INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

DURING PROLONGED (>90 min) continuous, moderate-intensity [70-80% of maximal oxygen uptake (O_{2 max})] cycling, the onset of fatigue is associated with very low (~100 mmol/kg dry wt) muscle glycogen concentrations (6, 8, 11) and hypoglycemia (6, 11). Muscle and liver glycogen stores can be increased above normal resting values by combining an exercise taper with the intake of large quantities of dietary carbohydrate (CHO) in the days before an endurance event (29). A recent review (18) concluded that such CHO loading improves endurance by ~20% during prolonged continuous exercise at a fixed submaximal work rate. Furthermore, performance of a fixed distance or workload of >90 min is also increased by ~2-3% after CHO loading.

The common assumptions from these studies are that 1) CHO loading acts exclusively by increasing muscle glycogen stores and 2) these studies establish the value of CHO loading for a wide variety of competitive sporting activities. However, the certainty of these conclusions is influenced by aspects of the experimental designs used in many of these studies.

First, to our knowledge, only one CHO-loading study (17) has included a placebo control in the research design. Without placebo control, the influence on subsequent exercise performance of psychological factors, in particular the expectation of an effect by both researchers and subjects, cannot be excluded. Interestingly, that placebo-controlled study failed to show any benefit of CHO loading on performance, albeit in a 1-h cycling time trial (TT) in which any effect would seem unlikely.

Second, some studies have tested performance under conditions of an overnight fast and/or intake of water during the endurance test. These conditions are neither typical nor recommended for optimizing performance in an endurance event.

Importantly, little is known about the effects of CHO loading on performance when there are continual changes in work rate, an exercise pattern typical of many competitive sports. For example, road cycling races are characterized by periods of sustained steady-state exercise punctuated by bouts of high- and low-intensity work in accordance with the course profile, terrain, environmental conditions, and riding strategies of the cycling group (25). Exercise tasks that include such alternating work bouts frequently elicit physiological responses that are different from those in which work, power, or duration is held constant (26). Thus stochastic exercise might be expected to cause a different pattern of fuel utilization compared with continuous, moderate-intensity work. It is only recently that laboratory studies have attempted to simulate actual race conditions in the laboratory (26, 28).

Accordingly, the aims of this study were to examine the reported ergogenic action of CHO loading on cycling performance tested on a reliable laboratory protocol (28) that simulates the demands of competitive road racing and, second, to exclude a possible placebo effect of this CHO loading on exercise performance. The study was also designed to include other dietary practices that are considered optimal for the performance of competitive road cycling.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. Seven endurance-trained male cyclists and triathletes accustomed to riding for prolonged periods (3-4 h) participated in this study. At the time of the investigation, these subjects were riding between 250 and 500 km/wk. Before commencement of the trial, all subjects were informed that the purpose of the investigation was to test two different sports supplements designed for race preparation. Subjects gave written informed consent in accordance with the guidelines outlined by the American College of Sports Medicine (2). The characteristics of the subjects are shown in Table 1.

Preliminary testing. On their first visit to the laboratory, subjects were tested for peak oxygen uptake (O_{2 peak}) (16) and peak sustained power output on their own bicycles, which were mounted on the Kingcycle ergometer (described below). After a 5- to 10-min warm-up at a self-selected intensity, the test commenced at a workload of 200 W; the load was then increased by 20 W/min until the subject could no longer maintain the required power output. The subject's peak power was taken as the highest average power during any 60-s period of the exercise test. During these incremental tests to exhaustion, subjects were requested to remain in a seated position.

Dietary intervention. Individual food plans were constructed for each subject on the basis of body mass (BM) and food preferences. Each subject was supplied with his food intake for the 72-h period before each trial. Menu plans were provided in written form, and the food assigned to each meal was individually prepared and packaged so that the need for further preparation by subjects was minimized. Subjects were required to record their actual food and fluid intake on dietary logs to account for any portion of meals left unconsumed or for any additional intake. Although some food items differed between subjects, each cyclist received identical breakfast, lunch, and dinner menus for both of his trials, with the CHO intake from these meals designed to provide 6 g CHO · kg BM¹ · day¹. Snacks were also provided for each day and provided the source of differentiation between trials. On the CHO loading diet, subjects received 1,200 ml of water each day and a number of sports bars (Gijima, Sasko, Paarl, South Africa) calculated to provide an additional 3 g CHO · kg BM¹ · day¹. Each sports bar had a composition of 27 g CHO, 6.5 g fat, and 2.7 g protein. The sports bars were used to mimic what is used by competitive athletes during training and competitive events. On the placebo trial, daily snacks were provided in the form of 1,200 ml of an artificially sweetened low-calorie drink that was described to subjects as a "CHO-loading" drink.

Experimental trials. Each subject completed a random crossover design of two experimental trials separated by 7 days. Subjects performed their TTs at the same time of day under standard laboratory conditions (~20°C, 55% relative humidity). Subjects were requested to perform the same type of training for the duration of the trial and to refrain from heavy physical exercise on the day before a TT. Training diaries were kept to assess compliance to this condition. On the morning of an experiment subjects reported to the laboratory between 0700 and 0800, 12-14 h after an overnight fast. At this time the Kingcycle ergometer was calibrated, and then subjects consumed a standard breakfast providing 2 g CHO/kg BM. After the subjects rested quietly for 105 min, a preexercise muscle biopsy sample was taken from the vastus lateralis of the right leg according to the technique of Bergstrom (5) as modified by Evans et al. (12). After this procedure, the subjects mounted their cycle and began a 5-min self-paced warm-up.

1

1

1

1

Analytic techniques. Muscle samples were subsequently freeze-dried; dissected free of blood, connective tissue, and fat; and weighed. Muscle glycogen content was determined after acid hydrolysis by a hexokinase method (21). The CV for this assay in our laboratory is <5% for duplicate glycogen assays of a single piece of muscle and <7% for assays of the glycogen content of separate pieces of the same muscle biopsy sample (17).

Statistical analyses. Differences in the sprint times and sprint power outputs and in the glycogen concentrations were examined by using repeated-measures ANOVA, whereas glycogen utilization, total 100-km time, and total power outputs were compared by using Student's t-tests. A Sheffé's post hoc test was used to assess differences revealed by the ANOVA. All data are reported as means ± SD, and all calculations were performed by using Statistica for Windows (version 6, Statsoft, Tulsa, OK).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Training diaries collected on the morning of each trial showed that all subjects complied with the pretrial training conditions: each subject undertook identical exercise sessions before each of his trials and achieved a training taper over this time. Reported dietary intake for the 72 h before each trial (Table 2) showed that subjects achieved the CHO intake goals predetermined for the study. Significantly greater CHO intake was consumed with the CHO-loading diet than with the placebo diet (9 vs. 5.8 g CHO · kg BM¹ · day¹), with this additional CHO being provided exclusively by the sports bar. Muscle glycogen levels were significantly higher after CHO loading (Table 3). Total energy, protein, and fat intake were also significantly higher with CHO loading (Table 2). This resulted from the high protein and fat content of the energy bar. Despite higher preexercise muscle glycogen concentrations after CHO loading, glycogen utilization during the 100-km TT was not significantly different between trials but tended to be greater after CHO loading. Postexercise glycogen concentrations were not significantly different between treatments. BM changes over the TT were not different between treatments (1.5 ± 0.9 vs. 1.1 ± 0.5 kg for CHO loading and placebo, respectively), indicating that a mild but similar degree of fluid deficit occurred over the TT.

There was no significant difference in overall TT performance between treatments (147.5 ± 10 vs. 149.1 ± 11.0 min, for CHO loading and placebo, respectively; P = 0.4). The time difference between the CHO-loaded and placebo trial was 1.1% (95% CI = 1.6-3.6%). The average power output over the total TT was not different between trials (258 ± 40 vs. 253 ± 40; P = 0.4). There were no differences between groups for HR and rating of perceived exertion data during each trial.

There were also no significant differences between groups for either the time to complete (Fig. 1) or the mean power (Fig. 2) of each of the nine 1- and 4-km sprints. Time and power data for the first and final of each of these sprints and for the total 100-km TT are presented in Table 4. In both groups, mean power output declined in both 1- and 4-km sprints over the duration of the 100-km TT, with a corresponding increase in time to complete the sprints.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Time taken by carbohydrate (CHO) and placebo groups in 4-km (A) and 1-km (B) sprints during 100-km time trial. Values are means ± SD.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Power output by CHO and placebo groups in 4-km (A) and 1-km (B) sprints during 100-km time trial. Values are means ± SD.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The major finding of this study is that, when tested against placebo, CHO loading did not improve performance during a prolonged (~2.5-h) self-paced TT that included high-intensity workbouts and CHO ingestion before and during the trial. This finding contrasts with the majority of studies that show that CHO loading enhances performance during prolonged exercise >90 min under a wide variety of laboratory and field conditions (18). A number of characteristics, unique to this trial, could explain this unexpected finding.

First was the use of a double-blind placebo-controlled design. This is an essential characteristic of any intervention study measuring an effect that might be influenced by psychological factors and is an accepted requirement in studies of CHO feeding during exercise. The failure of the majority of CHO-loading studies to include a placebo control group reduces the certainty of the conclusions inferred from their findings. Because endurance athletes are now well educated about the principles of CHO loading (9) and the reported benefits of this popular practice, it is likely that many subjects participating in current studies of CHO loading would expect to perform better after that intervention, thus introducing a psychological bias.

To date, only one other study of CHO loading has attempted to include a full placebo control, by providing subjects with their food intake during the pretrial period and disguising the true CHO-enriched menu by matching the other diet with a low-energy placebo supplement (17). That study also failed to find an improvement in performance associated with elevated preexercise glycogen stores, although their exercise protocol involved only ~60 min of cycling. Although in this trial a third nonplacebo group was not included, this does not alter the results of our present trial because subjects were blinded to the nature of the CHO or placebo ingestion.

Our present study therefore invites the possibility that the subjects' knowledge that they were CHO loading could be an important determinant of the measured ergogenic effect of CHO-loading studies that are not placebo controlled. Future studies of CHO loading must be appropriately controlled either to refute or support this unexpected interpretation.

A second characteristic that may have influenced the outcome of this trial was the exercise protocol used to evaluate cycling performance. The exercise task was designed to mimic the requirements of a 100-km road cycle race (25) rather than to have subjects exercise at fixed submaximal workloads to exhaustion as in previous studies of cycling (1, 6, 8, 20) or running (7, 14, 22). In other studies, CHO loading has been shown to reduce the time taken to complete a prolonged running or cycling task (21, 31, 32). Under these conditions, enhanced performance after CHO loading was achieved by the maintenance of greater average speeds because the rate of decline in power output during the latter stages of the task was reduced.

Although these protocols involved self-paced efforts and the possibility of a changing workload, it is unlikely that the typical pacing strategies used in these trials involved the random injection of extremely high-intensity efforts, as are characteristic of mass start road cycling events (25). Thus it is possible that elevated pretrial glycogen concentrations might increase performance in those forms of exercise in which the intensity is relatively more constant but not in the stochastic form of exercise evaluated in this trial, in which there are bouts of exercise of very high intensity. Indeed the inclusion of repeated high-intensity sprints in our TT protocol produced postexercise muscle glycogen concentrations as low as have been reported in any previous study (6, 8, 11).

Third, although preexercise glycogen concentrations were significantly elevated by ~20% at the beginning of the CHO-loaded TT compared with the placebo TT, this increase may not have been sufficient to ensure a physiological benefit to the performance of the repeated high-intensity sprints and therefore the overall time to complete the TT. However, the 20% increase in muscle glycogen concentrations induced by CHO loading is in agreement with values previously reported by other investigators (17, 22, 27) and was sufficient to enhance performance in these studies (22, 27). It should, however, be remembered that the original CHO-loading studies compared the performance of subjects with very large (~100%) differences in initial muscle (and liver) glycogen concentrations (6, 21) because the comparison was usually between high- and low-CHO diets rather than between high- and normal-CHO diets as in this study.

A fourth feature of our study design was the inclusion of several dietary strategies to optimize CHO availability during performance. CHO was consumed in a preevent meal and during the TT in accordance with current sports nutrition guidelines (3, 4) and the typical practices of competitive cyclists.

Several studies have partially or systematically studied the interaction of CHO loading and CHO ingestion during exercise on exercise performance. Flynn and co-workers (13) reported that CHO intake during cycling did not improve performance in CHO-loaded subjects during 2 h of self-paced cycling at ~80% O_{2 max}. In contrast, the study of Kang et al. (20) showed that CHO feeding in CHO-loaded subjects enhanced performance during ~3 h of exercise at 70% O_{2 max}. The study of Widrick et al. (31) evaluated performance during a self-paced 70-km TT when preexercise glycogen content was manipulated by CHO loading and CHO or placebo was ingested during the ride (31). CHO ingestion during the TT prevented a decline in blood glucose concentrations regardless of the starting muscle glycogen content. Performance was best in subjects who were CHO loaded and who ingested CHO during exercise and was worst in those who neither CHO loaded nor ingested CHO during exercise. However, performance differences were relatively small, especially between the CHO-loaded group who ingested placebo during exercise and the CHO-depleted group who ingested CHO during exercise. This occurred despite the fact that the CHO-depleted group had exercised for 45 min 24 h before the performance trial, whereas they rested for 48 h before the CHO-loaded trial. Thus, whereas CHO ingestion during exercise did not influence performance in CHO-loaded subjects, CHO ingestion enhanced the performance of CHO- depleted subjects, perhaps by preventing the development of hypoglycemia (mean blood glucose concentrations of ~3.2 mmol/l) and a synchronous fall in power output in subjects who did not CHO load before exercise.

Higher blood glucose concentrations with CHO ingestion than with placebo were also associated with superior exercise performance in the CHO-loading studies of Kang et al. (20). Similarly, superior performance in a 30-km run after CHO loading was associated with higher blood glucose concentrations during exercise than when only water was consumed (33). As in the study of Widrick et al. (31), reductions in running speed fell synchronously with the blood glucose concentrations in both CHO-loaded and control groups. In contrast, performance was not enhanced by CHO ingestion in the trial of Williams et al. (32), even though blood glucose concentrations were higher with ingestion of CHO than placebo during exercise. The opposite result was reported by Tsintzas et al. (30), where performance was enhanced by CHO ingestion during exercise even though blood glucose concentrations were essentially unaffected.

Accordingly, performance in this trial may have been the same in CHO-loaded and placebo groups because blood glucose concentrations may have been similar in both trials due to the ingestion of CHO during the ride. Hence the optimization of CHO ingestion before and during exercise in this trial may have negated any additional beneficial ergogenic effect of CHO loading alone.

Finally, although CHO loading did not produce a significant ergogenic effect in this trial, the results should be examined in light of what might be meaningful in a competitive sports event. In this study, CHO loading was associated with a 1.1% improvement in TT performance, with true differences in the performance of a similar population likely to range from a 1.6% decrement to a 3.6% improvement. At best, this improvement would not greatly change the outcome of the performances of the "back of the pack" cyclists; even a 4% improvement would not move these athletes to the front of the field. However, such an improvement, if real, is meaningful to the top cyclists in a race in that it would enhance the likelihood of an improvement in finishing order (19).

Hopkins and colleagues (19) have calculated that the smallest intervention likely to enhance the performance of a top-finishing athlete is 0.4-0.7% of the typical within-athlete CV in performance between events. The typical within-athlete CV for cyclists competing in such road events is currently not known. However, in other sports such as road running and track cycling, this CV is ~1-2% for good competitors and often <1% for the best international competitors (W. G. Hopkins, personal communication). Thus a 1.1% improvement in performance is likely to be worthwhile to an elite road cyclist. Should performance in mass- start road cycling races be less reliable than for other sports (e.g., CV >4%), then the impact of our effect would need to be recalculated. Power analysis shows that a sample size of ~30 subjects would be needed in a similar study to reduce the confidence intervals of the true performance change to approximately ±1%, that is ~0-2%, and therefore discount the likelihood of a meaningful effect.

In summary, this study shows that a CHO-loading protocol that increased preexercise muscle glycogen concentrations resulted in a minimal effect on the performance of a 100-km TT involving high-intensity sprints when CHO was ingested before and during the event according to contemporary sports nutrition guidelines. Although we cannot completely discount the possibility that CHO loading has a worthwhile effect on the performance of competitive cyclists under conditions similar to our trial, our data suggest that this effect is small and, at best, only likely to rearrange the finishing order of the top cyclists in the field. Furthermore, this study raises the possibility that part or all of the ergogenic effect of CHO loading reported in previous studies could result either from a placebo effect or from higher preexercise liver glycogen stores that could delay the onset of hypoglycemia during prolonged exercise. If this is the case, CHO ingestion during exercise, as included in this study, would minimize any ergogenic effect of CHO loading by preventing the development of hypoglycemia in persons beginning exercise with lower liver (and muscle) glycogen concentrations.