Very Low-Fat for Diabetes? Not So Fast...

Sometimes I receive emails that are a bit amusing. Yesterday, I received a number emails asking if I was up to the task of reporting on the results from a study mentioned in Sally Squires' column, the Lean Plate Club, in the Washington Post.

The article, 'Good' Carbs To the Rescue, highlighted findings published in the May 1 issue of the American Journal of Cardiology in the paper, Comparison of Coronary Risk Factors and Quality of Life in Coronary Artery Disease Patients With Versus Without Diabetes Mellitus. One email asked if I truly believed in reporting the findings from studies contrary to my own beliefs about diet and nutrition.

To answer that simply - yes, I can, I do and I have in the past in this column!

With regard to this particular study - I hadn't seen it previously and can say after reading through the paper, I did find the results intriguing, yet lacking in the detail expected when subjects with diagnoised diabetes are included in a study that specifically modifies diet as part of the design.

In the above study, researchers sought to investigate the effect of dietary modification in those with coronary artery disease (CAD), some of the subjects included also had diabetes. The subjects included 440 individuals - 347 men (55 with diabetes) and 93 women (36 with diabetes). All were placed on a very low-fat vegan/vegetarian diet (target 10% or less calories from fat), provided weekly counseling and support, and advised to exercise and practice stress management daily as they were followed for one-year.

For the purpose of this article, I am going to focus on the subjects with diagnoised diabetes since the article written by Ms. Squires focused on this group in her opening sentences - "People with Type 2 diabetes are advised to limit carbohydrates because of worries that too many carbs could overtax the body's dwindling insulin production and lessen its ability to process glucose...Now some scientists are asking if a very-low-fat diet rich in healthy carbohydrates -- whole grains, beans, fruit and vegetables -- might be another option."

Over the period of one-year, the researchers tracked changes in BMI, weight, cholesterol levels, blood pressure, exercise capacity, changes in medication requirement, and physical and emotional health.

Unfortunately, the researchers did not track, or at least did not publish, changes in fasting glucose, glucose tolerance, insulin levels, HbA1C or other risk markers considered important to evaluate in those with diabetes.

This lack of detail makes the data that is published in the study much more difficult to evaluate for efficacy in those with diabetes.

As I noted, the researchers did publish details about the number of subjects taking diabetic medications at baseline and at one-year. Of the 91 followed for the year, 5 dropped out so we have no data for their medication changes; 62 experienced no change in their medication regiment; 12 discontinued their diabetes medications; 6 no longer required insulin but did require an oral antiglycemic; and 6 experienced a worsening of their condition that required medication not taken at baseline.

Why lacking details are important:

• Of the 62 who did not change medication, 16 started with no medication and remained free of diabetic medication requirements. That's a good thing and I believe this sub-set should have been reported as a separate group.
• Of the remaining 46 reported as "no changes" in medication, we are provided no detail about dosing requirements! While their medication regiment (oral or insulin) may not have changed, we have no details as to whether their dosage increased, remained the same or decreased during the period of one year. This data is critical to understand if this type of dietary intervention is slowing progression of the disease or not.
• In the 6 subjects who experienced a worsening of their condition we are also not provided details to better understand why their condition may or may not have been made worse by the diet and/or lifestyle interventions of the study.

In her column, Ms. Squires writes, "Blood levels of low-density lipoprotein (LDL), "bad" cholesterol that raises heart disease risk, dropped. Levels of protective high-density lipoprotein (HDL) didn't drop, and unhealthy fats known as triglycerides didn't rise, as some researchers had feared."

However, when we read through the data, the reduction in LDL cholesterol did not reach statistical significance in those with diabetes; HDL remained unchanged in the diabetic women and declined slightly in the diabetic men (not statistically significant); and the triglycerides declined slightly in the diabetic men and rose in the diabetic women (not statistically significant), remaining well above the "high risk" target of 200mg/dl in both the men and women.

The last piece of critical data missing is details about the diet composition. We are provided only one small piece of data - the fat content, as percentage of calories. The fat content was 14.2% in the diabetic men and 19% in the diabetic women when they started the dietary modification (baseline). This was significantly modified in the trial - at one-year the men consumed 6.4% of total calories as fat and the women 8.9% of calories as fat.

What we don't know is just as important - how many calories did they consume at baseline and after one-year? How much protein? How many carbohydrates? How much fiber? What type of carbohydrates? What did the micronutrient profile look like at baseline and one-year? What did they eat to achieve the consumption pattern? How did calorie intake and composition change from baseline to one-year? There are more unknowns here than I believe is acceptable in a study modifying diet!

My last concern with the study is that it was a multi-faceted intervention: diet, exercise, stress management and intensive support. On its face, these modifications are all considered good interventions and worthy of investigation in a controlled trial. The problem lies with teasing out which ones were meaningful in the results. Was it the diet, the exercise, the reduction of stress, the support, the weight loss from diet and exercise, or a combination of two or more? We don't know because each potentially plays a role and each is considered a "confounding variable" if not controlled for specifically.

While I applaud the researchers for convincing a good number of individuals to stick with this trial for the year - a very low-fat vegan/vegetarian diet is extremely difficult to follow faithfully - the fact remains the published data remains lacking in sufficient data and control that could have been included.

Noticably absent is critical data for those with diabetes - levels of glucose, insulin and HbA1C; glossed over is the fact that in those with diabetes, LDL was not reduced significantly, HDL remained low in both men and women and men had a minor decrease in the year, and triglycerides remained at levels considered "high risk," and actually rose in the women with diabetes.

The lack of statistically significant improvements in those with diabetes means we should look at this study with caution because what we have here is study that is not as impressive as we're led to believe in the media reporting of the findings when we actually go to the data and read through the findings ourselves.