I think that it might be a good idea to state here that I'm a calories-in calories-out sort of a person. Arguing about metabolic advantage is pointless. When a person loses weight they convert tissue to energy, amputations excepted. Where the calories go, whether it is BMR, thermogenesis, glycosuria, increased spontaneous movement, shivering etc, the calories always go somewhere. As far as I am aware no one is suggesting that calories evaporate. Certainly I'm not.
I am on blog as stating that dietary fat, in common with carbohydrate, is stored in the aftermath of a meal. If you did not store your dietary fat it would sit there in your chylomicrons until you were as hyperlipaemic as a diabetic on an ADA approved low fat diet. All fat which is not used for on going metabolism must be stored. It may take a few hours to clear chylomicrons but they do get cleared. Mostly in to adipocytes.
EDIT: There is an update on this post here.
Okay, let's look at the Grey and Kipnis (GnK) 1971 paper.
If you take a human like ET, who is maintaining a bodyweight of 142kg by eating 4,200kcal per day, you can adjust her macronutrient ratio to pretty well whatever you like and she will remain weight stable, provided you insist that she always consumes 4,200kcal/d. Insulin will plummet on a high fat diet and this will release necessary fatty acids from adipocytes. These FFAs, the Spawn of Satan, will be both released and used at an increased rate. But will she lose weight? Of course not. She will just re-pack her adipocytes with the extra fat from her fixed, 4,200kcal, low carbohydrate diet. Many people, myself included, can maintain weight stability on a low carbohydrate diet for years.
There's a nice paper about fat accumulation under low insulin conditions on my hard drive. It happens. I will post about it when I feel like it, it looks interesting.
There was one participant in the GnK study, RP, who refused to comply with the study protocol. That's a 20% non compliance rate. She under ate and lost weight. We get no information about her caloric intake at any stage. We know relatively little about her diet (ie sugar vs starch) before the study. We have no idea what sort of carbohydrate was used in the study diet. The only information we have about which phase of the diet gave the weight loss is this graph:
But we certainly have a failure of compliance in this study where an obese subject refused scheduled food and subsequently lost weight. That's worth remembering.
Now let's look at hypocaloric conditions in the same paper. DB, SM and DM were put on to 1,500kcal/d (Keysian starvation) from 2,200kcal/d, 3600kcal/d or 3,800kcal/d respectively. They, err, lost weight. They lost weight pretty nigh on linearly over 12 weeks whether their fasting insulin was 40microIU/ml or 15microIU/ml, produced by adjustment of their carbohydrate intake from 240g/d, down to zero and back to 240g/d.
NB I am perfectly willing to accept these results as they stand but just as an aside; none of the individual records shows any suggestion of a weight shift related to to glycogen depletion/repletion on these changes in carbohydrate intake. The LC phase included zero carbohydrate. The HC version of 1,500kcal provided 72% of calories as some sort of carbohydrate, ie 240g/d. Maybe even Weight Watchers depletes liver glycogen following an overnight fast... A bit odd but probably irrelevant.
Back to the results. The caloric intake was fixed and low. Fat was stored in the LC/HF group and accessed easily because insulin was low. Under high carbohydrate intake calories were stored mostly as glycogen and glucose was metered out to avoid hypoglycaemia. Any fat lost by the lower-but-not-zero rate of lipolysis under high insulin levels was simply not replaced.
Just eat 1,500kcal/d and you will lose weight. Eat more than 1,500kcal and you're a pig.
The study was designed to get exactly these results.
But, out of only five subjects, one obese person became a food refusenick. Various studies have had similar compliance problems, with obese participants refusing food. Let's look at some of them.
I knew there had been a paper by Krauss looking at lipoprotein subgroups during weight stability on assorted carbohydrate intakes. It didn't measure plasma insulin but, if we accept anything from GnK's work, I think we have to accept that under weight stability we can dial fasting insulin by adjusting carbohydrate intake. At least between 4% and about 72% of calories.
Krauss looked at diets composed of 54%, 39% or 26% of energy from carbohydrate, with a bonus group on 26% carbohydrate and (gasp) 15% of total calories as saturated fat.... These folks were instructed to maintain weight stability. This quote had the LC brigade, myself included, giggling if not rolling around in the aisles.
"Despite our effort to maintain constant weight, the 26%-carbohydrate, low-saturated-fat diet group lost more weight than did the 54%-carbohydrate group during the stable-weight period. There was also a trend for a greater reduction in percentage body fat with the lower-carbohydrate diets (P < 0.02, analysis of variance)."
The numbers are in Table 2.
My browser squishes the table, these are the numbers that matter, simplified. Remember, everyone was supposed to be weight stable:
Some of these numbers made p<0.05 or even <0.02, shrug. What is more interesting is the trend in accidental weight loss. Oh and look, the sat fat group didn't lose the most weight, just the most fat. I like that.
My take home message is that the lower the carbohydrate intake (and it is reasonable to assume the lower the fasting insulin) the harder it is to consume enough calories to maintain the obese state. It's possible, but not easy.
Then there is this study looking at the HPA stress axis under fixed calorie low carbohydrate or medium carbohydrate diet conditions. I'll just look at weight and insulin because the changes in the processing of cortisol are about as lucid as Krauss' early papers on lipoprotein changes due to dietary saturated fat. It seems reasonable to assume things improved on LC, otherwise any deterioration would have been headline news. This is a crossover study, the same people did a month on low carb and a month on medium carb, in random order. This makes the results tables somewhat unintelligible but it still comes close to replication the 1,500ckal section of GnK's paper.
Under weight stable conditions fasting insulin was 16.6microIU/ml on 57% carbohydrate. Under LC (4% carbs) conditions, on a fixed 2000kcal intake insulin dropped to 7.3microIU/ml. Weight loss was 7.2kg in 4 weeks. Eating 2000kcal of which 35% was carbs gave 4.7kg weight loss on a fasting insulin of 9.2microIU/ml. After correction for water shifts under LC there was exactly the fat loss accounted for by a caloric deficit of 66kcal/d. This was, oddly enough, exactly the caloric count of the food REFUSED from the 2000kcal provided during the LC phase........... Which they had been asked to eat. Obese people refusing part of a 2000kcal ration.
These folks where in a residential diet study. They consistently refused food they had agreed to eat during the "fixed" caloric intake phase, but only if the macronutrient ratio lowered insulin. As an aside the insulin level on LC was statistically significantly lower than on weight stability but the drop on medium carbohydrate was not. I would argue that the difference between either 7.3microIU/ml or 16.6microIU/ml and 9.2 microIU/ml is biologically significant even if p is > 0.05.
The original report details the menus and a battery of psychological test. There is a mass of information in this study from Aberdeen. As we all know, people mostly seem to get depressed and stoopid on LC diets. In this instance they just stopped being hungry!
Here is one of the best quotes, from the results section:
"The 3-d maintenance diet was designed to 1) neutralize the ketogenic state and replete liver carbohydrate stores and 2) to return hunger to baseline levels— equivalent to the maintenance period 1, before ad libitum feeding—recognizing that a carryover effect from the weight-loss phase existed. This design is particularly relevant for the subjects who were given the LC ketogenic diet first and then the MC nonketogenic diet."
"to return hunger to baseline levels......" I like that. Hunger on the ketogenic diet is not at baseline levels, it is lower. Just supplying MORE calories INCREASES hunger, so long as the calories have carbohydrate at 57%. Hungry and weight stable or less hungry while losing weight. I really like that.
Then a brief quote from this paper, also a weight loss rather than weight stability study, but the quote is too entertaining to leave out.
Fasting insulin was 11.6microIU/ml and 14.4microIU/ml at 3 and 6 months under ad libitum calories but progressively less stringent carbohydrate restriction. Under rigid calorie restricted low fat eating it was steady around 18microIU/ml.
"Based on dietary records, the reduction in daily caloric intake was similar in the two groups. For the greater weight loss in the very low carbohydrate group to be strictly a result of decreased caloric consumption, they would have had to consume approximately 300 fewer calories/d over the first 3 months relative to the low fat diet group (28). Although the inaccuracy of dietary records for obese individuals is well documented (31, 32), it seems unlikely that a systematic discrepancy of this magnitude occurred between groups of subjects who were comparably overweight."
We know that obese people always under report their caloric intake, just ask any obesity expert. Why on earth should this particular group of obese people consistently over report their caloric intake? That's not what fat people do.
You could rephrase this to speculate that the LC group either ate and "used" an extra 300kcal/d or became pathological liars who exaggerated their food intake by 300kcal/d, presumably to wind up the experimenters. Ketosis does make you stoopid and depressed, why shouldn't it make you in to a practical joker too?
Is anyone seeing a pattern of people (or rats) refusing food under reduced insulin conditions?
In the real world people eat when they are hungry, because they are hungry. They don't drink fixed caloric intakes of mysterious liquid formulations from researchers who's rat models are based on either sucrose or vegetable oil.
Can people actually gain weight on high fat diets? Of course they can. You can accumulate fat without elevated insulin. But you are much less likely to gain weight if you are not hungry.
I note that Chris Voight was not on a fixed calorie intake and was not hungry while he lost an impressive amount of his excess weight. I think his fasting insulin was low and he was performing lipolysis at an impressive rate.
Let's make this clear. Fasting insulin determines weight loss. The effect is primarily through reduced dietary caloric intake secondary to lipolysis-mediated access to adipose tissue calorie stores.
Overfeeding in excess of preferred calorie intake breaks the system. GnK simply disabled the mechanism of appetite control by fixing caloric intake. Fine to prove a point. It's this sort of research that has got us where we are today.
NB I think this decrease in hunger probably only occurs in obesity. For those of us who have adopted a LC eating pattern without the need for weight loss (and still have little excess fat) there are clearly other factors coming in to play, as there will be when a previously overweight person approaches target/ideal weight, what ever that might be.