Diabetes progresses on LC/HF Diet

First the disclaimers:

  1. It's a rat study
  2. The OLETF rats used are a strain that become spontaneously diabetic and mildly obese -- however to the best of my knowledge this strain was not created by genetic manipulation (e.g. it's not a "knockout").  More here.
My excerpts will be from scattered around the article.
The long-term effect of low-carbohydrate/high-fat diets on the development of diabetes mellitus was studied in Otsuka Long-Evans Tokushima Fatty strain (OLETF) rats. Four groups of spontaneously diabetic (type 2) male rats at 10 weeks of age were pair-fed semi-purified powder diets containing different amounts of carbohydrate (80 %, 60 %, 40 %, 20 % of total calories) for 30 weeks. The carbohydrate content was isocalorically substituted for the fat content in the diet. At the onset of experimental feeding (10 weeks of age), an oral glucose tolerance test (OGTT) was normal in each group.

This was a fairly long term study for rats -- 40 weeks total. The first 10 weeks were simply to grow the rats. Although this is a spontaneously diabetic strain, the OGTT done at the start of the dietary intervention showed these rats were all still normal -- had not developed diabetes. 

The rats were divided into four groups, diets contained the same calories, and although not mentioned, protein content was held constant as well. Only the carb/fat proportions were altered from LC (20%) to HC (80%) carbs in 20% increments with fat reduced proportionally. I'll call these LC (20%), MLC (40%), MHC (60%) and HC (80%) 

Some background -- emphasis mine -- see text for citations.

Many epidemiological studies have already been performed to evaluate the relationship between dietary modification (low-carbohydrate/high-fat diet, LC/HF) and the development of DM. However, their results are contradictory. Some show negative associations, while others show positive associations. Even in animal experiments, some researchers studied the effect of high-fat feeding on the development of DM. However, the experimental duration of the test diet feeding was too short (less than 2 months) in most studies. Moreover, the experimental animals were allowed free access to the test diets in most of the studies. In general, diabetic animals select LC/HF as opposed to a high-carbohydrate/low-fat diet (HC/LF). In the previous studies, the animals consumed more LC/HF. Therefore, compared with the animals that were kept on the HC/LF, the calorie intake was higher in those fed LC/HF.  From the results of the previous studies, it is difficult to draw any conclusion about the effect of LC/HF on the development of DM.
This is interesting that ad libitum, rats gain MORE weight on LC than LF because they eat more. This is counter to low carb dogma.
After 15 weeks of the test diet feeding there was no significant difference in the glucose tolerance among the 4 groups, although most of the rats were diabetic. The body weight increased with the decrease of the carbohydrate intake and increase of the fat intake (p < 0.05), and the difference increased in proportion to age (p < 0.05). The severity of diabetes mellitus was also increased along with the lower carbohydrate intake and higher fat intake, when the carbohydrate intake was less than 60 % (in energy).
So even with isocaloric intake, THESE rats gained MORE weight on low carb than high carb.
On the other hand, there was a significant increase in the 20 % group in the postload plasma insulin levels as compared with the other 3 groups at 40 weeks of age. Fasting plasma free fatty acid levels were increased in the lower carbohydrate content groups (20 % and 40 %) as compared with the higher carbohydrate content groups (60 % and 80 %) at the end of the experiment. 
This is to be expected, and is so often ignored. Since these rats were only mildly obese, the elevated NEFA is probably an indication of some degree of insulin resistance in the fat cells.

Impairment of insulin secretion may be the cause of glucose intolerance induced by low carbohydrate intake rather than insulin resistance. These findings suggest that low-carbohydrate/high-fat diet aggravates diabetes mellitus in genetically diabetic rats, and that the development of diabetes mellitus is associated with the activation of the glucose-fatty acid cycle.

25 Weeks (15 Weeks Dietary Intervention) 

40 Weeks = End of Study 

The authors conclude that IGT is due to impaired insulin response (e.g. compromised beta cell function) and not skeletal muscle IR. I would have to question this because if you look at the tables above, we see that the LC group has a significantly higher insulin response to the glucose challenge at 30 min vs. all of the other groups. At 60 min, both the LC and MLC groups have higher insulin than the MHC and HC groups (insulin levels correlate inversely with carb), and at the 2 hour mark, only the LC group seems to have abnormally elevated insulin compared to the other groups. To me their results are consistent with IR -- in the fat tissue resulting in elevated NEFA, and peripherally resulting in the rats having to produce more insulin.


Again, we have to remember this is a rat model, and one in which the males of this strain are predisposed to diabetes. However there is a strong genetic component to the development of T2 diabetes in humans. This study seems to indicate that while the LC'er may have better glycemic control, this is achieved at the expense of elevating the other circulating compound that results in dysfunction and damage -- NEFA's. 

Bottom line: This study demonstrated that in genetically predisposed rats, diabetes developed and progressed further in those rats fed a LC diet than those fed a HC diet.


LeonRover said…
Here is the description of this rat (OLETF)Ref 13.

They are SO predisposed to T2DM, that I think one cannot draw any conclusions vis a vis humans and intake of low carbs in the 1st study; it could just as well be the high fat, and all a diet does is hasten or speed up the process. Show me something done on Wistars or Sprague-Dawleys and I begin to chew it over.

But that's me. If they me show something on rats, then on pigs, then I take it into account.

A spontaneously diabetic rat with polyuria, polydipsia, and mild obesity was discovered in 1984 in an outbred colony of Long-Evans rats, which had been purchased from Charles River Canada (St. Constant, Quebec, Canada) in 1982. A strain of rats developed from this rat by selective breeding has since been maintained at the Tokushima Research Institute (Otsuka Pharmaceutical, Tokushima, Japan) and named OLETF. The characteristic features of OLETF rats are 1) late onset of hyperglycemia (after 18 wk of age); 2) a chronic course of disease; 3) mild obesity; 4) inheritance by males; 5) hyperplastic foci of pancreatic islets; and 6) renal complication (nodular lesions). Histologically, the changes of pancreatic islets can be classified into three stages: 1) an early stage (6-20 wk of age) of cellular infiltration and degeneration; 2) a hyperplastic stage (20-40 wk of age); and 3) a final stage (at > 40 wk of age). These clinical and pathological features of disease in OLETF rats resemble those of human NIDDM.
CarbSane said…
I agree to some extent about the rat and the model. Yes, they are quite predisposed to diabetes. Perhaps not as predictably as in this rat model, but IR and certain tendencies (like ectopic accumulation of lipids) tend to be more prevalent in siblings and offspring of T2's. Clearly there is a genetic predisposition involved in developing T2. So while this result doesn't mean a regular rat will develop diabetes on LC, a rat predisposed to will develop it more rapidly. It is quite possible that LC could aggravate diabetes in humans who are predisposed. It is the high fat that causes the problems I think.
Peter said…
Hi Carb Sane,

These rats are very interesting. In the experiment they were fed 28.7 joules per day. Under strict isocaloric conditions the high fat fed rats gained about 120 grams more bodyweight than the high carbohydrate fed rats.

As I am uninterested in metabolic advantages either way I tend to view this as evidence that the high carb fed rats burned calories (uncoupling proteins or exercise) and the high fat rats stored more energy. They made less use of uncoupling proteins or were less active because they had stored more of their 28.7 joules.

I think that this is the first study I have seen where this has been achieved.

The implication of this, from the vantage of insulin as a storage hormone for fat, is that the 24h area under the curve of insulin in these rats increases with increasing fat content of the diet. At 15 weeks in to the experiment fasting insulins were identical across the groups so we are presumably looking at differences post prandially on their normal experimental diets.

In normal rats and normal humans this is the opposite of what I would expect, or what seems to happen.

So I have to accept that these rats become hyperinsulinaemic in direct proportion to fat intake and inverse proportion to carbohydrate intake.

The only way I can see this happening is, as you point out, by the development of muscle insulin resistance.

I fully accept that intramuscular di and tri glycerides inhibit the uptake of glucose by muscle cells, all other things being equal.

The problem for these rats appears to be that they accumulate sufficient IMTs to raise insulin levels to the point of fat storage despite a calorie limited diet. So something has to be broken on a fat processing basis, probably in muscle cells, resulting in insulin resistance greater than is appropriate for the metabolic fuel mix.

As fat storage rises I can see that adipose tissue insulin resistance will also increase, so FFAs will spill and the IMT condition will be aggravated. The effect is minor at 25 weeks of age and marked at 40 weeks of age.

The core question to me is how many people have this particular metabolic problem, and what does it have to do with fat intake and "normal" human T2 diabetes?

I know from occasional comments that Jenny Ruhl has made on her blog that there are indeed a small number of people with type two diabetes who respond to low carbohydrate, high fat diets by frankly deteriorating glycaemic control.

They are very very rare but they do exist.

I guess we could get them to marry and try breeding them up in the same way as the Otsuka Long-Evans Tokushima Fatty strain rats have been bred up but.....

It certainly is a very interesting paper and I've spent quite some time thinking about it. I think my ultimate conclusion is that the statement by the authors that the OLETF rat is a good model for human type 2 diabetes is simply wrong.

The OLETF rat very interesting. I'd love to know why it responds as it does. It might help a small number of people with a very rare form of diabetes.

Nigel Kinbrum said…
I have a couple of observations:-

1) The rats were always in caloric surplus as they were always gaining weight. I wonder what results would have been obtained if the rats had always been in caloric deficit? Humans usually go on LCHF diets to lose weight.

2) The rat OGTT glucose load was 2.0g/kg which is considerably higher than used in human OGTTs in the UK (the load in the UK being 75g). This increases the probability of the rats' 120-min post-load plasma glucose concentration being >11.2mmol/L.

Cheers, Nige.
Peter said…
Nigel, the rats gained weight because they stored more fat than they released. As you can't have fat calories free in your blood stream they all have to get stored. Their release from adipocytes is controlled by insulin. This does not require caloric excess. All of the rats stored the bulk of all of their calories. The hyperinsulinaemic rats failed to release enough FFAs from adipocytes to stay slim*. The stored fat will simply not have been used for activity of futile cycling..... They were probably cold as well as fat and lazy!

*until the fat cells became insulin resistant once overstuffed, late in the study FFAs rose in the fattest rats.

I ignored the OGTT as you would really need to carb adapt the rats before OGTTing them. And by this time they are spilling FFAs from insulin resistant adipocytes, so glucose adaptation might not be a simple matter in the presence of FFAs the way it is in a LC person who is actively losing weight so shrinking their adipocytes and can actually shut down FFA release with insulin and re-adapt to carbs as a major calorie source...

Peter said…
"activity OR futile cycling"

sorry for the typo, bed time!

CarbSane said…
@Nigel -- Hi there! Welcome and thanks for reading and commenting.

I've had some experience with lab rats (I guess you could say I was one and worked with them ;) ), so here is one limitation of them as a model: rats tend to continue to grow throughout their lives or at least for a longer relative period than humans. I've sac'd and autopsied more than a few older ones in my day and can attest that even on normal chow, just waiting around in a cage to be experimented upon, your basic Sprague Dawly rat will gain significant weight, and fat mass at that, in a matter of a few weeks. So yes, they tend to eat a positive caloric balance, at least some of the gain is "normal" growth (LBM), but they will gain fat mass too.

I do find it interesting that, counter to what we are led to believe, the LC rats maintained a higher positive caloric balance than did the HC ones.

But your point is valid regarding LCHF for weight loss. I guess my concerns are for the long term maintenance and/or for that group of low carbers that fail to lose weight on such plans (they do exist!). I have read many personal accounts of rather significant weight loss only for the person to regain despite maintaining LC. Or low carb maintenance is a 10 lb yo-yo of carb creep mild gains and buckling down meat and eggs loss cycles. How does this impact insulin sensitivity and/or pancreatic function?

I'm not sure I'm grasping what you're getting at in the first part of your comment. The rats were all in chronic positive caloric balance to gain consistently. You might be on to something about the futile cycling/body temp of these rats. Perhaps what predisposes them to diabetes is related to decreased futile cycles in rats. This is yet another problem with the rat model for humans, as we don't utilize these to anywhere near as great an extent to burn off caloric excesses except in cases of extreme overfeeding.

As to the OGTT results, even one low carb day prior can lead to impaired glucose tolerance in humans. So these results could be influenced by this. You are right, they should probably have carb adapted the rats for a few days before performing these tests to see if the IGT was "permanent". This is one of my concerns for treating T2 with VLC diets. Elevated NEFA seem to be just as, if not more deleterious as elevated blood glucose levels. Lots of low carbers have issues with 100% compliance and in a way are making themselves more diabetic.
Peter said…
Hi Carb sane,

Part one is just my Taubsian view of energy balance...

I would add that even a single high fat meal, let alone a LC meal, will impair glucose tolerance next day but that I don't view FFAs as lipotoxic under these circumstances. It happens even if you have perfectly functional adipocytes and normal muscles.

I agree this uses the same mechanism of muscle insulin resistance as occurs through inappropriate FFA release from overfilled adipocytes.

Sourcing FFAs from chylomicrons with a fully functional pancreas/liver produces a temporary switching of fuel source away from glucose to greater dependence on lipids in muscles. It's temporary. It's appropriate.

Sourcing FFAs from damaged adipocytes, which have been damaged by chronically elevated insulin to the point where they have stored so much dietary fat that they no longer respond to insulin, is a different problem. This is instigated and maintained by a failure of the liver to suppress glucose release. I keep coming back to an hepatic basis to "ordinary" human type 2 diabetics. That's how the adipocytes became over filled in the first place.

A non-diabetic LC eater will spike glucose badly with a big carb load. I have hit 12mmol/l after a bag of chips I ate at a leaving party at work. But I absolutely don't class myself as a diabetic because, after sensible LC adaptation, I run a perfectly normal OGTT. Okay, to be honest I get slight reactive hypoglycaemia which is asymptomatic at 3.4mmol/l.

I agree LC has the potential for hyperglycaemic damage on a lapse, but only for the first 2-3 days, possibly less (no data on adaptation time!). I maintain this risk is outweighed hugely by the chronic normoglycaemia of non lapse times. especially compared to the chronic slow death of high carb eating for diabetics giving an HbA1c of over 9%....

CarbSane said…
Peter, I remember reading an excellent post of yours regarding IR on your blog but don't seem to find it now. The gist was that IR is a "good thing" under certain conditions as it spares glucose for the tissues that need it most. I agreed with that at the time and what you say here makes logical sense. Not all IR is the same.

What concerns me at this point is that IR and T2 diabetes are characterized by BOTH elevated glucose and NEFA. Each of these are associated with deleterious systemic effects. Glucose, for example, seems to be particularly damaging to the nervous system. But NEFA's interfere with various receptors in organs and vascular endothelium and are directly responsible for deleterious effects in their own right.

Eating low carb to manage hyperglycemia at the expense of throwing hyperNEFAmia even further out of whack may not be the best strategy in the long run. I'm particularly concerned regarding hypertension and SCD.

I keep waiting to find SOMETHING to alleviate my concerns over elevated NEFA levels. Because a circulating free fatty acid is a circulating free fatty acid. It matters not to our bodies whether that FA was released from overstuffed fat cells or stripped off a chylo.

I don't know if you've read this post of mine:
The Progression of Insulin Resistance, or LynMarie Daye's similar offering Fat Fails First?.

My research seems to be pointing in this direction -- that the NEFA problem comes first, the hyperglycemia later. Both do damage.

I can rather effortly maintain my 80-100 lb weight loss eating fairly low carb most days (although if I don't keep added fats down I'm toast) with the occasional splurge for vacations and special occasions. Will this catch up to me, however? Will I, like some long term low carbers, reach a point where I can no longer tolerate any carbohydrate? Lately I've been working a few carbs in on a more daily basis -- mostly out of concerns I have over my recent reading material.
Peter said…
Hi Carb Sane,

To me both of the links above, and the post on "Mass MUST be conserved" (a sentiment I wholely endorse) do nothing to explain where the caloric surplus comes from. Glutton and sloth does not hit the spot for me. It's a bit like noting that alcoholism results from chronic excess consumption of alcohol. It's a statement of the obvious without explanantion.

WHY do people consume excess calories? This to me is a cross talk failure between pancreas, liver and adipocytes until the adipocytes fail. But they won't fail if the liver responds correctly to insulin because there will never then be the systemic glucose penetration nor the essential hyperinsulinaemia to manage this, subsequently expanding fat cells to the point of leaking FFAs. Liver comes first. Adipocytes fail subsequently...

Type two diabetics suffer post prandial hyperglycaemia for many years before fasting glucose rises. As the insulin rises to correct the systemic hyperglycaemia it locks dietary fat in to adipocytes. They don't start leaking until they are over full. They wouldn't get overfull without insulin and the insulin wouldn't have been there without failure of the liver to control systemic glucose.

Think hepatopathy. Any better explanation?

CarbSane said…
I don't think the caloric surplus is any great mystery. It is a combination of our obsession with caloric beverages, huge portions and eating "out" or pre-packaged foods too much. We ate carbs before we started adding 300-500 cal/day to our diets. It's a fallacy that we eat lower fat now than we did then. As a %, slightly, but some of those added calories are fat, although most admittedly came from more carbs. However hyperinsulinemia is in response to resistance to insulin's action. It's not from post prandial spikes. To me, "using" your insulin does not seem to be a viable mechanism for developing insulin resistance and diabetes. When I was a kid, "kid's" cereals were even more sugary and just as refined as they are today. Wonder bread was the staple sandwich bread for lunch. Twinkies and Hostess cupcakes were packed in many a lunch box. Kool Aid or milk and cookies were served at snack time. It's not the macronutrients that "started" this cascade IMHO.

I'm not sure there's a fixed progression, but we'll have to agree to disagree on this. To me, the stuffed "sick" fat spilling NEFAs that then act on the liver (hepatic lipid accumulation and/or signalling glucose production b/c NEFA's are released in fasting). It is certainly possible that hepatopathy as you call it occurs at the same time so that insulin doesn't suppress gluconeogenesis. Hyperglycemia is not the cause of obesity IMHO.

I've also blogged on direct pancreatic stimulation by LCFFA's. This is likely the mechanism by which increased fat mass increases basal insulin levels. It is the body's attempt to keep fats where they belong. The correlation of obesity and basal insulin seems to hold even in insulin-sensitive/non-diabetic obese.

Observationally speaking, this is why diabetes follows obesity, not the other way around.
James Krieger said…
It's also established that people selectively underreport "bad" foods. Given that fat was demonized in the 80's and 90's, fat intake is likely underreported in dietary surveys
CarbSane said…
Yeah, and then there's the American version of a salad for lunch or dinner instead of a "big meal". Thing is, it is easy to add 4-500 calories of dressing alone to that salad. Add crumbled cheese, a scoop of tuna salad, etc. It adds up. On TGIFriday's menu, the least caloric dinner salad weighs in at ~600 calories without the dressing, while the most caloric contains 1800 calories (dressing included in description). And yet many people think they are eating something more diet and health conscious by eating a salad. There are few items on the entire menu under 500 cal for a portion. Even on the "Right Portion Right Price" page, only 2 items come in under 500 calories (incidentally one low fat chicken, and one low carb shrimp dish). The rest start over 800 calories on up to almost 1200 calories. Most of the sides are over 400 cal (on up to over 900 for loaded mashed potatoes). It's crazy.

I don't eat there often (we have the menu from an emergency visit Christmas Day!), but the place is always packed and it is reasonable to assume once a week patronage for some of these folks. A "regular" meal (add a drink and ...) can easily top caloric needs for two days. I can't imagine preparing the same amount of food at home.
James Krieger said…
In line with what you're saying here, there's observational data that indicates that people who eat out more frequently are more likely to be overweight and gain weight over time
Anonymous said…
Excellent post. Keep posting such kind of info on your blog.
Im really impressed by your site.
Hello there, You've performed a fantastic job. I'll definitely digg it and in my opinion
recommend to my friends. I'm sure they'll be benefited
from this site.

my web site diets that work