*2.2. Study Design*

Using a randomized, double blind, counter-balanced cross-over design, low-GI and high-GI sport nutrition bars were consumed two hours before and at half time of a simulated soccer match during which soccer-specific skills (agility, ball dribbling, heading, and kicking accuracy) were assessed. Plasma glucose and insulin, non-esterified free fatty acids (NEFA), as well as fat and carbohydrate oxidation were assessed before and during the simulated soccer match.

#### *2.3. Preliminary Test*

Participants initially had their maximal aerobic power (VO2max) estimated by a shuttle-run test that involved running 20 m while the speed increased by 0.14 m/s each minute until volitional exhaustion [26]. The purpose of this test was to determine participants' aerobic fitness and to determine maximal running velocity, which was used to set speeds during the exercise tests used to evaluate the different feeding conditions. Participants then performed a familiarization test of the simulated soccer match. This test was identical to the test they performed during the sports bar conditions, but was used as a "practice" run. The purpose of this practice run was to minimize any learning effects from one test to another. The practice trial involved performing 10 six-minute sessions of running between two cones that were 20 m apart, where speed was alternated between sprinting, running, jogging, and walking to simulate the exercise performed during an actual soccer match [21,25]. Speeds of walking, jogging, and running were adjusted according to each participant's predicted maximal aerobic power [26]. The speeds were dictated by "beeps" emitted from a sound system that indicated when the participant was required to reach the next 20 m distance. The speed of walking, jogging, and running were set at 25%, 55%, and 95%, respectively, of the maximal speed reached in the initial maximal shuttle run test [27]. The Bitworks Team Beep Test software (Version 4.1, Bath, UK) was used to write the scripts for each participant. Each 6 min block alternated 60 m of jogging, running, and walking, and 20 m of sprinting. Performance was assessed by tests of either agility running/ball dribbling or kicking/heading of a soccer ball. These pairs of performance tests were alternately performed between the 6 min jogging-running-walking-sprinting sessions with the exception of the first and last 3 min periods of the test during which all four performance tests were completed [25]. The total time of the exercise test (i.e., the 10 six-minute intervals and the testing between intervals) was approximately 90 min which is the same duration as an actual soccer match. The study diagram is shown in Figure 1. This soccer test has been shown to be highly reproducible and is sensitive to improvement with carbohydrate feeding [25]. Separate performance scores were derived for agility, dribbling, kicking accuracy, and heading (Figure 2). Time to complete the agility and ball dribbling courses were recorded for assessment of performance of these tests. The highest vertical jump was recorded for heading performance. A vertical jump measuring device (Vertec, Power Systems (PS), LLC, Knoxville, TN, USA) was used; participants were instructed to use their heads rather than their hands to reach the vanes. Kicking accuracy was scored according to targets set up on a wall net.

**Figure 1.** Schematic of the simulated soccer match. "#REF" denotes fingertip blood collection time points for glucose.

**Figure 2.** Schematic of the scoring grids for kicking accuracy (**left**), ball-dribbling (**middle**), and agility (**right**) protocols. Arrows represent the distance between the cones. Adopted from Currell et al., [25].

#### *2.4. Experimental Test*

Participants reported to the lab on two different occasions after a 12 h fast for a low-GI lentil-based bar (Genki Foods Inc., Winnepegosis, MB, Canada) test and a high-GI bar (Clif Bar Inc. Berelely, CA, USA) test. Each condition was separated by at least one week. The GI of the lentil bar was 45 [28] and the GI of the Clif bar was 101 [29]. The crunchy peanut butter flavor Clif bar was used because it most closely matched the Genki Bar for macronutrients and calories. The characteristics of the nutrition bars for a 70 kg participant are shown in Table 1. The testing was double blind, that is, neither the participant nor the researchers knew what type of nutrition bar was consumed. The blinding was achieved by having a separate research assistant prepare the food and having the participant consume the food in an isolated room two hours before the exercise session. Wrappers were removed from the bars and an appropriate amount of bar was placed in plastic bags. The high- and low-glycemic index sport nutrition bars were similar in appearance (i.e., same color and consistency).


**Table 1.** Characteristics of the sport nutrition bars for a 70 kg participant.

GI, glycemic index.

On each testing day participants were given enough sports bars to consume 1.5 g/kg body mass available CHO, an amount of carbohydrate that was expected to improve performance when given two hours before high-intensity intermittent exercise [20]. This amount is also within the range of recommended CHO intake prior to endurance exercise performance [8]. Participants also consumed 0.38 g/kg available CHO from the bars at half time of the simulated soccer match. Participants had 20 min to consume the bars before the match and 15 min during half time. Furthermore, the exact amount of water consumption was documented in the first trial and then it was replicated in the second trial to minimize the impact of hydration status. The feeding and the simulated soccer match were separated by two hours. Blood glucose was assessed by fingertip sampling before the food consumption and at 5, 15, 30, 60, 90, and 120 min after consumption. Blood samples from an antecubital vein were taken immediately before, at half time, and after finishing the simulated soccer match for assessment of insulin and non-esterified free fatty acid (NEFA) levels. Fingertip blood samples were collected to assess glucose by using a glucose meter (AccuCheck Compact Plus Sarstedt, Nümbrect, Germany). Venous blood samples were maintained in 10 mL tubes (BD Vacutainer SST) for 30 min to clot. The serum was then separated by centrifugation for 15 min at 3500 rpm and stored at −80 ◦C. Insulin concentrations were determined using an enzyme-linked immunosorbent assay (ELISA) according to the manufacturer's directions (STELLUX®Chemi Human Insulin, Alpco Diagnostics, Salem, MA, USA). The serum NEFA assay was performed using a protocol with an oleic acid standard solution as per the manufacturer's directions (NEFAHR (2), Wako Diagnostics Inc., Richmond, VA, USA). The intra-assay coefficient of variations (CVs) for the insulin, and NEFA assays were <10%. Fingertip blood samples were taken after every second 6 min exercise interval during the simulated soccer match to measure glucose and lactate levels. Blood lactate measurement was assessed using BM-Lactate test strips and the Accutrend®Lactate analyzer (Roche Group; Mannheim, Germany). The K4 b2®(Cosmed USA, Chicago, IL, USA), a portable gas exchange system was used to measure oxygen consumption (VO2), and carbon dioxide output (VCO2). Respiratory gases were collected during every second 6 min exercise interval to estimate carbohydrate and fat oxidation. Carbohydrate and fat oxidation rates were estimated from VO2 and VCO2 by using stoichiometric equations [30]. Rating of perceived exertion, using the modified 10-point Borg scale, was collected after each 6 min interval [31].

#### *2.5. Dietary and Physical Activity Monitoring*

Participants recorded their dietary intake and physical activity for the 24 h before the feeding conditions. These were photocopied and given back to the participants so they could duplicate their diets and physical activity levels during subsequent feeding conditions. This ensured that participants arrived for each exercise feeding condition with similar diets and exercise the previous 24 h.

#### *2.6. Statistical Analysis*

All variables were analyzed with a two-factor repeated measures analysis of variance (ANOVA) with factors for food condition (low-GI lentil bar vs. high-GI Clif Bar) and time during the exercise test. When there was a time main effect or an interaction between condition and time, a Least Significant Difference (LSD) post-hoc test was used to determine differences between pairs of means. All variables were also assessed for order effects with a two-factor ANOVA with factors for order (first condition vs. second condition) and time during the exercise test. Significance was accepted at a *p*-value less than 0.05. All results are reported as means and standard deviations.

#### **3. Results**
