Standing up is not enough.

Taking breaks from long stretches of sitting is not, on its own, associated with a lower risk of dying from cancer, but replacing that sitting time with physical activity is, an accelerometer study of more than 91,000 adults showed.

Over more than a decade of follow-up, cancer death occurred in 2.8% of participants in the highest quartile of daily sedentary time compared with 1.3% in the lowest quartile, reported senior author Frederick K. Ho, PhD, of the University of Glasgow, Glasgow, Scotland. However, a model suggested that trading an hour of prolonged sitting for an hour of interrupted sitting was not associated with lower cancer mortality.

“My take is, try to incorporate more physical activity in our daily life, especially for those working at a desk job,” Ho told Medscape Medical News.

The study was published on July 2 in PLOS Medicine.

Sedentary time has been linked to cancer for more than a decade but largely on the basis of self-reported data, according to the investigators.

The World Health Organization recommends that adults limit sedentary time and replace it with activity of any intensity, including light activity; however, it does not offer exact guidance on ideal durations for sitting or taking breaks, citing a lack of evidence.

In the new study, Ho and colleagues analyzed data from 91,292 UK Biobank participants with no cancer at baseline who used wrist-worn accelerometers for 7 days between 2013 and 2015.

Median follow-up was 12.4 years. A machine-learning classifier divided sedentary time into prolonged bouts, defined as at least 30 minutes, and interrupted sedentary time, comprising shorter periods or longer ones broken up by movement.

Cox models were adjusted for age, sex, ethnicity, deprivation, education, smoking, alcohol, and diet, with a sensitivity analysis adding BMI and the number of long-term conditions. The models were not adjusted for physical activity.

Each additional hour per day of prolonged sedentary time was associated with a 9% increased risk for cancer death (hazard ratio [HR], 1.09), whereas each additional hour per day of interrupted sedentary time was associated with an 18% lower risk (HR, 0.82).

The above findings appear to suggest that interrupted sitting is associated with a lower risk for cancer death, but when the investigators held the whole 24-hour day constant, so that any hour added to one behavior was subtracted from another, replacing prolonged sitting with interrupted sitting was no longer associated with lower cancer mortality. Only replacing prolonged sitting with light or moderate activity was associated with a lower risk.

“In my opinion, what these people have proved has nothing to do with sitting,” said Gideon Meyerowitz-Katz, PhD, epidemiologist at the University of Wollongong in Wollongong, Australia, who was not involved in the study. “All they’ve shown is that exercising is good for your health.”

Simply standing up at a desk is not enough, Meyerowitz-Katz told Medscape Medical News. The benefit appears when sitting is replaced with exercise. “But that’s not really a comment on being sedentary,” he said. “It’s a comment on exercise.”

Ho did not dispute the point so much as reframe it. “We can’t change sedentary behavior without physical activity,” he said, noting that the benefits of physical activity are always mixed with the effect of displacing sedentary time.

How Do These Findings Differ From Related Research?

The closest predecessor to this new study was the REGARDS cohort study, which included approximately 8000 US adults. That study found that total sedentary time predicted cancer death after adjustment for moderate-to-vigorous activity, but sedentary bout duration did not (HR, 1.36; 95% CI, 0.96-1.93; highest vs lowest tertile).

Keith M. Diaz, PhD, Florence Irving associate professor of behavioral medicine at Columbia University Medical Center in New York City, who was senior author of the REGARDS analysis, pointed to a difference in the models.

“The researchers did not adjust for physical activity levels in their models so it’s hard to make a comparison here,” he said. Moderate-to-vigorous activity is well established to offset some of the risk associated with prolonged sedentary behavior, and “in our study, we saw this effect. It’s possible they might have seen a similar effect had they adjusted for moderate-vigorous intensity physical activity.”

REGARDS also used a hip-worn accelerometer, which detects posture; this study used wrist-worn devices, which detect movement.

Cause or Symptom?

Meyerowitz-Katz also raised concern about reverse causation: If people who were already ill sat more because of their illness, the sitting would be a symptom of disease rather than a cause.

Excluding the first 2 years of follow-up barely moved the association from an HR of 1.09-1.08, but Meyerowitz-Katz said that step cannot fully settle the question when the endpoint is death.

“Because it’s mortality, as opposed to something that’s sort of immediate, I don’t think it’s possible to fully exclude reverse causation,” he said. A person with a BMI of 40, diabetes, and cardiovascular disease may not die within 2 years, he said, but “you will likely be much more sedentary than someone who is of the same age and doesn’t have those health issues.”

Ho agreed that reverse causation cannot be entirely ruled out; however, it is unlikely to explain the findings completely, he added, citing UK data showing that the median interval from first symptom to cancer diagnosis ranges from about 26 days to 156 days, far short of 2 years.

The study was funded by the American Cancer Society and the US National Institutes of Health. Ho reported serving as a paid statistical consultant on the PLOS Medicine statistical board; the remaining authors reported no competing interests. Meyerowitz-Katz and Diaz reported no relevant financial relationships.