The Results Are In: Long Term Carb Cutting Works
July 17, 2008
Today in the New England Journal of Medicine a study was published that you are going to be hearing A LOT about because it could be the knockout punch that Low Carb diets have been waiting for.
Back in April I attended a nutrition conference in San Diego (you can read about my trip here). The most interesting part of the conference was a 2.5 hour session on carbohydrate restriction (science talk for low carb). The session was meant to highlight the controversy surrounding low carb diets but as leading low carb researcher Dr. Jeff Volek said to me that weekend (I’m paraphrasing)
It isn’t much of a controversy. The science shows low carb diets just work better.
And boy is he right. Unfortunately despite the ever growing pile of research surround cutting carbs, many people and organizations (USDA, American Heart Association, American Diabetics Association, leading nutrition researchers, policy makers, etc) have refused to truly acknowledge the effectiveness of carbohydrate restriction over a traditional low fat diet.
[Side note: You would think with the huge, and ever growing, obesity problem governing agencies would be willing to try some different recommendations as it is obvious that recommending a low fat diet isn't working.]
The one final straw that low fat diet lovers were holding on to in regards to fully acknowledging the effectiveness of low carb diets was the ‘long term study’ argument. Whenever low carb diets would be brought up you could almost time it on your watch how long it would take before someone would say
“Yes…well…there need to be more long term studies showing its safety first”
I normally have two reactions to that:
1. Where are the long term studies on ketchup’s safety?
2. But the biggest long term study (the American Public for the last 20 years) recommending low fat diets have failed miserably.
However, if you really sat down and looked at all the science, there was some truth to the long term study argument. Most carb restricted clinical trials (the heavily restrictive ones - Atkins style diets) only last 12 weeks (there are longer ones, I’ll briefly mention their failures below) - that’s not very long.
That brings us to today and the New England Journal of Medicine study. This study was 2 years long and
unlike other longer term low carb vs. low fat studies the adherence to the diets in this study was phenomenal - 95.4% at the end of one year (in a previous study adherence had been as low as almost 50%) and 84.6% at the end of the second year.
When all the assays were run, people weighed, and statistics analyzed the low carbohydrate diet came out on top in just about every way possible. Here’s a quick list in some of the areas that low carb out performed low fat:
- Greater Total Weight Loss
- Faster Initial Weight Loss
- Increased HDL (good) cholesterol
- Decrease in Triglycerides (almost 8x greater) - Triglycerides don’t get as much ‘media’ as a risk factor for heart disease but it is a very important marker to consider.
- C-Reactive Protein (CRP) - This is becoming an ever popular marker for heart disease and inflammation.
- Improved Liver Function
Pretty impressive list don’t you think?
You may be wondering if I am going to be a Atkins diet convert???
More on that tomorrow 
Have a good one,
Mike
P.S. If you want a carb restricted rapid weight loss plan then Warp Speed Fat Loss help you lose 10, 15, or even 20lbs before Labor Day. If you aren’t looking necessarily for rapid weight loss but want to lose weight and get many of the benefits discussed above head over to NakedNutritionGuide.com
Comments
8 Responses to “The Results Are In: Long Term Carb Cutting Works”
What do you think? Post a Comment and let us know!
-
Naked Nutrition Network 2.0
-
Subscribe To Mike's Blog RSS Feed
If I am reading the table in the study correctly, it appears the low carb group ate a macro ratio of 40/20/40 (carbs/protein/fat). Isn’t 40% not usually considered “low carb”?
[Reply]
Mike Roussell reply on July 17th, 2008 3:34 pm:
Lara,
Very good point…you sort of stole my thunder for tomorrow
This study does fail in the department of diet categorization. The low carb diet is not an Atkins style diet even those they say in the paper it is “based on the Atkins diet”. That doesn’t make any sense since they counseled the people in the low carb group to choose vegetarian sources of fat and protein (which isn’t very Atkins like).
The low fat diet group was 30% of calories from fat - which isn’t low fat either!
I’ll get into it more tomorrow.
-Mike
[Reply]
Lara reply on July 17th, 2008 4:01 pm:
Oops sorry about that
Looking forward to more info on this most interesting study.
[Reply]
An interesting study as studies go. I have a couple of issues with any study like this though. Food is greater than the sum of its parts. Being in a different country - their eating habits are naturally different than they are in the states. But to make this or any study useful I believe a couple of things must be documented.
1) The diet they are eating - what foods - any meds - supplements.
2) break down of processed foods vs. whole foods (fast food anyone)
3) restaurant dining or home cooked
4) daily journals of diet and exercise
5) total caloric intake per meal
All of this would be tremendously difficult of course, but with out it how can the study be beneficial to anyone?
The dietitian saw them basically every 6 weeks after things got going. The longer the study goes the more lax people become - human nature and all. How many people would be able to accurately record their eating habits even days earlier much less 6 weeks?
Then of course will people accurately record data when they break off the diet?
[Reply]
I think the most interesting point was the low-carb dieters didn’t have to count calories while the others we restricted calorie diets.
Yeah, looking at the table, the low-carb dieters didn’t achieve the low level of carbs the researchers instructed them to take (120 grams max per day after the inital 2-month 20 grams per day), but they did succeed getting their carbs down lower than the other groups which is probably what is important. Their lunch was the only meal that was carefully provided for them. The subjects had more freedom to mess up at home.
I speculate success with fat reduction is be somewhat proportional to how much carbs are cut with some people responding to carbs more than others. I think it is amazing that just cutting carbs a little more than the low-fat group made a substantial difference.
[Reply]
Mike:
This study raises many methodological/design issues, notably:
1. Is a 30% fat diet (avg US diet is 36%) really low fat? I suggest not; Ornish or Pritikin are truly low fat. The AHA diet this one is modeled after isn’t low fat, just lower fat. Also, one could also have labeled the Mediterranean diet “modified US diet” and been more accurate. The real story here is that the low carbohydrate diet yielded marginally better results than some variation of the traditional US diet.
2. Results from food intake questionnaires and most other data were only collected four times, including baseline- this study is basically four samples taken over two years. Self administered collection instruments are inaccurate.
3. Physical activity was assessed similarly, and converted to mets??? Why not convert to calories used per day etc?
3. Participants were stratified on the basis of statin intake, but not statin dose. Dose has an effect on lipids and related markers. Also, given the unique metabolic makeup of those with metabolic syndrome/hyperlipidemia etc, wouldn’t it make sense to stratify by some criteria there as well?
4. On page 230, the authors note that high attrition rates (15+percent) are a common limitation of dietary trials. Remember that attrition can skew a data set either way, not always in a negative fashion. On page 232, the authors give the rates for 24 month adherence:84.6% overall; 90.4% low fat (LF); 85.3% Mediterranean (MED); 78% Low Carbohydrate (LC). Draw your own conclusions.
5. Note in Table 1 the higher number and percent of smokers in the LF group, the higher number and percent of diabetics in the LC group, and the vastly higher number and percent of those on glycemic control meds in the LC group. This suggests that more of the diabetics in the LC group were medicated than in the other groups, which will definitely skew any lipid and glucose related measurements in the favor of the LC group.
6. Why was the LC group counseled to choose vegetarian sources of fat and avoid trans fats?
In addition, the authors seem to selectively present their data, and to ignore much of what they did present.
1. Table 2 does not tell us the baseline energy intake, only energy change from baseline. It does so in absolute number of calories, not percent, so we cannot determine how severe the caloric restriction was. (Note that on page 232 the authors state “there were no significant differences among the groups in the amount of decrease”, not percent decrease) The authors have the data (baseline was collected) but don’t present it. Any idea as to why? (Funded by Atkins)
2. Examine the macronutrient information in Table 2. The MED and LF diets reflect no change in macronutrient intake outside of decreased caloric intake (note the unfavorable p-values for most of the variables in the table). Are the diets that different from baseline (absolute calories and proportion of calories), or from each other?
3. Note the lack of confidence intervals supplied with much of the data. When it is supplied, it raises questions. For example, on page 235, the CIs for the mean wt change among women for the LF and LC groups both contain one, and also overlap. CIs that include one (the null value) indicate no difference. Also, I calculated CIs for mean wt change among those who finished the intervention and they are -3.371 to -2.09 (LF), -5.53 to -3.27 (MED) and -5.29 to -3.48 (LC). As you can see, these all overlap, suggesting no difference. (Many epidemiologists, biostatisticians and biomedical researchers pay as much, if not more, attention to CIs than to p-values.) These criticisms hold true for the mean wt change among men as well (-4.3 to -2.5 LF, -5.1 to -3.0 MED, -6.2 to -3.6 LC).
4. Note the use of standard error (SE) bars as opposed to CI bars in figures 2, 3 and 4. This is highly unusual, as the CI tells you much more about the data spread. Also note the overlap or near overlap of the bars on most variables. I would hazard a guess that, if you were inclined to calculate the CIs for this data, they would overlap and may contain one, all suggesting no difference.
5. In addition, the authors seem to be wed to statistical significance and avoid any discussion of clinical significance. For example, overall weight loss after two years was 6.3 lbs for LF, 9.7 lbs for Med and 10.3 for LC (Approx 3.1%, 4.8% and 5.12%, respectively- after two years). Are these results really that impressive, particularly given the widely accepted standard of 10% BW as significant weight loss?
This study is sloppily designed, to put it kindly, and the presentation and interpretation of data is questionable. In order to fairly test a LF, MED and LC diet it would be necessary to:
1. Use a true low fat diet, and a truly low carbohydrate diet.
2. Stratify for more (or different) variables.
3. Provide baseline measures for all variables involved
4. Provide CIs for the data
5. Address overlap (and excessive width) of CIs when appropriate.
Your thoughts,
PJ
ps-John Wallace is still doing great with YNNM and we are still applying alot of the warp speed training methods in his workouts.
[Reply]
Mike,
What do you think of this (another opinion on the study):
http://www.burnthefatblog.com/archives/2008/07/what_the_new_low_carb_study_re.php
Thanks!
[Reply]
The media has most definitely taken this research and turned it into their own media generating news story despite the inaccuracies of their stories.
The take home message from the study itself is that with some modifications in overall carb intake and other social support tools people can lose weight. All 3 diets were close in weight loss yet none of them really lost that much weight in 2 years to be impressive.
Just imagine if these participants were really counseled and given plenty of guidance to be strict with their respective programs. Maybe they lose 5x that amount.
[Reply]