By Ellen Feldman, MD
Altru Health System, Grand Forks, ND
- A review and meta-analysis of eight studies using accelerometer-generated records of physical activity reveals a non-linear dose-response relationship between all levels of patient-reported physical activity and reduced risk of all-cause mortality.
- In addition, nine and a half hours or more per day of sedentary time (excluding sleep time) was associated with higher risk of all-cause mortality.
- These associations held up across a variety of different types of studies and measures of physical activity; standardization techniques allow for inclusion of diverse studies, increasing the generalizability of the findings.
SYNOPSIS: This review of eight studies with 36,383 participants looks at objective measurements of physical activity and finds a dose-response association between any level of activity and reduced risk of death and between more sedentary time and higher risk of death.
SOURCE: Ekelund U, Tarp J, Steene-Johannessen J, et al. Dose-response associations between accelerometry measured physical activity and sedentary time and all-cause mortality: Systematic review and harmonized meta-analysis BMJ 2019;366:l4570
As evidence mounts that a sedentary lifestyle is linked to a development of chronic diseases, that physical inactivity is associated with premature death, and that more active lifestyles are associated with longer lifespans, public health guidelines emphasizing activity and movement are gaining traction.1,2 Many of the studies that provide evidence of these associations rely on people reporting their physical activity or sedentary time. This is a major limitation, since people often do not provide accurate estimates of their activity time, intensity, or sedentary time.3 Objective measurements of physical activity, such as accelerometers, provide a way to overcome this limitation and provide more precise estimates of activity levels. Notably, accelerometers offer an advantage over devices, such as the more common pedometer, as an accelerometer measures a change in velocity of an object over time and thus allow inference of intensity of action.4
However, Ekelund et al noted that, with few deaths, short follow-up periods, and diverse analytical methods, even studies using accelerometers have limitations in detecting associations between intensity of physical activity, sedentary time, and all-cause mortality. To overcome these limitations, Ekeland et al conducted a meta-analysis using specific “harmonization” techniques by asking authors of the included studies to reanalyze data in a consistent manner. In doing so, the participant pool expanded to more than 36,000 individuals with more than 240,000 person-years of follow-up, yielding more robust and generalizable results.
Out of 518 identified articles, a total of eight high-quality studies appropriate for inclusion in this harmonized meta-analysis emerged. Studies originated in the United States, United Kingdom, and Scandinavia. Three of the studies were from nationally representative samples, such as the Framingham Heart Study and the National Health and Nutrition Examination Survey. Four quartiles of activity level were created, with the first quartile signaling the least amount of activity and serving as a benchmark. Adjustments were made for multiple factors, including sex, age, socioeconomic status, smoking, and body mass index. For consistency, participant age was restricted to 40 years and older. Total wear time of the accelerometer ranged from three to seven days. To start addressing the issue of reverse causation, data from participants who died within the first two years of follow-up were excluded.
RESULTS
Looking at total physical activity and all-cause mortality, Table 1 shows the hazard ratios for all-cause mortality associated with the quartiles of activity level, beginning with the referent (least active) in quartile one. Adjustments are made to control for multiple variables.
Table 1. Hazard Ratios for All-Cause Mortality Associated with Quartiles of Activity Level
|
|
|
Quartile 1
(Least Active)
|
Quartile 2
|
Quartile 3
|
Quartile 4
(Most Active)
|
|
Hazard ratio
|
Referent = 1
|
0.48 (95% CI, 0.43-0.54)
|
0.34 (95% CI, 0.26-0.45)
|
0.27 (95% CI, 0.23-0.32)
|
|
Number of participants
|
9,096
|
9,105
|
9,096
|
9,086
|
|
CI: confidence interval
|
Looking at physical activity intensity and all-cause mortality, Table 2 shows the hazard ratios for all-cause mortality associated with quartiles of light physical activity and moderate-to-vigorous physical activity. Adjustments are made to control for multiple variables.
Table 2. Hazard Ratios for All-Cause Mortality Associated with Quartiles of Activity Level
by Intensity
|
|
|
Quartile 1
(Least Active)
|
Quartile 2
|
Quartile 3
|
Quartile 4
(Most Active)
|
|
Hazard ratio of light physical activity
(min/day)
|
Referent = 1
(n = 9,073)
|
0.60 (95% CI, 0.54-0.68)
(n = 9,101)
|
0.44 (95% CI, 0.38-0.51)
(n = 9,090)
|
0.38 (95% CI, 0.28-0.51)
(n = 9,119)
|
|
Hazard ratio of moderate-to-vigorous physical activity
(min/day)
|
Referent = 1
(n = 9,002)
|
0.64 (95% CI, 0.55-0.74)
(n = 9,153)
|
0.55 (95% CI, 0.40-0.74)
(n = 9,123)
|
0.52 (95% CI, 0.43-0.61)
(n = 9,105)
|
|
CI: confidence interval
|
Further analysis by Ekelund et al shows that maximal risk reduction is seen in slightly more than six hours daily of light physical activity and with 24 min/day of moderate-to-vigorous-intensity physical activity. With a goal of determining if sedentary time itself is associated with a high risk of premature mortality, sedentary time was analyzed independently of activity time. Table 3 shows the hazard ratios for all-cause mortality associated with quartiles of sedentary time (quartile one representing the least amount of sedentary time).
Table 3. Meta-Analysis for Mean Sedentary Time by Quartile
|
|
Mean Sedentary Time (Range)
|
448 min/day
(371-519)
|
520 min/day
(469-588)
|
578 min/day
(542-639)
|
650 min/day
(624-705)
|
|
Deaths/Year (National Database)
|
Referent = 1
|
1.28 (95% CI, 1.09-1.51)
|
1.71 (95% CI, 1.36-2.15)
|
2.63 (95% CI, 1.94-3.56)
|
|
CI: confidence interval
|
COMMENTARY
This ambitious review and meta-analysis begins to crack the code on the relationship between activity intensity, inactivity, and the risk of all-cause mortality. Performing a harmonized meta-analysis allowed researchers to merge results across studies and ultimately provide much-needed clarity to yield more powerful, actionable data. The results show that higher levels of physical activity and lower levels of sedentary time are associated with a lower risk of premature death. The dose-response curve for this respondent pool of age 40 years and older is non-linear, with the largest drops in risk occurring when comparing participants in the second quartile to the referent or first quartile.
The landmark London Busman study from the 1950s revealed that bus drivers had a higher rate of cardiovascular disease than bus conductors and sparked an interest in the relationship between sedentary lifestyle and development of chronic illness.5 Meanwhile, observational studies have repeatedly confirmed the association between physical activity and health.2,6 Ekelund et al present a review and meta-analysis to further investigate these relationships and pin down specifics of intensity of activity needed for health benefits and further understand the degree of sedentary behavior that confers a health risk.
The finding of a non-linear relationship between the intensity of physical exercise and all-cause mortality gives hope to those who are physically impaired and unable to exert full intensity of effort. Increasing activity enough to move from the first or least active quartile into the second quartile of activity yields significant benefit in terms of risk reduction for premature death. Notably, the largest risk for premature death linked to sedentary time is seen above 9.5 hours of sedentary time daily, excluding sleep time. While risk is elevated at lesser amounts of sedentary time, there is a substantial jump in risk from the third quartile to the fourth, most sedentary group. Results here again highlight the potential effect of fairly minor activity level adjustments on overall well-being.
Despite the fact that this study was able to overcome many limitations of prior research, there still are limitations present. To minimize reverse causation, researchers excluded participants who died within the first two years of follow-up. However, there remains a possibility that less-active participants at baseline had a greater likelihood of early mortality due to preexisting disorders that limited mobility and activity. Future studies may want to control for this possibility. Another clear limitation of this study is the reliance on four to seven days of accelerometer data. The use of an objective device to measure activity represents an improvement on self-reported measures of activity, but there is no information from this meta-analysis about the benefits or risks of continued activity in this age group. There also is no information about an association between ongoing activity levels and longevity. Studies looking at changes in activity level over time and association with health are needed. And finally, it is worth recognizing that correlation does not imply causality and that these results were generated from individuals ages 40 years and older from wealthy countries. Further studies are needed for the generalization of conclusions and recommendations.
For now, the primary care provider is on firm ground continuing to recommend movement as a pathway to health. This study may be used to reassure patients that even low-intensity movement or activity (activity during which an individual can comfortably talk and sing) for at least six hours of waking time daily is beneficial and that limiting sitting or sedentary time to under 9.5 hours daily is essential.
If the numbers seem daunting or overwhelming, it may be helpful to generate practical interventions with patients, such as combining walking, stretching, and sitting or using new technologies to provide feedback and positive reinforcement. When looking into the effect of movement on health, it appears that even small changes can confer significant benefit. n
REFERENCES
- U.S. Department of Health and Human Services. Physical activity guidelines for Americans. Updated Aug. 25, 2021. https://health.gov/our-work/nutrition-physical-activity/physical-activity-guidelines
- Centers for Disease Control and Prevention. Benefits of physical activity. Updated April 27, 2022. https://www.cdc.gov/physicalactivity/basics/pa-health/index.htm
- New York City Department of Health and Mental Hygiene. Epi Research Report: Self-Reported and Accelerometer-Measured Physical Activity: A Comparison in New York City. New York;2013. https://www1.nyc.gov/assets/doh/downloads/pdf/epi/epiresearch-pa_measures.pdf
- Medical Research Council. Accelerometers. https://dapa-toolkit.mrc.ac.uk/physical-activity/objective-methods/accelerometers
- Heady JA, Morris JN, Kagan A, Raffle PA. Coronary heart disease in London busmen. A progress report with particular reference to physique. Br J Prev Soc Med 1961;15:143-153.
- Cheval B, Sivaramakrishnan H, Maltagliati S, et al. Relationships between changes in self-reported physical activity, sedentary behaviour and health during the coronavirus (COVID-19) pandemic in France and Switzerland. J Sports Sci 2021;39:699-704
