No, You Don't Need 10 Hours of Exercise a Week

The number going around this week was 560. Or 610, depending on which headline caught you — either way, somewhere near ten hours a week of moderate-to-vigorous exercise, reported as the amount you now need before your heart sees a serious benefit. The 150 minutes you have been dutifully logging? Worth about eight percent, apparently.

It is an excellent headline. It is also, according to one of the people who builds this kind of dataset for a living, a misreading of the paper it came from. The study is real, large, and carefully done. The number that escaped from it is the least trustworthy thing in it — and the most interesting finding never made the headlines at all.

The study underneath the number

The paper is from Zhide Liang and colleagues at Macao Polytechnic University, published in the British Journal of Sports Medicine in late May 2026 and back in circulation this week after a fresh wave of coverage. It follows 17,088 UK Biobank participants — average age 57, 56% women — who each wore a wrist accelerometer for seven consecutive days and completed a cycle test used to estimate their VO₂max. Over a median 7.85 years, 1,233 cardiovascular events were recorded.

What makes it unusual is that it looks at movement and fitness together, rather than treating them as interchangeable. And the joint picture is genuinely non-linear: the researchers found a significant interaction between activity volume and fitness level. Meeting the 150-minute guideline was associated with roughly an 8–9% lower risk of a cardiovascular event. Getting to around 560–610 minutes a week was associated with more than 30%. About 12% of the cohort were in that upper band.

Figure — dose and response

Where the numbers in the headline actually come from

Only the three marked points are published values from Liang et al., BJSM 2026 (17,088 adults, 1,233 cardiovascular events). The line between them is interpolated to show the shape the authors describe, not measured data. The 20% point is where the least-fit and fittest groups landed, at roughly 370 and 340 minutes a week respectively.

Why 560 minutes is the weakest number in the paper

Start with what that figure is. It is not a threshold anyone discovered. It is the answer to a question someone chose to ask: where does this curve cross 30%? Ask where it crosses 25% and you get a smaller number. Ask where it crosses 10% and you are more or less back at the public-health guideline you were already being given.

Then there is where on the curve it sits. Roughly one in eight of this cohort were up in that band, and this is a cohort that skews healthier and more motivated than the population it was drawn from — people who volunteered for a research study, agreed to wear a device for a week, and turned up for a cycle test. The far right of a dose-response curve is always the thinnest part of the data and the most flattered by who chose to stand there.

Professor Aiden Doherty at Oxford, who works on exactly these accelerometer datasets, told the Science Media Centre that the press release misrepresented the research and that a 560–610 minute target is not a sensible public-health message — the paper, he argued, is consistent with the existing 150-minute guideline delivering meaningful benefit. Professor Steffen Petersen at Queen Mary University of London made a related point: 150 minutes still reads as a sensible floor, and the study's real strength is that activity was device-measured rather than self-reported.

The composite hiding inside "cardiovascular disease"

Here is a detail that changes how the whole curve reads. Of the 1,233 events, 874 were atrial fibrillation. A further 156 were heart attacks, 111 heart failure, 92 stroke. So roughly seven in ten of the events driving this "cardiovascular disease" curve were an irregular heart rhythm.

That does not make the analysis wrong. It does mean the phrase in the headline and the thing that was actually counted are not quite the same object — and it matters most at the high end, because the relationship between very large endurance volumes and atrial fibrillation is one of the genuinely unsettled questions in sports cardiology. A composite that lopsided is worth naming before anyone reorganises their week around its top corner.

The part of the study almost nobody reported

Bolted onto the cohort analysis was a Mendelian randomisation. The idea is elegant: you inherit your genetic variants at conception, before any of life's confounders arrive, so if variants associated with higher fitness also track with lower disease risk, that is harder to explain away as "healthy people exercise more." It is the closest an observational study gets to a randomised trial.

And the result pointed somewhere the coverage did not. Genetically higher cardiorespiratory fitness was associated with lower odds of heart failure — an odds ratio of 0.79. The genetic evidence for physical-activity traits was, in the authors' own description, weaker and less consistent.

Read that twice. The arm of the study built to get closest to causation found its signal in fitness, not in minutes. The headline picked the other one.

Minutes are the input. Fitness is the output.

This is the distinction the whole story keeps tripping over. Minutes are what you spend. Fitness is what you got for it. Two people log an identical 300 minutes and finish the year with meaningfully different VO₂max, because intensity, structure, sleep, recovery, age and plain genetic luck all sit in the gap between the two.

Fitness is also the variable with the longer evidential pedigree. A meta-analysis by Kodama and colleagues in JAMA (2009), pooling 33 studies, found that each 1-MET increment in cardiorespiratory fitness was associated with a relative risk of 0.87 for all-cause mortality and 0.85 for coronary and cardiovascular events. These are associations, not proof of cause — but they have replicated for fifteen years across very different populations.

The practical objection is that measuring fitness properly means a lab, a mask and a maximal test. That objection has weakened. A 2025 review in the Journal of Sport and Health Science pooled 42 studies and more than 3.8 million observations and found that estimated fitness predicted mortality about as well as directly measured fitness — pooled relative risks per 1 MET of 0.86 for all-cause and 0.84 for cardiovascular mortality, with no significant difference between measurement approaches. The estimate is not the laboratory. As a tracked trend, it is also not nothing.

What the curve says about where you are standing

Look again at the fitness-stratified result, because it is quietly the most encouraging number in the paper. For a 20% lower risk, the least-fit participants needed around 370 minutes a week; the fittest needed around 340. Thirty minutes apart.

That is a much smaller gap than the "deconditioned people face a steeper hill" framing implies. Starting unfit does not put the benefit out of reach; it shifts the requirement by about one extra walk. The bigger practical difference is not how many minutes the two groups need, but how much fitness each minute buys them — and that gap runs in favour of whoever has the most room to improve.

What to do with this in the next month

One change, and it is a change of scorekeeping rather than of schedule: stop grading yourself only on minutes, and pick a single fitness output to watch over eight to twelve weeks.

Any of three work. An estimated VO₂max trend, which is the most direct read on the thing the genetics arm pointed at. Heart-rate recovery in the minute after a hard effort, which needs no equipment beyond a watch and a stopwatch. Or resting heart rate, the least glamorous and most reliably available of the three. Then judge a training block by whether the output moved, not by whether the minutes were logged.

Two honest caveats. Any single reading is noise — the trend across weeks is the signal, which is the same reason a wearable's estimates are useful for direction and poor for absolutes. And these are wellness estimates, not medical measurements or a diagnosis; anything that concerns you belongs with your doctor. If you want somewhere to put the minutes themselves, the case for easy aerobic work and for the occasional hard interval hasn't changed this week, and the evidence that even one structured session beats none is still standing.

The bottom line

The guideline is a floor, not a ceiling. More activity does appear to track with lower cardiovascular risk well past 150 minutes a week, and that is worth knowing. But 560 minutes is a coordinate someone chose on a curve, not a bar you have failed to clear — and the part of this study designed to look hardest at cause found fitness, not minutes.

Which is the useful version of the headline, and a less demoralising one. The question is not how many hours you served. It is what those hours made of you — and unlike the hours, that is a thing you can actually watch change.