It’s been a busy week with no shortage of dietary fiction being reported by the media. In addition to the coconut oil is “pure poison” nonsense that I responded to, there are now lots of confused conversations about the purported dangers of low-carb diets because of a study that purports to show that both low-carb and high-carb eating lead to increased mortality over 25 years. You’ve seen the attention-grabbing headlines: “Low-carb linked to dying young.” “Low-carb kills.”
The several points I made in the “coconut is pure poison” blog post are the very same ones that apply here. This study is yet another in a long line of similar studies from Dr. Walter Willett and colleagues. All the data that these people generate is observational, not experimental. Because the issues with this awful study are the very same as those that lead to false conclusions such as “red meat causes cancer,” I shall re-post my comments on the dangers of trying to draw any cause-effect conclusions from observational data but applied to the low-carb question:
Let me talk for a moment about observational studies. An observational study is typically conducted like this: I have you fill out a dietary questionnaire that asks you to recall what you ate for, say, the last 3 days. I then recontact you 5, 10, 20, or 25 years later and see what happened to you: healthy, heart attack, dead. Problem: filling out a questionnaire at one moment in time is a lousy way to assess diet, people change diets for a variety of reasons over the years—to lose weight, financial troubles, diet fads, etc. Another problem: confounding factors. If you say, for instance, that you are a vegetarian, it also means that you don’t drink much, don’t smoke, don’t engage in other high-risk behaviors, probably take nutritional supplements, eat more whole foods over processed foods. In other words, eating a certain way means that there are other behaviors attached to that way of eating; any effect on outcomes cannot be assumed to be due to vegetarianism per se, but to the entire collection of behaviors. There are other sources of bias in observational studies that muddy the results. Bottom line: Observational studies cannot establish cause-effect relationships; they can only suggest an hypothesis. To prove whether or not low-carb intake does or does not impact mortality cannot be based on questionnaires; a randomized prospective study in which people are randomly assigned to an eating style are compared. Unfortunately, unlike the studies with saturated fat in which people randomized to low-fat diets did not have less cardiovascular events, there are no long-term randomized clinical studies on low-carb eating.
Observational data is like having no data at all. Time after time, the conclusions drawn from observational studies (and falsely reported by study authors or the media as definitive conclusion) have fallen apart in prospective studies. My favorite example is Premarin, horse estrogens prescribed to women for years. Observational data suggested that Premarin (that looks and acts NOTHING like human estrogens) reduced breast cancer, reduced endometrial cancer, and reduced heart disease. This was responsible for making Premarin the most widely-prescribed drug in the world for about a decade. Then the prospective, randomized HERS and Women’s Health Initiative trials were conducted. Conclusion: Premarin INCREASED breast cancer, INCREASED endometrial cancer, INCREASED cardiovascular death, even accelerated dementia. And this has been the story over and over again: Conclusions drawn in observational studies have proven to be flat wrong about 4 times out of 5. This hasn’t stopped people like Frank Sacks and Walter Willett, through the observational Physicians’ Health Study and Nurses’ Health Study to, time and again, declare observational findings as fact. Unfortunately, even the USDA buys this observational fiction, incorporating the findings of observational studies in their dietary guidelines.
So the observational data reported in this study cannot be used to draw conclusions. It would be nice if we could refute the claims being made with randomized, prospective data, but there are none. However, there are abundant data that look at various biomarkers that demonstrate the metabolic advantages of low-carb eating. Among improved biomarkers are:
- Reduced HbA1c, fasting glucose, insulin, reduced insulin resistance
- Reduced triglycerides (due to lack of carbs/sugars to fuel liver de novo lipogenesis, i.e., conversion of sugars to triglycerides )
- Reduction or elimination of small LDL particles
- Reduced severity of postprandial (after-meal) lipoproteins (especially VLDL)
- Reduced blood pressure
- Weight loss including from visceral fat
- Reduced inflammatory markers such as c-reactive protein, IL-2, TNF-alpha
- Reversal of fatty liver
Low-fat diets, in comparison, are typically accompanied by no drop in blood sugar or insulin or an increase, a marked increase in triglycerides and small LDL particles that cause heart disease, and no reversal of fatty liver.
The only problem with low-carb diets come when carbs are cut to very low levels of about 20 grams per day and ketosis sets in. I’ve discussed this before: There is nothing wrong with being ketotic, as it is a normal and natural physiological adaptation. The problems come, however, when you stay in ketosis for a prolonged periods of months to years. Because you have eliminated all prebiotic fibers (since they come as mostly carb-rich foods like legumes), dysbiosis and small intestinal bacterial overgrowth sets in, followed by constipation, diverticular disease, and increased risk for colon cancer. The initial metabolic improvements that derive from reducing carb exposure reverse over time with a rise in blood sugar and insulin, rise in blood pressure, rise in triglycerides. We know all this with confidence as thousands of children have been maintained on ketogenic diets to suppress intractable grand mal seizures. In addition to the above, kids stop growing (suggesting that something is fundamentally wrong with the diet), have exceptional risk for kidney stones, osteoporosis, and there are occasional instances of sudden cardiac death. This makes people in the ketogenic diet world angry, but there is simply no way to just dismiss these observations made over many decades.
But back to the “low-carb kills argument”: Ignore it. It is the same misleading and non-conclusive data along with plenty of other nonsense that comes from Willett, Sacks, et al, who deliver their hypotheses as conclusive scientific findings—which they are NOT. Be content and happy with the magnificent overall health, reflected by numerous biomarkers, that results when you banish all grains and sugars from the diet, do not restrict fat, and take efforts to cultivate bowel flora and address a handful of common nutritional deficiencies that derive from living a modern life (not from the diet).