The Stanford Twin Study has been receiving a lot of media attention, including a Netflix documentary chronicling the experiences of the 22 sets of identical twins who participated in this crude experiment.

In this study, 22 sets of adult twins agreed to be randomly assigned to either a “healthy vegan diet” or a “healthy omnivorous diet” for 8 weeks. Both diets included vegetables, legumes, fruits, whole grains, nuts, and seeds with no animal products (beef, pork, poultry, fish, shellfish, eggs, dairy), of course, in the vegan group. Both groups were encouraged to minimize processed foods, added sugars, and refined grains. Baseline measures were obtained, then repeated after the 8-week trial period.

Findings included:

  • 13.9 mg/dl reduction in (calculated) LDL cholesterol in vegans, 2.4 mg/dl in omnivores
  • 4.0 mg/dl reduction in HDL cholesterol in vegans, no change in omnivores
  • 8.3 mg/dl reduction in triglycerides in vegans, 7.9 mg/dl reduction omnivores, essentially no difference
  • 2.9 mIU/L reduction in fasting insulin in vegans, 0.9 mIU/L increase in omnivores
  • Vegans lost 1.9 kg (4.2 lbs) in total body weight, omnivores experienced no change
  • No differences in fasting glucose were seen between groups
  • Vegans reduced calorie intake by around 300 calories per day, omnivores reduced calorie intake by around 90 calories per day
  • Dietary cholesterol intake was dramatically higher in omnivores, negligible in vegans
  • Saturated fat intake was 62% greater in omnivores
  • Vitamin B12 intake plummeted in vegans, though insufficient to cause overt deficiency over the 8 weeks
  • Dietary iron intake was 25% lower in vegans.
  • Trimethyl amine oxide (TMAO) was 39% higher in omnivores

Of the above comparisons, only LDL cholesterol and weight achieved statistical significance; fasting insulin approached significance.

A few additional observations can be extracted, though not highlighted as important by the study authors:

  • Both diets were exceptionally high in carbohydrates with around 200 grams total carbs per day in both groups. The investigators and the media, both often using false logic, don’t hesitate to label the vegan diet as “ideal,” but this conclusion cannot be reached by this exercise. The most they can say is that, by LDL cholesterol measures, the vegan diet may be less harmful and potentially slightly more beneficial.
  • Insoluble fiber (mostly cellulose) was around 50% greater in vegans, soluble fiber slightly greater.
  • Consumption of legumes (as a source of the galactooligosaccharide prebiotic fiber) was dramatically less in omnivores, almost negligible, compared to vegans. This alone introduces significant potential for differences that have nothing to do with inclusion or exclusion of animal products.

There are a lot of problems here. First of all, we cannot make too much of the between-groups differences, good or bad, because the number of participants was small. We also ignore any differences in cholesterol intake because dietary cholesterol was abandoned many years ago as a meaningful factor in cardiovascular risk. We also do not care about saturated fat intake as that, likewise, has been debunked as a contributor to risk, despite the longstanding urgings of the American Heart Association to reduce consumption.

But let’s accept the above findings as more or less valid. I would point out that:

  • The reduction in calorie and protein intake in vegans suggests that the weight loss included significant amounts of muscle, since both dietary changes have been repeatedly associated with lost muscle mass in clinical trials. This is a potential negative outcome, not a positive result, with implications for metabolic health, independence, frailty, and longevity. This is also supported by the drop in HDL in vegans, though not achieving statistical significance over 8 weeks.
  • The investigators embrace the crude, outdated measure, LDL cholesterol obtained using the Friedewald calculation developed around 1960, as their primary outcome. We’ve known for years that, if you make major changes in diet, the Friedewald calculation is rendered invalid and LDL cholesterol becomes a useless marker. If we were to track calculated LDL cholesterol with the true gold standard measure, LDL particle number obtained using nuclear magnetic resonance (NMR), you would see, for instance, that the more changes you make in diet, the less LDL cholesterol approximates the real measure and the two can be wildly divergent, making calculated LDL cholesterol unreliable. This is further suggested by the drop in HDL. (Most studies of vegan or vegetarian diets also show a significant increase in triglycerides, though not shown here. The drop in HDL and rise in triglycerides conceals a marked increase in small LDL and VLDL particles, the real causes of heart disease.) Also, note that, despite dramatically higher saturated fat and cholesterol intake in omnivores, the increase in calculated LDL cholesterol was small. Had they measured NMR LDL particle number and VLDL, they would have likely revealed substantial reductions.
  • The investigators state that “eating a vegan diet can improve cardiovascular health.” Let’s be clear: There was NO measure of cardiovascular health made in this study. There was no measure of vascular reactivity, no measure of coronary atherosclerosis, no measure of left ventricular distensibility, no measure of intrinsic cardiac health such as serum atrial natriuretic peptide or intra-plaque fat composition or inflammation. They incorrectly label LDL cholesterol as a gauge of cardiovascular health—it is not. LDL cholesterol is a crude, indirect, and outdated gauge of risk for cardiovascular health, not of cardiovascular health itself.
  • The study was too brief to allow vitamin B12 deficiency to develop. But we know with confidence from both clinical studies and real-world experience that significant B12 deficiency will eventually develop, often sufficient to cause a macrocytic anemia and neurological impairment. (This is, of course, in addition to vegan deficiencies in absorbable heme iron, zinc, omega-3 fatty acids EPA + DHA, collagen, hyaluronic acid, and other nutrients)
  • The modest rise in TMAO was blamed on the inclusion of meats in the diets of omnivores, but ignoring the fact that the gastrointestinal microbiome is the principal determinant of this measure. The input—animal products containing choline and carnitine—is not as important as the composition of GI microbiome. Species of Proteobacteria that increase in colonic dysbiosis and small intestinal bacterial overgrowth, SIBO, are among the microbes that produce TMAO. This is why strategies such as supplementing with extra-virgin olive oil and polyphenols that strengthen the intestinal barrier and change microbiome composition reduce TMAO. What change in diet in omnivores or vegans was therefore responsible for the difference in TMAO—lack of galactooligosaccharides and other prebiotic fibers in omnivores? The reduction in calorie intake in vegans? We can’t tell from this study, but they cannot conclude that the difference is due to inclusion or exclusion of animal products.

Pitting vegan vs. omnivore diets against each other is an overly-simplistic dichotomy. It does not factor in intake of microbiota-accessible carbohydrates (prebiotic fibers and others), body composition (whether muscle is maintained or lost), hormonal changes, and other factors. It also tries to prove that following a diet that no free-living hunter-gatherer human population has ever followed throughout human history—i.e., a vegan diet—is preferable to the lifestyle programmed into the human genetic code. (India is, of course, a modern example of a vegetarian diet adopted for religious reasons, a lifestyle that has been plagued by impaired child growth and learning, as well as numerous nutrient deficiencies for the 3000 years this lifestyle has been followed.)

“You are what you eat”? It’s an odd choice for a title for a study that purports to demonstrate that humans should consume only plants—are you a plant? Homo sapiens and our primate ancestors have, for around 2-3 million years, been omnivorous with a diet that included plenty of animal organs, meats, skins, birds and birds’ eggs, reptiles, fish, and shellfish that resulted in the unique evolutionary divergence experienced by our species, adaptations that include a longer small intestine to accommodate digestion of animal products, shortening of the colon since we became less reliant on plant matter, and the expansion of the human brain permitted by diverting calorie consumption away from the colon and to the energy-intensive brain. And let’s apply some commonsense here: If you were to set out into the forest or jungle, how difficult would it be to survive only on plants, virtually all raw, as there is less need for fire and cooking? No wonder many of the participants assigned to the vegan diet in this were unhappy with diet assignment and resorted to marked increases in meat substitutes—they were forced to consume a diet that is unnatural and counter to the instincts and genetic code of our species.