It is not uncommon for athletes to express confusion over the timing and composition of precompetition meals. Some athletes may purposely consume a light meal as a result of precompetition nerves or concern over gastrointestinal comfort. Other athletes, however, may consume a very large meal with a view to providing adequate energy for the duration of an event. Consumption of a carbohydrate (CHO) source during competition may also be limited in some sports such as mountain biking, therefore warranting greater attention to precompetition nutrition preparation.
Carbohydrate ingestion 1–4 h before endurance exercise has been shown to enhance performance (10,23,25,26,28), show no effect (8,15), or even impair performance (9). Such variability of results may be attributed to carbohydrate quantity (5,23,24), timing, quality, and frequency of carbohydrate consumption before exercise (11).
In exercise lasting more than 90 min at 60–85% V̇O2max (moderate to high intensity), carbohydrate is the major source of energy (25). Ingesting small volumes of carbohydrate preexercise may not provide sufficient glucose to sustain moderate- to high-intensity prolonged exercise, particularly if initial glycogen stores are limited (11). Such conditions will lead to inability to maintain exercise at the same intensity or power output as time progresses (1,24). With the exception of feeding during exercise, limited carbohydrate availability due to the conditions described may account for the inability of some studies (8,15) to find significant improvements in performance.
Sherman et al. (24) investigated the performance effects after preexercise ingestion of 1.1 and 2.2 g CHO·kg−1 BM consumed 1 h before 90 min of cycling (i.e., 90 min at moderate to high intensity followed by 40 min at high intensity). No performance improvements were observed. However, ingestion of larger quantities of carbohydrate (ranging from approximately 2.7 to 5.0 g·kg−1 BM) 3–4 h before prolonged moderate to high-intensity exercise has resulted in significant improvements in endurance performance in comparison to a fasting placebo (19,23,28). Neufer et al. (19) found that feeding immediately before exercise and during exercise also improved performance. However, the addition of 200 g of carbohydrate 4 h before exercise produced the greatest improvement in exercise performance. The improvements in each study were attributed to greater carbohydrate availability, thus improving the ability to oxidize carbohydrate at higher rates late in exercise. Perhaps a failure to observe a performance benefit during the study by Sherman et al. (24) was due to insufficient reserves for an increase in carbohydrate availability and oxidation rates over a prolonged period. Coyle (5) suggests that a threshold of carbohydrate availability through oral intake (i.e., 200 g) must be reached for significant improvements in performance to occur.
Nutrition requirements for cross-country mountain bike (MTB) cyclists have not been previously researched. Therefore, current dietary recommendations are based on the assumption that MTB racing resembles the demands of other endurance events of similar duration (1.5–2.5 h) and intensity. Recent research into physiological demands of MTB cycling has revealed that it is intermittent in nature and raced at a very high intensity (12 and AIS unpublished data) and that increases in speed and grade during simulated off road cycling have been shown to significantly increase oxygen uptake, heart rate, and rating of perceived exertion (2). A study by Palmer et al. (22) showed that intermittent cycling incurs higher energy demands than steady-state cycling when performance is measured at the same average intensity. Furthermore, MTB cycling is technically demanding requiring upper body involvement for control and handling of the bicycle. Feeding can also be difficult during racing due to intensity and technical difficulty. The characteristics of MTB racing suggest that it is a unique endurance sport and that at least the physiological demands of the sport should be closely replicated during research in order to establish optimal nutrition strategies for the best performance.
The following study tests the hypothesis that a large quantity of carbohydrate ingested preexercise will produce a greater performance improvement in MTB cyclists during a MTB specific performance test compared with a smaller carbohydrate meal preexercise. Furthermore, some markers of metabolism were assessed to determine their responses to the different amounts of carbohydrate ingested preexercise.
The investigation was carried out with eight-trained male MTB cyclists whose profile expressed as mean (SD) include age 22 ± 6.3 yr, weight 69 ± 7.6 kg, height 179 ± 6.4 cm, skinfold sum 47 ± 6.8 mm, and V̇O2max 60 ± 3.7 mL·kg−1·min−1. Each subject competed in club level MTB races on a regular basis and was in either a general preparation/maintenance phase or endurance phase of training for the upcoming MTB season (more than 8 h training on the bike per week).
Each subject was given a Participant Information Sheet and an Informed Consent form to sign once they had agreed to the testing procedure. Parent guardian consent was also obtained for those subjects under 18 yr of age. The University of Wollongong Human Ethics Committee approved experimental procedures. Subjects were also verbally briefed about the testing procedure in addition to receiving dietary advice relating to preparation for each trial.
The study consisted of two familiarization trials and two experimental trials. The familiarization trials were used to determine the individualized workload/power outputs for the subsequent experimental trials and to introduce all experimental procedures including the intermittent nature of the MTB test in order to minimize potential learning effects. A placebo trial was not performed in this study. The rationale was that existing research clearly shows performance benefits when sufficient carbohydrate (i.e., >200 g) is consumed before exercise of durations 1–4 h, compared with consuming no carbohydrate preexercise (19,23,28).
The two experimental trials were undertaken in a randomized, double-blind design. Subjects were required to arrive at the laboratory 3 h before the scheduled commencement of their exercise test for insertion of a venous cannula and to consume the pretest meal. Subjects were given food diaries and asked to record food and fluid intake as well as activity conducted in the 48 h before the first familiarisation trial. Subjects were asked to repeat the first dietary pattern for all subsequent trials to ensure similar preparation. Food and fluid intake and activity was recorded for the 48 h before all trials and analyzed using a computer dietary analysis program (Serve Nutrition Management System, Version 2.0, M&H Williams Pty. Ltd., NSW, Australia) to verify controlled preparation. Subjects were instructed to conduct “light” training for 2 d before the MTB test and to fast for 7 h before each trial. All subjects were asked not to consume any caffeine-containing products and alcohol in the 24 h before each trial.
During the first visit to the laboratory, subjects underwent a graded exercise test (GXT) on a cycle ergometer (Lode Excalibur, Groningen, The Netherlands) to determine maximal oxygen uptake (V̇O2max) and power output at Dmax (4,30). Dmax was used as an objective measure to mathematically determine an individual’s power output at a threshold based on blood lactate levels (30). The protocol began at 150 W and increased by 50 W every 5 min until volitional exhaustion. Exhaustion was verified by a respiratory exchange ratio (RER) value that exceeded 1.1 and a plateau in V̇O2 (17). Subjects were asked to refrain from strenuous exercise for 2 d before the test. Pretest body mass (BM) was recorded to the nearest 0.05 g (Digital Weighing Scale, Teraoka Seiko, Tokyo, Japan) and used to determine the amount of carbohydrate to be consumed for each carbohydrate trial. Body fat levels were measured for each subject as a sum of seven skinfolds (bicep, tricep, subscapular, supraspinale, abdomen, thigh, and calf) by a trained anthropometrist according to the International Society for Anthropometry and Kinanthropometry (ISAK) Standards (20).
MTB performance test.
The MTB test protocol was based on power output data (SRM Training System, Germany) collected from an elite Australian male MTB cyclist during the third stage (circuit course) of a 3-d MTB tour. Power output during the MTB race was intermittent and variable. A laboratory test protocol was devised based on the percentage of time that the MTB cyclist spent at various power outputs during the race. The test protocol was 22.50 min in duration and was programmed in 10-s blocks (using Lode Excalibur Work Load Programmer) to correspond with intermittent nature of MTB racing. The course profile designed for each individual cyclist (Fig. 1) was scaled according to power output achieved at Dmax during the V̇O2max test. The course profile was designed so that power output was above 50% of Dmax at all times to ensure that intensity was maintained.
Each subject completed four laps of the 22.50-min protocol on an electrically braked cycle ergometer (Lode) with self-selected pedal cadence and automatically adjusted resistance. Testing was conducted in standard laboratory conditions (approximately 21°C and 30% relative humidity), and subjects were fan cooled using the same fan setting for each trial. The MTB test was preceded by a standard 9-min warm-up beginning at 100 W and increasing by 50 W every 3 min. Subjects used their own clipless MTB shoes and pedals, and they were positioned according to the seat height, handlebar height, and seat to head stem length on their usual race bike. The ergometer was fitted with a racing handlebar and seat. The same protocol and bike position for each individual was used for all experimental trials.
Performance was assessed by measuring the total amount of work in kilojoules completed on six “open sections”/short time trial during each lap of the MTB test (Fig. 1). In the open sections, subjects were able to produce a maximal or near maximal effort because the resistance from the cycle ergometer was proportional to the effort applied. This was designed to simulate parts of a MTB race where riders have an opportunity to gain time (e.g., open fire road), to chase another rider, or to ascend steep sections of a course. Each open section lasted 30 s to correspond with regular short-duration maximal or near maximal efforts during a MTB race. The sections were evenly spaced following the steady state section in the test protocol to allow for some recovery and collection of blood gases. Subjects were instructed to pace themselves as if they were competing in a MTB race. Equal encouragement was given to the subjects to ensure they completed each MTB trial.
An additional 30-s open section was placed at the start of the first 22.50-min lap to enable the cyclists to replicate the fast start in a typical MTB race. Subjects were informed that performance would not be measured at this point after lap 1. Heart rate was monitored continuously with a Polar Sports Tester heart rate monitor (Polar Electro OY, Kempele, Finland) during the MTB test.
Subjects arrived at the laboratory 3 h before the MTB test for both carbohydrate trials and consumed either the low- or high-carbohydrate meal (Table 1). The meals differed in carbohydrate concentration (1 g CHO·kg−1 BM (LC) vs 3 g CHO·kg−1 BM (HC)) but were equivalent in volume of food and fluid. The HC meal was supplemented with MODUCAL TM (Mead Johnson & Company, U.S.), a tasteless 100% maltodextrin powder that was added to the cordial and the tomato-based pasta sauce. The GI of the mixed meal was approximately 87 (21). The rice was pre-cooked and reheated using a microwave immediately before the meal was served, and a low-J cordial used in the LC trial. Subjects were not able to distinguish between the two meals. Subjects were given a total volume of 15 mL·kg−1 BM of artificially sweetened (placebo) drink during the MTB test to minimize dehydration.
Measurement of oxygen uptake (V̇O2).
The RER and V̇O2 during the experimental trials were measured in both steady state cycling (between 6 and 10 min) and a selected segment of stochastic (intermittent) cycling (between 18 and 21 min of each 22.50-min lap). The V̇O2 and RER were measured using a custom-built automated Douglas bag gas analysis system (AIS, Australia) that incorporated a CO2 and O2 analyser (Ametek, Applied Electrochemistry, U.S.) and twin Tissot gasometers (Warren. E. Collins Inc., Massachusetts) interfaced to a PC by Optical Rotary Encoders (RS 341-597, Switzerland). Carbohydrate oxidation was calculated from the mean V̇O2, and RER values in the steady state and from the peak value attained in the interval section during each stage of the experimental protocol by using a conversion table for nonprotein respiratory quotient (17).
Venous blood (5 mL) was collected via a cannula every 30 min for 2 h after meal consumption, during the steady-state period of the MTB test in each lap of HC and LC (at 10 min into each lap) and immediately before and after exercise. A 200-μL sample of blood was extracted for pH, PCO2, and PO2 analysis (CIBA-Corning 278 Blood Gas System, Ciba Corning Diagnostics Ltd., Suffolk, UK). The remaining blood was placed in a tube with serum separating gel then, after clotting, centrifuged for 5 min at 4500 rpm. The serum was divided into two tubes and stored at −80°C until analysis for free fatty acid (WAKO, Japan), glucose (Hitachi 911 automated analyzer), and insulin (spectrophotometric ELISA technique).
Gastrointestinal comfort and rating of perceived exertion (RPE).
Gastric comfort was measured before the preexercise meal, immediately pre-MTB test and at 18 min into the lap during each 22.50-min lap. The level of comfort/ discomfort in the stomach was established by asking the subject to point to a number on a scale (hungry, 1; comfortable, 2; slight discomfort, 3; uncomfortable, 4; very uncomfortable, 5) as outlined previously by Zachweija et al. (29). RPE was measured twice during each lap (at 10 min and 18 min into each lap) using the Borg scale (3).
Data were analyzed in a complete random block design using two-way and three-way analysis of variance with repeated measures. Post hoc Newman-Keuls test and critical ranges was used to determine the level of statistical significance between means and at specific times during the MTB test. A P < 0.05 was required for statistical significance. Correlation matrices were used to assess whether there was any relationship between variables. Statistical analysis was carried out using STATISTICA© for Windows Release 6.0 (1997, StatSoft Inc., OK).
The MTB test was shown to be physiologically and metabolically representative of true MTB racing based upon heart rate recordings that corresponded with the expected intensity induced by preprogrammed power outputs. The intensity achieved during the open sections of each lap also corresponded with near maximal efforts, which replicates the variable nature of MTB racing. Figure 2 represents the work done during each performance trial. Comparisons of the total work (kJ) done during each 30-s interval (“open section”) were not different between HC and LC trials. The overall performance during HC was 3% greater than LC equating to a 2 min 48 s time advantage during the HC trial; however, this did not reach statistical significance. There was a significant interaction observed for performance expressed in kilojoules over the four laps for HC and LC trials. Specifically, performance expressed as the mean work done during each lap was significantly greater during lap 1 of LC trial (12.0 ± 2.2 kJ vs 11.3 ±1.9 kJ, P = 0.03) but lower in lap 4 (10.7 ± 2.1 kJ vs 12.2 ± 1.5 kJ, P = 0.02) compared with the HC trial.
Blood glucose, insulin, and FFA.
Figure 3 represents the serum glucose responses across HC and LC trials that followed expected trends for the first 120 min. At this point, glucose concentration began to fall in the HC trial until a nadir at 10 min into the MTB test (190 min postprandial). In contrast, levels continued to rise in the LC trial after 120 min until 180 min and subsequently throughout the MTB test. The serum glucose level was significantly higher immediately pretest (4.8 mmol·L−1 and 3.8 mmol·L−1, respectively, P = 0.04) and at 10 min in the LC trial compared with the HC trial (4.7 and 3.7 mmol·L−1, respectively, P = 0.01). There were no significant differences in serum glucose levels from lap 1 onward.
Figure 4 represents the blood insulin response beginning premeal onward. Insulin levels peaked 30 min postprandial in both trials; however, the response was significantly greater in HC from 30 min until 120 min (P = 0.01). Insulin continued to decline to near basal levels at the commencement of exercise at 180 min, and there was a trend to remain higher in HC throughout the MTB test (NS).
There was no significant difference in FFA response to different quantities of carbohydrate ingestion (Fig. 5). The FFA levels tended to rise slowly throughout the exercise component of each trial. The FFA levels at 120 min in the LC trial increased to a peak at the commencement of exercise (180 min) but were suppressed in the HC trial.
Gastrointestinal comfort and rating of perceived exertion (RPE).
Gastric discomfort for both LC and HC over the duration of the MTB test duration increased significantly (P = 0.01) (Fig. 6). However, carbohydrate quantity had no effect on gastric comfort, as there was no difference between trials. The RPE also increased during both carbohydrate trials but was not affected by the quantity of carbohydrate ingested pre-MTB test (Fig. 7). Subjects indicated a higher RPE as they progressed through the MTB test, such that RPE at the end of the MTB test was significantly higher than during lap 1 for both HC and LC meals (P < 0.05).
Mean fluid intake during the exercise component of HC was 1001 mL ± 280 mL and for LC was 1250 mL ± 800 mL, with no significant difference between the trials. There was also no significant difference between trials as to the degree of dehydration incurred (average BM loss was 1.01 kg ± 0.28 kg for HC and 1.25 kg ± 0.31 kg for LC).
Analysis of food diaries showed that subjects met the recommended daily intake of macro and micronutrients (18) in the 2 d before each trial. In addition, there was no significant difference between the trials for average daily nutrient intake. The mean totals for macronutrient intake over the 2 d before each trial was HC total energy intake 12,409 kJ ± 1854 and 13,610 kJ ± 3017 for LC, total carbohydrate intake for HC 430 g ± 89 g and 472 g ± 121 g for LC, total protein intake for HC 117 g ± 8 g and 131 g ± 46 g for LC, and total fat intake 99 g ± 24 g and 107 g ± 23 g for LC.
Respiratory measurements and carbohydrate oxidation.
Total energy expenditure for HC and LC was similar for the duration of the MTB test (261.6 kJ for HC and 254.7 kJ for LC). RER was measured during the steady state of each lap, and there was a significant main effect of lap (P < 0.001) and diet (P = 0.031) where RER remains higher throughout HC, although the interaction was not significant (P = 0.29) (Fig. 8). There was also a significant main effect of lap (P = 0.002) and diet (P = 0.024) for CHO oxidation but again the interaction was not significant (P = 0.23) (Fig. 9). V̇O2 did not differ significantly between trials (Fig. 10). Fat oxidation appeared to be higher throughout LC trial compared with the HC trial although this was not significant (P = 0.18)
The results of this study have shown that a simulated MTB race performance was slightly but not statistically significantly improved by ingesting 3.0 g compared with 1.0 g CHO·kg−1 body weight. Although the overall performance gain (3% or 2.48 min) in the HC trial was relatively small, at the elite level this time difference could win or lose a race. The results of this study also indicate that carbohydrate intake 3 h before exercise influences pacing strategy. Because the typical senior male MTB race lasts 2 h 15 min to 3 h (27), extrapolating results from the last lap mean performance differences may become greater over time.
Carbohydrate ingestion influences blood substrate levels and muscular substrate utilization by increasing blood glucose oxidation and suppressing FFA utilization (7,9,13,23,28). These metabolic effects may help to explain the subtle performance differences observed in the current study between HC and LC. Borderline clinically low serum glucose levels (3.5 mmol·L−1) were observed in several subjects in HC immediately before and at 10 min into (190 min postprandial, Fig. 3) the MTB test. Although no assessment of symptoms was made, hypoglycemia may have contributed to a slower start in the HC trial. The borderline low blood glucose levels experienced during the HC trial did not appear to impair overall performance.
Despite a slower start during the HC trial, performance was maintained more consistently during laps 2 and 3 then improved rapidly in lap four (Fig. 2). The improvement in lap 4 coincided with an elevation in serum glucose. In contrast, subjects in the LC trial started out with a much harder effort but were not able to sustain the effort throughout the MTB test even though serum glucose was higher than the HC trial at the start and throughout exercise. This is not consistent with studies that show higher blood glucose as an indicator of carbohydrate availability. Kiens (13) suggested that the level of blood substrates might not be a true indicator of substrate usage at the muscular level. This would explain observations in the current study where both serum glucose and FFA levels remained slightly higher during LC, yet performance was relatively impaired overall. Because muscle substrates were not measured in the current study, actual substrate usage during the trials remains speculative.
Wright et al. (28) and Sherman et al. (23) showed that a large carbohydrate meal consumed 3–4 h preexercise elevated blood insulin levels before and during exercise. Typically, exercise dampens the insulin response and muscle glucose uptake is maintained by exercise-induced mechanisms (15). Blood insulin levels were suppressed by the onset of exercise in the current study; however, levels did remain slightly elevated throughout exercise during the HC trial compared with the LC trial. Increased blood insulin levels are a potent inhibitor of lipolysis, indicated by a suppression of blood FFA levels (16). Therefore, suppression of whole-body lipolysis as indicated by reduced FFA levels (13) may in turn facilitate increased carbohydrate availability during intense exercise.
Greater carbohydrate availability is thought to raise the carbohydrate oxidation rate during exercise (19,24). In the current study, carbohydrate oxidation rate was slightly, but not significantly, higher throughout HC compared with LC (Fig. 9). Even though the difference was subtle, a higher oxidation rate in HC may have enabled a better performance toward the end of exercise (Fig. 2) (23). Because FFA levels increased at a faster rate in LC toward the end of the MTB test, a shift toward greater fat utilization may have resulted from reduced carbohydrate availability and oxidation. Such substrate influences after ingestion of carbohydrate are consistent with findings from other studies (6,19,23,24,28). The results of this study show that ingestion of different quantities of carbohydrate preexercise can exert subtle changes in substrate use, which may explain variations in performance between HC and LC at different stages during exercise.
There were two factors that may have contributed to the subtle performance differences observed during this study. First, the nutrition analysis showed subjects had adequate nutrition during the days before testing in addition to the preexercise meal. It is possible that carbohydrate stores were sufficient to last 93 min of high-intensity intermittent exercise. Second, as mentioned previously, upper-body involvement through technical bike handling would undoubtedly contribute to overall energy expenditure in racing; it was not possible to simulate this in the laboratory-based MTB test protocol. Differences in performance during the current study may have been more obvious if the performance test had lasted longer or if there was field simulated upper-body involvement. This is further supported by extrapolating the serum glucose results from the last lap at 79 min (Fig. 3), indicating that glucose supply was adequate in HC and only slightly reduced in LC. The test duration was chosen because the subjects were in their winter off-season maintenance phase and could not expected to be in condition to complete summer MTB race times and distances (greater than 2 h).
In the current study, MTB racing was simulated using stochastic style cycling, and performance was measured throughout the trials. This made it possible to observe how carbohydrate ingestion may affect strategies during exercise as well as overall performance. Most other studies measure performance toward the end of steady state exercise or by the recording the amount of time that cyclists are able to ride above a certain intensity (19,24,28). Such methods do not replicate the true nature of MTB racing, which is high intensity and intermittent in nature. Palmer et al. (22) found that performance was better during steady-state cycling compared with stochastic cycling at the same average intensity, highlighting that performance can vary according to the nature of the cycling event. Furthermore, Laforgia et al. (14) found that the energy demand was higher for runners engaging in intermittent high-intensity exercise bouts compared with steady-state submaximal exercise. Using a parallel argument, it could be assumed that substrate use may alter according to a change of exercise intensity. Therefore, test specificity may provide a more comprehensive understanding of physiologic and metabolic responses during exercise.
High-intensity exercise performance may be influenced by gastrointestinal comfort. In this study, 3 h separated the meal from exercise to allow for digestion and absorption of the meal (28). The LC meal was chosen for this study because it represents a similar quantity of carbohydrate contained in a “light”/small meal or snack often consumed by MTB racers before an event. The HC meal was chosen for comparison because conceptually it is a tolerable amount to ingest 3 h before racing. Because there was no difference in the level of discomfort between trials (Fig. 6), it can be assumed that the carbohydrate quantity did not alter gastrointestinal comfort and hence did not influence performance. In contrast, gastrointestinal comfort may be related to RPE because this increased in parallel with the level of gastric discomfort. Further investigation is required to clarify this relationship.
In summary, this study has shown that a high-carbohydrate (3.0 g·kg−1 body weight) preexercise meal has a potentially greater benefit to performance than a smaller meal taken 3 h before intermittently intense endurance exercise (>90 min). The mechanism for overall improved performance appears to be attributable to a greater carbohydrate oxidation rate due to increased carbohydrate availability from serum glucose. The temporary decline in performance early during exercise in the HC trial was probably a result of low serum glucose and raised insulin levels immediately before and early during exercise. However, these results have raised the question of whether the physiologic changes may have promoted a pacing strategy thus contributing to potential performance benefits. Further research is warranted to investigate whether a high carbohydrate meal can benefit performance over true mountain bike race distances.
The Australian Institute of Sport (AIS) and the University of Wollongong (UoW) supported the research project. The authors would also like to thank H. Lee, E. Lawton, G. Slater, D. Pyne, and L. Burke for their contributions.
1. Anantaraman, R., A. Carmines, A. Gaesser, and A. Weltman. Effects of carbohydrate supplementation on performance during 1 hour of high-intensity exercise
. Int. J. Sports Med.
2. Berry, M., T. Koves, and J. Benedetto. The influence of speed, grade and mass during simulated off road bicycling. Appl. Ergon.
3. Borg, G. Perceived exertion: a note on “history” and methods. Med. Sci. Sports.
4. Cheng, B., H. Kuipers, A. C. Snyder, H. A. Keizer, A. Jeukendrup, and M. Hesselink. A new approach for the determination of ventilatory and lactate thresholds. Int. J. Sports Nutr.
5. Coyle, E. Timing and method of increased carbohydrate intake to cope with heavy training, competition and recovery, J. Sports Sci.
6. Coyle, E. Fuels for Exercise
, Recent Advances in Sports Nutrition. Gatorade Sports Science Institute, 1997, pp. 3–5.
7. Coyle, E., A. Coggan, M. Hemmert, R. Lowe, and T. Walters. Substrate usage during prolonged exercise
following a preexercise meal. J. Appl. Physiol.
8. Febbraio, M., and K. Stewart. CHO feeding before prolonged exercise
: effect of glycemic index
on muscle glycogenolysis and exercise
performance. J. Appl. Physiol.
9. Foster, C., D. Costill, and W. Fink. Effects of preexercise feedings on endurance performance. Med. Sci. Sports
10. Goodpaster, B., D. Costill, W. Fink, et al. The effects of preexercise starch ingestion on endurance performance. Int. J. Sports Med.
11. Hawley, J., and L. Burke. Effect of meal frequency and timing on physical performance, Br. J Nutr.
12. Impellizzeri, F., and A. Sassi, M. Rodriguez-Alonso, P. Mognoni, and S. Marcora. Exercise
intensity during off-road cycling competitions. Med. Sci. Sports Exerc.
13. Kiens, B. Effect of endurance training on fatty acid metabolism: local adaptations. Med. Sci. Sports Exerc.
14. La Forgia, J., R. Withers, N. Shipp, and C. Gore. Comparison of energy expenditure elevations after submaximal and supramaximal running. J. Appl. Physiol.
15. Marmy-Conus, N., S. Fabris, J. Proietto, and M. Hargreaves. Preexercise glucose ingestion and glucose kinetics during exercise
. J. Appl. Physiol.
16. Martin, W. Effect of endurance training on fatty acid metabolism during whole body exercise
. Med. Sci. Sports Exerc.
17. McArdle, W., F. Katch, and V. Katch. Exercise Physiology: Energy, Nutrition, and Human Performance
, 3rd Ed. Philadelphia: Lea and Febiger, 1991, pp. 150–157.
18. National Health and Medical Research Council. Recommended Dietary Intakes for Use in Australia.
Canberra, Australia: Australian Government Publishing Service, 1991, p. 26.
19. Neufer, P., D. Costill, M. Flynn, J. Kirwan, J. Mitchell, and J. Houmard. Improvements in exercise
performance: effects of carbohydrate feedings and diet. J. Appl. Physiol.
20. Norton, K., and T. Olds. Anthropometrica
. Australia: University of New South Wales Press, 1996, pp. 396–398.
21. O’Connor, H., J. Brand Miller, S. Colagiuri, and K. Foster-Powell. Pocket Guide to the G.I. Factor and Sports Nutrition
. NSW Australia: Hodder and Stoughton, 1997, pp. 32, 72–93.
22. Palmer, G., T. Noakes, and J. Hawley. Effects of steady-state versus stochastic exercise
on subsequent cycling performance. Med. Sci. Sports Exerc.
23. Sherman, W., G. Brodowicz, D. Wright, W. Allen, J. Simonsen, and A. Dembach. Effects of 4 h preexercise carbohydrate feedings on cycling performance. Med. Sci. Sports Exerc.
24. Sherman, W., M. Peden, and D. Wright. Carbohydrate feedings 1 h before exercise
improves cycling performance. Am. J. Clin. Nutr.
25. Thomas, D., J. Brotherhood, and J. Brand. Carbohydrate feeding before exercise
: effect of glycemic index
. Int. J. Sports Med.
26. Thomas, D., J. Brotherhood, and J. Brand Miller. Plasma glucose levels after prolonged strenuous exercise
correlate inversely with glycemic response to food consumed before exercise
. Int. J. Sports Nutr.
27. Wilber, R., K. Zawadzki, J. Keamey, M. Shannon, and D. Disalvo. Physiological profile of elite off-road and road cyclists. Med. Sci. Sports Exerc.
28. Wright, D., W. Sherman, and A. Dernbach. Carbohydrate feedings before, during, or in combination improve cycling endurance performance. J. Appl. Physiol.
29. Zachweija, J., D. Costill, J. Widrick, D. Anderson, and G. McConell. Effects of drink carbonation on the gastric emptying characteristics of water and flavoured water. Int. J. Sports Nutr.
30. Zhou, S., and S. B. Weston. Reliability of using the D-max method to define physiological responses to incremental exercise
testing. Physiol. Meas.
Keywords:©2004The American College of Sports Medicine
STOCHASTIC CYCLING; GLYCEMIC INDEX; EXERCISE; GASTROINTESTINAL COMFORT