Figures
Abstract
Population-based evidence for the role of habitual physical activity (PA) in the accumulation of visceral (VAT) and subcutaneous (SAAT) abdominal adipose tissue is limited. We investigated if usual patterns and types of self-reported PA and inactivity were associated with VAT and SAAT in a general white population. Total volumes of VAT and SAAT were quantified by magnetic resonance imaging in 583 men and women (61 ± 11.9 y; BMI 27.2 ± 4.4 kg/m2). Past-year PA and inactivity were self-reported by questionnaire. Exploratory activity patterns (APAT) were derived by principal components analysis. Cross-sectional associations between individual activities, total PA in terms of metabolic equivalents (PA MET), or overall APAT and either VAT or SAAT were analyzed by multivariable-adjusted robust or generalized linear regression models. Whereas vigorous-intensity PA (VPA) was negatively associated with both VAT and SAAT, associations between total PA MET, moderate-intensity PA (MPA), or inactivity and VAT and/or SAAT depended on sex. There was also evidence of a threshold effect in some of these relationships. Total PA MET was more strongly associated with VAT in men (B = -3.3 ± 1.4; P = 0.02) than women (B = -2.1 ± 1.1; P = 0.07), but was more strongly associated with SAAT in women (B = -5.7 ± 2.5; P = 0.05) than men (B = -1.7 ± 1.6; P = 0.3). Men (-1.52 dm3 or -1.89 dm3) and women (-1.15 dm3 or -2.61 dm3) in the highest (>6.8 h/wk VPA) or second (4.0–6.8 h/wk VPA) tertile of an APAT rich in VPA, had lower VAT and SAAT, respectively, than those in the lowest (<4.0 h/wk VPA) tertile (P ≤ 0.016; Ptrend ≤ 0.0005). They also had lower VAT and SAAT than those with APAT rich in MPA and/or inactivity only. In conclusion, our results suggest that in white populations, habitual APAT rich in MPA might be insufficient to impact on accumulation of VAT or SAAT. APAT including ≥4.0–6.8 h/wk VPA, by contrast, are more strongly associated with lower VAT and SAAT.
Citation: Fischer K, Rüttgers D, Müller H-P, Jacobs G, Kassubek J, Lieb W, et al. (2015) Association of Habitual Patterns and Types of Physical Activity and Inactivity with MRI-Determined Total Volumes of Visceral and Subcutaneous Abdominal Adipose Tissue in a General White Population. PLoS ONE 10(11): e0143925. https://doi.org/10.1371/journal.pone.0143925
Editor: Yvonne Böttcher, University of Leipzig, GERMANY
Received: June 13, 2015; Accepted: November 11, 2015; Published: November 30, 2015
Copyright: © 2015 Fischer et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: The work was supported by grants (EXC306, EXC306/2) from the Deutsche Forschungsgemeinschaft (http://www.dfg.de) Excellence Cluster “Inflammation at Interfaces” and through grants (01GR0468) from the German Federal Ministry of Education and Research (http://bmbf.de). The PopGen 2.0 network is supported by a grant (01EY1103) from the German Federal Ministry for Education and Research (http://bmbf.de). (Note to editors: All these grants were not recieved by a single author, but by the PopGen biobank). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Accumulation of abdominal adipose tissue (AAT) has been considered a consequence of modern lifestyle patterns, including habitual physical activity (PA) and inactivity (IA) [1, 2]. Both excess visceral (VAT) and subcutaneous (SAAT) AAT have recently been claimed to play a role in cardiometabolic disease etiology [1–3]. Whereas VAT has been shown to contribute to systemic inflammation [1, 4] and has been suggested to be the key causal driver of the cardiovascular risk associated with the metabolic syndrome [1, 4], the impact of SAAT on chronic diseases is less well understood [3–5].
A number of epidemiological studies have observed associations between PA [6–8] and/or IA [9–11] and simple anthropometric measures, such as waist circumference and waist-to-hip ratio as surrogate measures for total AAT. These measures correlate well with direct measurements of total AAT [12], but are unable to distinguish between VAT and SAAT [13, 14]. In contrast, specific amounts of VAT and SAAT can be directly measured by non-invasive imaging techniques such as magnetic resonance imaging (MRI) and computed tomography (CT) [15]. However, as the availability of imaging devices for the use in large-scale studies is limited [15], only a few population-based studies have used these techniques. Moreover, studies that did use imaging techniques usually applied single- or few-slice MRI or CT scans [16–18], which, as opposed to volumetric multi-slice or whole-body scans, do not cover the whole abdomen and, therefore, provide less accurate estimates of total volumes of VAT and SAAT [15, 19, 20].
Apart from several intervention studies [21–25] or population-based studies in specific population groups including women [26–32], men [33], adolescents [34], or children [35], to our knowledge only four studies on PA and directly measured VAT and/or SAAT were conducted in an adult population of both sexes using either single-slice CT [16, 17] or few- [18] and multi-slice [36] MRI, but not whole-abdomen scans. Moreover, whereas intervention studies have demonstrated that both the volume [25, 37] and intensity [25] of exercise or even light PA [38] can reduce VAT and SAAT in the short term, epidemiological evidence for the role of habitual activity patterns (APAT) or individual types of PA and IA in long-term accumulation of VAT and SAAT is limited. Importantly, previous studies have not investigated APAT in relation to AAT that integrate the interrelations and synergy effects of multiple aspects of PA and IA [39]. The aim of the present study was to explore if, and to what extent, overall APAT and individual types of habitual PA and IA were associated with MRI-determined total volumes of VAT and SAAT in a general white population. Moreover, we investigated whether these associations were independent of overall obesity, or modified by sex, age, and body mass index (BMI).
Materials and Methods
Participants and study design
Participants (n = 583; 59% men; 61 ± 11.9 y; BMI 27.2 ± 4.4 kg/m2) aged 25–83 years were recruited from the PopGen control cohort, which is a sample of 1316 individuals of the general population of Kiel, Northern Germany, recruited into the PopGen biobank as controls for case-control studies [40, 41]. Of the random sample of 23,000 local residents who were identified through official population registries and invited to participate in the study, 4,267 individuals (19%) agreed to be part of the PopGen control population. There were no specific characteristics of the non-responders in comparison to responders and the wider background population. Of the 4,267 individuals comprising the PopGen control population, 747 participants (18%) also agreed to further participate in the follow-up of the newly established control cohort. Together with 569 blood donors recruited by the University Medical Center Schleswig-Holstein in Kiel, Germany, these two groups constituted the 1316 individuals of the final PopGen control cohort. Baseline examinations of the cohort were carried out from 2005 to 2007 and comprised a questionnaire with sociodemographic and sociocultural questions as well as questions on medical history, family history of disease, prevalent cardiometabolic and other chronic diseases, use of medication and drugs, and lifestyle factors such as smoking status and alcohol consumption, anthropometric and medical examinations, blood sampling, and analyses of a range of biomarkers [41]. Between 2010 and 2012, a total of 930 cohort participants agreed to take part in the first follow-up that included questionnaires on medical history and demographic, lifestyle, and health-related characteristics such as diet, PA and IA, as well as physical examinations and bio sampling [41]. Of these subjects, 653 individuals with complete PA, IA, and dietary data agreed to undergo whole-body MRI to measure body fat distribution. From these, subjects who may have over- (>16.7 MJ [4000 kcal]/d women; >17.6 MJ [4200 kcal]/d men) or under- (<2.5 MJ [600 kcal]/d) reported their energy intakes [42] (n = 9) as well as subjects with incomplete data on MRI variables (n = 58) or any covariate of interest (n = 5) were excluded. After exclusions, a total of 583 individuals were included in the present cross-sectional study.
Ethics statement
The study protocol was in accordance with the standards for the use of human subjects in research as outlined in the Declaration of Helsinki and was approved by the ethics committee of the Medical Faculty of the University of Kiel, Germany. All participants gave their written informed consent to the study.
Assessment of habitual PA and IA
Habitual PA and IA were assessed using questions from a short, self-administered questionnaire used in the European Prospective Investigation into Cancer and Nutrition (EPIC) study [43] to estimate the average number of hours per week spent on common individual PA and IA over the past year. Contrary to the original questionnaire that was validated against repeated objective measures of fitness and energy expenditure [43], in the version of the questionnaire used in the present study, individual PA and IA asked were not specifically divided into occupational and recreational activities and the questionnaire was not revalidated. This simplification was done to reduce the burden of the participants who had to complete a number of different questionnaires during the first follow-up examination. Participants were asked to report the duration and frequency of time spent walking, cycling, engaging in sports, and gardening (all separately for summer and winter seasons, of which the average was calculated for analysis) as well as on household work and manual/do-it-yourself (DIY) work. In addition, subjects reported the number of flights of stairs climbed per day. The duration of stair climbing per week was calculated with the assumption of 20 steps/flight and that, on average, 72 steps/min were climbed [44]. In addition, participants were asked to report how many hours per day they spent watching TV and sleeping (average of sleep during the day and at night). Total PA reported (h/d) was calculated as the sum of reported time spent on housework, walking, gardening, DIY work, stair climbing, sports, and cycling. Total IA reported (h/d) was calculated as the sum of reported time spent sleeping and watching TV. Time that was not reported was assumed to be spent at rest [31]. Assumed overall 24-h IA (h/d) was calculated as the difference between 24 hours and total PA reported.
According to metabolic equivalents of tasks (METs) [45], individual activities were classified into four categories of intensity: 1) IA (<1.5 METs); 2) light-intensity PA (1.6–2.9 METs); 3) moderate-intensity PA (MPA; 3.0–6.0 METs); and 4) vigorous-intensity PA (VPA; >6.0 METs). To express the sum of all PA in terms of total MET-h/wk (PA MET) [45], hours per week spent on each PA were multiplied by the corresponding MET (in parentheses) [45] and summed for all individual PA: i.e. housework (3.5), walking (3.5), gardening (3.8), DIY work (4.5), stair climbing (4.0), sports (6.0), and cycling (6.8). Both sleeping (0.95) and watching TV (1.3) were considered as reported IA and thus were not included in the PA MET score. For a reference adult, one MET is equal to 1 kcal expended per kg body weight per hour of sitting at rest [45].
Assessment of covariates
Assessment of past-year dietary and energy intakes by an evaluated 112-item food-frequency questionnaire (FFQ) was described previously [46]. This FFQ was based on a validated FFQ used in the baseline assessment of the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam study. It was specifically designed for the follow-up of the EPIC-Potsdam study and adapted to yield high response rates and complete FFQ data, and to comprise particularly those food items that most discriminated participants with regard to food and nutrient intake.
Information on subject characteristics including age, sex, and smoking status was collected by self-administered questionnaires. Anthropometric and blood pressure measurements as well as biochemical analyses were described elsewhere [46].
Adipose tissue assessment
Analysis of whole-body MRI data [47] focusing on the abdominal region (from the top of the liver to the femur heads [13]) was performed by ATLAS (Automatic Tissue Labelling Analysis Software; inhouse-developed at the Department of Neurology, University of Ulm, Ulm, Germany) [14] segmenting total volumes of AAT into VAT and SAAT [14, 48] as described previously [46]. In short, after merging individual cross-sectional MRI volumes of defined body parts to a continuous three-dimensional dataset, the segmentation of AAT into VAT and SAAT was performed using the ARTIS (Adapted Rendering for Tissue Intensity Segmentation) algorithm. SAAT was defined as the sum of all AAT voxels underneath the skin layer surrounding the abdomen from the top of the liver to the femur heads. VAT was defined as the sum of all AAT voxels inside the abdominal muscular wall. Liver fat and fat in the intestinal loops as well as minor MRI artifacts (mainly caused by stents or hip implants) were manually excluded from segmented VAT during post-processing, whereas participants with major MRI artifacts were excluded from further analysis. All ATLAS-based analyses were performed by the same observer to minimize interrater variability. The accuracy of the semi-automatically ATLAS determined volumes of VAT and SAAT was validated against manually determined AAT volumes using the image analysis software slice-O-matic (version 4.2, Tomovision, Montreal, Canada) in a subset of 38 participants [49]. For both VAT (r = 0.996) and SAAT (r = 0.996), AAT volumes analyzed by ATLAS and slice-O-matic yielded high intraclass correlations.
Statistical analysis
Statistical analysis was performed using SAS version 9.3 (SAS Institute, Inc., Cary, North Carolina, USA). Differences in subject characteristics between men and women or subjects with low total PA (MET-h/wk < 96.1) and high total PA (MET-h/wk ≥ 96.1) were analyzed using a χ2 test for categorical and Student’s t test for continuous variables. For Student’s t test and linear regression analysis, distributions of continuous variables were examined and, if necessary, transformed to reach or (in the case of several activity variables) approach normality examined by skew, kurtosis, and Kolmogorov-Smirnov test.
Principal components analysis (PCA) with varimax orthogonal rotation on nine activity items (housework, walking, gardening, DIY work, stair climbing, sports, cycling, sleeping, and watching TV) was used to derive independent APAT variables that maximally explain the variation among the nine activity items. Principal components that met both the eigenvalue ≥1.0 criterion and Scree test were chosen to be representative of independent APAT, and the variation in the activity items explained by each APAT was determined.
Because the residuals of several activity variables were not normally distributed, even after transformation, robust multiple linear regression with MM estimation [50] was used to investigate associations between individual activity or APAT variables and VAT and SAAT. All models were adjusted for sex (except sex strata), age (except age strata), total energy intake, and smoking status (never [smoking period <3 mo], former [smoking period ≥3 mo], and current smoker), and either BMI (except BMI strata) (model 1) or height (model 2) to account for overall adiposity [51] or body size, respectively. As BMI was highly correlated with both VAT (r = 0.70) and SAAT (r = 0.80) [46], model 2 was considered the more appropriate and final model. Potential effect modification by sex, median age (<62.0 and ≥62.0 y), or median BMI (<26.7 and ≥26.7 kg/m2) was investigated by including a multiplicative interaction term in the models. Regression coefficients (B) and standard errors (SE) are reported for unstandardized variables. Generalized linear models (ANCOVA) adjusted for age, total energy intake, smoking status, and height (model 2) were used to calculate adjusted least-squares means for VAT and SAAT by sex and non-sex-specific tertile of individual activity or APAT variables. Statistical differences in unadjusted and adjusted mean VAT and SAAT values between tertiles were corrected for multiple comparisons by using the Tukey-Kramer procedure. Linear regression analysis was performed to test for a linear or quadratic trend across the tertiles by treating the median values of each category as continuous variables.
Reported P values are two-sided. P < 0.05 was considered statistically significant unless otherwise indicated. To account for multiple testing of 38 subject characteristics and 14 or 13 activity and APAT variables under possible dependence, the Benjamini-Yekutieli (B-Y) false discovery rate (FDR) [52] yielding significance thresholds of P ≤ 0.012, P ≤ 0.015, and P ≤ 0.016, respectively, was applied (critical P = α/Σ (1/k) where α = 0.05 and k = 38, 14, or 13 tests).
Results
Subject characteristics
Subject characteristics are presented by sex and non-sex-specific median PA MET (Table 1). Considering B-Y FDR significance (P ≤ 0.012), men with higher PA levels (≥96.1 MET) had lower plasma concentrations of C-reactive protein, higher intakes of dietary fiber, spent more hours on all types of reported PA and less hours on assumed overall 24-h IA, and had lower volumes of SAAT and subcutaneous trunk adipose tissue, but not VAT (P = 0.2), than men with lower PA levels (<96.1 MET). By contrast, women with higher PA levels (≥96.1 MET) were older, had a lower ratio of intakes of n-6 to n-3 fatty acids and higher intakes of dietary fiber, spent more hours on all types of PA reported (except for stair climbing; P = 0.04) and less hours on assumed overall 24-h IA, but did not have B-Y FDR significantly lower volumes of SAAT and subcutaneous trunk adipose tissue (both P = 0.026) or VAT (P = 0.3) than women with lower PA levels (<96.1 MET). Subject characteristics were also calculated for tertiles of three APAT (S1 Table) described below.
PA MET and VAT and SAAT
In multivariable analysis adjusted for BMI (model 1) (Table 2), PA MET was negatively associated with VAT, but not SAAT (P = 0.3). Moreover, the association with VAT was modified by sex, with a significant negative association in men, but not women. For both VAT and SAAT, the associations with PA MET were not modified by median age or median BMI.
In the corresponding multivariable analysis adjusted for height (final model 2, Table 2), similar to BMI adjustment, PA MET was negatively associated with VAT, but not SAAT (P = 0.08). Moreover, the associations between PA MET and both VAT and SAAT were not modified by sex, median age, or median BMI. In the respective sex, age, or BMI strata, however, the association with VAT was significant in men, subjects aged ≥62.0 y, and subjects with BMI < 26.7 kg/m2 only, whereas the association with SAAT was significant in women and subjects aged ≥62.0 y only.
Stratified by sex and tertile category of PA MET (S2 Table and Fig 1), in multivariable analysis adjusted for height (final model 2), there was a non B-Y FDR significant inverse linear trend (Ptrend < 0.05) for the association between PA MET and VAT and SAAT in both men and women. However, compared with subjects in the lowest tertile (PA MET < 74.1), volumes of VAT or SAAT, respectively, were only significantly lower among men (-0.61 dm3 [-12.0%] or -0.71 dm3 [-11.2%]) and women (-0.44 dm3 [-14.2%] or -1.56 dm3 [-17.2%]) in the highest tertile (PA MET > 121.9), but not among subjects in the second tertile.
Data (n = 583 subjects) are LSM (± SE) of total volumes of (1) VAT or (2) SAAT from generalized linear models adjusted for age, total energy intake (kcal/d), smoking status (never, former, and current smoker), and height (final model 2) by tertile category and sex of PA MET and PCA-derived activity patterns (for exact data see also S2 Table). Multiple comparisons assessing statistical differences in LSM between tertiles were corrected by using the Tukey-Kramer procedure. Mean values without sharing a common superscript letter (a-c) were statistically different at P < 0.05. APAT 1–3; activity pattern 1 to 3 derived by PCA; B-Y FDR, Benjamini-Yekutieli false discovery rate; F, females; LSM, least-square mean; M, males; PA MET, metabolic equivalent hours per week of total physical activity (including housework, walking, gardening, do-it-yourself work, stair climbing, sports, cycling); PCA, principal components analysis; SAAT, subcutaneous abdominal adipose tissue; VAT, visceral abdominal adipose tissue.
APAT derived by PCA
Three major principal components representative of independent APAT (APAT-1 to APAT-3) were identified by PCA (Fig 2). With regard to activities loading high on a specific pattern, APAT-1 included cycling and sports; APAT-2 encompassed DIY work and gardening; and APAT-3 comprised housework, walking, sleep, and watching TV. Stair climbing loaded similarly high on both APAT-1 and APAT-2 and, therefore, was regarded as representative of both patterns. The variation (%) in activity variables explained by APAT 1–3 was 16.8, 15.4, and 13.3 (overall 45.5), respectively.
Identified by PCA, on the circular plot the first three principal components that met the Scree test and eigenvalue >1.0 criterion are shown by different colored lines (referred to as APAT pattern 1 to 3). For each of the 7 physical activity and 2 inactivity items, the component loadings of the individual 3 patterns are indicated on the circular axis (component loading scores ranging between -0.6 and +0.7). Each component represents an independent activity pattern including all activity items that yielded a component loading ≥0.5 for this pattern. Activity items that did not obtain a component loading ≥0.5 for any of the principal components were assigned to the pattern the component loading of which was highest. Activity items with similarly high loadings for two patterns were regarded as representative of both patterns. The variation (%) in activity variables explained by APAT 1–3 was 16.8, 15.4, and 13.3 (overall 45.5), respectively. PCA, principal components analysis.
Types of reported PA, IA, or APAT and VAT and SAAT
In multivariable analysis adjusted for BMI (model 1) (Table 3), considering B-Y FDR significance (P ≤ 0.015), VAT was negatively associated with time spent cycling and APAT-1, and positively with time spent watching TV and total IA reported. The associations with APAT-1, watching TV, and total IA reported were modified by sex, with APAT-1 revealing a stronger negative association in men than women, and watching TV (P = 0.02) and total IA reported (P = 0.07) showing a stronger non B-Y FDR significant positive association in men than women. Similarly, the association between stair climbing and VAT was modified by sex, with a non B-Y FDR significant negative association in men (P = 0.04), but not women. For SAAT, only effect modification by sex for DIY work was revealed with a non B-Y FDR significant negative association in men (P = 0.02), but not women.
In the corresponding multivariable analysis adjusted for height instead of BMI (final model 2, Table 3), both VAT and SAAT were negatively associated with time spent cycling and APAT-1, and positively with time spent watching TV and total IA reported. VAT was additionally negatively associated with time spent stair climbing and time spent on sports. The associations between time spent watching TV or total IA reported and VAT were modified by sex, though, with a much stronger positive association in men than women. Moreover, the association between DIY work and SAAT was modified by sex with a positive association in women, but not men.
Tertiles of reported IA, PA, or APAT and VAT and SAAT
In multivariable analysis adjusted for height (final model 2), stratified by sex and tertile category of individual activities (S2 Table and Fig 3) and considering B-Y FDR significance (P ≤ 0.016), there was a quadratic trend (Ptrend ≤ 0.002) between time spent on housework and VAT and SAAT among women, but not men. Compared to women spending <3 h/wk on housework, women had lower VAT and SAAT when 3–9 h/wk, but not when >9 h/wk, were spent on housework. As against subjects spending no time on DIY work, women, but not men, had higher SAAT when >1.5 h/wk, but not when ≤1.5 h/wk were spent on DIY work. For gardening, there was a linear trend for the association with SAAT among men (Ptrend = 0.001), but not women. Compared with men spending <0.4 h/wk gardening, men had higher SAAT when >2.8 h/wk, but not when 0.4–2.8 h/wk, were spent gardening. As against subjects spending <3.0 h/wk walking, women, but not men, had lower VAT and SAAT when ≥3.0 h/wk were spent walking. Regarding stair climbing, there was a linear trend for the association with VAT in men (Ptrend = 0.005), but not women. Compared to men climbing <1.0 stairs/d, men who climbed >4 stairs/d, but not those climbing 1–4 stairs/d, had lower VAT. Among both men and women, there was a linear trend for the association of cycling with VAT, and for men also with SAAT (Ptrend ≤ 0.004). Compared with subjects spending <0.4 h/wk cycling, men and women had lower VAT or SAAT when at least ≥0.4 h/wk were spent cycling. Similarly, there was a linear trend of lower VAT with increasing hours spent on sports among men and women, and for men also of lower SAAT (Ptrend ≤ 0.013). As against subjects spending <0.9 h/wk on sports, men and women spending >3.0 h/wk, but not those spending 0.9–3.0 h/wk, on sports had lower VAT, and males also lower SAAT. Compared with subjects spending <7.3 h/d sleeping, men had lower VAT when 7.3–8.2 h/d, but not when >8.2 h/d were spent, whereas women had higher VAT when ≥7.3 h/d were spent sleeping. Among both men and women, there was a linear trend (Ptrend ≤ 0.007) for the association between watching TV and VAT and SAAT. As against subjects spending <2.0 h/d watching TV, men and women had higher VAT or SAAT when >3.0 h/d, but not when 2.0–3.0 h/d, were spent watching TV.
Data (n = 583 subjects) are LSM (± SE) of total volumes of (1) VAT or (2) SAAT from generalized linear models adjusted for age, total energy intake (kcal/d), smoking status (never, former, and current smoker), and height (final model 2) by tertile category and sex of individual physical activity variables (for exact data see also S2 Table). Multiple comparisons assessing statistical differences in LSM between tertiles were corrected by using the Tukey-Kramer procedure. Mean values without sharing a common superscript letter (a-b) were statistically different at P < 0.05. DIY, do-it-yourself; F, females; LSM, least-square mean; M, males; SAAT, subcutaneous abdominal adipose tissue; VAT, visceral abdominal adipose tissue.
For both men and women, there was a highly significant linear trend (Ptrend ≤ 0.0005) of lower VAT and SAAT with increasing scores of APAT-1 (S2 Table and Fig 1), which loaded high on VPA (Fig 2). Regarding the midpoint between the VPA values (h/wk) of two respective APAT-1 tertiles (S1 Table) as the boundary value for VPA between these two tertiles, compared to subjects in the lowest tertile (<4.0 h/wk VPA), men and women in the second tertile (4.0–6.8 h/wk VPA) as well as men (-1.52 dm3 [-26.8%] or -1.89 dm3 [-27.0%]) and women (-1.15 dm3 [-32.5%] or -2.61 dm3 [-27.3%]) in the third tertile (>6.8 h/wk VPA) had lower VAT or SAAT, respectively. However, in both men and women, there was no association of APAT-2 or APAT-3 loading high on MPA or MPA and IA, respectively (Fig 2 and S1 Table), with either VAT or SAAT except for APAT-3 and SAAT in women. Compared to tertiles of total PA MET that quantitatively summarizes different types of PA in terms of energy expenditure, of the APAT that qualitatively group individual PA according to related PA behaviors, particularly tertiles of APAT-1 (explaining most of the variation in PA behaviors) showed a similar pattern of association with both VAT or SAAT (S2 Table and Fig 1). However, for both men and women, respectively, the tertiles of APAT-1 captured more of the variance in VAT (26.8% and 32.5%) and SAAT (27.0% and 27.3%) than the tertiles of PA MET with regard to VAT (14.2% and 14.2%) and SAAT (17.7% and 17.2%).
Overall, about two third of the 17 activity item or pattern variables investigated were significantly related to VAT or SAAT in BMI- or height-adjusted analysis (Table 4). Most results from models including either an individual PA or IA item or an APAT variable containing this item obtained similar results, i.e. individual PA or IA loading high on a APAT found to be significantly associated with VAT or SAAT were also related to the respective AAT when the respective PA or IA was analyzed individually.
Discussion
The present study is the first to report associations of APAT and individual types of PA and IA with total volumes of VAT and SAAT by sex for a general population sample. Whereas habitual VPA were negatively associated with VAT and SAAT in both white men and women, associations for habitual MPA and IA were usually depended on sex and/or a specific threshold or range for a specific activity volume. Moreover, habitual APAT including ≥4.0–6.8 h/wk VPA, but not APAT rich in MPA only, were associated with lower total volumes of VAT and SAAT.
PA intensity and VAT and SAAT
In line with our study, a number of intervention studies have reported that VPA is able to reduce VAT and SAAT in both men and women [21–25]. Few observational studies also found a negative association of VPA [17, 27] or moderate-to-vigorous PA (MVPA) [17, 18] with VAT in adults, and even of sporadic MVPA [18, 32], i.e. bouts lasting <10 min, similar to the negative association between stair climbing and VAT in men in our study. As with our study showing a linear association between VPA and VAT, a meta-analysis of intervention studies found a dose-response relation between aerobic exercise and VAT reduction in healthy obese subjects [53]. With regard to SAAT, however, results of associations with higher-intensity PA are conflicting [17]. Dependent on ethnicity, a negative association [17], but also no association [17, 18] of VPA or MVPA with SAAT have been reported. For MPA and light-intensity PA, associations with VAT and SAAT are also not clear. We found that associations for MPA were dependent on sex. Similarly, whereas two intervention studies in men [37] or both sexes [54] did not find a reduction in VAT and SAAT due to MPA, two cross-sectional studies in women [55] or both sexes [17] found a negative association between MPA and VAT [17, 55], but not SAAT [17]. Likewise, in line with the negative association of walking with VAT and SAAT in women, but not men, in our study, an intervention study including ~73% women reported that light-intensity walking was capable of reducing VAT [38].
Sex and other factors in the association between PA and AAT
In our study, we revealed a stronger negative association of PA MET or higher-intensity PA with VAT in men, whereas with SAAT in women. Similarly, overall PA was stronger negatively associated with waist circumference in men than women in a prospective cohort study [7]. Biologically, a stronger association between MVPA or VPA and VAT in men than women may be explained by an interplay of 1) the presence of higher VAT in men than women [1, 2, 16]; 2) the observation that men usually perform the same MVPA [56] and probably also VPA at higher intensities than do women; and 3) the physiology of visceral adipocytes that are more responsive to exercise-induced lipolytic catecholamines than subcutaneous adipocytes [4, 16]. As with our results, given that men usually carry greater VAT than women, whereas women have more SAAT than men [1, 2, 16], associations between PA and percentage values of AAT may not be directly comparable between men and women. In addition, given that there is a possible sex difference in the actual intensity of PA defined as moderate or vigorous, analyses on PA and health-related outcomes should always be performed separately for men and women. Apart from dependence on sex, inconsistencies in associations between MPA and VAT or SAAT may result from a specific threshold for an activity volume necessary to impact AAT as we found in our study. Such a threshold has also been reported for associations between MPA [38, 53] or PA MET [16] and VAT and/or SAAT. Generally, there seems to be an interplay between activity intensity and activity volume for the impact of PA on AAT [37, 57]. Ethnicity has also been found to influence the relation between PA and VAT or SAAT. Similar to our study, Europeans [17] and white and black women [30], but not Chinese and South Asians [17], have shown a consistent negative linear relationship [30].
When adjusting for BMI to investigate whether associations were independent of overall obesity, in our study the associations were usually weaker compared to height-adjustment, but especially for VAT in part still significant suggesting that PA may be associated with VAT independent of total body fat. Similarly, adjusted for total fat mass, associations between total PA [35] or VPA [58] and AAT were substantially weakened, but as opposed to SAAT still significant for VAT [35].
IA and VAT and SAAT
We did not find any association of assumed overall 24-h IA and APAT rich in IA with both VAT and SAAT. Similarly, in a cross-sectional study, sedentary behavior was not associated with VAT or SAAT [18]. However, it has to be taken into account that our 24-h IA variable assumed that time that was not reported was spent at rest. Because of the high average age of the study population (61.0 ± 11.9 y), it is likely that the majority of the time that was not reported was indeed spent at rest or otherwise on light-intensity or MPA. Still, it cannot be ruled out that some subjects spent a substantial part of the time they did not report also on VPA, e.g. if their leisure or working time included heavy manual activities that were not covered by the questionnaire. Equally to our study showing a marked positive association of reported IA or spending ≥3 h/d watching TV with both VAT and SAAT, other cross-sectional studies reported that subjects with physical disabilities had higher VAT and SAAT [16], and subjects spending >3 h/d watching TV had higher waist circumference [59]. The latter observation was reported to be not related to a reduction in overall leisure-time PA and, therefore, may partially explained by food and beverage consumption during TV viewing [59]. Similar to other studies [60, 61], our study suggests an optimal range for sleep duration of around 8 h/d for men and possibly 7 h/d for women to be associated with lower VAT. This implies that having both inadequate (e.g. due to chronic stress) and too long sleep duration may result in accumulation of VAT compared to adequate sleep duration of 7–8 h/d.
APAT and VAT and SAAT
Given that during moderate but not greater weight loss, preferential loss of VAT relative to SAAT has been reported [22] and accumulation of VAT is associated with greater metabolic disturbances than an increase in SAAT [1, 4], habitual long-term activity patterns resulting in stable, lower weight may be more effective to reduce VAT and improve metabolic health than short-term exercise interventions that result in rapid and more severe loss of weight. Therefore, our finding that an activity pattern including ≥4.0–6.8 h/wk VPA was associated with significantly lower AAT in white men and women may be interpreted as a recommendation for habitual PA being capable of achieving stable, low levels of VAT and SAAT. Likewise, for MPA, weekly expenditure of 1601–2283 kcal (i.e. about 6.4–9.1 h/wk MPA assuming MPA with expenditure of 250 kcal/h) has been identified to prevent excess VAT [55]. Regarding the current WHO recommendations for adults claiming that 2.5–5.0 h/wk MPA or 1.25–2.5 h/wk VPA should be spent to achieve or maintain good health, these recommendations would not be sufficient to prevent excess VAT or ensure a stable, low level of VAT.
Strengths and limitations
Our study combines the strengths of analyzing APAT and different types of PA and IA by sex as well as the accurate assessment of total volumes of AAT. Still, the study has several limitations. First, the cross-sectional design that is prone to reverse causation and does not allow for cause-effect inference. Second, the use of a questionnaire, which is susceptible to measurement error and limited in accuracy, to estimate self-reported activities. Self-reported answers may suffer from a number of biases and articfacts including over- or understatement due to embarrassement to give true answers or recall bias resulting in over- or underestimation because of inaccurate or incomplete recollections by study participants regarding events from the past. Moreover, for individual self-reported PA that are assigned a specific MET there might be differences in the actual intensity of activities between men and women. Third, our analysis only included common activities and did not differentiate subtypes of activities. Moreover, the rather small sample sizes in stratified analysis may have attenuated statistical significance. Lastly, prospective large-scale studies are needed to confirm our results.
Conclusions
Our findings suggest a linear negative association of habitual VPA especially with VAT, but also SAAT, in white men and women. Associations for habitual MPA and IA, however, may depend on sex and/or a threshold or range for a specific activity volume. In white populations, habitual APAT including ≥4.0–6.8 h/wk VPA, but not APAT rich in MPA only, may in the longer term help prevent excess accumulation of VAT and SAAT, which, however, remains to be confirmed in prospective studies. Given that excess VAT is a major cardiometabolic risk factor, current WHO and other national guidelines do not recommend sufficient VPA to have a meaningful beneficial impact on VAT.
Supporting Information
S1 Table. Characteristics of white adult study subjects by tertiles of exploratory activity patterns derived by principal components analysis.
https://doi.org/10.1371/journal.pone.0143925.s001
(PDF)
S2 Table. Total volumes of visceral and subcutaneous abdominal adipose tissue for tertiles of PA MET, individual types of physical activity or inactivity, and activity patterns by sex in Northern German adults.
https://doi.org/10.1371/journal.pone.0143925.s002
(PDF)
Acknowledgments
The authors would like to thank the participants of the PopGen control cohort for their cooperation as well as the staff of the PopGen biobank for their assistance and support in data collection. The popgen 2.0 network is supported by a grant from the German Ministry for Education and Research (01EY1103).
Author Contributions
Conceived and designed the experiments: KF UN. Performed the experiments: DR GJ. Analyzed the data: KF. Contributed reagents/materials/analysis tools: GJ WL HPM JK. Wrote the paper: KF. Provided methodological adaptions of the ATLAS-software: HPM JK. Provided data: WL UN. Quantified adipose tissue variables: DR.
References
- 1. Tchernof A, Despres JP. Pathophysiology of human visceral obesity: an update. Physiol Rev. 2013;93:359–404. pmid:23303913
- 2. Wajchenberg BL. Subcutaneous and visceral adipose tissue: their relation to the metabolic syndrome. Endocr Rev. 2000;21:697–738. pmid:11133069
- 3. Abate N, Chandalia M. Role of subcutaneous adipose tissue in metabolic complications of obesity. Metab Syndr Relat Disord. 2012;10:319–20. pmid:22816652
- 4. Ibrahim MM. Subcutaneous and visceral adipose tissue: structural and functional differences. Obes Rev. 2010;11:11–8. pmid:WOS:000273022000003
- 5. Porter SA, Massaro JM, Hoffmann U, Vasan RS, O'Donnel CJ, Fox CS. Abdominal subcutaneous adipose tissue: a protective fat depot? Diabetes Care. 2009;32:1068–75. pmid:19244087
- 6. Trichopoulou A, Gnardellis C, Lagiou A, Benetou V, Naska A, Trichopoulos D. Physical activity and energy intake selectively predict the waist-to-hip ratio in men but not in women. Am J Clin Nutr. 2001;74:574–8. pmid:11684523
- 7. Ekelund U, Besson H, Luan J, May AM, Sharp SJ, Brage S, et al. Physical activity and gain in abdominal adiposity and body weight: prospective cohort study in 288,498 men and women. Am J Clin Nutr. 2011;93:826–35. Epub 2011/02/25. pmid:21346093
- 8. Steeves JA, Bassett DR Jr., Thompson DL, Fitzhugh EC. Relationships of occupational and non-occupational physical activity to abdominal obesity. Int J Obes (Lond). 2012;36:100–6. Epub 2011/03/24. pmid:21427697
- 9. Czernichow S, Bertrais S, Preziosi P, Galan P, Hercberg S, Oppert JM, et al. Indicators of abdominal adiposity in middle-aged participants of the SU.VI.MAX study: relationships with educational level, smoking status and physical inactivity. Diabetes Metab. 2004;30:153–9. pmid:15223987
- 10. Arsenault BJ, Rana JS, Lemieux I, Despres JP, Kastelein JJ, Boekholdt SM, et al. Physical inactivity, abdominal obesity and risk of coronary heart disease in apparently healthy men and women. Int J Obes (Lond). 2010;34:340–7. pmid:19918249
- 11. Lakerveld J, Dunstan D, Bot S, Salmon J, Dekker J, Nijpels G, et al. Abdominal obesity, TV-viewing time and prospective declines in physical activity. Prev Med. 2011;53:299–302. Epub 2011/08/03. pmid:21807021
- 12. Browning LM, Mugridge O, Dixon AK, Aitken SW, Prentice AM, Jebb SA. Measuring abdominal adipose tissue: comparison of simpler methods with MRI. Obesity facts. 2011;4:9–15. Epub 2011/03/05. pmid:21372606
- 13. Machann J, Thamer C, Schnoedt B, Haap M, Haring HU, Claussen CD, et al. Standardized assessment of whole body adipose tissue topography by MRI. J Magn Reson Imaging. 2005;21:455–62. pmid:ISI:000228029900019
- 14. Muller HP, Raudies F, Unrath A, Neumann H, Ludolph AC, Kassubek J. Quantification of human body fat tissue percentage by MRI. NMR Biomed. 2011;24:17–24. Epub 2010/07/31. pmid:20672389
- 15. Shuster A, Patlas M, Pinthus JH, Mourtzakis M. The clinical importance of visceral adiposity: a critical review of methods for visceral adipose tissue analysis. Brit J Radiol. 2012;85:1–10. pmid:WOS:000298398900007
- 16. Riechman SE, Schoen RE, Weissfeld JL, Thaete FL, Kriska AM. Association of physical activity and visceral adipose tissue in older women and men. Obes Res. 2002;10:1065–73. Epub 2002/10/12. pmid:12376588
- 17. Lesser IA, Yew AC, Mackey DC, Lear SA. A Cross-Sectional Analysis of the Association between Physical Activity and Visceral Adipose Tissue Accumulation in a Multiethnic Cohort. J Obes. 2012;2012:703941. Epub 2012/10/11. pmid:23050128
- 18. McGuire KA, Ross R. Incidental physical activity and sedentary behavior are not associated with abdominal adipose tissue in inactive adults. Obesity (Silver Spring, Md). 2012;20:576–82. Epub 2011/10/01. pmid:21959343
- 19. Shen W, Punyanitya M, Wang ZM, Gallagher D, St-Onge MP, Albu J, et al. Visceral adipose tissue: relations between single-slice areas and total volume. Am J Clin Nutr. 2004;80:271–8. pmid:ISI:000222912800006
- 20. So R, Sasai H, Matsuo T, Tsujimoto T, Eto M, Saotome K, et al. Multiple-slice magnetic resonance imaging can detect visceral adipose tissue reduction more accurately than single-slice imaging. Eur J Clin Nutr. 2012;66:1351–5. pmid:WOS:000312083600013
- 21. Smith SR, Zachwieja JJ. Visceral adipose tissue: a critical review of intervention strategies. Int J Obes. 1999;23:329–35. pmid:WOS:000079597900001
- 22. Chaston TB, Dixon JB. Factors associated with percent change in visceral versus subcutaneous abdominal fat during weight loss: findings from a systematic review. Int J Obes (Lond). 2008;32:619–28. pmid:18180786
- 23. Kay SJ, Fiatarone Singh MA. The influence of physical activity on abdominal fat: a systematic review of the literature. Obes Rev. 2006;7:183–200. Epub 2006/04/25. pmid:16629874
- 24. Ohkawara K, Tanaka S, Miyachi M, Ishikawa-Takata K, Tabata I. A dose-response relation between aerobic exercise and visceral fat reduction: systematic review of clinical trials (vol 32, pg 395, 2008). Int J Obes. 2008;32:395–. pmid:WOS:000253239200025
- 25. Vissers D, Hens W, Taeymans J, Baeyens JP, Poortmans J, Van Gaal L. The effect of exercise on visceral adipose tissue in overweight adults: a systematic review and meta-analysis. PLoS One. 2013;8:e56415. pmid:23409182
- 26. Hunter GR, Kekes-Szabo T, Treuth MS, Williams MJ, Goran M, Pichon C. Intra-abdominal adipose tissue, physical activity and cardiovascular risk in pre- and post-menopausal women. Int J Obes Relat Metab Disord. 1996;20:860–5. pmid:8880355
- 27. Samaras K, Kelly PJ, Chiano MN, Spector TD, Campbell LV. Genetic and environmental influences on total-body and central abdominal fat: the effect of physical activity in female twins. Ann Intern Med. 1999;130:873–82. Epub 1999/06/22. pmid:10375335
- 28. Kanaley JA, Sames C, Swisher L, Swick AG, Ploutz-Snyder LL, Steppan CM, et al. Abdominal fat distribution in pre- and postmenopausal women: The impact of physical activity, age, and menopausal status. Metabolism. 2001;50:976–82. Epub 2001/07/28. pmid:11474488
- 29. Greenfield JR, Samaras K, Jenkins AB, Kelly PJ, Spector TD, Campbell LV. Moderate alcohol consumption, estrogen replacement therapy, and physical activity are associated with increased insulin sensitivity: is abdominal adiposity the mediator? Diabetes Care. 2003;26:2734–40. Epub 2003/09/30. pmid:14514572
- 30. Dugan SA, Everson-Rose SA, Karavolos K, Avery EF, Wesley DE, Powell LH. Physical activity and reduced intra-abdominal fat in midlife African-American and white women. Obesity (Silver Spring, Md). 2010;18:1260–5. Epub 2009/10/31. pmid:19876007
- 31. Hayes L, Pearce MS, Firbank MJ, Walker M, Taylor R, Unwin NC. Do obese but metabolically normal women differ in intra-abdominal fat and physical activity levels from those with the expected metabolic abnormalities? A cross-sectional study. BMC Public Health. 2010;10:723. Epub 2010/11/26. pmid:21106050
- 32. Ayabe M, Kumahara H, Morimura K, Sakane N, Ishii K, Tanaka H. Accumulation of short bouts of non-exercise daily physical activity is associated with lower visceral fat in Japanese female adults. Int J Sports Med. 2013;34:62–7. Epub 2012/08/21. pmid:22903316
- 33. Bisschop CN, Peeters PH, Monninkhof EM, van der Schouw YT, May AM. Associations of visceral fat, physical activity and muscle strength with the metabolic syndrome. Maturitas. 2013;76:139–45. Epub 2013/07/23. pmid:23870830
- 34. Labayen I, Ruiz JR, Ortega FB, Huybrechts I, Rodriguez G, Jimenez-Pavon D, et al. High fat diets are associated with higher abdominal adiposity regardless of physical activity in adolescents; the HELENA study. Clin Nutr. 2013;33:859–66. Epub 2013/11/05. pmid:24182766
- 35. Saelens BE, Seeley RJ, van Schaick K, Donnelly LF, O'Brien KJ. Visceral abdominal fat is correlated with whole-body fat and physical activity among 8-y-old children at risk of obesity. Am J Clin Nutr. 2007;85:46–53. Epub 2007/01/09. pmid:17209176
- 36. Molenaar EA, Massaro JM, Jacques PF, Pou KM, Ellison RC, Hoffmann U, et al. Association of lifestyle factors with abdominal subcutaneous and visceral adiposity: the Framingham Heart Study. Diabetes Care. 2009;32:505–10. Epub 2008/12/17. pmid:19074991
- 37. Sasai H, Katayama Y, Nakata Y, Eto M, Tsujimoto T, Ohkubo H, et al. The effects of vigorous physical activity on intra-abdominal fat levels: a preliminary study of middle-aged Japanese men. Diabetes Res Clin Pract. 2010;88:34–41. Epub 2010/01/16. pmid:20074828
- 38. Herzig KH, Ahola R, Leppaluoto J, Jokelainen J, Jamsa T, Keinanen-Kiukaanniemi S. Light physical activity determined by a motion sensor decreases insulin resistance, improves lipid homeostasis and reduces visceral fat in high-risk subjects: PreDiabEx study RCT. Int J Obes (Lond). 2013;38:1089–96. Epub 2013/11/29. pmid:24285336
- 39. Ocke MC. Evaluation of methodologies for assessing the overall diet: dietary quality scores and dietary pattern analysis. Proc Nutr Soc. 2013;72:191–99. pmid:23360896
- 40. Krawczak M, Nikolaus S, von Eberstein H, Croucher PJ, El Mokhtari NE, Schreiber S. PopGen: population-based recruitment of patients and controls for the analysis of complex genotype-phenotype relationships. Community Genet. 2006;9:55–61. pmid:16490960
- 41. Nothlings U, Krawczak M. [PopGen. A population-based biobank with prospective follow-up of a control group]. Bundesgesundheitsblatt, Gesundheitsforschung, Gesundheitsschutz. 2012;55:831–5. pmid:22736164
- 42. Imamura F, Lichtenstein AH, Dallal GE, Meigs JB, Jacques PF. Generalizability of dietary patterns associated with incidence of type 2 diabetes mellitus. Am J Clin Nutr. 2009;90:1075–83. pmid:19710193
- 43. Wareham NJ, Jakes RW, Rennie KL, Mitchell J, Hennings S, Day NE. Validity and repeatability of the EPIC-Norfolk Physical Activity Questionnaire. Int J Epidemiol. 2002;31:168–74. pmid:11914316
- 44. Suijkerbuijk KP, Van Duijnhoven FJ, Van Gils CH, Van Noord PA, Peeters PH, Friedenreich CM, et al. Physical activity in relation to mammographic density in the dutch prospect-European prospective investigation into cancer and nutrition cohort. Cancer Epidemiol Biomarkers Prev. 2006;15:456–60. pmid:16537701
- 45. Ainsworth BE, Haskell WL, Herrmann SD, Meckes N, Bassett DR Jr., Tudor-Locke C, et al. 2011 Compendium of Physical Activities: a second update of codes and MET values. Med Sci Sports Exerc. 2011;43:1575–81. pmid:21681120
- 46. Fischer K, Moewes D, Koch M, Müller HP, Jacobs G, Kassubek J, et al. MRI-determined total volumes of visceral and subcutaneous abdominal and trunk adipose tissue are differentially and sex-dependently associated with patterns of estimated usual nutrient intake in a northern German population. Am J Clin Nutr. 2015;101:794–807. pmid:25833977
- 47. Bosy-Westphal A, Kossel E, Goele K, Blocker T, Lagerpusch M, Later W, et al. Association of pericardial fat with liver fat and insulin sensitivity after diet-induced weight loss in overweight women. Obesity (Silver Spring, Md). 2010;18:2111–7. Epub 2010/03/13. pmid:20224561
- 48. Lindauer E, Dupuis L, Muller HP, Neumann H, Ludolph AC, Kassubek J. Adipose Tissue Distribution Predicts Survival in Amyotrophic Lateral Sclerosis. PLoS One. 2013;8:e67783. ARTN e67783 pmid:WOS:000321150000098
- 49. Ruttgers D, Fischer K, Koch M, Lieb W, Muller HP, Jacobs G, et al. Association of food consumption with total volumes of visceral and subcutaneous abdominal adipose tissue in a Northern German population. Br J Nutr. 2015:1–12. pmid:26439793
- 50. Yohai VJ. High Breakdown-Point and High-Efficiency Robust Estimates for Regression. Annals of Statistics. 1987;15:642–56. pmid:WOS:A1987J106800011
- 51. McKeown NM, Troy LM, Jacques PF, Hoffmann U, O'Donnell CJ, Fox CS. Whole- and refined-grain intakes are differentially associated with abdominal visceral and subcutaneous adiposity in healthy adults: the Framingham Heart Study. Am J Clin Nutr. 2010;92:1165–71. Epub 2010/10/01. pmid:20881074
- 52. Benjamini Y, Yekutieli D. The control of the false discovery rate in multiple testing under dependency. Ann Stat. 2001;29:1165–88. pmid:WOS:000172838100012
- 53. Ohkawara K, Tanaka S, Miyachi M, Ishikawa-Takata K, Tabata I. A dose-response relation between aerobic exercise and visceral fat reduction: systematic review of clinical trials. Int J Obes. 2007;31:1786–97. pmid:WOS:000251173200003
- 54. Coker RH, Williams RH, Kortebein PM, Sullivan DH, Evans WJ. Influence of exercise intensity on abdominal fat and adiponectin in elderly adults. Metab Syndr Relat Disord. 2009;7:363–8. pmid:19196080
- 55. Pitanga CP, Pitanga FJ, Beck CC, Gabriel RE, Moreira MH. [Level of physical activity in the prevention of excess visceral fat in postmenopausal women: how much is needed?]. Arq Bras Endocrinol Metabol. 2012;56:358–63. Epub 2012/09/20. pmid:22990639
- 56. Tripette J, Murakami H, Ando T, Kawakami R, Tanaka N, Tanaka S, et al. Wii Fit U intensity and enjoyment in adults. BMC Res Notes. 2014;7:567. pmid:25155382
- 57. Slentz CA, Duscha BD, Johnson JL, Ketchum K, Aiken LB, Samsa GP, et al. Effects of the amount of exercise on body weight, body composition, and measures of central obesity: STRRIDE—a randomized controlled study. Arch Intern Med. 2004;164:31–9. pmid:14718319
- 58. Roemmich JN, Clark PA, Walter K, Patrie J, Weltman A, Rogol AD. Pubertal alterations in growth and body composition. V. Energy expenditure, adiposity, and fat distribution. Am J Physiol Endocrinol Metab. 2000;279:E1426–36. pmid:11093932
- 59. Cleland VJ, Schmidt MD, Dwyer T, Venn AJ. Television viewing and abdominal obesity in young adults: is the association mediated by food and beverage consumption during viewing time or reduced leisure-time physical activity? Am J Clin Nutr. 2008;87:1148–55. Epub 2008/05/13. pmid:18469233
- 60. Hairston KG, Bryer-Ash M, Norris JM, Haffner S, Bowden DW, Wagenknecht LE. Sleep duration and five-year abdominal fat accumulation in a minority cohort: the IRAS family study. Sleep. 2010;33:289–95. Epub 2010/03/27. pmid:20337186
- 61. Chaput JP, Bouchard C, Tremblay A. Change in sleep duration and visceral fat accumulation over 6 years in adults. Obesity (Silver Spring, Md). 2014;22:E9–12. Epub 2014/01/15. pmid:24420871