09/08/2026
The standard advice about sitting treats it as a quantity problem. Sit less, or offset it with a workout. The evidence points somewhere more specific: what matters is not only how many hours you sit, but whether those hours are continuous.
The mechanism runs through skeletal muscle, and it does not require insulin. When muscle contracts, it moves glucose transporters to the cell surface and pulls glucose out of the blood directly. That pathway operates independently of the insulin signal, which is why movement lowers blood sugar in people whose insulin signalling works poorly, and why the effect appears within minutes rather than requiring weeks of training. Prolonged inactivity works in the other direction. Work in rats and in humans found that a single day of reduced muscle activity suppressed the activity of lipoprotein lipase, the enzyme responsible for clearing fat from the bloodstream, and that the suppression was not reversed by a bout of vigorous exercise elsewhere in the day. Sitting is not simply the absence of exercise. It is its own signal.
A 2026 systematic review pooled the experimental evidence. Fifty-three randomised crossover trials met the inclusion criteria, all of them acute laboratory studies lasting under 24 hours, and 39 were included in the meta-analysis. Each compared an uninterrupted sitting period against the same period broken by short activity bouts of under ten minutes, at least three times across three hours, with matched meals in both conditions.
Interrupting sitting lowered the post-meal rise in glucose and the post-meal rise in insulin by a similar amount, both standardised mean differences of about 0.30, which is a small effect by convention. Two subgroup findings sharpen it. Walking breaks produced larger effects than other modes, particularly for insulin, where the standardised mean difference reached 0.44. And frequency mattered: breaking every 15 to 20 minutes produced the largest reduction in glucose of any schedule tested, at a standardised mean difference of 0.51.
The foundational experiment behind this literature is worth knowing. In 2012, 19 overweight adults sat for five hours after a test drink in three conditions: uninterrupted sitting, sitting broken by two minutes of light walking every 20 minutes, and the same schedule at moderate intensity. Both walking conditions lowered the glucose and insulin responses substantially, and light walking worked about as well as moderate. The dose was two minutes, twenty times over.
Almost none of this evidence extends past a single day. Every trial in the pooled analysis was acute and laboratory based, meaning participants ate a standardised meal, sat in a controlled environment, and were prompted to move on a schedule. Glucose and insulin responses to one meal are markers, not outcomes. Nobody has shown that a person who adopts this pattern at their own desk, unprompted, for a year, ends up with better long-term glycaemic control or fewer cardiovascular events. The longer term evidence that exists is largely observational, such as the finding that people with more breaks in their sedentary time had smaller waists and lower markers of metabolic risk, which cannot separate the breaks from everything else that differs between people who move often and people who do not.
The comparison with structured exercise is not settled, and the honest version is more interesting than either side claims. A randomised crossover study found that replacing sitting with light activity spread through the day improved 24-hour glucose and insulin sensitivity, in some respects more than a single bout of moderate exercise, but the two approaches were not identical in their effects. Breaks and workouts appear to do overlapping but not interchangeable things. The reasonable position is that frequent movement is a distinct target rather than a substitute, and that someone who trains regularly and then sits unbroken for nine hours has addressed one of the two.
Practically, the parameters from this literature are unusually specific. The bouts that worked were under ten minutes and typically two to three. The intensity that worked was light, meaning ordinary walking pace rather than anything that requires changing clothes. The frequency that worked best was every 15 to 20 minutes, which is roughly three or four times an hour. That is a demanding cadence to maintain, and the sensible way to read it is as a direction rather than a prescription. Every hour you break up is better than the one before it, and the mode that did most of the work was simply walking.
PMID 42070794
Dunstan et al., Diabetes Care 2012;35(5):976 · PMID 22374636
Dempsey et al., Diabetes Care 2016;39(6):964 · PMID 27208318
Healy et al., Diabetes Care 2008;31(4):661 · PMID 18252901
Bey and Hamilton, J Physiol 2003;551(2):673 · PMID 12815182
Remie et al., Diabetologia 2021;64(12):2817 · PMID 34510226
