“Dr. Spiegel, why is your institute trashing the Atkins diet?” This was the question angrily posed to me by Congresswoman Anne Northup (R-KY) at the 2003 House Appropriations hearing for the National Institutes of Health (NIH). As director of the National Institute of Diabetes and Digestive and Kidney Diseases, I testified annually (1999–2006) at this hearing, and was keenly aware of the importance of being responsive to elected officials who controlled the NIH budget. I mumbled a conciliatory response, even though I was clueless about the basis for her anger.
I had certainly heard of the Atkins diet, promoted in a series of best-selling books, with Dr. Atkins’ Diet Revolution selling 15 million copies. I was aware that Robert Atkins had died that same year at the age of 72, following a fall on an icy New York City sidewalk. His death touched off a controversy covered in multiple obituaries. Critics of the low-carbohydrate/high-fat diet contended that it was unhealthy, and alleged that Atkins died an obese man with heart disease. His supporters insisted that his heart disease was due to a viral infection. His wife refused permission for an autopsy, so we may never know the truth about his personal health. But what about the health implications for the estimated 20 million people who by 2003 had embraced his eponymous diet?
And, more importantly, could any low-carbohydrate diet provide a possible solution to the obesity epidemic in the United States? Figures from the Centers for Disease Control show that obesity affects 39.8 percent of adults. These individuals are susceptible to obesity-related conditions that include heart disease, stroke, type 2 diabetes, and some forms of cancer. A look at the origins of these diets may help us determine whether they are worth further evaluation.
Before trying to answer these questions, we must provide some background about human nutrition and metabolism. We can ingest three basic “macronutrients”: carbohydrates, fat, and protein. Carbohydrates include “simple” sugars and starches, and “complex” carbohydrates in fiber-containing foods. The latter are absorbed more slowly in the gut, provoking a slower rise in blood sugar (glycemic index) and less insulin secretion. Fats include fatty acids of varying chain lengths and degrees and types of saturation (hydrogenation). Each macronutrient has a defined energy (calorie) content, with fat containing nine calories/gram versus four in carbohydrates or proteins.
Macronutrients and Weight Loss
One question has dominated the “great diet debate”: What is the optimal macronutrient composition of a diet to promote weight loss and then maintain healthy weight? Atkins’ low-carb diet prescription (fewer than 30 grams of carbohydrates per day and ad libitum amounts of fat and protein) was not his invention. He popularized his low-carb/high-fat diet after reading studies by Alfred W. Pennington, an internist working for DuPont. Pennington, describing the basis for successful weight loss on the low-carb diet he tested, wrote: “The mechanism by which increased amounts of fat are made available for fuel is by means of the ketosis that develops when carbohydrate is in short supply” (“A Reorientation on Obesity,” New England Journal of Medicine, 1953). When the body was deprived of carbohydrates, he argued, endogenous fat was broken down into ketones to supply energy.
Why was the low-carb diet largely forgotten until Atkins’ marketing campaign? A major factor was a public-health campaign mounted by federal agencies that demonized fat. Dietary fat, the campaign maintained, was the cause of coronary artery disease and deaths from heart attacks, both occurring in record numbers after World War II. Unsurprisingly, proponents of low-fat diets, such as Dean Ornish, harshly criticized the Atkins diet. But others highlighted the inadequate scientific evidence for the health benefits of a low-fat diet (e.g., Gary Taubes, “The Soft Science of Dietary Fat,” Science, 2001) and speculated that the obesity epidemic that took off in the 1970s in the United States might actually have been caused by the substitution of high-carbohydrate foods for dietary fat.
The Causes of Confusion about Diets and Nutrition
Needless to say, contradictory conclusions by “the experts” and weekly media stories about studies touting the health effects of a particular dietary component (often reaching opposite conclusions) left the public thoroughly confused. John Ioannidis of Stanford has written a scathing critique of the types of nutritional epidemiologic studies that are the usual basis for these media reports (“The Challenge of Reforming Nutritional Epidemiology Research,” Journal of the American Medical Association, 2018). He points out multiple methodologic flaws in most epidemiologic studies of the effect of a particular dietary component on morbidity and mortality: inadequate methods to determine actual consumption over long time periods, difficulty in isolating effects of a single nutritional variable, failure to account for bias in group comparisons, and the cardinal principle that correlation does not prove causality.
For a rigorous comparison of various diets, we need carefully controlled, randomized clinical trials. Two of the first trials comparing the Atkins diet to a low-calorie/low-fat diet both showed significantly greater weight loss after six months with the Atkins diet, and most surprisingly, overall improvements in heart-disease risk factors (G. D. Foster et al.; F. F. Samaha et al., New England Journal of Medicine, 2003). The DIRECT trial randomized subjects to a low-fat/low-calorie diet, the Mediterranean diet (high fiber, high unsaturated fat, low calorie), and the low-carb Atkins diet (calories unrestricted) for two years (Shai et al., New England Journal of Medicine, 2008). The Atkins subjects showed the greatest weight loss and an improved blood-lipid profile at two years (significantly better than the low-fat diet, but not significantly better than the Mediterranean diet).
How might the Atkins diet work? Suggestions include increased energy expenditure and reduced caloric intake, the latter facilitated by greater satiety with higher protein intake and with ketosis. David Ludwig of Boston Children’s Hospital, a prominent proponent of the low-carb/ketogenic diet, argues that there is no human requirement for dietary carbohydrate or fiber (Journal of Nutrition, 2019). Ketosis occurs not only with low-carb diets but also with intermittent fasting regimens (an apt subject for a subsequent blog). In addition to providing fuel, “ketones are potent signaling molecules” that “regulate proteins known to influence health and aging,” possibly explaining the beneficial effects of intermittent fasting (De Cabo and Mattson, New England Journal of Medicine, 2019).
What remains to be determined are the long-term sustainability and overall health effects of low-carb diets. As Ludwig writes, high-quality research is urgently needed to answer these questions. The threat posed by our current obesity epidemic demands no less.