Why We Get Sick: The Hidden Epidemic at the Root of Most Chronic Disease–and How to Fight It

Overall score

Scientific accuracy

Reference accuracy

Healthfulness

How hard would it be to apply the book's advice? Fairly difficult

Book published in 2020

Published by BenBella Books

First edition Edition, Paperback

Review posted December 9, 2022

Primary reviewer: Shaun Ward

Peer reviewer: Morgan Pfiffner

Table of contents

1. Introduction
2. Scientific accuracy
3. Reference accuracy
4. Healthfulness
5. Most unusual claim
6. Other
7. Conclusion
8. Updates

Claim 1

Insulin resistance is the root cause of most chronic diseases.

Supporting quote(s) and page number(s)

Book Subtitle: The “Hidden Epidemic at the Root of Most Chronic Disease“ is referring to insulin resistance, as shown by the named sections of the book: ‘Part I: The Problem: What is Insulin Resistance and Why Does it Matter?’; ‘Part II: Causes: What Makes Us Insulin Resistant in the First Place?’; ‘Part III: The Solution: How Can We Fight Insulin Resistance?’.

Page xi: “Over the past few decades, all these conditions have been on the rise. But why? You’re about to learn that a lot of it comes down to one root cause: insulin resistance and hyperinsulinemia (meaning too much insulin in the blood).”

Page xiii: “In Why We Get Sick, Ben tackles a similar question, but on a broader scale, by identifying insulin as what leads us to develop chronic disease.”

Page xix: “But don’t despair, despite all the serious chronic diseases stemming from [insulin resistance]…”

Page xv: “Though [disorders] may seem unrelated, all of these disorders and more have one thing in common: to varying degrees, insulin resistance is causing the problem or making it worse.”

Page xvii: “And if other biomedical professionals weren’t aware of insulin resistance as a single cause of the most common chronic diseases, I figured that the average person would be almost completely in the dark.”

Page 11: “Most people with insulin resistance will ultimately die from heart disease or other cardiovascular complications; others will develop Alzheimers disease, breast or prostate cancers, or any number of other lethal diseases.”

Page 11: Why We Get Sick posits that “Type 2 diabetes is insulin resistance” and that these terms can be interchanged synonymously.

Page 13: “It is the puzzle—insulin resistance and cardiovascular disorders are almost inseparable.”

Criterion 1.1. How well is the claim supported by current evidence?

3 out of 4

This claim received a score of 3, indicating that relevant evidence is intrinsically convincing and partly aligns with the author’s claim. It is true that insulin resistance is widespread and associated with a spectrum of diseases, and in some cases there is strong evidence that it contributes to causing those diseases, but there is not enough evidence to suggest it is the primary (or root) cause of most chronic diseases.

If we define the term “root cause” as the main cause that contributes early in the cascade of biological events leading to disease, then insulin resistance is clearly a root cause of type 2 diabetes. Virtually all people destined to develop type 2 diabetes will first become insulin resistant and it is considered one of two necessary causal factors in this disease (the other being failure of the cells that produce insulin, beta cells).

Another disease that insulin resistance may be considered a cause of is cardiovascular disease. Numerous meta-analyses of observational studies provide compelling evidence that insulin resistance is a strong independent risk factor in people with and without diabetes. According to studies which examine groups with variable glucose tolerance, such as the Insulin Resistance Atherosclerosis Study, the groups with the lowest insulin sensitivity have over double the odds of developing cardiovascular disease compared to the group with the highest insulin sensitivity.

But while insulin resistance is generally accepted as a good predictor of developing cardiovascular disease, its relative contribution compared to other causal risk factors is yet to be investigated with much rigor. One study suggested that diabetes and metabolic syndrome (both featuring insulin resistance) are among the strongest risk predictors of premature CHD in women; however, this type of analysis directly comparing risk factors should be interpreted with a healthy degree of skepticism since estimates are based on mutually adjusted models (where all variables are considered simultaneously) and is not built to reliably measure independent causal effects. This limitation also applies to mathematical models attempting to compare the relative contributions of each risk factor, too—although some models indicate that insulin resistance is the single most important cause of cardiovascular disease, the models are based on external data and rely on strong causal assumptions, leaving much room for error. So while there are undoubtedly independent associations between insulin resistance and the risk of cardiovascular disease, often strong, how insulin resistance compares to other causal risk factors requires more robust data.

One way to untangle the long-term causal impact of different risk factors on disease is genetics research. For example, Mendelian randomization uses the random assortment of genes between people to figure out which biological processes drive disease. These studies suggest that insulin resistance is an important driver of cardiovascular disease, and the effect size is about the same as LDL cholesterol. Broader genetic studies have found multiple “root causes” of cardiovascular disease, including insulin resistance, blood lipids, clotting tendency, inflammation, and blood vessel tone (see figure 3 here). This is partially consistent with the argument in Why We Get Sick, but we think the causes of cardiovascular disease are more complex than presented in the book.

In comparison to type 2 diabetes and cardiovascular disease, there is considerably less evidence suggesting that insulin resistance is the root cause of other chronic diseases. The book discusses the causal role of insulin resistance for cancer, aging, gender-specific conditions (erectile dysfunction and polycystic ovary syndrome), gastrointestinal and neurodegenerative diseases, among others, but the type of evidence presented is not particularly convincing. There are indeed consistent associations and plausible mechanisms that could indicate a causal link between insulin resistance and many other diseases—for example, cancer and non-alcoholic fatty liver disease—but there remains a lack of strong, prospective evidence, or reliable data comparing the relative contribution of insulin resistance to other established causes. It is therefore difficult to strongly support the claim that insulin resistance is the root cause of most diseases.

Considering this gap in the evidence base, perhaps the strongest argument that insulin resistance could be the root cause of most chronic diseases is a logical one involving the potentially central role of insulin resistance in metabolic syndrome. This is not a disease state per se, but metabolic syndrome is a common term used to categorize a cluster of metabolic abnormalities linked to chronic diseases: elevated waist-to-hip ratio, fasting glucose, blood pressure, triglycerides, and low high-density lipoprotein concentrations. When Gerald Reaven introduced the concept of metabolic syndrome (originally ‘Syndrome X’) in 1988, he presented a series of cross-sectional studies showing strong associations between high insulin levels and other aspects of the syndrome. A range of mechanistic studies have since added to this pool of evidence to suggest that insulin resistance may precede the development of other components of metabolic syndrome and therefore be central to its development (a position accepted by many scientists).

If insulin resistance were the primary cause of metabolic syndrome, then it would be reasonable to think of it as a contributing factor in most diseases. Not necessarily a root cause, but a cause nonetheless. However, even if we grant this logical argument, we do not consider it sufficient to strongly support the claim. First, it is clear that certain elements of metabolic syndrome can form in the absence of insulin resistance. Second, some established causes of disease, such as high LDL cholesterol, are not necessarily part of the metabolic syndrome.

We also add that, despite cross-sectional associations and mechanisms indicating that insulin resistance precedes metabolic syndrome, prospective human data exploring the directionality of this relationship is mixed. Post-hoc analyses of the San Antonio Heart Study and the Kuopio Ischemic Heart Disease Risk Factor Study suggest that insulin resistance precedes the development of metabolic syndrome components, while other prospective data suggests the opposite: that some aspects of the syndrome (abnormal blood lipids) occur before the onset of insulin resistance.

We found one interesting paper stating that differences in who is included vs. excluded between studies is probably important to explain the mixed findings. Excluding subjects with baseline traits of insulin resistance will favor a result suggestive of insulin resistance preceding metabolic syndrome, whereas including subjects with baseline traits of insulin resistance will bias metabolic syndrome preceding insulin resistance. Considering that body fatness is also central to the development of insulin resistance, statistical analyses adjusting for body mass (i.e. remove its effect) could hide the true effects of insulin resistance in metabolic syndrome.

Overall, there is evidence suggesting that insulin resistance may be the root cause of some chronic diseases, but we don’t think the evidence strongly supports the idea that insulin resistance is the root cause of chronic diseases in general. There are remaining uncertainties and gaps in the relevant evidence base that stop us from scoring this criterion as a 4.0 (strong evidence). Perhaps we would score this claim higher if there was more prospective evidence for a wider range of chronic diseases, and the relative contribution of insulin resistance in these diseases was assessed with robust statistical models.

Criterion 1.2. Are the references cited in the book to support the claim convincing?

2 out of 4

Overall, the references in the book are intrinsically convincing and weakly support the author’s claim. The book discusses a number of studies linking insulin resistance to many chronic diseases, sometimes enough to justify a causal statement, but the author ultimately does not present a study (or collection of studies) to show that insulin resistance is the root cause or primary driver of most chronic diseases.

With the exception of type 2 diabetes, the apparent logic behind the author’s claim is to refute or downplay the importance of other causes of disease (for example, the causal role of LDL in heart disease), and then present a combination of epidemiological data (mostly non-prospective, which is a weaker design), animal, and cell studies that link insulin resistance to the disease, so that it appears to be the missing link.

We do not think this line of reasoning is enough to strongly support the claim, which is ultimately one that requires data reliably comparing the relative contributions of different causes against one another. This type of evidence was not presented in the book and, to our knowledge, the only studies that have attempted to calculate this have large causal assumptions built into their statistical models (as discussed in criterion 1.1). Also, other causes can have effects as large as insulin resistance when examining the differences between studies, such as genetic studies examining the contribution of either insulin resistance or LDL-C to cardiovascular disease.

In criterion 1.1 we suggested the potentially central role of insulin resistance in metabolic syndrome might indirectly link it to most chronic diseases, but the author did not outline this argument in the book. We also discussed why we still would not consider this argument to strongly support the claim.

Criterion 1.3. How well does the strength of the claim line up with the strength of the evidence?

2 out of 4

The evidence suggests moderately smaller and more uncertain effects relative to the author’s claim. There is certainly evidence that insulin resistance has a strong causal role in type 2 diabetes and cardiovascular disease, but the relative contribution of insulin resistance to cancer, aging, gender-specific conditions (erectile dysfunction and polycystic ovary syndrome), gastrointestinal and neurodegenerative diseases, among others, is not as clearly established compared to other causes. Focusing almost exclusively on insulin resistance throughout the book to explain why we get sick is overly simplistic and does not reflect the totality of scientific evidence.

 

Overall (average) score for claim 1

2.3 out of 4

Claim 2

Elevated insulin is a main cause of insulin resistance.

Supporting quote(s) and page number(s)

Page xi: “Over the past few decades, all these conditions have been on the rise. But why? You’re about to learn that a lot of it comes down to one root cause: insulin resistance and hyperinsulinemia (meaning too much insulin in the blood). But, wait–isn’t that actually two root causes? No, they are the same thing…).”

Page 8: “And while diabetes types 1 and 2 share the symptom of excess glucose, they completely diverge when it comes to insulin. Whereas type 1 diabetes is caused by having too little insulin (or none), type 2 is caused by having too much.”

Page 95: “Perhaps this won’t surprise you by now, but, to state the obvious: too much insulin causes insulin resistance.”

Page 96: “However, because insulin causes insulin resistance..”

Page 98: “The evidence is clear that too much insulin drives insulin resistance at several parts of the body, including muscle and fat tissue..”

Page 146: “Because elevated insulin is one of the most relevant factors in developing insulin resistance, it makes sense to follow a dietary plan that incorporates periods of time throughout the day when insulin is low.”

Page 151: “Once we appreciate that too much insulin is a main driver of insulin resistance, the chain of events suggesting a solution is too obvious: eating fewer carbohydrates = reduced blood glucose = reduced blood insulin = improved insulin sensitivity.”

Page 183: “The most scientifically sound diet to improve insulin resistance is one that keeps insulin low. Every aspect of your diet—what you eat, when you eat it—is a means to that end.”

 

Criterion 1.1. How well is the claim supported by current evidence?

2 out of 4

The question of whether elevated insulin is a cause or consequence of insulin resistance (or both) is still debated. As hormones can be regulated by negative feedback loops, one hypothesis is that insulin itself may cause body tissues to start resisting its signals. However, though insulin probably does contribute to insulin resistance, how large of a role it plays is still relatively unknown and it’s likely to be smaller than other causes.

One of the main issues with supporting elevated insulin as a main cause of insulin resistance is that the bulk of evidence is from animal and human infusion studies which do not reflect normal physiology—participants are typically infused with insulin for prolonged periods and in high doses. The external validity of these studies and their relevance to normal living conditions is contentious; notably, even if chronically high insulin levels are capable of increasing insulin resistance, it is unlikely that a person can have chronically high insulin levels without insulin resistance being present (because otherwise they would be chronically hypoglycemic, which is rare). This leads to a ‘chicken-egg’ debate about whether typical exposure to insulin (i.e. what can be expected from a typical diet) is enough to cause insulin resistance, or whether other factors cause insulin resistance, which then results in chronically high insulin levels.

Other forms of evidence supporting elevated insulin as a cause of insulin resistance include insulin therapy in type 2 diabetics (typically worsening their insulin resistance) and the role of insulin-spiking carbohydrates as a dietary cause. We consider both of these explanations to support the claim but only weakly, and we expand on the evidence in more detail in the next criterion and in our analysis of the next claim.

To contend with elevated insulin as a main cause of insulin resistance is the Twin-Cycle hypothesis, first described in 2008 and now considered the leading theory of type 2 diabetes pathology. This theory posits that a chronic energy surplus, not insulin, is the primary cause of insulin resistance and type 2 diabetes. As stated in a comprehensive review of insulin resistance, although “any effort to understand insulin resistance with a unified, succinct, and straightforward model may be a fool’s errand… The fundamental element linking all putative mediators of insulin resistance is a relationship to nutrient oversupply.”

While fat cells can buffer excess energy to maintain good health, when they enlarge they eventually reach a “critical threshold” and can no longer increase their size or number for additional storage. In response, as supported mechanistically by cell and animal studies, fat cells start rejecting additional fat from the bloodstream, and one way they do this is by resisting the actions of insulin. This increases energy storage in other tissues such as muscle, liver, and the pancreas. When insulin resistance is systemic across these metabolically active tissues, there is a greater risk that insulin-producing cells in the pancreas will stop being able to secrete enough insulin (due to chronically high exposure to energy substrates), blood sugar levels will rise, and type 2 diabetes ensues.

The Twin-Cycle hypothesis is most strongly supported by a series of prospective human studies—Counterpoint, Counterbalance, and the DiRECT trial—which together show that (1) pancreatic and liver fat stores correlate with insulin sensitivity, (2) sufficient weight loss reduces pancreatic and liver fat, normalizes glucose handling, and restores insulin sensitivity and beta-cell function, and (3) the more body weight that is lost, the greater the likelihood of diabetes remission.

The DiRECT trial, in particular, showed that 86.1% of type 2 diabetics achieved remission if they lost 15 kg of their body weight or more at 12 months. In fact, weight loss was the only strong predictor of type 2 diabetes remission in this clinical trial, whereas fasting insulin levels did not predict remission. Long-term studies of weight loss (bariatric) surgery have reached similar conclusions; glucose control improves within days of surgery due to preferential and rapid mobilization of fat from the liver and pancreas.

This concept has also been supported by human genetics research that has investigated the causes of insulin resistance by examining the functions of genes that cause insulin resistance in the general population. This research suggests that the main cause of insulin resistance is fat cells hitting their storage limit.

The cumulation of evidence has led to a conceptual model of the Personal Fat Threshold, explaining that each person has a ceiling to how much fat they can store without developing type 2 diabetes (of which insulin resistance features). The more body weight that is gained or lost, the more likely it is that an individual will cross their personal fat threshold. Genetics and early developmental factors are thought to be the main determinants of ones personal fat threshold, although the role of lifestyle and environmental factors needs clarification.

There are certainly lifestyle factors that reduce the risk of type 2 diabetes independently of body weight, implying they either lower the personal fat threshold or reduce fat accumulation in the visceral organs. Exercise, for example, can shuttle more energy to muscle cells instead of fat cells and effectively reduce visceral fat. Or, unsaturated fats, as another example, store less energy as liver fat compared to saturated fats. Even in these cases, though, energy oversupply and body fatness appear to be the key mediating variables in insulin resistance and type 2 diabetes.

Criterion 1.2. Are the references cited in the book to support the claim convincing?

2 out of 4

The book presents numerous lines of evidence which are generally consistent with the claim but are intrinsically weakly convincing.

One line of evidence is that insulin resistance occurs when you infuse insulin into a person or expose extracted human muscle cells to high amounts of insulin. The author cited two relevant studies here: one study infused healthy insulin-sensitive men with exogenous insulin for 96 hours, and another study exposed extracted human muscle cells to high insulin levels for 3 days. Both studies found a reduction in glucose uptake in response to insulin exposure, suggesting that insulin impaired insulin action. We consider this to support the claim but only weakly, as it is not clear whether the resulting insulin resistance was due to elevated insulin itself, excess energy exposure to tissues, or both. In the human study, for example, the insulin was infused along with enough glucose to prevent hypoglycemia, so the protocol was driving excess glucose into tissues over 96 hours. As previously discussed, excess energy exposure is a probable contributor to insulin resistance.

In addition, infusion studies and cell culture studies have limited external validity and do not directly assess whether typical exposure to insulin (i.e. what can be expected from a typical diet) is enough to achieve such effects in people living their normal lives.

Another line of evidence is that insulin injections in people with type 2 diabetes worsen their insulin resistance (pages 96 / 97). The reference to support this was an intensive 6-month programme of conventional insulin injections in 14 people with type 2 diabetes, indeed showing progressive hyperinsulinemia from insulin therapy. But much like our criticism of the first line of evidence, we do not consider artificial exposure to insulin to be reflective of typical insulin exposure. Because the participants (people with type 2 diabetes) were also insulin resistant at baseline and throughout the study, hypothesizing about the causes of the insulin resistance from this data alone requires strong assumptions. We also note that weight gain tends to be a strong confounding factor in these types of insulin therapy studies; for instance, the participants gained almost 10 kg from baseline in this study, and body weight increases correlated significantly with changes in total insulin dose and mean insulin levels in the blood. Insulin therapy might have contributed to the weight gain (possibly linking indirectly to insulin resistance) but how much so is rather speculative.

The final line of evidence in the book was that hypothalamic obesity (caused by dysfunctions in the hypothalamic region of the brain) causes insulin resistance independent of weight gain (page 96). To support this, the author referenced a study that demonstrated rats become hypersensitive to insulin just 1 week after the lesion of their ventromedial hypothalamus. But we are unsure how this study relates to the claim in the book. The hypothalamus is central to appetite regulation and VMH-lesioned rats tend to gain substantial amounts of body weight immediately. In this study, body weight increased significantly more in VMH-lesioned rats compared to controls—meaning that it does not support claims of insulin resistance occurring independent of weight gain. The study also reported insulin resistance in muscle tissue only; at least initially, insulin sensitivity increased in adipose tissue after the lesion.

Criterion 1.3. How well does the strength of the claim line up with the strength of the evidence?

1 out of 4

The strength of the claim is substantially overstated based on current evidence. It may be true that insulin contributes to insulin resistance but its causal role is relatively unknown and likely to be smaller than other causes. The current prevailing scientific theory is that insulin resistance occurs when an individual stores more energy and body fat than can be tolerated, and that high insulin levels are more a consequence of insulin resistance than a cause.

Overall (average) score for claim 2

1.7 out of 4

Claim 3

A high carbohydrate diet causes insulin resistance.

Supporting quote(s) and page number(s)

Page 146: “Because elevated insulin is one of the most relevant factors in developing insulin resistance, it makes sense to follow a dietary plan that incorporates periods of time throughout the day when insulin is low.”

Page 151: “Once we appreciate that too much insulin is a main driver of insulin resistance, the chain of events suggesting a solution is too obvious: eating fewer carbohydrates = reduced blood glucose = reduced blood insulin = improved insulin sensitivity.”

Page 151/152: “Thus, a diet that limits the insulin spiker (carbohydrates, especially refined) and increases the insulin dampeners (protein and fat, especially unrefined) is one that should improve insulin sensitivity. And—as we will see—it does.”

Page 153: “Clinical research since the 1990s has provided compelling evidence that carbohydrate restriction prevents or improves insulin resistance.”

Page 183: “The most scientifically sound diet to improve insulin resistance is one that keeps insulin low. Every aspect of your diet—what you eat, when you eat it—is a means to that end.”

Criterion 1.1. How well is the claim supported by current evidence?

2 out of 4

This claim received a score of 2 out of 4, meaning it’s weakly supported by current evidence.

In terms of observational research, the evidence is heavily in favor of carbohydrates having no significant association with the prevalence of insulin resistance. Based on cross-sectional examinations in the Framingham Offspring Study, Inter99 study, and the Insulin Resistance Atherosclerosis Study, totalling near 10,000 participants, no study found a significant association between total carbohydrate intake and insulin resistance. In the latter two studies, a protective effect on insulin resistance was found for dietary fiber (non-digestible carbohydrates).

We found a similar lack of association between total carbohydrate intake and the risk of type 2 diabetes in the largest, prospective observational studies (this tends to be the most rigorous type of observational study). The Nurses’ Health Study, the Health Professionals Follow-up Study, and the Iowa Women’s Health Study, totalling over 100,000 participants and each with 6 years of follow-up time, all found no significant association between total carbohydrate consumption and the risk of developing type 2 diabetes. The main analysis of a recent systematic review and meta-analysis of 13 prospective cohort studies also found that the risk of type 2 diabetes did not significantly differ when comparing the highest versus lowest categories of dietary carbohydrate intake.

Despite the above, it is worth noting that the type of carbohydrate and fat matters more for insulin resistance than the relative contributions of either macronutrient to total calorie intake. A meta-analysis of controlled feeding trials, for example, found that replacing carbohydrates with unsaturated fats reduced insulin resistance, whereas replacing carbohydrates with saturated fats had a neutral effect. This suggests that switching carbohydrates for certain fats may be beneficial; however, the authors of this analysis mentioned there was limited information to classify subgroups of carbohydrates (simple vs complex) as they did for fats, so the results should not be extrapolated to the potential effects of less processed carbohydrate sources such as fruits, legumes, and whole grains.

The type of carbohydrate in question becomes more of a crucial point when considering that intervention studies including a high carbohydrate (> 50 % total energy intake) and fiber intake consistently report either an improved or neutral effect on insulin sensitivity when compared to a calorie-matched high fat diet. Meta-analyses of dietary fiber RCTs also tend to find that higher intakes significantly improve insulin sensitivity.

On the other hand, and as mentioned in Why We Get Sick, refined carbohydrates in the form of free and added sugars have more uncertain effects on insulin resistance. While there is evidence that refined types of carbohydrates can increase insulin resistance, they probably do so to the extent they promote overeating and body fatness, not to the extent they increase insulin. A couple of randomized controlled trials that controlled for body weight and compared low to high glycemic index carbohydrate sources found no significant differences in markers of insulin resistance. A systematic review also stated that “…although some intervention studies suggested favorable effects of low [glycemic index] diets on fasting glucose, findings regarding effects on insulin, insulin sensitivity, or HbA1c are equivocal.”

Overall, the bulk of observational and interventional evidence does not support the claim that high carbohydrate diets are a major cause of insulin resistance. As the primary cause of insulin resistance appears to be excess accumulation of body fat driven by an oversupply of energy (as we outlined in claim 2), high carbohydrate diets probably contribute to insulin resistance to the extent they contribute to excess energy intake and body fatness, which is not unique compared to other diets such as a high-fat diet.

Likewise, improvements in insulin sensitivity from a low carbohydrate diet appear to be driven mostly by weight loss, rather than any particularly unique nutrient effect, at least when the most accurate measure of insulin sensitivity is used in the form of clamps. It does seem possible for dietary composition to affect insulin sensitivity in the absence of weight loss, but this pertains less to the proportion of carbohydrates and fats in the diet and more to the types of these nutrients consumed (for example, low vs high fiber carbohydrate foods; saturated vs unsaturated fats). We decided to score this claim a 2 instead of a 1 to account for some remaining uncertainties regarding the impact of free and added sugars on insulin resistance.

Criterion 1.2. Are the references cited in the book to support the claim convincing?

1 out of 4

Why We Get Sick references 4 studies comparing lower versus higher carbohydrate diets to support the claim that high carbohydrate diets cause insulin resistance. However, none of these studies are convincing and most findings are reported inaccurately.

The first reference was a 2-year trial that randomized 322 adults with obesity to one of three diets: a calorie-restricted low fat or Mediterranean diet, or a non-restricted low-carbohydrate diet. Why We Get Sick claims that the unrestricted low-carbohydrate group lowered their insulin resistance more than the other diets, but this is not completely accurate. First, this study didn’t directly measure insulin resistance; it measured it indirectly using fasting insulin and glucose levels, which hasn’t been validated in the context of low-carbohydrate diets. The main results of the study were that insulin resistance decreased significantly from baseline in all diet groups and there were no statistically significant differences between groups in the amount of decrease. The low-carbohydrate diet lowered insulin resistance more in non-diabetics, and the Mediterranean diet lowered insulin resistance the most in diabetics.

Why We Get Sick claims the second reference is a clinical trial that found insulin resistance dropped over three times more with a low-carbohydrate diet compared to a low-fat diet. However, the reference is not a clinical trial but a review article on the impact of carbohydrate restriction on atherogenic dyslipidemia, fatty acid partitioning, and metabolic syndrome.

We believe reference 27 in the review is the clinical trial mentioned in the book. If we are correct, the low carb group in this study indeed reduced their insulin resistance score significantly more than the low fat group. However, it is important to consider that the low carb group had a baseline insulin resistance score that was 70% greater than the low fat group. Because of this, both diet groups had virtually the same level of insulin resistance at the end of the study (low carb: 1.3; low fat: 1.4). This raises concerns about regression toward the mean, which can yield misleading results. Also, like the previous study, this study used an indirect measure of insulin resistance that hasn’t been validated in the context of low-carbohydrate diets.

The third reference was a retrospective follow-up of participants who had previously taken part in a low carbohydrate dietary intervention. Why We Get Sick claims that after four years, the low carbohydrate diet was “significantly superior” at improving health and led to almost half of the patients getting off insulin. But we do not consider this study as strong evidence for claims about insulin resistance because this outcome was not measured directly (only HbA1c, which loosely correlates with insulin resistance), there was no long-term data for the control group, and the low carbohydrate intervention also included advice to eat between 1600 – 1800 kcal per day, to avoid eating between meals, to walk 30 minutes a day, and to supplement a daily multivitamin with calcium supplement.

The fourth study referenced in the book was a randomized crossover trial that had insulin-resistant participants consume a relatively high (60% total calories) and low (30% total calories) carbohydrate diet, each for 3 weeks. Why We Get Sick claims that insulin sensitivity increased more with the low carbohydrate diet, but this is not true. To quote the results section of the study, “Multiple comparisons, however, revealed no difference in plasma glucose excursions between the diets at either day 4 or 21 of the meal tolerance test (Table 6). Insulin responses to the test meals were not significantly different between the two periods… Fasting values of plasma glucose and insulin were similar”.

Criterion 1.3. How well does the strength of the claim line up with the strength of the evidence?

1 out of 4

The strength of the claim in the book fluctuates but is generally substantially overstated. At some points, the author focuses on the negative effects of more refined carbohydrates and added sugars, as opposed to whole food sources of carbohydrates. At others, the authors claims a solution to insulin resistance is as simple as “eating fewer carbohydrates = reduced blood glucose = reduced blood insulin = improved insulin sensitivity” [page 151], that every aspect of the diet must keep insulin low [page 183], and that you probably should not eat more than 100 grams of carbohydrates per day even if insulin-sensitive [page 184]. The clear benefits of dietary fiber in insulin resistance are also questioned in the book, and summarized as beneficial for most people only when replacing sugars and starches, not fat [page 146].

Overall, we do not consider the strength of the claim to be reflective of the literature, as summarized in criterion 1.2. The book provides mixed messages but not enough attention or clarity is given to the critically important point of varying effects from different types of carbohydrates, and that high-fat diets are not demonstrably superior for insulin resistance than calorie-matched diets with a high amount of carbohydrate and fiber.

Overall (average) score for claim 3

1.3 out of 4

Overall (average) score for scientific accuracy

1.8 out of 4

Reference 1

Reference

Chapter 4, Ref 35. Walker WH. 2011. Spermatogenesis. 1(2):116-120.

Associated quote(s) and page number(s)

Page 45: “If testosterone is below normal levels, sperm production cannot occur”

Criterion 2.1. Does the reference support the claim?

1 out of 4

This reference received a score of 1. It points to a review paper examining how testosterone affects spermatogenesis. The only claim in the review which loosely tied with the claim was, “However, it has been established that spermatogenesis does not proceed in the absence of relatively high levels of testosterone (>70 nM in the rat)”. But since the evidence was in animals and there was no mention about how this level of testosterone translates to humans, we do not consider this to convincingly support the claim.

Reference 2

Reference

Chapter 15, Ref 41. US Centers for Disease Control and Prevention

Associated quote(s) and page number(s)

Page 153: “And we responded—we eat relatively less fat now than we did 50 years ago”

Criterion 2.1. Does the reference support the claim?

2 out of 4

This reference is intrinsically convincing but only weakly supports the claim, scoring a 2. In this report of National Health and Nutrition Examination Surveys from 1971 to 2000, the percentage of total calories sourced from dietary fat was found to significantly reduce during this period. However, it was also mentioned that there was a significant increase in total energy intake during this time such that “The decrease in the percentage of kcals from fat during 1971–1991 is attributed to an increase in total kcals consumed; absolute fat intake in grams increased”.

We interpreted the use of the term “relatively” in the claim to mean relatively less fat in the context of total calories, which is false. It is also worth pointing out that the cited NHANES data was restricted to before the year 2000 and does not reflect the dietary intake of now or when the book was published, in 2020.

Reference 3

Reference

Chapter 4, Ref 21. Friedrichsen M. 2012. Eur J Endocrinol. 167(6):829-38.

Associated quote(s) and page number(s)

Page 40: “Yes, it’s true—infants born below the normal weight are the most likely to become obese and insulin resistant”

Criterion 2.1. Does the reference support the claim?

1 out of 4

This reference scored 1 as the results do not support the claim even weakly. The study did compare the muscle inflammatory status and insulin sensitivity of 20 males with low birth weight (birth weight < 10th percentile) to 20 age-matched males with normal birth weight (birth weight > 10th percentile), before and after 9 days of bed rest, but it found no significant differences in BMI, insulin sensitivity, or any measure of adiposity between groups. The data also cannot support claims about the incidence of obesity since no participant was obese (BMI > 30).

 

Reference 4

Reference

Chapter 2, Ref 3. Goff Jr et al. Diabetes Care. 26(3):805-9. 2003.

Associated quote(s) and page number(s)

Page 14: “There’s no debate: insulin resistance and high blood pressure are related. When patients consistently have both, that’s evidence of a clear association, and almost all people with hypertension are insulin resistant”

Criterion 2.1. Does the reference support the claim?

2 out of 4

This reference scored 2 as it convincingly supports one part of the claim (that insulin resistance and high pressure are related) but not the other (that almost all people with hypertension are insulin resistant). The study is a post-hoc analysis of insulin sensitivity and hypertension from the prospective Insulin Resistance Atherosclerosis Study—an observational study designed to assess the relationships between insulin resistance, insulinemia, glycemia, and other components of cardiovascular disease in 1,600 adults. It found the risk of developing incident hypertension over 2 – 5 years was 11% lower for every unit increase in insulin sensitivity. This effect corresponded with the most insulin sensitive group having a 33% lower risk of developing hypertension compared to the least insulin sensitive group.

However, this study does not provide evidence that almost everyone with hypertension is insulin resistant. The study did not report how many participants were clinically regarded as insulin resistant at baseline or follow-up, and 19.6% of participants with insulin sensitivity above the median value at baseline still developed hypertension (compared to 27.9% of participants with insulin sensitivity below the median value at baseline).

Reference 5

Reference

Chapter 2, Ref 5. DiNicolantonio JJ. 2014. Am J Cardiology 114(7):1126-8. and Stamler J. 1997. Am J Clin Nutr. 65(1 Suppl):338S-365S.

Associated quote(s) and page number(s)

Page 15: “It also explains why carbohydrates, which increase insulin more than other nutrients, so effectively increase blood pressure..”

Criterion 2.1. Does the reference support the claim?

2 out of 4

This reference received a score of 2. Two references were included within one citation and neither offers strong support for the claim that carbohydrates effectively increase blood pressure. The first reference is merely a commentary responding to an editorial on sugar-sweetened beverages and blood pressure, contending the notion that associations between these variables are mediated by salt intake. The commentary mentions a few randomized trials on sugar-sweetened beverages and blood pressure, but none examine other forms of carbohydrates or compare carbohydrates to other macronutrients.

The second reference is more relevant to the claim. It presents data on the relationship between dietary variables and blood pressure in men randomly assigned to the Multiple Risk Factor Intervention Trial—a 9 year prevention trial that randomized 12,866 men with a high risk of cardiovascular mortality to either a “special intervention” program or usual medical care. The special intervention included counseling for smoking cessation and cholesterol-lowering (including dietary changes), along with antihypertensives. Total carbohydrate and dietary starch intake were significantly and positively correlated with changes in diastolic and systolic blood pressure; but conversely, there was a significant inverse relationship between fiber intake (a non-digestible carbohydrate and a proxy for complex carbohydrates) and changes in systolic and diastolic blood pressure. So although this study does provide some evidence for the claim, it also suggests that carbohydrate type may be a determining factor in whether blood pressure increases or decreases. The claim did not provide this context that would have otherwise improved its reference accuracy.

Reference 6

Reference

Chapter 6, Ref 4: Bonafe M et al. 2003. J Clin Endocrinol Metab. 88(7):3299-304.

Associated quote(s) and page number(s)

Page 58: “The theory [that caloric restriction extends lifespan] took another hit when another similar study in monkeys in 2012 found no lifespan benefit”

Criterion 2.1. Does the reference support the claim?

1 out of 4

This reference received a score of 1 as it is irrelevant to the claim. The study did not assess caloric restriction in monkeys; the study tested the hypothesis that polymorphic variants of IGF-I response pathway genes play a role in systemic IGF-I regulation and human longevity. We believe this claim was paired with the wrong citation.

Reference 7

Reference

Chapter 14, Ref 3. Lehmann et al. 1995. Diabetologia. 38(11):1313-9.

Associated quote(s) and page number(s)

Page 135: “One study had insulin-resistant individuals engage in moderate-intensity walking for three months. Even over the course of the relatively brief study, the people lost an average of 2% body fat, which appeared to largely come from visceral fat. A 2% change isn’t much, yet it was still enough to improve participants’ insulin sensitivity.”

Criterion 2.1. Does the reference support the claim?

3 out of 4

This reference received a score of 3, indicating that it offers moderate support for the claim. The study was a controlled trial assessing the effect of 3 months of exercise in 16 elderly patients with type 2 diabetes, compared to 13 non-exercising patients with type 2 diabetes that were matched by age and sex. The claim in the book is correct that exercising patients lost an average of 2 % body fat compared to baseline, and this corresponded with a significant reduction in the patient’s waist-to-hip ratio (a marker of visceral fat). But to contest the second part of the claim, the exercise intervention did not result in significant changes in any marker of insulin sensitivity (HbA1C, fasting plasma glucose, and fasting plasma insulin) compared to baseline values. We decided to give this claim a total score of 3 because the study strongly supports one part of the claim (body fat reduction) but not the other (insulin sensitivity).

Reference 8

Reference

Chapter 6, Reference 20. Thomas DM et al. 1998. Bone. 23(3):181-6.

Associated quote(s) and page number(s)

Page 63: “However, very recent findings suggest that insulin resistance could be the cause [of fibromyalgia]. In “Is Insulin Resistance the Cause of Fibromyalgia: A Preliminary Report,” researchers revealed that people with fibromyalgia are significantly more likely to struggle with insulin and glucose control.

Criterion 2.1. Does the reference support the claim?

1 out of 4

The reference received a score of 1 as it is not relevant to the claim. Reference 20 is not a preliminary report about the cause of fibromyalgia; it is an in vitro study on insulin receptor expression in primary and cultured osteoclast-like cells. We believe this claim was paired with the wrong citation.

Reference 9

Reference

Chapter 10, Reference 2. Pontiroli AE. 1992. Diabetologia. 35(3):294-5.

Associated quote(s) and page number(s)

Page 96: “The [insulinoma] patients with the highest degree of insulin production from the insulinoma become highly insulin resistant, whereas the patients with the lower insulin level become mildly insulin resistant. But, in the end, they always develop insulin resistance”

Criterion 2.1. Does the reference support the claim?

2 out of 4

This reference received a score of 2 as it offers evidence that is intrinsically weakly convincing and is consistent with the author’s claim. The reference is a 500-word letter to the editor of a journal mentioning two studies which, in total, evaluated insulin resistance in 19 patients: 13 patients with insulinoma and 6 patients with non-tumoural hyperinsulinaemia. The letter reports a significant association between serum insulin levels and insulin resistance based on an analysis of 16 of the 19 patients that were tested for insulin resistance during a 24-hour fasting period, appearing to support the book’s claim.

Yet on closer inspection, only 10 of the 16 patients tested for insulin resistance had insulinoma, so it is difficult to speculate about associations in this population alone. The letter did not communicate the results of the 3 insulinoma patients that had insulin resistance measured by the gold standard method, a euglycaemic hyperinsulinaemic clamp; the letter also did not mention how many patients were clinically regarded as insulin resistant. Because of these limitations, it is difficult for this letter to inform us about the prevalence of insulin resistance in insulinoma patients.

Reference 10

Reference

Chapter 15, Reference 74. Sharman MJ et al, J Nutr, 132(7):1879-85. 2002.

Associated quote(s) and page number(s)

Page 163: “This could be why people adhering to a calorie-unrestricted low carbohydrate diet can lose more fat than people following the classic calorie-restricted low-fat diet, even if potentially eating significantly more calories”

Criterion 2.1. Does the reference support the claim?

2 out of 4

This reference received a score of 2 because the reference offers evidence that is not completely relevant and only weakly supports the author’s claim. The study was an intervention examining how a ketogenic diet affects serum biomarkers in 20 normal-weight, normolipidemic men. Twelve men switched from their habitual diet to a ketogenic diet (30% protein, 8% carbohydrate and 61% fat) for 6 weeks, and 8 men continued their habitual diet.

There are a couple of issues with the authors interpretation of this study. First, calorie intake was not unrestricted during the ketogenic intervention; participants were given personalized dietary instructions, were encouraged to maintain their calorie intake, and dietary counseling was provided when body weight changed by 1 kg or more. Second, body fat was not measured in this study, and there was not a statistical test for changes in body mass between groups.

We assume the claim in the book that low carbohydrate diets can help people to lose more body fat despite eating more calories is based on the ketogenic group losing 2.2 kg of body mass, despite consuming 205 kcal more than the control group, who did not lose body mass. But aside from body mass and body fat being different outcomes, the ketogenic group actually consumed 511 kcal more than the control group at baseline such that both diet groups lowered their calorie intake similarly during the study (intervention group, – 205 kcal; control group, – 214 kcal).

Overall (average) score for reference accuracy

1.7 out of 4

Summary of the health-related intervention promoted in the book

Why We Get Sick recommends combining exercise and dietary change to manage insulin resistance. The primary exercise recommendations are to engage in physical activity most days (6 days a week for at least 20 minutes), focus mostly on full-body and high-intensity movements, and try to increase exercise intensity over time. The author notes, too, that exercise suggestions should be tailored to the individual and the main goal is to “just do it” [Page 138]. A sample weekly exercise plan is provided (Appendix A) including 4 days of resistance exercise—with brief suggestions for specific exercises and set and repetition ranges—and 1 day of cardiovascular exercise.

The dietary recommendations in the book are more comprehensive, with the overarching aim to keep insulin low. The author says that every aspect of the diet should be “a means to that end” (Page 183). In what are labeled as the four pillars of dietary change, the main advice is to control carbohydrates (up to 100 grams per day), prioritize protein (1.0 – 1.5 grams per kg of body weight), fill with fat (55 – 70% of total calorie intake), and watch the clock (fast for 18 hours a few times a week, and for 24 hours every 2 – 4 weeks). More specific recommendations for each pillar are outlined in the book.

Practically speaking, then, and as shown in Appendix B at the end of the book, the foods to prioritize include unprocessed meat and fish, eggs, dairy products (other than skimmed milk), high-fat foods (butter, lard, heavy cream, certain oils, etc.), vegetables that grow above the ground, fatty fruits, and fermented foods (kimchi, sauerkraut, pickles, etc.). Restrictions to 2 or less servings per day are placed on nuts and seeds, processed meats, and vegetables with a notable amount of carbohydrates. It is also recommended to avoid consuming seed oils, low-fat dairy, high-sugar fruits, alcohol, and carbohydrate-rich condiments.

Condition targeted by the book, if applicable

Insulin resistance is the condition targeted in the book.

Apparent target audience of the book

The book does not specify a target audience; however, since the book focuses on insulin resistance, which is said to be the reason we get sick, the target audience is presumably anyone trying to improve their health or prevent disease.

Criterion 3.1. Is the intervention likely to improve the target condition?

3 out of 4

The intervention is likely to moderately improve insulin resistance in the medium-to-long term, relative to typical lifestyle patterns.

The recommendation to be consistent with physical activity will likely have a powerful effect on managing insulin resistance. Insulin resistance within muscle is closely tied to physical activity levels, and exercise is regarded as one of the strongest tools to prevent type 2 diabetes. Systematic reviews and meta-analyses highlight that exercise usually improves insulin sensitivity within weeks to months, with the largest changes found when exercise intensity is moderate and performed 3 times or more per week for at least 30 minutes.

Other studies also indicate that physical activity has a dose-response effect on type 2 diabetes risk, with each 500 kcal per week increase in physical activity reducing the risk of type 2 diabetes by ~ 9%. Of course, the primary factor determining whether an individual benefits from exercise is the sustainability of the exercise regime, but this is the case for any lifestyle change. Finding a type and amount of exercise that is enjoyable and suits one’s lifestyle is key, as the benefits of insulin sensitivity can be achieved via a range of activities.

Similarly, for as long as an individual can sustain a low-carbohydrate diet, this dietary change is likely to benefit insulin resistance. A low-carbohydrate diet is recognized by the American Diabetes Association as one approach to improve insulin sensitivity and manage the risk of type 2 diabetes. If a low carbohydrate diet helps someone to lose weight—with plenty of strong evidence suggesting this approach is an effective option for weight loss—improvements in insulin sensitivity are expected.

A meta-analysis of randomized trials evaluating low-carbohydrate diets for at least 12 weeks in type 2 diabetics found that, on the basis of low to moderate certainty evidence, patients adhering to this diet for 6 months had significantly greater rates of diabetes remission and clinically important improvements for insulin sensitivity compared to control diets. That said, limited data assessed diabetes remission and insulin sensitivity at 12 months and, if anything, the improvements seemed to diminish.

We do not feel that a possible lack of long-term benefit is unique to this dietary approach, though; all diets also tend to have reduced effectiveness over time, mainly due to sustainability issues. For example, a meta-analysis of 29 long-term weight loss maintenance studies (up to five years) demonstrated that just over half of all weight lost during intervention periods is regained by 2 years, and 80% by 5 years. Dietary adherence is a major problem for long-term weight loss and the recommendations in the book are no exception.

Criterion 3.2. Is the intervention likely to improve general health in the target audience?

2 out of 4

Relevant evidence suggests the proposed intervention is likely to slightly improve general health in the medium-to-long-term (6+ months). Most of this benefit will probably hinge on the increased physical activity levels and reduced intake of refined carbohydrates and highly processed foods. Some health benefits are also expected from weight loss—at least for individuals that are overweight—and from increasing the consumption of healthy foods that are relatively underconsumed among the population: leafy greens, fish, and plant-based proteins.

But again, a diet is only beneficial if it can be followed, and many people will probably find it difficult to stick with the book’s recommended diet. Beyond the recommended elimination of almost every processed food, the diet restricts sweeteners, most fruits, milk, nuts, seeds, legumes, and certain vegetables and oils. The author also advocates for a maximum carbohydrate intake of 100 grams per day (less if not ‘insulin-sensitive’) and fasting for at least 18 hours 2 – 3 times per week, and for 24 hours every 2 – 4 weeks. While some people may find this approach fairly easy to stick to, we believe that most people will be overwhelmed by the rigid carbohydrate recommendation, the restrictions placed on foods commonly (and in our opinion, rightly) labeled as healthy, and the fasting regimen.

Other than adherence, our primary concern with the dietary intervention is the high saturated fat and low omega-6 fat recommendation, and how this affects cardiovascular health. As shown by numerous meta–analyses of randomized controlled trials, a lowering of saturated fat intake to below 10% of total energy intake reduces the risk of combined cardiovascular events. There is also evidence from dietary interventions that replacing saturated fat with an equivalent amount of polyunsaturated fat significantly reduces the risk of coronary heart disease. For this reason, a few foods that are recommended to be ‘eaten until satisfied’ in the book are likely not conducive to positive long-term health outcomes—namely butter, heavy cream, and lard. The suggestion that omega-6 fats are also “best avoided” (page 193) is also potentially harmful to long-term health.

Criterion 3.3. Does the diet portion of the intervention promote an adequate nutrient intake for general health in the target audience?

3 out of 4

The diet received a score of 3, indicating that it is more than nutritionally adequate. As the only foods to eliminate from the diet are highly processed foods (other than some fruits), most people following the book’s recommendations should be able to consume an adequate amount of essential nutrients for good health.

The only reason we have deducted 1 point is due to concerns that some of the restricted foods are rich in essential nutrients: legumes, nuts and seeds, tubers, most fruits, and certain vegetables. The intervention allows for up to 2 servings per day for most of these restricted foods, but the restriction still poses the risk that many people will substantially reduce their intake to emphasize the priority foods in the book (predominantly animal-based foods). To what extent these limitations risk deficiency in essential micronutrients or beneficial phytochemicals will depend on the composition of the diet as a whole and the extent of restriction that an individual places on these foods.

 

Overall (average) score for healthfulness

2.7 out of 4