The authors have declared that no competing interests exist.
Conceived and designed the experiments: HHCMS MK NCS. Performed the experiments: BMFMD MAB GvC. Analyzed the data: BMFMD GvC HHCMS MAB MK PPCAM. Wrote the paper: BMFMD GvC HHCMS MAB MK NCS PPCAM.
Epidemiological studies suggest that excessive sitting time is associated with increased health risk, independent of the performance of exercise. We hypothesized that a daily bout of exercise cannot compensate the negative effects of inactivity during the rest of the day on insulin sensitivity and plasma lipids.
Eighteen healthy subjects, age 21±2 year, BMI 22.6±2.6 kgm−2 followed randomly three physical activity regimes for four days. Participants were instructed to sit 14 hr/day (sitting regime); to sit 13 hr/day and to substitute 1 hr of sitting with vigorous exercise 1 hr (exercise regime); to substitute 6 hrs sitting with 4 hr walking and 2 hr standing (minimal intensity physical activity (PA) regime). The sitting and exercise regime had comparable numbers of sitting hours; the exercise and minimal intensity PA regime had the same daily energy expenditure. PA was assessed continuously by an activity monitor (ActivPAL) and a diary. Measurements of insulin sensitivity (oral glucose tolerance test, OGTT) and plasma lipids were performed in the fasting state, the morning after the 4 days of each regime. In the sitting regime, daily energy expenditure was about 500 kcal lower than in both other regimes. Area under the curve for insulin during OGTT was significantly lower after the minimal intensity PA regime compared to both sitting and exercise regimes 6727.3±4329.4 vs 7752.0±3014.4 and 8320.4±5383.7 mU•min/ml, respectively. Triglycerides, non-HDL cholesterol and apolipoprotein B plasma levels improved significantly in the minimal intensity PA regime compared to sitting and showed non-significant trends for improvement compared to exercise.
One hour of daily physical exercise cannot compensate the negative effects of inactivity on insulin level and plasma lipids if the rest of the day is spent sitting. Reducing inactivity by increasing the time spent walking/standing is more effective than one hour of physical exercise, when energy expenditure is kept constant.
Balancing energy intake and expenditure is the current paradigm in promoting lifestyle related health behaviour and is the basis for many physical activity (PA) guidelines
Insulin resistance is thought to play a central role in the development of type 2 diabetes. Several lines of evidence indicate that physical inactivity can lead to skeletal muscle insulin resistance and possibly to lipid abnormalities
Twenty healthy volunteers (students of the Maastricht University, 17 females and 3 males) were recruited via advertisement. To be included in the study, participants had to perform physical exercise less than 1 hr/week, their BMI should be between 20–30 kg/m2 and their age between 18 to 30 years. Exclusion criteria were any drug use (except oral contraceptives); diseases which interfered with physical activities; frequent alcohol use (more than two units/day); fasting triglycerides >3.0 mmol/l and a fasting plasma glucose >6.0 mmol/l. The study complied with the Declaration of Helsinki and was approved by the Local Ethics Committee of Maastricht University Medical Centre; all participants gave written informed consent. The study was registered as NCT01299311 at ClinicalTrials.gov.
The study was performed under free living conditions and all participants were instructed to follow three activity regimes of four days each. A counterbalanced, randomised crossover design was used, in which participants served as their own controls (
Graphical overview of time spent in different activity categories (sleeping, sitting, standing, MVPA cycling and activity (walking)) in the three regimes followed by the participants.
Subjects were instructed to consume the same caloric intake during each regime and to maintain their usual dietary habits during the three activity regimes but dietary intake was not controlled, e.g. by providing meals or food ingredients. Subjects were not restricted in foods consumed except that they were asked to refrain from alcohol. During each activity regime participants kept a food diary in which daily intake was entered and after each activity regime they filled out a questionnaire on changes in health, medication and impact of the study on daily activities.
During the four days of a regime participants wore continuously (24 hours a day) an ActivPAL™ activity monitor (PAL Technologies, Glasgow, Scotland) to quantify daily PA and postural allocation. The monitor was attached to the skin on the anterior aspect of the thigh using Tegaderm (3M™); non-wear was therefore not an issue. Waterproof wrapping of the monitors allowed wearing during water activities such as bathing. This accelerometer-based activity monitor discriminates time spent sitting or lying, standing and active. In addition stepping bouts and activity intensity were determined and energy expenditure was estimated. Validity and reliability of the ActivPAL in assessing activity pattern of free living healthy subjects has been shown previously
In the exercise regime, participants cycled for 1 hr at Maastricht University Medical Centre+ on a bicycle ergometer (Bodyguard cardiocycle 975). To control intensity and energy expenditure of cycling the participants’ heart rate was monitored continuously (Polar, Kempele, Finland). The model by Hiiloskorpi et al.
Based on the diaries the sleeping time was determined. To calculate sitting time the sleeping time was subtracted from the ActivPAL class ‘sitting/lying’. In addition to the posture allocation, the ActivPAL assessed energy expenditure as metabolic equivalents (MET). By multiplying MET-values by estimated basal metabolic rate (BMR, Harris-Benedict equation), estimated energy expenditure as kcal was obtained. For the exercise condition 450 kcal spent cycling was added. Data of posture allocation and energy expenditure were averaged over four days for each regime.
Measurements of insulin sensitivity (oral glucose tolerance test, OGTT) and plasma lipids were performed in the fasting state, the morning after the 4 days of each activity regime at the Clinical and Translational Research Centre facility. The OGTT was chosen as a measure for insulin sensitivity because of its relative simplicity enabling a large number of measurements and its acceptable correlation with the gold standard (i.e. hyperinsulinemic euglycemic clamp). The minimum time between the last bicycle exercise bout during the exercise regime and the OGTT was 16 hours (mean interval 20±2.6hours). An i.v. catheter was placed in an antecubital vein for blood sampling. At baseline blood was sampled for analysis of glucose, insulin, C-peptide, triglycerides, total cholesterol, high- (HDL-C) and low-density-lipoprotein cholesterol (LDL-C), non-HDL cholesterol, apolipoprotein A-I and B (apo A-I and apo B). After ingestion of 75 g of glucose in 250 ml of water, blood samples were drawn for glucose, insulin and C-peptide levels at 15, 30, 45, 60, 90, and 120 minutes.
Blood samples for glucose, total cholesterol, HDL-C, LDL-C, non-HDL cholesterol and triglycerides were determined the same day. Samples for insulin, C-peptide, apo A-I and apo B were stored at −20°C until analysis after the end of the study. Plasma glucose, total cholesterol, HDL-C, triglycerides were colometric analysed on a Synchron LX20 Pro (Beckman Coulter). Insulin was measured with a double antibody radioimmunoassay Auto-Delfia (Perkin Elmer) and C-peptide with a double chemiluminiscent immunometric Immulite 2000 (Siemens). Apo A-I and apo B were nefelometric determined with a BN ProSpec (Siemens). LDL-C was calculated using the Friedewald formula
If in the series of seven OGTT sample points one or two values missed, polynomial regression was used to assess the best fitting second or third degree polynomial through the available sample points. The best fitting polynomial was used to determine the missing sample points. For each of the OGTT measurement intervals, the product of the duration of the interval and the average insulin, glucose and C-peptide level respectively was calculated. The area under the curve for insulin, glucose and C-peptide curves for the 2 hour period of the OGTT was calculated as the sum of these intervals. As a measure of insulin sensitivity, the insulin sensitivity index (ISI) was assessed
All statistical analyses were executed with SPSS (SPSS 18, Chicago, IL, USA). Values are reported as mean±standard deviations. Variables were tested for normality and homogeneity. Repeated measures ANOVA was applied to evaluate the influence of the different regimes on plasma lipids, on areas under the curve (AUC) of insulin, glucose and C-peptide and on ISI. P-values of ≤0.05 were considered statistically significant. If the repeated measures ANOVA revealed a statistically significant effect of the intervention, conditions were pairwise compared using a Least Significant Difference (LSD) test. Since the LSD test does not correct for multiple testing, only p-values less than 0.017 (0.05/3) were considered significant in the pairwise comparison. To test whether changes in insulin sensitivity were associated with adaptations in plasma lipids, Pearson’s correlation coefficients were calculated between changes in triglyceride concentration over the regimes and changes in ISI.
Two subjects (one male, one female) withdrew before completing the protocol. The participants were on average 21 years of age, had a normal BMI with normal plasma lipid and glucose values (
Variables | Means ± SD |
|
18 |
|
21±2 |
|
1.68±0.07 |
|
63.9±7.8 |
|
22.6±2.6 |
4.61±0.31 | |
4.64±0.70 | |
0.89±0.25 | |
1.45±0.34 | |
2.77±0.56 |
n = 17.
The number of hours slept did not differ between the regimes and the study succeeded in manipulating independently inactivity time, walking/standing time and physical exercise (
Sitting regime | Exercise regime | Minimal intensityPA regime | p-value | p sit vs exerc. | p sit vsMIPA | p exerc vs MIPA | |
|
1539(427) | 1477(352) | 1394(292) | 0.136 | |||
|
61.1(14.8) | 59.7(13.5) | 55.6(13.4) | 0.165 | |||
|
54.5(14.7) | 50.2(19.6) | 50.1(12.2) | 0.248 | |||
|
199.0(68.9) | 196.7(48.9) | 180.0(51.2) | 0.227 | |||
|
1934(88) | 2407(100) | 2486(121) | <0.001 | <0.001 | <0.001 | 0.022 |
|
13.6(1.2) | 12.7(1.7) | 7.4(1.3) | <0.001 | 0.002 | <0.001 | <0.001 |
|
0.99(0.50) | 1.08(0.48) | 3.08(0.88) | <0.001 | 0.166 | <0.001 | <0.001 |
|
0.81(0.29) | 1.01(0.26) | 4.85(0.63) | <0.001 | 0.001 | <0.001 | <0.001 |
|
8.58(0.74) | 8.17(1.37) | 8.65(0.93) | 0.200 | |||
|
4324(1485) | 6049(1402) | 27590(3724) | <0.001 | <0.001 | <0.001 | <0.001 |
|
0.90(0.26) | 0.85(0.35) | 0.70(0.23) | 0.007 | 0.326 | 0.002 | 0.029 |
|
4.20(0.67) | 4.11(0.60) | 3.96(0.50) | 0.171 | |||
|
1.26(0.34) | 1.27(0.28) | 1.30(0.30) | 0.686 | |||
|
2.94(0.47) | 2.84(0.57) | 2.65(0.48) | 0.011 | 0.275 | 0.007 | 0.048 |
|
2.53(0.51) | 2.45(0.57) | 2.34(0.49) | 0.094 | |||
|
1.57(0.24) | 1.57(0.21) | 1.55(0.21) | 0.905 | |||
|
0.75(0.12) | 0.70(0.16) | 0.69(0.14) | 0.022 | 0.052 | 0.005 | 0.627 |
|
20.4(8.2) | 22.8(9.9) | 26.3(11.7) | 0.052 | 0.246 | 0.051 | 0.036 |
|
4.6(0.4) | 4.5(0.3) | 4.5(0.4) | 0.681 | |||
|
11.5(9.0) | 9.4(4.4) | 8.5(4.0) | 0.310 | |||
|
715.7(135.7) | 765.8(115.9) | 754.9(141.8) | 0.171 | |||
|
7752.0(3015.4) | 8320.4(5383.7) | 6727.3(4329.4) | 0.005 | 0.841 | 0.010 | 0.002 |
|
217.4(76.6) | 219.2(67.4) | 193.0(63.7) | 0.104 |
Plasma lipids, glucose, insulin and C-peptide levels were assessed in fasting state. Second, third and fourth column contain average values and standard deviations for each of the regimes. The fifth column represents the level of significance for repeated measurements ANOVA. Column six to eight give the statistical significance for pairwise comparisons of the regimes (Least Significant Differences, p-values were not corrected for multiple testing). For pairwise comparing, p-values less than 0.017 were considered significant.
In six of 54 insulin and C-peptide curves and in 7 of 54 glucose curves one or two sample points were missing, these data were inputted using polynomial regression. In one glucose curve three sample points were missing, the remaining data were not used in the analyses.
Insulin levels differed significantly between the regimes, insulin sensitivity index was nearly significant (p = 0.052). The ISI showed a trend for improvement after the minimal intensity PA regime. Pairwise comparison revealed that the AUC for insulin in the OGTT was significantly smaller after the minimal intensity PA regime than after the sitting (p = 0.010) and the exercise regime (p = 0.002;
Average insulin levels for each of the three regimes (blue: sitting, red: exercise, green: minimal intensity PA) during the oral glucose tolerance tests that were performed after each regime (left hand panel) and average area under the curve for each of the three regimes (right hand panel). Area under the curve was in the minimal intensity PA regime significantly smaller than in both other conditions.
Triglycerides (p = 0.007), non-HDL cholesterol (p = 0.011) and apo B concentrations (p = 0.022) were significantly affected by the different regimes; pairwise comparison revealed that, in comparison to the sitting regime, these lipid measures were significantly reduced after the minimal intensity PA regime (with approximately 22%, 10% and 8%, respectively,
Changes over conditions in triglycerides concentration and ISI did not correlate. Pearson’s correlation coefficient for changes in triglycerides and ISI between sitting and minimal intensity PA was −0.113 (p = 0.665); for the changes between sitting and exercise it was −0.388 (p = 0.112).
A sedentary lifestyle has become a major health threat in our affluent society
In the present study subjects were instructed during a run-in phase about the activity pattern and they received daily feedback. Subjects with a sedentary lifestyle were selected; both the ActivPAL data during the run-in phase and the questionnaires obtained at the end of the study suggested that the sitting regime reflected their daily activities. During the sitting regime they took approximately 4300 steps/day; less than 5000 steps/day is considered sedentary
In line with earlier studies, we observed a positive, non-significant effect of physical exercise on triglycerides, non-HDL cholesterol and apo B as well as a (non-significant) 12% increase in insulin sensitivity. Physical exercise is currently seen as one of the cornerstones in the treatment of (sedentary) subjects with the metabolic syndrome and type 2 diabetes. However, MVPA seems to be a bridge too far for many of these subjects, due to lack of motivation, lack of time or physical impairments
Several epidemiological studies suggest that too much inactivity is detrimental for health
To our knowledge this is the first study that separately manipulated sitting time, physical exercise and DEE in healthy, but sedentary subjects and the novel finding was that a 1 hour bout of physical exercise cannot completely compensate for the negative effects of inactivity on insulin, triglycerides, apo B and non-HDL cholesterol levels. From a traditional exercise physiological point of view, the results of this study might appear surprising; walking at a leisurely pace and standing were more effective than a high intensity physical exercise alternative. As argued by Hamilton et al.
Reducing sitting time with approximately 6 hours resulted in this study in a marked 15% reduction in insulin levels and a non-significant 11% reduction in C-peptide levels. The lack of statistically significant differences in C-peptide levels was probably caused by a lack of statistical power due to the variability of the responses to the OGTT, as discussed above. Moreover, in comparison to the sitting regime, triglycerides, non-HDL-cholesterol and apo B levels were 22%, 10% and 8% lower during minimal intensity PA. How, i.e. by which mechanism, inactivity and minimal intensity PA affect insulin sensitivity and plasma lipids remains to be determined. Given the short duration of each (in)activity regime in our study, changes in microvascular perfusion or mitochondrial function seem less likely. The reduction in triglycerides is compatible with a beneficial effect of minimal intensity PA on free fatty acids (FFA) clearance and/or lipid oxidation and impaired lipid oxidation is thought to be one of the fundamental steps in inactivity induced insulin-resistance
In previous exercise studies, the activities during the rest of the day were often not controlled, in the present study we strictly controlled (in)activity behaviour 24 hr/day. However, the duration of the interventions in the present study was relatively short (4 days) and in future studies the effects of the duration of inactivity need to be addressed, preferably also over longer periods. Moreover, more detailed assessment of insulin sensitivity, such as hyperinsulinemic euglycemic clamp techniques, should be used to unravel the underlying mechanisms.
One hour of daily physical exercise cannot compensate for the negative effects of inactivity on insulin sensitivity and plasma lipids if the rest of the day is spent sitting. Reducing inactivity by low intensity activities such as walking at a leisurely pace and standing is more effective than physical exercise in improving these parameters in sedentary subjects. Our study suggests that in addition to health interventions that stress the importance of spending enough energy to maintain a neutral energy balance, a minimal daily amount of non-sitting time should also be promoted.
We acknowledge R. Jeuken for assisting in the data collection. We express our gratitude to all individuals who participated in our study.