Abstract
Background
Isotemporal substitution examines the effect on health outcomes of replacing sedentary time with light‐intensity physical activity or moderate‐to‐vigorous intensity physical activity; however, existing studies are limited by cross‐sectional study designs.
Methods and Results
Participants were 1922 adults from the CARDIA (Coronary Artery Risk Development in Young Adults) study. Linear regression examined the associations of sedentary, light‐intensity physical activity, and moderate‐to‐vigorous intensity physical activity at year 20 (2005–2006) with waist circumference, blood pressure, glucose, insulin, triglycerides, high‐density lipoprotein cholesterol, and a composite risk score at year 30 (2015–2016). Models then examined change in activity with change in cardiometabolic risk over the same 10‐year period. Replacing 30 min/day of sedentary time with 30 min/day of light‐intensity physical activity at year 20 was associated with a lower composite risk score (−0.01 SD [95% CI, −0.02, −0.00]) at year 30, characterized by lower waist circumference (0.15 cm [95% CI, −0.27, 0.02]), insulin (0.20 μU/mL [95% CI, −0.35, −0.04]), and higher high‐density lipoprotein cholesterol (0.20 mg/dL [95% CI, 0.00, 0.40]; all P<0.05). An increase of 30 min/day in MVPA from year 20 to year 30, when replacing an equivalent increase in sedentary time, was associated with a decrease in the composite risk score (−0.08 [95% CI, −0.13, −0.04]) over the same 10 years, characterized by a decrease in waist circumference (1.52 cm [95% CI, −2.21, −0.84]), insulin (−1.13 μU/mL [95% CI, −1.95, −0.31]), triglycerides (−6.92 mg/dL [95% CI, −11.69, −2.15]), and an increase in high‐density lipoprotein cholesterol (1.59 mg/dL [95% CI, 0.45, 2.73]; all P<0.05).
Conclusions
Replacement of sedentary time with light‐intensity physical activity or moderate‐to‐vigorous intensity physical activity is associated with improved cardiometabolic health 10 years later.
Keywords: cardiometabolic risk, epidemiology, isotemporal substitution, physical activity, sedentary time
Subject Categories: Epidemiology, Exercise, Cardiovascular Disease
Clinical Perspective
What Is New?
Few large, epidemiological studies have repeated measures of accelerometer‐measured sedentary time, light‐intensity physical activity, and moderate‐to‐vigorous intensity physical activity (MVPA), which allows for more‐accurate quantification of the rest‐activity behavioral profile compared with self‐reported questionnaires.
Using isotemporal substitution, this study examined longitudinal associations of accelerometer‐measured sedentary time, light‐intensity physical activity, and moderate‐to‐vigorous intensity physical activity in 2005–2006 with cardiometabolic risk factors assessed in 2015–2016, as well as change in activity with change in cardiometabolic risk factors over the same 10‐year period.
What Are the Clinical Implications?
Study findings indicate that replacement of sedentary time with light‐intensity physical activity or moderate‐to‐vigorous intensity physical activity is associated with improved cardiometabolic health 10 years later.
In addition to promoting moderate‐to‐vigorous intensity physical activity, clinicians and public health practitioners should consider encouraging patients to replace sedentary time with light‐intensity physical activity for cardiometabolic benefit, which may be a more‐feasible lifestyle modification strategy for highly sedentary adults.
Introduction
High prevalence of sedentary behaviors in today's society is a growing public health concern. Excess sedentary time is associated with many markers of cardiometabolic risk, including larger waist circumference and higher blood pressure, glucose, insulin, triglycerides, and lower high‐density lipoprotein cholesterol (HDL‐C).1, 2, 3, 4, 5 Evidence suggests that meeting guidelines for physical activity may not be sufficient for chronic disease prevention if accompanied by high levels of sedentary time.6 Given that the number of hours in a day is fixed, participation in 1 activity means that a different activity is displaced, and each activity may have differential associations with cardiometabolic risk factors depending on the type of activity displaced. For example, replacing sedentary time with light‐intensity physical activity (LPA) may have heterogeneous effects on cardiometabolic risk factors as compared with replacing sedentary time with moderate‐to‐vigorous intensity physical activity (MVPA). It is therefore important to assess the associations of substituting sedentary time with physical activity of various intensity levels to better inform health promotion efforts.
Isotemporal substitution has been used to examine the associations of alternating allocations of time in 1 behavior with another while holding total time constant.7 This approach was originally developed for nutritional epidemiology and has recently been used in the physical activity literature to examine associations when replacing sedentary time with physical activities of varying types and intensities across a variety of outcomes, including self‐rated health,8, 9, 10 cardiometabolic risk factors,7, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 and all‐cause mortality.21, 22, 23, 24 Although there has been increasing use of the isotemporal substitution method, these studies are largely limited by their cross‐sectional study designs. A small number of studies haves used isotemporal substitution to examine the associations of physical activity and sedentary behaviors with prospective measures of health outcomes. However, to our knowledge, the isotemporal substitution method has not been used to examine change in objectively measured physical activity with concurrent change in cardiometabolic risk factors.
The CARDIA (Coronary Artery Risk Development in Young Adults) Fitness and Activity ancillary studies provide an opportunity to fill this existing gap in knowledge, as a subset of participants wore an ActiGraph accelerometer for 7 consecutive days in early (ages 38–50 years) and late (ages 48–60 years) midlife, corresponding to the year 20 and 30 CARDIA exams. The objective of this study was to examine the associations of time spent in accelerometer‐measured activity (sedentary, LPA, and MVPA) at year 20 with a composite cardiometabolic risk score (primary outcome) and each individual cardiometabolic risk factor (secondary outcomes: waist circumference, blood pressure, glucose, insulin, triglycerides, and HDL‐C) at year 30, using the isotemporal substitution method. We then examined associations between change in time spent in accelerometer‐measured activity from year 20 to year 30, with change in the composite risk score and individual risk factors from year 20 to year 30. We hypothesized that replacing sedentary time with LPA or MVPA at year 20 will result in more‐favorable cardiometabolic risk factors at year 30, and that an increase in LPA or MVPA from year 20 to year 30, when substituted for sedentary time, will be associated with more‐favorable 10‐year changes in cardiometabolic risk factors.
Methods
Requests to access the data set, analytical methods, and study materials may be sent to the CARDIA Coordinating Center. Contact information can be found on the CARDIA website.25
Study Participants
The CARDIA study is an ongoing longitudinal cohort of 5115 black and white men and women, aged 18 to 30 years, who completed an in‐person clinical examination in 1985–1986 (year 0) from 1 of 4 field centers: Birmingham, Alabama; Minneapolis, MN; Chicago, IL; or Oakland, CA. Additional in‐person clinic examinations were held approximately every 2 to 5 years, including a year 20 (2005–2006) and 30 (2015–2016) exam, with 72% and 71% participant retention, respectively. Details on eligibility criteria, methods of participant selection, and follow‐up have been previously reported.26
All ambulatory participants in the core study were invited to take part in the ancillary CARDIA Fitness Study during the year 20 exam, where physical activity was measured for 7 days with the ActiGraph 7164 (N=2332). As previously described, individuals who did not take part in the year 20 CARDIA Fitness Study for medical or nonmedical reasons had a higher risk of cardiovascular events and/or all‐cause mortality as compared with those who took part in the Fitness Study.27 At year 30, all ambulatory participants were again invited to take part in the CARDIA Activity Study, where physical activity was measured using the ActiGraph wGT3X‐BT (N=1397). For the present study, participants were excluded if they had unreasonably high values for MVPA (>14 h/day, N=3, all 7164 devices), or were missing data on the cardiometabolic risk factors of interest at the year 20 (N=25) or year 30 examinations (N=318), or were missing data on potential confounders of interest from year 20 (N=64), resulting in a final sample of 1922 for the analyses examining associations of year 20 physical activity with year 30 cardiometabolic outcomes. Participants who were excluded from analyses because of missing data on cardiometabolic risk factors or other covariates (N=410) were more likely to be black, with less years of education, without health insurance, reporting medication use, and were also more likely to be current smokers as compared with those included in the analyses (see Table 1). Of the 1922 participants included in analyses, 913 also had accelerometer data from the year 30 examination and were included in the analyses examining change in activity and cardiometabolic risk from years 20 to 30. The smaller sample with accelerometer data at year 30 was primarily the result of a shortened data collection period (because of funding mechanism) that began approximately midway through the year 30 exam. The institutional review board at each center approved all study protocols for the primary CARDIA exam, as well as the CARDIA Fitness and Activity studies. The institutional review boards at the University of Alabama at Birmingham, Northwestern University, University of Minnesota, and Kaiser Permanente approved the study. Written informed consent was obtained at each exam, separately for the primary and ancillary studies.
Table 1.
Characteristics of CARDIA Participants Included and Excluded From Analyses
| Characteristics | Included Participants (N=1922) | Excluded Participants (N=410)* | P Value† |
|---|---|---|---|
| Age, y | 45.3±3.5 | 45.2±3.8 | 0.509 |
| Female, n (%) | 1123 (58.4) | 221 (54.0) | 0.102 |
| White, n (%) | 1143 (59.5) | 215 (52.4) | 0.009 |
| Education, y | 15.3±2.6 | 14.6±2.6 | <0.001 |
| Unemployment, n (%) | 206 (10.7) | 56 (13.8) | 0.072 |
| Health insurance, n (%) | 1716 (89.3) | 333 (81.4) | <0.001 |
| Medication use, n (%)‡ | 375 (19.5) | 106 (25.9) | 0.004 |
| Smoking status, n (%) | <0.001 | ||
| Current | 282 (14.7) | 92 (23.4) | |
| Former | 411 (21.4) | 17 (18.1) | |
| Never | 1229 (63.9) | 230 (58.5) | |
| Alcohol consumption, mL/day | 2.4 (14.5) | 0.0 (13.6) | 0.117 |
| Body mass index, kg/m2 | 28.8±6.4 | 29.6±9.0 | 0.101 |
| Accelerometer measured activity, min/day§ | |||
| Total wear time | 886.5±86.9 | 878.4±100.7 | 0.093 |
| Sedentary | 490.1±102.4 | 472.1±112.8 | 0.003 |
| LPA | 360.9±86.3 | 360.5±89.3 | 0.938 |
| MVPA | 29.8 (29.8) | 30.7 (33.7) | 0.916 |
Data presented as mean±SD or median (interquartile range), unless otherwise specified. CARDIA indicates Coronary Artery Risk Development in Young Adults; LPA, light‐intensity physical activity; MVPA, moderate‐to‐vigorous intensity physical activity.
Participants excluded because of unreasonably high values for MVPA (N=3), or if they were missing data on the cardiometabolic risk factors of interest at the year 20 (N=25) or year 30 examinations (N=318), or were missing data on potential confounders of interest from year 20 (N=64).
P‐value testing for differences among those who were included vs excluded from analyses using independent samples t tests, Wilcoxon–Mann–Whitney tests, or chi‐square tests, as appropriate.
Self‐reported medication use for blood pressure, cholesterol, or diabetes mellitus.
Freedson cut point thresholds defined sedentary time in counts/minute as <100, LPA as 100–1951, and MVPA as ≥1952.
Physical Activity and Sedentary Behavior
The ActiGraph 7164 (Year 20) and ActiGraph wGT3X‐BT (Year 30) were initialized to start collecting data at 12:00 am on the day following the in‐person examination and were worn on an elastic belt during waking hours. The 7164 model was initialized to collect data in 60‐second epochs, and for the wGT3X‐BT devices, raw triaxial data were sampled at 40 Hz. Data from both monitors were processed using ActiLife6 software, and raw data from the wGT3X‐BT were reintegrated to a 60‐second epoch (with the low‐frequency extension applied). Files from both devices were screened for wear time using the Troiano algorithm.28 Weekly summary physical activity and sedentary behavior estimates were averaged across days for all participants with at least 4 valid days of ≥10 hours of wear time based on criteria established for (National Health and Nutrition Examination Survey 2003–2004 and 2005–2006 cycles.28 Total and average accelerometer counts per day were calculated using summed counts detected over wear periods and time (ie, minutes) spent in different intensity levels using standardized cut‐point threshold values.28, 29, 30 Freedson cut‐point threshold values were selected given their broad use in physical activity research; sedentary time was defined as ≤100 counts per minute, LPA as 101 to 1951 counts per minute, and MVPA as 1952+ counts per minute.30 We chose to combine moderate‐intensity physical activity and vigorous‐intensity physical activity because of the low prevalence of vigorous activity detected in this cohort.
Given that different monitors were used in the 2 exams, we examined the comparability of the 7164 and wGT3X‐BT models by having a subset of participants (N=100) wear both monitors simultaneously during the year 30 examination.31 There were significant differences between monitors for sedentary, LPA, and MVPA. Therefore, to allow for data harmonization between the 2 monitors, we developed and applied a calibration factor to the wGT3X‐BT values. After calibration, no differences were observed between the 7164 and wGT3X‐BT in min/day for sedentary (513.2±93.6 versus 509.6±98.6; P=0.23), LPA (335.3±81.5 versus 338.7±81.1; P=0.22), or MVPA (33.1±24.6 versus 32.0±26.0; P=0.13). For the present analyses, we report the calibrated values for the wGT3X‐BT; no correction was applied to the 7164 values.
Cardiometabolic Risk Factors
The same protocols and procedures for all cardiometabolic risk factors were used at both the year 20 and year 30 examinations. Participants were asked to fast for at least 12 hours before the study examination and refrain from smoking or engaging in vigorous physical activities for at least 2 hours earlier. Waist circumference was measured in duplicate to the nearest 0.5 cm at the minimal abdominal girth and then averaged. Blood pressure was measured 3 times on the right arm using an automated sphygmomanometer (Omron HEM907XL) in 1‐minute intervals after the participant was seated for 5 minutes. The average of the second and third blood pressure readings were used for analysis. Serum glucose was obtained from fasting venous blood samples and assessed in the Molecular Epidemiology and Biochemistry Research Laboratory (Minneapolis, MN) with quality control using standards from the National Institute of Standards and Technology. Fasting insulin was measured using radioimmunoassay by Linco Research (St. Louis, MO), and triglycerides and HDL‐C were measured using an enzymatic assay by Northwest Lipids Research Laboratory (Seattle, WA).
Composite risk score was estimated separately at year 20 and year 30 by standardizing and summing the following variables: waist circumference, average blood pressure ([systolic+diastolic]/2), fasting blood glucose, insulin, triglycerides, negative HDL‐C, then dividing by 6 to create a z‐score, as previously reported in the literature, so that higher scores represent poorer cardiometabolic status.32, 33, 34, 35 Before standardization, log transformations were performed on glucose, insulin, and triglycerides because of skewed distributions. Standardization was done by subtracting the sample mean from the individual mean and dividing by the SD of the sample mean. Sex‐specific standardizations were performed for waist circumference and HDL‐C. Both the year 20 and 30 composite scores were standardized using the sample mean and SD of the year 20 variables to better illustrate the change in score over 10 years, and to reduce random error and improve regression estimate precision.
Covariates
Study covariates from the year 20 examination included field center and self‐reported age, sex, race (black/white), years of education completed, employment status (yes/no), health insurance for the past 2 years (yes/no), and medication use for blood pressure, cholesterol, or diabetes mellitus (yes/no). Smoking status was obtained from a tobacco‐use questionnaire that has previously been validated by a study using serum cotinine levels36 and categorized as never, former, or current. Alcohol consumption was calculated in milliliters per day using self‐reported intake of wine, beer, and hard liquor. Body mass index (BMI) was calculated using measured height and weight (kg/m2).
Statistical Analyses
Descriptive statistics were calculated and stratified by year 20 physical activity quartiles, standardized to total wear time. Quartile cut points were similar for men and women; therefore, the same cut points were used for both sexes. Means and SDs or medians and interquartile ranges were calculated for accelerometer measures and cardiometabolic risk factors at the year 20 and year 30 examinations. Differences between exam years were examined using paired samples t tests or Wilcoxon signed‐ranks test. Spearman correlations were used to examine bivariate associations for continuous variables between accelerometer measures at year 20 with cardiometabolic risk factors at year 30.
Multivariable linear regression isotemporal substitution models were used to examine how accelerometer‐measured activity at the year 20 exam (sedentary, LPA, and MVPA) was associated with cardiometabolic risk factors 10 years later, during the year 30 exam. We then repeated this process, examining how changes in accelerometer‐measured activity between the 2 exams (eg, Δsedentary=year 30 sedentary−year 20 sedentary) were associated with changes in the dependent variable, or change in cardiometabolic risk factors, from the same 10‐year period (eg, Δrisk score=year 30 risk score−year 20 risks core). Isotemporal substitution models estimated the effect of replacing time from 1 activity for an equal amount of time from another activity. This was done by entering all variables that collectively characterize total (wear) time (sedentary+LPA+MVPA or Δsedentary+ΔLPA+ΔMVPA) and the individual activity categories to the models simultaneously, while dropping the specific activity of interest from the model. These models represent the effect of replacing the dropped activity variable with the other variables in the model. Inclusion of a total time variable (or change in total time) in the model constrains time and allows for direct comparison between activities and their impact on cardiometabolic risk score. Before entry in the models, all activity categories were divided by 30, such that a unit increase in the activity represented an increase of 30 min/day within the given category, as commonly reported in the isotemporal literature.8, 11, 20, 22, 24 For example, if sedentary time is dropped from the model, then the interpretation of the beta coefficient for LPA is the effect on the cardiometabolic risk factor of substituting 30 min/day of sedentary time with 30 minutes of LPA.
All models were adjusted for year 20 demographics (center, sex, race, age, education, employment status, health insurance, self‐reported medication use for blood pressure, cholesterol, or diabetes mellitus), lifestyle behaviors (smoking status, alcohol consumption, and BMI), and the year 20 cardiometabolic risk factor of interest, to account for differences in baseline risk. In the models examining change in activity with change in cardiometabolic risk, we additionally adjusted for year 20 sedentary time and MVPA to account for differences in baseline activity. BMI was not included in the waist circumference or composite risk score model because of the high correlation with waist circumference (r=0.87). Because of skewed distributions, we examined the glucose, insulin, and triglyceride models using both original and log‐transformed data. Findings did not materially differ; therefore, to facilitate interpretation, we report the results using the original data only. All models were tested for multicollinearity (variance inflation factor <2.0).
In sensitivity analyses, we also adjusted for an overall diet‐quality score assessed at the year 20 exam.37 Results were unchanged, and because of the high level of missing data (N=205), we report findings without adjustment for diet. A second set of sensitivity analyses was also conducted where individuals with past cardiovascular disease, self‐reported at the year 20 exam, including history of myocardial infarction, angina, peripheral vascular disease, stroke, or transient ischemic attack at year 20, were excluded (N=45). Results again did not materially differ; therefore, we present results with the full study sample. In a third set of sensitivity analyses, we additionally adjusted for year 20 cardiorespiratory fitness, assessed using a maximal graded exercise treadmill test. Finally, we tested for interactions between physical activity and sex as well as physical activity and race; none were statistically significant, and therefore we only present models adjusted for sex and race. All statistical analyses were conducted using SAS software (version 9.4; SAS Institute, Inc, Cary, NC). All tests were 2‐sided, with statistical significance set at P<0.05.
Results
Participant characteristics at the year 20 exam, stratified by quartiles of accelerometer‐measured physical activity are shown in Table 2. There were significant differences across physical activity quartiles by sex, race, education, smoking status, alcohol consumption, and BMI. Percentages of women and whites decreased across quartiles of physical activity, as did years of education and BMI, whereas alcohol consumption was positively associated with activity quartiles. Cardiometabolic risk factors were generally more favorable across physical activity quartiles, with waist circumference, diastolic blood pressure, glucose, insulin, triglycerides, and the composite risk score going down and HDL‐C going up across activity groups.
Table 2.
Participant Characteristics at Year 20 by Physical Activity Quartiles, the CARDIA Study, 2005–2006 (N=1922)
| Characteristics | Quartiles of Physical Activity* | P Value† | |||
|---|---|---|---|---|---|
| Q1 ≤333 Min/Day (N=480) | Q2 334 to 391 Min/Day (N=481) | Q3 392 to 454 Min/Day (N=481) | Q4 ≥455 Min/Day (N=480) | ||
| Age, y | 45.5±3.5 | 45.3±3.7 | 45.2±3.4 | 45.2±3.5 | 0.100 |
| Female, n (%) | 250 (52.1) | 191 (39.7) | 155 (32.2) | 203 (42.3) | <0.001 |
| White, n (%) | 313 (65.2) | 277 (57.6) | 305 (63.4) | 248 (51.7) | <0.001 |
| Education, y | 15.8±2.5 | 15.5±2.5 | 15.4±2.6 | 14.5±2.4 | <0.001 |
| Unemployment, n (%) | 55 (11.5) | 48 (10.0) | 49 (10.2) | 54 (11.3) | 0.842 |
| Health insurance, n (%) | 432 (90.0) | 437 (90.9) | 433 (90.0) | 414 (86.3) | 0.094 |
| Medication use, n (%)‡ | 99 (20.6) | 96 (20.0) | 89 (18.5) | 91 (18.0) | 0.839 |
| Smoking status, n (%) | <0.001 | ||||
| Current | 58 (12.1) | 63 (13.1) | 59 (12.3) | 102 (21.3) | |
| Former | 103 (21.5) | 93 (19.3) | 119 (24.7) | 96 (20.0) | |
| Never | 319 (66.5) | 325 (67.6) | 303 (63.0) | 282 (58.5) | |
| Alcohol consumption, mL/day | 2.4 (12.4) | 2.4 (12.7) | 2.7 (14.3) | 4.9 (17.0) | 0.011 |
| Body mass index, kg/m2 | 29.5±6.7 | 29.0±6.5 | 28.3±6.3 | 28.5±6.1 | 0.009 |
| Accelerometer measured activity, min/day§ | |||||
| Total wear time | 887.6±89.8 | 895.4±86.3 | 890.5±83.7 | 872.6±86.3 | 0.005 |
| Sedentary | 597.1±68.0 | 528.2±54.2 | 468.0±45.0 | 367.2±62.4 | <0.001 |
| LPA | 264.6±44.4 | 335.2±40.2 | 385.9±44.9 | 457.7±65.5 | <0.001 |
| MVPA | 21.6 (22.6) | 27.3 (27.8) | 31.0 (28.1) | 41.4 (38.1) | <0.001 |
| Cardiometabolic risk factors | |||||
| Waist circumference, cm | 93.7±15.7 | 90.6±14.4 | 88.0±14.1 | 89.7±13.5 | <0.001 |
| Systolic blood pressure, mm Hg | 115.4±13.3 | 113.7±13.1 | 113.5±15.1 | 114.8±13.6 | 0.446 |
| Diastolic blood pressure, mm Hg | 72.2±10.5 | 70.9±10.1 | 70.5±11.0 | 70.8±10.5 | 0.038 |
| Glucose, mg/dL | 95.6 (12.7) | 95.6 (11.7) | 94.6 (11.7) | 94.6 (12.7) | 0.008 |
| Insulin, μU/mL | 8.9 (6.6) | 8.9 (5.9) | 7.6 (5.2) | 7.6 (5.2) | <0.001 |
| Triglycerides, mg/dL | 93.0 (74.5) | 84.0 (72.0) | 84.0 (59.0) | 83.0 (61.0) | <0.001 |
| HDL‐C, mg/dL | 51.2±15.4 | 54.8±16.4 | 57.4±16.7 | 55.1±16.2 | <0.001 |
| Composite risk scoreǁ | 0.14±0.70 | 0.02±0.66 | −0.09±0.66 | −0.07±0.62 | <0.001 |
Data presented as mean±SD or median (interquartile range), unless otherwise specified. CARDIA indicates Coronary Artery Risk Development in Young Adults; HDL‐C, high‐density lipoprotein cholesterol; LPA, light‐intensity physical activity; MVPA, moderate‐to‐vigorous intensity physical activity.
Accelerometer measured light‐, moderate‐, and vigorous‐intensity physical activity, standardized for total wear time.
P‐value testing for differences across quartiles of physical activity using chi‐square test, 1‐way ANOVA, or Kruskal–Wallis test, as appropriate.
Self‐reported medication use for blood pressure, cholesterol, or diabetes mellitus.
Freedson cut‐point thresholds defined sedentary time in counts/min as <100, LPA as 100 to 1951, and MVPA as ≥1952.
Composite risk score was calculated by standardizing and summing waist circumference, average blood pressure ([systolic+diastolic]/2), log glucose, log insulin, log triglycerides, and negative HDL‐C, then dividing by 6 to create a z‐score.
Accelerometer wear time did not significantly differ between the year 20 and year 30 exams (see Table 3). Sedentary min/day increased, on average, by 36.8 min/day between the 2 exams. LPA and MVPA decreased, on average, by 29.6 and 5.4 minutes, respectively (all P<0.001). There were also significant differences observed in cardiometabolic risk factors across the 2 exams (Table 4). Waist circumference, systolic blood pressure, diastolic blood pressure, glucose, insulin, and the composite risk score were significantly higher at the year 30 exam compared with the year 20 exam, indicating adverse changes over time. HDL‐C increased over time, indicating a health improvement, which has been previously described in the CARDIA.38 The increase in HDL‐C was observed in both men and women, and persisted after excluding individuals who reported cholesterol medication use at either exam (data not shown). There were no differences between the 2 exams in triglycerides.
Table 3.
Accelerometer‐Measured Total Wear Time, Sedentary, Light‐, and Moderate‐to‐Vigorous Intensity Physical Activity by Examination Year, the CARDIA Study, 2005–2016
| Accelerometer Measures, Min/Day* | Year 20 2005–2006 (N=1922) | Year 30 2015–2016 (N=913) | Year 30−20 Difference (N=913) | P Value† |
|---|---|---|---|---|
| Total wear time | 886.5±86.9 | 894.1±87.8 | 0.2±101.3 | 0.961 |
| Sedentary | 490.1±102.4 | 532.3±102.7 | 36.8±110.6 | <0.001 |
| LPA | 360.9±86.3 | 332.4±84.2 | −29.6±85.2 | <0.001 |
| MVPA | 29.8 (29.8) | 24.1 (28.8) | −5.4 (23.3) | <0.001 |
Data presented as mean±SD or median (interquartile range). CARDIA indicates Coronary Artery Risk Development in Young Adults; LPA, light‐intensity physical activity; MVPA, moderate‐to‐vigorous intensity physical activity.
Freedson cut‐point thresholds defined sedentary time in counts/min as <100, LPA as 100 to 1951, and MVPA as ≥1952.
P‐value testing for differences between year 20 and year 30 accelerometer measures using paired samples t tests or Wilcoxon signed‐ranks test.
Table 4.
Cardiometabolic Risk Factors by Examination Year, the CARDIA Study, 2005–2016 (N=1922)
| Cardiometabolic Risk Factors | Year 20 2005–2006 | Year 30 2015–2016 | Year 30−20 Difference | P Value* |
|---|---|---|---|---|
| Waist circumference, cm | 90.5±14.6 | 94.7±15.6 | 4.2±7.2 | <0.001 |
| Systolic blood pressure, mm Hg | 114.3±13.8 | 119.7±16.1 | 5.4±15.1 | <0.001 |
| Diastolic blood pressure, mm Hg | 71.1±10.6 | 73.3±10.6 | 2.2±10.3 | <0.001 |
| Glucose, mg/dL | 94.6 (12.7) | 95.0 (14.0) | 0.4 (12.6) | 0.034 |
| Insulin, μU/mL | 8.3 (5.2) | 9.7 (9.3) | 1.1 (6.6) | <0.001 |
| Triglycerides, mg/dL | 86.0 (67.0) | 86.0 (60.0) | 1.0 (47.0) | 0.409 |
| HDL‐C, mg/dL | 54.6±16.3 | 60.4±18.9 | 5.7±11.6 | <0.001 |
| Composite risk score† | −0.00±0.66 | 0.11±0.77 | 0.11±0.50 | <0.001 |
Data presented as mean±SD or median (interquartile range). CARDIA indicates Coronary Artery Risk Development in Young Adults; HDL‐C, high‐density lipoprotein cholesterol.
P‐value testing for differences between year 20 and year 30 cardiometabolic risk factors using paired samples t tests or Wilcoxon signed‐ranks test.
Composite risk score was calculated by standardizing and summing waist circumference, average blood pressure ([systolic+diastolic]/2), log glucose, log insulin, log triglycerides, and negative HDL‐C, then dividing by 6 to create a z‐score.
Correlations between the year 20 accelerometer measures and year 30 cardiometabolic risk factors are presented in Table 5. Sedentary time at year 20 was positively associated with year 30 waist circumference, glucose, insulin, triglycerides, and the composite risk score and inversely associated with HDL‐C. LPA and MVPA were inversely associated with waist circumference, average blood pressure (MVPA only), glucose, insulin, triglycerides, HDL‐C (LPA only), and the composite risk score.
Table 5.
Spearman Correlations of Accelerometer‐Measured Activity at Year 20 and Cardiometabolic Risk Factors at Year 30, the CARDIA Study, 2005–2016 (N=1922)
| Accelerometer Measured Activity | Cardiometabolic Risk Factors | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Sedentary | LPA | MVPA | Waist | BP | Glucose | Insulin | Triglycerides | HDL | |
| Accelerometer measured activity | |||||||||
| LPA | −0.608* | ||||||||
| MVPA | −0.187* | 0.157* | |||||||
| Cardiometabolic risk factors | |||||||||
| Waist | 0.073* | −0.101* | −0.131* | ||||||
| BP | 0.004 | 0.002 | −0.088* | 0.276* | |||||
| Glucose | 0.045* | −0.048* | −0.085* | 0.472* | 0.205* | ||||
| Insulin | 0.068* | −0.090* | −0.209* | 0.675* | 0.250* | 0.503* | |||
| Triglycerides | 0.064* | −0.096* | −0.066* | 0.430* | 0.186* | 0.357* | 0.466* | ||
| HDL‐C | −0.101* | 0.108* | 0.016 | −0.543* | −0.102* | −0.349* | −0.473* | −0.532* | |
| Composite score | 0.070* | −0.096* | −0.193* | 0.777* | 0.462* | 0.638* | 0.829* | 0.680* | −0.667* |
BP indicates average blood pressure; CARDIA, Coronary Artery Risk Development in Young Adults; HDL, high‐density lipoprotein cholesterol; LPA, light‐intensity physical activity; MVPA, moderate‐to‐vigorous intensity physical activity.
Statistically significant (P<0.05).
Isotemporal substitution models examining associations of physical activity at year 20 with cardiometabolic risk factors from year 30 are presented in Table 6. Replacing 30 minutes of sedentary time with 30 minutes of LPA resulted in a 0.15 cm lower waist circumference, a 0.20 μU/mL lower insulin, a 0.20 mg/dL higher HDL‐C, and a −0.01 SD lower composite risk score. Replacing the same duration of sedentary time with MVPA resulted in a 0.73 μU/mL lower insulin and a 4.57 mg/dL lower triglycerides. Replacing LPA with MVPA resulted in a 0.54 μU/mL lower insulin and a 4.51 mg/dL lower triglycerides. In sensitivity analyses with additional adjustment for year 20 cardiorespiratory fitness (data not shown), the observed associations were unchanged with 3 exceptions; replacement of sedentary time with LPA was no longer associated with HDL‐C (0.18 mg/dL; P=0.08), and replacement of sedentary time with MVPA, and LPA with MVPA, was no longer associated with insulin (−0.48 μU/mL, P=0.07; and −0.27 μU/mL, P=0.35, respectively), although associations remained in the expected direction.
Table 6.
Replacing 30 Min/Day in Sedentary and LPA With 30 Min/Day in LPA or MVPA at Year 20 With Cardiometabolic Risk Factors at Year 30, the CARDIA Study, 2005–2016 (N=1922)
| Cardiometabolic Risk Factors | Replace Sedentary With LPA | Replace Sedentary With MVPA | Replace LPA With MVPA | ||||||
|---|---|---|---|---|---|---|---|---|---|
| β | 95% CI | P Value | β | 95% CI | P Value | β | 95% CI | P Value | |
| Waist circumference, cm | −0.15* | −0.27, −0.02* | 0.020* | −0.36 | −0.77, 0.05 | 0.084 | −0.21 | −0.65, 0.23 | 0.347 |
| Average blood pressure, mm Hg | −0.08 | −0.27, 0.11 | 0.404 | −0.36 | −0.97, 0.26 | 0.254 | −0.28 | −0.94, 0.39 | 0.414 |
| Glucose, mg/dL | −0.42 | −0.85, 0.03 | 0.065 | 0.66 | −0.78, 2.10 | 0.369 | 1.08 | −0.49, 2.64 | 0.177 |
| Insulin, μU/mL | −0.20* | −0.35, −0.04* | 0.012* | −0.73* | −1.23, −0.24* | 0.004* | −0.54* | −1.08, −0.00* | 0.049* |
| Triglycerides, mg/dL | −0.06 | −1.04, 0.91 | 0.900 | −4.57* | −7.75, −1.39* | 0.005* | −4.51* | −7.96, −1.06* | 0.010* |
| HDL‐C, mg/dL | 0.20* | 0.00, 0.40* | 0.045* | −0.21 | −0.85, 0.44 | 0.528 | −0.41 | −1.11, 0.29 | 0.252 |
| Composite risk score† | −0.01* | −0.02, −0.00* | 0.020* | −0.03 | −0.05, 0.00 | 0.070 | −0.02 | −0.05, 0.02 | 0.307 |
CARDIA indicates Coronary Artery Risk Development in Young Adults; HDL‐C, high‐density lipoprotein cholesterol; LPA, light‐intensity physical activity; MVPA, moderate‐to‐vigorous intensity physical activity.
Statistically significant (P<0.05). Models adjust for year 20 center, age, race, sex, employment, health insurance, medication use, smoking status, alcohol consumption, body mass index (excluding waist circumference and composite risk score models), total wear time, and the cardiometabolic risk factor of interest (eg, model predicting change in waist circumference is adjusted for year 20 waist circumference).
Composite risk score was calculated by standardizing and summing waist circumference, average blood pressure ([systolic+diastolic]/2), log glucose, log insulin, log triglycerides, and negative HDL‐C, then dividing by 6 to create a z‐score.
Isotemporal substitution models examining change in physical activity with change in the cardiometabolic risk factors from year 20 to year 30 are shown in Table 7. A 30‐minute increase in MVPA, when replacing a 30‐minute increase in sedentary time, was associated with a 1.52‐cm decrease in waist circumference, a 1.13‐μU/mL decrease in insulin, a 6.92‐mg/dL decrease in triglycerides, a 1.59‐mg/dL increase in HDL‐C, and a 0.08‐SD decrease in the composite risk score. Similarly, a 30‐minute increase in MVPA, when replacing an increase in LPA, was associated with a 1.47‐cm decrease in waist circumference, a 0.93‐μU/mL decrease in insulin, a 6.52‐mg/dL decrease in triglycerides, a 1.35‐mg/dL increase in HDL‐C, and a 0.08‐SD decrease in the composite risk score. There were no significant associations when replacing a 30‐minute increase in LPA with the same duration of sedentary time. With additional adjustment for cardiorespiratory fitness (data not shown), the findings were unchanged with 1 exception: replacement of LPA with MVPA was no longer associated with a decrease in insulin, although the direction of the association was unchanged (−0.71 μU/mL; P=0.12).
Table 7.
Replacing a 30 Min/Day Increase in Sedentary and LPA With a 30 Min/Day Increase in LPA or MVPA From Year 20 to Year 30 on Change in Individual Cardiometabolic Risk Factors From Year 20 to Year 30, the CARDIA Study, 2005–2016 (N=913)
| Change in Cardiometabolic Risk Factors | Replace ΔSedentary With ΔLPA | Replace ΔSedentary With ΔMVPA | Replace ΔLPA With ΔMVPA | ||||||
|---|---|---|---|---|---|---|---|---|---|
| β | 95% CI | P Value | β | 95% CI | P Value | β | 95% CI | P Value | |
| ΔWaist circumference, cm | −0.06 | −0.25, 0.13 | 0.557 | −1.52* | −2.21, −0.84* | <0.001* | −1.47* | −2.19, −0.74* | <0.001* |
| ΔAverage blood pressure, mm Hg | 0.05 | −0.24, 0.33 | 0.751 | 0.68 | −0.34, 1.70 | 0.188 | 0.64 | −0.45, 1.72 | 0.249 |
| ΔGlucose, mg/dL | 0.18 | −0.45, 0.81 | 0.573 | −1.18 | −3.43, 1.07 | 0.304 | −1.36 | −3.75, 1.03 | 0.265 |
| ΔInsulin, μU/mL | −0.19 | −0.42, 0.04 | 0.098 | −1.13* | −1.95, −0.31* | 0.007* | −0.93* | −1.80, −0.06* | 0.035* |
| ΔTriglycerides, mg/dL | −0.40 | −1.73, 0.94 | 0.559 | −6.92* | −11.69, −2.15* | 0.005* | −6.52* | −11.6, −1.45* | 0.012* |
| ΔHDL‐C, mg/dL | 0.25 | −0.07, 0.56 | 0.133 | 1.59* | 0.45, 2.73* | 0.006* | 1.35* | 0.13, 2.56* | 0.030* |
| ΔComposite risk score† | −0.01 | −0.02, 0.01 | 0.352 | −0.08* | −0.13, −0.04* | <0.001* | −0.08* | −0.12, −0.03* | 0.002* |
CARDIA indicates Coronary Artery Risk Development in Young Adults; HDL‐C indicates high‐density lipoprotein cholesterol; LPA, light‐intensity physical activity; MVPA, moderate‐to‐vigorous intensity physical activity.
Statistically significant (P<0.05). Models adjust for year 20 center, age, race, sex, employment, health insurance, medication use, smoking status, alcohol consumption, body mass index (excluding waist circumference and composite risk score models), sedentary time, MVPA, change in total wear time, and the cardiometabolic risk factor of interest (eg, model predicting change in waist circumference is adjusted for year 20 waist circumference).
Composite risk score was calculated by standardizing and summing waist circumference, average blood pressure ([systolic+diastolic]/2), log glucose, log insulin, log triglycerides, and negative HDL‐C, then dividing by 6 to create a z‐score.
Discussion
The primary finding of this population‐based cohort study is that substituting sedentary time with LPA in early midlife (ages 38–50 years, CARDIA year 20) was associated with a significantly lower composite cardiometabolic risk score in late midlife (ages 48–60 years, CARDIA year 30), which was primarily characterized by lower waist circumference and insulin and higher HDL‐C. Additionally, an increase in MVPA from early to late midlife, when replacing an equivalent increase in sedentary time or LPA, was associated with a significant decrease in the composite cardiometabolic risk score from the same 10‐year period. This decrease in cardiometabolic risk was characterized by a corresponding decrease in waist circumference, insulin, and triglycerides and an increase in HDL‐C. These associations were largely robust after adjustment for baseline (year 20) demographics, lifestyle behaviors, the cardiometabolic risk factor of interest, activity level (in the change models), and cardiorespiratory fitness. This study contributes to the existing literature by using isotemporal substitution to examine the associations of accelerometer‐measured physical activity with a prospective assessment of cardiometabolic risk. Furthermore, this study is novel for using the isotemporal substitution paradigm to examine change in activity with change in cardiometabolic risk.
The majority of studies in the isotemporal literature examining cardiometabolic outcomes have focused on individual risk factors, whereas few have additionally included a composite or clustered cardiometabolic risk score. Nilsson et al examined whether sedentary behavior, when replaced with LPA or MVPA, influenced a composite risk score (including waist circumference, blood pressure, glucose, HDL‐C, and triglycerides) in 120 women aged 65 to 70 years.18 They found that replacing 10 min/day of sedentary time or LPA with MVPA was associated with a decrease in the composite risk score; however, associations were attenuated after adjustment for total accelerometer wear time. Knaeps et al also examined the theoretical substitution of sedentary time with LPA and MVPA on a composite risk score among 410 Flemish adults, using the same risk factors as Nilsson et al. The investigators found that replacing 30 min/day of sedentary time or LPA with MVPA resulted in a significantly lower composite risk score. There was no observed benefit of replacing sedentary time with LPA, which differs from our study findings. However, Ekblom‐Bak et al found that replacing 10 minutes of sedentary time with LPA, MPA, or VPA was associated with significantly lower metabolic syndrome prevalence among 836 participants from the Swedish Cardiopulmonary Bioimage Study, with lower odds in a dose‐response manner with higher physical activity intensity.14 Across studies, health benefits were observed when replacing sedentary time with physical activity, with the strongest associations found for MVPA. However, these studies are limited by their cross‐sectional study design. The present analysis contributes to the literature by using the isotemporal substitution paradigm to examine associations of physical activity with an assessment of cardiometabolic risk 10 years later.
In addition to prospectively assessing cardiometabolic risk factors, we examined whether change in activity was associated with change in cardiometabolic risk factors over the same 10‐year period. A 30 min/day increase in MVPA over 10 years, when replacing a 30 min/day increase in sedentary behavior or LPA, was associated with a significant decrease in the composite risk score. In our sample, we observed an average increase in sedentary time of ≈40 min/day and a concurrent decrease in MVPA of ≈5 min/day from early to late midlife. Furthermore, we observed adverse changes in the majority of the cardiometabolic risk factors over the same 10‐year period. Our findings indicate that if we replace the traditional increase in sedentary time with an increase in MVPA, we can prevent the usual decline in cardiometabolic health observed over time. This supports the notion that even in midlife, it is not too late to become physically activity to improve cardiometabolic health.
When examining the individual cardiometabolic risk factors, we observed a greater health benefit for insulin and triglycerides when replacing sedentary time with MVPA, as compared with replacing sedentary time with LPA, which is consistent with the existing literature.11, 12, 14, 16, 20 However, when replacing sedentary time with LPA, we observed significant reductions in waist circumference and insulin, and a significant increase in HDL‐C 10 years later, illustrating the importance of LPA for future cardiometabolic risk reduction. In comparison to MVPA, less is known about the role of LPA on health outcomes.39 Our findings add to the growing body of evidence regarding the health benefits of lower‐intensity physical activity. This prospective study extends the work of LaMonte et al, who recently reported favorable cross‐sectional associations between accelerometer‐measured LPA with many markers of cardiometabolic risk, including HDL‐C, triglycerides, glucose, C‐reactive protein, BMI, and waist circumference among 4832 women enrolled in the Objective Physical Activity and Cardiovascular Health Study.40 The investigators also found that LPA was favorably associated with the Reynolds Risk Score, which predicts 10‐year cardiovascular disease risk.41, 42 Taken together, our findings have important implications for health promotion efforts given that individuals may be more willing to increase LPA as compared with MVPA. However, it is important to note that the lack of a dose‐response relationship (significant findings for the year 30 composite risk score when replacing sedentary time with LPA, but not when replacing sedentary time with MVPA) raises concerns about biological plausibility, and thus findings should be interpreted with caution.
Strengths of this study include the use of objectively measured physical activity and cardiometabolic risk factors at 2 time points, when participants were in early and late midlife, and the use of a relatively large biracial sample. This study is also novel for using isotemporal substitution to examine change in physical activity with change in cardiometabolic risk factors. This provides the opportunity to present a novel methodological approach that should be considered in other longitudinal physical activity studies. However, there are several study limitations that should be noted. First, isotemporal substitution is a statistical model that does not reflect real‐time reallocation. Second, we were unable to account for sleep duration or quality in the analyses, as done in several previous studies.11, 20, 21, 43 Isotemporal substitution does not require representation from the full 24‐hour day; therefore, omission of sleep is acceptable within this statistical framework. However, this does not eliminate the possibility of confounding attributed to differences in sleep duration or quality across individuals. Third, with any observational study, there is a potential for residual confounding. However, this possibility is minimized through use of the well‐characterized CARDIA data set. Last, this study was limited to black and white adults, and therefore findings may not be generalizable to individuals of other races/ethnicities and different age groups.
In conclusion, this study provides evidence that replacing sedentary behaviors with LPA is associated with a lower cardiometabolic risk score 10 years later. This study also makes a novel contribution to the isotemporal literature by examining the associations of change in activity with change in cardiometabolic risk over a 10‐year period. We found that an increase in MVPA from early to late midlife, when replacing an increase in sedentary time or LPA, was associated with a significant decrease in the cardiometabolic risk score over the same time period. Our findings illustrate that improvements in cardiometabolic health may be possible as one ages from early to late midlife by reallocating sedentary time to LPA or MVPA. Clinicians and public health practitioners should consider the potential benefits to cardiometabolic health of encouraging LPA in addition to MVPA among midlife adults.
Sources of Funding
The Coronary Artery Risk Development in Young Adults Study (CARDIA) is supported by contracts HHSN268201300025C, HHSN268201300026C, HHSN268201300027C, HHSN268201300028C, HHSN268201300029C, and HHSN268200900041C from the National Heart, Lung, and Blood Institute (NHLBI), the Intramural Research Program of the National Institute on Aging (NIA), and an intra‐agency agreement between the NIA and NHLBI (AG0005). Accelerometer data collection was supported by grants R01 HL078972 and R56 HL125423 from the NHLBI. Whitaker was supported by T32 HL007779 from the NHLBI.
Disclosures
The views expressed in this manuscript are those of the authors and do not necessarily represent the views of the National Heart, Lung, and Blood Institute; the National Institutes of Health; or the U.S. Department of Health and Human Services.
Acknowledgments
The authors thank the investigators, the staff, and the participants of the CARDIA study for their valuable contributions.
(J Am Heart Assoc. 2019;8:e010212 DOI: 10.1161/JAHA.118.010212)
This article was handled independently by Peter Wilson, MD, as a guest editor. The editors had no role in the evaluation of the manuscript or in the decision about its acceptance.
References
- 1. Brocklebank LA, Falconer CL, Page AS, Perry R, Cooper AR. Accelerometer‐measured sedentary time and cardiometabolic biomarkers: a systematic review. Prev Med. 2015;76:92–102. [DOI] [PubMed] [Google Scholar]
- 2. Altenburg TM, Lakerveld J, Bot SD, Nijpels G, Chinapaw MJ. The prospective relationship between sedentary time and cardiometabolic health in adults at increased cardiometabolic risk—the Hoorn Prevention Study. Int J Behav Nutr Phys Act. 2014;11:90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Kronenberg F, Pereira MA, Schmitz MK, Arnett DK, Evenson KR, Crapo RO, Jensen RL, Burke GL, Sholinsky P, Ellison RC, Hunt SC. Influence of leisure time physical activity and television watching on atherosclerosis risk factors in the NHLBI Family Heart Study. Atherosclerosis. 2000;153:433–443. [DOI] [PubMed] [Google Scholar]
- 4. Thorp AA, Healy GN, Owen N, Salmon J, Ball K, Shaw JE, Zimmet PZ, Dunstan DW. Deleterious associations of sitting time and television viewing time with cardiometabolic risk biomarkers: Australian Diabetes, Obesity and Lifestyle (AusDiab) study 2004–2005. Diabetes Care. 2010;33:327–334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Wijndaele K, Healy GN, Dunstan DW, Barnett AG, Salmon J, Shaw JE, Zimmet PZ, Owen N. Increased cardiometabolic risk is associated with increased TV viewing time. Med Sci Sports Exerc. 2010;42:1511–1518. [DOI] [PubMed] [Google Scholar]
- 6. Thorp AA, Owen N, Neuhaus M, Dunstan DW. Sedentary behaviors and subsequent health outcomes in adults a systematic review of longitudinal studies, 1996–2011. Am J Prev Med. 2011;41:207–215. [DOI] [PubMed] [Google Scholar]
- 7. Mekary RA, Willett WC, Hu FB, Ding EL. Isotemporal substitution paradigm for physical activity epidemiology and weight change. Am J Epidemiol. 2009;170:519–527. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Buman MP, Hekler EB, Haskell WL, Pruitt L, Conway TL, Cain KL, Sallis JF, Saelens BE, Frank LD, King AC. Objective light‐intensity physical activity associations with rated health in older adults. Am J Epidemiol. 2010;172:1155–1165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. van Roekel EH, Bours MJ, Breedveld‐Peters JJ, Willems PJ, Meijer K, Kant I, van den Brandt PA, Beets GL, Sanduleanu S, Weijenberg MP. Modeling how substitution of sedentary behavior with standing or physical activity is associated with health‐related quality of life in colorectal cancer survivors. Cancer Causes Control. 2016;27:513–525. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Balboa‐Castillo T, Leon‐Munoz LM, Graciani A, Rodriguez‐Artalejo F, Guallar‐Castillon P. Longitudinal association of physical activity and sedentary behavior during leisure time with health‐related quality of life in community‐dwelling older adults. Health Qual Life Outcomes. 2011;9:47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Buman MP, Winkler EA, Kurka JM, Hekler EB, Baldwin CM, Owen N, Ainsworth BE, Healy GN, Gardiner PA. Reallocating time to sleep, sedentary behaviors, or active behaviors: associations with cardiovascular disease risk biomarkers, NHANES 2005–2006. Am J Epidemiol. 2014;179:323–334. [DOI] [PubMed] [Google Scholar]
- 12. Healy GN, Winkler EA, Owen N, Anuradha S, Dunstan DW. Replacing sitting time with standing or stepping: associations with cardio‐metabolic risk biomarkers. Eur Heart J. 2015;36:2643–2649. [DOI] [PubMed] [Google Scholar]
- 13. Yates T, Henson J, Edwardson C, Dunstan D, Bodicoat DH, Khunti K, Davies MJ. Objectively measured sedentary time and associations with insulin sensitivity: importance of reallocating sedentary time to physical activity. Prev Med. 2015;76:79–83. [DOI] [PubMed] [Google Scholar]
- 14. Ekblom‐Bak E, Ekblom O, Bergstrom G, Borjesson M. Isotemporal substitution of sedentary time by physical activity of different intensities and bout lengths, and its associations with metabolic risk. Eur J Prev Cardiol. 2016;23:967–974. [DOI] [PubMed] [Google Scholar]
- 15. 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;47:2070–2075. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Hamer M, Stamatakis E, Steptoe A. Effects of substituting sedentary time with physical activity on metabolic risk. Med Sci Sports Exerc. 2014;46:1946–1950. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Healy GN, Winkler EA, Brakenridge CL, Reeves MM, Eakin EG. Accelerometer‐derived sedentary and physical activity time in overweight/obese adults with type 2 diabetes: cross‐sectional associations with cardiometabolic biomarkers. PLoS One. 2015;10:e0119140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Nilsson A, Wahlin‐Larsson B, Kadi F. Physical activity and not sedentary time per se influences on clustered metabolic risk in elderly community‐dwelling women. PLoS One. 2017;12:e0175496. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Huang WY, Wong SH, He G, Salmon JO. Isotemporal substitution analysis for sedentary behavior and body mass index. Med Sci Sports Exerc. 2016;48:2135–2141. [DOI] [PubMed] [Google Scholar]
- 20. 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;15:197–203. [DOI] [PubMed] [Google Scholar]
- 21. Stamatakis E, Rogers K, Ding D, Berrigan D, Chau J, Hamer M, Bauman A. All‐cause mortality effects of replacing sedentary time with physical activity and sleeping using an isotemporal substitution model: a prospective study of 201,129 mid‐aged and older adults. Int J Behav Nutr Phys Act. 2015;12:121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Fishman EI, Steeves JA, Zipunnikov V, Koster A, Berrigan D, Harris TA, Murphy R. Association between objectively measured physical activity and mortality in NHANES. Med Sci Sports Exerc. 2016;48:1303–1311. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Matthews CE, Keadle SK, Troiano RP, Kahle L, Koster A, Brychta R, Van Domelen D, Caserotti P, Chen KY, Harris TB, Berrigan D. Accelerometer‐measured dose‐response for physical activity, sedentary time, and mortality in US adults. Am J Clin Nutr. 2016;104:1424–1432. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Dohrn I, Kwak L, Oja P, Sjostrom M, Hagstromer M. Replacing sedentary time with physical activity: a 15‐year follow‐up of mortality in a national cohort. Clin Epidemiol. 2018;10:179–186. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Coronary Artery Risk Development in Young Adults: New Investigators. University of Alabama at Birmingham. Available at: https://www.cardia.dopm.uab.edu/invitation-to-new-investigators. Accessed December 6, 2018.
- 26. Friedman GD, Cutter GR, Donahue RP, Hughes GH, Hulley SB, Jacobs DR Jr, Liu K, Savage PJ. CARDIA: study design, recruitment, and some characteristics of the examined subjects. J Clin Epidemiol. 1988;41:1105–1116. [DOI] [PubMed] [Google Scholar]
- 27. Pettee Gabriel K, Whitaker KM, Duprez D, Sternfeld B, Lewis CE, Sidney S, Knell G, Jacobs DR Jr. Clinical importance of non‐participation in a maximal graded exercise test on risk of non‐fatal and fatal cardiovascular events and all‐cause mortality: CARDIA study. Prev Med. 2018;106:137–144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008;40:181–188. [DOI] [PubMed] [Google Scholar]
- 29. Matthews CE. Calibration of accelerometer output for adults. Med Sci Sports Exerc. 2005;37(11 Suppl):S512–S522. [DOI] [PubMed] [Google Scholar]
- 30. Freedson PS, Melanson E, Sirard J. Calibration of the Computer Science and Applications, Inc. accelerometer. Med Sci Sports Exerc. 1998;30:777–781. [DOI] [PubMed] [Google Scholar]
- 31. Whitaker KM, Pettee Gabriel K, Jacobs DR Jr, Sidney S, Sternfeld B. Comparison of two generations of ActiGraph accelerometers: the CARDIA study. Med Sci Sports Exerc. 2018;50:1333–1340. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Ekelund U, Griffin SJ, Wareham NJ. Physical activity and metabolic risk in individuals with a family history of type 2 diabetes. Diabetes Care. 2007;30:337–342. [DOI] [PubMed] [Google Scholar]
- 33. Franks PW, Ekelund U, Brage S, Wong MY, Wareham NJ. Does the association of habitual physical activity with the metabolic syndrome differ by level of cardiorespiratory fitness? Diabetes Care. 2004;27:1187–1193. [DOI] [PubMed] [Google Scholar]
- 34. Ekelund U, Brage S, Franks PW, Hennings S, Emms S, Wareham NJ. Physical activity energy expenditure predicts progression toward the metabolic syndrome independently of aerobic fitness in middle‐aged healthy Caucasians: the Medical Research Council Ely Study. Diabetes Care. 2005;28:1195–1200. [DOI] [PubMed] [Google Scholar]
- 35. Whitaker KM, Buman MP, Odegaard AO, Carpenter KC, Jacobs DR Jr, Sidney S, Pereira MA. Sedentary behaviors and cardiometabolic risk: an isotemporal substitution analysis. Am J Epidemiol. 2018;187:181–189. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Wagenknecht LE, Cutter GR, Haley NJ, Sidney S, Manolio TA, Hughes GH, Jacobs DR. Racial differences in serum cotinine levels among smokers in the Coronary Artery Risk Development in (Young) Adults study. Am J Public Health. 1990;80:1053–1056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Sijtsma FP, Meyer KA, Steffen LM, Shikany JM, Van Horn L, Harnack L, Kromhout D, Jacobs DR Jr. Longitudinal trends in diet and effects of sex, race, and education on dietary quality score change: the Coronary Artery Risk Development in Young Adults study. Am J Clin Nutr. 2012;95:580–586. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Schreiner PJ, Jacobs DR Jr, Wong ND, Kiefe CI. Twenty‐five year secular trends in lipids and modifiable risk factors in a population‐based biracial cohort: the Coronary Artery Risk Development in Young Adults (CARDIA) study, 1985–2011. J Am Heart Assoc. 2016;5:e003384 DOI: 10.1161/JAHA.116.003384. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Fuzeki E, Engeroff T, Banzer W. Health benefits of light‐intensity physical activity: a systematic review of accelerometer data of the National Health and Nutrition Examination Survey (NHANES). Sports Med. 2017;47:1769–1793. [DOI] [PubMed] [Google Scholar]
- 40. LaMonte MJ, Lewis CE, Buchner DM, Evenson KR, Rillamas‐Sun E, Di C, Lee IM, Bellettiere J, Stefanick ML, Eaton CB, Howard BV, Bird C, LaCroix AZ. Both light intensity and moderate‐to‐vigorous physical activity measured by accelerometry are favorably associated with cardiometabolic risk factors in older women: the Objective Physical Activity and Cardiovascular Health (OPACH) study. J Am Heart Assoc. 2017;6:e007064 DOI: 10.1161/JAHA.117.007064. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 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;297:611–619. [DOI] [PubMed] [Google Scholar]
- 42. Ridker PM, Paynter NP, Rifai N, Gaziano JM, Cook NR. C‐reactive protein and parental history improve global cardiovascular risk prediction: the Reynolds Risk Score for men. Circulation. 2008;118:2243–2251. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. 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:e0139984. [DOI] [PMC free article] [PubMed] [Google Scholar]