Exercise Energy Expenditure and Postprandial Lipemia in Girls

Keith Tolfrey; Alex Engstrom; Caoileann Murphy; Alice Thackray; Robert Weaver; Laura A. Barrett


Med Sci Sports Exerc. 2014;46(2):239-246. 

In This Article



After approval from the University Ethical Advisory Committee, 18 girls volunteered for the study and provided their written assent. Written informed consent was also obtained from a parent or guardian. All participants were physically active generally, in good overall health, not taking any medications or dietary supplements known to affect lipid or carbohydrate metabolism, and had no contradictions to exercise participation. Physical and physiological characteristics are presented in Table 1.

Anthropometry and Physical Maturation

All anthropometric measurements were conducted with participants wearing shorts, T-shirt, and socks. Stature was measured to the nearest 0.01 m using a fixed stadiometer (Holtain, Crosswell, UK). Body mass was quantified to the nearest 0.1 kg using a balance beam scale (Avery, Birmingham, UK). Body mass index was calculated as body mass (kg) divided by stature (m) squared. Skinfold thickness was measured to the nearest 0.2 mm using Harpenden calipers (John Bull, St. Albans, UK), with measurements taken on the right-hand side of the body at two sites (triceps and subscapular). The median of three measurements at each site was used to estimate percent body fat.[31] Lean body mass was estimated as follows:

Participants provided a self-assessment of their level of physical maturity using drawings depicting the five stages of breast and pubic hair development.[32] The scale ranges from 1, indicating prepubescent, to 5, indicating full sexual maturation, and participants were required to select the stage most closely resembling their current level of sexual development. We did not control for the menstrual cycle because menstruation had not started for 11 of the 18 girls, and 4 of the remaining 7 girls reported that their menstrual cycle was sporadic and irregular. This variability is a limitation of working with a mixed maturation sample.

Preliminary Exercise Measurements

During the first visit to the laboratory, participants were familiarized with walking and running on the treadmill (Technogym Runrace, Gambettola, Italy) before completing two preliminary exercise measurements. The first test involved a 16-min incremental treadmill protocol divided into 4 × 4 min stages. The treadmill started at a speed of 4 km·h−1 and increased 1 km·h−1 at the start of each subsequent stage, with the gradient set at 1% throughout.[19] This enabled the individual steady-state relationship between treadmill speed, oxygen uptake (V̇O2), and heart rate to be established. Participants were given a standardized 10-min passive rest period before completing the second exercise measurement to determine peak V̇O2. The girls ran at a fixed individual speed (6.8–9.5 km·h−1), identified from the incremental treadmill protocol, while the treadmill gradient was increased 1% each minute until volitional exhaustion. Throughout both tests, heart rate was monitored continuously via short-range telemetry (PE4000; Polar-Electro, Kempele, Finland), and Borg's ratings of perceived exertion were recorded at standardized intervals.[5] Expired air samples were monitored continuously using an online breath-by-breath gas analysis system (Cosmed K4b2, Rome, Italy), which was fully calibrated according to the manufacturer's instructions before each use. Participants wore an appropriate size facemask (Hans Rudolf, Shawnee, KS), which was checked for leaks and connected to the online system via a flowmeter before each exercise test began. At least two of the following criteria were satisfied by all participants during the peak V̇O2 test to confirm maximal effort was achieved: a plateau in V̇O2 (≤3%) with an increase in treadmill gradient, a peak heart rate ≥95% of age-predicted maximum (220 - chronological age), and a respiratory exchange ratio ≥1.10. An average of the breath-by-breath V̇O2 data were taken every 10 s, and peak V̇O2 was defined as the highest 30-s rolling average. Data from the incremental and peak V̇O2 protocols were used to determine the treadmill speed required to elicit ~55% peak V̇O2 during the main experimental exercise conditions.

Experimental Design

Similar to previous studies with boys,[2,30,34,36] a within-measures, counterbalanced crossover design was adopted whereby participants completed three, 2-d experimental conditions: resting control (CON), 30 min of intermittent treadmill exercise (EX30), and 60 min of intermittent treadmill exercise (EX60). The conditions were separated by a standardized period of 14 d. The study design is shown schematically in Figure 1.

Figure 1.

Schematic of the 2-d study protocol. TAG, triacylglycerol; CON, rest control condition; EX30, 30-min intermittent exercise condition; EX60, 60-min intermittent exercise condition. Evening meal replicated from first condition.

Day 1

The girls arrived at the laboratory at 1530 h, and all measures were completed by 1730 h. Body mass was recorded at the start of each experimental condition to standardize the test breakfast provided on day 2 (see next section). During CON, participants rested for 110 min in the laboratory. During the exercise conditions, participants completed either 30 min (EX30) or 60 min (EX60) of moderate-intensity intermittent treadmill exercise designed to elicit ~55% peak V̇O2. The exercise conditions were completed in intervals of either 3 × 10 min or 6 × 10 min with a standardized 10-min period of passive rest between each interval. Oxygen uptake and heart rate were monitored continuously throughout each 10-min exercise interval as described previously, and the treadmill speed was adjusted accordingly to ensure the target exercise intensity was achieved. Participants provided RPE at standardized intervals as described previously. The exercise EE and the oxidation rate of carbohydrate and fat were estimated via indirect calorimetry,[10] assuming that participants reached a physiological steady state and that the urinary nitrogen excretion rate was negligible.

Day 2

After a 12-h overnight fast, the girls arrived at the laboratory at ~0745 h, and a fasting capillary blood sample was taken after a 10-min seated rest. A high-fat milkshake was consumed for breakfast within 10 min, marking the start of the postprandial period (0800 h), and six further capillary blood samples were taken at hourly intervals throughout the 6-h postprandial period (Fig. 1). Participants rested throughout the postprandial period and were able to read, watch DVD films, and play nonactive computer games. Participants consumed water ad libitum during the first condition and replicated the ingested volume during the subsequent conditions.

Standardization of Diet, Physical Activity, and Test Milkshake

Participants recorded their dietary intake, and all physical activity categorized according to intensity level in a detailed diary during the 48-h period preceding day 2 of the first experimental condition. They were also asked to minimize their physical activity during this period and to replicate this diet and physical activity pattern before the second and the third experimental conditions, which was confirmed verbally by the lead investigator. The overnight fasting period was standardized by asking participants to consume a small carbohydrate snack at 1945 h on day 1 of each experimental condition (Kellogg's® Nutri-Grain Strawberry, 502 kJ, 3 g fat, 24 g carbohydrate, 2 g protein). After 2000 h, the participants were allowed to drink plain water but were asked not to consume any other drinks or food before arriving at the laboratory on day 2.

The milkshake provided for breakfast on day 2 contained vanilla dairy ice cream and double cream, at a ratio of 3:1, with 10 g of either powdered strawberry or chocolate flavor added to ensure the milkshake was palatable. Participants consumed the same milkshake flavor on all visits for consistency. The milkshake quantity was adjusted relative to body mass so that it provided 1.50 g of fat (70% of total energy), 1.20 g of carbohydrate (25%), 0.21 g of protein (5%), and 80 kJ·kg−1 body mass.[34]

Analytical Methods

After the hand was prewarmed in water heated to 40°C for 5 min, the fingertip was pierced (Unistick 3 Extra, Owen Mumford, UK), and between 300 and 600 μL of whole capillary blood was collected into potassium-EDTA coated Microvette CB 300 tubes (Sarstedt Ltd., Leicester, UK). The whole blood samples were centrifuged immediately for 15 min at 12,800g (Eppendorf 5415c, Hamburg, Germany). An automatic pipette was used to dispense 20 μL of plasma into a 1.5-mL Eppendorf tube (Fisher Scientific Ltd., Loughborough, UK), which was then diluted 50 times by adding 980 μL of ice-cold saline (0.9% NaCl)[2] (Hamilton Microlab 500 series, Reno, NV). This procedure was repeated so that two aliquots of diluted plasma were stored at –80°C for up to 2 months before subsequent analysis.[34] All samples were analyzed for [TAG] and glucose concentration ([glucose]) by enzymatic, colorimetric methods (Randox Laboratories Ltd., Crumlin, UK) using a centrifugal analyzer (Cobas Mira Plus; Roche, Basel, Switzerland). To account for the predilution of the plasma samples, the dilution step of the assay was eliminated, and three times the sample volume stated in the Randox kit assay procedure was used to ensure the concentration of the sample for analysis was the same as the original assay procedure.[2,34,36] The within-batch coefficients of variation for [TAG] and [glucose] were 3.4% and 2.2%, respectively. Hemoglobin concentration and hematocrit were also quantified in the fasting and postprandial samples to estimate the acute change in plasma volume.[9] Hemoglobin was measured in duplicate by hemophotometry (HemoCue AB, Ängelholm, Sweden), and hematocrit was assessed in duplicate using a microhematocrit centrifuge and reader (Haematospin 1300 Microcentrifuge; Hawksley & Sons Ltd., Sussex, UK).

Statistical Analyses

Data were stored and analyzed using the IBM SPSS statistics software for Windows version 19 (IBM Corporation, Armonk, NY). Descriptive statistics outlining the physical and physiological characteristics of participants were determined. Normality of the data was confirmed by Shapiro–Wilk tests, and the homogeneity of variances was checked by Mauchly's test of sphericity, with a Greenhouse–Geisser correction of the degrees of freedom applied if sphericity was violated. Student's paired t-tests were used to identify differences between the EX30 and EX60 responses. The total 6 h area under the plasma concentration versus time curves for TAG (TAUC-TAG) and glucose (TAUC-glucose) were calculated using the trapezium rule. Incremental versions of these were also calculated after accounting for respective fasting concentrations (iAUC-TAG and iAUC-glucose) across the conditions. The AUC responses, fasting blood concentrations, and estimated changes in plasma volume were compared using separate one-way within-measures ANOVA. Differences in postprandial [TAG] and [glucose] were identified using separate 3 × 7 (condition by time) within-measures ANOVA. The 95% confidence intervals (95% CI) for the mean absolute pairwise differences between experimental conditions were calculated using the t-distribution and degrees of freedom (n – 1), and absolute standardized effect sizes (ES) are provided to supplement the findings. In the absence of a clinical anchor, an ES of 0.2 was considered the minimum important difference for all outcome measures, 0.5 moderate and 0.8 large.[6] Bivariate correlations identifying possible determinants of the exercise-induced changes in TAUC-TAG were quantified using linear regression. The interpretation of the data will be based on 95% CI and ES.