How Different Diets Impact Your Health | Dr. Christopher Gardner

KEYPOINTS

• 01

  Gardner's research points to whole foods, mostly plants

  Gardner sums up his studies with Michael Pollan's 'eat food, not too much, mostly plants'. Animal foods could make up 25%, 10% or none of the diet. He stresses that the right comparison depends on the goal: weight loss, cardiometabolic markers or something else.

• 02

  Differences within a diet dwarf the differences between diets

  In the A to Z study (Atkins, Traditional, Ornish, Zone; 311 women for a year, published in JAMA in 2007), groups ended only a few pounds apart. The only significant gap was between Atkins and Zone, the two low-carb diets. In each group of about 75, some women lost 30–50 pounds and others gained 5–10, which sent Gardner looking for what predicts individual response.

• 03

  DIETFITS found no winner and no predictor

  In this NIH-funded trial, with extra support from Peter Attia and Gary Taubes' Nutrition Science Initiative, 600 people were randomized to healthy low-fat or healthy low-carb diets for a year. Average weight loss was the same, with individuals ranging from losing 60 pounds to gaining 20 in both arms. Neither insulin resistance (measured by oral glucose tolerance test) nor a three-SNP genotype pattern from Interleukin Genetics predicted who did better.

• 04

  Both arms in a fair comparison should be the best version

  Taubes told Gardner he had 'screwed up' DIETFITS by banning added sugar and refined grains from the low-fat arm. Gardner says that was the point: a fair test gives each diet its best shot. His group used the same approach for vegan vs omnivore, keto (Jeff Volek and Steve Phinney's well-formulated version) vs Mediterranean, and Beyond Meat vs pasture-raised red meat from Good Eggs in the SWAP-MEAT trial.

• 05

  The protein RDA isn't the average requirement

  The 0.8 g/kg/day figure comes from nitrogen-balance studies, including Doris Calloway's 'penthouse' metabolic ward at Berkeley, where participants (largely conscientious objectors) collected all their outputs. The RDA adds two standard deviations to the average need, so 97.5% of people meet or exceed their requirement at that intake. Gardner accepts the argument that it's a minimum rather than an optimum. Huberman questions how artificial those inactive, confined conditions were.

• 06

  Lactose tolerance is the clearest diet–gene adaptation

  Most of the world stops making lactase after weaning, so most adults are lactose intolerant. Some northern Europeans who kept dairy herds evolved to keep producing lactase into adulthood. Gardner calls this the only well-established example of a recent evolutionary diet adaptation.

• 07

  The NOVA ultra-processed label is useful but imperfect

  Carlos Monteiro's NOVA system from Brazil underlies a decade of research on ultra-processed food. Gardner notes problems with it and argues that US rules on additives such as dyes, emulsifiers and glazing agents are too loose. Changing what manufacturers do would have more impact than teaching consumers to read ingredient labels.

• 08

  A 'protein flip' is a better frame than 'plant-based'

  Gardner and Huberman both prefer the idea of moving meat from the center of the plate to the side. Gardner says his group never pushed veganism; the vegan arm in the twin study was a design requirement. He sees taste, sourcing and convenience as what will actually shift how Americans eat.