Shai and Diabetes

I thought I would post the following graphic from the Shai Study that compared Low Fat, Low Carb and Mediterranean diets over 2 years.  Just to refresh regarding weight loss, when one compares LC to MDTN, at around 1 year the average weight loss was comparable and essentially the same at the 2 year mark.  Therefore the changes in diabetic markers are (at least on average) independent of weight loss.  As a limitation, although there were roughly 100 participants in each group, only 11-13 of them were T2 diabetics, so the sample size here is quite small.  Also, to refresh, the LC group reduced carb intake by around 125g and caloric intake by ~550 cal on average while the MDTN group's intakes were reduced by around 45-50g and 350 cal respectively.


The legend doesn't give stats for the non-diabetics, but there was no difference between the three diets for FBG, there does not appear to be much difference for HOMA-IR (measure of insulin resistance), but a fasting insulin does appear to have been significantly reduced by LC.  I'll leave that result for another post.

Many people believe carbohydrate restriction is superior for treatment of T2 diabetes.  Remember, the purple triangles are LC and the orange squares are MDTN.  At the 2 year mark the MDTN diet beats LC for all three markers.  What MAY be going on is that with improved insulin sensitivity, the MDTN's produced less basal insulin, and the insulin they do produce is "seen" by the liver suppressing gluconeogenesis resulting in lower fasting blood glucose levels.  I wish they assessed NEFA's!

Comments

Fred Hahn said…
Carbsane you said:

"At the 2 year mark the MDTN diet beats LC for all three markers."

Here's what the LC diet was:

"Low-Carbohydrate Diet
The low-carbohydrate, non–restricted-calorie diet aimed to provide 20g of carbohydrates per day for the 2-month induction phase and immediately after religious holidays, with a gradual increase to a maximum of 120g per day to maintain the weight loss. The intakes of total calories, protein, and fat were not limited. However, the participants were counseled to choose vegetarian sources of fat and protein"

So they increase the carbs to a level of carbs that no low carb advocate would consider low carb. Nice.

Then they tell the LC subjects to eat vegetarian sources of fat and meat, meaning, more carbohydrates and processed foods like soy burgers and the like. No LC advocate would EVER eat like this.

Notice in figure 2 that as carbs gradually increased, so too did weight gain. The LC diet was winning the weight loss race by far up until month 6.

HDL was best in the LC group but ceased to skyrocket upwards at month 6 - when carbohydrates started increasing.

Triglycerides decreased the most in the LC group until carbs started getting added in - right at month 6!

LDL rose in the LC group that is until carbs started to get added in at the - you guessed it - the 6 month mark. LDL cholesterol often increases in response to a VLCD because the particles are changing from small dense to large buoyant making for a better LDL profile even though total LDL rises.

And the ratio was also the best in the LC group and would have been even better had it not been for - you guessed it - month 6.

My guess is had they calorie restricted the LC group or allowed the MED non restricted calories and certainly of the LC subjects actually ate a low carb diet (<60g carbs per day) you'd see a very different picture emerge.

This is not a good study to use as evidence that a LC diet is not always the best choice or the only choice to combat diabetes.
CarbSane said…
Carbs were added in at the TWO month mark. Month 6 was when the regain began.

The carb numbers in the dietary reportings are for DEFICIT from baseline. Therefore from 6 months on, LC slightly reduced carbs if they are to be believed. You don't have to guess, the caloric intake is reported. Again, they are to be taken with a grain of salt, but Gary uses this study to support his theory that all diets work (when they do) because they restrict carbs. Not true, again, by his own "evidence".

The Med group restricted carbs and calories the least and had the best results for those with the most "deranged" metabolisms.

You really should let Gary know that the low carbers in this study weren't really eating LC! LOL
Patrick said…
Hey CarbSane... for the most part, I appreciate your contrarian viewpoint and I think you keep the "low carb groupthink" down, but I think you're missing some perspective on this study as it applies to diabetes (possibly because Diabetes isn't so far as I know your main focus.) Notice the following text:

"Among the participants with diabetes, the proportion of glycated hemoglobin at 24 months decreased by 0.4±1.3% in the low-fat group, 0.5±1.1% in the Mediterranean-diet group, and 0.9±0.8% in the low-carbohydrate group. The changes were significant (P<0.05) only in the low-carbohydrate group (P=0.45 for the comparison among groups)."

Glycated hemoglobin (commonly referred to as "A1c") is a measure of your average blood sugar over a 2-3 month period, and many of us would argue that it is the *most* significant marker of overall diabetes control in the absence of detailed glucometer logs.

The other thing I noticed is that, if you look at the graph, fasting BG was comparable in the LC and Mediterranean groups until the 2 year point, at which time LC shot up for some reason. There are a number of reasons why this makes me kind of suspicious. First, the small sample size. Second, the fact that low carb had the lowest A1c. Third, the fact that all it would take would be one person in the low-carb group to go on a binge the night before to throw the fasting BG off.... or be sick, or be on prednisone, or you name it. There are a multitude of things that can cause a single high fasting BG.

The sample size is just too small for conclusions to be drawn, and fasting blood glucose in diabetics is famous for bouncing up and down without explanation. To me, the A1c is just much more credible.

Moreover, I would argue that glycation of hemoglobin (i.e. A1c) is a better predictor of diabetes complications -- which are, after all, primarily caused by glycation of other tissues -- than is fasting blood glucose.
CarbSane said…
Hi Patrick,

Thanks for your comments. I've actually done a ton of research on diabetes because a main line of research (free fatty acids) I got started down leads to a whole wealth of diabetes research.

I would agree that A1c is a better measure of glycemic control for most, but I don't see it as a "trump card" in this study. In her recent interview with Jimmy Jenny Ruhl discussed having a relatively normal A1c and FBG but soaring postprandial BG levels.

A1c can be an effective measure of overall glycemic levels, but the other three measures where Med outperformed LC are those that are reflective of the degree of a person's degree of insulin resistance.

Basically I see the narrow view of diabetes as merely blood sugar levels as ... well, a narrow view. The consequences of diabetes extend past glycation and this is what the fasting levels indicate. Insulin resistance itself is a CVD risk factor with or without hyperglycemia.

I would also doubt that one binging LC'er could throw the average FBG for that group off 30 points. That is quite a remarkable difference for the Med group. Meanwhile the HbA1c CHANGE (from baseline) is significant only for LC, but it is also NOT significantly different between LC vs. Med, or any other between group comparisons.

I keep coming across (or being alerted to) more and more cases of diabetes being cured -- as in restoration of normal insulin secretion and glucose tolerance -- by methods other than carb restriction, certainly not to anywhere near that advocated as the optimal. While Shai was not a VLC diet, we still see that there appear to be somewhat diminishing returns over the long term with this approach. I believe these are due to the fact that carb restriction only "treats" one aspect of the underlying metabolic problem, and absent weight loss, exacerbates the other.

Otherwise, how does one explain the remission of diabetes in some instances -- in some cases despite returning to a "normal" diet -- while VLC'ers seem to have to watch carbs more and more closely and become more and more intolerant as time wears on, often despite maintaining considerable weight loss?