Chapter 10 — Key Takeaways
One page. The card for every diet argument you'll ever be dragged into.
🚪 The Threshold Concept
Across head-to-head trials, the differences between named diets are small and diminish with time. The variation between individuals on the same diet is far larger than the variation between diets. The best diet is the one you will actually keep doing.
And the qualifier that stops it being a cop-out: composition sets the floor — adequate protein, adequate fiber, adequate micronutrients, low ultra-processed food. Among nutritionally reasonable diets, differences are small. Composition sets the floor; adherence determines whether you ever stand on it.
The Trials
| Trial | Design | Result |
|---|---|---|
| DIETFITS (2018) | 609 adults, healthy low-fat vs. healthy low-carb, 12 months | No significant difference (~5–6 kg both). Genotype patterns and insulin secretion both failed to predict who did better. Within-group spread >> between-group difference. |
| A TO Z (2007) | Atkins / Zone / LEARN / Ornish | Small differences; individual variation dominated |
| POUNDS LOST (2009) | 811 adults, four ratios, 2 years | No meaningful difference |
| Network meta-analyses | Named diets | Small at 6 months, smaller at 12, gone by 24 |
Why They Converge
- The assigned diet is not the eaten diet — adherence drifts
- Dropout is 20–40% and non-random — the worst performers leave
- Every reasonable diet removes the same foods — soda, pastries, snacking, takeaway
- Everything runs through energy balance (Ch 4)
What the Winners Share
Whole, minimally processed food · vegetables · legumes · adequate protein · high fiber · nuts and seeds · no sugary drinks · little ultra-processed food · home cooking.
What they do NOT share: a macronutrient ratio. No two of them agree.
Diet by Diet
| Diet | Gets right | Gets wrong |
|---|---|---|
| Low-carb / keto | Removes UPF and sugary drinks; raises protein; suppresses appetite; real evidence for glycemic control in T2D | Metabolic advantage claim; early loss is glycogen + water; high dropout |
| Low-fat | Energy density; plant-based versions have decent CVD evidence | The 1980s implementation replaced fat with refined carbohydrate |
| Mediterranean | Best long-term randomized evidence (PREDIMED); additive; complete; flexible; not low-fat | Loosely defined outside trials |
| Vegan / plant-based | Fiber; phytochemicals; environmental case; favourable cohorts | B12 non-negotiable; iron, zinc, iodine, calcium, omega-3 need planning; healthy-user bias |
| Paleo | Removes UPF and refined grain; improves diet quality | 🟠 The premise — no single Paleolithic diet; modern foods differ; humans kept evolving |
| Carnivore | Effective elimination; removes UPF, sugar, alcohol at once; high protein | ⚗️ No long-term controlled trials exist. Zero fiber. |
The 24-Month Trajectory
Rapid loss (0–3 mo) → plateau (6–12 mo) → gradual regain (12–24 mo). Same shape for every diet.
Separation between diets is widest at 3–6 months — exactly when diet books stop reporting.
And the within-group spread band is taller than the distance between the lines.
But "a fraction of peak loss" is not zero — modest sustained loss carries real benefits in blood pressure, glycemia, lipids and function that persist even with partial regain.
Why Diets "Fail"
They don't. They stop being executable. "The diet failed" and "I failed the diet" are both wrong; the accurate framing is mismatch between a diet and the life it had to survive in.
Reject any regime with a reset-on-failure rule. It writes the what-the-hell effect into the terms and conditions.
How to Choose
- Establish the floor — protein, fiber, micronutrients, low UPF. Anything failing this is disqualified regardless of its story.
- Ask the adherence questions: Would I still be doing this in a year? Does it survive a bad week? Can I eat with other people? Can I afford it, in money and time? Does it require me to become a different person? Does it have a reset-on-failure rule?
- Match the mechanism to your temperament — rules or flexibility. Better predictor than any genotype.
- Prefer additive to eliminative.
- Run the reversibility test: what is my plan for the day this ends? "I won't stop" is not an answer.
- Default to Mediterranean if you have no strong preference.
Cost — the dimension nobody mentions
| Pattern | Annual, one adult |
|---|---|
| Beans-and-grains plant-based | ~$1,800–$2,600 |
| Mediterranean | ~$2,900–$3,900 |
| Standard mixed | ~$3,100–$4,200 |
| Low-carb / keto | ~$4,400–$6,200 |
| Carnivore | ~$5,200–$8,300 |
The patterns with the best long-term evidence are among the cheapest.
Why the Argument Is So Bitter
A diet is an identity, not a behaviour — so evidence against it is evidence against you. Everyone has a genuine personal success story that feels more compelling than any trial. Conversion costs are high, and sunk costs are hard to revise.
This chapter's conclusion offers the least identity of any position in the argument — no in-group, no enemy, nothing to post. That's a real disadvantage in an attention economy and none at all in a clinic.
Verdict Summary
| Claim | Verdict | Why |
|---|---|---|
| Paleo reflects what our ancestors evolved to eat | 🟠 Probably false | No single Paleolithic diet; modern foods differ; lactase persistence and amylase copy number show continued evolution. The diet is fine; the rationale isn't. |
| Humans thrive on an all-meat carnivore diet | ⚗️ Untested | No controlled trials in either direction. Improvements reported are consistent with any aggressive elimination. |
| A DNA test can identify your best diet | ❌ Not supported | DIETFITS tested it prospectively and pre-specified. It failed. |
⚠️ When to See a Professional
Very low-carbohydrate approaches with diabetes medication — insulin and sulfonylureas can cause hypoglycaemia; SGLT2 inhibitors carry euglycemic DKA risk (Ch 6 §6.8). Any restrictive diet in pregnancy or breastfeeding. Any elimination diet in a child — growth is at stake and a child cannot consent to a nutritional experiment. Any diet undertaken by someone with a history of disordered eating — restriction is both the intervention and the illness (Ch 34).
What We Still Don't Know
How to predict who will adhere to what. This is the largest gap in weight management, it's a psychology question, and we mostly find out by trying.
And there are no very-long-term randomized comparisons — five or ten years, hard outcomes — for any of these diets. PREDIMED is the closest and it's one trial in one population. Every confident claim about which diet is best over a lifetime is an extrapolation.
One Thing to Remember
If you rank Theo's diets by weight lost, keto wins. If you rank them by weight lost twelve months later, they all tie at zero.