How Often Should You Get Up From Your Desk?

That is the honest version. Below is why the honest version has so many hedges in it, what the newest and largest study actually measured, and a guide telling you what to do based on how long you sit.

The biggest study yet, and what it can and cannot prove

In 2026 the British Journal of Sports Medicine published Diaz et al., much the largest study of this question to date. It recruited 19,342 US adults through NPR’s Body Electric Challenge. Of those, 11,484 completed the protocol: five-minute movement breaks, at intervals they chose themselves, for 14 days.

The results:

• Fatigue fell. Low mood fell. Positive mood rose.

• Work engagement improved by 4 to 7%.

• Self-reported performance moved 1 to 3%.

• Hourly was the sweet spot. It was nearly as effective as breaking every 30 minutes and far more sustainable over the two weeks.

Now the caveats, because they matter. This was uncontrolled and self-reported. There was no comparison group sitting still, everyone knew they were in a study about movement breaks, and every outcome came from participants rating themselves. That design cannot separate the effect of the walking from the effect of paying attention to your own mood twice a day for a fortnight.

So read the 1 to 3% performance number carefully. In an uncontrolled self-report design, that is not evidence that breaks make you more productive. It is decent evidence that breaks did not cost anyone any output, which is the claim worth defending anyway. The strong signal here is the sustainability finding: given a free choice, people converged on hourly and stuck with it.

If hourly is where real people land, that is a scheduling problem with an obvious fix. A 50-minute focus block with a five-minute movement break is one Pomodoro cycle with the numbers turned up. You can run exactly that in the PomoFit timer without configuring anything.

The dose-response study everyone quotes, and its sample size

Almost every article on this topic leans on Duran et al. 2023, from Columbia, published in Medicine & Science in Sports & Exercise. It is genuinely good work and it had 11 participants. Eleven. Keep that in your head.

Those 11 adults sat for eight hours under five different conditions: one minute of walking every 30 minutes, one minute every hour, five minutes every 30 minutes, five minutes every hour, and no walking at all.

What it found:

• Five minutes of walking every 30 minutes cut the post-meal glucose spike by 58%, and it was the only dose that significantly improved glucose.

• Every break dose lowered systolic blood pressure by 4 to 5 mmHg.

• Every dose except one-minute-hourly reduced fatigue and improved mood.

• No walking regimen improved cognition. Not one.

That last line is the one that gets left out of the summaries. A small Columbia study with a tightly controlled crossover design tested whether walking breaks made people think better, and the answer was no.

The 58% figure is the number you have seen on LinkedIn. It is real, it is from 11 people in a lab eating a standardised meal, and no large-scale trial has reproduced that specific figure. Treat it as a promising signal, not a law.

What happens when you pool the trials

Yin et al. 2024, in the Scandinavian Journal of Medicine & Science in Sports, pooled 13 crossover trials to find the best interval. The finding: breaking up sitting at least every 30 minutes beat less frequent breaks, but only for glucose. Other outcomes did not separate cleanly.

And then the authors did the thing that should make you trust them: they rated the overall quality of the evidence as LOW. Small samples, short durations, lab conditions, and a lot of heterogeneity between studies.

Dunstan et al. 2012, in Diabetes Care, is the older cousin of these trials. Nineteen adults, breaks every 20 minutes with two minutes of walking, post-meal glucose down 24 to 30% and insulin down 23%. Same shape of result, same small-lab caveat.

So the literature clusters like this. Frequent breaks reliably move short-term metabolic markers in small lab studies. They reliably move fatigue and mood. They do not move cognition. And the whole body of it is graded low certainty.

Why no health body will give you a number

Because the evidence does not support one yet.

The WHO 2020 guidelines say to limit sedentary time. They explicitly declined to set a number of hours or a break interval. If you see a page claiming the WHO recommends standing every 30 minutes, that page is making it up. The same goes for the ACSM. Neither organisation has published a sitting-break interval.

What the WHO does specify is the weekly total: 150 to 300 minutes of moderate activity or 75 to 150 minutes of vigorous, plus muscle strengthening on two or more days a week. Movement breaks feed the first number and, if you use them for actual strength work, the second one too. They do not replace either.

One thing that is settled: the old ten-minute minimum is gone. The US Physical Activity Guidelines for Americans, 2nd edition (HHS, 2018) removed the rule that activity had to come in bouts of at least ten minutes to count. The headline message became “move more and sit less”. A 90-second set of squats now counts toward your weekly total. That single change is what makes short movement breaks worth taking seriously at all.

Standing up is not the same as getting up

This is the part that trips up everyone who bought a standing desk and expected results.

Look at what the trials above actually tested. Duran tested walking. Dunstan tested walking. Wennberg et al. 2016 in BMJ Open tested three minutes of light walking every 30 minutes, and found significantly lower fatigue at both four and seven hours (and, again, no effect on executive function or memory). None of them tested standing still.

The mechanistic explanation usually given is that skeletal muscle contraction is what pulls glucose out of the blood, and a standing desk holds you upright without contracting much of anything. That mechanism is plausible and it is not what these trials measured, so hold it loosely. But it explains why every researcher in this field designed a walking protocol rather than a standing one.

The most useful practical finding on this comes from Radwan et al. 2022 in Cogent Engineering, six trials and 232 office workers. Active micro-breaks reduced musculoskeletal discomfort more than passive ones, and they did not reduce typing speed or work quality. The best protocol they tested was 20 minutes sitting, 8 minutes standing, 2 minutes walking. Standing is in there. It is just not the whole intervention.

While we are demolishing office folklore: for desk-worker neck and shoulder pain, a review of 27 randomised trials (Chen et al., 2018) found that strengthening exercise works at moderate-quality evidence, while stretching was no better than not stretching. Ergonomic changes alone did not work. Education alone did not work. So the ten-second neck roll, the entire genre of deskercise, is not the thing. Squats are the thing.

How long is too long to sit?

Patterson et al. 2018 in the European Journal of Epidemiology pooled 34 cohorts and 1.33 million people. All-cause mortality risk starts rising above eight hours of sitting a day. Cardiovascular mortality risk starts rising above six hours.

Which is awkward, because you probably clear both. One analysis from the Stanford Center on Longevity in 2021 found full-time remote workers sat 9.2 hours a day against 7.3 for non-remote workers, though that was self-reported and pandemic-era. Schöne et al. 2025 in BMC Public Health, pooling 38 studies and 282,264 people, found working from home adds about 31 minutes a day of sedentary time during work hours and costs about 2,564 steps a day, and roughly doubles the odds of sitting eight or more hours. Most of those underlying studies were also pandemic-era, so the size of the effect may not hold now. The direction almost certainly does.

If you sit X hours a day, do Y

Find your daily sitting total, then follow the break pattern next to it.

• Under 4 hours. Below both risk thresholds. One break an hour is maintenance. Use it for strength work, not walking, since you already move.

• 4 to 6 hours. Below both thresholds. Hourly break, 3 to 5 minutes. Two of those breaks a week should be resistance work to hit the WHO strengthening target.

• 6 to 8 hours. Above the cardiovascular threshold. Hourly, five minutes, non-negotiable. Add a deliberate walk outside the workday.

• 8 to 10 hours. Above the all-cause threshold. Aim for every 30 to 45 minutes on your worst days. Hourly on the rest. Prioritise consistency over the perfect interval.

• Over 10 hours. Well past both thresholds. Every 30 minutes where the job allows it. Also fix the schedule itself, because no break protocol out-runs a 12-hour sitting day.

The interval matters less than whether it happens. An hourly break you actually take beats a 30-minute protocol you abandon on Wednesday. That is precisely what the study found when it let people choose.

Making it survive a real week

Three things make the difference between knowing this and doing it.

Attach the break to something you already stop for. The reason a standalone reminder dies is that it fires mid-thought and you kill it. A break at the end of a focus session you chose was already coming. That is the entire design of PomoFit’s exercise breaks: the timer runs your work block, the break arrives on schedule, and a bodyweight move is already on screen.

Decide the exercise in advance. A blank five-minute break becomes your phone. Ten air squats on screen becomes ten air squats. If deciding is the friction, run one of the ready-made PomoFit plans and let it choose. Weekly desk warrior is built for exactly this pattern, back pain desk relief leans on the strengthening evidence above, and energy boost cardio exists for the 3pm slump.

Set expectations correctly. Micro-breaks of ten minutes or less reliably raise vigour and reduce fatigue (Albulescu et al. 2022, 22 samples, N=2,335, d≈0.35 for both). In the same meta-analysis the effect on performance was not statistically significant, and for cognitively demanding tasks it was effectively zero. So do not expect to think better after a set of squats. Expect to feel less wrecked at 4pm, and expect it not to cost you anything. On the evidence, that is a good trade.

If you work from home and are trying to claw back some of those 31 minutes, Pomodoro for remote workers covers the schedule side. If you want the break itself designed for you, start a session and take the first one now.