The 2023 statistics from the American Heart Association and National Institutes of Health show that cardiovascular diseases (CVD) remain the leading cause of global mortality, accounting for 32% of deaths worldwide.1 CVD not only cause high healthcare expenses, but also reduces life expectancy, and quality of life.1 It is well-documented that physical activity (PA) is a potent protective factor against CVD,2 while sedentary time (ST) is a significant risk factor for CVD and all-cause mortality.3–6
High ST has been linked to an increased risk of CVD and all-cause mortality, statistically independent of PA level based on a mutually adjusted approach.3,7–12 Previous meta-analyses have reinforced these findings in pooled analysis.13–15 In addition, PA level has been inversely associated with risk of CVD.16 Complementary, using stratified and joint analyses, more recent meta-analyses with pooled individual-level data examining the continuum of ST and PA demonstrated that very high PA levels might attenuate or even eliminate the associations of high ST levels with all-cause mortality and only one with CVD mortality.17–20 Therefore, the current evidence suggests that despite the mutually adjusted approach cannot be discarded, it is not fully explaining the health effects of ST and PA due the potential existence of effect modification.21
Considering that different approaches have been used to investigate the association of ST and PA levels on CVD outcomes, including independent, stratified, and joint analyses, this systematic review aimed to map out the following: (1) the existing conceptual definitions of independent, stratified, and joint analyses of ST and PA levels in studies focusing on CVD outcomes; (2) the assessment methods employed to assess ST and PA levels; (3) the operational approaches used to define the comparison groups with different ST and PA levels; and (4) the analytical approaches employed to investigate the association of ST and PA levels with CVD outcomes. Last, this review aimed to summarize the main findings obtained from independent, stratified, and joint analyses of ST and PA levels, in relation to CVD outcomes, and to discuss their implications on public health messages.
Methods
The protocol of this systematic review and search strategy are available in the OSF (https://osf.io/zpq25/) and the PROSPERO database (registration number: CRD42021281493). This study was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis.22 The study was previously designed to investigate the joint associations of ST and PA levels with CVD outcomes based on a harmonized analysis (pooled analysis). However, considering mainly the methodological heterogeneity observed among the included studies, pooled analyses were no longer performed, and we conducted a systematic review with results syntheses to better understand the main findings in different analytical approaches used to investigate the relationship of ST and PA with CVD outcomes.
Eligibility Criteria
Inclusion criteria of studies were: (1) addressing adults aged 18+ years; (2) prospective cohort studies that assessed ST and PA levels by self-reported methods or accelerometery; (3) data on individual level exposure, CVD mortality, and nonfatal major adverse cardiovascular events; (4) reported effect estimates as hazard ratios, odds ratios, or relative risks with 95% confidence intervals for CVD mortality and nonfatal CVD; and (5) language: English, Spanish, and Portuguese. Our exclusion criteria were: (1) no information on ST and PA levels with CVD outcomes and (2) brief communications, research letters, systematic reviews, narrative reviews, scoping reviews, randomized clinical trials, cross-sectional studies, case-control studies, or other different study designs. In the first-step screening (title and abstracts), articles with clear information about the independent, stratified, and/or joint associations of ST and PA levels with CVD outcomes were included. In the second-step screening (full-text analysis), it was considered the inclusion of the study that met all eligibility criteria. We also checked the reference list to find additional studies for this review.
Literature Search Strategy
A literature search was conducted in 4 databases: (1) PubMed, (2) Web of Science, (3) Embase, and (4) Scopus. Initially, all available records retrieved from the databases were uploaded in EndNote x9 to remove duplicates and to the Rayyan QCRI website for the screening process.23 Two independent reviewers (de Lucena Alves and Delpino) performed the screening phase, selected potentially eligible articles, checked titles, abstracts, and full text. Discrepancies were reviewed until a consensus was reached. Full-text screening followed eligibility criteria. Eligible and ineligible articles were listed with exclusion reasons (Supplementary Table S1 [available online]). Missing or unclear data prompted 2 email requests to authors. The search strategy is available at OSF (https://osf.io/zpq25/).
Three approaches on the association analyses of ST and PA with CVD outcomes were considered based on previous meta-analyses.11 Independent association of ST and PA levels with CVD outcomes were considered if the studies adjusted their analysis for ST and PA.3,7–12 Stratified associations were examined when studies compared all categories of one behavior (eg, ST) across different levels of the other behavior (eg, PA). This approach entailed a reference group for each ST or PA stratum. A joint association was considered when there was only one reference group for the combination of ST and PA categories17–20 (Figure 1).
—Schematic representation illustrating the independent, stratified, and joint associations between sedentary time and physical activity levels with the risk of CVD. CVD indicates cardiovascular diseases.
Citation: Journal of Physical Activity and Health 21, 10; 10.1123/jpah.2024-0019
Outcomes
The primary outcome was defined as CVD mortality. The secondary outcome was nonfatal CVD, which specifically included hospitalization due to acute myocardial infarction, acute ischemic stroke, cardiovascular-related interventions (eg, percutaneous coronary intervention and coronary artery bypass graft), heart failure, and peripheral artery disease.24,25 Acute myocardial infarction was defined as an admission using International Classification of Diseases Tenth, Revision (ICD-10) diagnostic code of I21 or I22. Acute ischemic stroke was defined as an admission with an ICD-10 diagnostic code of I63. Coronary revascularization was defined by a cardiovascular revascularization procedure history identified in the claims database. In addition, a new diagnosis of heart failure (ICD-10-CM) and peripheral artery disease (ICD-9-CM) were considered. We considered any reported methods for the outcome (eg, self-reported methods, diagnosis by a physician or other healthcare).
Risk of Bias
The Risk of Bias In Non-Randomized Studies—of Exposure26 scale was used to assess the methodological quality of nonrandomized clinical trials based on 7 domains: (1) confounders, (2) selection of participants into the study, (3) classification of interventions, (4) deviations from the intended intervention, (5) missing data, (6) measurement of outcomes, and (7) selection of the reported result. The evaluation was made by 2 reviewers (de Lucena Alves and Leão) and any discrepancies were reviewed until a consensus was reached.
Data Extraction
Data were extracted and coded in Excel (Microsoft Corp) by one author (de Lucena Alves). In case of doubt, a second author was consulted. The following variables were extracted: name of the first author; year of publication; country; number of participants; age of participants; sex; years of follow-up; CVD mortality and nonfatal major adverse cardiovascular events; health status; assessment methods of ST and PA levels (including different domains and type of information collected; eg, television viewing, leisure time, etc); and covariates included in the final adjusted model, the results of the main analyses performed, the conceptual definitions (eg, concept of joint analysis), operational (ie, how the groups of different ST and PA levels were created), and analytical (ie, statistical analyses employed) approaches used to investigate the independent, stratified, and/or joint associations of ST and PA levels with CVD outcomes. All data are presented as tabular form.
Results
Literature Search and Screening Process
The final literature search was performed in April 2023. A total of 15 230 studies were retrieved. After removing 2871 duplicates, 12 235 of the 12 359 remaining studies were excluded at the title and abstract screening. The 124 remaining studies were screened at the full-text level, leading to the exclusion of an additional 79 (Supplementary Table S1 [available online]). Therefore, 45 studies were included in this systematic review.3,4,7–12,24,25,27–61 Figure 2 shows the flowchart that summarizes the literature search and screening process.
—Flowchart of the literature search process. *Data from MEDLINE (via PubMed), Web of Science, Embase, and Scopus.
Citation: Journal of Physical Activity and Health 21, 10; 10.1123/jpah.2024-0019
Characteristics of the Included Studies
Tables 1 and 2 describe the characteristics of the studies that investigated the independent, stratified, and/or joint associations of ST and PA levels with CVD outcomes. Overall, the included studies were published between 2009 and 2022
Characteristics of the Included Studies That Investigated the Independent Association of ST and PA With Cardiovascular Outcomes in Adults (n = 31)
| Reference | Country | Sample size | Age | Sex | Follow-up |
|---|---|---|---|---|---|
| Katzmarzyk et al27 | Canada | 17,013 | 18–90 y | Females and males | 12.9 y |
| Wijndaele et al28 | United Kingdom | 12,608 | 61.4 (9) y | Females and males | 6.9 y (mean) |
| Herber-Gast et al30 | Australia | 6154 | 52.5 (1.5) y | Female | 9.9 y (mean) |
| Ensrud et al35 | United States | 2918 | Aged 71 y and older | Males | 4.5 y (mean) |
| Matthews et al34 | United States | 63,308 | 40–79 y | Females and males | 6.4 y (mean) |
| Chau et al37 | Australia | 50,817 | ≥20 y | Females and males | 3.3 y (mean) |
| Holme and Anderssen36 | Norway | 14,846 | 52.5 (2.7) y | Males | 12 y (mean) |
| Allesøe et al38 | Denmark | 12,096 | 45–64 y | Females | 15 y (mean) |
| Evenson et al41 | United States | 3809 | ≥40 y | Females and males | 6.7 y (median) |
| Lee et al39 | Korea | 336,560 | 39.7 y (mean) | Females and males | 6.1 y (median) |
| Johnsen et al40 | Sweden | 9961 | 42.7 y (mean) | Females and males | 13.1 y (mean) |
| Doukky et al42 | United States | 902 | 65.1 (14) y | Females and males | 3 y |
| Schnohr et al43 | United States | 12,314 | 20–98 y | Females and males | 33 y |
| Cabanas-Sánchez et al7 | Spain | 2657 | ≥60 y | Females and males | 11.7 y (median) |
| Dohrn et al45 | Sweden | 851 | 66.7 (10) y | Females and males | 14.2 y (mean) |
| Dohrn et al46 | Sweden | 851 | 66.7 (10) y | Females and males | 14.2 y (mean) |
| Hupin et al47 | France | 1011 | ≥65 y | Females and males | 15 y (mean) |
| Cuthbertson et al49 | United States | 13,534 | 54.0 (5.7) y | Females and males | 27.2 y (median) |
| Bellettiere et al5 | United States | 5638 | 79 (7) y | Females | 4.9 y (mean) |
| Dohrn et al48 | Sweden | 1220 | 45.3 (14.5) y | Females and males | 14.4 y (mean) |
| Hamer et al52 | United Kingdom | 479,658 | 56.5 (8.0) y | Females and males | 10.4 y (mean) |
| Holtermann et al24 | Denmark | 104,046 | 56 (14) y | Females and males | 10 y (median) |
| Gardner et al56 | United States | 528 | 69.0 (8.8) y | Females and males | 9.9 y (median) |
| Ballin et al4 | Sweden | 3343 | 70.5 (0.1) y | Females and males | 2.7 y (mean) |
| Ekblom-Bak et al57 | Sweden | 3693 | 60 y old | Females and males | >20 y |
| Yerramalla et al53 | France | 3319 | 68.9 y (mean) | Females and males | 6 y (mean) |
| Kim et al54 | United States | 5207 | 48.3 y (mean) | Females and males | 10.6 y (mean) |
| Yerramalla et al58 | France | 3321 | 60–83 y | Females and males | 6.2 y (mean) |
| Hooker et al60 | United States | 7607 | 63.4 (8.5) y | Females and males | 7.4 y (mean) |
| Dempsey et al61 | United Kingdom | 7671 | 70.2 (7.5) y | Females and males | 6.4 y (mean) |
| Paudel et al12 | United Kingdom | 328,228 | 55.9 (8.1) y | Females and males | 12.1 y (mean) |
Abbreviations: PA, physical activity; ST, sedentary time.
Characteristics of the Included Studies That Investigated the Stratified and Joint Associations of ST and PA With Cardiovascular Outcomes in Adults (n = 14)
| Reference | Country | Sample size | Age | Sex | Follow-up |
|---|---|---|---|---|---|
| Warren et al25 | United States | 7744 | 20–89 y | Males | 21 y (mean) |
| Matthews et al29 | United States | 240,819 | 50–71 y | Females and males | 8.5 (1.7) y |
| Chomistek et al31 | United States | 71,018 | 50–79 y | Females | 12.2 y (median) |
| Petersen et al33 | Denmark | 71,363 | 18–99 y | Females and males | 5.4 y (mean) |
| Seguin et al3 | United States | 92,234 | 50–79 y | Females | 12 y (mean) |
| Young et al32 | United States | 82,695 | ≥45 y | Males | 7.8 y (mean) |
| Bennett et al44 | England | 487,334 | 30–79 y | Females and males | 7.5 y |
| Celis-Morales et al8 | Scotland | 390,089 | 40–69 y | Females and males | 5.0 y (median) |
| Stamatakis et al21 | Australia | 149,077 | ≥45 y | Females and males | 8.9 y (median) |
| Liu et al50 | China | 93,110 | 52.8 (12.3) y | Females and males | 5.8 y (median) |
| LaMonte et al51 | United States | 80,982 | 50–79 y | Females | 9 y (mean) |
| Joundi et al55 | Canada | 143,180 | >60 y | Females and males | 9.4 y (median) |
| Jung et al59 | United States | 17,730 | ≥20 y | Females and males | 7 y (mean) |
| Li et al11 | China | 105,677 | 50.4 (9.6) y | Females and males | 11.1 y (median) |
Abbreviations: PA, physical activity; ST, sedentary time.
Independent Association of ST or PA With CVD Outcomes
Follow-up time varied between 2.7 and 33 years (mean). The median of sample size was 5.880 participants, ranging from 528 to 479.658. Most of the studies were conducted in the United States (n = 10) and Sweden (n = 6), followed by United Kingdom (n = 4), and France (n = 3). The age of the participants ranged from 38.6 (15.3) to 71.1 (7.2) years at baseline. A total of 27 studies included both sexes, 3 studies included only females, and 2 studies included only males. All studies (n = 32) included mixed samples of apparently healthy adults and with chronic diseases at baseline (eg, hypertension, obesity, type 2 diabetes). All studies presented sensitivity analysis excluding participants with chronic diseases at baseline or provided an adjustment for chronic disease at baseline, as well as deaths during the first and second year of follow-up.
Stratified and Joint Associations of ST and PA With CVD Outcomes
Follow-up time ranged from 5.4 to 21 years (mean). The median sample size was 92.234 participants, ranging from 7744 to 487.334. Most of the studies were conducted in the United States (n = 7), followed by Australia, Canada, China, Denmark, England, and Scotland (n = 6). The age of the participants ranged from 20.0 (1.3) to 63.8 (7.1) years at baseline. A total of 7 studies (61.5%) included females and males, 3 studies included only females (23%), and 2 studies (15.8%) included only males. Finally, all studies (n = 13) included mixed samples of apparently healthy adults and with chronic diseases at baseline. More details are reported in Table 2.
Analytical Approaches
Supplementary Tables S2 and S3 (available online) shows the conceptual definitions, analytical, operational approaches, as well as assessment methods and main results of the included studies that investigated the independent, stratified, and/or joint associations of ST and PA levels with CVD mortality and nonfatal CVD.
Independent Association of ST or PA With CVD Outcomes
All the studies (n = 31) used Cox proportional hazard models to estimate the hazard ratios and 95% confidence intervals for the independent association. Thirteen studies evaluated the mutually adjusted association of both ST and PA levels with CVD outcomes; 11 studies evaluated the association of PA with CVD outcomes, and 8 studies evaluated the association of ST with CVD outcomes. In total, 23 studies included statistical adjustments for ST or PA levels, while 9 studies did not include or report any statistical adjustment for ST or PA levels in the independent association analyses. Supplementary Table S2 (available online) presents a summary of the findings from these studies.
Stratified and Joint Analyses of ST and PA With CVD Outcomes
All studies (n = 14) analyzed the stratified and/or joint associations using Cox regression analysis. Among these, 3 studies used both stratified and joint association approach, where they examined the associations between ST and CVD outcomes with one reference group within each PA stratum (stratified analysis) and deriving a combined variable of ST and PA levels in several different groups, where the lowest ST and highest moderate-to-vigorous physical activity group served as the reference category (joint association). Three studies presented only stratified analysis, and 8 studies used only joint association. Supplementary Table S3 (available online) presents a summary of the findings from these studies.
Assessment Methods and Operational Approaches of ST and PA Levels
Independent Association of ST or PA With CVD Outcomes
Twenty studies (65.6%) and 11 studies (34.4%) that examined the independent associations between ST and PA with CVD assessed exposures by self-report and accelerometery, respectively. In the studies considering stratified and joint associations analyses, all used self-reported measures to assess ST and PA levels (n = 13). Eleven studies modeled ST as dichotomous variable; while 6 studies used ST as categorical variable (eg, low sedentary, highly sedentary, etc); 11 studies categorized participants into tertiles (6 studies) and quartile (5 studies), whereas 3 studies modeled ST as a continuous variable. One study provided no information on adjustment. Seven studies modeled PA as a continuous variable, while another 11 studies categorized participants into various predefined groups (eg, inactive, moderately inactive, moderately active, and active). Others grouped participants into PA tertiles (6 studies), quartiles (4 studies) and dichotomous variable (eg, meeting the recommendations of PA; n = 4).
Stratified and Joint Analyses of ST and PA With CVD Outcomes
All studies (n = 14) have created groups combining ST and PA levels and evaluated associations among the groups with CVD outcomes. For instance, CVD outcomes among those with high ST and low PA levels were compared to other combinations such as low/moderate ST and high/moderate PA. Nevertheless, as detailed in Supplementary Tables S2 and S3 (available online), there was a large heterogeneity on how each study categorized participants in these group.
CVD Outcomes
For the independent analyses, 14 studies (43.8%) investigated the associations between ST and/or PA with CVD mortality, while 18 (56.2%) evaluated incidence of both fatal and nonfatal CVD disease (eg, myocardial infarction, heart disease, stroke, and coronary heart diseases). For the stratified and joint approach, 8 studies included CVD mortality (61.5%), while 5 studies included the incidence of nonfatal coronary heart disease (38.5%).
Independent Associations Between ST or PA With CVD Outcomes
Supplementary Table S2 (available online) summarizes the findings regarding the independent associations between ST or PA levels and CVD outcomes. With the exception of 4 studies,12,30,37,61 all others7,9,27,28,36,37,41,42,46,48,52,58,60 studies reported a positive association between ST and adverse CVD outcomes (hazard ratio ranging from 1.01 to 1.62). Conversely, all4,24,34,35,38–40,43,47,49,54,56–58,60 except 3 studies12,36,40 studies reported a protective association between PA and CVD outcomes (hazard ratio ranging from 0.28 to 0.92). Importantly, these associations were mutually adjusted for ST and PA, and they were consistent regardless of the methods used to assess the movement behaviors (ie, accelerometery or self-reported measures).
Stratified and/or Joint Associations Between ST and PA With CVD Outcomes
Supplementary Table S3 (available online) summarizes the findings related to the stratified and/or joint associations between ST and PA levels with CVD outcomes. In terms of stratified association analysis,3,10,11,25,55,62 studies have showed a more pronounced magnitude of association between ST and CVD outcomes within the lower strata of PA levels. In addition, even among participants who reported levels of PA that approximated or exceeded currently recommended PA levels (150 min/wk), the effect of higher ST was attenuated, but remained associated with an increased risk of CVD outcomes.25,55 There were no statistically significant interactions observed in most studies addressing both stratified3,10,11,62 and joint8,10,31,33,50 associations. Based on the joint association analysis,8,10,11,29,31–33,44,50,55,59 5 studies11,29,32,55,59 found the highest risk for CVD outcomes when comparing participants with the highest ST and lowest PA levels (unhealthiest profile) with the healthiest profile, characterized by the lowest ST and highest PA levels. In the remaining 6 studies, the highest risk for CVD outcomes was identified in other intermediate combinations.8,10,31,33,44,50
Risk of Bias
The risk of bias was evaluated using the Risk of Bias In Non-Randomized Studies-of Exposure tool. The overall risk of bias showed that 29/45 articles presented some concerns; 12/45 presented low risk of bias and 4/45 as having high of risk bias. Details are reported in the Figure 3 and Supplementary Figure S1 (available online).
—Summary plot of the risk of bias (n = 45).
Citation: Journal of Physical Activity and Health 21, 10; 10.1123/jpah.2024-0019
Discussion
This study aimed to systematically review the literature on the independent, stratified, and joint associations of ST and PA with CVD outcomes in adults. Our main findings were: (1) studies lacked clear conceptual definitions for independent, stratified, and joint associations between ST and PA levels; (2) the most common analytical approaches used were Cox regression for independent, stratified, and joint analyses; (3) common operational approaches included dichotomous variables, tertiles, or quartiles for ST and/or PA; (4) most studies used self-reported methods to assess ST and PA levels; (5) the primary focus of the investigation was CVD mortality; (6) in mutually adjusted analyses, studies have consistently reported a positive association between ST and CVD outcomes, as well as an inverse association between PA and CVD outcomes; (7) in stratified analyses, there was a higher magnitude of association between ST and CVD outcomes within the lower strata of PA levels. Moreover, the effect of ST for CVD outcomes is attenuated but not eliminated in participants who reported levels of PA that approximated or exceeded currently recommended PA levels; and (8) in joint analysis, the highest risk for CVD outcomes was usually observed in adults with the unhealthiest profile, marked by highest ST and lowest PA levels, when compared to the healthiest profile, which featured lowest ST and highest PA levels. Also, the highest risk for CVD outcomes was found in various intermediate combinations.
A better understand of the relationship between ST, PA, and CVD outcomes require a better understanding of the 3 different approaches, their differences, and the complementary characteristics. Statistically independent effects of ST and PA on CVD outcomes, by mutually adjusted analyses, can basically estimate the association of one variable irrespective the levels of the other. However, this method falls short of capturing the true independence of the effects of both behaviors in real-world scenarios. Therefore, stratified and joint associations become crucial in enhancing the comprehension of the impact of both ST and PA on health outcomes.
Stratified analysis uses one reference group in each PA or ST stratum. This analysis is useful when researchers intend to understand whether the associations of ST and a given outcome vary by levels of PA, and vice versa. In this approach, we can investigate whether (or how much) PA can attenuate the harmful effect of ST on CVD outcomes, because we are assessing the effect of ST among different strata of PA level and observing if it happens and in which strata. In the study of Li et al,11 for example, the deleterious effect of high ST was only observed in the group of participants accumulating lower than 600 METs × minute per week, meaning that the harmful effect of high ST was not observed among participants with higher levels of PA.
Joint analyses are similar to the stratified approach, but it differs conceptually, being based on comparisons of a single reference group with all other combinations of time spent ST and PA. Here the research question being addressed is slightly different, providing more detail about the combined effect of both behaviors and its magnitude (potentially higher due to comparisons with the single healthier or unhealthier group of combination). Li et al,11 using joint analyses, showed the highest hazard ratios for CVD outcomes when the combination of highest ST and lowest levels of PA was compared with the healthiest group (ie, lowest ST and highest levels of PA).
Considering the complementary perspective of independent, stratified, and joint approaches, if researchers are focused on comprehensively addressing associations of ST and/or PA levels with CVD outcomes the following steps might be recommended: (1) to perform statistical analysis to investigate the independent association of ST and/or PA with CVD outcomes, adding potential confounders previously defined by a direct acyclic graph or similar; (2) in addition to other confounders, mutually adjustment can be informative; (3) to perform stratified analyses (fitting an interaction term between ST and PA can be relevant; both additive and multiplicative scales are useful given that different approaches generate important changes in effect measure); and (4) to implement the joint association approach by creating different combinations of PA and ST and comparing them with only one reference group (unhealthiest or healthiest combination).
The other methodological characteristic relevant in this research topic is the ST and PA measurement. Most studies relied on self-reported methods. While we acknowledge that accelerometers can be costly and logistically challenging in large cohort studies, and present also limitations, self-reported variables are prone to important measurement errors. In terms of ST, for example, systematic reviews identified that it generally underestimating ST by approximately 2.5 hours, in addition to a fair to moderate agreement when compared with accelerometers.63,64 Moreover, most studies used television viewing as a proxy measure of ST, which may be also problematic because some sedentary activities (eg, work-related screen time, television viewing) may have different magnitudes of association of CVD outcomes compared with total ST.25,27–29 Considering that all studies included in our review using stratified and joint analyses were based on self-report methods, the impact of measurement method was not addressed. However, it is important to note that further than potential information bias, the different methods of measurement also can impact on the public health messages. Categories of analyses, as well as the total amount of PA that can eliminate or attenuate the deleterious risk of ST will be heavily influenced by what kind of methods are being applied.
In both independent and joint association analyses, findings suggest large heterogeneity in terms of categories of ST, where most studies used median (eg, <8 h/d for low ST and > 8 h/d for high ST), and others different an arbitrary defined cutoff, such as <1 hour per day, 2 to 3 hours per day, and >4 hours per day. Despite the cutoff points of ST are not clear, some classifications seem to be not reasonable and applied for real life (eg, <1 h/d), given that significant associations are found mainly in people with more than 8 hours per day of ST (eg, high ST associated with a high risk for CVD outcomes).15 In the operational approach for PA, time spent in moderate to vigorous PA was the information mostly used, being classified in tertiles or quartiles, while others reported PA as “low,” “medium,” “high,” “light,” “moderate,” or “vigorous” intensity. Although those approaches are highly prevalent in the literature, they can create difficulties for interpretation mainly due to the high heterogeneity. For instance, tertiles of PA in a prospective cohort of a low-middle-income country probably are different from a prospective cohort from a high-income country.
Establishing a standardized cutoff point for ST and PA offers potential advantages in individual studies, facilitating the comparisons of different studies, enabling to address public health recommendations, and providing decision makers with clearer messages. Sample size limitations and consistency with the research question must be also considered in the operational definitions. Furthermore, meta-analyses with pooled individual-level information, which are an important method to deal with the heterogeneity of operational definitions of ST and PA, might prioritize the distribution-based categories (tertiles or quartiles) considering the heterogeneity of ST and PA measurement. In this sense, international research collaboration platforms of cohorts to explore the effects of PA on a wide range of health outcomes, such as the initiatives the Prospective Physical Activity, Sitting and Sleep consortium,65 are highly relevant. Additionally, the adult accelerometer consortium is equally relevant.19
Based on independent associations, with the exception of 5 studies,11,29,32,55,59 all other studies have demonstrated a positive association between ST and CVD outcomes. This finding is in accordance with previous studies that demonstrated a positive association in a dose–response manner of ST and adverse health-related outcomes.15,66,67 On the other hand, most of included studies have showed an inverse association between PA and CVD outcomes. This finding is in accordance with previous meta-analyses that showed an inverse nonlinear dose–response association between PA and CVD outcomes, suggesting a substantial protective role of PA against this adverse health-related outcomes.16,19,68 It is important to highlight that the abovementioned associations were independent of the methods used to assess the movement behaviors (ie, accelerometery or self-reported measures). Of note, both associations were mutually adjusted.
In terms of stratified association analysis,3,10,11,25,55,62 most of the studies showed a more pronounced magnitude of association between ST and CVD outcomes within the lower strata of PA levels. This result is in accordance with previous harmonized meta-analyses that reported that PA may attenuated, but not eliminate the deleterious effect of high ST in each stratum of PA17,18,20 regarding CVD and all-cause mortality.
In the joint association analysis,8,10,11,29,31–33,44,50,55,59 5 studies11,29,32,55,59 found the highest risk for CVD outcomes when comparing adults with the highest ST and lowest PA levels (unhealthiest profile) with the healthiest profile (lowest ST and highest PA levels). In other studies, highest risk for CVD outcomes was identified only in other intermediate combinations. Nevertheless, our results are in accordance with previous harmonized meta-analyses that showed that higher ST is associated with higher mortality risk in less active individuals. Last, there were no statistically significant interactions observed for most studies have shown both stratified3,10,11,62 and joint8,10,31,33,50 associations.
Our systematic review has limitations that should be mentioned. First, our search strategy was limited to studies published in English, Spanish, and Portuguese languages. Second, we did not search the gray literature. Consequently, there is a possibility that some relevant studies may not have been included in our systematic review. Third, we were unable to conduct a meta-analysis of the included studies due to their significant heterogeneity.
Conclusions
In summary, our findings demonstrate a lack of consensus on definitions, assessment methods, and operational and analytical approaches to investigate the association of ST and PA levels with CVD outcomes. In mutually adjusted analyses, ST was positively linked to CVD outcomes, while PA showed a protective association. Stratified analyses highlighted stronger ST-CVD links in low PA individuals. High PA mitigated ST’s impact but did not eliminate it. Jointly, highest ST and lowest PA posed the greatest CVD risk, with elevated risk in intermediate combinations.
Independent, stratified, and joint associations analyses may bring different and complementary public health messages for cardiovascular health promotion. Therefore, we encourage the development of future research to reduce heterogeneity and improve device-based assessment of ST and PA for future meta-analysis/harmonized analysis using individual participant data on the independent, stratified, and joint associations of ST and PA levels with CVD. The research area would also benefit from more transparent reports of operational definitions, and adequate interpretations of findings according to the conceptual approach.
Amendments to the Research Protocol
Instead of pursuing a harmonized meta-analysis, as initially outlined in the systematic review protocol, we opted to present the main findings in a textual syntheses approach and to increase the methodological appraisal. This decision was necessary mainly due to several challenges faced during this study, mostly driven by 3 analytical approaches observed: independent analyses, stratified, and joint associations. Therefore, our revised aims included (1) to examine the conceptual definitions of independent, stratified, and joint analyses of ST and PA levels in studies concentrating on CVD outcomes, (2) evaluate the assessment methods utilized for measuring ST and PA levels, (3) investigate the operational approaches used to define comparison groups with varying ST and PA levels, and (4) examine the analytical approaches employed to explore the association of ST and PA levels with CVD outcomes. This review ultimately seeks to summarize the results obtained from independent, stratified, and joint analyses of ST and PA levels concerning CVD outcomes and discuss their implications for public health messaging.
Acknowledgments
Funding: de Lucena Alves is financed in part by CAPES (Finance Code 001). Mielke is supported by a National Health and Medical Research Council Investigator Grant (APP2008702). Costa is supported by the National Council for Scientific and Technological Development (CNPq research productivity grant: 306537/2022-2). Delpino is financed in part by the National Council for Scientific and Technological Development (CNPq) with a postdoctoral fellowship. Availability of Data and Materials: The authors followed the recommendations from the International Committee of International Journal Editors for data sharing. Data, statistical code, and research materials are publicly available at Open Science Framework. Author Contributions: Concept and design: de Lucena Alves, Delpino, Ekelund, Costa, Crochemore-Silva. Data acquisition: de Lucena Alves, Delpino. Interpretation: de Lucena Alves, Leão, Delpino, Mielke, Ekelund, Costa, Crochemore-Silva. Drafting of the manuscript and critically revising the manuscript: All authors. All authors gave their final approval and agreed to be accountable for all aspects of the work.
References
- 1.↑
Tsao CW, Aday AW, Almarzooq ZI, et al. Heart disease and stroke statistics—2023 update: a report from the American Heart Association. Circulation. 2023;10:1123. https://www.ahajournals.org/doi/10.1161/CIR.0000000000001123
- 2.↑
Bull FC, Al-Ansari SS, Biddle S, et al. World Health Organization 2020 guidelines on physical activity and sedentary behaviour. Br J Sports Med. 2020;54(24):1451–1462. doi:
- 3.↑
Seguin R, Buchner DM, Liu J, et al. Sedentary behavior and mortality in older women: the women’s health initiative. Am J Prev Med. 2014;46(2):122–135. doi:
- 4.↑
Ballin M, Nordström P, Niklasson J, Nordström A. Associations of objectively measured physical activity and sedentary time with the risk of stroke, myocardial infarction or all-cause mortality in 70-year-old men and women: a prospective cohort study. Sport Med. 2021;51(2):339–349. doi:
- 5.↑
Bellettiere J, LaMonte MJ, Rillamas-Sun E, et al. Sedentary behavior increases risk for cardiovascular disease in older women: The Objective Physical Activity and Cardiovascular Health (OPACH) study. Circulation. 2018;137:1036–1046. doi:
- 6.↑
Dempsey PC, Larsen RN, Dunstan DW, Owen N, Kingwell BA. Sitting less and moving more implications for hypertension. Hypertension. 2018;72(5):1037–1046. doi:
- 7.↑
Cabanas-Sánchez V, Guallar-Castillón P, Higueras-Fresnillo S, Rodríguez-Artalejo F, Martínez-Gómez D. Changes in sitting time and cardiovascular mortality in older adults. Am J Prev Med. 2018;54(3):419–422. doi:
- 8.↑
Celis-Morales CA, Lyall DM, Steell L, et al. Associations of discretionary screen time with mortality, cardiovascular disease and cancer are attenuated by strength, fitness and physical activity: findings from the UK Biobank study. BMC Med. 2018;16(1):1063. doi:
- 9.↑
Bellettiere J, Lamonte MJ, Evenson KR, et al. Sedentary behavior and cardiovascular disease in older women: the OPACH study. Circulation. 2019;139(8):1036–1046. doi:
- 10.↑
Stamatakis E, Gale J, Bauman A, Ekelund U, Hamer M, Ding D. Sitting time, physical activity, and risk of mortality in adults. J Am Coll Cardiol. 2019;73(16):2062–2072. doi:
- 11.↑
Li S, Lear SA, Rangarajan S, et al. Association of sitting time with mortality and cardiovascular events in high-income, middle-income, and low-income countries. JAMA Cardiol. 2022;7(8):796–807. doi:
- 12.↑
Paudel S, Ahmadi M, Phongsavan P, Hamer M, Stamatakis E. Do associations of physical activity and sedentary behaviour with cardiovascular disease and mortality differ across socioeconomic groups? A prospective analysis of device-measured and self-reported UK Biobank data. Br J Sports Med. 2023;57:921–929. doi:
- 13.↑
Wu JJ, Yang LL, Jing Y, Ran LL, Xu YQ, Zhou N. Sedentary time and its association with risk of cardiovascular diseases in adults: an updated systematic review and meta-analysis of observational studies. BMC Public Health. 2022;22:286. doi:
- 14.
Hermelink R, Leitzmann MF, Markozannes G, et al. Sedentary behavior and cancer-an umbrella review and meta-analysis. Eur J Epidemiol. 2022;37(5):447–460. doi:
- 15.↑
Pandey A, Salahuddin U, Garg S, et al. Continuous dose-response association between sedentary time and risk for cardiovascular disease a meta-analysis. JAMA Cardiol. 2016;1(5):575–583. doi:
- 16.↑
Wahid A, Manek N, Nichols M, et al. Quantifying the association between physical activity and cardiovascular disease and diabetes: a systematic review and meta-analysis. J Am Heart Assoc. 2016;5:495. doi:
- 17.↑
Ekelund U, Brown WJ, Steene-Johannessen J, et al. Do the associations of sedentary behaviour with cardiovascular disease mortality and cancer mortality differ by physical activity level? A systematic review and harmonised meta-analysis of data from 850 060 participants. Br J Sports Med. 2019;53(14):886–894. doi:
- 18.↑
Ekelund U, Tarp J, Fagerland MW, et al. Joint associations of accelero-meter measured physical activity and sedentary time with all-cause mortality: a harmonised meta-analysis in more than 44 000 middle-aged and older individuals. Br J Sports Med. 2020;54(24):1499–1506. doi:
- 19.↑
Ekelund U, Tarp J, Steene-Johannessen J, et al. Dose-response associations between accelerometry measured physical activity and sedentary time and all cause mortality: systematic review and harmonised meta-analysis. BMJ. 2019;366:570. doi:
- 20.↑
Ekelund U, Steene-Johannessen J, Brown WJ, et al. Does physical activity attenuate, or even eliminate, the detrimental association of sitting time with mortality? A harmonised meta-analysis of data from more than 1 million men and women. Lancet. 2016;388(10051):1302–1310. doi:
- 21.↑
Stamatakis E, Ekelund U, Ding D, Hamer M, Bauman AE, Lee I-M. Is the time right for quantitative public health guidelines on sitting? A narrative review of sedentary behaviour research paradigms and findings. Br J Sports Med. 2019;53(6):377–382. doi:
- 22.↑
Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.
- 23.↑
Ouzzani M, Hammady H, Fedorowicz Z, Elmagarmid A. Rayyan—a web and mobile app for systematic reviews. Syst Rev. 2016;5(1):384. doi:
- 24.↑
Holtermann A, Schnohr P, Nordestgaard BG, Marott JL. The physical activity paradox in cardiovascular disease and all-cause mortality: the contemporary Copenhagen general population Study with 104 046 adults. Eur Heart J. 2021;42(15):1499–1511. doi:
- 25.↑
Warren TY, Barry V, Hooker SP, Sui X, Church TS, Blair SN. Sedentary behaviors increase risk of cardiovascular disease mortality in men. Med Sci Sports Exerc. 2010;42(5):879–885. doi:
- 26.↑
ROBINS-E Development Group, Higgins J, Morgan R, et al. Risk of bias in non-randomized studies-of exposure (ROBINS-E). https://www.riskofbias.info/welcome/robins-e-tool
- 27.↑
Katzmarzyk PT, Church TS, Craig CL, Bouchard C. Sitting time and mortality from all causes, cardiovascular disease, and cancer. Med Sci Sport Exerc. 2009;41(5):998–1005. doi:
- 28.↑
Wijndaele K, Brage S, Besson H, et al. Television viewing and incident cardiovascular disease: Prospective associations and mediation analysis in the EPIC norfolk study. PLoS One. 2011;6(5):e20058. doi:
- 29.↑
Matthews CE, George SM, Moore SC, et al. Amount of time spent in sedentary behaviors and cause-specific mortality in US adults. Am J Clin Nutr. 2012;95(2):437–445. doi:
- 30.↑
Herber-Gast GCM, Jackson CA, Mishra GD, Brown WJ. Self-reported sitting time is not associated with incidence of cardiovascular disease in a population-based cohort of mid-aged women. Int J Behav Nutr Phys Act. 2013;10(1):2–6. doi:
- 31.↑
Chomistek AK, Manson JE, Stefanick ML, et al. Relationship of sedentary behavior and physical activity to incident cardiovascular disease: results from the women’s health initiative. J Am Coll Cardiol. 2013;61(23):2346–2354. doi:
- 32.↑
Young DR, Reynolds K, Sidell M, et al. Effects of physical activity and sedentary time on the risk of heart failure. Circ Hear Fail. 2014;7(1):21–27. doi:
- 33.↑
Petersen CB, Bauman A, Grønbæk M, Helge JW, Thygesen LC, Tolstrup JS. Total sitting time and risk of myocardial infarction, coronary heart disease and all-cause mortality in a prospective cohort of Danish adults. Int J Behav Nutr Phys Act. 2014;11(1):13–11. doi:
- 34.↑
Matthews CE, Cohen SS, Fowke JH, et al. Physical activity, sedentary behavior, and cause-specific mortality in black and white adults in the Southern community cohort study. Am J Epidemiol. 2014;180(4):394–405. doi:
- 35.↑
Ensrud KE, Blackwell TL, Cauley JA, et al. Objective measures of activity level and mortality in older men. J Am Geriatr Soc. 2014;62(11):2079–2087. doi:
- 36.↑
Holme I, Anderssen SA. Increases in physical activity is as important as smoking cessation for reduction in total mortality in elderly men: 12 years of follow-up of the Oslo II study. Br J Sports Med. 2015;49(11):743–748. doi:
- 37.↑
Chau JY, Grunseit A, Midthjell K, et al. Sedentary behaviour and risk of mortality from all-causes and cardiometabolic diseases in adults: evidence from the HUNT3 population cohort. Br J Sports Med. 2015;49(11):737–742. doi:
- 38.↑
Allesøe K, Holtermann A, Aadahl M, Thomsen JF, Hundrup YA, Søgaard K. High occupational physical activity and risk of ischaemic heart disease in women: the interplay with physical activity during leisure time. Eur J Prev Cardiol. 2015;22(12):1601–1608. doi:
- 39.↑
Lee JY, Ryu S, Cheong ES, Sung KC. Association of physical activity and inflammation with all-cause, cardiovascular-related, and cancer-related mortality. Mayo Clin Proc. 2016;91(12):1706–1716. doi:
- 40.↑
Johnsen AM, Alfredsson L, Knutsson A, Westerholm PJM, Fransson EI. Association between occupational physical activity and myocardial infarction: a prospective cohort study. BMJ Open. 2016;6(10):692. doi:
- 41.↑
Evenson KR, Wen F, Herring AH. Associations of accelerometry-assessed and self-reported physical activity and sedentary behavior with all-cause and cardiovascular mortality among US adults. Am J Epidemiol. 2016;184(9):621–632. doi:
- 42.↑
Doukky R, Mangla A, Ibrahim Z, et al. Impact of physical inactivity on mortality in patients with heart failure. Am J Cardiol. 2016;117(7):1135–1143. doi:
- 43.↑
Schnohr P, O’Keefe JH, Lange P, Jensen GB, Marott JL. Impact of persistence and non-persistence in leisure time physical activity on coronary heart disease and all-cause mortality: the Copenhagen city heart study. Eur J Prev Cardiol. 2017;24(15):1615–1623. doi:
- 44.↑
Bennett DA, Du H, Clarke R, et al. Association of physical activity with risk of major cardiovascular diseases in Chinese men and women. JAMA Cardiol. 2017;2(12):1349–1358. doi:
- 45.↑
Dohrn I-M, 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:
- 46.↑
Dohrn IM, Sjöström M, Kwak L, Oja P, Hagströmer M. Accelerometer-measured sedentary time and physical activity—a 15 year follow-up of mortality in a Swedish population-based cohort. J Sci Med Sport. 2018;21(7):702–707. doi:
- 47.↑
Hupin D, Raffin J, Barth N, et al. Even a previous light-active physical activity at work still reduces late myocardial infarction and stroke in retired adults aged >65 years by 32%: the PROOF cohort study. Front Public Health. 2019;7:51. https://www.frontiersin.org/article/10.3389/fpubh.2019.00051/full
- 48.↑
Dohrn IM, Welmer AK, Hagströmer M. Accelerometry-assessed physical activity and sedentary time and associations with chronic disease and hospital visits—a prospective cohort study with 15 years follow-up. Int J Behav Nutr Phys Act. 2019;16(1):878. doi:
- 49.↑
Cuthbertson CC, Tan X, Heiss G, et al. Associations of leisure-time physical activity and television viewing with life expectancy free of nonfatal cardiovascular disease: the ARIC study. J Am Heart Assoc. 2019;8:e012657.
- 50.↑
Liu Q, Liu F, Li J, et al. Sedentary behavior and risk of incident cardiovascular disease among Chinese adults. Sci Bull. 2020;65(20):1760–1766. doi:
- 51.↑
LaMonte MJ, Larson JC, Manson JE, et al. Association of sedentary time and incident heart failure hospitalization in postmenopausal women. Circ Hear Fail. 2020;13:508. https://www.ahajournals.org/doi/10.1161/CIRCHEARTFAILURE.120.007508
- 52.↑
Hamer M, Ding D, Chau J, Duncan MJ, Stamatakis E. Association between TV viewing and heart disease mortality: observational study using negative control outcome. J Epidemiol Community Health. 2020;74(4):391–394. doi:
- 53.↑
Yerramalla MS, McGregor DE, van Hees VT, et al. Association of daily composition of physical activity and sedentary behaviour with incidence of cardiovascular disease in older adults. Int J Behav Nutr Phys Act. 2021;18(1):83. doi:
- 54.↑
Kim D, Murag S, Cholankeril G, et al. Physical activity, measured objectively, is associated with lower mortality in patients with nonalcoholic fatty liver disease. Clin Gastroenterol Hepatol. 2021;19(6):1240–1247.e5. doi:
- 55.↑
Joundi RA, Patten SB, Williams JVA, Smith EE. Association between Excess leisure sedentary time and risk of stroke in young individuals. Stroke. 2021;52(11):3562–3568. doi:
- 56.↑
Gardner AW, Addison O, Katzel LI, et al. Association between physical activity and mortality in patients with claudication. Med Sci Sport Exerc. 2021;53(4):732–739. doi:
- 57.↑
Ekblom-Bak E, Halldin M, Vikstrom M, et al. Physical activity attenuates cardiovascular risk and mortality in men and women with and without the metabolic syndrome—a 20-year follow-up of a population-based cohort of 60-year-olds. Eur J Prev Cardiol. 2021;28(12):1376–1385. doi:
- 58.↑
Yerramalla MS, Van Hees VT, Chen M, Fayosse A, Chastin SFM, Sabia S. Objectively measured total sedentary time and pattern of sedentary accumulation in older adults: associations with incident cardiovascular disease and all-cause mortality. J Gerontol A Biol Sci Med Sci. 2022;77:842–850. https://www.scopus.com/inward/record.uri?eid=2-s2.0-85128522403&doi=10.1093%2Fgerona%2Fglac023&partnerID=40&md5=2f11d0209de7f10915b6592fdf8e55fd
- 59.↑
Jung J, Lee J, Bae E, et al. Association between behavioral patterns and mortality among US adults: national health and nutrition examination survey, 2007–2014. PLoS One. 2022;17:480. https://www.scopus.com/inward/record.uri?eid=2-s2.0-85124919957&doi=10.1371%2Fjournal.pone.0264213&partnerID=40&md5=d8319f766e8be790c569338ebd5a480d
- 60.↑
Hooker SP, Diaz KM, Blair SN, et al. Association of accelerometer-measured sedentary time and physical activity with risk of stroke among US adults. JAMA Netw Open. 2022;5(6):e2215385. doi:
- 61.↑
Dempsey PC, Strain T, Winkler EAH, et al. Association of accelerometer-measured sedentary accumulation patterns with incident cardiovascular disease, cancer, and all-cause mortality. J Am Heart Assoc. 2022;11(9):845. doi:
- 62.↑
Lamonte MJ, Larson JC, Manson JAE, et al. Association of sedentary time and incident heart failure hospitalization in postmenopausal women. Circ Hear Fail. 2020;13:E007508. doi:
- 63.↑
Prince SA, Cardilli L, Reed JL, et al. A comparison of self-reported and device measured sedentary behaviour in adults: a systematic review and meta-analysis. Int J Behav Nutr Phys Act. 2020;17:31. doi:
- 64.↑
Bakker EA, Hartman YAW, Hopman MTE, et al. Validity and reliability of subjective methods to assess sedentary behaviour in adults: a systematic review and meta-analysis. Int J Behav Nutr Phys Act. 2020;17(1):75. doi:
- 65.↑
Stamatakis E, Koster A, Hamer M, et al. Emerging collaborative research platforms for the next generation of physical activity, sleep and exercise medicine guidelines: The Prospective Physical Activity, Sitting, and Sleep consortium (ProPASS). Br J Sports Med. 2020;54(8):435–437. doi:
- 66.↑
Wilmot EG, Edwardson CL, Achana FA, et al. Sedentary time in adults and the association with diabetes, cardiovascular disease and death: systematic review and meta-analysis. Diabetologia. 2012;55(11):2895–2905. doi:
- 67.↑
Jingjie W, Yang L, Jing Y, Ran L, Yiqing X, Zhou N. Sedentary time and its association with risk of cardiovascular diseases in adults: an updated systematic review and meta-analysis of observational studies. BMC Public Health. 2022;22(1):728. doi:
- 68.↑
Garcia L, Pearce M, Abbas A, et al. Non-occupational physical activity and risk of cardiovascular disease, cancer and mortality outcomes: a dose-response meta-analysis of large prospective studies. Br J Sports Med. 2023;57(15):979–989. doi:
