Low Carb, Eco-Atkins Diet after 6 Months
In 2009, a 4-week clinical trial putting overweight people on a low-carbohydrate vegan diet, known as the Eco-Atkins, was released. I mentioned it in my article Of Oil and Ethics.
It was promising because after 2 weeks, LDL cholesterol levels went from 174 to 134 mg/dl. After 4 weeks, LDL cholesterol levels appeared to be stuck, at an average of 136 mg/dl. Participants also lost weight and had some other improvements, and they reported feeling more satisfied than did the participants in the control diet (a lacto-ovo, semi-low-fat vegetarian diet).
It took the researchers awhile to publish it, but they just released a report on what happened to the participants after 6 months (1). Unfortunately…not much. Their LDL cholesterol levels were back up to 157 mg/dl at the end of 6 months. They had still lost some body weight (15 lbs) and their risk for heart disease had improved over baseline.
The analysis was done on an intention-to-treat basis which means that people who didn’t stick with the diet were included in the final results (or their numbers were estimated). Overall diet compliance was fairly low at only about 34% of the recommended foods.
For those on Eco-Atkins, percentage of fat went up from their normal diet, but only from 34.4 to 36.0% of calories. And when you consider that their calories went down, from 1,840 to 1,388, their total fat intake actually went down, from 70 to 56 g.
The authors suggested that increases in monounsaturated fat (MUFA) could account for some of the improvements in heart disease risk factors, but their MUFA fat intake, while increasing slightly on a percentage basis, actually went down in total, from 27 to 23 g.
When you add in the fact that fiber increased from 12 to 21 g, it seems that all of the improvements could simply be attributed to a lower intake of calories and an increase in fiber.
Still, the Eco-Atkins did better than the 28% fat, lacto-ovo vegetarian diet. However, at baseline, the participants in the lacto-ovo group had an average daily caloric intake of 1,598, which was 242 calories less than the Eco-Atkins dieters. During the study, the lacto-ovo vegetarian dieters ate a very similar amount of calories to the Eco-Atkins dieters (1,347 and 1,388 respectively). As such, the Eco-Atkins dieters had a lot more room for improvement which could possibly explain why they did somewhat better.
In conclusion, this trial provides evidence that fiber is good and calories are bad for lowering cholesterol and losing weight.
References
1. Jenkins DJ, Wong JM, Kendall CW, Esfahani A, Ng VW, Leong TC, Faulkner DA, Vidgen E, Paul G, Mukherjea R, Krul ES, Singer W. Effect of a 6-month vegan low-carbohydrate (‘Eco-Atkins’) diet on cardiovascular risk factors and body weight in hyperlipidaemic adults: a randomised controlled trial. BMJ Open. 2014 Feb 5;4(2):e003505. | link
February 14th, 2014 at 10:34 pm
Hi Jack,
is the Low Carb, Eco-Atkins diet practicable for people who need a lot of calories, for example enduring-athletes?
I ask because I do a lot of cycling and need at least 3000 calories a day. I get about 20% from protein, 35 from fat and 45 from carbohydrates. My total cholesterol is 100.
February 15th, 2014 at 8:38 am
Hi Stefan,
If it seems to be working for you, then I would hesitate to change it. My only concern would be if you’re getting enough carbohydrate. I just got out the American Dietetic Association, Dietitians of Canada, and the American College of Sports Medicine’s 2009 position paper on sports nutrition and they say this about carbohydrate:
“For example, when energy intake is 4000–5000 kcal/d, even a diet containing 50% of the energy from carbohydrate will provide 500–600 g of carbohydrate (or approximately 7–8 g/kg [of body weight] (3.2–3.6 g/lb) for a 70-kg (154 lb) athlete), an amount sufficient to maintain muscle glycogen stores from day to day. … Conversely, when energy intake is less than 2000 kcal/d, a diet providing 60% of the energy from carbohydrate may not be sufficient to maintain optimal carbohydrate stores (4–5 g/kg or 1.8–2.3 g/lb) in a 60-kg (132 lb) athlete.”
They also say this about protein and endurance athletes:
“Nitrogen balance studies suggest that dietary protein intake necessary to support nitrogen balance in endurance athletes ranges from 1.2 to 1.4 g/kg/d. These recommendations remain unchanged, although recent studies have shown that protein turnover may become more efficient in response to endurance exercise training.”
At 3,000 kcal, it appears that your protein is probably good, but that your carbohydrate intake might be a bit higher (and fat lower) for better performance.
February 15th, 2014 at 7:02 am
>However, at baseline, the participants in the lacto-ovo group had an average daily caloric intake of 1,598, which was 242 calories less than the Eco-Atkins dieters.
That’s a beautiful example of chance bias in small randomized trials. The smaller the trial (especially N
> The analysis was done on an intention-to-treat basis which means that people who didn’t stick with the diet were included in the final results (or their numbers were estimated).
I support the use of intention-to-treat. I think it’s the only way. Data imputation, however, can also be fraught with problems if data are not missing at random (e.g. if patients who dropped out would have had worse profiles had they stayed in, in which case a “worst case scenario” analysis should be done).
February 15th, 2014 at 9:39 am
> I think it’s the only way.
Really? I’m more interested in what happens to people who actually follow a prescribed diet rather than what happens to a group of people who are a mixture of following and not following it. For public policy, I think intention-to-treat is very useful, but papers would ideally give both ITT results and then results of the people who followed the diet with a high rate of compliance.
Here’s an analogy. Let’s say we take 100 people and ask them to quit smoking and then we follow them for 5 years. After 2 weeks, 75% of them have gone back to smoking (I hear it’s pretty hard to quit!) and by year 5, it’s 95%. Are we to then take the data on health from all 100 people and then determine that quitting smoking does no good, when, in fact, most of the people just didn’t quit?
Maybe the Eco-Atkins researchers could have just asked the people to follow Eco-Atkins rather than follow a low-calorie Eco-Atkins (which was the case – I forgot to mention that in my post; probably should have), and a lot more people would have stuck with it.
In any case, based on this ITT analysis and the very low compliance, I have to come away from this study with just about no idea what following an Eco-Atkins diet does for someone. Not what I was hoping for when I started reading the study.
February 15th, 2014 at 11:37 am
If you don’t follow ITT, you end up with an observational study, not a randomized trial. Doing a “per protocol”, “adherence”, “on treatment” or “efficacy” analysis (all synonyms), means you subvert randomization and introduce healthy adherer bias. Then, the patients who are inherently healthier or better educated or whatever in the active group, or who, because of baseline characteristics are more likely to tolerate active treatment, are just those who remain in that arm. So what happens is you end up magnifying bias in your results, because you don’t know which effects are the results of the diet and which effects are the result of the fact that you have enriched the sample with healthier, better educated, more “with it” people (and eliminated everyone else). The “healthy adherer bias” is the reason that patients who are more adherent to placebo have better outcomes for virtually any disease ever looked at (including death!) than patients who are less adherent to placebo. Thus one can’t make any comment about the effects of the diet. The thing about ITT is that it is inherently conservative, but in large enough samples, dropouts don’t make as much of a diluting difference on the results – which is one reason to do very large trials.
This is why ITT has become the gold standard for reporting the primary analysis in journal articles – per CONSORT guidelines and editorial policies at various scientific journals.
I remain a die-hard ITTer because I know how lousy “on treatment” analyses are.
Another thing about ITT is that it replicates the real world, where many patients are unlikely to persist with diets in the long run. Thus it provides you with a good working average for what happens when you recommend a diet in clinical practice.
February 16th, 2014 at 10:09 pm
Dan,
Are you suggesting that the mere fact that someone is prone to stay on a diet will lower their LDL cholesterol and would have nothing to do with their actual diet?
> I remain a die-hard ITTer because I know how lousy “on treatment” analyses are.
How do you know this?
And what would be wrong with doing both and ITT and a high-compliance analyses?
I am still very much interested in what the Eco-Atkins diet would do *for people who are actually on it.*
March 4th, 2014 at 12:13 pm
Reading the “Diets” section of the paper, I found it interesting that the authors went into great detail about what comprised the low-carbohydrate diet (which also included supplements, apparently not all of which are delineated), but all they said about the foods in the high-carbohydrate diet was:
“The high-carbohydrate diet was a low-fat lacto-ovo vegetarian diet (58% carbohydrates, 16% protein, and 25% fat) using low-fat or skim milk dairy products and liquid egg whites or egg substitute to ensure a low–saturated fat and low-cholesterol intake.”
http://archinte.jamanetwork.com/article.aspx?articleid=415074
Seems quite biased to me.