Is sitting really 'the new smoking,' and what can you do about it?

Why sitting may matter. Resting muscles use little glucose. After a meal, blood sugar and insulin stay higher when you sit for hours than when you break the sitting up. In a trial of 19 overweight adults, 2-minute walking breaks every 20 minutes cut the post-meal blood-sugar and insulin response compared with sitting straight through⁴. A pooled analysis of seven such crossover trials found light walking breaks beat standing breaks for both glucose and insulin, and neither changed systolic blood pressure⁵.

How much is too much. Across 34 studies and about 1.3 million people, the extra risk of dying from any cause was small up to 8 hours of sitting a day (about 1% per added hour) and rose to about 4% per added hour beyond that. For cardiovascular death the turning point was about 6 hours, and for TV viewing about 3 to 4 hours².

Activity offsets much of it. In a harmonized analysis of more than 1 million adults, people who sat over 8 hours a day but did about 60 to 75 minutes of moderate activity daily had no clear excess risk. High TV time was only partly offset³. In the 45 and Up Study (149,077 adults), sitting mattered mainly among those with no moderate-to-vigorous activity; meeting the guidelines largely erased the link⁶.

Light movement counts. Accelerometer data from 36,383 adults showed any activity, even light, went with lower mortality, and the most sedentary quarter had about 2.6 times the risk of the least sedentary⁷. Long unbroken sitting seems worse: in 7,985 adults, the longest average sitting bouts carried higher risk than shorter ones⁸.

Standing desks. Sit-stand desks cut office sitting by about 100 minutes per workday in the short term, but the evidence is low quality and runs out after about a year⁹. Standing burns only about 0.15 kcal per minute more than sitting¹⁰, and a job spent mostly standing was not safer for the heart¹¹.

Practical rule. Move every 20 to 30 minutes, add a brisk walk after meals (see Daily Steps and After-Meal Walks), and build toward the WHO target of 150 to 300 minutes of moderate activity a week. WHO could not set a sitting limit, but it recommends reducing sedentary time¹². For desk setup, see The Modern Posture Crisis.

Sitting versus smoking. A public-health commentary compared effect sizes: the most recent meta-analysis of sedentary behavior gave a hazard ratio of 1.22 (95% CI 1.09 to 1.41) for all-cause mortality, while current smokers had a relative risk of 2.80 in men and 2.76 in women, and 4.08 to 4.41 for heavy smokers. In absolute terms, the heaviest sitters had about 190 excess deaths per 100,000 people a year versus more than 2,000 for the heaviest smokers¹.

Harmonized cohorts. The Lancet analysis pooled 13 cohorts with sitting data (1,005,791 people, 84,609 deaths, follow-up 2 to 18 years). Against a reference of under 4 hours of sitting and the most active quartile, those sitting over 8 hours and in the least active quartile had a hazard ratio of 1.59 (95% CI 1.52 to 1.66). Those who sat over 8 hours but were in the most active quartile had 1.04 (0.99 to 1.10). Even people who sat least but were least active had 1.27 (1.22 to 1.31). For TV viewing (six cohorts, 465,450 people), 3 or more hours a day raised mortality regardless of activity, except in the most active group, where only 5 or more hours did (HR 1.16, 1.05 to 1.28)³. In the 45 and Up Study (149,077 adults followed for 8.9 years), among people reporting no moderate-to-vigorous activity the all-cause HR for more than 8 versus under 4 hours of sitting was 1.52 (1.13 to 2.03), while the link was weak and inconsistent in those meeting 150 minutes a week or more⁶.

Device-measured cohorts. Pooling 8 accelerometer studies (36,383 adults, median follow-up 5.8 years, 2,149 deaths), mortality hazard ratios across increasing sedentary-time quarters were 1.28, 1.71 and 2.63 (the last: 1.94 to 3.56), and across increasing total activity quarters 0.48, 0.34 and 0.27⁷. In a US national cohort (7,985 adults aged 45 or older, 340 deaths in about 4 years), the most sedentary quartile had an HR of 2.63 (1.60 to 4.30) and the longest-bout quartile 1.96 (1.31 to 2.93)⁸. A dose-response meta-analysis of 34 studies confirmed thresholds near 6 to 8 hours of total sitting².

Trials. The 19-person crossover trial found post-meal glucose areas under the curve of 5.2 and 4.9 mmol/L per hour with light and moderate walking breaks, versus 6.9 with uninterrupted sitting⁴. The seven-trial meta-analysis found standing breaks lowered post-meal glucose modestly (effect size -0.31) and walking breaks more (-0.72 for glucose, -0.83 for insulin)⁵. These are single-day studies of blood markers, not heart attacks or deaths. A Cochrane review of 34 workplace studies (3,397 participants) found sit-stand desks reduced sitting by 100 minutes a day at up to 3 months (95% CI 84 to 116) and 57 minutes at 3 to 12 months, with low-quality evidence and no longer follow-up⁹. A meta-analysis of 46 studies (1,184 people) found standing used 0.15 kcal per minute more than sitting, or about 54 kcal a day for 6 hours of standing in a 65 kg person¹⁰.

Null and conflicting results. Sitting data are mostly self-reported or device-measured for about a week, and healthier people sit less, so reverse causation and confounding are likely. Accelerometer analyses treat sitting and activity as trading time, so sitting effects cannot be separated cleanly from missing activity⁷. The Ontario cohort of 7,320 workers found predominantly standing jobs had about twice the heart-disease risk of sitting jobs over 12 years, which is a different exposure from short spells of standing at a desk¹¹. WHO's expert group concluded evidence was insufficient to quantify a sedentary threshold¹².

Bottom line (confidence 3 of 5). Sitting is not smoking, but more than about 8 hours a day likely adds risk, mainly if you are otherwise inactive, and about an hour of moderate activity a day seems to offset it. Interrupting sitting helps blood sugar in the short term. Treat standing desks as a modest aid, and put effort into walking and regular movement instead.