Well I’d intended to get up this morning and write a run of the mill rant about the stupid article accompanying an interview the other day on NPR Morning Edition with my buddy John Durant. Here’s the link to the thing, along with 7 minutes of audio that’s substantially better than the article. There’s now 441 comments. When last I scanned through them, around 360 of them or something, it was about the most idiotic comment thread I’ve ever seen.
OK, so no rant this time.
I got wind of this study this morning via a link to Glenn Reynlds from a friend of mine, then chasing down rabbit holes. This appears to be the full text: Effects of Dietary Composition on Energy Expenditure During Weight-Loss Maintenance. I haven’t scrupulously picked through it but figured why not let everyone who want to go through it just like I will.
Here’s a couple of media reports for those wanting just a summary.
- USA Today: Low-carb diet burns the most calories in small study
- The New York Times: Which Diet Works?
And from the Abstract of the actual study:
Objective To examine the effects of 3 diets differing widely in macronutrient composition and glycemic load on energy expenditure following weight loss.
Design, Setting, and Participants A controlled 3-way crossover design involving 21 overweight and obese young adults conducted at Children’s Hospital Boston and Brigham and Women’s Hospital, Boston, Massachusetts, between June 16, 2006, and June 21, 2010, with recruitment by newspaper advertisements and postings.
Intervention After achieving 10% to 15% weight loss while consuming a run-in diet, participants consumed an isocaloric low-fat diet (60% of energy from carbohydrate, 20% from fat, 20% from protein; high glycemic load), low–glycemic index diet (40% from carbohydrate, 40% from fat, and 20% from protein; moderate glycemic load), and very low-carbohydrate diet (10% from carbohydrate, 60% from fat, and 30% from protein; low glycemic load) in random order, each for 4 weeks.
Main Outcome Measures Primary outcome was resting energy expenditure (REE), with secondary outcomes of total energy expenditure (TEE), hormone levels, and metabolic syndrome components.
[…]
Conclusion Among overweight and obese young adults compared with pre–weight-loss energy expenditure, isocaloric feeding following 10% to 15% weight loss resulted in decreases in REE and TEE that were greatest with the low-fat diet, intermediate with the low–glycemic index diet, and least with the very low-carbohydrate diet.
~~~
Alright, here’s some of my initial, off the cuff thoughts having just skimmed through. I believe I saw somewhere that the participants were eating food prepared for them by the research team; if so, that does add a bit of credibility. Also, if REE and TEE (resting and total energy expenditure) was less with the LF diet compared to the LC diet it seems more plausible to me that the LF, high-carb diet would generally be prone to more cheating, hence more caloric intake, hence an even more curious result if that were the case.
Here’s a blurb from the comments I found interesting:
The results of our study challenge the notion that a calorie is a calorie from a metabolic perspective. During isocaloric feeding following weight loss, REE was 67 kcal/d higher with the very low-carbohydrate diet compared with the low-fat diet. TEE differed by approximately 300 kcal/d between these 2 diets, an effect corresponding with the amount of energy typically expended in 1 hour of moderate-intensity physical activity.
The physiological basis for the differences in REE and TEE remains subject to speculation. Triiodothyronine was lowest with the very low-carbohydrate diet, consistent with previously reported effects of carbohydrate restriction23 ; thus, changes in thyroid hormone concentration cannot account for the higher energy expenditure on this diet. The thermic effect of food (the increase in energy expenditure arising from digestive and metabolic processes) dissipates in the late postprandial period and would not affect REE measured in the fasting state. Because the thermic effect of food tends to be greater for carbohydrate than fat,24 – 25 it would also not explain the lower TEE on the low-fat diet. Although protein has a high thermic effect of food,16 the content of this macronutrient was the same for the low-fat and low–glycemic index diets and contributed only 10% more to total energy intake with the very low-carbohydrate diet compared with the other 2 diets. Furthermore, physical activity as assessed by accelerometry did not change throughout the study. Alternative explanations for the observed differences in REE and TEE may involve intrinsic effects of dietary composition on the availability of metabolic fuels13 – 14 or metabolic efficiency, changes in hormones (other than thyroid) or autonomic tone affecting catabolic or anabolic pathways, and (for TEE) skeletal muscle efficiency as regulated by leptin.26 – 29 Regarding the last possibility, the ratio of energy expenditure to leptin concentration has been proposed as a measure of leptin sensitivity,30 and this ratio varied as expected in our study among the 3 diets (very low carbohydrate>low glycemic index>low fat).
Maybe I’m missing something here, but if the participants were eating isocaloric diets (just the different macro ratios), they were in weight maintenance (no participants gaining or losing), yet the LC participants had total energy expenditure 300 kcal per day greater than the LF participants, and…“Furthermore, physical activity as assessed by accelerometry did not change throughout the study”...then what’s the point?
What am I missing?
OK, perhaps I’ll put up an addendum or another post once I’ve gone through this more, but don’t let that stop any of your science, metabolism or science geeks in comments.
Update: Synthesis: Guyenet, Colpo, Calories Count, Food Quality Matters, Macronutrient Ratios are Qualitative (my extensive follow-up post on the matter)
“The low-fat diet pro-
duced changes in energy expenditure
and serum leptin42-44 that would pre-
dict weight regain. In addition, this con-
ventionally recommended diet had un-
favorable effects on most of the
metabolic syndrome components stud-
ied herein. In contrast, the very low-
carbohydrate diet had the most benefi-
cial effects on energy expenditure and
several metabolic syndrome compo-
nents, but this restrictive regimen may
increase cortisol excretion and CRP.”
Study Finds that Carbs Prevent Energy Use Readers here may also enjoy Barry Groves take on the findings.
The full text of the paper is online from which I see the length of trialImmediately before the 3-day inpatient hospital admission, the assessments under free-living conditions were conducted over 14 (total energy expenditure) or 7 (physical activity) days.
were probably insufficient to ensure those on a VLCD were fully ketoadapted
maybe there are details I’m missing as the eMethods link provided is not working for me.
OK, that’s great, Ted, but what I don’t get is that if these are isocaloric diets, if physical activity didn’t change and if weight remained the same (maintenance), then where did those 300 calories go and perhaps more importantly, why does it really matter.
So, I’m missing something or am very confused (both highly possible).
Don’t get me wrong, I think LC to MC is definitely the best way to get fat off and keep it off and feel great and not be too hungry, but I’m still confused about the study result.
did the vlc group lose weight during the trial?300 more calories”lost” should show up as continued weight loss,no?
See, cary, that’s what I can’t figure but as time permits am trying to find it. They say it’s crossover and isocaloric, so each participant ate the same calories daily on all three diets. They also say there was no difference in physical activity.
So, you have on all three diets:
Same calories
Same physical activity
300 more calories total energy expenditure on LC vs LF
Seems like that should come out to an average of 1/2 pound lost per week.
Actual differences in weight could have been too small to be significant. And since it was cross-over, anything lost in one might have been regained in the next.
I don’t think they can measure any significant weight change in 5 days. I’d assume that if they kept up these conditions for longer that the low-carbers would lose more weight, but that’s just conjecture. I’d assume they could only do it for 5 days because this part of the study was inpatient, correct?
Jared
Apparently, according to my email exchange with Dr. Ludwig cited below, weight was measured over the 4 weeks for each intervention (4 weeks each diet). It’s implied that there was more additional weight lost on the LC intervention but not statistically significant, but also “trending.” He said it would take six months on each to make a conclusive determination.
Stephan Guyenet tweeted this study and said he’d be posting an analysis of it this week. And I’m sure Anthony Colpo will have something to say about this, as the whole “metabolic advantage” idea is his area of expertise. Looking forward to those posts.
I emailed it to both Stephan and Anthony this morning (Perter too), though I’m sure they all had gotten wind of it.
one thing I noticed immediately is that the protein portion in the low carb diet was 10% higher than all the others. pretty sure protein has the highest “metabolic advantage”of all macros. why didn’t they keep it protein controlled at 20%?
The point was to study different compositions. The results suggest doing further studies, to try to isolate what’s happening in each case. Kinda frustrating, but I don’t entirely blame them. Studies are expensive.
I could study that abstract for the next 10 years and still not have a clue as to what it means.
Have you given the full text a look?
Yeah but I got a C+ in the “Chemistry For Idiots” course in college so ….
I totally buy the result, having been VLC for an extended period of time and then switching to 700-800 calories of starch a day. I think my resting energy expenditure was higher on VLC, probably because I was always short on glucose. Over the course of a day 300 calories isn’t that much.
Just sent the following email to the listed point of contact.
~~~
Dear Dr. Ludwig:
I’m a blogger who blogs regularly on weight loss and health issues surrounding the “Paleo Diet,” including issues involving Low-Carb diets.
I was interested in the study just published yesterday:
Effects of Dietary Composition on Energy Expenditure During Weight-Loss Maintenance
http://jama.jamanetwork.com/article.aspx?articleid=1199154
I have blogged about it order to generate discussion.
https://freetheanimal.com/2012/06/jama-effects-of-dietary-composition-on-energy-expenditure-during-weight-loss-maintenance.html
However, I have a glaring question I have been unable to answer after reading the study and looking at the figures and tables:
Where did the 300 calories go LC vs. LF? As I understand it, participants ate isocalorically in crossover fashion, so the only change was the macro ratios. It’s also stated that physical activity did not change. So if they ate the same calories, had the same physical activity, maintained the same weight loss throughout, then how are the 300 daily calories accounted for if not in additional weight loss beyond the initial loss?
Perhaps I could learn more from the eMethods supplement as cited in the article. However, the link to the PDF is broken ( ). If you have that PDF I’d appreciate it, and of course any light you could shed on my question so I can pass it on to my readers and commenters.
Anticipating that the answer might be something like: increased bodily metabolism (higher heat rate, respiration, etc., etc), then my follow on question might be what’s the point, the advantage?
Richard Nikoley
Free the Animal.Com
Looks like AndrewS in comments wins the prize. Quick response from Dr. Ludwig.
“The point is that 4 weeks isn’t long enough to translate a 300 kcal/d
difference into statistically significant weight change, especially when one
considers that body weight normally fluctuates by a kilogram or two though
the course of a week, based on differences in hydration status, the time of
the last bowel movement, etc. We’d need 6 months to reliably see this
effect. Nevertheless, there was a slight, and not statistically significant
difference in the hypothesized direction, with body weight highest on the
low fat diet (data included in the results section).”
A 300 kcal/d will become approximately 30 lbs. in a year, ceteris paribus.
The metabolic rate can also be different without (big) change in weight short term, when fat mass is traded for lean mass. Augmenting muscles, protein mesh in bones and conjonctive tissue are more expensive than adding fat in fat cells. I’m not saying that it was that that was measured in that study, but a lot of people have observed that in their own n=1 experiment with low-carb.
The link to PDF works from side-bar (Data Supplement)
It’s http://jama.jamanetwork.com/data/Journals/JAMA/24277/JPC120005_supp.pdf
Thanks Pat. Stephan emailed it to me a few hours ago but I haven’t had a chance to look.
Damn me. I looked on the sidebar yesterday too….
As far as I can tell, the patients were not weighed at the conclusion of the study, or that data was not given.
The pre-trial weight study is just an average with SD, and I don’t see anything about weight loss.
I think the authors of the study assumed that with an isocaloric diet no weight would be loss. Seems like a huge and silly assumption, but I can’t find any evidence that points to that assumption not being true.
i know you and evelyn aint heading off to the bar together anytime soon, but she does have a nice post about this one. and, wonder of wonders, you both ask the same question, “If TEE differed and feeding was isocaloric but they were weight stable, howzzat?” (she has a few more questions that are worth reading, too!)
http://carbsanity.blogspot.com/2012/06/modest-proposal-for-peer-review.html
Colpo weighs in
http://anthonycolpo.com/finally-a-study-that-proves-a-low-carb-metabolic-advantage-yeah-right/
My view is that switching from yes carb to no carb is physically stressful, and it is reasonable to expect physical stress to result in an increased resting energy expenditure, which is a good thing if your goal is to lose 50 pounds of unsightly fat … imo nothing Colpo or anyone else says disproves that.
Colpo is one of the smartest dudes I have ever read when it comes to nutrition – and he’s funny too
I’m just getting started drafting a post, synthesizing both Colpo’s and Guyenet’s takes on this.
I completely agree. I have been reading independent researchers for years, and Anthony Colpo is the most meticulous one. The same as Richard -and unlike many- he has changed his views through the years when the evidence dictated so, instead of blindly adhering to an initial dogma.
The only thing to do when you find yourself being stupid and ignorant is to stop being stupid and ignorant.
looking forward to it.
Ugh. I read through some of the comments, but just couldn’t take more than a few minutes of it.
I think I’m quickly going from wanting to spread the joys of a paleo/primal diet to all who would listen to not wanting to share it with anyone at all…and that bothers me. And yet, why should it? Why do I care whether or not others embrace it? (In fact, if the other side is correct that the lifestyle is not sustainable for the entire planet, it’s actually better for me if most of the people stay on the other side of the fence.)
Maybe I’ll try a modified approach. I’ll share it with anyone *once*, but will ignore the inevitable fight that follows. I really don’t want to be “that guy”, but I think it’s better than becoming the other “that guy”.
*shrug*
I suggest reading up on agriculture — Lierre Keith’s book or Pollan’s Omnivore’s Dilemma or any of the many anti-wheat books — make a strong argument that animal foods *are* sustainable, but not through monoculture farms and factory-raised animals.
The more that embrace this movement, the more widespread and available the food becomes, the easier it will be to find. I definitely want more people to come on over to the Paleo side of the fence.
No need for me to research further; I’m already convinced.
However, I’m not as convinced about your “increased demand will increase supply” argument. I’m reasonably comfortable with the availability of the food I want to consume right now. If it catches on mainstream, then we will be able to obtain our food from bigger players such as Walmart, Costco, etc…which will inevitably drive at least some of the smaller farmer/ranchers out of business, but it will be “more widespread and available” (and we will be at the mercy of the claims of the new larger corporate “farmers”).
Don’t misunderstand, I’m not a big corporation hater kind of guy…in fact, I’m generally pro corporation/pro economies of scale/pro efficiency, but as for my current food choices, I’m comfortable paying the “mom & pop tax” for the intangible benefits.
The planet isn’t sustainable because of over population and the need to ship apples 4000 miles. I’m with you on the preaching – its gets old fast.
Just hold on – this will soon turn into eat balanced carbs/fats/proteins – no grains/veg oils/processed crap diet. The bodybuilders got this right 100 years ago and still do. It would do everyone good to hit their nutritional forums – even if you don’t pound iron like those freaks.
Low carbohydrates probably force the body into higher utilization of gluconeogenesis in order to make blood sugar levels stable. That is an energetically expensive pathway compared with eating glucose, and I would bet that’s where the calories (at least some) went.
Thnaks for the discussion and you other guys for posting the links to the other discussions. I saw this tweeted yesterday and was wondering WTF also. My initial thought was to short of study to show trends lie someone else had mentioned.
I just wanted to mention that the CRP levels did not really rise with the low carb diet. They were just reduced less than with the other diets, but were still lower than the pre-wieght loss baseline. These are the numbers provided by the study:
Pre-weight loss baseline CRP: 1.75 (0.44 to 4.61)
Low fat diet CRP: 0.78 (0.38 to 1.92)
Low glycemic CRP: 0. 76 (0.50 to 2.20)
Low carb: 0.87 (o.57 to 2.69)
So, it is not accurate to say that the low carb diet “raised” CRP as some of the articles published on the media say. The levels are still lower than pre-weight loss, just nos quite as low by a very very small difference (anything under 1 is considered normal CRP)