers accurately reflect their findings in their publications; it means we must not accept shoddy work or warnings of impending doom that are based on nothing more than assumptions.

Because the researchers did not review history or previous test results in their population (HbA1c, cholesterol, blood pressure) to determine if the reported dietary pattern is exacerbating the observed diabetic complications, it is nothing more than an assumption that the reported baseline diet is the reason for their complications.

Quite frankly the data included actually suggests that both quality and absolute gram intake of carbohydrate matters. But saying that would mean suggesting the ADA minimum intake of 130g is too simplistic and perhaps even too high. Remember, they excluded those already consuming a low-carb diet because they would not realize a benefit to their HbA1c in the study underway.

We know - this is something not argued within the diabetes community at all - that carbohydrates directly influence blood sugars; they raise blood sugars. As evidenced by the findings at baseline - when consumption of carbohydrates is such that glycemic index and load is frighteningly high, while absolute gram intake is moderate and above current recommendations, unacceptably high HbA1c (indicative of chronic hyperglycemia), hyperlipidemia, hypertension and other complications may be indicative of an overall poor dietary pattern in patients trying to control blood sugars while meeting ADA minimum intake of 130g or more carbohydrate each day.

But the researchers don't take this approach.

Nor do they acknowledge what they don't know - was this level of carbohydrate and fat, even with higher than desired GI and GL, resulting in improvement or progression of diabetes since diagnosis?

They don't know.

We don't know.

They didn't ask.

They should have.

Or they should have at least acknowledged it was premature to make the assumption the dietary pattern reported for carbohydrates and saturated fat was the cause of the complications observed at baseline in this population.

The bottom line is that this paper is an intellectual disgrace and impedes the progress of science. Perhaps, as a friend of mine suggested this morning an erratum needs to be included to say...the purpose of this study is to examine whether Nutrition is a properly peer-reviwed publication.


  1. A true disgrace. Shame, shame, shame on them...and the journal. Perhaps the desperate acts of desperate people.

  2. (A typo in your 2nd to last word.)

    This is so typical. What's kind of ironic is that some fields of scientific research, such as parapsychology, have had this kind of BS foisted on them ad nauseum and as publicly as possible for decades. It wasn't until Atkins that people in "traditional" science fields started realizing how much of science is literally nothing but propaganda.

  3. makes me wonder who paid for this study..........

  4. Yes, always ask yourself: for what reason is this report being published. Does anyone really believe that scientists have fewer vested interests than anyone else? Thanks for the article. Anyone interested in Hoodia Gordonii cactus for weight loss?

  5. Studies are absolutely useless. This is why it's wrong to quote any dietary study to support any position, whether it's pro low-carb or not, and should rather be based on life experiences and physiology. Shoddy research and questionable funding is always an issue.

  6. Let me guess, they determined dietary intake with Food Frequency Questionnaires? Am I just stupid, or did they report average caloric consumption of 1800 for people with average BMI of 35? Seems to me the folks should have been loosing weight. Given calorie deficits around 1000. The lack of critical thinking that passes for “research” is depressing.

  7. Were they doctors or spin doctors? When I did a unit called Basic Research Methods for my original; degree back in the late '70s, we learnt hot to construct and UNBIASED sample.

    Sounds like these pseudo scientists need to go back to school and do that unit!