← Longevity & Supplement Guides
Biological age fell in four weeks. Read the authors’ own caveat.
A headline came round again this week, the way good headlines do: four weeks of changing what older adults ate made them biologically younger. It is a real trial, competently run, published in a serious journal. And the researchers who ran it wrote, into the published abstract, an explicit warning against reading it the way almost every retelling has read it.
That warning is the interesting part.
The headline came back around
The study is the Nutrition for Healthy Living trial, run out of the University of Sydney’s Charles Perkins Centre and published in Aging Cell by Caitlin Andrews and colleagues. It first travelled in the spring. In mid-September the science aggregators picked it up again, and it went round a second time to an audience that had never seen it.
Which is its own small lesson about how longevity news works. A finding does not have to be new to be trending. It only has to be re-posted somewhere with a large enough feed, and the second lap is usually louder than the first, because by then nobody is linking to the paper.
What the trial actually did
One hundred and four Australians aged 65 to 75 were randomised, in a two-by-two design, to one of four diets for four weeks. Protein was pinned at 14 per cent of energy in every arm — that is the clever bit, and the reason the trial can say anything at all. What varied was where the protein came from (omnivorous or semi-vegetarian) and how the remaining energy was split between fat and carbohydrate.
Before and after, the team calculated a biological-age estimate from a panel of routine blood biomarkers, then subtracted each participant’s actual age from it. The leftover — call it the gap — is the number that moved.
Three of the four arms saw that gap shrink relative to the omnivorous high-fat diet, which was the arm closest to what these people already ate and which barely budged. The omnivorous high-carbohydrate arm’s reduction was statistically significant. The two semi-vegetarian arms moved in the same direction, and not all of them cleared the bar.
The sentence the headlines dropped
The abstract does not finish on the result. It finishes on a caution, and the caution is unusually blunt for a journal: the authors warn against interpreting these shifts as evidence that biological age was reversed, because what moved may be nothing more than the body responding, acutely and correctly, to different food — rather than any change in the trajectory of ageing underneath. Judging anything about long-run risk, they add, would take a far longer trial.
I cannot improve on that. Scientists wrote please do not call this what you are about to call it, and then watched it get called that, twice, five months apart.
A clock built from blood tests will move like blood tests
The instrument matters here, so it is worth two minutes. The trial used the Klemera–Doubal method, first described in 2006 and since validated against illness and death in large cohorts. It is not a genetic clock and it does not read anything exotic. It takes a set of ordinary biomarkers — the sort of chemistry a standard blood workup returns — asks which combination best tracks chronological age across a reference population, and then reports the age at which your particular values would be unremarkable.
That construction is both its strength and its trap. The strength is that the resulting gap carries genuine information about a person’s near-term health. The trap is that every ingredient in it is a fast variable. Blood lipids shift within weeks of a changed diet. Fasting insulin answers to a fortnight of different carbohydrate. Inflammatory markers can move in days.
So a score like this falling after four weeks of supervised meals has not told you that someone aged more slowly. It has told you their blood chemistry looked different at the end than at the start — which was more or less guaranteed the moment the menu changed. The number did exactly what a well-built number should do. It is the translation into years that smuggles in the claim.
How to read the number
One score, two very different questions
what the word ageing meanswhat a blood-panel score readsthe trial window
The fast chemistry
Blood lipids, fasting insulin, inflammatory markers. This is most of what composite biological-age scores are assembled from, and it is what a four-week menu change moves.
The trainable layer
Cardiorespiratory fitness, muscle and grip strength, body composition, resting blood pressure. Real, measurable, and on a training timescale rather than a dinner timescale.
The thing being claimed
Accumulated change in tissues and organ reserve — the ageing the headline is actually talking about. Four weeks is not long enough to see it move, in either direction.
The three cards group markers by how quickly they respond, not by how much they matter; this is a timescale taxonomy, not a ranking or a measurement. The trial moved the first card.
The finding nobody repeated
Here is the part that did not make a single headline I saw, and it is the most genuinely surprising thing in the paper.
Ask most people who follow longevity talk which of four diets would shift the panel furthest, and you will get a confident answer: less animal protein, and fewer carbohydrates. The arm that actually reached significance was omnivorous and high-carbohydrate. Meat still on the plate, and more of the energy from carbohydrate rather than fat.
Two honest brakes on that. The comparison arm was the one closest to habitual eating, so part of what is being measured is simply the difference between a changed diet and an unchanged one. And 104 people split four ways is around 26 per arm over four weeks — a mechanistic signal, not a dietary instruction, and nobody should reorganise their kitchen around it.
But as a corrective it is useful. Hold protein constant, and the composition that moved the markers most was the one contemporary wellness culture would have picked last. That is worth sitting with before the next confident thread about macronutrients and ageing.
Why a fast clock is still worth having
None of this makes the measurement worthless. It makes it a different instrument than the one being advertised.
There are two separate questions hiding under the phrase biological age. The first is how much wear a body has accumulated — slow, largely invisible, the thing the word is supposed to mean. The second is what a body’s physiology is doing right now. A score that can move in twenty-eight days is answering the second question, and answering it rather well.
Sold as a verdict on the first, it misleads. Understood as a reading on the second, it is a feedback instrument — and a feedback instrument you can move in a month is more practically useful than a verdict you can only receive. This is the same distinction we have written about before in what biological age tests can honestly tell you and in the biomarkers that actually matter: the value of a number depends entirely on knowing which question it answers.
What to take from this
Four things, none of them a protocol.
Watch the inputs, not the composite. A single figure distilled from twenty markers conceals which of them moved. If a service gives you a year and no list, the year is the product. Ask what went in.
Note the size of the exposure. The thing that produced this result was four weeks of a defined, controlled diet with protein held steady — not a supplement, not a fast, not one swap at breakfast. Studies are exposures, and the exposure is the instruction, if there is one at all.
Read direction, not years. If you do retest a marker panel after a genuine change, the useful information is which way it went and by how much. The conversion into biological years is a presentation layer.
Keep the slow layer in view separately. Fitness, strength and body composition sit on the middle card of the figure above — they take months, they are measurable, and no four-week trial will tell you about them. They are also where most of the durable benefit lives; see how to build a longevity protocol and protein and healthy ageing for what that looks like in practice. And the broader habit of asking what a reading is physically capable of knowing applies far beyond blood panels — what your wearable can and cannot measure is the same argument in a different instrument.
Anyone managing a diagnosed condition, or acting on their own blood results, should take that conversation to their own clinician rather than to a trial in 104 Australians.
The bottom line
A four-week dietary change moved a biomarker-based age estimate in older adults, most clearly in the arm that kept animal protein and leaned toward carbohydrate over fat. That is a legitimate, interesting result, and the authors were careful to say what it is not: evidence that anybody’s ageing was slowed, redirected, or undone.
The honest reading is less thrilling and more useful. A number that answers in a month is telling you about this month. That is a good thing to be able to see — provided nobody converts it into years and sells it back to you as time.
The clock moved. The calendar did not.
