– by Austin Bergquist –
Type 2 diabetes (T2D) is a growing health concern here in America. By 2030, it is estimated that 1 in 3 Americans will be living with T2D, or prediabetes. Over the past hundred years, the role that nutrition plays in the development, progression, and reversal of T2D has become more and more evident. To this end, the American Diabetes Association has been producing periodic evidence-based dietary recommendations aimed at educating individuals on what to eat to help mitigate the negative consequences of the disease. Interestingly, the current dietary recommendations highlight that diet is not one-size-fits-all and that a variety of diets can be effectively adopted to manage T2D. However, common themes among the recommended diets include: 1) consumption of whole foods over processed foods, 2) limited intake of added sugars, refined grains and overall carbohydrate volume, with an emphasis on consuming only non-starchy vegetables, 3) further reduction of carbohydrates for select individuals not meeting blood glucose goals (i.e. induce a ketogenic state), 4) Lastly, energy intake (i.e. calories) should be reduced in overweight individuals. The emphasis on low or extremely low carbohydrate consumption in these recommendations seems sensible. Since diabetes is a blood glucose regulation issue, it follows that limiting carbohydrate rich foods (i.e. sources of glucose in the diet) should help limit the potential for large spikes in blood glucose levels (i.e. hyperglycemia). Management of blood glucose is of critical importance since we know that chronic hyperglycemia wreaks havoc on vascular endothelial function through oxidative stress and inflammation and predisposes individuals to cardiovascular disease, stroke and premature death. Herein, we will review some of the relevant literature that demonstrates the effectiveness of low-carbohydrate and very-low-carbohydrate (ketogenic) diets on addressing many of the “symptoms” of diabetes. Further, we will discuss whether a low-carbohydrate or ketogenic diet addresses the “cause” of diabetes, as well as some factors about the research which may make you “think twice” about adopting a carbohydrate restricted diet yourself or about recommending one to your patients.
To assist in assessing the literature, it is important to define the criteria by which diabetes is diagnosed. When establishing a diagnosis of prediabetes or diabetes, there are 4 main tests: 1) HbA1C test, 2) Fasting blood glucose test, 3) Glucose tolerance test and 4) Random blood glucose test. The HbA1c test is a measure of your average blood sugar over a 3-month period. An HbA1C below 5.7% is considered healthy, between 5.7% and 6.4% indicates prediabetes, and 6.5% or greater indicates diabetes. The fasting blood glucose test measures your blood glucose after an overnight of not eating. A fasting blood sugar level of 99 mg/dL or lower is considered normal, 100 to 125 mg/dL indicates prediabetes, and 126 mg/dL or greater indicates diabetes. Following a fasting glucose test, a glucose tolerance test can be conducted by having the patient drink a specified amount of glucose (75-g) dissolved in water, and monitoring blood glucose levels for 2 hours thereafter. In response to this glucose challenge, blood glucose levels of 140 mg/dL or lower are considered normal, 140 to 199 mg/dL indicates prediabetes, and 200 mg/dL or greater indicates diabetes. Lastly, the random blood glucose test measures your blood sugar at any point in time during the day, without fasting. A blood glucose level exceeding 200 mg/dl at any point during the day, even after a meal, indicates diabetes. Importantly, failing any one of these tests would support a diagnosis of either prediabetes or diabetes.
Potential Benefits of Carbohydrate Restricted Diets in People Living with T2D
There are several studies that support positive effects of carbohydrate restricted diets in people living with T2D. In a 12-week randomized controlled trial (RCT) of 49 patients living with T2D, researchers compared the efficacy of a low-carbohydrate diet (39% carbohydrate; 42% fat) versus a low-fat diet (56% carbohydrate; 26% fat) and found that only those following a low-carbohydrate diet had a significant reduction in their HbA1C values. No differences in fasting blood glucose were observed. In a longer, 52-week, RCT of 115 obese participants living with T2D, researchers randomized participants to either a low-carbohydrate, high-unsaturated fat diet (14%carbohydrates; 58% fat) or a high-carbohydrate diet (53% carbohydrates; <30% fat) meant to reflect the dietary guidelines at the time. Both diets resulted in equivalent reductions in HbA1C and fasting blood glucose. However, participants on the low-carbohydrate diet were found to spend less time in the hyperglycemic state. In the short-term (<1-year), low-carbohydrate diets can produce some favourable changes in blood glucose markers for people living with T2D.
When carbohydrates are further lowered intentionally to induce ketosis (i.e. a metabolic state when fat is used as the primary energy source), there may be even further improvements in observed glucose markers among people living with T2D. In a small 90-day pre-post intervention study of 11 women living with T2D, the effects of a ketogenic diet (5% carbohydrate; 75% fat) on the symptoms of T2D were assessed. All 11 participants lost substantial weight (~20 lbs) and HbA1c levels decreased from diabetic (8.9%) to non-diabetic (5.6%) levels. In a larger 24-week RCT of 49 people living with T2D, researchers allocated participants to either a ketogenic diet (less than 20 g of carbohydrates per day) or a low glycemic reduced calorie diet (500 kcal reduction per day). Both groups lost significant weight and reduced their HbA1c within the first 12-weeks, and this reduction was maintained for the remainder of the trial. In an even larger 24-week RCT of 89 obese participants living with T2D, researchers allocated participants to either a standard low-calorie diet (restriction of 500-1000 kcal/day; ~55% carbohydrate; ~30% fat) or to a ketogenic diet (less than 50 g of carbohydrates per day).10 Participants in both groups lost weight and reduced their HbA1c from diabetic (6.8%) to prediabetic (6.0% to 6.4%) levels. Finally, one of the longest prospective trials to date (2-years), researchers allocated 349 overweight participants to either a ketogenic diet (less than 30 g of carbohydrates per day) or a standard American Diabetes Association diet.
Only participants in the ketogenic diet experienced reductions in weight, HbA1c, and fasting glucose. For this group, HbA1c levels dropped from 7.7%, but remained at diabetic levels (6.7%) by the end of 2 years. Fasting glucose dropped from 163 mg/dL, but also remained at diabetic levels (134 mg/dL) by the end of 2 years. Together, the data from these studies indicate that both lowcarbohydrate and ketogenic diets can assist in weight-loss and decrease both HbA1c and fasting blood glucose in people living with T2D.
“Thinking Twice” About Low-Carbohydrate and Ketogenic Diets in People Living with T2D
While low carbohydrate and ketogenic diets improve key “symptoms” of T2D (e.g. high blood sugar), these diets do not address the underlying “cause” of the disease, and likely make it worse. For some time now, it has been known that diabetes begins with insulin resistance, and that insulin resistance arises due to either excess calorie consumption and/or dietary fatty acid consumption that accumulates in skeletal muscle and liver.12, 13 Saturated fatty acids, and not mono- or polyunsaturated fatty acids, are specifically implicated in insulin dysregulation.14, 15 When saturated fatty acids accumulate in muscle and liver, they interfere with insulin signalling, such that insulin cannot do its job, and glucose begins to accumulate in the blood. Then more and more insulin is required to transport glucose within the cell. The pancreas struggles, chronic elevated blood sugar results and the person presents with diabetes. Low-carbohydrate and ketogenic diets cut the dietary source of glucose which can help with some of the “symptoms” of T2D (e.g. lower HbA1c, lower fasting glucose), but do so without correcting the underling cause (i.e. insulin resistance). In a small (n=9) prospective cross-over study design,16 researchers had participants follow a low carbohydrate diet (20% carbohydrate; 69% fat) and a “normal” diet (67% carbohydrate; 22% fat) each for 3 days, and then assessed blood glucose following an oral glucose tolerance test. The oral glucose tolerance test is the only diabetes classification criteria that truly assesses the underlying cause of diabetes, by applying a glucose challenge and seeing how the system responds. After following a low-carbohydrate, high-fat diet for only 3 days, individuals who consumed 75 g of sugar dissolved in water had higher spikes in blood glucose that lasted longer (i.e. insulin was less effective or was not being produced), compared to eating the “normal” diet for 3 days. Whether living in an insulin-resistant state itself is inherently unhealthy will require further long-term research. However we know that diets low in carbohydrates and high in saturated fat raise LDL cholesterol9, 17, 18 increase the risk of dying from all causes by 31% and increases the risk of developing T2D by 37%.
The glycemic benefits of a ketogenic diet stem largely from the weight-loss associated with what inevitably is a hypocaloric diet. We know that a hypocaloric diet (600 kcal per day) alone, without a change in macronutrient composition, can reverse abnormal insulin sensitivity and reduce blood lipids (LDL and triglycerides) in participants with T2D. In RCTs that test the efficacy of low-carbohydrate and ketogenic diets, improvements in HbA1c and fasting glucose always parallel weight-loss. For example, in one of the longest trials to date,11 participants lost weight on a ketogenic diet during the first year of the study, and their HbA1c and fasting glucose levels dropped concurrently. However, when followed into the second year of the study, participants started to regain their lost weight and observed parallel increases in HbA1c and fasting glucose. Therefore, we do not know if the benefits of a ketogenic diet are the result of weight-loss, or the result of the macronutrient composition of the diet. Further, the view that ketogenic diets allow for greater weight-loss than other diets is misrepresented. A meta-analysis of 13 trials lasting up to 2 years demonstrate that there are no differences in weight-loss between low-carbohydrate and “low-fat” diets by the 2-year mark.23 This means that the short-term benefits of the ketogenic diet are not likely to be sustained in the long-term, while the known harms remain.
It is difficult to find clinical studies that support a low-carbohydrate or ketogenic diet in the prevention of T2D that are NOT tied to industry funding. Of the 6 papers highlighted in this essay, all 6 were either directly funded or the authors were tied to industry that stands to gain from the positive outcome (e.g. Atkins Foundation; Virta Health Corp., DietDoctor; Insulin IQ; Unicity International, Pronokal Protein Supplies). This is problematic because we know that positive clinical research findings are 2-4 times more likely to come from industry versus not-for-profit funded research. This is not to say that industry funded research is not to be trusted entirely. Rather, it indicates that the research should be reviewed with an especially keen eye. This does present a problem because not everyone reads research closely, nor do they have the expertise to be able to discriminate between high- and low-quality research within a given field of study. Whether explicit or implicit, researchers may misrepresent their data in favor of their sponsor (i.e. the act of reciprocity) under the pressure to generate positive research findings. Consider the Athinarayanan et al 2019 study funded by Virta Health Corp, a company that sells ketogenic dietary programs for people with T2D. When you look to the data of the study at the 2-year mark, we find that patients had on average an HbA1c of 6.7%, and a fasting blood glucose of 134mg/dL. According to the American Diabetes Foundation, both of these levels would indicate that patients remained diabetic. However, the concluding sentence of the abstract states that the diet was, “…effective in the resolution of diabetes and visceral obesity…”. What is even more surprising about this statement is that patients in the study remained on glucose lowering medication (e.g. metformin). Along with their blood work, the finding that patients remained on diabetic medication indicates that patients did not come to a “resolution” of their disease.
In the short term (<1-year), adopting a low-carbohydrate or ketogenic diet can confer some health benefits to people living with T2D. Participants generally lose weight, and this weight reduction is associated with improvements in blood glucose measures like HbA1c and fasting blood glucose. However, diets low in carbohydrates and high in fat, especially saturated fat, lack long-term efficacy. Based on the best balance of the current evidence, I would advise my family and patients to avoid an animal based ketogenic diet, high in saturated fat. In long-term observational studies, avoidance of whole grains, fruits, and vegetables, required to sustain a low carbohydrate or ketogenic diet, demonstrates poor health outcomes among its adopters.