314 ‒ Rethinking nutrition science: obesity treatment, GLP-1 agonists, protein, & better research

KEYPOINTS

• 01

  Decades of obesity research have produced little public-health progress.

  Allison, a statistician and obesity researcher who began his career around 1991 at the New York Obesity Research Center, says clinical treatment improved moderately but public-health efforts achieved virtually nothing. Peter uses an aviation analogy: if planes fell from the sky more often every decade, nobody would claim we understood aerodynamics—yet obesity has risen steadily despite enormous effort.

• 02

  The 'epidemic' was declared from NHANES data in the early 1990s.

  NHANES III (roughly 1988–94) midpoint data released around 1991 showed what looked like a jump and the word 'epidemic' took hold. Allison notes skinfold data show less of a jump and an earlier, more gradual rise; historical data on 18th-century British naval cadets, French cadets and Civil War soldiers help show long-run trends.

• 03

  The environment changed, not our genes.

  People ate ad libitum 50, 100 or 200 years ago without today's obesity rates, so something environmental shifted. Peter notes type 2 diabetes rose by roughly a log-fold in his lifetime. Allison points to multiple plausible contributors, including air conditioning—a buffet is more tempting at 110°F in Austin if you're cool indoors—and is wary of single-cause zealotry, including from low-carb advocates.

• 04

  Classic public-health tools like 'nudges' have shaky evidence.

  Eat-less-move-more messaging, school programs and nudges have weak data. Cluster-randomized school-based childhood obesity trials tend to be quite poor in quality. After repeated failures, Allison argues, we should say 'enough' and change approach.

• 05

  GLP-1 success forces a paradigm shift—but we didn't get fat from a GLP-1 shortage.

  The drugs work, which some in public health resist because it seems to imply lifestyle doesn't matter. Allison wants a balanced approach: pharmacology plus realistic environmental change, while studying the mechanisms we still don't fully understand. Long-term safety is uncertain because no one has taken them for 40 years.

• 06

  Risk means uncertainty, not just known probabilities.

  Following FDA thinking, Allison stresses safety involves unknowns. That shapes the debate over whether someone 'subjectively 10 pounds overweight' should take GLP-1s versus someone with a BMI of 40 and diabetes, where benefits clearly outweigh unknowns.

• 07

  Peter declines GLP-1s for himself despite their appeal.

  He'd love to be 10 pounds lighter and never hungry, and the drugs improve glycemic control, but after watching many patients he doesn't accept side effects like the heart-rate increase for a cosmetic goal; for 40 pounds of medically significant excess, he'd take it 'any day of the week'.

• 08

  'Healthy food' depends on the goal.

  Someone wanting to be 10 pounds thinner, lift 10 more pounds or live 10 years longer may need different diets, so Allison dislikes blanket healthy/unhealthy labels.

• 09

  Protein needs are probably above the RDA, but thresholds are uncertain.

  The RDA of 0.8 g/kg probably isn't optimal. Allison questions claims that 20–30 g per sitting is required to be anabolic—a sigmoid curve is plausible but not proven. A study co-authored by Luc van Loon using up to 100 g suggested the old 40 g ceiling may apply only to fast-digesting whey.

• 10

  Upper limits for protein are poorly established.

  Claims that more than about 3 g/kg risks kidney damage in healthy people rest on thin evidence. 'Rabbit starvation' reports go back a century; one or two men ate only meat for 6–12 months and were fine, and a 1928 study put two Polish men on nothing but potatoes for six months, also fine—'semi-anecdotes', not RCTs.

• 11

  Physics-style humility: we may not understand obesity mechanisms for decades.

  Allison compares obesity science to other fields where progress was steady because the underlying science (physics, chemistry) kept improving and being applied. He imagines a future, perhaps 30 years out, where obesity rates fall back toward levels of 50–100 years ago—driven partly by drugs—even while we still don't fully understand the mechanisms, much as the 1970s 'population bomb' fears were defused by agricultural advances rather than by the solutions people then proposed.

• 12

  Bias in trials is subtle and often misunderstood.

  Allison distinguishes genuine bias from selection inherent in a goal—being cut from a basketball team for being short isn't 'bias' because height is intrinsic to basketball. In nutrition RCTs, he urges clear definitions before labeling a study biased, and notes funding source alone doesn't determine quality.