Abstract
Objective
This crossover experimental study examined the acute effects of high sugar/low fiber (HSLF) vs. low sugar/high fiber (LSHF) meals on sedentary behavior (SB) and light-plus activity (L+) in minority adolescents with overweight and obesity.
Methods
87 Latino and African American adolescents (mean age = 16.3 ± 1.2 years, mean BMI z-score = 2.02 ± 0.52, 56.8% Latino, 51.1% male) underwent two experimental meal conditions during which they consumed HSLF or LSHF meals. Physical activity and sedentary behavior were measured using accelerometers and blood glucose and insulin were collected every 30 minutes over 5 hours. Mixed models were used to examine the temporal trends of SB and L+, whether the temporal trends of SB and L+ differed by meal condition, and the influence of blood glucose and insulin on the activity behaviors.
Results
SB and L+ fluctuated over time during the HSLF condition, but were stable during the LSHF condition. SB and L+ were influenced by the blood glucose response to the HSLF meals. Insulin did not influence SB or L+ in either meal condition.
Conclusions
Sugar and fiber content of meals can have differing acute impacts on activity behaviors in minority adolescents with overweight and obesity, possibly due to differing metabolic responses.
Keywords: dietary sugar, dietary fiber, sedentary behavior, physical activity, minority adolescents
Introduction
Little is known about how consumption of sugar, a major component of the U.S. diet (1), and consumption of fiber, which is under-consumed in the U.S. diet (2), may influence physical activity and sedentary behavior. Studies have indicated that diets high in simple carbohydrates may be associated with feelings of fatigue and low energy, which could negatively influence levels of physical activity and sedentary behavior (3, 4). Additionally, studies have found correlations between high sugar intake and low physical activity (5, 6). However, little else is known about how activity behaviors are influenced by sugar and fiber consumption.
Adolescents are one population for which elucidating this relationship may be important. Sugar consumption has risen in recent decades among U.S. adolescents (1, 7). Adolescents also tend to have diets low in fiber (1). The typical diets of African American and Latino adolescents are less healthy than the typical diets of adolescents from other ethnic groups. African American and Latino youths have dietary fiber intake from whole grain, fruit and vegetable sources below national dietary guidelines, and sugar intake above guidelines (8, 9). Adolescents from these populations also do not meet physical activity recommendations (10). The age-related decline in physical activity that is exhibited by adolescents (11, 12) is more pronounced in Latino and African American adolescents than youths from other ethnic groups (13). Adolescence is a critical time for establishing health behaviors, as both dietary patterns and physical activity habits in this period of life track into adulthood (14, 15). It is clearly important to understand how dietary intake may influence activity behaviors in these populations.
A previous exploratory pilot study by our research group examined the influence of high sugar and low sugar breakfasts on physical activity in a group of ten Latino adolescent girls with obesity (16). This preliminary in-lab crossover design study found overall differences in physical activity and sedentary behavior between meal conditions (16), but yielded a small subject-level sample size and insufficient prompt-level data to examine temporal changes in activity behaviors. The current study aimed to expand upon the pilot study findings by extending the time period for data collection, increasing the sample size, and examining potential underlying mechanisms for differences in activity behaviors by meal condition.
This randomized crossover study investigated the metabolic and activity responses to high-sugar/low-fiber (HSLF) and low-sugar/high fiber (LSHF) conditions. Our research aim was to examine the dynamics of sedentary behavior (SB) and of light-plus activity (L+) over time in response to meals with differing dietary sugar and fiber profiles, and to determine if the metabolic responses to meals with differing dietary sugar and fiber profiles influenced activity behavior change over time. Therefore, we describe the influence of meal type on temporal changes of SB and L+ over five hours, whether these differed by meal condition, and the possible influence of blood glucose and insulin on the growth curves of these activity behaviors.
Based on findings from previous studies and our pilot study (16, 17, 18), we hypothesized that compared to the LSHF condition, the HSLF condition would be associated with a steeper decrease in L+ and a steeper increase in SB over the course of the in-lab observation period. Additionally, we hypothesized that a more pronounced metabolic response to the HSLF would mean that blood glucose and insulin would have a greater impact on the growth curves of the activity behaviors in the HSLF condition compared to the LSHF condition.
Methods
Participant recruitment
Participants were recruited from the Los Angeles area from 2007-2010. The inclusion criteria were: African American or Latino ethnicity, male or female 14 to 17 years old, and body mass index ≥ 85th percentile for age and sex. The exclusion criteria were: diagnosis of diabetes, participation in a weight loss or exercise program, use of medications that influenced body weight or insulin sensitivity, or diagnosis of a syndrome that influences body composition. Prior to study procedures, informed written parental consent and participant assent were obtained. This study was approved by the Institutional Review Board of the University of Southern California. Participants were provided with tiered compensation ($100 for the first visit, $125 for the second visit).
Experimental design
This study used a randomized crossover design. Two experimental meal conditions were employed in a 5-hour in-lab setting: a high sugar/low fiber (HSLF) condition and a low sugar/high fiber (LSHF) condition. The HSLF and LSHF conditions were conducted during two in-lab visits separated by a 2 to 4 week washout period. Participants underwent both conditions and were randomly assigned to a HSLF/LSHF or a LSHF/HSLF visit order using a stratified block design randomization procedure.
Experimental meal conditions
The HSLF and LSHF meals were developed using data from focus groups conducted with a representative sample of youth. HSLF meals consisted of a Pop-tart (Kellogg NA Co.), calcium-enriched string cheese (Sargento Mootown Light String Cheese, Sargento Foods Inc.), and Tampico juice (Tampico Beverages). LSHF meals consisted of a whole-wheat bagel, margarine (I Can't Believe It's Not Butter Light, Unilever PLC/Unilever N.V.), and water treated with a soluble fiber powder (Benefiber Powder, Novartis Consumer Health, Inc.). Nutrient compositions were determined using the Nutrient Data System for Research (NDS-R 2010). The meals were isocaloric and matched for macronutrients; only dietary fiber and sugar contents varied. Table 1 shows the nutrient contents of each meal. Portions were approximately 20% of the participant's daily caloric need, which was calculated with the Dietary Reference Intake Guidelines Estimation of Energy Expenditure for children ages 3-18 with overweight (19). Each meal was provided for breakfast and lunch. A Registered Dietician (RD) designed the meals and prepared or supervised the preparation of all meals.
Table 1.
Nutrient compositions for HSLF and LSHF breakfast and lunch meals
Macronutrient g (% kcal) | HSLF Meal | LSHF Meal |
---|---|---|
54.0g Poptart | 61.0g whole wheat bagel | |
42.0 string cheese | 14.0g margarine | |
247.0g juice | 10.5g Benefiber supplement | |
Fat | 11.0 (24%) | 9.5 (24%) |
Carbohydrate | 64.0 (61%) | 61.0 (68%) |
Protein | 14.0 (13%) | 10.0 (11%) |
Sugar | 41.0 (39%) | 7.0 (8%) |
Fiber | 1.0 (1%) | 16.0 (18%) |
HSLF: high sugar/low fiber meal condition; LSHF: low sugar/high fiber meal condition; kcal: kilocalories; g: grams
Demographic and baseline measures
Data was collected from 2008-2010. Potential participants attended a screening exam at the Clinical Trials Unit at the USC University Hospital. Eligible participants completed an inpatient visit where insulin sensitivity, body weight, height, and demographic data were collected. Height (cm) and weight (kg) were measured in triplicate by a Registered Nurse. Body mass index (BMI) percentile was calculated based on the CDC age- and sex- specific growth charts (20).
Insulin Sensitivity (SI) was assessed at a baseline visit via three-hour Frequently Sampled Intravenous Glucose Tolerance Test (FSIVGTT) (21). Blood samples were collected at −15, −5, 2, 4, 8, 19, 22, 30, 40, 50 70, 100, and 180 minutes. Glucose (25% dextrose, 0.3 g/kg body weight) was administered intravenously at time 0 and insulin [0.02 units/kg body weight, Humulin R regular insulin for human injection; Eli Lilly] at 20 minutes. SI was calculated using the minimal model from the FSIVGTT results (22) [MINMOD MILLENIUM 2002 computer program, Version 5.16; Richard Bergman, Los Angeles, CA (23)].
In-lab procedures
Participants began each visit after a 10-hour overnight fast. A saline lock intravenous catheter was placed into the forearm for blood sampling. Participants were given 15 minutes to complete the meal, then instructed to choose from activities available in the laboratory for the observation period. Options were based on feedback from the aforementioned focus groups. Active options included Nintendo Wii, treadmill, small trampoline, jump rope, hula-hoops, free weights, Rock Band, and Dance Dance Revolution. Sedentary options included books, movies, arts and crafts center, listening to music, and movies.
In-lab activity behavior measures
Activity behaviors were assessed via accelerometry. Participants wore a uniaxial Actigraph GT1M accelerometer on the right hip using an elastic belt during each visit. The accelerometer collected activity data in 60-second epochs. Accelerometer data was processed using SAS code developed by the National Cancer Institute for use with NHANES data (http://riskfactor.cancer.gov/tools/nhanes_pam). Time spent in activity levels was calculated by summing each minute spent below the cut point for sedentary behavior, and above the cut points for light physical activity and moderate-to-vigorous physical activity. A previously defined and validated cut point of 100 counts per minute was used to define sedentary behavior (24). The thresholds for light intensity physical activity (<4 METs) and moderate or greater intensity physical activity (4+ METs) were age-adjusted using the criteria from Freedson and colleagues (25). Time spent in each activity category was calculated by summing the minutes of counts above or below the appropriate threshold. Minutes in each activity category were summed for each 30-minute interval over 5 hours. Due to a very low amount of moderate-to-vigorous physical activity in this sample, the minutes of light and moderate-to-vigorous physical activity data were summed to create a “light-plus activity” variable.
In-lab metabolic measures
Blood samples were taken at -5 minutes prior to the first meal and every 30 minutes after for a total of 5 hours during each visit. Samples were centrifuged in a microfuge on-site within one hour of blood draw, placed on ice, and transported on dry ice to the on-site lab where they were stored at −70 C until assayed. Blood glucose was analyzed on a Dimension Clinical Chemistry system and an in vitro Hexokinase method (Dade Behring, Deerfield, IL). Insulin was assayed with an automated random access enzyme immunoassay system Tosoh AIA 600 II analyzer (Gibbco Scientific, In. Coon Rapids, MN) using an immunoenzymemetric assay (IEMA) method.
Statistical analyses
Analyses were conducted using SAS v.9.1 (SAS Institute, Cary, NC). Significance levels for statistical tests were set at α = 0.05. Data were analyzed using repeated measures analyses via PROC MIXED with an approach similar to dyadic multilevel modeling (26). This approach was used because the design of the study led to interdependence of outcome measures in each condition within each participant, analogous to the interdependence of actor and partner data in a dyadic study. The dyadic multilevel modeling approach expands standard multilevel modeling by accounting for statistical dependency of data between paired observations (or in this case, between conditions), as well as statistical dependency of data within individuals (27, 28). This approach also yielded simultaneous estimation of growth curves for activity behaviors in each condition separately (26, 28). In order to conduct this type of multilevel analysis, the data was structured so that the meal condition variable was separated into two dichotomous variables, one to indicate the HSLF condition (1 = HSLF, 0 = LSHF) and the other to indicate the LSHF condition (1 = LSHF, 0 = HSLF) (26, 29, 30). SB and L+ outcomes were non-normally distributed. A square transformation was used on SB and square root transformation was used on L+ (31, 32).
First, the ideal functional form (i.e., effect of time) describing the change of SB and L+ over the in-lab observation periods was identified. This approach involved modeling the growth curves for the outcome with time and its higher-order function (without including covariates) for each condition separately in order to find the best fitting functional forms for the change in activity behavior outcomes over time.
Second, multilevel models were run to determine 1) if the growth curves of the activity behavior outcomes differed between condition, and 2) if these growth curves were influenced by blood glucose or insulin measures. In order to determine whether the growth curves differed by condition, meal condition by time interactions were included in the models. An interaction term for metabolic measure (blood glucose or insulin) by time by meal condition was included in the models to examine the influence of metabolic measures on the activity behavior outcomes. A separate model tested for each activity behavior outcome and for each metabolic predictor for a total of four models (the influence of blood glucose on the SB growth curve, the influence of insulin on the SB growth curve, the influence of blood glucose on the L+ growth curve, and the influence of insulin on the L+ growth curve).
Blood glucose and insulin were time-varying predictors that varied both between- and within-person. In order to differentiate the effects of the between- and within-person variance of the metabolic measures on the activity behavior outcomes, the metabolic variables were disaggregated into their respective between- and within-person effects (33). These between- and within-person effects were included as separate variables in each model. A priori time-invariant covariates included age, sex, ethnicity, BMI z-score, baseline insulin sensitivity, and randomization order.
For all models, the variance components were estimated using the restricted maximum likelihood (REML) estimation method. A Kronecker product structure was specified for the residuals. Random effects were included for the intercepts of both the HSLF and LSHF conditions. A no intercept (“NOINT”) option was included in the model statements to suppress the intercept for the aggregate data and allow the estimation of the intercepts for each condition separately (27, 34). Post-hoc contrast statements were added in all models to empirically test whether the estimates for the intercepts and growth curves for each activity behavior outcome differed by condition. Estimates from the model results were used to plot the functional forms of the activity behaviors over time.
Results
Descriptive Statistics
Baseline sample characteristics are displayed in Table 2. A total of 87 participants were included in the study (56.8% Latino, 51.1% male). There was no participant attrition between visits. The mean age of the participants was 16.3 ± 1.2 years and the mean BMI z-score was 2.02 ± 0.52. All participants had overweight or obesity (77.3% had obesity). Preliminary analysis of the growth curves for SB and L+ revealed that a cubic functional form was optimal for describing the change in both SB and L+ over the course of the observation periods. Results from the multilevel models are shown in Table 3 and Table 4. The general level 1 and level 2 equations for the analyses are shown in Table 5.
Table 2.
Sample Characteristics (n = 87)
Variable | Mean (SD) |
---|---|
Age (years) | 16.3 (1.2) |
BMI z-score | 2.02 (0.52) |
Weight status | |
Overweight | 5.7% (5) |
Obese | 94.3% (83) |
Insulin Sensitivity (SI) | 1.6 (1.1) |
Sex1 | |
Male | 48.9% (43) |
Ethnicity1 | |
African American | 43.2% (38) |
Latino | 56.8% (50) |
Mean and (SD) reported, unless otherwise indicated
Frequencies (N)
SD = standard deviation; BMI = body mass index
Table 3.
Multilevel model results for growth curves of sedentary behavior and light-plus activity during HSLF and LSHF meals (blood glucose)a
Sedentary Behavior b | |||
---|---|---|---|
HSLF condition | LSHF condition | Comparison of HSLF vs. LSHF estimates | |
Variable | B (SE) | B (SE) | t-valuec |
Intercept | 1046.23 (148.43) *** | 820.43 (114.48) *** | 1.53 |
Meal*Time | −164.40 (64.26) * | −6.70 (51.99) | −1.92 |
Meal*time2 | 23.83 (8.72) ** | −0.41 (7.38) | 2.13* |
Meal*time3 | −1.05 (0.37) ** | 0.05 (0.32) | −2.24* |
Meal*blood glucosebs | 0.91 (1.25) | 0.52 (1.27) | 0.28 |
Meal*blood glucosews | −12.84 (5.76) * | −6.73 (8.05) | −0.62 |
Meal*time*blood glucosews | 6.67 (2.90) * | 3.43 (3.94) | 0.67 |
Meal*time2*blood glucosews | −0.99 (0.43) * | −0.51 (0.57) | −0.67 |
Meal*time3*blood glucosews | 0.05 (0.02) * | 0.02 (0.03) | −0.67 |
Light-plus Activity d | |||
Intercept | 0.18 (0.74) | 1.07 (0.57) | 1.32 |
Meal*Time | 0.74 (0.32) * | 0.03 (0.26) | 1.73 |
Meal*time2 | −0.11 (0.43) * | 0.004 (0.04) | −1.96 |
Meal*time3 | 0.005 (0.002) * | −0.004 (0.002) | 2.09* |
Meal*blood glucosebs | −0.005 (0.007) | −0.003 (0.01) | −0.26 |
Meal*blood glucosews | 0.06 (0.03) * | 0.03 (0.04) | 0.62 |
Meal*time*blood glucosews | −0.33 (0.01) * | −0.02 (0.02) | −0.64 |
Meal*time2*blood glucosews | 0.005 (0.002) * | 0.003 (0.003) | −0.62 |
Meal*time3*blood glucosews | −0.0002 (0.0001) * | −0.0001 (0.0001) | 0.62 |
models control for age, ethnicity, baseline insulin sensitivity, sex, BMI-z score, and randomization order (estimates not reported)
estimates for square transformed sedentary behavior
t-value from contrast tests comparing estimates between HSLF and LSHF meal conditions
estimates for square root transformed light-plus activity
p<0.05
p<0.001
p<0.0001
SE = standard error; Meal = meal condition variable (HSLF or LSHF); HSLF = high sugar meal condition; LSHF = high fiber meal condition; blood glucosebs: between-subject blood glucose; blood glucosews: within-subject blood glucose; time2: quadratic time trend; time3: cubic time trend
Table 4.
Multilevel model results for growth curves of sedentary behavior and light-plus activity during HSLF and LSHF meals (insulin)a
Sedentary Behavior b | |||
---|---|---|---|
HSLF condition | LSHF condition | Comparison of HSLF vs. LSHF estimates | |
Variable | B (SE) | B (SE) | t-valuec |
Intercept | 987.37 (135.48) *** | 826.57 (101.98)*** | 1.30 |
Meal*Time | −157.12 (59.74) ** | −19.24 (47.82) | −1.81 |
Meal*time2 | 24.47 (8.23) ** | 2.05 (6.94) | 2.09 * |
Meal*time3 | −0.08 (0.31) | −0.12 (0.14) | −2.25 * |
Meal*insulinbs | −0.12 (0.14) | −0.15 (0.14) | 0.23 |
Meal*insulinws | −2.95 (1.18) * | −3.11 (1.59) | 0.08 |
Meal*time * insulinws | 1.66 (0.58) ** | 1.75 (0.79) * | −0.08 |
Meal*time2*insulinws | −0.25 (0.09) ** | −0.27 (0.12) * | 0.15 |
Meal*time3*insulinws | 0.01 (0.004)** | 0.01 (0.01) * | 0.15 |
Light-plus Activity d | |||
Intercept | 0.53 (0.68) | 0.99 (0.51) | −0.95 |
Meal*Time | 0.67 (0.30) * | 0.11 (0.24) | 1.48 |
Meal*time2 | −0.11 (0.04) * | −0.01 (0.03) | −1.78 |
Meal*time3 | 0.005 (0.002) ** | 0.0003 (0.002) | 1.95 |
Meal*insulinbs | 0.001 (0.001) | 0.001 (0.001) | −0.34 |
Meal*insulinws | 0.01 (0.01) * | 0.02 (0.01) * | −0.25 |
Meal*time*insulinws | −0.01 (0.003) * | −0.01 (0.004) * | 0.28 |
Meal*time2*insulinws | 0.001 (0.0004) ** | 0.001 (0.001) * | −0.37 |
Meal*time3*insulinws | −0.00001 (0.00002) * | −0.0002 (0.00003) * | −0.37 |
models control for age, ethnicity, baseline insulin sensitivity, sex, BMI-z score, and randomization order (estimates not reported)
estimates for square transformed sedentary behavior
t-value from contrast tests comparing estimates between HSLF and LSHF meal conditions
estimates for square root transformed light-plus activity
p<0.05
p<0.001
p<0.0001
SE = standard error; Meal = meal condition variable (HSLF or LSHF); HSLF = high sugar meal condition; LSHF = high fiber meal condition; blood glucosebs: between-subject blood glucose; blood glucosews: within-subject blood glucose; time2: quadratic time trend; time3: cubic time trend
Table 5.
General level 1 and level 2 equations for second set of models examining the influence of metabolic measures (blood glucose or insulin) on activity behavior growth curves
Level 1: |
Yij = π0i HSLFij + π1i HSLFij + π2i HSLFij * Timeij + π3i LSHFij * Timeij + π4i HSLFij * Time2ij + π5i LSHFij * Time2ij + π6i HSLFij * Time3ij + π7i LSHFij * Time3ij + π8i HSLFij *Metabolicwsij + π9i LSHFij * Metabolicwsij + π10i HSLFij * Timeij * Metabolicwsij + π11i LSHFij * Timeij * Metabolicwsij + π12i HSLFij * Time2ij * Metabolicwsij + π13i LSHFij * Time2ij * Metabolicwsij + π14i HSLFij * Timeij3* Metabolicwsij + π15i LSHFij * Timeij3* Metabolicwsij + εij |
Level 2: |
π0i = γ00 + γ01Metabolicbsi + γ02Ethnicityi + γ03Sexi + γ04BMI-zi + γ05Agei + γ06SIi + γ07Randomizationi + ζ0j |
π1i = γ10 + γ11Metabolicbsi + γ12Ethnicityi + γ13Sexi + γ14BMI-zi + γ15Agei + γ16SIi + γ17Randomizationi + ζ1j |
Where: |
Yij = sedentary behavior or light-plus activity; HSLF = high sugar/low fiber meal condition; LFHF = low sugar/high fiber meal; Metabolicws = within-subject effect of blood glucose or insulin; Metabolicbsi = between-subject effect of blood glucose or insulin; BMI-z = body mass index z-score; SI = baseline insulin sensitivity; Randomization = randomization order |
Effect of Blood Glucose and Insulin on the Growth Curves of Sedentary Behavior
In the HSLF meal, there were significant linear, quadratic, and cubic effects of time on SB (Table 3 first column and Table 4 first column). This indicated that SB initially decreased, then increased, and subsequently decreased again over the course of the day during the HSLF condition. There were no significant fluctuations of SB when participants were in the LSHF condition (Table 3 second column and Table 4 second column).
In the HSLF condition, the effects of blood glucose on SB varied as a function of time (Table 3 first column). Figure 1 shows the cubic functional form of SB during the HSLF condition for when blood glucose was at one's mean level (centered at overall person mean), as well as when blood glucose was higher and lower than one's mean level. In the HSLF condition, higher than mean blood glucose was initially associated with lower SB, and lower than mean blood glucose was initially associated with higher SB. Between 90 minutes post-breakfast and the start of lunch at 240 minutes, higher than mean blood glucose was associated with higher SB and lower than mean blood glucose was associated with lower SB. After 240 minutes, higher blood glucose was again associated with lower SB and lower blood glucose was again associated with higher SB. Insulin did not impact SB during the HSLF condition.
Figure 1. Influence of blood glucosea and insulinb on sedentary behavior during HSLF meal Condition.
Figure 1 caption for top graph:
a: Higher blood glucose than usual was initially associated with lower sedentary behavior, and lower blood glucose than usual was initially associated with higher sedentary behavior. Between 90 minutes after breakfast and lunch (at 240 minutes), higher blood glucose was associated with higher sedentary behavior, and lower blood glucose was associated with lower sedentary behavior. After lunch, higher blood glucose was again associated with lower sedentary behavior and lower blood glucose was again associated with higher sedentary behavior
Figure 1 caption for bottom graph:
b: Insulin did not have impact sedentary behavior.
Effect of Blood Glucose and Insulin on the Growth Curves of Light-plus Activity
In the HSLF condition, there were significant linear, quadratic, and cubic effects of time on L+ (Tables 3 first column and Table 4 first column). This indicated that L+ initially increased, then decreased, and subsequently increased again over the course of the day during the HSLF condition. There were no significant fluctuations of L+ when participants were in the LSHF condition (Table 3 second column and Table 4 second column).
In the HSLF condition, the effects of blood glucose on L+ varied as a function of time (Table 3 first column). Figure 2 shows the cubic functional form of L+ during the HSLF condition for when blood glucose was at one's mean level, as well as when blood glucose was higher and lower than one's mean level. In the HSLF condition, higher than mean blood glucose was associated with higher L+ during the first 90 minutes, and lower than mean blood glucose was associated with lower L+ during the first 90 minutes. After 90 minutes post-breakfast, higher than mean blood glucose was associated with lower L+, and lower than mean blood glucose was associated with higher L+. Insulin did not impact L+ during the HSLF condition.
Figure 2. Influence of blood glucosea and insulinb on light-plus activity during HSLF meal condition.
Figure 2 caption for top graph:
a: Higher blood glucose than usual was initially associated with higher light-plus activity, and lower blood glucose than usual was initially associated with lower light-plus activity. After 90 minutes post-breakfast, higher than usual blood glucose was associated with lower light-plus activity, and lower than usual blood glucose was associated with higher light-plus activity. Figure 2 caption for bottom graph:
b: Insulin did not impact light-plus activity.
Discussion
This study aimed to examine the acute effects of two meal conditions that differed in sugar and fiber content on temporal trends of sedentary behavior and light-plus activity, and to determine if metabolic responses to these meals influenced activity behaviors. We found that meal type and blood glucose influenced activity in the HSLF condition only. Sedentary behavior and light-plus activity appeared to change subsequent to meal consumption in the HSLF condition. Consumption of simple carbohydrates during the HSLF condition after an overnight fast could have temporarily lead to a relief in feelings of fatigue or may have relieved depletion of skeletal muscle energy, which could have influenced activity behaviors (35). This response could have occurred again after the HSLF lunch meal. While this explanation provides possible support for our findings, studies examining the influence of dietary sugar content on sedentary behavior and physical activity over the long-term are needed to elucidate the broader impact of high-sugar diets on activity behaviors.
Another potential explanation for our findings is that an acute decrease in sedentary behavior and increase in light-plus activity in response to a high-sugar meal could represent a compensatory response by the body to regulate blood glucose levels. Individuals with overweight and obesity tend to have blunted insulin responses, which result in high blood glucose levels in response to high-sugar foods (36). Skeletal muscle activity aids in uptake of blood glucose from the bloodstream (37), so increasing activity may be a way in which the body compensates to regulate glucose levels after consumption of high amounts of sugar in individuals with overweight. This may provide support to the influence of blood glucose on the functional forms of sedentary behavior and light-plus activity levels we found in the HSLF condition. Neuronal control of energy balance may provide another explanation for our findings. Orexin neurons in the brain are glucose sensing, and the neurotransmitter orexin A has the ability to induce spontaneous physical activity (38, 39). Enhanced orexin signaling prompted by the sensing of glucose by orexin neurons may induce activity to maintain short-term energy balance (38). This may be another potential mechanism for the influence of the HSLF condition on sedentary behavior and light-plus activity.
The LSHF condition findings may also be supported by the notion of a compensatory response to a high-sugar meal. Perhaps such a compensatory response using physical activity to regulate blood glucose levels would not have been necessary in the LSHF condition. Fiber intake stabilizes blood sugar and insulin concentrations (40). Therefore, there may be less of a compensatory response to regulate blood glucose levels after consumption of high-fiber, low-sugar meals. This may explain why there was a lack of change in participant activity behaviors throughout the day in the LSHF condition.
This study has several strengths. The controlled environment of the in-lab visits allowed for objective measurement of food intake and observation of activity behaviors, and allowed us to determine the temporal relationship between food consumption and the influence on activity behaviors. Since this study used a crossover design, we were able to compare the influences of both HSLF and LSHF meal consumption on activity behaviors in the same participants. Focus groups of Latino and African American participants were used to determine the foods included in the test meals, so the meals represented breakfast foods that were typical of these populations. The focus groups were also used to determine the in-lab activity options, so both the active and sedentary options reflected activities that are typically enjoyed by this demographic.
This study also has limitations that should be noted. While the in-lab crossover design was a strength, it may have limited the generalizability of the findings to free-living situations. Habitual dietary intake may play an important role in influencing activity behavior, and a short-term feeding study may not be sensitive enough to demonstrate larger variations in activity potentially influenced by dietary intake. Participants were aware that they were being observed in the lab setting, which could have influenced the activities they chose. It is possible that participants could have figured out the purpose of the study since they were required to consume different types of foods and subsequently be observed in the lab setting. The models for both sedentary behavior and light-plus activity did not allow the time terms to vary randomly because the models would not converge with inclusion of these terms in the random effects statements.
Conclusion
This is the first study to our knowledge to demonstrate that high-sugar/low-fiber foods can acutely influence activity behaviors in minority adolescents. Future research should aim to elucidate the longer-term influence of sugar and dietary fiber on activity behaviors as well as the influence of dietary intake on activity in free-living settings. Findings from this study and future studies may have implications for multiple behavior weight loss interventions aimed at reducing sugar intake and sedentary behavior, and increasing fiber intake and physical activity.
Study Importance.
What is already known about this subject?
Diets high in simple carbohydrates may be associated with feelings of fatigue and low energy, which may impact physical activity.
The associations of dietary sugar and fiber with sedentary behavior and physical activity are not well understood.
Hispanic and African American adolescents typically have diets high in sugar and low in fiber and do not meet sedentary behavior and physical activity guidelines.
What does this study add?
This crossover, laboratory-based experimental feeding study provides insight into the acute effects of meals high in sugar vs. meals high in fiber on sedentary behavior and physical activity in minority adolescents.
Acknowledgments
Funding: This study was supported by the National Institute for Minority Health and Health Disparities (NIMHD) as part of the USC Minority Health Center of Excellence (NCHMD P60 MD002254) and the National Institutes of Cancer (NCI), NCI Centers for Transdisciplinary Research on Energetics and Cancer (TREC, U54 CA 116848) as part of the USC Center for Transdisciplinary Research on Energetics and Cancer. Work on this manuscript was supported by the National Cancer Institute to G.A.O. (T32CA009492).
Footnotes
Disclosure: The authors do not have any conflicts of interest to disclose.
References
- 1.Briefel RR, Johnson CL. Secular trends in dietary intake in the United States. Annu Rev Nutr. 2004;24:401–431. doi: 10.1146/annurev.nutr.23.011702.073349. [DOI] [PubMed] [Google Scholar]
- 2.Clemens R, Kranz S, Mobley AR, Nicklas TA, Raimondi MP, Rodriguez JC, et al. Filling America's fiber intake gap: summary of a roundtable to probe realistic solutions with a focus on grain-based foods. The Journal of nutrition. 2012;142:1390S–1401S. doi: 10.3945/jn.112.160176. [DOI] [PubMed] [Google Scholar]
- 3.Thayer RE. The biopsychology of mood and arousal. Oxford University Press; 1989. [Google Scholar]
- 4.Thayer RE. Calm energy: How people regulate mood with food and exercise. Oxford University Press; 2003. [Google Scholar]
- 5.Pate RR, Heath GW, Dowda M, Trost SG. Associations between physical activity and other health behaviors in a representative sample of US adolescents. American journal of public health. 1996;86:1577–1581. doi: 10.2105/ajph.86.11.1577. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ranjit N, Evans MH, Byrd-Williams C, Evans AE, Hoelscher DM. Dietary and activity correlates of sugar-sweetened beverage consumption among adolescents. Pediatrics. 2010;126:e754–e761. doi: 10.1542/peds.2010-1229. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Ludwig DS, Peterson KE, Gortmaker SL. Relation between consumption of sugar-sweetened drinks and childhood obesity: a prospective, observational analysis. The Lancet. 2001;357:505–508. doi: 10.1016/S0140-6736(00)04041-1. [DOI] [PubMed] [Google Scholar]
- 8.Reynolds KD, Spruijt-Metz D. Translational research in childhood obesity prevention. Eval Health Prof. 2006;29:219–245. doi: 10.1177/0163278706287346. [DOI] [PubMed] [Google Scholar]
- 9.Mendoza JA, Drewnowski A, Cheadle A, Christakis DA. Dietary energy density is associated with selected predictors of obesity in US Children. The Journal of nutrition. 2006;136:1318–1322. doi: 10.1093/jn/136.5.1318. [DOI] [PubMed] [Google Scholar]
- 10.Caspersen CJ, Pereira MA, Curran KM. Changes in physical activity patterns in the United States, by sex and cross-sectional age. Med Sci Sports Exerc. 2000;32:1601–1609. doi: 10.1097/00005768-200009000-00013. [DOI] [PubMed] [Google Scholar]
- 11.Goran MI, Gower BA, Nagy TR, Johnson RK. Developmental changes in energy expenditure and physical activity in children: evidence for a decline in physical activity in girls before puberty. Pediatrics. 1998;101:887–891. doi: 10.1542/peds.101.5.887. [DOI] [PubMed] [Google Scholar]
- 12.Stone EJ, Baranowski T, Sallis JF, Cutler JA. Review of behavioral research for cardiopulmonary health: emphasis on youth, gender, and ethnicity. Journal of Health Education. 1995;26:S9–S17. [Google Scholar]
- 13.Eaton DK, Kann L, Kinchen S, Ross J, Hawkins J, Harris WA, et al. Youth risk behavior surveillance,ÄîUnited States, 2005. Journal of school health. 2006;76:353–372. doi: 10.1111/j.1746-1561.2006.00127.x. [DOI] [PubMed] [Google Scholar]
- 14.Baranowski T, Mendlein J, Resnicow K, Frank E, Cullen KW, Baranowski J. Physical activity and nutrition in children and youth: an overview of obesity prevention. Preventive Medicine. 2000;31:S1–S10. [Google Scholar]
- 15.Li J, Wang Y. Tracking of dietary intake patterns is associated with baseline characteristics of urban low-income African-American adolescents. J Nutr. 2008;138:94–100. doi: 10.1093/jn/138.1.94. [DOI] [PubMed] [Google Scholar]
- 16.Spruijt-Metz D, Belcher B, Anderson D, Lane CJ, Chou CP, Salter-Venzon D, et al. A high-sugar/low-fiber meal compared with a low-sugar/high-fiber meal leads to higher leptin and physical activity levels in overweight Latina females. J Am Diet Assoc. 2009;109:1058–1063. doi: 10.1016/j.jada.2009.03.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Pasman WJ, Blokdijk VM, Bertina FM, Hopman WP, Hendriks HF. Effect of two breakfasts, different in carbohydrate composition, on hunger and satiety and mood in healthy men. Int J Obes Relat Metab Disord. 2003;27:663–668. doi: 10.1038/sj.ijo.0802284. [DOI] [PubMed] [Google Scholar]
- 18.Thayer RE. Energy, tiredness, and tension effects of a sugar snack versus moderate exercise. Journal of Personality and Social Psychology. 1987;52:119. doi: 10.1037//0022-3514.52.1.119. [DOI] [PubMed] [Google Scholar]
- 19.Otten JJ, Hellwig JP, Meyers LD. DRI, dietary reference intakes: the essential guide to nutrient requirements. National Academy Press; 2006. [Google Scholar]
- 20.Kuczmarski RJ, Ogden CL, Guo SS, Grummer-Strawn LM, Flegal KM, Mei Z, et al. CDC Growth Charts for the United States: methods and development. Vital and health statistics Series 11, Data from the national health survey. 20002002:1–190. [PubMed] [Google Scholar]
- 21.Spruijt-Metz D, Belcher BR, Hsu Y-W, McClain AD, Chou C-P, Nguyen-Rodriguez S, et al. Temporal Relationship Between Insulin Sensitivity and the Pubertal Decline in Physical Activity in Peripubertal Hispanic and African American Females. Diabetes care. 2013;36:3739–3745. doi: 10.2337/dc13-0083. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Cobelli C. Quantitative estimation of insulin sensitivity. Am J Physiol. 1979;236:E667–E677. doi: 10.1152/ajpendo.1979.236.6.E667. [DOI] [PubMed] [Google Scholar]
- 23.Pacini G, Bergman RN. MINMOD: a computer program to calculate insulin sensitivity and pancreatic responsivity from the frequently sampled intravenous glucose tolerance test. Computer methods and programs in biomedicine. 1986;23:113–122. doi: 10.1016/0169-2607(86)90106-9. [DOI] [PubMed] [Google Scholar]
- 24.Matthews CE, Chen KY, Freedson PS, Buchowski MS, Beech BM, Pate RR, et al. Amount of time spent in sedentary behaviors in the United States, 2003–2004. American journal of epidemiology. 2008;167:875–881. doi: 10.1093/aje/kwm390. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Freedson PS, Melanson E, Sirard J. Calibration of the Computer Science and Applications, Inc. accelerometer. Medicine and science in sports and exercise. 1998;30:777–781. doi: 10.1097/00005768-199805000-00021. [DOI] [PubMed] [Google Scholar]
- 26.Lyons KS, Sayer AG. Longitudinal dyad models in family research. Journal of Marriage and Family. 2005;67:1048–1060. [Google Scholar]
- 27.Bolger N, Shrout PE. Accounting for statistical dependency in longitudinal data on dyads. Citeseer. 2007 [Google Scholar]
- 28.Luhmann M, Weiss P, Hosoya G, Eid M. Honey, I got fired! A longitudinal dyadic analysis of the effect of unemployment on life satisfaction in couples. Journal of personality and social psychology. 2014;107:163. doi: 10.1037/a0036394. [DOI] [PubMed] [Google Scholar]
- 29.Campbell L, Kashy DA. Estimating actor, partner, and interaction effects for dyadic data using PROC MIXED and HLM: A user–friendly guide. Personal Relationships. 2002;9:327–342. [Google Scholar]
- 30.Lyons KS, Sayer AG. Using multilevel modeling in caregiving research. Aging & mental health. 2005;9:189–195. doi: 10.1080/13607860500089831. [DOI] [PubMed] [Google Scholar]
- 31.Brage S, Wedderkopp N, Ekelund U, Franks PW, Wareham NJ, Andersen LB, et al. Features of the Metabolic Syndrome Are Associated With Objectively Measured Physical Activity and Fitness in Danish Children The European Youth Heart Study (EYHS). Diabetes care. 2004;27:2141–2148. doi: 10.2337/diacare.27.9.2141. [DOI] [PubMed] [Google Scholar]
- 32.Hurling R, Catt M, De Boni M, Fairley BW, Hurst T, Murray P, et al. Using internet and mobile phone technology to deliver an automated physical activity program: randomized controlled trial. Journal of medical Internet research. 2007;9 doi: 10.2196/jmir.9.2.e7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Curran PJ, Bauer DJ. The disaggregation of within-person and between-person effects in longitudinal models of change. Annual review of psychology. 2011;62:583. doi: 10.1146/annurev.psych.093008.100356. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Gao F, Thompson P, Xiong C, Miller JP. Analyzing multivariate longitudinal data using SAS®.. Proceedings of the Thirty-first Annual SAS Users Group International Conference; Citeseer. 2006.pp. 187–131. [Google Scholar]
- 35.Gibson EL. Carbohydrates and mental function: feeding or impeding the brain? Nutrition Bulletin. 2007;32:71–83. [Google Scholar]
- 36.McLaughlin T, Allison G, Abbasi F, Lamendola C, Reaven G. Prevalence of insulin resistance and associated cardiovascular disease risk factors among normal weight, overweight, and obese individuals. Metabolism. 2004;53:495–499. doi: 10.1016/j.metabol.2003.10.032. [DOI] [PubMed] [Google Scholar]
- 37.Baron AD, Steinberg H, Brechtel G, Johnson A. Skeletal muscle blood flow independently modulates insulin-mediated glucose uptake. American journal of Physiology. 1994;266:E248–E248. doi: 10.1152/ajpendo.1994.266.2.E248. [DOI] [PubMed] [Google Scholar]
- 38.Kotz CM. Integration of feeding and spontaneous physical activity: role for orexin. Physiol Behav. 2006;88:294–301. doi: 10.1016/j.physbeh.2006.05.031. [DOI] [PubMed] [Google Scholar]
- 39.Levine JA, Kotz CM. NEAT–non - exercise activity thermogenesis–egocentric & geocentric environmental factors vs. biological regulation. Acta Physiologica Scandinavica. 2005;184:309–318. doi: 10.1111/j.1365-201X.2005.01467.x. [DOI] [PubMed] [Google Scholar]
- 40.de Leeuw JA, Jongbloed AW, Verstegen MWA. Dietary fiber stabilizes blood glucose and insulin levels and reduces physical activity in sows (Sus scrofa). The Journal of nutrition. 2004;134:1481–1486. doi: 10.1093/jn/134.6.1481. [DOI] [PubMed] [Google Scholar]