Search This Blog

Sunday, 15 May 2016

Dietary fat type - saturated or unsaturated - does it make a difference to glycaemic control?

This is a section from a paper I'm writing about hepatic glycogen control, this part concerns the effect of dietary fat type on the insulin response. Spoiler alert: you will be surprised how little sound evidence there is on a subject about which so many pronounce so confidently.

Carbohydrate feeding stimulates the release of glucagon from delta cells in the gut and pancreatic alpha cells.[1] Glucagon is the hormone that elevates blood glucose by stimulating gluconeogenesis, but this is a delayed response; the most immediate glucose-elevating effect of glucagon is to induce glycogenolysis. In healthy metabolism, after eating a carbohydrate meal the paracrine effect of the phase 1 insulin response rapidly suppresses this glucagon release and the hepatic endocrine action of insulin inhibits the action of glucagon in the hepatic parenchymal cell, so that both gluconeogenesis and glycogenolysis are fully inhibited.[2,3]

Figure 1: Showing glucagon and insulin response to carbohydrate in normal metabolism

In type 2 diabetes, the delayed insulin response to a carbohydrate meal results in a longer elevation of glucagon; hepatic insulin resistance also reduces the inhibitory effect of insulin on glucagon action in the liver.
What is the value of this normal brief glucagon response to carbohydrate feeding? Glycogenolysis is a glycolytic process (glycogen -> glucose-6-phosphate -> lactate) which generates ATP in the glycogen-storing parenchymal cell; a brief and minor increase in glycogenolysis might be a preparatory adaptation, priming the cell for rapid glycogen synthesis from incoming glucose.
The delayed insulin peak from the beta cell of the diabetic pancreas (suggested mechanisms include ectopic fat accumulation in the beta cell, and/or cytokine interference with its function) allows a longer action of glucagon that is maladaptive in the context of a carbohydrate meal, and therefore the consumption of carbohydrate causes post-prandial hyperglycaemia by stimulating the release of glucose from glycogen and inhibiting its non-oxidative disposal in persons with type 2 diabetes.
This is an immediate cause of elevated PPPG that is rapidly corrected once carbohydrate is restricted.
In a study of 6 subjects with diabetes a simulated phase 1 and phase 2 insulin release during a hyperglycaemic clamp resulted in a 90% suppression of hepatic glucose production at 20 minutes, compared to a 50% suppression at 60 minutes from a simulated phase 2 response alone.[4]
However, a study of enhanced phase 1 insulin response in 14 elderly patients with diabetes found that phase 1 insulin response was not important in the regulation of hepatic glucose output or peripheral glucose disposal in these patients.[5]

1:02 The differential effect of fat type on the phase 1 insulin response

Does the type of fat in the diet influence the phase 1 insulin response? Below is the insulin response to a mixed meal containing two different fats – butter (SFA) and olive oil (MUFA) in 10 women with gestational diabetes mellitus. It will be seen that the butter-containing meal provoked a more rapid insulin response, and as a result both insulin and glucose area-under-the-curve (UAC) was reduced with the butter meal, and post-prandial plasma glucose at 2 and 3 hours was significantly lower compared with the olive oil meal.[6]

Figure 3: Plasma glucose response to a meal with olive oil (MUFA) or butter (SFA) in women with gestational diabetes

This difference may be due to other factors present in the fats, as butter contains 3% c9t11 CLA and olive oil supplies 11% linoleic acid (LA), compared to 2% in butter. c9t11 CLA improves insulin sensitivity compared to LA in prediabetic men.[7] Elevated plasma levels of trans-palmitoleic acid, mainly found in dairy and ruminant fat, are also associated with a reduced incidence of diabetes and insulin resistance.[8,9]
Wistar rats fed soybean oil (60% LA) for 4 weeks had significantly lower glucose-stimulated insulin responses compared to rats fed lard (10% LA) whose insulin responses were similar to those of rats fed a low fat control diet.[10] A study of inhibition of fasting FFAs by nicotinic acid (NA), replaced by soybean oil (Intralipid) and heparin, in 10 healthy male subjects found that FFAs were essential for insulin response to glucose in fasting humans.[11] A further study in rats in which serum FFAs were inhibited by NA and replaced by infusions of soybean oil or lard with heparin found that serum saturated fatty acids were essential for the first-phase insulin response to glucose, which was suppressed by high levels of unsaturated fatty acids, which only supported a second-phase response.[12]

1.03 The differential effect of fat type on insulin sensitivity

While some feeding studies show that meals high in saturated fat result in higher glucose levels than meals high in monounsaturated fat, others show the opposite, while yet other studies find no difference, as summarized in Jackson et al 2005.[13] The saturated fat source most likely to be used in such feeding studies is palm oil, which is the dietary fat with the highest concentration of palmitic acid, which was mixed with cocoa butter, the dietary fat with the highest concentration of stearic acid, in the saturated fat arm of the feeding study in that paper, which showed higher glucose AUC in the saturated fat arm. Palmitic and stearic acids are the main endogenous saturated fatty acid products of de novo lipogenesis (DNL) and serum levels of these fatty acids are known to be correlated with the carbohydrate content of the diet. Thus such a study may not accurately represent the effects of the mixture of fats found in normal diets, especially in the context of a low carbohydrate diet. Of randomised long-term studies, the LIPGENE study found no effect of fat type, whereas the KANWU study, a study cited as showing a worsening of insulin sensitivity (albeit non-significant) after feeding saturated fat compared to monounsaturated fat for 3 months, noted that the favourable effects of substituting a MUFA diet for a SFA acid diet on insulin sensitivity were only seen at a total fat intake below median 37E%.[14,15]

1.04 Recommendations regarding fat type in very low carbohydrate diets

The 2006 experiment by Krauss et al was a test of the hypothesis that saturated fat in a carbohydrate-restricted diet would influence the effect of the diet on the atherogenic dyslipidemia produced by hyperinsulinaemia in the context of insulin resistance.[16] Men (n=178) with a mean BMI of 29.2 (+/- 2) were randomized to four different diets – 54% CHO, 39% CHO, 29% CHO with 9% SFA, and 29 % CHO with 15% SFA, for twelve weeks, including a 5 week period of calorie restriction followed by a 4 week period of weight stabilization.
Concentrations of apo B, a measure of total atherogenic particle concentrations, as well as total:HDL cholesterol, an integrated measure of CVD risk, decreased similarly with both the higher- and lower-saturated-fat diets. Moreover, the changes in LDL cholesterol for both the lower- and higher-saturated-fat diets (−11 and 1 mg/dL, respectively) were considerably more beneficial than were those predicted on the basis of studies that used diets with a more conventional macronutrient composition (−1 and 9 mg/dL, respectively). The difference in LDL cholesterol between the two diets was due to the appearance of larger, less atherogenic LDL particles in those on the 15% SFA diet; both diets saw similar reductions in levels of atherogenic small, dense LDL (sdLDL) particles. The ratio between triglycerides and HDL cholesterol correlates with serum insulin and insulin sensitivity; the TG/HDL ratio was the same with both 9% and 15% SFA at 29% CHO.[17]

Fig 3: glucose response to fasting and carbohydrate-free diet

It is considered that very low carbohydrate diets partially mimic the fasting state. In a 2015 randomised cross-over study by Nuttall et al, 7 men and women with untreated type 2 diabetes were placed on a control diet (55% CHO, 15% PRO, 30% FAT), a carbohydrate-free diet (3% CHO, 15% PRO, 82% FAT), or fasted for 3 days.[18] On the third day of the carbohydrate-free phase, overnight fasted blood glucose concentrations were 160 mg/dl compared with 196 mg/dl in the standard diet and 127 mg/dl in the fasting phases. Carbohydrate restriction also led to a rapid drop in post-prandial glucose concentrations and glucose area-under-the curve decreased by 35% in the carbohydrate-free phase compared to the standard diet. It was found that carbohydrate restriction accounted for 50% of the reduction in overnight glucose concentrations and 71% of the reduction in integrated glucose concentrations in the fasted phase compared with the standard diet phase. It is notable that human depot fat, which is the major fuel source in the fasting state, consists of (approximately) 55% monounsaturated fat and 30% long-chain saturated fat, with the remainder consisting of smaller amounts of polyunsaturated fats and medium-chain saturated fats. It has been noted that a 50:50 mixture of ghee and olive oil has a fatty acid composition of 32% saturated fat (some of which is short and medium chain fatty acids, leaving 25-28% from the long-chain saturated fats, palmitic and stearic acids), 50% monounsaturated fat, and 7% polyunsaturated fat, approximating reasonably well the composition of human depot fat. Thus there is insufficient evidence to support recommendations restricting saturated fat in very low carbohydrate diets. However, there is some evidence for preferring full-fat dairy foods to other sources of saturated fat in the diet, with regard not only to glycaemic control but also cardiovascular risk, based on observational studies [19,20,21].
Adherence to diets is likely to be greatest when the rationale for choices is simple and convincing, when the diet is adequately nutritious, and when food is culturally appropriate – that is, when the diet is made up of foods that are already familiar and liked.
It should also be noted that both carbohydrate-free diets and fasting appear to be well-tolerated in the feeding studies we have described, with no adverse events reported during or after any study.


[1] Lund A, Bagger JI, Wewer Albrechtsen NJ et al. Evidence of Extrapancreatic Glucagon Secretion in Man. Diabetes. 2015 Dec 15. pii: db151541. [Epub ahead of print]

[2] Raskin P, Unger RH. Hyperglucagonemia and Its Suppression — Importance in the Metabolic Control of Diabetes. N Engl J Med 1978; 299:433-436.

[3] Sonksen P, Sonksen J. Insulin: understanding its action in health and disease. Br. J. Anaesth. (2000) 85 (1): 69-79.

[4] Luzi L, DeFronzo RA. Effect of loss of first-phase insulin secretion on hepatic glucose production and tissue glucose disposal in humans.
American Journal of Physiology - Endocrinology and Metabolism Published 1 August 1989 Vol. 257 no. 2, E241-E246

[5] Meneilly GS, Elahi D. Physiological importance of first-phase insulin release in elderly patients with diabetes. Diabetes Care. 1998 Aug;21(8):1326-9.

[6] Ilic et al, Comparison of the effect of saturated and monounsaturated fat on postprandial plasma glucose and insulin concentration in women with gestational diabetes mellitus. American Journal of Perinatology 1999

[7] Rubin D, Herrmann J, Much D, et al. Influence of different CLA isomers on insulin resistance and adipocytokines in pre-diabetic, middle-aged men with PPAR╬│2 Pro12Ala polymorphism. Genes & Nutrition. 2012;7(4):499-509. doi:10.1007/s12263-012-0289-3.

[8] Mozaffarian D, Cao H, King IB, et al. Trans-palmitoleic acid, metabolic risk factors, and new-onset diabetes in U.S. adults: a cohort study. Ann Intern Med. 2010 Dec 21;153(12):790-9.

[9] Yakoob MY, Shi P, Willett WC, Rexrode KM, Campos H, Orav EJ, Hu FB, Mozaffarian D. Circulating Biomarkers of Dairy Fat and Risk of Incident Diabetes Mellitus Among US Men and Women in Two Large Prospective Cohorts. Circulation AHA.115.018410 Published online before print March 22, 2016

[10] Dobbins RL, Szczepaniak LS, Myhill J, et al.  The composition of dietary fat directly influences glucose-stimulated insulin secretion in rats. Diabetes June 2002 vol. 51 no. 6 1825-1833.

[11] Dobbins RL, Chester MW, Daniels MB et al. 1998: Circulating fatty acids are essential for efficient glucose-stimulated insulin secretion after prolonged fasting in humans. Diabetes. 1998;47(10): 1613-1618,

[12] Stein DT, Esser V, Stevenson BE, et al. Essentiality of circulating fatty acids for glucose-stimulated insulin secretion in the fasted rat. J Clin Invest. 1996 Jun 15; 97(12): 2728–2735.

 [13] Jackson KG, Wolstencroft EJ, Bateman PA, Yaqoob P, Williams CM. Acute effects of meal fatty acids on postprandial NEFA, glucose and apo E response: implications for insulin sensitivity and lipoprotein regulation? Br J Nutr. 2005 May;93(5):693-700.

[14] Tierney AC, McMonagle J, Shaw DI et al. Effects of dietary fat modification on insulin sensitivity and on other risk factors of the metabolic syndrome--LIPGENE: a European randomized dietary intervention study. Int J Obes (Lond). 2011 Jun;35(6):800-9.

[15] Vessby B, Uusitupa M, Hermansen K et al. Substituting dietary saturated for monounsaturated fat impairs insulin sensitivity in healthy men and women: The KANWU Study. Diabetologia. 2001 Mar;44(3):312-9.

{16] Krauss RM, Blanche PJ, Rawlings RS, Fernstrom HS, Williams PT:
Separate effects of reduced carbohydrate intake and weight
loss on atherogenic dyslipidemia. Am J Clin Nutr 2006,

[17] Feinman RD, Volek JS. Low carbohydrate diets improve atherogenic dyslipidemia even in the absence of weight loss. Nutrition & Metabolism 2006;3:24.

[18] Nuttall FQ, Almokayyad RM, Gannon MC. Comparison of a carbohydrate-free diet vs. fasting on plasma glucose, insulin and glucagon in type 2 diabetes. Metabolism - Clinical and Experimental. 2015;64(2):253 – 262.

[19] Ericson, U, Hellstrand, S, Brunkwall, L, Schulz, C-A, Sonestedt, E, Wallstr├Âm, P, et al. Food sources of fat may clarify the inconsistent role of dietary fat intake for incidence of type 2 diabetes. AJCN 2015;114.103010v1

[20] Praagman J, Beulens JWJ, Alssema M et al. The association between dietary saturated fatty acids and ischemic heart disease depends on the type and source of fatty acid in the European Prospective Investigation into Cancer and Nutrition–Netherlands cohort. Am J Clin Nutr. ajcn122671

[21] De Oliveira Otto MC, Mozaffarian D, Kromhout D et al. Dietary intake of saturated fat by food source and incident cardiovascular disease: the Multi-Ethnic Study of Atherosclerosis. The American Journal of Clinical Nutrition. 2012;96(2):397-404. doi:10.3945/ajcn.112.037770.


Peter said...


George Henderson said...


Here's a good paper on the glucagon response to glucose in diabetic humans

Abnormal response of pancreatic glucagon to glycemic changes in diabetes mellitus.

Remember the butter-with-potatoes study you parsed recently; that was ad lib showing improved satiety (hey maybe insulin= satiety hormone in that one), then there are these similar studies showing higher insulin, lower BG when spuds and butter added.(thanks Grant for these).

We might be starting to explain the dairy fat vs diabetes observational link. Diabetes in epidemiology can be diagnosed with OGTT but is often just based on fasting PG and HbA1c cut offs. So if this effect of lower BG is maintained for decades, it could account for the differential diagnosis (which is not huge, about 13%). But what this needs, is for this increased insulin response not to be detrimental to the pancreas. Plainly if butter is not associated with increased T2D then such a boost in insulin response to a given amount of CHO is not detrimental long-term.
There's another factor - in that first butter-spud feeding study, the effect maxes out at 15g butter vs 5g - if 50g butter is added, there's no more difference to insulin or glucose.
So - plenty extra calories and satiety but no effect on glycaemia. Pity there's no olive oil control in this paper.

Why don't the human RCTs show the effects seen in the NA treated rats? Humans have a free flow of FFAs. What they ate months ago, what they made by DNL, is all part of the mix. And the RCTs are testing MUFA oils, not so much soybean. I only found one human NA study, with FFAs suppressed, and the only thing tested was soybean oil vs no fat. Soy was better than nothing, but it would be great to compare it with lard. Or indeed butter oil. (follow up to previous butter-potato study)

"Our data indicate that the glucose and insulin response to butter is unique when compared with the three other fat sources varying in their fatty acid composition."

Also - note the cognitive dissonance in those final sentences

George Henderson said...

That last link should be

Passthecream said...

George, I've never heard any of your other musical performances but this is a virtuoso solo.

The mention of depot fat composition - any clues as to which dietary regime that might be associated with?

I am one of those, eating fairly low carb and restricting mealtimes to about an 8 hour band, who finds fasting pre-p glucose high, to ~8mml and which drops quickly as soon as i have e.g. a cup of coffee with cream. I'm trying to decode this as a mix of physiological ir versus low fasting insulin or liver ir >> glucagon dominance. But?

George Henderson said...

@ Passthecream - thanks.

I preferred 1960s depot fat studies as showing a more natural mix of fats. Stephan Guyenet has documented a huge increase in LA in the US recent decades
and I didn't want that polluting the thesis.

Cream gives a very slight but real insulin rise; coffee also elevates insulin but this is a round-about response to what is (meant to be) an increase in epinephrine and hence glucose.
I can only guess that the large somatostatin and small insulin response to cream is blocking lipolysis and proteolysis, that if you're habituated to caffeine there isn't much epinephrine. It's a mystery! It makes sense to me that when energy comes in from the gut hepatic energy output should be reigned in, and glycogen release is that currency.

George Henderson said...

That is, it's a mystery - but I'm glad to hear that it happens!

Robert Simms said...

Regarding part 1.02, the difference in insulin response between butter and olive oil could be partly explained by the protein in butter. Dr Jason Fung has a page on the incretin effect:

George Henderson said...

Thanks for that link - I'm not sure that there's enough protein in butter, 0.9%, I think there might be a threshold for these effects. If not, the gut hormones would respond to proteins from the gut lining?

Rattus said...

random question: if the entire purpose of a low CHO diet is to mimic the fasted state, why are we not attempting to eat lipid ratios that mimic stored body fat, and instead favoring SFA over MUFA?

George Henderson said...

Excellent question Rattus.
Short answer is, we should try to mimic depot fat (this gives enough SFA for any known benefit of higher SFA).
Long answer is, that SFA and MUFA are somewhat interconvertible so there is some leeway.
I make depot fat (going back to before the seed oil craze) to be roughly 30% SFA, 60% MUFA, and 10% various PUFA and MCSFAs.
This is equivalent of a 50/50 mix of butter and olive oil, with a little coconut oil thrown in, or rather more beef fat and less olive oil, or pork fat with coconut. The PUFAs will be in the protein foods and veges so we needn't worry about those.

Rattus said...

I like the idea of creating the ratios by mixing fats.

Another question is, assuming we are burdened with an unnatural amount of pufa in our depots, would it be wise to try to keep the percentage as close to 0 as possible? In my experience, it doesn't take much pufa to experience negative effects, especially in the context of a high fat diet. In other words, looking at total volume of pufa consumed, it will automatically be higher due to the advent of seed oils AND that we are on a high fat diet. Olive oil has quite a bit of omega 6, as do egg yolks. Obviously, it's impossible to avoid a certain amount.

George Henderson said...

To me it makes sense to start a diet correction by emphasising fats very low in PUFA, with a few exemptions for foods high in cholesterol, AA, EPA, and DHA such as egg yolks.
This means using tallow, ghee, coconut. Then once depot PUFA has been drawn on for a while it should be okay to increase PUFA from nuts, seeds, olive oil.
However in practice it is probably enough to avoid each every source of vegetable seed oils forever after.