I’ve got to apologize in advance for the length of this post, but in order to thoroughly do what needs to be done, it took the space.
Readers of this blog who have been around for a couple of years have been through the Anthony Colpo (AC) fiasco with me. For those of you who weren’t around at the time, I’ll give a brief – a very brief – overview of what happened so you’ll understand what this is all about.
I wrote a post in September 2007 describing two different diets and their outcomes. The first was designed by Ancel Keys and was a 1500+ calorie low-fat, high-carb diet; the other, designed by John Yudkin, was a 1500+ calorie low-carb, high-fat diet. The subjects following the two diets experienced drastically different results.
This post, for whatever reason, inspired AC, a trainer and self-taught nutritional guru from Australia, to go into mad-dog attack mode. I wasn’t the first person he had gone after, but I became the first to fight back.
Around the same time AC took it upon himself to attack me, he had just published an online book on weight loss that he was beginning to promote called The Fat-Loss Bible. A more cynical person than I might have thought AC picked this fight in an effort to get some free publicity for himself and his book. If that was indeed his motivation, he may have gotten a little more publicity than he had bargained for.
I took a look at his book – which I hadn’t realized even existed prior to this kerfuffle – and found it to be much like the ad for the educational software pictured above to the left. At first glance, it looked reasonable, but upon closer inspection, it had some problems.
I made the offer to readers to dissect AC’s book if that’s what they wanted. Or I could ignore the whole thing and continue with my regular posting. A majority in the comments section voted for me to dissect. I dug into the book, pulled all the papers cited, but subsequently got involved in other stuff and forgot about AC and his book. He more or less dropped from sight, but has surfaced lately. I had forgotten all about him, his book and the whole situation, but his new antics have stirred a few readers to ask about the dissection that I promised but never came through with.
So, with that preamble, here it is.
The crux of AC’s objection to me (and a few other people, namely Gary Taubes, Richard Feinman and Gene Fine) is that I (and they) believe there is a metabolic advantage that becomes manifest during low-carb dieting. AC has taken the position that my idea of the low-carb driven metabolic advantage means that people following low-carb diets can eat all the calories they want and lose massive amounts of weight as long as they keep their carbs reduced. He accuses me of leading people astray by encouraging them to eat, eat, eat as long as carbs stay low.
I don’t know where he got this idea because I have certainly never said such a thing anywhere. The metabolic advantage brought about by low-carb dieting is probably somewhere in the neighborhood of a 100-300 calories, which isn’t all that much. This few hundred calories don’t even come into play until the 1500-2000 calorie range of consumption. I’ve written about this numerous times and have always used these figures, so, as I say, I don’t know where the idea that I believe the metabolic advantage allows low-carb dieters to eat huge numbers of calories and still lose weight.
I don’t plan to go through The Fat-Loss Bible in its entirety or this post would take on the dimensions of War and Peace. I’m going to limit my comments to Chapter 1, titled “Myth 1: Don’t Count Calories.” This first chapter is the one that tells why AC so fervently believes there is no metabolic advantage.
AC sells his book online, but (at least the last time I checked) it can be downloaded only on a PC. At the time this dispute started I had a PC, which I used to download the book. Since then, my PC has given up its ghost and I now use Macs exclusively. So, the copy I have is about two years old. I don’t know if AC has changed it since; consequently, I don’t know if my critique applies to the book as it exists today. AC changes his book all the time, updating here and there, and I don’t blame him for it. I do it with this blog all the time. I find typos in old posts and sentences that I don’t like. I change these things all the time and the blog is the better for it, so I don’t blame him if he does the same thing. But I just want everyone to know that I’m critiquing the book as it was when he launched his attack.
AC firmly believes that a calorie is a calorie is a calorie. He believes that people lose the same amount of weight dieting irrespective of the composition of whatever diet they’re on. He believes that a given person will lose exactly the same amount of weight on, say, a 1600 calorie diet whether that diet is a low-carb diet or a low-fat diet or any other kind of diet. It is the calories that set the weight loss, not the macronutrient composition or any other factor.
I don’t know if AC came to this conclusion then went looking for studies to confirm his bias or if he came to this conclusion because of the studies he read. The first chapter of his book contains a number of studies he trots out to ‘prove’ his idea that only calories count.
There have been many out patient studies that have shown a metabolic advantage and many that haven’t. Overall a greater number of studies demonstrating a metabolic advantage exist than studies showing no such metabolic advantage. The first part of the first chapter of The Fat-Loss Bible goes into great detail describing why such studies are worthless. He makes a fairly plausible argument as to why people on low-carb diets might tend to overreport consumption while those on low-fat diets may underreport. If correct, this difference in reporting would create the appearance of a metabolic advantage where none exists.
To solve this problem, AC turns to what he calls
strict ‘metabolic ward’ studies in which, for the entire duration of the study, the participants are confined to a research facility where they can only eat the foods supplied by the researchers.
On the surface this seems to make sense. Put the subjects under lock and key, give them just the food you want them to eat, and see what happens. You’re going to have some individual variation, but if evaluate enough subjects and they all end up losing the same amount of weight irrespective of macronutrient composition, then you’ve got some pretty good evidence that there probably isn’t a metabolic advantage.
But as obvious as this appears at first glance, there are problems with this approach.
The first problem is a problem of measurement. Newton derived his gravitational laws and everything scientists measured obeyed them. These laws became sacrosanct. If some observation didn’t conform to Newton’s laws, then the observation was faulty because Newton’s laws were infallible. Those quirky movements of planets way out on the edge of the solar system were off a little from Newton’s predictions, but, hey, it’s got to be a measurement error somehow. Then Einstein came along with his theory of relativity, and all the weird deviations conformed to Einstein’s laws. Newton had been superseded. Because the caloric differences brought about by a metabolic advantage (at least as I see it) are so small, weighing subjects in pounds and kilograms may miss it.
That’s the first problem. But there is a problem much greater than that. One that AC isn’t aware of because he doesn’t really have any real-world experience in doing nutritional studies in a hospital.
When subjects are studied in ‘metabolic wards’ they aren’t locked away and under constant observation. In fact, often enough, they aren’t even in a hospital at all. A ‘metabolic ward’ is simply a part of the hospital set aside to do nutritional studies. And often it isn’t even a specific part of the hospital. Subjects can be scattered about among the other patients. Subjects can have visitors, can roam through the hospital, can even go to the cafeteria. A ‘metabolic ward’ study can mean anything from: careful observation; to check into the hospital for a couple of days; to get trained on the diet then follow it at home; to check in, go to work all day, then come stay in the hospital all night. They are definitely not the strictly-controlled studies AC thinks they are. He confuses them with ‘metabolic chamber’ studies, which are a horse of a different color.
The opportunities to cheat in a ‘metabolic ward’ study are, for the most part, as great as the opportunities to cheat in an outpatient study, especially since many of the subjects are outpatients most of the time. There is a difference though. When people are on outpatient studies they are more likely to at least admit their cheating and record what they cheat with than they are in ‘metabolic ward’ studies. Some of the studies AC sites are formula diet studies in which shakes made of specific caloric and macronutrient composition are provided to subjects throughout the day. (Or are given to them to consume outside the hospital at work or wherever.) These are the kinds of programs you wouldn’t want to report cheating on. And these subjects do without question cheat. The fact that the data is reported as coming from a ‘metabolic ward’ study gives it a veneer of accuracy that it doesn’t really deserve.
AC gathered up a bunch of these ‘metabolic ward’ studies – 17 to be exact – that he uses to prove his point that there is no metabolic advantage and that only calories count. He lists these studies in a chart (reproduced below), then proceeds to go through them one at a time.
On the ones that confirm his bias, he spends little time. Just a brief description typical of this one describing the first study.
In a paper aptly titled ”Calories Do Count”, Kinsell and co-workers admitted five obese subjects to a hospital metabolic ward, then fed them liquid formula diets. The diets ranged in protein content from 14 to 36 percent, fat from 12 to 83 percent, and carbohydrate from 3 to 64 percent. The calorie content of the various diets was held constant for each patient irrespective of diet composition. As they switched from one diet to another, each patient continued to lose weight at a similar pace. Concluded the researchers: “…it appears obvious that under conditions of precise consistency of caloric intake, and essentially constant physical activity, qualitative modification of the diet with respect to the amount or kind of fat, amount of carbohydrate, and amount of protein, makes little difference in the rate of weight loss. [Italics in the original]
This is a great study to start with because it contains many, many flaws that AC is blinded to by his own confirmation bias. It’s a terrible study. Let me show you why.
Here is the first paragraph of the study. And I’m not kidding. This is directly quoted from the paper.
The accumulation of excess adipose tissue is a malady which affects many people. That undue preoccupation with the pleasures of the table contributes to the disease has geen [sic] generally accepted in most quarters; or, to express the matter differently, majority opinion has held that the first law of thermodynamics applies to the human machine quite as predictably as it does to inanimate machines. Despite this body of “official opinion” one finds many obese individuals who are either convinced that their food intake completely fails to explain their adiposity, or who spend time and money in the search for the magic potion or pill which will enable them to consume food in any quantity but still maintain or achieve a slim figure.
Do you think there might be just a little bias in this author and his co-workers? From this first paragraph one sees by the reference to the first law of thermodynamics the set of the sail of these researchers. Plus it’s pretty clear that these researchers don’t like overweight people and think obesity comes from a “preoccupation with the pleasures of the table…” How do you suppose their data is going to turn out?
First of all, were these five subjects inpatients in a metabolic ward or did they just pick up their formula and take it home. Did the live in the hospital or just spend the night? No information is given.
Here is the sum total of the information given on the ‘metabolic ward’ status of the first patient described:
His weight on admission to the metabolic ward was 270 pounds.
Was he admitted to the ward where he stayed full time for the full 70 days of the study? I doubt it, and I’ll describe why in a bit. Or was he admitted for his initial workup then released to continue his diet at home. I suspect the latter. Whatever the situation, this is all the study says about it.
Here are the descriptions of how the rest of the subjects entered the study:
Weight on admission to the study was 227 1/2 pounds…
At the time the study was undertaken her weight was 199 pounds…
At the time the study was undertaken, her weight was 211 1/2 pounds…
Patient GTAY was a 61 year old white female with a history of diabetes for more than 20 years. She had received insulin in the past but could be maintained in a satisfactory diabetic control with diet and tolbutamide. Milky fasting plasma was discovered in July 1962. Other findings included evidence for coronary and peripheral atherosclerosis, and diabetic retinopathy. She had partial removal of a goiter 40 years ago, but was essentially euthyroid during her stay in the metabolic ward.
The study in this patient was actually directed toward evaluation of her hyperlipidemia, but she is included in this report since she was maintained on quantitatively constant, eucaloric regimens containing high fat and high carbohydrate respectively, and also received both saturated and unsaturated fat.
This last patient wasn’t even accepted into the study as a subject for a diet study but more or less added after the fact.
There were five subjects in this study that lasted for anywhere from 65 to 77 days. We can’t really tell which subjects went how long. Nor can we really tell if it was an inpatient study or just one where the subjects checked in. Nor do we know how much weight each lost over how long a period. We know the starting weights and that’s about it.
The data as displayed looks like data collected in an inpatient study, but the paper itself only implies that it is. As you might imagine, inpatient studies are tremendously expensive, and, consequently, authors tend to make sure readers of the study know they are inpatient studies. In this paper, we have to guess.
If these are truly inpatient studies for 65 to 77 days, we need to address another point: the quality of the subjects in such studies. Who do you know who would have the time or inclination to spend two to two and a half months in a hospital full time? People who are willing to spend the time in such facilities are usually not the most reliable. They are typically unemployed with little education and, for the most part, are imbued with a lack of understanding as to how important their rigid adherence to the protocol truly is. I will be the first to say that not everyone who has ever volunteered for such a study falls into this category, but, unfortunately, many do. I’ll let a couple of the authors of these metabolic ward studies expound on this fact a little later.
The age range of these subjects is from 25 to 61. All of the subjects in this trial save one have serious medical problems and are under treatment with multiple drugs. The one who doesn’t have serious problems is a 25 year-old male who has “been grossly obese since childhood.” These are not the subjects you would want in a study of this nature.
The subjects getting the most calories got 1200 per day while those getting the least consumed 800 calories per day. As I’ve written before, if calories are kept ultra low, all the calories – irrespective of composition – are going to be used for energy. And under those circumstances, you would expect there to be no metabolic advantage. And you would expect weight loss to pretty much follow a trajectory driven solely by caloric deficit, which is pretty much what happens in this study. But it’s difficult to tell because of how terrible this study is presented. There is a starting weight, but no ending weight for the subjects. And, although the Methods section reports that the study lasted from 65 to 77 days, my calculations based on the data provided shows the study lasted from 64 to 82 days. Which are we to believe? Without an ending weight for the subjects and a precise number of days under caloric restriction, how do we really know how much they lost verses how much they should have lost given the number of calories they were getting?
And we have this other little tidbit thrown in when discussing the results of one patient, RTEA, who was a 26 year old female with “a history of resection of a cystic chromophobe adenoma of the pituitary…followed by radiation”:
Rate of weight loss was greater during the last 2 weeks on the high fat, high protein intake than during either of the other 2 dietary periods. This probably does not have significance on view of the “stair case pattern” of weight loss.
Say what? So they do have a subject that shows greater weight loss (and late in the program rather than early), yet they toss off the data with a bunch of weasel words implying that it probably isn’t significant.
I suggest you pull down the full text of this study at the bottom of this post so you can see for yourself how terrible it is.
I’m certainly not going to go through all 17 of the studies in this fashion because this post would then truly gargantic, but I wanted to go into this one at length to show that so-called ‘metabolic ward’ studies, those AC terms the ‘gold standard’ of medical research can be very, very flawed. I, for one, would not want to be making any categorical statements based on the data contained in this study we just evaluated, that’s for sure. If AC weren’t so blinded by his own confirmation bias, he would have laughed this study off. If I had used it to ‘prove’ a metabolic advantage – based on the one patient described above who had more weight loss on the high-fat diet – he would have had a field day.
Next, let’s turn our attention to the Liebel et al study. It’s number 11 down the chart if you’re counting. Here’s what AC says about it:
Leibel and co-workers took 13 subjects, determined how many daily calories each needed to maintain his/her weight, then proceeded to feed them, in crossover fashion, diets differing in their macronutrient content. Despite wide variations in protein, fat, and carbohydrate intake, the subjects maintained their weight irrespective of diet type. This included two subjects who followed low- and high-carb diets (15 percent and 75 percent carbohydrate, respectively) for a minimum of 34 days each.
That’s it. That’s AC’s commentary on the study. I suppose readers are meant to believe that this study showed that it was all a matter of calories with no difference in terms of weight lost versus macronutrient composition of the diet.
The Leibel et al paper is a great one because it shows just how sloppy AC is in his presentation of data and, no doubt, in his own evaluation of the medical literature.
Go back and reread AC’s description of how the study was done. Looks like Leibel et al did a hands-on study of these subjects, right. Well, that’s not exactly how it worked. Here is what really happened as reported by Leibel et al:
The records of all subjects studied by the Lipid Laboratory of the Rockefeller University Hospital between 1955 and 1965 who were fed lipid-formula diets of various carbohydrate (CHO) and fat composition were reviewed.
Leibel et al didn’t do squat in terms of studying subjects. They went back through 40-year old records of subjects who had undergone formula feeding in the 1950s and 1960s to drag out records of 13 subjects (they actually drug out 16, but three were of children) who met their experimental parameters. They weren’t looking for evidence of a metabolic advantage; they were looking to see if fat intake irrespective of calories made people gain weight.
Out of the countless studies done in those early years, they wanted to see if any could show that fat intake increased weight gain to a greater extent than the calories consumed as fat. As they put it in the Introduction to their paper:
One group of investigators concluded that “fat intake may play a role in obesity that is independent of energy intake.”
The Leibel et al paper was published in 1992, the time in which the low-fat mantra was at its zenith. It was a time that many people who should have known better were telling us we could eat all we wanted as long as we limited fat. Fat makes us fat, we were told. Cut it and you lose. What Leibel et al were trying to show in this paper was that the weight gain or loss effects of fat were a function of the calories contained in the fat, not some other magical property that makes people gain weight above and beyond calories.
Before we get to the interesting data in this study, let’s take a look at what the guy who actually did this work had to say. Leibel’s group went through old formula feeding studies done by Edward H. Ahrens, M.D., the head of the formula feeding lab at the time and the lead author of all the old papers referenced by Leibel. Says Dr. Ahrens about the subjects in the inpatient studies:
Thirty-eight of forty patients were observed continuously under strict metabolic ward conditions; four of the forty [I know, the math doesn’t add up] were sufficiently motivated and intelligent to follow the regimen at home. (Ahrens EH et al 1957)
A couple of points here. First, if four subjects out of 40 were “sufficiently motivated and intelligent” to be sent home with formula and instructions, what does that say about the other 36 (or 38)? Which is to my point earlier about the quality of subjects recruited into metabolic ward studies. Second, were some of the patients whose data was used for the Leibel paper those who were sent home? If so, it blows AC’s notion of being unable to rely on any data gathered from free-living subjects.
Dr. Ahrens in another paper describing his 15 years of experience using formula diets says this about cheating in metabolic ward studies:
Such cheating is a natural (but dismaying) consequence when a patient’s dissatisfactions with any part of the ward routine are not quickly enough appreciated by the ward personnel. Anticipation of the discontent is the clinician’s daily concern. The closer the relationship between the patient and his medical attendants, the less likely cheating is to occur. We have detected [my italics] cheating in only eight patients; undoubtedly others have gone undetected, but we feel the problem has been surprisingly minor. (Ahrens, EH 1970)
These are the subjects under lock and key. The people running the study have to maintain constant vigilance to prevent cheating. How about those who only check into the metabolic ward to sleep and spend the rest of their days at work or home? And those are the subjects who make up most of the metabolic studies you read about.
One last interesting point about the Leibel paper. The subjects they looked up in their retrospective analysis had undergone experiments during which they were given formula in amounts sufficient to maintain their weight. As they lost or gained weight, their caloric intake was increased or decreased to compensate so that their weight stayed about the same. According to the old papers about the original studies, the researchers tried to keep the subjects from fluctuations greater than one kg. One kg equals two pounds. If there was a metabolic advantage, it would probably show up within this two pound range and would be considered insignificant in terms of how this study was presented.
Some of the subjects, however, did lose or gain weight. Leibel et al then adjusted their caloric intake on paper to compensate for the weight differential. In other words, if a patient lost weight on a given number of calories of a precise formula in the original study, Leibel et al would adjust the intake (40 years after the fact) to compensate for the weight loss.
One subject, a 55-year-old male with a BMI of 32, maintained his weight on a high-carb formula at 2871 calories per day. The same subject then required 3501 calories to maintain his weight on a 70% fat, 15% carbohydrate diet. Sounds like a metabolic advantage to me.
There were two papers in AC’s list of 17 that did show what could be considered a metabolic advantage. In other words, subjects on the low-carb diet lost greater amounts of weight than subjects on low-fat, high-carb diets of the same number of calories. These are two of the three studies by Rabast et al that are the 4th and 6th studies on the list of 17 shown above.
How did AC deal with this seeming refutation of his notion that no metabolic advantage exists? By typical AC flimflammery.
In their 1981 study, Rabast et al observed significantly greater potassium excretion on the low-carbohydrate diets during weeks one and two. A considerable amount of potassium inside our bodies is bound up with glycogen, so the greater potassium losses in Rabast’s low-carbohydrate dieters may indeed be a reflection of greater glycogen, and hence water losses. Until recently, potassium excretion was often used a a marker or lean tissue loss; in Rabast’s study, this would indicate that the low-carbohydrate diet subjects lost more lean tissue. As lean tissue holds a considerable amount of glycogen, this would again point to glycogen-related water loss as the explanation for the allegedly “significant” differences in weight loss. [Italics in the original] If the low-carbohydrate groups maintained greater lean tissue and/or glycogen losses at the end of the study, then this would easily explain their greater weight loss.
Regardless of whether Rabast et al’s findings were the result of water loss from glycogen depletion, pure chance, or some other unidentified factor, they should be regarded for what they are: An anomaly that has never been replicated by any other group of researchers. For a research finding to be considered valid, it must be consistently reproducible when tested by other researchers. As proof of the alleged weight-loss advantage of low-carbohydrate diets, the findings by Rabast and colleagues fail dismally on this key requirement.
Wow! Where do we start?
First, AC didn’t mention Rabast’s 1979 study in which 117 patients were admitted to the hospital and studied on formula diets. I assume these subjects were hospitalized round the clock because in the body of the paper it states:
…and as the patients were under constant supervision differences in food intake between the two groups could be excluded.
Unlike the Kinsell study (the first of AC’s 17 I described in detail above), the authors of this study were expecting a different outcome. As discussed, Kinsell was obviously biased going in against the notion of anything other than calories count. Rabast et al went in biased against low-carb diets:
The popularity of so-called ‘fad’ diets, low in carbohydrates and relatively high in fat, has continued to spread, especially among lay groups. The caloric intake is only slightly limited, if al all; alcohol is allowed most of the time, and fat is consumed in the form of saturated fatty acids. However, this kind of dieting, which must always be carried out on a long-term basis, has proved harmful. The cholesterol intake can lead to severe health damage and clearly contributes to atherosclerosis.
After keeping the 117 subjects on low-carb vs high-carb diets of the same number of calories for 25 – 50 days, and probably hoping to find that those on the low-carb diet didn’t lose any more weight than those on the low-fat diet, the subjects on the low-carb formula diet lost considerably more weight than those on the low-fat diets. Here are the graphs from the paper.
After going through all the data, Rabast et al conclude
Differences in fluid and electrolyte balance could not be measured but marked fluctuations can occur. However, the change in body water and electrolytes could only be considered in short-term studies as the cause of the differences in weight loss. Variation in the depletion of the glycogen pool is also a feasible explanation, as up to now, sufficiently long-term studies have not been reported. However, the glycogen pool can be restored even under fasting conditions. Therefore, an increased rate of metabolism presents itself as the most feasible explanation. [my italics]
The 1981 Rabast study that AC does comment upon refutes his commentary on the difference being due to greater fluid loss from the low-carb diet.
Potassium excretion during the low-carbohydrate diets was significantly greater for as long as 14 days, but at the end of the experimental period the observed differences no longer attained statistical significance. At no time did the intake and loss of fluid and the balances calculated therefrom show significant differences. From the findings obtained it appears that the alterations in the water and electrolyte balance observed during the low-carbohydrate diets are reversible phenomenon and should thus not be regarded as causal agents.
As to AC’s comment that the work of Rabast et al should be ignored because it has never been replicated by another group of researchers, I’ll leave to you to decide the validity of that. There have been a number of such studies, including ones (as I’ll describe in a moment) in AC’s own list that confirm what Rabast found. The 1979 Rabast paper discussed earlier lists 17 of them.
Hang in there; we’re almost through. If I have to read all these papers and type all this stuff, the least you can do is stick with me ‘til the end.
Most of these studies don’t list the amounts of weight lost by the subjects because most of them aren’t designed to really look at weight loss. Most are designed to look at other metabolic parameters such as protein sparing or branch chain amino acid use or nitrogen balance and the authors weren’t particularly interested in how much weight the subjects lost. The authors mention that the two groups of subjects lost similar amounts of weight. Other than the Rabast studies that we’ve already discussed, only four studies listed the weight lost over the course of the study by the subjects on either low-carb or high-carb diets. In none of these cases did the weight loss difference reach statistical significance, so AC is presenting them as if there is no difference.
But in reality, there was a difference. It just wasn’t statistically significant.
Statistical significance as it pertains to weight loss is a function of both number of subjects and amount of weight loss. If I enroll 10 obese subjects in a weight-loss study and put five subjects on one diet and five on another, observe them for four weeks, and find that one group has lost an average of 2 pounds more than the other, that probably won’t be a statistically significant difference. Why? Because with only five subjects in each arm of the study, it requires a much larger weight loss to show a statistically significant difference.
If I do the same exact study, but enroll 100 subjects with 50 in each arm, and get exactly the same results – a two pound differential – then I achieve statistical significance. The more subjects, the smaller the difference in outcomes it takes to reach significance.
In the case of these metabolic ward studies, the numbers of subjects are small. As we’ve discussed, it is extremely expensive to keep subjects hospitalized 24 hours per day. Consequently, most metabolic ward studies don’t enroll very many subjects.
I went through all the papers in AC’s list and found four (aside from the Rabast that we’ve already discussed) that list both starting and ending weights for the subjects. I’ve listed them in the chart below.
As you can see, the study with the largest number of subjects had only 22 subjects in each arm. These studies all use a caloric intake that is lower than would be expected to produce any kind of a metabolic advantage because all are at an almost starvation level. Yet, as you can see, three out of the four show a greater weight loss in the low-carb arm than in the low-fat arm of the study. Equal caloric intake, greater weight loss with the low-carbohydrate diet. But, due to the small number of subjects, the difference doesn’t reach statistical significance.
If we had these same findings and same difference in weight loss between the two diets with a larger number of subjects, we would indeed have a significant difference. If we did a meta-analysis of these studies, we might find that adding the subjects together would end up showing a significantly difference in weight loss. Even though these differences don’t add up to statistical significance given the number of subjects involved, you can see the definite trend.
But what about the Piatti study, the one that showed the low-fat diet producing more weight loss than the low-carb? I have it marked with an asterisk for a reason. The paper by Piatti et al titled Hypocaloric High-Protein Diet Improves Glucose Oxidation and Spares Lean Body Mass: Comparison to Hypocaloric High-Carbohydrate Diet looked at how 25 obese women fared in terms of lean body mass and insulin sensitivity. They were put on 800 kcal diets for 21 days. It was found that the low-carb diet spared more muscle tissue and improved insulin sensitivity more than the low-fat diet of an equal number of calories.
Since the authors weren’t specifically studying weight loss, they didn’t really randomize the subjects by weight but did so by other parameters. As it turned out, the group on the low-fat, high-carb diet were much heavier than those that ended up in the low-carb arm. The average starting weight of the subjects in the low-fat arm was 213 pounds (96.8 kg) whereas the starting weight of those on the low-carb arm was 191 pounds (86.8 kg), a significant difference. It would stand to reason that subjects starting off at 213 pounds on a 800 calorie diet would lose more over 21 days than subjects starting out at 191 pounds and following the same diet, and indeed they did.
This post has gone on way, way too long, but I think it’s pretty obvious that these studies fail to ‘prove’ that a metabolic advantage does not exist. I would say, if anything, that they ‘prove’ just the opposite.
Just so you can go through these studies yourselves if you so desire, I’ve put them all up on Scribd. The links are below to the full text of all.
The next post will a) be much, much shorter and will b) go into detail on a beautiful study that AC totally disses in his book. We’ll look at his diss and what the study really says. That should put paid to AC.
All the papers referenced by AC listed below. All full text.