Journal Logo


Sleep, Sedentary Behavior, Physical Activity, and Cardiovascular Health: MESA


Author Information
Medicine & Science in Sports & Exercise: April 2021 - Volume 53 - Issue 4 - p 724-731
doi: 10.1249/MSS.0000000000002534


Sleep, sedentary behavior, and physical activity are each independently associated with cardiovascular health (CVH) (1,2). Evidence-based guidelines recommend seven or more hours of sleep, low levels of sustained sedentary behavior, and high levels of physical activity for optimal CVH (3–5). A significant limitation of research to date is that the relationship of these behaviors with CVH is investigated in isolation, although a change in any given behavior will affect time spent in others.

The 24-h activity cycle was developed as a paradigm for exploring the effects of time spent in each of four domains: sleep, sedentary behavior, light-intensity physical activity (LIPA), and moderate to vigorous physical activity (MVPA) on health outcomes (6). Isotemporal substitution modeling can be used to estimate the effect of substituting a specific duration of one type of activity with an equivalent duration of another (7). For example, Buman et al. (8) used this analytic approach to demonstrate that substituting 30 min·d−1 of sedentary time to sleep, LIPA, or MVPA is associated with improvements in various cardiovascular disease (CVD) risk biomarkers (8). Importantly, although many of these studies have used objectively measured physical activity information, few have used objectively measured sleep, and many have focused on single cardiometabolic risk factors on health outcomes rather than overall CVH.

To our knowledge, no study has evaluated the complex interrelationships of accelerometer-measured sleep, sedentary behavior, LIPA, and MVPA with overall CVH. In addition, these relationships have not been explored in an ethnically diverse cohort of middle-age and older U.S. adults, a population known to have high rates of sedentary behavior and a low prevalence of ideal CVH (9,10). In light of these gaps in the literature, this study aimed to use partition and isotemporal substitution modeling to understand how relationships between these health behaviors influence CVH assessed using the American Heart Association Life’s Simple 7 (AHA LS7) score (11) in the Multi-Ethnic Study on Atherosclerosis (MESA) Sleep Ancillary Study (12,13). We hypothesized that substituting sedentary behavior for an equivalent amount of sleep or physical activity would result in a higher overall CVH score, facilitated by increases in its component metrics.


Study Population

MESA is an ongoing prospective cohort study designed to investigate the prevalence and progression of subclinical CVD and to identify CVD risk factors in middle-age and older adults (45–84 yr). A total of 6814 participants free of clinical CVD from four racial/ethnic groups (38% Caucasian, 12% Chinese American, 28% African American, and 22% Hispanic) (12) were recruited from six field centers across the United States: Baltimore, MD; Chicago, IL; Los Angeles, CA; New York, NY; Saint Paul, MN; and Winston-Salem, NC. Six clinical examinations to collect data on clinical, sociodemographic, and lifestyle factors have been completed since the inception of this cohort in 2000. At clinical examination 5 (2010–2013), 4077 participants were invited to participate in the MESA Sleep Ancillary Study, which was designed to examine objectively measured sleep characteristics in relation to CVD risk. A total of 2261 participants consented to participate in the Sleep Ancillary Study, of whom 2156 wore an accelerometer for 7 d and nights on their wrist. Participants without valid accelerometer data (n = 218) were not included in the analyses. Among the remaining 1938 individuals, 214 did not have sufficient data to compute the CVH score, and an additional 6 did not have data on key covariates, resulting in a final sample of 1718 individuals. MESA was approved by the Institutional Review Board at each site, and all participants provided written informed consent.

Assessment of Sleep, Sedentary Behavior, and Physical Activity

Sleep duration, sedentary behavior, and physical activity duration and intensity were ascertained using the Actiwatch Spectrum accelerometer (Philips Respironics, Murrysville, PA) worn on each participants’ nondominant wrist for seven consecutive days. After removing non–wear periods, individuals without 600 daily minutes of wear time on at least 3 d were excluded from analyses. These parameters were chosen based on previous literature (14,15). Sleep duration was estimated by scoring these data during 30-s epochs as “sleep” or “wake” by Actiware-Sleep v.5.59 analysis software (Mini Mitter Co., Inc., Bend, OR) after manually editing the sleep period using sleep diary data and event and light markers. A validated algorithm was used to calculate the activity count for each epoch (16). The average sleep duration was computed from nighttime sleep onset (sleep start time) and morning wakening (sleep end time), averaging sleep duration across all nights that the accelerometer was worn. Records were assessed by blinded, trained scorers at the Brigham and Women’s Hospital Sleep Reading Center, with high inter- and intraclass correlations for sleep duration metrics.

To calculate daily minutes of sedentary time, LIPA, and MVPA, nonsleep data were rescored based on 60-s epochs using published wrist-worn cut points to isolate minutes spent in sedentary behavior (<178.5 counts per minute), LIPA (178.6–562.4 counts per minute), and MVPA (>562.4) (17). After non–wear periods were removed, total minutes of nighttime sleep, sedentary time, LIPA, and MVPA were summed and averaged across days to produce average daily minutes of each variable.

Assessment of CVH Metrics

Body mass index (BMI) (kg·m−2) was calculated from weight and height measurements obtained by trained personnel during MESA examination 5. A validated 128-item food frequency questionnaire was administered to collect data on habitual dietary intake (12,18). Seated systolic blood pressure (SBP) and diastolic blood pressure (DBP) were assessed from three readings after participants had rested for 5 min and the average of the last two readings was recorded. Total cholesterol (including HDL and LDL cholesterol) and blood glucose levels were measured from fasting blood samples obtained during the study visit. Smoking status (currents vs former vs never), BP, cholesterol, and diabetes medication were assessed by self-report using a standardized questionnaire.

Operationalization of the CVH Score

Lifestyle behaviors, anthropometric measures, and biomarkers that constitute the AHA LS7 score assessed at examination 5 were used to compute the CVH score. Published definitions (11) for assessing CVH using the AHA LS7 were adapted based on data available in the MESA Sleep Ancillary Study and MESA examination 5 and are consistent with previous literature on the AHA LS7 score (19,20). For the present analyses, physical activity was excluded from the CVH score given its status as an exposure variable of interest. We computed the CVH score based on meeting guidelines for smoking, diet, BMI, BP, total cholesterol, and fasting glucose. Briefly, participants were assigned a score based on their level of meeting recommendations for these six CVH metrics: ideal (2 points), intermediate (1 point), or poor (0 point). Participants received a score of 2 if they met the following criteria: 1) were never smokers, 2) met four to five of diet guidelines (11), 3) had BMI <25 kg·m−2, 4) had untreated SBP < 120 mm Hg and DBP < 80 mm Hg, 5) had untreated total cholesterol <200 mg·dL−1, and 6) had untreated fasting glucose <100 mg·dL−1. They received a score of 1 for each CVH metric if they met the following criteria: 1) were former smokers, 2) met two to three of diet guidelines, 3) had BMI of 25–29.9 kg·m−2, 4) had SBP 120–129 mm Hg or DBP <80 mm Hg or treated to ideal level, 5) had total cholesterol 200–239 mg·dL−1 or treated to ideal level, and 6) had fasting glucose of 100–125 mg·dL−1 or treated to normal level. Otherwise, they received a score of 0 for each CVH metric. The individual component scores were summed to create a global CVH score ranging from 0 to 12, such that higher scores represented a more favorable CVH profile. Participants with a score of 0–5, 6–9, and 10–12 were considered to have poor, moderate, and high CVH, respectively (20,21).

Statistical Analysis

Descriptive statistics were calculated for the sociodemographic, lifestyle, and cardiometabolic risk factors. Next, a series of partition models were conducted to examine the linear relationship between the average minutes of sleep, sedentary time, LIPA, MVPA, and CVH score. Because of the high correlation between sedentary time and LIPA, we ran additional sensitivity analyses. The first excluded average sedentary time while retaining average time spent in LIPA, and the second excluded time spent in LIPA while including sedentary time. All partition models controlled for sex (male vs female), age (yr), marital status (married vs unmarried), education attainment (college educated vs less than college education), and race (White vs non-White).

Isotemporal substitution models were conducted to examine the effect of substituting 30 min of sedentary time for an equal duration of LIPA, MVPA, or sleep (7). First, a total time variable was computed by summing all time variables (i.e., total time = sedentary + LIPA + MVPA + sleep). Next, all variables were divided by 30 to facilitate examination of a 30-min substitution. Finally, all behaviors except sedentary time plus the total time variable (i.e., LIPA, MVPA, sleep, and total time) were included with the covariates of interest. Including total time in the model constrains time, and as a result, the regression estimate for each included behavior represents the estimated mean change in levels of the outcome that is observed when that behavior is increased by 30 min and sedentary time is simultaneously decreased by 30 min. Our estimates are based on cross-sectional associations, and thus results cannot be interpreted causally or interpreted as temporal changes within individuals.

The main substitution model of interest included the CVH score as the dependent variable. We then conducted exploratory models for continuous measures of glucose, SBP, DBP, total cholesterol (as well as HDL and LDL cholesterol separately), and BMI. We also included a model stratified by sex. Models assessing diet and smoking status were not analyzed because they are not continuous (7) or binary (22) variables. All substitution models controlled for sex, age, marital status, education, and race. In addition, models were stratified by sex but not race because of power constraints. All models also included BMI as a covariate except when CVH or BMI was the dependent variable of interest. The model examining substitution effects on glucose controlled for diabetes medication usage, both BP models controlled for hypertension medication usage, and the cholesterol models accounted for the use of lipid-lowering medication. Assumptions of linearity, homoscedasticity, multicollinearity, outlying residuals, and normality of residuals were assessed for all models, and all deviations are noted in the results section. All analyses were conducted using SPSS version 25 (IBM Corp., Armonk, NY). Statistical significance was determined at P < 0.05.


Participant Characteristics

Study demographics are displayed by quintiles of physical activity in Table 1. In brief, the mean age of participants was 68 yr, 54% were female, and on average participants were overweight (mean BMI, 28.8 kg·m−2). Most participants were White (39%), followed by Black (27%), Hispanic (24%), and Chinese (10%). Accelerometer mean wear time was 1329 min. After accounting for non–wear time (111 min), there was 499 min of sedentary time, 415 min of LIPA, 26 min of MVPA, and 388 min of nighttime sleep averaged over a 24-h period, with an overall mean CVH score of 5.91. Participants in the highest physical activity quintile were generally younger, more likely to be female, and had higher CVH scores.

TABLE 1 - Descriptive characteristics of the study population (N = 1718).
Physical Activity Quintile (min·d−1)
Participant Characteristics (N = 1718)
≤319.83 319.84–403.33 403.34–472.67 472.68–555.67 ≥555.68
Age (yr) 72.2 (71.2–73.2) 70.0 (69.1–70.8) 68.9 (68.0–69.9) 66.7 (65.8–67.6) 63.9 (63.2–64.7)
Married 188 (57.1) 200 (58.3) 206 (59.0) 218 (62.8) 222 (63.4)
College education 161 (48.9) 141 (41.1) 144 (41.3) 126 (36.3) 124 (35.4)
 White 147 (44.7) 145 (42.3) 133 (38.1) 125 (36.0) 113 (32.3)
 Chinese American 27 (8.2) 31 (9.0) 42 (12.0) 38 (11.0) 42 (12.0)
 Black 100 (30.4) 89 (26.0) 98 (28.0) 90 (25.9) 91 (26.0)
 Hispanic 55 (16.7) 78 (22.7) 76 (21.8) 94 (27.1) 104 (29.7)
Male 186 (56.5) 156 (45.5) 156 (44.7) 155 (44.7) 131 (37.4)
Diabetes medication use 71 (21.6) 48 (14.0) 55 (15.8) 51 (14.7) 33 (9.4)
Hypertension medication use 216 (65.7) 193 (56.3) 187 (53.6) 175 (50.4) 143 (40.9)
Lipid-lowering medication use 151 (45.9) 146 (42.6) 122 (35.0) 115 (33.1) 106 (30.3)
Activity parameters a
 Sedentary time 634.9 (623.6–646.3) 563.5 (555.7–571.3) 510.4 (502.7–518.0) 444.7 (437.5–451.9) 351.9 (344.0–359.9)
 LIPA 250.0 (244.8–255.2) 350.3 (347.8–352.8) 416.7 (414.2–419.1) 478.7 (475.6–481.7) 570.2 (563.7–576.6)
 MVPA 6.5 (5.7–7.3) 12.7 (11.4–13.9) 19.8 (18.2–21.3) 32.7 (30.6–34.89) 57.0 (52.8–61.2)
 Sleep 409.5 (399.7–419.2) 403.1 (395.1–411.2) 382.4 (374.1–390.7) 384.9 (377.2–392.6) 362.9 (355.3–370.4)
BMI (kg·m−2) 29.3 (28.6–29.9) 29.6 (29.0–30.1) 28.7 (28.2–29.3) 28.5 (27.9–29.0) 28.1 (27.5–28.7)
SBP (mm Hg) 123.4 (121.2–125.5) 123.1 (121.1–125.0) 123.5 (121.3–125.7) 122.6 (120.4–124.8) 119.2 (117.3–121.2)
DBP (mm Hg) 67.4 (66.7–68.5) 68.0 (67.0–69.0) 68.1 (67.0–69.1) 68.8 (67.7–69.9) 69.0 (68.0–70.1)
Cholesterol (mg·dL−1) 174.6 (170.4–178.8) 184.6 (180.6–188.5) 185.5 (181.8–189.3) 184.4 (180.8–188.0) 190.0 (186.2–193.9)
Glucose (mg·dL−1) 105.2 (101.7–108.6) 99.7 (97.1–102.3) 101.6 (99.0–104.2) 101.4 (98.4–104.4) 99.2 (96.3–102.1)
CVH score 5.36 (5.11–5.60) 5.68 (5.45–5.91) 5.90 (5.66–6.14) 6.15 (5.91–6.38) 6.41 (6.16–6.65)
Variables are presented as n (95% CI) or mean (%).
aActivity parameters in minutes per day.

Partition Models

Results from the partition models are displayed in Table 2. When all temporal variables were included, LIPA, MVPA, and sleep were each significantly and positively associated with CVH score (model 1). These effects were maintained when sedentary behavior was removed (model 2). There was no significant association between sedentary time and CVH score when LIPA was removed (model 3).

TABLE 2 - Partition models for associations of sleep, sedentary behavior, and physical activity on CVH.
Activity Parameter Model 1 Model 2 Model 3
Sedentary 0.001 ± 0.002 Dropped −0.001 ± 0.002
LIPA 0.002 ± 0.002* 0.001 ± 0.002** Dropped
MVPA 0.017 ± 0.004** 0.016 ± 0.004** 0.016 ± 0.004**
Sleep 0.003 ± 0.002** 0.003 ± 0.002** 0.002 ± 0.002**
Values are presented as β ± 95% CI, which represent the effect of a 1-min increase in each activity variable on the CVH score.
All models control for sex, age, marital status, education, and race.
*P ≤ 0.05.
**P ≤ 0.01.
Model 1, all temporal variables included; Model 2, all temporal variables minus sedentary behavior included; Model 3, all temporal variables minus LIPA included.

Isotemporal Substitution Models

Results from the isotemporal substitution models are provided in Table 3 and visually represented as a forest plot in Figure 1.

TABLE 3 - Isotemporal substitution models for associations of sleep, sedentary behavior, and physical activity with CVH.a,b
Activity Parameter CVH Score Glucose (mg·dL−1) c SBP (mm Hg) d DBP (mm Hg) d Total Cholesterol (mg·dL−1) e BMI (kg·m−2)
Sedentary Dropped Dropped Dropped Dropped Dropped Dropped
LIPA 0.039 ± 0.033* −0.091 ± 0.363 0.231 ± 0.290 0.147 ± 0.143* 0.534 ± 0.502* −0.134 ± 0.080**
MVPA 0.485 ± 0.127** 0.279 ± 1.419 −1.205 ± 1.133* −0.858 ± 0.557** −0.049 ± 1.929 −1.179 ± 0.310**
Sleep 0.077 ± 0.045** −0.111 ± 0.506 −0.182 ± 0.404 −0.077 ± 0.198 0.493 ± 0.694 −0.262 ± 0.112**
*P ≤ 0.05.
**P ≤ 0.01.
aβ ± 95% CI represents the effect of substituting 30 min of sedentary time for LIPA, MVPA, or sleep.
bAll models are adjusted for sex, age, marital status, education, race, and total daily time on monitor. All models except for those predicting the CVH score and BMI control for BMI.
cAdditionally adjusted for diabetes medication usage.
dAdditionally adjusted for hypertension medication.
eAdditionally adjusted for lipid-lowering medication.

Forest plot depicting the effect of substituting 30 min of sedentary time for LIPA, MVPA, or sleep. *Values represent β and 95% CI.

CVH score

Substituting 30 min of sedentary time to LIPA, MVPA, and sleep was associated with a 0.039-point (95% confidence interval [CI] = 0.033), a 0.485-point (95% CI = 0.127), and a 0.077-point (95% CI = 0.045) increase in CVH score, respectively.


No significant associations were observed between any substitution model and glucose.


Substituting 30 min of sedentary time to MVPA was associated with a 1.205-mm Hg (95% CI = 1.133) decrease in SBP.


Substituting 30 min of sedentary time to LIPA and MVPA was associated with a 0.147-mm Hg (95% CI = 0.143) increase in DBP and a 0.858-mm Hg (95% CI = 0.557) decrease in DBP, respectively.

Total cholesterol

Substituting 30 min of sedentary time to LIPA was associated with a 0.534-mg·dL−1 (95% CI = 0.502) increase in total cholesterol. No significant associations were observed when substituting sedentary time for MVPA or sleep.


Substituting 30 min of sedentary time to LIPA, MVPA, and sleep was associated with a 0.134-kg·m−2 (95% CI = 0.080), a 1.179-kg·m−2 (95% CI = 0.310), and a 0.262-kg·m−2 (95% CI = 0.112) decrease in BMI, respectively.

Stratification by sex

In women, substituting 30 min of sedentary time to LIPA, MVPA, and sleep were all associated with an increase in CVH score and reduction in BMI. Substituting sedentary time to MVPA and sleep was associated with a decrease in DBP and increase in total cholesterol, respectively. No significant associations were observed between any substitution model and glucose or SBP (see Table, Supplemental Digital Content 1, Isotemporal substitution models for associations of sleep, sedentary behavior, and physical activity with cardiovascular health in women,

In men, substituting 30 min of sedentary time to MVPA and sleep was associated with an increase in CVH score and a reduction in BMI, although the relationship between these variables and LIPA was nonsignificant. Substituting sedentary time to LIPA was associated with an increase in DBP, whereas substitution to MVPA was associated with a decrease in DBP. No significant associations were observed between any substitution model and glucose, SBP, or total cholesterol (see Table, Supplemental Digital Content 2, Isotemporal substitution models for associations of sleep, sedentary behavior, and physical activity with cardiovascular health in men,

Stratification of total cholesterol

When stratifying total cholesterol into LDL and HDL cholesterol, no significant associations were observed between any substitution model and LDL cholesterol. However, significant increases in HDL cholesterol were observed when substituting 30 min of sedentary time to LIPA and MVPA, but not sleep (see Table, Supplemental Digital Content 3, Isotemporal substitution models for associations of sleep, sedentary behavior, and physical activity with LDL and HLD cholesterol, All stratified models are exploratory in nature. The interpretation of these models should be done with caution given the decreased sample size necessary to conduct these analyses.

When 14 cases with outlying residuals were censored, the effect for total cholesterol remained in the positive direction (0.434 mg·dL−1 higher) but was no longer significant (P = 0.08). There were no significant substitution effects when glucose was included as a dependent variable, and effects did not meaningfully change when individuals with diabetes were removed from the analysis. There were outlying residuals (i.e., >3 SD) present in the models for SBP, DBP, BMI, and glucose. Sensitivity analyses censoring outlying cases produced results that did not meaningfully differ from those presented.


Our study estimates that substituting 30 min of sedentary time for LIPA, MVPA, or sleep using isotemporal substitution modeling was associated with an improvement in overall CVH, assessed using metrics included in the AHA LS7 score, using data from the MESA cohort. Moreover, substituting sedentary time for LIPA, MVPA, or sleep was associated with reductions in BMI, whereas substituting sedentary time for MVPA was also associated with lower SBP and DBP. However, substituting sedentary time for sleep, LIPA, or MVPA was not associated with significant improvements in total cholesterol or blood glucose.

Much of the literature has focused on the benefits of MVPA on cardiovascular biomarkers and all-cause mortality (23,24). Our results were similarly robust, showing the most favorable outcomes when substituting sedentary time for MVPA. In addition, we uncovered an increase in the CVH score when substituting sedentary time for LIPA, which highlights the clinical and public health value of LIPA. When discussing MVPA in a clinical setting, patients often believe they must engage in more structured exercise programs, which can be difficult to achieve given common barriers such as lack of time, motivation, and cost (25). Importantly, both LIPA and MVPA can be achieved during routine activities of daily living, including walking and engaging in housework (26,27), which are more convenient than traditional exercise. Moreover, LIPA can be accumulated in large volumes with little perceived effort, offsetting the negative health effects of sedentary behavior (28), while still achieving important health benefits including reductions in CVD, cardiovascular, and all-cause mortality (26). U.S. adults spend roughly 30% of their time in LIPA compared with 2% of their time in MVPA over a 24-h time period; thus, there is more potential public health benefit to be gained from LIPA (29).

More favorable SBP, cholesterol, and BMI were also observed when substituting sedentary time with either LIPA or MVPA, consistent with the literature on the benefits of physical activity on cardiometabolic risk factors (30–32). However, we did not find a significant benefit on blood glucose in our model. A similar study using isotemporal substitution modeling from Whitaker et al. (33) also reported improvements in a composite cardiovascular risk score when substituting sedentary time with LIPA and MVPA, but no significant reduction in blood glucose. Results using isotemporal substitution have been mixed, with some studies showing no reduction in blood glucose when substituting sedentary time with physical activity (34,35), whereas others have shown a significant reduction (8,36). Differences in results may be due to the characteristics of the cohorts used in each study. Participants were generally younger and healthier, with less diabetes, in the studies that demonstrated a reduction in blood glucose with substitution, compared with the studies that demonstrated no reduction in blood glucose. Thus, the potential benefit of substituting sedentary time with physical activity on blood glucose may have been offset by the high prevalence of other cardiometabolic risk factors present in this older cohort.

We found an unexpected increase in DBP and total cholesterol when substituting sedentary time with LIPA. The relationship between physical activity and BP is complex, dependent on frequency, duration, type, and intensity of activity, which we may not have fully captured in our models. This is further acknowledged by the 2018 Physical Activity Guidelines for Americans, which states “no conclusions can be made regarding the influence of intensity on the BP response to physical activity (26).” Given previous data that suggest physical activity increases HDL cholesterol (33,37), we stratified total cholesterol into HDL and LDL cholesterol and found that substituting sedentary time for both LIPA and MVPA resulted in increased HDL cholesterol, verifying our suspicion that the increase in total cholesterol was driven by increased HDL cholesterol.

Sleep is an important and necessary determinant of health and well-being, with evidence to suggest that both short and long sleep duration are linked to an increased risk of several cardiometabolic risk factors, cardiovascular disease, and all-cause mortality (38,39). However, few studies have evaluated sleep in relation to sedentary behavior or physical activity. Our results are consistent with the literature on the importance of sleep in improving CVH.

A significant reduction in BMI and improvement in CVH was observed when substituting sedentary time for sleep, although we did not observe a similar benefit in SBP, DBP, total cholesterol, or blood glucose was observed. Our results are largely consistent with a recent study that used isotemporal substitution modeling in the Woman’s Health Initiative (40). Full et al. (40) found that substituting 91 min of sedentary time with an equivalent amount of sleep was associated with a lower waist circumference. However, similar substitutions revealed no significant benefit in insulin level, HOMA-IR, cholesterol, triglycerides, or Reynolds Risk Score, a validated tool used to predict the risk of heart attack, stroke, angioplasty, coronary artery bypass surgery, or death related to heart disease in women (41). Differences in substitution time or differences in gender may explain why we observed a significant improvement in our CVH score compared with this study.

The strengths of this study include the use of objectively assessed habitual sleep and physical activity data and the rigorous assessments of the outcomes of interest. Furthermore, this represents one of the largest and earliest reports to use partition and isotemporal substitution modeling to examine the effect of substituting sedentary time with sleep, LIPA, or MVPA on overall and individual metrics of CVH in a population-based sample of middle-age and older U.S. adults at risk for CVD. This study is also novel because it uniquely examines physical activity and sleep concomitantly on its relationship with CVH, whereas previous studies have focused on these factors separately.

This study has limitations that we acknowledge. First, isotemporal substitution is a statistical model that does not necessarily reflect the effects of real-time substitution. Second, the MESA Sleep Ancillary Study is cross sectional, so we were unable to examine associations of change in sleep and PA with changes in CVH over time within individuals. Third, dietary intake, which was included as a metric of CVH, was self-reported and is therefore prone to measurement error. Fourth, although we were able to adjust our analyses for several key confounders, we cannot rule out the presence of residual confounding. Fifth, the Actiwatch accelerometer was used to capture physical activity in the MESA Sleep Ancillary study, whereas the ActiGraph and ActivPAL devices are more common in physical activity research. There are limited published cut points for the Actiwatch, requiring the use of cut points calibrated to a later device (Motionwatch 9, CamNtech, Cambridge, UK). However, our resulting activity data align well with other published accelerometer data (42); thus, we believe that the objectively measured physical activity information captured from the Actiwatch proves useful insight into the potential effect of substituting low-active states (a common proxy for sedentary time) for higher active states (i.e., LIPA and MVPA) and sleep. Sixth, our sleep variable accounted for sleep duration but not measures of sleep architecture or fragmentation. There are data that increasing N3 (slow wave) sleep is associated with favorable cardiometabolic outcomes, independent of sleep duration (43). Thus, further studies assessing not only sleep duration but also quality and character are warranted to fully understand these complex relationships. Seventh, daytime napping may have occurred, which would have been captured as sedentary behavior and not sleep time. Finally, we did not have sufficient power to fully examine differences by race/ethnicity.

In conclusion, our study suggests that LIPA, MVPA, and sleep were all associated with more favorable CVH, and that substituting 30 min of sedentary time with an equivalent amount of sleep, LIPA, or MVPA was associated with improvements in overall CVH and several key cardiometabolic risk factors. These findings underscore the importance of lifestyle modifications in improving CVH. Further studies targeted at elucidating the complex relationships between sleep, sedentary behavior, and physical activity are warranted to inform personalized treatment plans on an individual level for those at risk of CVD.

The authors thank the other investigators, the staff, and the participants of the MESA study for their valuable contributions. A full list of participating MESA investigators and institutions can be found at Polysomnography, actigraphy, and sleep questionnaire data can be accessed from the MESA section of the National Sleep Research Resource Web site:

This research was supported by contracts HHSN268201500003I, N01-HC-95159, N01-HC-95160, N01-HC-95161, N01-HC-95162, N01-HC-95163, N01-HC-95164, N01-HC-95165, N01-HC-95166, N01-HC-95167, N01-HC-95168, and N01-HC-95169 from the National Heart, Lung, and Blood Institute, and by grants UL1-TR-000040, UL1-TR-001079, and UL1-TR-001420 from the National Center for Advancing Translational Sciences. The MESA Sleep Study was supported by NHLBI grant HL56984. S.R. was partially supported by R35HL135818.

Dr. Abdalla receives support through 18AMFDP34380732 from the American Heart Association and from the NIH/NHLBI (K23 HL141682-01A1).

The funders had no role of in the design and conduct of the study; in the collection, analysis, and interpretation of the data; and in the preparation, review, or approval of the manuscript.

The results of the present study do not constitute endorsement by the American College of Sports Medicine. The results of the study are presented clearly, honestly, and without falsification, or inappropriate data manipulation.


1. Kwok CS, Kontopantelis E, Kuligowski G, et al. Self-reported sleep duration and quality and cardiovascular disease and mortality: a dose-response meta-analysis. J Am Heart Assoc. 2018;7(15):e008552.
2. Katzmarzyk PT. Physical activity, sedentary behavior, and health: paradigm paralysis or paradigm shift? Diabetes. 2010;59(11):2717–25.
3. Piercy KL, Troiano RP, Ballard RM, et al. The physical activity guidelines for Americans. JAMA. 2018;320(19):2020–8.
4. St-Onge MP, Grandner MA, Brown D, et al. Sleep duration and quality: impact on lifestyle behaviors and cardiometabolic health: a scientific statement from the American Heart Association. Circulation. 2016;134(18):e367–86.
5. Young DR, Hivert MF, Alhassan S, et al. Sedentary behavior and cardiovascular morbidity and mortality: a science advisory from the American Heart Association. Circulation. 2016;134(13):e262–79.
6. Rosenberger ME, Fulton JE, Buman MP, et al. The 24-hour activity cycle: a new paradigm for physical activity. Med Sci Sports Exerc. 2019;51(3):454–64.
7. Mekary RA, Willett WC, Hu FB, Ding EL. Isotemporal substitution paradigm for physical activity epidemiology and weight change. Am J Epidemiol. 2009;170(4):519–27.
8. Buman MP, Winkler EA, Kurka JM, et al. Reallocating time to sleep, sedentary behaviors, or active behaviors: associations with cardiovascular disease risk biomarkers, NHANES 2005–2006. Am J Epidemiol. 2014;179(3):323–34.
9. Diaz KM, Howard VJ, Hutto B, et al. Patterns of sedentary behavior in US middle-age and older adults: the REGARDS study. Med Sci Sports Exerc. 2016;48(3):430–8.
10. Benjamin EJ, Virani SS, Callaway CW, et al. Heart disease and stroke statistics—2018 update: a report from the American Heart Association. Circulation. 2018;137(12):e67–492.
11. Lloyd-Jones DM, Hong Y, Labarthe D, et al. Defining and setting national goals for cardiovascular health promotion and disease reduction: the American Heart Association’s strategic impact goal through 2020 and beyond. Circulation. 2010;121(4):586–613.
12. Bild DE, Bluemke DA, Burke GL, et al. Multi-ethnic study of atherosclerosis: objectives and design. Am J Epidemiol. 2002;156(9):871–81.
13. Chen X, Wang R, Zee P, et al. Racial/ethnic differences in sleep disturbances: the Multi-Ethnic Study of Atherosclerosis (MESA). Sleep. 2015;38(6):877–88.
14. Hart TL, Swartz AM, Cashin SE, Strath SJ. How many days of monitoring predict physical activity and sedentary behaviour in older adults? Int J Behav Nutr Phys Act. 2011;8(1):62.
15. Zhang GQ, Cui L, Mueller R, et al. The National Sleep Research Resource: towards a sleep data commons. J Am Med Inform Assoc. 2018;25(10):1351–8.
16. Oakley N. Validation with Polysomnography of the Sleepwatch Sleep/Wake Scoring Algorithm Used by the Actiwatch Activity Monitoring System. Mini Mitter, Cambridge Neurotechnology: Bend; 1997.
17. Landry GJ, Falck RS, Beets MW, Liu-Ambrose T. Measuring physical activity in older adults: calibrating cut-points for the MotionWatch 8©. Front Aging Neurosci. 2015;7:165.
18. Mayer-Davis EJ, Vitolins MZ, Carmichael SL, et al. Validity and reproducibility of a food frequency interview in a multi-cultural epidemiology study. Ann Epidemiol. 1999;9(5):314–24.
19. Kulshreshtha A, Vaccarino V, Judd SE, et al. Life’s simple 7 and risk of incident stroke: the reasons for geographic and racial differences in stroke study. Stroke. 2013;44(7):1909–14.
20. Ogunmoroti O, Oni E, Michos ED, et al. Life’s Simple 7 and Incident Heart Failure: the multi-ethnic study of atherosclerosis. J Am Heart Assoc. 2017;6(6):e005180.
21. Lloyd-Jones DM. Improving the cardiovascular health of the US population. JAMA. 2012;307(12):1314–6.
22. Mekary RA, Lucas M, Pan A, et al. Isotemporal substitution analysis for physical activity, television watching, and risk of depression. Am J Epidemiol. 2013;178(3):474–83.
23. Loprinzi PD. Dose-response association of moderate-to-vigorous physical activity with cardiovascular biomarkers and all-cause mortality: considerations by individual sports, exercise and recreational physical activities. Prev Med. 2015;81:73–7.
24. Loprinzi PD, Loenneke JP, Ahmed HM, Blaha MJ. Joint effects of objectively-measured sedentary time and physical activity on all-cause mortality. Prev Med. 2016;90:47–51.
25. Hoare E, Stavreski B, Jennings GL, Kingwell BA. Exploring motivation and barriers to physical activity among active and inactive australian adults. Sports (Basel). 2017;5(3):47.
26. Services USDoHaH. 2018 Physical Activity Guidelines Advisory Committee. Physical Activity Guidelines Advisory Committee Scientific Report; 2018.
27. Murphy MH, Donnelly P, Breslin G, Shibli S, Nevill AM. Does doing housework keep you healthy? The contribution of domestic physical activity to meeting current recommendations for health. BMC Public Health. 2013;13:966.
28. Owen N, Healy GN, Matthews CE, Dunstan DW. Too much sitting: the population health science of sedentary behavior. Exerc Sport Sci Rev. 2010;38(3):105–13.
29. Chastin SF, Palarea-Albaladejo J, Dontje ML, Skelton DA. Combined effects of time spent in physical activity, sedentary behaviors and sleep on obesity and cardio-metabolic health markers: a novel compositional data analysis approach. PLoS One. 2015;10(10):e0139984.
30. Liu X, Zhang D, Liu Y, et al. Dose-response association between physical activity and incident hypertension: a systematic review and meta-analysis of cohort studies. Hypertension. 2017;69(5):813–20.
31. Glazer NL, Lyass A, Esliger DW, et al. Sustained and shorter bouts of physical activity are related to cardiovascular health. Med Sci Sports Exerc. 2013;45(1):109–15.
32. Fan JX, Brown BB, Hanson H, Kowaleski-Jones L, Smith KR, Zick CD. Moderate to vigorous physical activity and weight outcomes: does every minute count? Am J Health Promot. 2013;28(1):41–9.
33. Whitaker KM, Pettee Gabriel K, Buman MP, et al. Associations of accelerometer-measured sedentary time and physical activity with prospectively assessed cardiometabolic risk factors: the CARDIA study. J Am Heart Assoc. 2019;8(1):e010212.
34. Knaeps S, De Baere S, Bourgois J, Mertens E, Charlier R, Lefevre J. Substituting sedentary time with light and moderate to vigorous physical activity is associated with better cardiometabolic health. J Phys Act Health. 2018;15(3):197–203.
35. Falconer CL, Page AS, Andrews RC, Cooper AR. The potential impact of displacing sedentary time in adults with type 2 diabetes. Med Sci Sports Exerc. 2015;47(10):2070–5.
36. Ekblom-Bak E, Ekblom O, Bolam KA, Ekblom B, Bergstrom G, Borjesson M. SCAPIS pilot study: sitness, fitness and fatness—is sedentary time substitution by physical activity equally important for everyone’s markers of glucose regulation? J Phys Act Health. 2016;13(7):697–703.
37. Kodama S, Tanaka S, Saito K, et al. Effect of aerobic exercise training on serum levels of high-density lipoprotein cholesterol: a meta-analysis. Arch Intern Med. 2007;167(10):999–1008.
38. Buxton OM, Marcelli E. Short and long sleep are positively associated with obesity, diabetes, hypertension, and cardiovascular disease among adults in the United States. Soc Sci Med. 2010;71(5):1027–36.
39. Cappuccio FP, D’Elia L, Strazzullo P, Miller MA. Sleep duration and all-cause mortality: a systematic review and meta-analysis of prospective studies. Sleep. 2010;33(5):585–92.
40. Full KM, Gallo LC, Malhotra A, et al. Modeling the cardiometabolic benefits of sleep in older women: exploring the 24-hour day. Sleep. 2020;43(1):zsz205.
41. Ridker PM, Buring JE, Rifai N, Cook NR. Development and validation of improved algorithms for the assessment of global cardiovascular risk in women: the Reynolds risk score. JAMA. 2007;297(6):611–9.
42. Hansen BH, Kolle E, Dyrstad SM, Holme I, Anderssen SA. Accelerometer-determined physical activity in adults and older people. Med Sci Sports Exerc. 2012;44(2):266–72.
43. Javaheri S, Zhao YY, Punjabi NM, Quan SF, Gottlieb DJ, Redline S. Slow-wave sleep is associated with incident hypertension: the Sleep Heart Health Study. Sleep. 2018;41(1):zsx179.


Supplemental Digital Content

Copyright © 2020 by the American College of Sports Medicine