Tuesday, May 20, 2008

Physiological insulin resistance; Dawn Phenomenon

What is the Dawn Phenomenon (DP)? A nice simple definition is available from here:

"The dawn phenomenon is a term used to describe hyperglycemia or an increase in the amount of insulin needed to maintain normoglycemia, occurring in the absence of antecedent hypoglycemia or waning insulin levels, during the early morning hours. To be clinically relevant, the magnitude of the dawn increase in blood glucose level should be more than 10 mg/dL or the increase in insulin requirement should be at least 20% from the overnight nadir. Controversy exists regarding the frequency, reproducibility, and pathogenesis of the dawn phenomenon. Approximately 54% of patients with type 1 diabetes and 55% of patients with type 2 diabetes experience the dawn phenomenon when the foregoing quantitative definition is used"

OK, I lied about the simplicity.

If you go back to 1988 this group seemed to think that the Dawn Phenomenon was pretty straight forward and amenable to pharmacological management. In the early hours of the morning humans have a growth hormone (GH) surge. GH causes lipolysis, lipolysis releases free fatty acids. No muscle wants glucose when it has access to free fatty acids. Muscle thus becomes insulin resistant and blood glucose rises. They studied type one diabetics as doing this eliminates all of those messy insulin responses to glucose that normal people produce.

Give an anticholinergic, block the GH surge and you block the DP. All nice and simple.

Then this group, in 1992, went out to check if this was true and gave a bolus of GH, again to some type one diabetics. Any old time of day as far as I can tell.

As expected GH caused a rise in FFAs but no insulin resistance in this paper! Quite how they managed this is a bit beyond me. FFAs should produce insulin resistance. Both papers report comparable peaks in GH but neither gives an AUC for GH. It's difficult to compare FFA changes between papers.

Then, when you get down to the nitty gritty, you find that the second paper was quite careful to produce only a SHORT physiological burst of GH, shorter than occurs at night in humans. They were ONLY looking for the effect of GH on insulin resistance, so they had to keep FFA changes to a minimum. This looks to be a very carefully crafted sentence to me, it's the "summing up" in the abstract (the section of the paper which actually gets read):

"Since no significant effect on glucose metabolism was recorded, we do not presently find evidence to support a primary role for small surges of GH in the pathogenesis of the dawn phenomenon"

A translation might read:

"A bolus of GH which is significantly lower than that needed to produce the lipolysis necessary for the Dawn Phenomenon does not produce the Dawn Phenomenon"

People don't work like this! The nightly GH surge in humans does crank up FFAs. Eliminating this effect from your study won't help elucidate what's happening in the DP, except to say it's not GH per se, but it's still the down stream effects of GH that matter. The 1988 paper looks far more convincing to me.

Current thinking seems to have forgotten about FFAs but does come up with the concept that there is no drop in insulin levels involved. This is interesting in so far as Dr Bernstein suggests that rapid hepatic breakdown of insulin is the cause of the problem. I've yet to see any evidence of this mechanism, unusual for a Dr B idea.

The other main support for FFA involvement is the roll of eating in terminating the DP. This has been discussed many times on the Bernstein forum. Many type two diabetics develop a vicious DP on LC eating and, if they continue to fast through the morning, their blood glucose will just keep going up and up. Eating CARBS stops this, presumably the carbs get insulin high enough to get ahead of the effects of FFAs on muscle, while the extra insulin can shut down hormone sensitive lipase and so drop FFA production... This is physiological insulin resistance taken to pathological extremes.

It's interesting to speculate whether it is the facility to indulge in lipolysis to excess, possibly related to the absolute fat mass available or insulin resistance in those adipocytes, or failure of the cross talk between alpha and beta cells in the pancreas or failure of hepatic vagal nerve supply which makes the DP such a big problem in some diabetics but not in others. I don't know.

Also the how and why of bed time alcohol blocking the DP effect is another big unknown, but it seems to work for many diabetics. Whether this is an hepatic effect, a GH effect or a lipolysis effect seems wide open. Wine lovers with DP just seem grateful when it works for them. Evening alcohol incidentally also drops my morning glucose in to the 4 point something range.

So it looks to me as if the DP is an insulin resistance phenomenon by which GH induced lipolysis bumps up morning glucose. I think this is physiological. When lipolysis is one step ahead of the extra insulin production needed to keep blood glucose in an "acceptable" range you can easily end up with an elevation of FFAs, insulin and glucose all at the same time. In fact, broken insulin sensitivity might well do this, and so you end up with type two diabetes markedly worsened by a GH surge, every night...



Unknown said...


My understanding is that cortisol levels are cyclical, with the peak being just before rising. Does this enter into your analysis of the situation?


Peter said...

Hi Porter,

As I understand it for diurnal animals the cortisol nadir is at 4am and the peak at noon. This would tend to drop insulin resistance at the time when it is increasing so doesn't appear to be a factor...

Obviously rats and mice are nocturnal...

BTW My wife assures me that some dogs, as diurnal animals, show a marked dawn phenomenon, noted when she was running 24h glucose curves on them back when she was an intern.


ItsTheWooo said...

Interestingly this is the first time I've even heard the theory that growth hormone causes DP. I had learned via research and again in school that the morning increase in cortisol (also circadian pattern) is what causes DP. It does make sense that GH could also contribute.

The reason growth hormone mediated increase in FFAs did not necessarily produce IR may be because GH being anabolic also encourages cell division and repair; cellular need for energy may have increased to the point that increased FFAs were not hogging up mitochondria (thus no compensatory IR). On the other hand, even small amounts of cortisol will increase blood glucose it is catabolic. A larger dose of GH, however, will induce insulin resistance but small doses do not do it as easily as cortisol.

Eating even before bed will minimize the DP the next morning. Something about fasting modifies how both GH and cortisol works on the liver (and liver production of glucose is a major cause of diabetic hyperglycemia). I don't agree with Dr B that the problem is insulin being degraded, it seems to be that fasting causes the liver to actually make more sugar from normal release of GH and cortisol. The balance of insulin and glucogon seem to modulate how GH and cortisol both work; when glucagon, is more dominant as in fasting then the body is more catabolic, and the liver starts making more sugar in response to normal pulses of GH and cortisol. When insulin is dominant (and/or hepatic insulin sensitivity in the liver is normal) then this does not happen.

Diabetics with the worst DP probably have the greatest hepatic insulin resistance, or deficiencies in insulin production. The former would be well served to eat more to bump up insulin production, or to try more drugs that inhibit liver gluconeogenesis. The latter need to get on or tweak their insulin therapy.

I had long observed that the ol' booze triggered insane hypoglycemia in me (I'm a reactive hypoglycemic and fasting my sugar is barely above 60) that can last many hours after I've sobered up. Originally my thought was alcohol might induce insulin resistance and hyperinsulinemia. When I did research on it, I came to learn alcohol inhibits gluconeogenesis because metabolizing the alcohol interferes with this process. For people with hypoglycemia, you don't feel the sugar drop until you stop drinking because alcohol is preferentially metabolized and can provide energy needs. However once blood alcohol level starts to drop, your body greedily sucks out all of your sugar. The liver is unable to support more sugar production because it is getting rid o' toxins. End result? Massive hypoglycemia after drinking ends. Everyone finishes off a night of drinking with a carb binge... this is why. The lower sugar can persist for many hours after the blood alcohol returns to normal because the liver is still cleaning up.

Bad for a hypoglycemic on a low carb diet. Good for diabetics or prediabetes.

Anonymous said...

Peter, I think alcohol prevents the dawn phenomenon, because it reduces liver glycogen, preventing the rise in blood sugar from occurring. I've investigated the effects of alcohol esp as regarding low-carb diets and ketosis. I concluded that alcohol's highly ketogenic. It depletes liver and muscle glycogen, forcing a more rapid adaptation to ketosis. I have speculated that drinking alcohol in lieu of carbs would provide lots of benefits (including weight loss). I found a study backing up my theory, too. Subjects were fed iso-caloric, low-calorie diets. One got 10% wine and the other 10% grape juice. Wine drinkers lost 25% more weight. Here is the study's abstract.


ItsTheWooo said...
This comment has been removed by the author.
ItsTheWooo said...

The problem with the theory that alcohol is ketotic is that alcohol does not seem to actually induce fat burning, all it does is cause glucose depletion.

Eating long chain fatty acids like butter and beef is ketogenic, because this automatically gives me a "balanced feeling" and greatly satiates hunger for a long time which, for me, is a classic sign of burning fats very well. On the other hand, alcohol merely depletes my blood sugar without augmenting fat burning, sending me into severe hypoglycemia and prowling through the fridge for something, anything to eat.

EDIT: If the problem is hyperglycemia and insulin resistance in the liver then alcohol might actually help lose weight.
But, most people become at least a little hypoglycemic after drinking alcohol as evidenced by the rate of carb binges that seem to occur after a night of drinking (my local bar scene is studded with fast food restaurants for a reason). Most people do not have fasting hypoglycemia, which is why.

Peter said...

The more I think about it the more I would agree with you, ItsTheWoo, the problem is more likely glucose dumping from the liver rather than peripheral insulin resistance. There's another post on mice coming which would fit in with this better.

Re alcohol: Zero carb people tend to be very cautious about alcohol. If you don't have any glycogen to start with, depleting it may not be an option and the hypoglycaemia may be unpleasant. I remember anecdote, Barry Groves I think, that modest alcohol is neutral re weight loss and high dose counter productive. Don't follow the metabolism on that particular idea.

The wine for weight loss paper is interesting. Of course, with my biases the obvious explanation comes from this paper

White wine doesn't spike insulin. Grape juice does. Back to effects on hormone sensitive lipase...


Anonymous said...

"The problem with the theory that alcohol is ketotic is that alcohol does not seem to actually induce fat burning, all it does is cause glucose depletion."

Ketosis isn't just fat-burning, but ketone burning. Alcohol breaks down to acetone, which is a ketons, thus alcohol is ketogenic. Going back to the very first low-carb dieter, Mr. William Banting had 4-6 drinks/day, and lost weight on a low-carb diet. The theory that the alcohol prevent fat-burning is something Dr. Atkins made up, I think. Dr. Wolfgang Lutz said alcohol is fine if you limited carbohydrates. Alcoholics are often lean, because the body can't create fat from alcohol efficiently (if at all), nor can it make glucose.

Maybe you are right that alcohol is bad for hypoglycemics, but I do not think alcohol should be blamed. The hypoglycemic ruined their liver and pancreas with carbs, PUFA oils, and hydrogenated fats. The zero-carbers are probably in the same boat. They have extremely messed-up bodies and we shouldn't base our diets on what someone else experiences.

Bingeing on carbs is simply a habit controlled by will power. I noticed years ago that it's better to eat a high-fat food, like cheese or nuts, when I felt sick or light-headed. A typical doctor would tell diabetics to eat candy or soft drinks to beat a hypoglycemic attack. And they say that eating fats could kill you, in that situation. As with most of the medical advice today, nothing could be further from truth IMO.

What I would eat to stabliize blood sugar fast is pure fat like butter, semi-refined coconut oil, macadamia oil, cream cheese, or suet. I don't get hypoglycemia any more thanks to my low-PUFA diet. Nor do I feel any effects from alcohol now.

ItsTheWooo said...

I think we need to focus on these concepts:

1) SOME people have fasting hyperglycemia.

2) SOME people have fasting hypoglycemia.

3) Alcohol is hypoglycemic, because it inhibits gluconeogenesis

Assuming insulin production is in tact, the difference between these two people is usually insulin resistance, especially hepatic insulin resistance. The first group makes too much blood sugar too often. The second group does not, and may even have innate defects in gluconeogenesis (or, excessive insulin production as in very mild cogenital hyperinsulinemia).

If we consider what alcohol does (hypoglycemic, due to inhibition of gluconeogenesis)
What do you think will happen to the first group of people? They'll lose weight better, because less sugar means less insulin. Their diabetes may be better controlled for the same reasons.

What do you think will happen to the second group? As a member of the second group, I'll tell you: massive hypoglycemia, and compensatory ravenous eating. My sugar is low, and my hunger ravenous, even a day after drinking if I drank enough. This is how I gained 160 pounds, because eating carbs smashes my blood sugar causing lows and highs and then lows again. Alcohol is just another hammer as far as I"m concerned.

Alcohol make break down into acetone but it is not ketogenic in the sense that low carb eating is ketogenic. Low carbing does NOT inhibit gluconeogenesis, whereas alcohol does.

Pardon me but I did not "ruin my liver and pancreas" by "binging on carbs". The irony is, the hyperglycemic who benefits from alcohol is usually the one who has long term damage from carbohydate (hepatic insulin resistance). A tendency to fasting hypoglycemia is more of a default and healthy state, and a genetic trait. If my fasting sugar is low, it means my insulin sensitivity and insulin production is normal. If I don't adulterate my body with toxins such as alcohol and simple carbs I have no unpleasant symptoms.
On the other hand, hyperglycemia is is almost always acquired (although genes control who acquires it and who doesn't).

Anyway, I think we need to just accept that different blood sugar states are caused by different issues, and those issues require different interventions. I am not attacking alcohol, if it helps your blood sugar, enjoy it in moderation. But I'm merely defending the other point: alcohol is NOT ketogenic, it is merely hypoglycemic via inhibiting sugar production. It's positive effect on fat burning is only present in those people who have abnormally excessive hepatic gluconeogenesis. FOr people like myself with normal or mildly deficient liver glucose production, alcohol offers zero benefits and may encourage binging on carbs (which seems to be the more COMMON outcome, as MOST people are not fasting hyperglycemics).

Anonymous said...

"Alcohol is hypoglycemic, because it inhibits gluconeogenesis." (Woo)

So maybe it would be a good idea to eat a small amount of carbohydrates with alcohol, like William Banting. He never ate zero-carb diet, nor do many people here. He ate small bits of bread and potatoes.

"Alcohol [may] break down into acetone but it is not ketogenic in the sense that low carb eating is ketogenic. Low carbing does NOT inhibit gluconeogenesis, whereas alcohol does."

If you eat a high-fat diet, like JK uses, it will limit gluconeogenesis more than the high-protein low-carb diets. Gluconeogenesis isn't needed for ketosis. Inhibiting it might be good for some people who eat excess protein and/or carbs.

"If I don't adulterate my body with toxins such as alcohol and simple carbs I have no unpleasant symptoms."

You are assuming that alcohol is a toxin, but toxin's a relative term. If you eat corn starch, sugar, and corn oil, then casein may be toxic. Whereas if you ate a low-carb diet or a diet with no high-PUFA oil or refined carbs it might be harmless. Maybe your body is contaminated by toxins and that causes you to have reactions to alcohol. Others don't have those reactions.

I don't have any reaction to 1.5 oz of gin with 102 proof. In the past, I'd feel effects from less alcohol. Now, I feel nothing. This says that my body has been changed by my diet and is now more resistant to toxins as a result. There are many reasons why this could be true.

Ed said...

Anecdotally, Robb Wolf says that he used to be a big fan of intermittent fasting, but after observing it on his clients, he has become less-so. He says that about half his clients do well, but the other half do poorly.

The body is wonderfully complex...

Ken said...

Peter, here is the work (Gerich et al) from which Dr. B's (incorrect, I think -- but you can decide) conclusions are derived:
There are three regions of very independent insulin/glucagon hormonal regulation in the tissues:
1. peripheral, including muscle and adipose, at 1x
2. portal vein, at 40x (relative to peripheral)
3. islets, at 400x (relative to peripheral)
In the islets, alpha-cell secretion of glucagon is regulated by:
a. local insulin levels and transients (the latter most importantly for prandial response and BG drops and so forth)
b. portal BG levels (this is the basal, but not transient, regulator)
ALL diabetics (with the one exception being Gck-monogenic, aka MODY-2, for which there is no insulin secretion defect but only a glucose-sensing defect) have an insulin-secretion defect, and this in turn leads to hyperglucagonemia. It is the excessive glucagon secretion that is responsible for essentially ALL hyperglycemia in a diabetic (the meal stimulates the disregulated hormone secretion, but is not itself responsible for the BG rise to any significant degree).
Glucagon is the master hormone for whole-body energy storage and distribution in the liver. Insulin has no effect upon the liver EXCEPT when glucagon secretion is shut down (to zero) -- this is the non-diabetic response to carb, but in the diabetic the glucagon secretion is perversely increased. This is why diabetics are so intolerant of carb's, and cannot compensate with insulin. Whereas diabetics CAN compensate for dietary protein with insulin.
I am a diabetic. Such will observe much large hyperglycemic swings in response to:
a. infection
b. diurnal variations in hormones
c. exercise or anything else stimulating fight-or-flight (catabolic) hormones
By the way, this is also the reason for the sometimes-higher BGs of non-diabetics on low-carb diets -- their basal insulin levels are sometimes low enough to allow the stress hormones to raise BG above normal fully-fed homeostatic levels.
Almost ALL hormones in the body raise BG, and many of these have strong diurnal cyclic variation.
The dawn phenomenon, most fundamentally, is caused by the (basal, in this case) insulin deficiency of the diabetic. All diabetic phenomena derive from this defect, indirectly. Those who think otherwise are ill-informed. T2DM is the same -- not caused by IR, although higher genetic IR in the muscle tissues (which are dominantly responsible for BG clearance flux) is probably one requirement for susceptibility to acquisition.
I have a monogenic type (HNF1-alpha) of diabetes, and have very high insulin sensitivity as do ALL diabetics of this type (and some similar types). There is no known historical example of coexistence of HNF1-alpha and T2 diabetes. HNF1a types get all the complications, though.
I have two genetic conditions, both of which come from Nordic ancestry in regions where no plant food was available. One is CVID and the other is the diabetes. In evolution, "use it or lose it" often applies.

Peter said...

Thanks for the input Ken. Some interesting thoughts. Could you just clarify your comment about diabetics compensating for protein with insulin. This doesn't ring true for my canine diabetic patients who will spike glucose badly on a carbohydrate free diet...


PCTServices said...

I read a paper that stated an early pathology of T2, is the failure of FFA's to increase insulin secretion. So it would seem, that if FFA's shot up in the morning, and they failed to increase insulin release, while the muscles were running on those fatty acids, glucose would increase in the blood. Perhaps normal people increase FFA's AND insulin release so see much less effect. This would mean, it's not a drop in actual insulin release, but a failure to increase which would make it 'relatively' low for that situation.

Ken said...

Hi Peter,
Regarding your canines, I think our observations are in agreement despite mixing species. I was just pointing out that in my form of diabetes (and also in T1DM) dietary protein alone causes prandial hyperglycemia as you have observed in dogs. Since I eat almost no carb's at all, protein dominates my requirement for bolus insulin.
I probably should have spec'd "exogenous" (injected) insulin to be more lucid.
It is interesting that T2Ds retain normal (i.e. non-diabetic levels of) amino-acid stimulated insulin secretion (AASIS) typically for years after first diagnosis, and hence can maintain normal BG with diet alone. The early overt/diagnosed T2D has heavily degenerated GSIS but has not yet lost AASIS. Researchers studied this a bit in the 1970s, but seem to have lost interest.
I think a lot of the loss of function in T2DM is due to deranged, but still extant, beta cells. In my form of diabetes there is simply underexpression of insulin-related genes and proteins. In T1DM the surviving beta cells are normal-functioning. So retention of AASIS in early stages of overt T2DM should offer some clues about the pathology.