Pre-Exercise Carbohydrate or Protein Ingestion Influences Substrate Oxidation but Not Performance or Hunger Compared with Cycling in the Fasted State

utritional intake can in uence exercise metabolism and performance, but there is a lack of research comparing protein-rich pre-exercise meals with endurance exercise performed both in the fasted state and following a carbohydrate-rich breakfast. The purpose of this study was to determine the effects of three pre-exercise nutrition strategies on metabolism and exercise capacity during cycling. On three occasions, seventeen trained male cyclists (VO 2peak62.2 5.8 mL kg 1 min 1 , 31.2 12.4 years, 74.8 9.6 kg) performed twenty minutes of submaximal cycling (4 5 min stages at 60%, 80%, and 100% of ventilatory threshold (VT), and 20% of the difference between power at the VT and peak power), followed by3 3 min intervals at 80% peak aerobic power and3 3 min intervals at maximal effort, 30 min after consuming a carbohydrate-rich meal (CARB;

74.8 9.6 kg) performed twenty minutes of submaximal cycling (4 5 min stages at 60%, 80%, and 100% of ventilatory threshold (VT), and 20% of the difference between power at the VT and peak power), followed by3 3 min intervals at 80% peak aerobic power and3 3 min intervals at maximal effort, 30 min after consuming a carbohydrate-rich meal (CARB; 1 g/kg CHO), a protein-rich meal (PROTEIN; 0.45 g/kg protein + 0.24 g/kg fat), or water (FASTED), in a randomized and counter-balanced order. Fat oxidation was lower for CARB compared with FASTED at and below the VT, and compared with PROTEIN at 60% VT. There were no differences between trials for average power during high-intensity intervals (367 51 W,p= 0.516). Oxidative stress (F 2-Isoprostanes), perceived exertion, and hunger were not different between trials. Overall, exercising in the overnight- fasted state increased fat oxidation during submaximal exercise compared with exercise following a CHO-rich breakfast, and pre-exercise protein ingestion allowed similarly high levels of fat oxidation. There were no differences in perceived exertion, hunger, or performance, and we provide novel data showing no in uence of pre-exercise nutrition ingestion on exercise-induced oxidative stress. Keywords:nutrition; exercise; fat oxidation; oxidative stress; isoprostanes 1. Introduction Nutritional intake before exercise can in uence performance and the physiological responses to an exercise session [1]. Exercise performed with reduced carbohydrate (CHO) availability can increase fat oxidation, increase the activation of cell signaling pathways, and promote oxidative adaptations in skeletal muscle [2,3]. At the same time, suf cient CHO ingestion before and/or during exercise is recommended for exercise sessions requiring a high quality, duration, and/or intensity [4]. It is therefore suggested that CHO ingestion be varied according to the goals and type of each exercise session to optimize both training adaptations and acute performance, yet there is wide variance among athletes regarding appropriate nutritional intake before exercise [5]. Strategies to vary CHO availability before exercise include ingesting high- or low- CHO meals, and exercising in the overnight-fasted state. We recently reported nearly Nutrients2021,13, 1291.

training adaptations and acute performance, yet there is wide variance among athletes regarding appropriate nutritional intake before exercise [5]. Strategies to vary CHO availability before exercise include ingesting high- or low- CHO meals, and exercising in the overnight-fasted state. We recently reported nearly Nutrients2021,13, 1291.

Nutrients2021,13, 1291 2 of 15 two-thirds of endurance athletes (63%) train in the overnight-fasted state, while 72% consume CHO before some or all training sessions, and only 28% ever consume low-CHO meals before exercise [5,6]. Athletes perform fasted-state training primarily to increase fat oxidation and improve gut comfort during exercise, while athletes that avoid fasted training do so because they feel their workout quality deteriorates and/or they will be too hungry during exercise [6]. It is well established that performing low-to-moderate intensity exercise in the overnight-fasted state can induce higher levels of fat oxidation compared with exercise performed following ingestion of CHO [3]. However, several studies have shown fat oxidation during exercise to be similar following protein ingestion compared with a placebo (fasted) condition [7,8]. Therefore, pre-exercise protein ingestion may be an alternative to performing fasted-state training that could reduce hunger while maintaining high levels of fat oxidation. Additional research is needed to better understand differences in substrate oxidation between CHO-fed, protein-fed, and fasted-state training, as previous studies using pre-exercise protein ingestion have either not had a CHO control group [7,8], performed extended exercise at a single intensity [7–9], or provided very large (>1000 kcal) pre-exercise meals [10,11]. From a performance standpoint, fed-state exercise generally enhances prolonged (>60 min), but not shorter duration (<60 min) aerobic exercise compared with exercising in the fasted state [12], although few studies have used a high-intensity interval training (HIIT) model to measure performance despite HIIT being performed by virtually all endurance athletes [13]. Total work performed during HIIT has been reported to be increased in the fed, compared with the fasted state during some [14,15] but not all [16] studies. To our knowledge, no studies have compared pre-exercise CHO, protein, and fasted-state training on HIIT work capacity. More important for athletes and coaches than an acute training session are the longer- term training adaptations. Exercise-induced oxidative stress provides a key signal for the adaptative response to an exercise session with greater exercise-induced oxidative stress being associated with improved adaptations [17,18], but it is unknown how this might be affected by various pre-exercise meals. At

HIIT work capacity. More important for athletes and coaches than an acute training session are the longer- term training adaptations. Exercise-induced oxidative stress provides a key signal for the adaptative response to an exercise session with greater exercise-induced oxidative stress being associated with improved adaptations [17,18], but it is unknown how this might be affected by various pre-exercise meals. At rest, a high-CHO meal can evoke a greater postprandial oxidative stress response compared with a high-fat meal [19], while whey protein can enhance endogenous antioxidant enzyme activity [20]. Furthermore, CHO ingestion before and during exercise can decrease exercise-induced oxidative stress during longer-duration moderate-intensity cycling [21]. Therefore, an understanding of how nutrition in uences exercise-induced oxidative stress would be valuable as it could inform pre-exercise nutrition strategies. Research comparing pre-exercise protein with pre-exercise CHO ingestion and fasted- state training across a range of exercise intensities is needed and could help endurance athletes and coaches make better pre-exercise nutrition choices. Accordingly, the aim of this crossover study was to determine the effects of three different pre-exercise nutrition strategies on substrate oxidation, performance during HIIT, and exercise-induced oxidative stress. We hypothesized that the fasted and protein conditions would have the highest fat oxidation, while cycling power during HIIT would be highest following CHO ingestion. Secondary outcomes were to determine the in uence of the pre-exercise meal on cycling gross ef ciency, heart rate (HR), and rating of perceived exertion (RPE) during moderate and high intensity cycling, along with hunger and gut comfort before and after exercise. 2. Materials and Methods Participants: Seventeen trained cyclists and triathletes participated in this study (31.2 12.4 years , 181.9 6.4 cm, 74.8 9.6 kg, VO 2peak62.2 5.8 mL kg 1 min 1 , peak aerobic power 425 55 W/5.7 0.6 W kg 1 , average weekly training volume13.6 3.1 h ). Participants were required to be 18–55 years old, with a training history of at least 8 h per week for the previous two years, and a VO 2peak> 55 mL/kg/min. One participant completed only two of the three trials due to an injury unrelated to this

55 W/5.7 0.6 W kg 1 , average weekly training volume13.6 3.1 h ). Participants were required to be 18–55 years old, with a training history of at least 8 h per week for the previous two years, and a VO 2peak> 55 mL/kg/min. One participant completed only two of the three trials due to an injury unrelated to this study. Using the PASS 15 software and a 3x3 cross-over design, we estimated that 16 subjects would

Nutrients2021,13, 1291 3 of 15 allow us to detect a difference in fat oxidation of 0.21 g min 1 among the three within- subject conditions, with 90% power and a type I error rate of 0.05. This is based on an F Test, a between-subject standard deviation of = 0.2, and a conservative autocorrelation among the repeated measurements of 0.2. We based our assumed means across the three conditions (carbohydrate = 0.34, protein = 0.55, fasted = 0.55) on previous work [22,23]. The standard deviation of the hypothesized means is m= 0.099, equating to a Cohen'sƒ effect size of 0.495. Using the same parameters, we estimated that 12 subjects were required to detect a difference in respiratory exchange ratio (RER), using means (0.91, 0.86, 0.86) and a between-subject standard deviation of = 0.04 ( m= 0.0236; Cohen'sƒ= 0.59) also estimated from previous work [22,23]. All study protocols and materials were approved by the Auckland University of Technology Ethics Committee (19/420). Participants reported to the laboratory on four occasions, seven days apart. Partici- pants were asked to refrain from exercise, caffeine, and alcohol 24 h before each visit and kept a 24 h food log in order to replicate dietary intake prior to each testing day. Instructions on keeping a food log were provided. Visit 1: After obtaining written informed consent and completing a health screening, a graded exercise test was performed to determine maximal oxygen consumption (VO 2peak). Participants cycled on an electronically braked cycle ergometer (Excalibur Sport, Lode BV, Groningen, The Netherlands) at 60 W for three minutes followed by a 30 W per minute increase until volitional fatigue. Expired gas was collected and analyzed continuously using a computerized metabolic system with mixing chamber (TrueOne2400, ParvoMedics, Sandy, UT, USA), with the VO 2peakrecorded as the highest 15-s average. Peak power (Wmax) was determined by the workload in the last completed stage plus the workload relative to the time spent in the last incomplete stage [power of completed stage + (30*(seconds at uncompleted stage/60)]. The ventilatory threshold (VT) was identi ed as the work rate where the ventilatory equivalent for

with the VO 2peakrecorded as the highest 15-s average. Peak power (Wmax) was determined by the workload in the last completed stage plus the workload relative to the time spent in the last incomplete stage [power of completed stage + (30*(seconds at uncompleted stage/60)]. The ventilatory threshold (VT) was identi ed as the work rate where the ventilatory equivalent for oxygen (V E.V O2 1) began to increase in the absence of changes in the ventilatory equivalent for carbon dioxide (V E.V CO2 1), with 15 W deducted to account for the lag in VO2during the incremental test [24]. Following a 10-min rest, participants were familiarized with the HIIT protocol using 3 3-min intervals. The rst interval was set at 80% Wmax, performed in a cadence- independent manner, while subsequent intervals used the cadence-dependent linear mode set to produce a workload of 80% Wmaxat their preferred cadence. Three intervals were deemed appropriate to minimize the likelihood of training effects and because the par- ticipants were fatigued from the VO 2peaktesting. Participants were asked at the start of the session about their weekly training volume, recorded as self-reported hours per week, and how often they perform exercise in the overnight-fasted state (i.e., without ingesting any calorie-containing foods or beverages). All trials were conducted under standard laboratory conditions (17–19 C, 40–65% relative humidity), with participants fan cooled during exercise. Visits 2–4: Participants reported to the laboratory in an overnight-fasted state (~10 h), with each visit at the same time of day. Upon arrival, participants completed a ve- question survey that assessed fatigue, sleep quality, muscle soreness, stress, and mood on a ve-point scale (scores 1 to 5), with overall well-being determined by summing the ve scores [25]. Participants also rated their subjective sensations of hunger and gut discomfort upon arrival and again at the end of each session using paper-based visual analogue scales (VAS) with written anchors of “not hungry at all”/“no discomfort” and “extremely hungry”/“extreme discomfort” placed 0 and 100 mm, respectively [26]. A urine sample was obtained upon arrival (before meal consumption) and within ve minutes of completing the exercise

sensations of hunger and gut discomfort upon arrival and again at the end of each session using paper-based visual analogue scales (VAS) with written anchors of “not hungry at all”/“no discomfort” and “extremely hungry”/“extreme discomfort” placed 0 and 100 mm, respectively [26]. A urine sample was obtained upon arrival (before meal consumption) and within ve minutes of completing the exercise session. In a randomized and counter-balanced order, participants received one of three meals to be consumed within a 5-min window. A CHO-rich meal (CARB; 1 g/kg CHO), a protein- rich meal (PROTEIN; 0.45 g/kg protein + 0.24 g/kg fat), or 500 mL water (FASTED). A 70-kg person received 51 g white bread (Tip Top, New Zealand) with 19 g raspberry jam (Barkers, New Zealand) and 500 mL of a 7% CHO-electrolyte drink [4:1 glucose-to-fructose

Nutrients2021,13, 1291 4 of 15 ratio; Replace, Horleys, New Zealand] for CARB, and 25 g whey protein isolate (ICE, Horleys, New Zealand) with 33 g peanut butter (Forty Thieves, New Zealand) and 500 mL water for PROTEIN. A small amount of fat was included with PROTEIN to keep the two trials isocaloric and mimic real-world application. Total energy content of the CARB and PROTEIN meals was 4 kcal per kg body mass (299 38 kcal). All groups consumed 500 mL uid and could drink water ad libitum during the remainder of the session. Thirty minutes after ingestion of the meal, participants began the sub-maximal cycling portion of the testing which included 4 5-min stages at a power equivalent to 60%, 80%, and 100% of VT (VT60, VT80, VT100, respectively), and 20% of the difference between VT and Wmax(VTD20), to measure substrate oxidation, energy expenditure, heart rate (HR), and perceived exertion (RPE) (Figure). Expired gas was continuously measured using a metabolic cart (TrueOne2400, ParvoMedics, Sandy, UT, USA), with average values during the nal two minutes of each stage analyzed. Intensity was normalized to the VT to reduce inter-subject variability in the physiological and perceived responses to exercise compared with using a percentage of VO 2peak[27]. Figure 1. Schematic overview of cycle testing sessions using wattage from an example participant. The 5-min intervals were performed at intensities equivalent to 60%, 80%, and 100% of their ventilatory threshold (VT 60, VT80, VT100, respectively), and 20% of the difference between the ventilatory threshold and peak power (Wmax, VTD20), followed by a 3-min cool down at 100 W. The 3-min intervals included a lead-in of 1 min at 100 W and 1 min at 150 W, followed by intervals 1–3 at 80% of Wmax, and intervals 4–6 performed as maximal efforts. Rate of energy expenditure (EE) was calculated using the formulas of Jeukendrup and Wallis [28], with cycling gross ef ciency (GE) calculated as GE(%) = (mechanical work (kcal/min)/energy expenditure (kcal/min)) 100. Whole-body rates of CHO and fat oxidation were calculated using standard equations, assuming 9.75 kcal/g fat and 4.07 kcal/g CHO [28]. Following a

intervals 4–6 performed as maximal efforts. Rate of energy expenditure (EE) was calculated using the formulas of Jeukendrup and Wallis [28], with cycling gross ef ciency (GE) calculated as GE(%) = (mechanical work (kcal/min)/energy expenditure (kcal/min)) 100. Whole-body rates of CHO and fat oxidation were calculated using standard equations, assuming 9.75 kcal/g fat and 4.07 kcal/g CHO [28]. Following a 3-min static rest, participants performed 6 3-min cycling intervals with 3 min of active recovery (100 W) between each interval (Figure). The rst three intervals were performed at 80% of Wmax, in a cadence-independent manner. Intervals 4–6 used the cadence-dependent linear mode set to produce a workload of 80% Wmaxat their preferred cadence, with participants instructed to produce their maximal power output across intervals 4–6 by increasing the cycling cadence. Immediately following intervals 3 and 6, a 0.3 L blood sample was collected from the left index ngertip and analyzed immediately using a portable blood lactate analyzer (Lactate Pro 2, Carlton, Australia). Power output (W) during HIIT was analyzed as mean power (W) for each interval. Heart rate was measured using a chest-strap (Polar T31, Polar, Inc., Kempele, Finland), with average values during the nal 30 s of each interval analyzed. Rating of perceived exertion

Nutrients2021,13, 1291 5 of 15 (RPE) was recorded following each submaximal stage and each high-intensity interval using Borg's 6–20 scale [29], and at the end of the session (SRPE) using a 10-point scale [30]. We chose to have the work rate “clamped” for the rst three intervals to compare HR, RPE, and lactate across conditions at a xed cycling power, and have three intervals performed as maximal efforts to be used to determine work capacity during HIIT. A competitive immunoassay was used for the quantitation of urinary F2-isoprostanes (Kit #51635, Cayman Chemicals, Ann Arbor, MI, USA) as previously described [31]. Sam- ples were puri ed using solid-phase extraction cartridges. For standardizing urine dilution, creatinine levels were measured using a commercially available kit (Kit #500701, Cay- man Chemicals, Ann Arbor, MI, USA). Due to technical problems the number of samples analyzed was n = 12 for CARB, n = 12 for FASTED, and n = 11 for PROTEIN. Statistical analysis: A series of linear mixed models were used to estimate differences in the exercise-induced changes between the three treatment conditions (CARB, PROTEIN, FASTED). These were t using the lme4 R package. For the submaximal portion, intensity (four levels: VT60, VT80, VT100, VTD20) was added as a xed effect, while interval (three levels) was considered a xed effect for the high-intensity portion. When examining differences pre-post exercise (for hunger, gut comfort, and oxidative stress measures), time point (two levels; pre and post) was added as a xed effect. For all models, treatment order was included as a xed effect (given the crossover design) and participant ID was speci ed as a random intercept. Interactions between the treatment and other xed effects were explored, and the optimal, best- tting model for each outcome was decided based on the likelihood ratio test. The t of each model was checked by visualizing the Q–Q and other residual plots to ensure approximate residual normality and heteroscedasticity, using the performance R package. Model-estimated means were calculated using the emmeans R package and presented as estimated means 95% con dence interval (CI). Contrasts between each treatment (within each

This study examines the effects of pre-exercise nutrition on metabolism and performance in trained cyclists.