Author + information
- Received February 15, 2013
- Revision received March 9, 2013
- Accepted March 13, 2013
- Published online June 11, 2013.
- Andrea K. Chomistek, ScD*,* (, )
- JoAnn E. Manson, MD, DrPH†,
- Marcia L. Stefanick, PhD‡,
- Bing Lu, MD, DrPH†,
- Megan Sands-Lincoln, PhD§,
- Scott B. Going, PhD⋮,
- Lorena Garcia, PhD¶,
- Matthew A. Allison, MD#,
- Stacy T. Sims, PhD‡,
- Michael J. LaMonte, PhD**,
- Karen C. Johnson, MD†† and
- Charles B. Eaton, MD‡‡,§§
- *Department of Nutrition, Harvard School of Public Health, Boston, Massachusetts
- †Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts
- ‡Department of Medicine, Stanford Prevention Research Center, Stanford University, Stanford, California
- §Center for Sleep and Circadian Neurobiology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania
- ⋮Department of Nutritional Sciences, University of Arizona, Tucson, Arizona
- ¶School of Medicine, University of California, Davis, California
- #Division of Preventive Medicine, Department of Family and Preventive Medicine, University of California, San Diego, California
- **Department of Social and Preventive Medicine, University at Buffalo, The State University of New York, Buffalo, New York
- ††Department of Preventive Medicine, University of Tennessee Health Science Center, Memphis, Tennessee
- ‡‡Departments of Family Medicine and Epidemiology, Alpert Medical School of Brown University, Providence, Rhode Island
- §§Center for Primary Care and Prevention, Memorial Hospital of Rhode Island, Pawtucket, Rhode Island
- ↵*Reprint requests and correspondence:
Dr. Andrea K. Chomistek, Department of Nutrition, Harvard School of Public Health, 655 Huntington Avenue, Building 2, Boston, Massachusetts 02115.
Objectives The aim of this study was to examine the independent and joint associations of sitting time and physical activity with risk of incident cardiovascular disease (CVD).
Background Sedentary behavior is recognized as a distinct construct beyond lack of leisure-time physical activity, but limited data exist on the interrelationship between these 2 components of energy balance.
Methods Participants in the prospective Women’s Health Initiative Observational Study (n = 71,018), 50 to 79 years of age and free of CVD at baseline (1993 to 1998), provided information on sedentary behavior, defined as hours of sitting/day, and usual physical activity at baseline and during follow-up through September 2010. First CVD (coronary heart disease or stroke) events were centrally adjudicated.
Results Sitting ≥10 h/day compared with ≤5 h/day was associated with increased CVD risk (hazard ratio: 1.18, 95% confidence interval: 1.09 to 1.29) in multivariable models including physical activity. Low physical activity was also associated with higher CVD risk (p for trend < 0.001). When women were cross-classified by sitting time and physical activity (p for interaction = 0.94), CVD risk was highest in inactive women (≤1.7 metabolic equivalent task-h/week) who also reported ≥10 h/day of sitting. Results were similar for coronary heart disease and stroke when examined separately. Associations between prolonged sitting and risk of CVD were stronger in overweight versus normal weight women and women 70 years of age and older compared with younger women.
Conclusions Prolonged sitting time was associated with increased CVD risk, independent of leisure-time physical activity, in postmenopausal women without a history of CVD. A combination of low physical activity and prolonged sitting augments CVD risk.
Lack of leisure-time physical activity is a major risk factor for coronary heart disease (CHD), stroke, and increased cardiovascular mortality (1–4). Humans are spending increasingly more time in sedentary behaviors, and this global trend is likely to continue, given the increasing availability and popularity of personal computers and television, automation of chores at home, increase in sedentary occupations, and transportation trends (5,6). Time spent in sedentary behavior displaces time spent in higher-intensity activities (e.g., activities of daily living), contributing to an overall reduction in total energy expenditure (7).
Emerging evidence suggests that sedentary behavior has independent effects on human metabolism, physical function, and potentially on health outcomes from low leisure-time physical activity. Sedentary behavior defined in various ways (e.g., sitting, television-watching, energy expenditure of 1.0 to 1.5 metabolic equivalent tasks [METs]) has been associated with increased risk of obesity, metabolic syndrome, type 2 diabetes, and cardiovascular disease (CVD) mortality, suggesting that it be treated as a construct distinct from physical activity (8,9). Thus, even among individuals who meet current physical activity guidelines, excessive sedentary behavior might have adverse metabolic and prognostic implications, particularly in older adults (10).
Sitting, a unique aspect of human behavior, might not simply represent the extreme low end of the physical activity continuum (7). We previously reported that both low levels of recreational physical activity and higher sitting time were associated with elevated CVD risk in the WHI-OS (Women’s Health Initiative Observational Study) (4). However, the independent and joint associations of sedentary time and physical activity with CVD events were not reported. We now extend follow-up of the cohort for an additional 10 years, update sitting and activity variables during follow-up, and examine the interrelationship of sitting time and physical activity in detail. Thus, the purpose of this study was to examine the independent and joint associations of sedentary time and physical activity with risk of incident CVD in older women. Given the limited data available on this subject, the WHI-OS affords an excellent opportunity to elucidate the complex interplay of these 2 separate components of energy balance.
The multiethnic WHI-OS cohort of 93,676 postmenopausal women, 50 to 79 years of age at study entry, was enrolled between 1994 and 1998 across 40 U.S. clinical centers. Details of the scientific rationale, study design, eligibility requirements, and baseline characteristics of the cohort have been previously published (11). Exclusion criteria included the presence of any medical condition associated with predicted survival of <3 years (e.g., class IV congestive heart failure, obstructive lung disease requiring supplemental oxygen, or severe chronic liver or kidney disease), alcoholism, mental illness, or dementia. Additional exclusions for the current analysis included history of CVD or cancer at baseline, reporting an inability to walk at least 1 block, or missing sedentary time or physical activity data, leaving 71,018 women for these analyses.
Each WHI-OS participant had a baseline clinic visit at which she completed self-administered questionnaires related to medical history, physical activity, smoking, diet, and other lifestyle-related factors and had her height, weight, waist and hip circumferences, and blood pressure measured. Additionally, participants completed periodic health forms and repeated baseline clinic assessments 3 years after enrollment. All women provided written informed consent, and the study protocol was approved by the institutional review board of each center.
Sedentary behavior was assessed by questionnaire at baseline and twice during follow-up with the following question: “During a usual day and night, about how many hours do you spend sitting? Be sure to include the time you spend sitting at work, sitting at the table eating, driving or riding a car or bus, and sitting up watching TV or talking.” Eight categories were provided for the response, ranging from <4 h/day to 16 or more h/day.
Leisure-time physical activity was assessed at baseline and during follow-up by a detailed questionnaire on walking, including the frequency of walks outside the home, average duration, and pace of each walk, and other types of activity (strenuous, moderate, and mild), including the frequency (days/week) and duration of each type. Examples of strenuous activities, defined as activities that result in increased heart rate and sweating, were aerobics, aerobic dancing, jogging, tennis, and swimming laps. Examples of moderate activity included biking outdoors, using an exercise machine (such as a stationary bicycle or a treadmill), calisthenics, easy swimming, and popular or folk dancing. Examples of mild exercise were slow dancing, bowling, and golf. Each type of activity was assigned a MET score on the basis of its energy cost (12), and physical activity-related energy expenditure (MET-h/week) was computed as the summed product of frequency, duration, and intensity.
Ascertainment of endpoints
The primary endpoints for this analysis were incidents of CHD, including nonfatal myocardial infarction (MI) and fatal CHD, and stroke. These endpoints were also combined to examine the associations of sedentary behavior and physical activity with incident CVD. Outcomes were identified on the basis of annual mailed questionnaires (response rates >95%), with permission to obtain and review medical records. Physicians blinded to exposure data confirmed self-reported diagnoses.
Nonfatal MI was confirmed according to standardized criteria of diagnostic electrocardiography changes or elevated cardiac enzymes or both (13). Stroke was confirmed by diagnosis of a typical neurological defect of sudden or rapid onset lasting ≥24 h or until death attributed to a cerebrovascular event. Fatal CHD was confirmed by documentation in hospital or autopsy records or if coronary disease was listed as the cause of death on death certificates and evidence of previous coronary disease was available.
All analyses were performed with SAS statistical software (version 9.3, SAS Institute, Inc., Cary, North Carolina). Eligible participants contributed person-time from return of baseline questionnaires to the first diagnosis of CVD, death from any cause, loss to follow-up, or September 2010.
Baseline descriptive characteristics were compared according to categories of sitting time and physical activity with linear models. For each baseline covariate, medians of sitting time or physical activity categories were modeled as a function of the covariate of interest, adjusted for age, with the p value from the resulting type 3 test used to determine statistical significance for the covariate presented.
Cox proportional hazards models were used to estimate hazards ratios (HRs) and 95% confidence intervals (CI) for outcomes. Given the strong association between age and CVD, all models were stratified on age in years with the STRATA statement for PHREG in SAS. Tests for linear trend were computed with the medians for categories modeled as an ordinal variable. Statistical significance was defined as p < 0.05. The possibility of nonlinear relations between sitting time, physical activity, and CVD were examined nonparametrically with restricted cubic splines (14). Tests for nonlinearity used the likelihood ratio test, comparing the model with only the linear term with the model with linear and cubic spline terms.
Both exposures were included simultaneously in the model, to assess independent associations of sitting time and physical activity with CVD risk. Simple updated levels of sitting time and physical activity, in which outcomes were predicted from the most recent questionnaire, were used. For example, events that occurred between baseline and Year 3 of follow-up were examined in relation to exposures reported on the baseline questionnaire, events occurring between Year 3 and Year 6 of follow-up were examined in relation to exposures reported on the Year 3 questionnaire, and so forth. Sitting time and physical activity were first examined as categorical variables and then as continuous variables to show the impact on outcomes/1-U increment of each exposure. Sitting was categorized as ≤5 h/day, 5.1 to 9.9 h/day, and ≥10 h/day, which were approximate tertiles on the basis of distribution of the data. For physical activity, women were classified into 4 categories: inactive (≤1.7 MET-h/week); low (1.8 to 8.3 MET-h/week); medium (8.4 to 20 MET-h/week); and high activity (>20 MET-h/week)—with accumulating 150 min/week of moderate-intensity exercise (i.e., the minimum dose of activity recommended by the federal government) (15) being equivalent to at least 8.4 MET-h/week of exercise. To assess the joint association of sitting time and physical activity with risk of CVD, participants were cross-classified into 12 groups according to the levels of sitting time and physical activity. The interaction was assessed by the difference in −2 log likelihood between the model containing the cross-classified sitting time–physical activity variables and the main effects model.
All multivariable models were stratified by age in years and included the following covariates: race/ethnic group; family income; education; marital status; smoking; parental history of premature MI; depression; alcohol intake; hours of sleep; and intake of total calories, saturated fat, and fiber. Covariates that were reassessed during follow-up were updated over time with the most recent value. Secondary analyses were additionally adjusted for history of hypertension, diabetes, high cholesterol levels, and body mass index (BMI).
Whether associations between sedentary time and CVD were modified by BMI (<25 vs. ≥25 kg/m2), age (<70 vs. ≥70 years), or employment status (employed vs. unemployed) was also examined. Interactions were tested by examining the difference in −2 log likelihood between the model containing interactions with the potential effect modifiers and the main effects model.
As a secondary analysis, the association between change in sitting time and CVD risk was examined. Change in sitting time between baseline and Year 3 was used to examine events that occurred between Year 3 and Year 6 of follow-up; change in sitting time between Year 3 and Year 6 was used to examine events that occurred from Year 6 on. Change in sitting time was examined as a continuous variable and with the following categories: decreased sitting by more than 2 h/day; no change; and increased sitting by more than 2 h/day.
During a median follow-up of 12.2 years (interquartile range: 8.7 to 14.0 years), 2,411 incident cases of CHD, 2,050 incident cases of strokes, and 4,235 first CVD events were documented. Baseline sitting time and physical activity in relation to other potential risk factors for CVD are presented in Table 1. Women who reported ≥10 h/day of sitting were more likely to be white, to have attended college, and to have a higher income compared with women who reported ≤5 h/day of sitting; whereas, the opposite was true of women reporting less physical activity compared with highly active women. Additionally, greater time spent sitting and less physical activity were positively associated with current smoking, higher BMI, and self-reported depression (Table 1).
In multivariable-adjusted analyses, increased sitting time and decreased physical activity were positively associated with risk of CHD and stroke (Table 2). Because the results for CHD and stroke were similar, the endpoints were combined to examine total CVD. In multivariable models that included physical activity, the HR for total CVD comparing ≥10 h/day of sitting time with ≤5 h/day was 1.18 (95% CI: 1.09 to 1.29) (p for trend <0.001). With high physical activity (i.e., >20 MET-h/week) as the reference group, the CVD risks for medium (8.4 to 20 MET-h/week), low (1.8 to 8.3 MET-h/week) activity, and inactive (≤1.7 MET-h/week) groups were 1.16 (95% CI: 1.06 to 1.27), 1.30 (95% CI: 1.18 to 1.42), and 1.47 (95% CI: 1.33 to 1.62), respectively (p for trend <0.001). All associations were mildly attenuated after adjustment for BMI and history of comorbidities but remained statistically significant (Table 2). When examined as continuous variables, each hour/day of sitting time was associated with 2% higher risk of CVD (HR: 1.02, 95% CI: 1.01 to 1.03), and each MET-h/week of physical activity was associated with a 1% lower risk of CVD (HR: 0.990, 95% CI: 0.987 to 0.992). There was no evidence of nonlinearity when applying restricted cubic splines to the association between sitting time (p = 0.87) or physical activity (p = 0.60) and risk of CVD.
When participants were cross-classified on the basis of both sitting time and leisure-time physical activity, only 18% of the high activity group reported sitting for ≥10 h/day compared with 32% of the physically inactive group. Except in the most active women, more time spent sitting increased CVD risk in each physical activity group (Fig. 1), with CVD risk being highest in physically inactive women who also reported ≥10 h/day of sitting (HR: 1.63, 95% CI: 1.39 to 1.90); however, the interaction between sitting time and physical activity was not statistically significant (p for interaction = 0.94).
There were significant interactions of sitting time with CVD risk within subgroups defined by BMI and age but not by employment status (Fig. 2) (p for interaction = 0.044, 0.026, and 0.22, respectively). Time spent sitting was associated with increased CVD risk in women with BMI ≥25 kg/m2 (HR: 1.26, 95% CI: 1.13 to 1.40 for sitting ≥10 h/day; p for trend <0.001) but not in women with BMI <25 kg/m2 (HR: 1.02, 95% CI: 0.88 to 1.19; p for trend = 0.85). Sitting time was also associated with a higher CVD risk among women 70 years of age and older (HR: 1.22, 95% CI: 1.09 to 1.36 for sitting ≥10 h/day; p for trend <0.001) but not among younger women (HR: 1.08, 95% CI: 0.94 to 1.25; p for trend = 0.23). Although the interaction was not statistically significant (p for interaction = 0.22), more sitting time was associated with higher CVD risk in unemployed women (HR: 1.21, 95% CI: 1.10 to 1.34; p for trend <0.001) but not employed women (HR: 1.07, 95% CI: 0.89 to 1.30; p for trend = 0.36).
Finally, an increase in sitting time over a 3-year period was associated with an increased risk of CVD. Compared with participants who reported no change in sitting time, the HR for participants who increased sitting time by more than 2 h/day was 1.18 (95% CI: 1.07 to 1.31) after adjusting for covariates, whereas the HR for participants who decreased sitting time by more than 2 h/day was 1.01 (95% CI: 0.91 to 1.13). When examined continuously, a 1-h/day increase in sitting time over a 3-year period was associated with a 1.4% increase in CVD risk (HR: 1.014, 95% CI to 1.001 to 1.027, p = 0.03) in the multivariable-adjusted model.
In this large, prospective study of post-menopausal U.S. women, prolonged sitting time was associated with increased risk of incident CHD, stroke, and total CVD, independent of leisure-time physical activity; however, low levels of leisure-time physical activity were also strongly associated with increased CVD risk, after adjusting for sitting time. Women who were physically inactive and reported ≥10 h/day of sitting time, comprising 6% of our study population, were at 63% greater CVD risk than highly active women who reported sitting ≤5 h/day, after adjusting for several cardiovascular risk factors. Associations between prolonged sitting and risk of CVD were stronger in overweight and obese versus normal weight women and women 70 years of age and older compared with younger women.
Few studies have examined the association between sitting time and risk of incident cardiovascular disease; to date, the outcome of interest has been largely limited to CVD mortality (4,10,16–18). Thus, the current analysis, which includes incident CHD and stroke separately in relationship to sitting time and physical activity habits within a large well-characterized cohort of older postmenopausal women as well as combined CVD events, is an important contribution to the published data. Our results are also consistent with prior studies that have reported an increased risk of CVD mortality as time spent sitting increases, independent of usual leisure-time physical activity.
The association between physical activity, hours spent sitting, and CVD has been previously reported among WHI-OS participants (4). Manson et al. (4) reported that high levels of recreational physical activity were associated with 30% to 40% reductions in CVD risk over a mean 3.2 years of follow-up; and, with baseline measures of sitting time, women who spent at least 16 h/day sitting had an increased CVD risk (HR: 1.68, 95% CI: 1.07 to 2.64), compared with those who spent <4 h/day sitting. The joint effects of sitting time and physical activity were not reported in this previous study.
Although the interaction between sitting time and physical activity was not statistically significant in the present analysis, clinically significant attenuation of CVD risk was found in all women, except for those in the high activity group, who reported sitting <5 h/day. Because the least active women were also the most likely to report prolonged time spent sitting, reducing sitting time could potentially reduce CVD risk substantially among less active women. The finding that the association between prolonged sitting and CVD is attenuated among the most active women is similar to studies of mortality (16,18) and shows that it might be beneficial to participate in regular exercise despite engaging in other potentially detrimental behaviors, like prolonged sitting. The most active group in our study reported >20 MET-h/week of physical activity, which exceeds the physical activity guidelines substantially. However, women engaging in 8.4 to 20 MET-h/week of physical activity, which meets the physical activity guidelines, were still at increased CVD risk if they reported prolonged sitting.
The association between prolonged sitting time and increased CVD risk was significant only in overweight women and women 70 years and older. Previous studies (16,18) have reported similar results, where the positive association between time spent sitting and mortality was stronger in overweight and obese versus normal weight individuals. These findings emphasize the importance of limiting time spent sitting, particularly in overweight and obese individuals. The fact that the association between sitting time and CVD was stronger in older compared with younger women might be due to sitting time including both occupational and leisure sitting in this study. Thus, in older women, sitting might comprise more television watching, because they are more likely to be retired compared with younger women. Television watching has previously been found to be more detrimental than overall sitting time (10).
Additionally, increasing sitting time over a 3-year period was associated with a higher risk of CVD. Reducing time spent sitting, however, was not associated with CVD risk in this study. This finding might indicate that change in sitting time needs to be assessed over a longer period of time. Alternatively, a better measurement of sitting time, like accelerometers, or a randomized controlled trial might be necessary to truly examine the effect of changing sedentary behaviors.
Physiological responses associated with prolonged sitting, such as suppression of skeletal muscle lipoprotein lipase activity, which is necessary for triglyceride uptake and high-density lipoprotein cholesterol production, and reduced glucose uptake (9,19,20) might explain the independent effect of sedentary behavior on CVD risk factors and also on CVD risk. Sedentary behavior might also be related to overweight and obesity through increased energy intake and decreased energy expenditure (21).
In addition to reducing sitting time, taking breaks during prolonged periods of sitting (i.e., by standing up and taking short walks) might be beneficial in lowering CVD risk. In accelerometer studies, having a higher number of breaks in sedentary time was inversely associated with waist circumference, BMI, triglycerides, 2-h plasma glucose, and C-reactive protein, independent of total time spent sitting and physical activity (22,23). Additionally, a recent randomized trial showed that interrupting sitting time with short bouts of light- or moderate-intensity activity lowers postprandial glucose and insulin levels in overweight adults (24).
Strengths of our study include its prospective design; the large, multiethnic, geographically diverse cohort of postmenopausal women; detailed assessment of physical activity; and availability of several relevant covariates for analysis. Additionally, this study is the first to investigate the association with CVD using serially obtained measurements of sitting time and physical activity, thereby enabling us to update exposure status during follow-up. We were also able to assess the independent associations between sitting time and usual recreational physical activity with CVD risk.
Our study also has several limitations, including the fact that analyses were limited to postmenopausal women; therefore, findings cannot necessarily be generalized to men or younger women. Sitting time and physical activity were self-reported, although it is worth noting that physical activity measures were validated in this cohort (11,25). Moreover, measurement error is unlikely to bias our results, because sedentary time and physical activity were assessed prospectively so any reporting errors would be nondifferential with respect to subsequent disease status. Nonetheless, device-based measures of sitting time and physical activity, like accelerometers, might provide more accurate assessments of these variables. Another limitation of this study is that we were unable to examine the associations with different types of sedentary behavior, like television watching or transportation, because the questionnaire only assessed total sitting time. Finally, as with most such studies, the possibility of residual confounding by other lifestyle and behavioral factors must be considered.
The findings from this large multiethnic group of postmenopausal women indicate that prolonged sitting time significantly increases risk of CVD, independent of leisure-time physical activity. Given the projected population growth of U.S. women 65 years of age and older and the relatively high prevalence of physical inactivity, the present findings have important public health implications. Reducing sitting time among older women who are less active could potentially reduce risk of CHD and stroke, major causes of morbidity in older women. Moreover, for individuals who are unable or averse to exercise, amount of time spent sitting might be more amenable to change than increasing levels of physical activity.
The authors thank the Women’s Health Initiative investigators and staff for their dedication and the study participants for their continued cooperation and participation (Online Appendix).
The Women's Health Initiative program is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, U.S. Department of Health and Human Services through contracts HHSN268201100046C, HHSN268201100001C, HHSN268201100002C, HHSN268201100003C, HHSN268201100004C, and HHSN271201100004C. Dr. Chomistek was supported by grants CA152904 from the National Cancer Institute and HL035464 from the National Heart, Lung, and Blood Institute and an institutional training grant (DK007703-17) from the National Institute of Diabetes and Digestive and Kidney Diseases. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Peter Kokkinos, MD, served as Guest Editor for this paper.
- Abbreviations and Acronyms
- body mass index
- coronary heart disease
- confidence interval
- cardiovascular disease
- hazard ratio
- metabolic equivalent task
- myocardial infarction
- observational study
- Women’s Health Initiative
- Received February 15, 2013.
- Revision received March 9, 2013.
- Accepted March 13, 2013.
- American College of Cardiology Foundation
- Sattelmair J.,
- Pertman J.,
- Ding E.L.,
- Kohl H.W. III.,
- Haskell W.,
- Lee I.M.
- Eaton C.B.,
- Medalie J.H.,
- Flocke S.A.,
- Yaari B.A.,
- Zyzanski S.J.,
- Goldbourt U.
- Church T.S.,
- Thomas D.M.,
- Tudor-Locke C.,
- et al.
- Matthews C.E.,
- George S.M.,
- Moore S.C.,
- et al.
- Physical Activity Guidelines Advisory Committee
- Patel A.V.,
- Bernstein L.,
- Deka A.,
- et al.
- Zderic T.W.,
- Hamilton M.T.
- Hamilton M.T.,
- Hamilton D.G.,
- Zderic T.W.
- Healy G.N.,
- Dunstan D.W.,
- Salmon J.,
- et al.
- Healy G.N.,
- Matthews C.E.,
- Dunstan D.W.,
- Winkler E.A.,
- Owen N.
- Dunstan D.W.,
- Kingwell B.A.,
- Larsen R.,
- et al.