The Verification Venue · pointed at a thing everyone gets wrong
The Metabolism That Didn't Slow
Everyone blames it: a metabolism that quietly slows through your 30s and 40s, which is why the weight creeps on. The largest measurement of human energy burn ever assembled says it isn't true. Adjust for body composition and your daily burn is flat as a board from 20 to 60. The real decline waits until your sixties.
In 2021 a team led by Herman Pontzer pooled the gold-standard measurement of energy expenditure — the doubly labelled water method, which tracks how fast your body clears two harmless isotopes — for 6,421 people from 29 countries, aged eight days to 95 years. They asked a simple question: how does total daily energy expenditure really change across a life? The answer broke the most repeated fact in diet folklore.
The trick is that bigger bodies burn more, and most of that is fat-free mass — muscle and organs, the metabolically expensive tissue. The paper's single governing equation is a power law:
So the honest question isn't "how much do you burn at 45?" — it's "how much do you burn for your body composition at 45?" Drag the age slider. Then press the toggle and watch the two stories diverge: the raw curve falls (the myth), and the adjusted curve goes flat (the truth).
At your chosen age
age 45 · ×1.00 adjusted
life phase: adult plateau (20.5–63 y)
Slope of the 20–60 plateau (adjusted)
0.00 %/yr
flat as a board — no age-driven slowdown
Slide from infancy to 95. The four shaded bands are the paper's four metabolic life phases, split at the breakpoints it found: ~0.7 y, 20.5 y and 63.0 y.
The raw curve falls across middle age only because fat-free mass falls with age (the page assumes a representative ~0.4 kg/yr lean-mass loss — a declared choice). Per kilogram of working tissue, the engine never slows. Turn the raw curve off and on to see the whole illusion appear and vanish.
For a quarter of a century the rule of thumb was that metabolism dropped year after year through adulthood, so a 50-year-old's engine was meaningfully cooler than a 25-year-old's. The data say no. Birth to ~1 year is a furnace — size-adjusted burn peaks about +50% above the adult level. It cools through childhood, settles to the adult plateau at 20.5 years, and then nothing happens to the rate until about 63, where it finally starts to ebb at roughly 0.7% a year. Pregnancy doesn't move it either.
Weight gain in middle age is real — but it isn't a falling metabolic rate. It's lost muscle, changing activity, and what's on the plate, meeting a per-kilogram engine that is, the data insist, running exactly as hot as it did at 25.
The check — every number recomputed in front of you
The adjusted curve f(age) is the size-adjusted expenditure as a fraction of the adult plateau, pinned to the paper's reported landmarks. For your chosen age, computed live:
The headline claim is the flat plateau. A least-squares line fit through f(age) on 20.5–60 y has slope:
| landmark | age | adjusted f | vs plateau |
|---|
The verifier recomputes the power law, the curve, the zero plateau-slope, the falling raw curve, and the −0.7%/yr decline — and confirms they match these on-screen numbers. Run it yourself: node research/the-metabolism-that-didnt-slow/verify-the-metabolism-that-didnt-slow.mjs.
What's idealised here, and what's exactly true
Exactly true (from the paper). Total expenditure scales with fat-free mass as TEE = 0.677·FFM0.708 (r² = 0.83); body size explains ~83% of the variation. Size-adjusted expenditure was stable from age 20 to 60 — no significant slope — even during pregnancy. The four life phases and their breakpoints (~0.7 y, 20.5 y, 63.0 y with 95% CI 60.1–65.9) are the paper's segmented-regression result. Older adults decline at −0.7 ± 0.1% per year; the nineties sit about 26% below middle age; the neonate peak is about +50% near age 1.
The shape of the curve between landmarks is interpolated. The page draws a simple rise-and-fall through the published landmarks (linear up to the +50% peak, linear back down to the plateau at 20.5 y), then a flat plateau, then compounding −0.7%/yr. The landmarks are the paper's; the smooth path joining them is a faithful sketch, not the paper's exact spline.
The older-adult depth is honestly under-modelled. A constant −0.7%/yr from 63 only reaches the paper's "~26% below" figure near age ~106 — so the real decline must steepen with age. The page shows the verified constant rate (so the curve reads ~17% below at 90) and declares that this under-predicts the nineties; it does not pretend to reproduce the 26%. The 26% is quoted as the paper's measured fact, not as something this simple model regenerates.
The fat-free-mass trajectory is the page's choice, not Pontzer's. To show the raw curve falling, the page assumes a representative adult who loses lean mass at ~0.4 kg/yr (≈0.8%/yr of a 50 kg start). Real lean-mass loss varies hugely with sex, training and health; the exact raw numbers shift with it. What does not shift is the lesson: any reasonable lean-mass loss makes raw burn fall while the per-kilogram engine stays flat. That is the whole illusion.
What this does NOT say. It does not say individuals can't have genuinely low metabolic rates, nor that menopause, thyroid disease or medication don't matter. It refutes one specific belief: a steady, age-driven slowdown of size-adjusted expenditure through the 30s, 40s and 50s. The database is large and globally diverse but is an assembled, partly cross-sectional collection, not a single lifelong cohort — it describes the population trajectory, not any one person's future.