On to the second and, mercifully, final part of the critical review of the metabolic advantage as presented by A Colpo in his book The Fat-Loss Bible. As discussed in the previous post, our friend, like the kid to the left, is focused so intently on his refusal to believe in even the possibility of the existence of a metabolic advantage that he can’t read the literature correctly – not even the very literature he uses to try to prove his own position. His bias has hypnotized him to the point that he can’t see anything that doesn’t confirm his what he already believes. And this same bias prevents him from even taking a scientific approach to the problem.
We all fall victim to the confirmation bias and have to fight it constantly. Gary Taubes thinks I may even have succumbed a little in the earlier post on AC and the metabolic advantage. He emailed me saying he had read the post and thought it was great up to the point right at the end where I wrote that the data on the whole showed that, if anything, there was a metabolic advantage. Gary thought the data presented in all the studies in AC’s chart was ambiguous and that I was going out on a limb a little in making the statement that I thought, if anything, that the papers argued for a metabolic advantage.
I decided to base this critique not on the scientific literature at large, but instead on only the papers that AC mustered for his argument. I intended to make the critique much like a court case in which one side presents the information and the other attempts to counter it. I didn’t want to go out myself and gather a bunch of papers that confirmed my viewpoint, because then we would have had nothing but a bunch of dueling Ph.Ds, a bunch of he saids, she saids, that wouldn’t prove anything. I stuck with the papers AC used and presented my arguments as to why I didn’t think his papers proved his case. After going back and rereading the post, I still feel that if this ‘evidence’ were presented to a jury, the verdict would come back in favor of my arguments. If anything, AC’s own ‘evidence’ argues for the existence of a metabolic advantage, and, at worst, certainly doesn’t ‘prove’ that one doesn’t exist.
Since I posted the first part of my critique, AC has responded using his customary restraint and understated gentility designed to appeal to his sort of reader. His response – as I figured it would be – is merely a listing of even more papers he believes substantiate his claims. Instead of undertaking a serious scientific inquiry, he is looking for more white swans. Let me explain.
I wrote a long post a couple of years ago on Sir Karl Popper and the metabolic advantage. Popper set the standards by which hypotheses should be structured. A well-stated hypothesis should be able to be falsified. That doesn’t mean it will be falsified, but it should be structured in a way that it can be. And real scientists – of which, sadly, there are all too few in the field of nutrition – don’t try to confirm their hypotheses: they try to refute them.
One of the examples Popper used in explaining how a hypothesis should be established involved swans – white and black. He used the following as an example of a good hypothesis: All swans are white. He made the case that this hypothesis cannot be confirmed by simply pointing out more and more white swans. The hypothesis can be strengthened by doing so, but it can’t be proven. It can, however, be disproved by the discovery of even a single black swan. Popper argued that scientists should be working to find black swans instead of simply adding more and more white swan sightings to their data. The more effort scientists expend to find a black swan without finding one, the more their hypothesis is strengthened. Diligently searching for black swans is a much more valid scientific endeavor than simply looking for more white swans.
Many scientists don’t want to hunt for black swans, however, because they don’t want to blow up their hypotheses. The easy way to bolster their hypotheses is to continue to tally up all the white swans they find and forget about looking for black ones.
Which, of course, is what our young friend AC has done and written about in his latest missive. He tallies up a bunch more white swans and ignores the black ones, even the black ones in hiding in plain sight in his own list of papers. This failure of his to try to puncture his own hypothesis leads me to believe there exists a large chasm of incomprehensibility between the way AC thinks and the scientific method.
To give but one example of this, AC argues in his book that the studies by Rabast that clearly show a metabolic advantage aren’t valid because, as AC puts it,
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.
(In other words, AC is saying: that black swan over there isn’t really a black swan, because all the other swans I’ve pointed out are white. And since all the others are I’ve pointed out are white, that one can’t be black. It’s impossible.)
In point of fact, Rabast’s group in Germany has performed a number of studies showing a significant metabolic advantage in subjects in metabolic wards who follow low-carb, high-fat diets as compared to those taking in the same number of calories as high-carb, low-fat diets. This group pursued this line of inquiry and published a number of studies showing this metabolic advantage. Suddenly, however, they quit publishing on this subject and turned their attention elsewhere.
While in the research phase for Good Calories, Bad Calories, Gary Taubes interviewed Dr. Rabast about his group’s work, and here is what he said. They were inspired by an old scientific paper (more about which later) that offered up some data they found interesting and wanted to test themselves. They did the studies using formula diets, so they could more easily control intake and confirmed the data from the old study. They continued to perform these studies, all with similar outcomes, until Dean Ornish published his paper on dietary fat and heart disease. Dr. Rabast and his group decided that Ornish might be correct. They felt that although their own data showed that high-fat diets brought about substantially better weight loss than low-fat diets of equal calories, their work might encourage people to consume more fat, which, thanks to Ornish and the low-fat movement, they had come to believe may cause heart disease. So, they abandoned their research on high-fat diets and moved on to other interests.
The study that inspired them to study high-fat diets? An study from the 1950s done by a couple of British researchers, Dr. Alan Kekwick and Dr. G.L.S. Pawan. Their famous paper showed a definite metabolic advantage, a black swan writ large, as it were. And their famous paper is well known to AC, who has a few things to say about it. As you might suspect, given the results of this study, he declares it not worthy of consideration. Here is what he says in his book after he’s gone through his list of white swan studies, which, of course, are all worthy of mention.
There is one metabolic ward trial that due to its short duration did not qualify for inclusion in Table 1a, but still warrants a mention. Incessantly cited by supporters of low-carb diets, this is the famous metabolic ward study conducted in the 1950s by Kekwick and Pawan. The London researchers conducted two experiments. In one of these, they claimed that patients maintained or gained weight on a typical mixed diet of 2,000 calories, yet consistently lost weight when placed on a 2,600 calorie low-carbohydrate diet for periods ranging from 4 to 14 days. In the second of their experiments, they had 14 patients alternate between four different 1,000 calorie diets, spending a grand total of 5-9 days on each diet: 1) 90 % protein; 2) 90% fat; 3) 90% carbohydrate, and; 4) a mixed diet. According to Kekwik and Pawan, all of the subjects in the protein, fat, and mixed diet groups lost weight, with the high-fat group experiencing the greatest weight loss of all. However, despite the very low calorie intake, many of the patients reportedly gained weight during the high-carbohydrate diet! Not surprisingly, the Kekwik and Pawan study is frequently cited by supporters of low-carbohydrate nutrition. That they ignore the studies in Table 1a, yet eagerly embrace a short-term study conducted over 50 years ago, speaks volumes about their complete disregard for rational scientific inquiry. [Italics in the original]
Here’s why: Firstly, it has long been known that in the first week or two of low-carbohydrate dieting, there is often a far greater reduction in water weight due to excretion of sodium and/or glycogen, both of which bind water in the body. Therefore, studies of such short duration are next to useless as indicators of the comparative longer-term weight loss effects of these diets.
Secondly, the Kekwik and Pawan study was a poorly controlled mess. The researchers were even driven to denigrate their study participants, writing: “The first and main hazard was that many of the patients had inadequate personalities. At worst they would cheat and lie, obtaining food from visitors, from trolleys touring the wards, and from neighbouring patients. (Some required almost complete isolation.)” [Italics in the original]
Given that protein and fat have been shown numerous times to exert satiating effects, while low-fat, high-carbohydrate diets (especially the liquid, low-fiber variety!) typically result in ravenous hunger, it’s not hard to guess during which diet the participants may have ‘cheated’ the most!
The researchers also wrote: “The results we report are selected, a considerable number of known failures in discipline being discarded”. Note how the researchers included the words “known failures”; how many failures did they not know about? How many of the patients were crafty enough to sneak extra food without being caught? Why should we trust Kekwik and Pawan’s unlikely results, given their study’s numerous flaws? The answer is simple: Unless you are a famous low- carb diet ‘guru’ who has made millions promising people they will lose extra weight at the same calorie intake by cutting carbs, we shouldn’t! At least not if we believe good science mandates a tightly controlled process of investigation. [Italics in the original]
As we shall see shortly, this commentary is all so much piffle.
(Here is the full-text version of the Kekwick and Pawan study so that you can pull it down and follow along with the rest of the discussion if you like.)
Let us begin.
It is apparent from his critique that AC read the first part of this study, found a black swan, used a bunch of incorrect gibberish and swagger to try to say it wasn’t really a black swan and moved on without ever getting to the important part of the paper. Or, an alternative explanation is that, as with the Leibel study mentioned in my first critique, he either didn’t really read the paper thoroughly or he seriously misunderstood what he read.
Drs. Kekwick and Pawan start off by explaining why they undertook this study in terms that any of us who have struggled with excess weight and found different results with different diets can understand.
Many different types of diet have been successfully used to reduce weight in those considered obese. The principle on which most of them are constructed is to effect a reduction of calorie intake below the theoretical calorie needs of the body. Experience with these patients has suggested, however, that this conception may be too rigid. Many of them state that a very slight departure from the strict diet which can hardly affect calorie intake results in them failing to lose for a time. Though it is realized that evidence from such patients is notoriously inaccurate owing to their approach to this particular condition, it is too constant a belief among them to be entirely discarded.
Drs. K & P did a number of experiments. First they kept hospitalized subjects on diets of similar macronutrient composition but differing calories and found that reducing calories made the subjects lose weight. And, unsurprisingly, the more the calories were cut, the more weight the subjects lost. Next, the good doctors decided to see if changing the macronutrient composition of the diets made a difference. They started the subjects on 1000 calorie per day diets of one of the following three structures: 90 percent of calories as carbohydrate; 90 percent of calories as protein; or 90 percent of calories as fat. The structure of the diets made an enormous difference in how much weight the subjects lose. As Drs. K & P wrote:
So different were the fates of weight-loss on these isocaloric diets that the composition of the diet appeared to outweigh in importance the intake of calories.
In an effort to confirm their findings, Drs. Kekwick and Pawan went on to a third series of experiments as described here:
…patients…were put on to 2000-calorie diets of normal proportions to show that their weight could be maintained while in hospital at this level and then placed on high-fat, high-protein diets providing 2600 calories per day. It was demonstrated that these patients on the whole could maintain or gain weight on 2000-calories but, except in one instance, lost weight consistently on a 2600 daily calorie intake.
It’s easy to see why AC doesn’t like this paper. And we haven’t even gotten to the good stuff yet, which AC doesn’t make mention of in his book. We’ll get to that in a bit, but before we do, let’s take a look at AC’s critique of this much of the study (which is, apparently, all he bothered to read). You can read along from the above quote in his book.
His first complaint is that the study is over 50 years old. I find this a strange complaint, since the first study he lists in his chart of studies ‘proving’ his point was published a mere eight years after this Kekwick and Pawan study. The Kinsell paper was published in 1964, 46 years ago. Is there some magic cutoff date at 50 years that makes scientific papers unreliable?
Second, he claims that on low-carb diets all the weight loss from the first two weeks is water, and since these studies lasted less than two weeks, the difference was all water.
Kekwick and Pawan were a little smarter than Anthony gives them credit for being. They understood well the notion of water loss. (As we will see shortly, they understood it vastly better than our young friend.) They pointed out the following:
During these periods [the different diet studies] the patients were weighed daily and in some of them balance studies were carried out in respect of water, nitrogen, fat, sodium, chloride, and potassium. Total body-water and the basal metabolic rate were estimated weekly or at the end of each period on the diet.
If you look at the full-text version of the study I linked to above, you can see graphically how this all plays out. In these studies the weight loss was definitely not all water.
In an effort to be meticulously accurate, not only did K & P monitor all the above carefully, they even went further. Since these patients were not on formula diets but were on real foods instead, making it more difficult to accurately determine caloric intake, the staff would take representative samples of the foods eaten, blend them into a soup, then analyze samples to make sure the protein, carbohydrate and fat content were as estimated in the food tables. It was hardly a “poorly controlled mess” of a study.
AC next attacks the study because the researchers admitted as to how difficult it is – even in hospitalized studies – to prevent cheating.
In such a study the difficulties are formidable. The first and main hazard was that many of these patients had inadequate personalities. At worst they would cheat and lie, obtaining food from visitors, from trolleys touring the wards, and from neighbouring patients. (Some required almost complete isolation.) At best they cooperated fully but a few found the diet so trying that they could not eat the whole of their meals. When this happened the rejected part was weighed, and the equivalent calories and foodstuffs were added to a meal later in the day. The results we report are selected, a considerable number of known failures in discipline being discarded.
Kekwick and Pawan simply wrote of the difficulties in preventing cheating. They were on the lookout for it, threw out data they knew was compromised, and compensated for episodes of cheating of which they were aware. I believe the fact that they recognized cheating as going on and were keeping an eagle eye out for the cheaters makes their data more accurate, not less.
I also find it strange that AC is more than willing to toss data because of cheating in this study and is more than willing to accept data from other studies in which there was probably just as much – if not more – cheating that the authors neglected to mention either by design or because they didn’t realize it was happening.
One other thing that points to the degree into which K & P watched over this study is one that all female readers who have had trouble losing will be familiar with.
Another factor of importance which could not be eliminated was that many patients were women, in whom the retention and the losses of water associated with the menstrual cycle affected the daily weight and the estimation of total body-water. We were surprised to find how great such factors could be, amounting in one woman to the retention of more than 3 litres of water.
Only a fool or a seeker of white swans only would think the good doctors didn’t monitor this study closely.
Now to the fun part, the part AC probably didn’t read. And the part that really demonstrates the metabolic advantage.
The first part of this paper, the part AC has critiqued, is only a minor part of the paper. The majority of the paper is devoted to the efforts the Drs. K & P made to determine what happened to the excess weight lost in dieters on the higher-fat diet. They checked fat loss in the stool, they checked (as mentioned previously) water loss, they checked about everything they could think of. You can read in the full version how careful they were.
After sifting through all the data and finding no reason that their results should have been invalid, the docs checked yet one more item. They looked at insensible water loss.
Insensible water loss is the loss of water we all experience minute by minute that we not aware of. We know we lose water when we urinate and/or defecate, and we know we lose some water when we visibly sweat, but we are not aware of the large amount of water we are getting rid of through our breath and via sweating that we don’t notice. And this amount of water we lose is fairly large.
Do this experiment. Get an accurate scale and weigh yourself immediately before going to bed. Go ahead and urinate (and do anything else you might need to do) before weighing. Don’t drink or eat anything, hop in the sack and sleep through the night, then get up and weigh before you urinate in the morning. I absolutely guarantee that you’ll weigh less than before you went to bed.
If you breathe on a mirror, you will fog it from the water vapor in your breath. This vapor is water that you lose every single time you take a breath. You breathe approximately 12 times per minute (while resting), which means you breathe 720 times per hour and 17,280 times per day. And that’s if you’re at rest. If you are active, you take a lot more breaths than that. Probably something in the neighborhood of 20,000-23,000 breaths per day, depending upon activity level. Each one of these breaths contains water vapor that you are losing from your body, which is why you drink liquids throughout the day. If you didn’t replace this water, you would become dehydrated.
If you have a fever or if you exercise, you breathe a lot more rapidly and lose a lot more fluid. Thus, one of the things doctors have to be concerned about in very sick patients with high fevers is dehydration.
You also lose insensible water through constant perspiration. When you awaken in the morning, if you’ve slept tightly covered up, you’ll notice you’re a little damp. Not a lot, unless you’ve had a fever, but a little. This is insensible water that you lost.
I remember how amazed I was the first time I ever looked at my own hand under a dissecting microscope. Looking at my hand with my naked eye, it appeared normal and dry. When I stuck it under the scope and looked, I could see little volcanoes of perspiration bubbling up from unseen pores. It’s part of the way we regulate our temperature, and unless we work up a visible sweat, we never notice.
This loss of insensible water is why we lose weight overnight. In eight hours of sleep, we breathe out about 5,760 breaths filled with water vapor and we sweat all night. This water weight usually ends up being between 1 to 2 pounds or even a little more.
If I were to take a bunch of thyroid hormone or take an amphetamine, I can assure you that my metabolic rate would rise and that my insensible water loss would increase. In fact, insensible water loss is a surrogate for metabolic rate. If your metabolic rate rises, your insensible water loss rises. And since insensible water loss can be easily measured, the metabolic rate can be easily estimated without having to do metabolic chamber studies.
Which is exactly what Drs. Kekwick and Pawan did with several subjects on the various diets.
They kept the subjects isolated and under supervision and weighed them on extremely accurate scales throughout the day.
Measurements were made by weighing the patient at intervals of one hour on scales specially constructed for this purpose by Messrs. W. & T. Avery Ltd. which are sensitive to 2 g. over the range of weights concerned. During these hours no food was taken and neither urine nor faeces voided, and errors due to temperature, activity, and air draughts were avoided as far as possible.
(Scales that are sensitive to 2 g are extremely sensitive. Two grams weighs about seven one hundredths of an ounce.)
So, here is what the researchers did. They first fed the subjects the standard diet available to the patients on the ward and discovered what the insensible water losses were throughout the day. You can see how this came out in the graph below, Fig. 11.
When Drs. K & P put a single patient on the different diets – 90 percent fat, 90 percent protein or 90 percent carbohydrate – and measured the insensible water loss throughout the day, the table below, Fig. 12 shows what happened. There was an increase in insensible loss with the high-protein diet as compared to the high-carb diet, and a much greater increase in insensible water loss with the high-fat diet.
The area of the chart that I colored in is the difference between insensible water loss, which represents a change in metabolism, between the high-carb diet and the other two diets. This colored part of the chart represents the metabolic advantage of the high-protein and high-fat diets compared to the high-carb diet of the same number of calories. The peach colored part of the chart represents the metabolic advantage of the high-fat diet as compared to the high-protein diet while the grayish color represents the metabolic advantage, as measured by increased insensible water loss, between the high-protein and high-carb diets.
The researchers wanted to make sure this wasn’t an isolated phenomenon, so they analyzed three other patients and created the graph below, Fig. 13, which mirrors the results in Fig. 12 and demonstrates that this wasn’t an outcome isolated to just one subject.
The ever cautious Drs. Kekwick and Pawan interpreted their findings thus:
The rate of insensible loss appears to be much affected by the type of food, provided that the water and sodium intakes are kept constant throughout the period of observation; whether this increased rate of insensible loss is a measure of bodily metabolic activity must remain in question. Even if metabolic activity cannot be measured directly, the difference in weight responses seen with these diets does not seem to be completely due either to an altered state of hydration or to a simple deficiency of calories. We suggest that the rate of katabolism of body-fat may alter in response to changes in the composition of the diet.
And their summary:
As the rate of weight-loss varied so markedly with the composition of the diets on a constant calorie intake, it is suggested that obese patients just alter their metabolism in response to the contents of the diet. The rate of insensible loss of water has been shown to rise with the high-fat and high-protein diets and to fall with high-carbohydrate diets. This supports the suggestion that an alteration in metabolism takes place.
If you haven’t already, I would encourage you to read this entire study and make your own judgment. I’m sure you won’t find it the “poorly controlled mess” that AC does. In fact, I suspect you’ll find just the opposite. Unlike most of the studies published today, this one is not loaded with incomprehensible jargon, is delightfully well written and is extremely accessible to those with little medical or scientific knowledge. You can see for yourself how precise these researchers were and now meticulously they looked for anything that might confound their results. It would be great if more studies were done this carefully today and written this clearly.
This is the end. I am through with AC. I’ll leave it to the readers of this post and the previous one on this subject to make their own decisions as to whether or not a metabolic advantage exists for low-carb, higher-fat diets. I won’t be provoked again into jumping into the mud and wrestling around. So this is my black swan song on the subject.
I read a quote a few days ago by Nassim Taleb, the author, appropriately enough, of the book The Black Swan and, for my money, the infinitely better Fooled by Randomness that is apropos to this situation:
A good foe is far more loyal, far more predictable, and, to the clever, far more useful than any admirer.
So, to you, Anthony Colpo, I raise my hat. Had you not attacked me out of the blue, I would be less knowledgeable than I am today. I wouldn’t have bothered to dig into all the ‘white swan’ papers you posted trying to figure out why these researchers got the results they got. I, like you, would still be mired in the notion that metabolic ward studies are squeaky clean without any hint of sullied data as a consequence of cheating. Like you, I would still probably be confusing metabolic ward studies with metabolic chamber studies, which are horses of a much different color. Also, I thank you because I had kind of blown off the Kekwick and Pawan papers (there are others besides this one from The Lancet) as being too old to be worth studying. You forced me to take another look, and I was delighted at what I found. And, sad to say, like you, I, too, had read only the first part of the these studies, the parts about the diet comparisons. It wasn’t until your attack that I actually read this paper all the way through and found the gold mine in the latter pages.
So, AC, I sincerely hope the best for you; I thank you for pushing me into this exercise and wish you godspeed on your journey through life.