The poor, misunderstood...glycerol molecule

As an outsider to the field of nutrition, it’s puzzling to me when commonplace foodstuffs are surrounded by fear, uncertainty, and doubt. High profile additives like food dyes and artificial sweeteners get lots of attention in the mainstream media. These ingredients can be very profitable for giant food conglomerates, so while the corporate shills and conspiratorial kooks both spread their own particular brand of misinformation and confirmation bias, somebody somewhere is probably doing some real science to get at the real answer...eventually...hopefully.  High profile foodstuffs have a much better chance of being thoroughly investigated, tested, and their results peer reviewed. Relatively low profile additives may not get the same level of attention.

Glycerin is a very odd ingredient, indeed. I came across it during my foray into the science of low carb ice cream. Its metabolism in the human body is difficult for a layman like me to understand. Considering that edible glycerin is derived from animal or plant fats and glycerol is part of fat metabolism, one might assume it would be metabolized like a fat. Sometimes it is; but under different conditions it can be converted into glucose by the liver. Wait, what?! Is this stuff going to raise my blood sugar and insulin or not? Am I going to be kicked out of ketosis if I ingest this stuff? I’m afraid to say that the answer isn’t clear cut.

Glycerin can also be called glycerol, glycerine, food additive E422, and a few other possibilities. These names will be used interchangeably throughout this article depending on the quoted reference source. Just keep in mind that it’s referring to the same thing.

CAUTION: Not every form of glycerin is fit for human consumption. For example, diethylene glycol is toxic and should not be consumed. Always verify the product is safe to use!

Classifying glycerin

The FDA classifies glycerin as a carbohydrate. This nutritional information sheet explains:

According to James E. Hoadley, Ph.D. of FDA’s Office of Food Labeling, glycerin is labeled as carbohydrate and, if any claim is made regarding sugar content, also as a sugar alcohol.

Ah, so it’s a sugar alcohol.

When part of the fat molecule...glycerin is included in the weight of total fat in nutrition labeling. However, when added to a food as a separate ingredient, glycerin is labeled as part of total carbohydrate.

Uh, so...I count the glycerin that’s part of the fat as fat, but if glycerin is added, it’s counted as a carbohydrate. Uh...

There is no rational basis to consider glycerin as anything but carbohydrate. 

But you said...

A sugar alcohol (or polyol) is the hydrogenated derivative of a sugar. As noted above, glycerin is the hydrogenated derivative of glyceraldehydes and dihydroxyacetone, and thus is a sugar alcohol.

Okay! It’s a carbohydrate. Fine. So, how much of a glucose and insulin response is there? Certainly someone has tested this. I searched and found this information from David Mendosa’s article on net carbohydrates:

Atkins Nutritionals says that glycerine is another carbohydrate that has “a minimal impact on blood sugar.” Dr. Thomas Wolever, professor and acting chair of the department of nutritional sciences at the University of Toronto, confirms this in personal correspondence with me. He also heads a company, Glycaemic Index Testing Inc., which has ascertained the GI value of hundreds of foods.

“We did a study on glycerine at GI Testing, but the data don’t belong to me so I cannot publish it — except it was published in abstract form — and up to 75g glycerine had a negligible effect on blood glucose and insulin in normal subjects.’ He cites his article, “Oral glycerine has a negligible effect on plasma glucose and insulin in normal subjects” in Diabetes 2002;51(Supplement 2):A602. Some others believe, however, that it might have a greater impact on people with type 2 diabetes who have overactive livers.

(Side note: I can’t find the abstract for this paper on the Internet. Anyone have better luck?)

(UPDATE: I found the article! You can read it here: http://gpinzone.blogspot.com/2014/03/lost-article-on-insulin-response-to.html)

This all sounds promising, but it’s a bit of a hand wave in my opinion. What’s considered negligible? What happens to people who are type 2 diabetics that ingest glycerol? What about people who don’t have diabetes, but are obese or insulin resistant? You know, the people who tend to count carbs to lose weight? What about people who are no longer insulin resistant or obese, but want to stay ketogenic?

Basic metabolism

Low carbers know that ketones and fatty acids are the preferred energy source for the heart and an important energy source for skeletal muscle during prolonged exertion. Fat metabolism (lipolysis) during fasting provides most of the body’s energy needs during fasting. This means fat tissue is catabolized into free fatty acid and glycerol. The free fatty acid is metabolized in the liver and peripheral tissue via ß-oxidation into acetyl CoA and the glycerol is used by the liver for triglyceride synthesis or for gluconeogenesis.

But what happens when glycerin is introduced directly into the bloodstream or digestive system?

The most comprehensive evaluation and explanation of this topic I found in the paper, “Glycerol: Its Metabolism and Use as an Intravenous Energy Source.” It covers administration of glycerin orally, subcutaneously, and intravenously in both experimental (mammalian) animals and humans. It covers topics like toxicity, uses in medical treatments, and metabolism. It was published in 1983 so there may be newer research that modifies these findings, but it’s the best I can find. I’ll attempt to provide a summary of the main points of concern as best I can.  Please feel free to correct me if you know more.

Human levels of glycerol

Adult serum concentrations of glycerol are normally between 0.05 and 0.201 mM (0.46-1.85 mg/dl.^6,31-33 Glycerol levels in women tend to be slightly higher than those in men.³⁴ Prolonged fasting,^35,36 exercise,^37-38 and catecholainines^35,40 can elevate the level several fold. Obese or diabetic persons often have glycerol levels that exceed the normal range by a factor of two to three.³⁷ As observed in animals,^20,21 insulin in human subjects has been shown to lower the serum glycerol concentrations.^36,41 All of these changes are associated with the degree of fat mobilization or disposition. The conditions are also reflected by serum levels of FFA.³⁵ The basal turnover rate in the adult is approximately 0.1 mol/min or 550 g/hr. The removal of glycerol from the blood varies in proportion to concentrations over a wide range,^35,37,38 and may be affected by other factors such as exercise, blood ethanol levels, or alcoholic liver disease.^42,43

My interpretation of this is that glycerol levels are tied insulin in the sense that the glycerin is a product of lipolysis. If insulin levels are elevated, then lipolysis is inhibited, and therefore, less glycerin (and free fatty acids). Fasting and exercise can cause lipolysis to occur, so glycerol (and free fatty acids) will rise. I assume this doesn’t take into account glycerin from intravenous or oral administration. I think I’m getting the hang of this!

Insulinogenicity

It is quite clear that glycerol participates in gluconeogenesis as demonstrated by various experimental results.^{11,15,25,44,65,66} However, its action on insulin release remains unclear, especially as it relates to the route of glycerol administration. It has been found in humans that plasma insulin levels are inconsistently increased following intravenous administration of glycerol,⁷² while in rats a significant increase in blood insulin has been observed.⁷³ Continuous intravenous infusion of glycerol in fasting human subjects at a dose of approximately 1 g/kg/day did not result in elevated insulin levels.^23,74 In fasting human subjects, insulin release was influenced by orally administered glycerol at 1 g/kg body weight. Insulin reached the highest blood concentration of 63 µU/ml at 90 min after glycerol ingestion.⁷⁵ Since plasma glucose concentrations were not modified when plasma glycerol was highest, it was suggested that glycerol may either directly stimulate pancreatic ß-cells to release insulin, or act indirectly through the secretion of secretin, pancreozymin, cholecystokinin, and gastrin, which are active insulin release agents. The latter hypothesis is supported by the lack of an insulin response when glycerol is given intravenously in humans.⁷⁴

Now we’re getting somewhere! An oral dose of 1 gram of glycerin per kilogram of human weight did have an insulin response. To put that in perspective, a 180 pound human was given about 13 teaspoons of glycerin, which is approximately 285 calories. The paper this was referenced from is, “Stimulation of insulin secretion in man by oral glycerol administration,” and is unavailable for free. The abstract states:

The effects of an orally administered glycerol load (1 g/Kg body weight) on blood glucose, plasma FFA, and plasma insulin levels have been determined in eight normal fasting or glucose loaded (1 g/Kg body weight) volunteers. Blood glucose levels were not affected by glycerol loading while glicemia followed the same pattern of a glucose tolerance test in the group treated with glucose plus glycerol. Plasma FFA were significantly lowered only 90 min after glycerol loading while they had markedly and persistently decreased by glycerol plus glucose per os. Finally, though glicemia did not change, insulinemia was markedly increased by glycereol, 90 min after loading; moreover, plasma IRI was significantly higher in the group treated with glycerol plus glucose than in the group treated with glucose alone. These data suggest that the release of insulin may be stimulated by a very small increment of blood glucose, which derives from glycerol.

Hmmm... Blood sugar didn’t rise when glycerol was administered alone. Glucose tolerance tests of glucose plus glycerin versus glucose alone were the same. However, insulin was “markedly increased” by glycerol and was higher in the glycerol plus glucose group than the glucose only group. The reason given is that the body is converting glycerin to glucose, but blood sugar levels are not impacted. Huh?

I think what this might mean is that oral glycerin will (partially) be converted to glucose by the liver. However, it doesn’t dump glucose into the bloodstream rapidly. It probably shuts down gluconeogenesis from protein, but replaces it with gluconeogenesis from glycerin. Since insulin levels are higher, I’ll also assume the amount of glucose from glycerin gluconeogenesis is higher than that from a typical state of carbohydrate restriction induced ketosis.

Antiketogenicity

Glycerol engages in carbohydrate metabolism by entering glycolysis at the triose level and therefore is antiketotic. Alloxan-diabetic sheep given glycerol have been shown to exhibit a considerable reduction in blood acetate and ketone levels.⁷⁶ Significant reductions in plasma ketone body concentrations have been observed after oral glycerol administration to sheep made ketotic by starvation.⁷⁶ The decreases in plasma ketone body concentrations were accompanied by significant increases in both plasma glucose and glycerol levels.⁷⁷ In rats, glycerol is antiketogenic during all stages of development,⁷⁸ in contrast to the antiketogenic effect of alanine, which is not evident in neonates.⁷⁹ Differences in the sensitivity of 3-hydroxybutyrate and acetoacetate levels to reduction by glycerol have been found in rats.⁷⁸ Acetoacetate was much more sensitive to regulation by glycerol in mature rats, while its level in suckling animals was not affected.⁷⁸ Thus, the two ketone bodies may be regulated by different mechanisms in the rat. Reductions in ketone body concentrations have also been reported for diabetic humans, and in patients receiving glycerol to reduce intracranial pressure after cerebral infarctions.⁸¹ In humans, continuous intravenous infusion of glycerol in combination with isotonic amino acids resulted in significantly lower levels of circulating ketones when compared to infusion of amino acids alone.⁸²

I think this supports my understanding of glycerin metabolism. Glycerin inhibits ketosis because it’s creating glucose. Makes sense. The question is, does this imply that glycerin can “kick you out” of ketosis or is the glucose produced small enough to keep a person keto-adapted?

Toxicity

The authors also studied 14 college students given 30 ml of glycerol (1.3-2.2 g/kg body weight) in orange juice with each meal for 50 days. The subjects were observed for changes in weight, erythrocyte and; leucocyte counts, and hemoglobinuria. No abnormalities were noted. The basal metabolic rates and temperatures were not changed significantly in these students. Elevated urinary output resulting from glycerol ingestion was not observed in this study in contrast to the observations of other investigators.^101,102 Based on the observations described above, it was concluded that oral glycerol was safe in man.

It appears for people whom I assume have normal metabolisms that glycerin is safe as an additive and doesn’t cause weight gain. However, this gives us no insight into how glycerin can effect dietary induced ketosis in humans.

Hepatic fat accumulation and hypertriglyceridemia.

Since glycerol is an integral part of fat, it is reasonable to expect alterations in lipid metabolism from excessive glycerol intake. Early work concerning the ingestion of glycerol by humans and animals has not generally indicated any major pathology.¹⁰⁰ More recently, enhancement of hypertriglyceridemia has been observed in normal subjects receiving glycerol.^123,124 It has been reported that rats fed a diet containing 53.4% glycerol exhibited reduced weight gain, enlarged livers and kidneys, and increased enzyme activities along with increased gluconeogenesis and lipogenesis.¹¹ No toxicity was observed in the glycerol-fed rats as judged by the fact that the animals appeared healthy and grew, although somewhat slower than controls as a result of slightly reduced intake. In another study, animals fed a 25% glycerol diet had increased plasma lipids, especially triglycerides, as well as increases in the enzymes which regulate lipogenesis.¹²⁴

In a long-term study in which rats were fed a fat-free diet confining 30% glycerol for 13 to 20 wk, liver triglycerides, liver cholesterol, serum cholesterol phospholipids, and serum triglycerides were significantly higher compared to animals receiving the same diet without glycerol.¹⁰⁷ These results demonstrated that fat-free diets containing, glycerol could promote blood and liver alterations in lipid content. However, it was also found that high levels of saturated and unsaturated fatty acids could produce some of the liver alterations seen with glycerol. Similar observations for serum triglycerides and cholesterol concentrations have been made in baboons fed glycerol-containing diets.¹²⁵ Lipid accumulation has been found to accompany protein repletion in rats fed a diet containing 10% glycerol for up to 7 days after 14 days depletion.¹²⁶ Although food intake was similar, rats fed a repletion diet containing glucose did not show liver lipid accumulation. When dietary glycerol was fed to animals not subjected to protein depletion, no alterations in the lipid content of the liver were found. It was concluded that the formation of fatty liver induced by glycerol feeding was primarily due to a marked increase in food consumption during the transitional state from depleted to repleted.¹²⁶ The cause of lipid accumulation in the liver may be attributed to the stimulatory effect of dietary glycerol on enzymes catalyzing lipid synthesis. Dietary glycerol has been shown to dramatically increase the activities of citrate cleavage enzyme, fatty acid synthetase, and malic enzyme in rat livers.¹²⁷

While the accumulation of lipid in the livers of animals under conditions in which the caloric requirements of the test species were met have been observed, the effect of glycerol on lipid synthesis under hypocaloric conditions has not been reported.

Yikes! Extremely high levels of glycerin can cause fatty liver and increased blood triglycerides in rats. Nothing here specifically about humans. Better check to see if... oh... Look at this: “Pseudohypertriglyceridemia from oral glycerine.”

A 74 year-old Thai woman was found to have hypertriglyceridemia possibly due to an increased blood glycerol level. Her previous serum triglyceride (TG) levels were 65 and 99 mg/dl. After 2 months of taking glycerine at a dose of 50 ml orally every 8 to 12 hours for treatment of glaucoma, her serum TG concentrations increased from 77 to 1,815 and 2,693 mg/dl, but decreased rapidly to 72 and 59 mg/dl on days 3 and 6 following withdrawal of glycerine treatment. There were no other causes of hypertriglyceridemia. High blood glycerol level can interfere with enzymatic methods commonly used in the measurement of TG in most laboratories and result in falsely elevated levels of TG.

WHAT?! A blood test that lists my triglyceride level at 2,693 mg/dl would give me a heart attack whether it was real or not! 50ml is 10 teaspoons of glycerin every 8-12 hours. That’s a lot of glycerin. Since this condition is called “pseudohypertriglyceridemia,” I’m hoping triglyceride levels weren’t really that high. Remind me never to eat anything with glycerin a few days before a blood test!

I couldn’t find any study that linked glycerin with fatty liver in humans. On the contrary, glycerin is used to alleviate fatty liver in ketotic dairy cows. (If I can’t reliably extrapolate results from rats, I’m certainly not going to attempt to tease anything out from a cow study.) I don’t know if long term human use of glycerin in significant quantities would result in fatty liver disease. There’s just no data on the topic I can find.

Summary

Glycerol is an integral part of lipids, and has a caloric content of 4.32 kcal/g. In mammalian species, this quantity of calories becomes biologically available through known metabolic pathways common to carbohydrate and fat metabolism. Therefore, under hypocaloric conditions, glycerol can be metabolized to generate energy and conserve body proteins. Although experimental data indicates that blood glycerol concentrations can be depressed by insulin, the insulin response to exogenous glycerol has been found to be small and inconsistent, and appears to be dependent on the route of administration. In human subjects, continuous intravenous administration of glycerol at a dose of 1 g/kg body weight did not result in increased serum insulin concentrations. Therefore it is expected: that fat mobilization during nutritional therapy with glycerol will occur. However, the possibility of a direct antilipolytic effect of glycerol has not been studied, and is subject to further investigation. In addition, transport of glycerol across cell membranes does not seem to require insulin as does glucose. Even though the possible impact of the mild ketosis observed in patients receiving isotonic amino acid solutions has not been fully evaluated, ketosis has traditionally been considered to be undesirable. The use of glycerol to curtail ketone body generation as a result of excessive fat oxidation as well as to reduce gluconeogenesis from amino acids may be beneficial in patients receiving postoperative protein-sparing therapy.
...
From the evidence in the literature, it can be concluded that glycerol is gluconeogenic and inhibits gluconeogenesis from amino acids, is insulinogenic to a small degree, antiketogenic, chemically compatible with amino acids, and higher in caloric density than glucose.

Intravenous administration of glycerin doesn’t raise insulin levels, but oral administration does slightly. Glycerin suppresses ketone production and gluconeogenesis from protein (vice gluconeogenesis from glycerin), but it’s not clear if this will knock a person out of ketosis. Doses in most studies were 1 gram for each kilogram the person weighed. In other words, a 180 pound person would be ingesting approximately 13 teaspoons of glycerin at that level.

A quart of low carb ice cream from my research recipes needs up to 3 teaspoons of glycerin to adequately depress the freezing point. Most nutrition label serving size for ice cream is ½ cup, which implies 8 servings per quart. Let’s be generous and estimate 5 servings per quart. You’d be ingesting less than a 20th of the dose per serving as in those studies.

In conclusion, go ahead and eat your low carb ice cream. Yeah, sorry. I know it’s a hand wave, but my gut (and purple urine ketone sticks) tell me that this amount of glycerin should be fine on a low carb diet.

UPDATE 2014-01-30: I contacted Mitzi Elkes to get a copy of the letter from James Hoadley about glycerine and she went to the trouble of finding this 15 year old document, scanning it, and emailing it to me. Kudos! Here’s the link:

FDA letter from James Hoadley on Glycerine dated December 24, 1998

References

• “Glycerol.” Wikipedia. Wikimedia Foundation, 24 Jan. 2014. Web. 29 Jan. 2014.
• Institute of Food Technologists Food Laws and Regulations Division Newsletter, Vol. 9, No. 1, Winter/Spring 1999. Archived from http://www.ift.org/divisions/food_law/nl_v9n1.htm
• Mendosa, David. “Net Carbs.” mendosa.com. N.p., n.d. Web. 29 Jan. 2014.
• Tao, R., R. Kelley, N. Yoshimura, and F Benjamin. “Glycerol: Its Metabolism And Use As An Intravenous Energy Source.” Journal of Parenteral and Enteral Nutrition 7.5 (1983): 479-488. Print.
• Zanoboni, Alberto, Schwarz, Daniele, and Zanoboni-Muciaccia, Wanda. “Stimulation Of Insulin Secretion In Man By Oral Glycerol Administration.” Metabolism 25.1 (1976): 41-45. http://www.ncbi.nlm.nih.gov/pubmed/1246207. Web. 29 Jan. 2014.
• Charoenhirunyingyos W, Vannasaeng S. “Pseudohypertriglyceridemia from oral glycerine.” J Med Assoc Thai. 2010 Jul;93(7):870-2. http://www.ncbi.nlm.nih.gov/pubmed/20649069. Web. 29 Jan. 2014.