Whether or not insulin is to blame for the obesity epidemic is one of the hot questions being debated on heath and diet blogs. On the surface, this seems like an arcane question that would mainly interest physiologists and diet researchers. After all, who really cares about the underlying mechanisms of fat storage and release? Most of us just want to know some practical steps we can take to lose excess weight and keep it off and, beyond that, to stay healthy.
It seems like a simple yes-or-no question of fact that you could settle by studying populations and doing lab studies. But it’s not so much a question about facts as one about causation. Questions of causation are often the thorniest ones. This particular question has taken on almost political or religious overtones, provoking emotion and acrimony in the diet blogosphere. On one side are defenders of the Carbohydrate/Insulin Hypothesis, like Gary Taubes and Michael Eades. This is laid out in detail in Taubes’ book Good Calories, Bad Calories (2007), and more compactly in “Why We Get Fat: And What To Do About It” (2010). On the other side are opponents such as James Krieger and CarbSane, who find the Carbohydrate/Insulin Hypothesis to be oversimplified and deeply flawed, citing recent scientific advances. People tend to chose up sides in this debate. I’ve been participating in this debate myself (while still learning a lot) on the websites of Jimmy Moore, James Krieger, and CarbSane. I won’t rehash all the technical details here. Instead, I’d like to propose a “frameshift” that recognizes and integrates the strong points from each side, attempting to overcome their shortcomings.
First, here’s an overview of what each side has to say:
Proponents of the Carbohydrate/Insulin Hypothesis, as articulated by Taubes, posit four main points:
- Obesity is a disorder of excess fat accumulation, not voluntary overeating or inactivity, caused by an imbalance in hormonal regulation of adipose tissue and fat metabolism.
- Insulin is the primary regulator of fat storage. When insulin levels are elevated–either chronically or after a meal–we accumulate fat in adopose tissue. When insulin levels fall, we release fat and oxidize it for fuel.
- Elevated blood insulin levels increase hunger and the drive to eat, while decreasing energy expenditure and activity
- By stimulating insulin secretion, carbohydrates make us fat and ultimately cause obesity
In short: Carbohydrates drives insulin, which drives fat.
Opponents of the Carbohydrate Hypothesis challenge each of the above points. I’ve paraphrased four main counterpoints here:
- Fat accumulation and obesity result from positive caloric balance (more calories consumed than expended), without regard to the macronutrient class of calorie (carbohydrate, protein, or fat).
- Your body can store fat even when insulin is low, via the action of the hormone ASP (acylation stimulating protein)
- Insulin doesn’t make you hungry; rather, it suppresses appetite. (The critics proffer that low carb diets may work because protein is more satiating than carbohydrates, but they merely report this observation and don’t attempt to explain it).
- Carbohydrate doesn’t uniquely stimulate insulin; many proteins are equally or more insulinogenic.
In short: Calories in minus calories out drives fat.
On the surface of it, these two models of fat metabolism appear to be diametrically opposed. But are they really? There is at least one large point on which both sides appear to agree:
Obesity, particularly of the abdominal type, is associated with insulin resistance.
What that means is that people with abdominal obesity (the characteristic “apple” or pot belly shape, rather than those with “pear” shaped backsides or extra subcutaneous fat) tend to secrete more insulin after eating and have high basal insulin levels, ultimately leading to elevated blood glucose, triglycerides, elevated blood pressure, unfavorable cholesterol ratios, and a host of other issues associated with metabolic syndrome or “Syndrome X”. Nobody seems to deny this. Sometimes leptin resistance is also cited as an independent or alternative marker of obesity. But I’ll focus here primarily on insulin resistance, because it seems to be more closely involved with regulation of nutrient partitioning than is leptin.
Where the two sides disagree, however, is on the causal chain behind the association between obesity and insulin resistance. Advocates of the carbohydrate/insulin hypothesis tend to arrange the causal the order, from causes to effects, as:
carbohydrates > insulin spikes > hyperinsulinemia > insulin resistance > obesity
Whereas Krieger and CarbSane argue that the order of causality should be:
positive caloric balance > obesity > insulin resistance > hyerinsulinemia
When you look more deeply, however, there is acknoweldgement on both sides that insulin resistance is not a simple monocausal condition, but is likely multifactorial. There is evidence of many contributing factors, including:
- specific dietary components: fructose, sucrose, saturated fats, gluten, lectins, dairy, allergens
- micronutrient deficiencies: vitamin D, magnesium, omega-3 fatty acids
- metabolites: triglycerides, free fatty acids (“FFA”, also called non-esterified fatty acids or “NEFA”)
- inflammatory conditions
- lack of physical activity and exercise (particularly strenuous exercise)
There is as yet no broad scientific consensus as to the relative importance of each of these factors in causing insulin resistance. But it is almost certain that there is no single cause. Regardless of the cause, however, it is important to understand what insulin resistance is on a cellular level: a reduction in the number and sensitivity of insulin receptors, such as GLUT4 receptors. Different tissues can experience different degrees of insulin resistance. Typically, muscle tissues are the first to become insulin resistance and fat tissue is one of the last. Insulin resistance in different organs like the brain or the skin can have different effects. Some have argued that certain pathologies such as Alzheimer’s disease and acne are associated with organ-specific insulin resistance. I’ve proposed elsewhere on this blog (“Change your receptors, change your set point“) that receptor number and sensitivity can serve as a kind of dynamic “set point” for weight and other physiogical states governed by hormone-receptor and neurotransmitter-receptor balances.
So here is where I think that a frameshift in the debate about insulin can reconcile the two sides, at least in good measure:
Insulin resistant (IR) individuals respond in a qualititatively different way to carbohydrates and fats in their diet.
Let’s see what that means specifically:
First, consider insulin resistant (IR) individuals, regardless of how they got that way. IR individuals have elevated basal insulin levels, usually defined as a fasting insulin of at least 15 μIU/mL, or perhaps higher. If you have a protruding belly, high triglycerides and a high blood pressure, you are probably in this category. Under these conditions, dietary carbohydrate, and to a lesser extent protein, add fuel to the fire by spiking an already elevated insulin level. And let us grant here the point of Krieger and CarbSane that ASP is a potent faciliator of fat storage. It is known than insulin significantly enhances the action of ASP. In addition, insulin upregulates lipoprotein lipase (LPL) a fat-storage promoting enzyme and inhibits the action of hormone sensitive lipase (HSL) an enzyme that favors hydrolysis of stored lipids to free fatty acids. Combine all three effects and we should expect that IR individuals store dietary fat easily, even with moderately low carbohydrate diets.
For these individuals, the elevated levels of basal insulin will tend to shift the balance of glucose and fatty acids from the blood stream into the tissues. (Krieger and CarbSane are correct that insulin may not play a big direct role in driving fat sequestration, but its indirect stimulatory effects on ASP and LPL and inhibitory effect on HSL are quite significant, reducing the concentration of fatty acids in the blood stream by shifting the equilibrium towards the adipocytes). This will also tend to stimulate appetite and eating, leading to more fat storage and a worsening IR condition. Sugarholics and those with carbohydrate cravings tend to be insulin resistant. Appetite has a large conditioned component, whereby preprandial levels of insulin, ghrelin, and other hormones are secreted based upon temporal cues and specific sensory cues. It has been found that this pre-prandial secretion is much more pronounced in overweight, IR individuals.
One of the best ways to break this cycle is to go on a very low carbohydrate diet, something like the Atkins induction diet. Since there is no insulin response to dietary fat, a high fat, very low carb, moderate protein diet will allow basal insulin level to gradually drift down. This will shift the balance, reducing (but not eliminating) the actions of ASP and LPL, and disinhibiting the action of HSL. This will increase release of glucose and fatty acids, supplying energy and providing satiety, further lessening the drive to eat. The vicious cycle is replaced by a virtuous one. Unfortunately, a reduced calorie, high carb diet will not work for IR individuals, because their appetite is so easily triggered by any increase in insulin, which leads to a faster than normal drop in blood glucose. Note that blood glucose does not have to be “low” to induce hunger. There is evidence that hunger is triggered merely by a rapid drop in glucose levels. On the Deconditioning Diet page of this blog, I describe a method for extinguishing this conditioned pre-prandial insulin response.
Claims that insulin suppresses appetite is based on studies involving central administration of insulin while artificially infusing glucose. Krieger is correct about the “central” effect of insulin within the hypothalamus and upon the vagal afferent fibers. However, as with many hormones, insulin can have opposing effects at different locations and times. We need to consider the important appetite-inducing effect of insulin secreted into the “periphery”, without the simultaneous supplementation of glucose or other nutrients. This is a particular issue for IR individuals who are vulnerable to insulin-induced cravings, and less of an issue for those with good blood sugar control.
Now let’s consider insulin sensitive (IS) individuals. These are people with less than 10 μIU/mL, ideally less than 5 μIU/mL insulin. The situation is quite different for these folks. As a result of much lower basal insulin levels, they have more stable blood glucose and fatty acid levels, because the lower insulin levels reduce inhibition of glucose and fatty acid release from glycogen and adipose tissue. So IS individuals are less prone to hunger cravings, because they can access their own energy stores more easily. They are much better able to tolerate higher levels of carbohydrate in the diet, because their insulin response is well controlled and glucose readily gets to the cells and brain after eating.
This may also provide a plausible explanation for why certain populations such as the Okinawans, the Kitavans, and other cultures remain lean on a relatively high carbohydrate diet: their low basal insulin levels and high insulin sensitivity permit them to handle carbohydrates easily. According to Stephan Guyunet’s Whole Health Source blog:
Grains, refined sugar, vegetable oils and other processed foods are virtually nonexistent on Kitava. They get an estimated 69% of their calories from carbohydrate, 21% from fat, 17% from saturated fat and 10% from protein. Most of their fat intake is saturated because it comes from coconuts. They have an omega-6 : omega-3 ratio of approximately 1:2. Average caloric intake is 2,200 calories per day (9,200 kJ). By Western standards, their diet is high in carbohydrate, high in saturated fat, low in total fat, a bit low in protein and high in calories.
While this is a “high carbohydrate” diet, the carbohydrates are not typical western foods: The Kitavan diet consists mainly of foods like tubers, fruit, coconut, fish and vegetables. Even with the high carbohydrate levels, their insulin levels are much lower than that of typical Westerners. One could argue that these foods have low levels of fructose and sugars, and are generally quite non-inflammatory, so they should promote insulin sensitivity. According to Lindeberg, their fasting insulin levels averaged 3.12 and 3.29 IU/ml for males and females, respectively. This is about half the basal insulin levels of Swedes: 6.98 and 6.65 IU/ml for males and females, respectively. Fasting blood glucose levels for the Kitavan’s were about 27% lower than that of the Swedes.
Furthermore, IS individuals should be able to lose fat quite easily by restricting carbohydrate, intermittent fasting and/or exercise. With resulting very low basal insulin levels, it should be even easier to release fat from adipose tissue and oxidize it for energy, or to go into ketosis. It is known that Type 1 diabetics, who have no insulin, shed fat readily and have trouble holding onto it without injections. But someone with low basal insulin can achieve a naturally lean state easily, while also being able to handle insulinogenic meals without difficulty. Based on my own experience over time, as my fasting insulin level has dropped, intermittent fasting and even fasted workouts become easy, and this does not preclude a reasonable level of carbohydrates in my diet.
Now let’s ask the question of whether insulin sensitive (IS) individuals can accumulate body fat on a high-fat, low carb diet. According to Krieger and CarbSane, this should be no more difficult than on a high-carb diet. You just have to eat a “caloric surplus” of fat, with no or little carbohydrate, and ASP will do the job, even without insulin. But will this really have the predicted effect? Without doing the study, it is hard to know for sure. But my prediction would be that it is unlikely to play out as they suggest, for several reasons:
- Despite the claims that ASP works without any insulin, the primary sources don’t show this. For example in the paper by Saleh et al., which CarbSane cites in support, there is still some insulin and carbohydrate present to stimulate ASP, with or without the action of chylomicrons.
- Even assuming that the ASP could drive fat accumulation without insulin present, the lack of insulin would also favor downregulation of LPL and activation of HSL, which will tend to balance ASP’s action by liberating fatty acids from the adipocyte.
- Under low insulin conditions, even with excess fatty acids being fixed within the adipocytes, one would expect a reasonably high equilibrium level of free fatty acids in the blood stream. This would favor satiety, so that eating the fat meal would be self-limiting. This contrasts with the action of insulin which, when elevated, will tend to deplete the blood stream of glucose and fatty acids.
I will conclude with the following synthesis between the above opposing positions:
- Obesity is a disorder of excess fat accumulation resulting from insulin resistance (and leptin resistance), which stimulates appetite and naturally leads to caloric imbalance, including overconsumption of both carbohydrates and fats.
- Insulin and ASP together regulate fat storage and release. While ASP acts directly to transport and fix fatty acids within fat cells, insulin acts to induce fat storage via ASP and LPL, and to inhibit fat release via HSL and epinephrine and norepinephrine. Reduced levels of both insulin and ASP favor lipolysis and fat loss. The synergy of insulin and ASP further explain why the combination of dietary carbohydrate and fat is particularly fattening.
- In insulin resistant individuals, elevated blood insulin levels stimulate hunger and the drive to eat; this effect is largely absent for insulin senstive individuals due to superior blood glucose control
- In insulin resistant individuals, the pancreas compensates for reduced receptor sensitivity by secreting more insulin, leading to hyperinsulinemia.
So the answer to the question is to shift the blame from the hormone insulin to the condition of the insulin receptors. Insulin spikes at meal time are no problem, so long as basal insulin remains low. Restriction of dietary carbohydrate is one very effective strategy, which should be chosen not for the short term benefits in weight loss, so much as the longer term benefits in improving insulin sensitivity and reducing basal insulin. With the focus on “regrowing” and “reconditioning” insulin receptors, we should look at the full arsenal of tools, including intermittent fasting, nutrients such as vitamin D, magnesium and fish oil, and high intensity interval training.
Let me emphasize here that my proposed explanation is meant as a tentative conceptual framework rather than a conclusive scientific analysis. I’m still learning about the details and I fully expect that our understanding of the underlying mechanisms of fat metabolism will continue to be revised and evolve. But I do think that there has been too much emphasis placed on hormones and neurotransmitters, which fluctuate every day, and not enough on receptor health, which is something we can can influence over the long term by commitment to scientifically informed practices.
If you are interested in this general framework for diet and how it fits into my overall philosophy of Hormetism, check out my podcast interview with Jimmy Moore which just went live today.