Eur J Appl Physiol (1992) 65 : 18-24 European ou.o, A p p l i e d P h y s i o l o g y

and Occupational Physiology © Springer-Verlag 1992

The effect of a high carbohydrate diet on running performance during a 30-km treadmill time trial

Clyde Williams, John Brewer, and Moya Walker

Department of Physical Education, Sports Science and Recreation Management, Loughborough University, Loughborough, LEll 3TU, UK

Accepted January 29, 1992

Summary. The purpose of the present study was to examine the influence of a high carbohydrate diet on running performances during a 30-km treadmill time trial. Eighteen runners (12 men and 6 women) took part in this study and completed a 30-kin time trial on a level treadmill without modifying their food intake (trial 1). The runners were then randomly assigned to a control or a carbohydrate (CHO) group. The CHO group supple- mented their normal diets with additional carbohydrate in the form of confectionery products during the 7 days before trial 2; the control group matched the increased energy intake of the CHO group by consuming addi- tional fat and protein. The mean (SEM) carbohydrate intake of both groups was 334 (22) g before trial 1, after which the CHO group consumed 566 (29) g. day-1 for the first 3 days and 452 (26) g. day- 1 for the remaining 4 days of recovery. Although there was no overall differ- ence between the performance times for the two groups during trial 2, the CHO group ran faster during the last 5 km of trial 2 than during trial 1 [3.64 (0.24) m's -1 vs 3.44 (0.26) m ' s - 1; p < 0.05]. Furthermore, the 6 men in the CHO group ran the 30 km faster after carbohydrate loading [131.0 (5.4) min vs 127.4 (4.9) rain; P<0.05], whereas there was no such improvement in times of the men in the control group. Blood glucose concentrations of both groups decreased below pre-exercise values dur- ing trial 1 (P<0.001), but only the control group had a decrease in blood glucose concentrations during trial 2 (P< 0.001). There were no differences between the con- centrations of plasma catecholamines of the control group during the two trials. However, the adrenaline concentrations of the CHO group were lower (P< 0.05) during trial 2 than during trial 1, even though they ran faster during trial 2. These results confirm that dietary carbohydrate loading improves endurance performance during prolonged running and that confectionery can be used as'an effective means of supplementing the normal carbohydrate intake in preparation for endurance races.

Offprint requests to: C. Williams

Key words: Diet - Endurance - Carbohydrate metabol- ism - Catecholamines - Running

Introduction

The benefits of increasing the carbohydrate stores of skeletal muscles and liver before prolonged exercise were reported over two decades ago (for review, see Hultman 1967). The focus of these studies was the influence of dietary carbohydrate loading on endurance capacity rather than on endurance performance. Endurance ca- pacity is defined as the time to fatigue at a fixed exercise intensity whereas, endurance performance is the time to complete a prescribed distance or work load. The one early study on the influence of carbohydrate loading on endurance performance showed that running times, dur- ing a 30-km cross-country race, improved after carbohy- drate loading (Karlsson and Saltin 1971). In contrast, a more recent study reported no improvements in the run- ning times of well-trained runners competing over a flat 20.9-km indoor course after carbohydrate loading (Sher- man et al. 1981). Their method of carbohydrate loading differed from the traditional approach in that it did not include 2-3 days on a low carbohydrate diet. Instead, their subjects tapered their training and increased their carbohydrate intake during the 4 days before competi- tion. The muscle glycogen concentrations achieved were the same as the values obtained after the traditional method of carbohydrate loading (Astrand 1967).

The question about what types of carbohydrates are most effective in carbohydrate loading diets is one which is often raised. Costill and colleagues suggested that simple carbohydrates are just as effective in replac- ing muscle glycogen stores during the first few days of recovery, but that complex carbohydrates may be more effective thereafter (Costill et al. 1981). On the other hand, Roberts et al. (1987) suggest that simple carbohy- drates are as effective as complex carbohydrates in res- toring muscle glycogen concentrations after prolonged exercise. Of course, the timing of carbohydrate intake as

19

well as the influence of different carbohydrates on blood glucose concentration (glycaemic index) play an important role in dictating the rate of glycogen resynthe- sis (Coyle 1991).

In a previous study we showed that endurance capac- ity during treadmill running can be improved by supple- menting normal mixed diets with either complex or sim- ple carbohydrates (Brewer et al. 1988). Simple carbohy- drates, in the form of confectionery, were used because on many occasions sportsmen and sportswomen do not have the opportunity to increase the carbohydrate con- tent of their meals. Confectionery products, therefore, provide a convenient way of increasing carbohydrate in- take on these occasions.

The purpose of the present study was to examine the effects of increasing the carbohydrate content of normal mixed diets with confectionery products on endurance performance during a simulated race over 30 km on a level treadmill. This distance is sufficiently long to de- tect fatigue, but short enough for the subjects to main- tain their motivation to achieve personal best times. A 7-day recovery period was chosen because this is often the maximum time an athlete has between races during the height of the road racing season.

Methods

Subjects. The 18 subjects (12 men and 6 women) who volunteered to take part in this study were experienced endurance runners and familiar with racing over marathon and half-marathon distances. The subjects were fully informed about the nature of the experi- ments and what was required of them before they volunteered to take part in this study. The experimental procedures employed in this study were approved by the Ethical Advisory Committee of Loughborough University.

Experimental design. The subjects were required to complete two treadmill runs (trial 1 and trial 2) over a distance of 30 km (18.64 miles), separated by a period of 7 days. During these trials, the subjects were encouraged to complete the 30-km runs as fast as possible. They controlled the treadmill speed with a small hand- held switch. After completing trial 1, the subjects were randomly assigned to one of two dietary groups (6 men and 3 women in each group). For the 7 days after the first run, one group (CHO group) was prescribed a diet high in simple carbohydrates which was de- signed to increase their carbohydrate consumption by 70°7o during the first 3 days and by 35% during the remaining 4 days. This was achieved by supplementing their normal diets with simple carbo- hydrates in the form of confectionery. The subjects in the control group maintained their normal carbohydrate intake during the 7- day period between the two trials but they consumed additional protein and fat in order to achieve energy intakes that were isoca- loric with the diets of the CHO group.

Procedures. The subjects were fully familiarized with laboratory procedures and completed two preliminary tests before undertak- ing the two 30-km time trials. Each individual's maximum oxygen uptake value (1702max) (Table 1) was measured on a motorized treadmill, using methods previously described (Williams et al. 1990). The second test assessed the oxygen cost of running over a range of submaximl speeds, and consisted of 16 min continuous running on a level treadmill. The running speed was increased ev- ery 4 min, and for the last minute of each 4-min period, expired air was collected through a low resistance respiratory valve and lightweight, wide bore (40 mm) tubing into 150-1 plastic Douglas

Table 1. The physiological characteristics of the 18 subjects as- signed to the control and carbohydrate (CHO) groups; values are means (SEM)

Group Age Mass J / 'O2 max ~rE . . . . Maximum (years) (kg) (ml. kg - 1 (1. min - 1) heart rate

• rain - 1) (beats • min -1)

Control 24 66.1 60.5 120.8 190 (1.2) (1.9) (1.6) (4.9) (2)

CHO 28 66.0 58.7 116.2 190 (1.9) (0.6) (1.6) (4.7) (3)

~rE . . . . Maximum expiratory volume; l?Oz . . . . maximum oxygen consumption

bags. The results from these two preliminary tests were used to calculate the running speeds equivalent to 70°7o of each individu- al's I20~ . . . .

Following the 3 days on their "normal" diets, the subjects ar- rived at the laboratory in the morning after an overnight fast. They were weighed and had chest electrodes attached, for moni- toring heart rates. Resting samples of expired air were collected for a period of 5 min, followed by a venous blood sample from an ante-cubital vein and capillary blood samples from the thumb of a pre-warmed hand. The subjects completed a 5-min "warm up" on the treadmill at running speeds equivalent to 60O/o I202m~. A 60-S expired air collection was obtained at the end of the warm-up and then the treadmill speed was increased to 70°7o 1202 . . . . for the start of the 30-kin run. This speed was an appropriate "guide-line" for each subject and it was maintained for the first 5 km. There- after, the runners could change the treadmill speed at any time during the time trials. They were, however, encouraged to com- plete the distance in the shortest possible time. The chosen speeds, distances and time elapsed were displayed on a computer screen in front of the treadmill and in view of the subjects (Williams et al. 1990).

Throughout the run, expired air samples were collected at 5- km intervals, along with capillary blood samples from the thumb. Immediately after completion of the 30-kin run, a venous blood sample was taken. Throughout the run, heart rates were moni- tored and recorded every 30 s. Stride rate was also measured every 5 km by recording the small changes in treadmill speed which oc- curred each time the feet of the subjects made contact with the treadmill belt. This caused a slight fluctuation in the voltage pass- ing through the treadmill speed indicator, which was amplified and linked to a flat bed chart recorder (model CR503: Lloyd). Stride length was calculated from stride frequency and running speed. Water, placed adjacent to the treadmill, was allowed ad li- bitum throughout the run, and a moist sponge was also available for cooling purposes. After completing the run the masses of the subjects were recorded before they were allowed to ingest further amounts of fluid or consume any food.

Mean (SEM) laboratory temperatures for trial 1 were 16.3 (0.4) °C and 15.0 (0.4) °C for the control and CHO groups, re- spectively. During trial 2 the corresponding values were 16.3 (0.5) ° C and 15.9 (0.6) ° C. Relative humidity values during trial 1 and trial 2 were 52 (2) O7o and 54 (2) °7o for the control group; the corre- sponding values for the CHO group trials were 66 (4)°7o and 59 (3) %, respectively. There were no significant differences between these values for the two groups.

Nutritional status. Two to 3 weeks before trial 1 the subjects com- pleted 7-day weighed food intake diaries• They were analysed to provide a quantitative description of each subject's "normal" dai- ly food intake (Paul and Southgate 1978). Using this information, the subjects were prescribed their normal diets during the 3 days before trial 1 to ensure that they did not change their CHO intake. In order to achieve this degree of nutritional control the subjects

20

Table 2. Daily energy and macro-nutrient intakes of the two groups during the 3 days before trial 1 (pre-T1), during the conse- cutive 3 (lst-3 days) and 4 days (2nd-4 days) before trial 2; values represent means (SEM)

Observation Energy Fat CHO Protein period (MJ) (g) (g) (g)

Control Pre-T1 12.1 126 334 109

(0.9) (11) (22) (6) lst-3 days 18.2" 248* 364 181"

(1.3) (19) (23) (15) 2nd-4 days 15.0' 187" 349 145"

(1.0) (15) (20) (9)

CHO Pre-T1 11.3 114 334 105

(0.7) (8) (18) (8) 1st-3 days 16.9" 158" 566* 120"

(0.9) (9) (29) (8) 2nd-4 days 13.9" 133" 452* 105

(0.8) (8) (26) (7)

* Denotes significant increases (P< 0.001) compared to pre-trial 1 values

also completed weighed food intake diaries during the 3 days be- fore trial 1, and during the 7 days between trial 1 and trial 2.

The dietary patterns for the two groups during the study are shown in Table 2. During the 3-day period immediately after trial 1, the CHO group increased their carbohydrate intake by 70 (3) % (P<0.001). The consumption of this additional food increased overall energy intake by 50 (3)% (P< 0.001). Over this same peri- od, there was no change in the carbohydrate intake of the control group, but they increased their energy intake by the same amount as the CHO group, which was largely the result of a 99 (6.3)% increase in fat consumption (P<0.001). Nevertheless, there were no significant differences between the energy intakes of 18.2 (1.3) MJ and 16.9 (0.9) MJ of the control and CHO groups respective- ly, over this 3-day period. During the following 4 days which pre- ceded trial 2, the CHO group reduced their carbohydrate intake to 452 (26) g-day -1. This change in their carbohydrate and energy intakes was still 35 (2)% and 24 (1)% (P<0.001) respectively, above their normal values. Over the same 4-day period, the con- trol group also reduced their energy intake to 25 (3)% (P< 0.001) above their normal intake, by reducing their intake of fat, but without changing their carbohydrate intake. There was no change in the body masses of the control group during the 7-day period between the two trials whereas, the CHO group gained 0.83 (0.2) kg (P< 0.001).

Analyses. The percentages of oxygen and carbon dioxide in the expired air samples were determined using a paramagnetic oxygen analyser (model 570A; Sybron Taylor) and an infra-red carbon dioxide analyser (model 303; Mines Safety Appliances). Prior to, and during each series of analyses, both of the analysers were cal- ibrated with a known calibration gas, room air and nitrogen. The volume of each expired air sample was determined by evacuating the contents of each Douglas bag through a dry gas meter (Parkin- son Cowan), which had been previously calibrated with a 600-1 Tissot spirometer (Collins, USA). The venous blood samples were collected in lithium heparin tubes and samples analysed for hae- moglobin concentrations by the cyanmethemoglobin method (Boehringer, Mannheim, FRG), and packed cell volumes using a microcentrifuge (Hawksley). Plasma samples were obtained by centrifugation of the venous blood at 2 ° C, and analysed for free fatty acids (FFA) and glycerol by methods previously reported (Williams et al. 1990), as well as for plasma catecholamines (Brooks et al. 1988). Capillary blood samples (25 ~tl) were depro-

teinised in 0.4 mol . l - I perchlorid acid, frozen at - 20 ° C and later analysed for lactate and glucose (Maughan 1982). Changes in plas- ma volume were calculated from the pre- and postexercise haemo- globin and haematocrit values, according to the method described by Dill and Costill (1974).

Statistical analyses. The analyses of the results were based on standard statistical techniques. Tests of homogeneity were carried out on the results. Parametric t-tests were used to examine differ- ences between results which were distributed normally, whilst the non-parametric Wilcoxon matched pairs test was used to examine differences between sets which were not distributed normally. Pages L trend analyses was used to examine trends in data over a period of time (Cohen and Holliday 1982). In all analyses, the 95% level of confidence was taken to be indicative of statistical significance. Throughout the text, tables and figures, values are reported as means (SEM).

Results

The per formance times for the control group dur ing trials 1 and 2 were 135.0 (4.5) m i n and 135.3 (4.7) min respectively (NS). The C H O group covered the 30-km run in 137.5 (5.5) min dur ing trial 1 and in 134.9 (5.5) min dur ing trial 2, representing an improvement of 2.6 rain (1.9%). In the C H O group, 8 of the 9 subjects ran the 30-km run faster dur ing trial 2 than dur ing trial 1. The r u n n i n g speeds over each successive 5 km of trial 1 and trial 2 are shown for the two groups in Fig. 1. There were no differences for the control group dur ing the two trials. However, the r unn i ng speed of the CHO group over the last 5 km dur ing trial 2 was faster than the speed they achieved over the last 5 km dur ing trial 1

4.2

A 4.0

3.8

3.6

3.4

3.2

3.0

Control Group

' ' , ' o ' 'o '6 5 1 5 2 2 30

Distance ( k m )

g

¢o

==

4.2

4.O

3.8

3.6

3.4

3.2

CHO Group

3.0 . . . . . 5 1 0 1 5 ' 2 0 ' 2f5 ' 30

Distance ( k m )

Fig. 1. Mean (SEM) running speeds during trial 1 (--[]--; T1) and trial 2 (--I~--; T2) for the control group and the carbohydrate group (CHO) during the two 30-km treadmill time trials. ** P < 0.001

[3.64 (0.24) m.s -1 vs 3.44 (0.26) m . s - I ; P<0.001] . g 0.9, Trend analysis revealed significant decreases in running ~ 0.92 speed for the control and CHO groups during both trials

¢: 0.90 1 and 2. A closer examination of the overall perform- ~. ance times of the men and the women in this study $ o.88 showed that the men in the CHO group had faster run- ning times during trial 2 than trial 1 [127.4 (4.9) min vs _ ~ o., 131.0 (5.4) min; P<0 .05] , but there were no such differ- ~ o.8, ences between trials for the men in the control group [129.6 (4.1) min vs 128.4 (3.6) min; NS]. The small num- ber of women in this study prevented a similar analysis of their performance times, o.96

The oxygen uptakes of the control group during both go trials were equivalent to 70.6% P-Oamax [41.5 (2.1) ~ o.94 m l ' k g - l ' m i n -1] during the first 5 km. By the end of ~ 092

the time trial running speed had fallen to the equivalent ~ 0.9o of 62.0070 ~rO2max (P< 0.005). There were no significant ~" 0.= differences between the oxygen uptake values for the CHO group over the first 5 km during the two trials [trial 1:41.3 (0.4) vs trial 2:41.7 (0.1) ml .kg -~ .min-1] . °.84 Heart rates ranged between 169 and 180 beats .min -1 for both groups, during the two trials. No differences were found in the heart rate responses of either group on trial 2 when compared to trial 1. During trial 1, both groups decreased their average stride lengths ( P < 0.01); the CHO group decreased from 1.33 (0.06) m to 1.24 (0.08) m, whilst the control group decreased from 1.39 (0.05) m to 1.15 (0.08) m. Only the control group de- 5.0 creased their stride lengths ( P < 0.001) during trial 2. An =

- - - 5.5 examination of the relationships between performance times, stride length, and ~rO2max values for the whole g 5.0 group (n = 18) during trial 1 showed a stronger correla- o~ 4.5 tion between performance times and average stride

4.0

lengths ( r= -0 .94 ; P<O.O1) than between performance times and DrO2raax values ( r= -0 .75 ; P<0 .01) . ~ ~.9

The respiratory exchange ratios (R) of the control 5.0 and CHO groups decreased (P<0.001) during both trials (Fig. 2). There were no differences between the R values on trial 1 and trial 2 for the control group. How- ever, there was a trend towards higher R values in the CHO group during trial 2 compared with trial 1, and this trend became significant at 5 k m and 20km (P< 0.05).

There were no differences in blood glucose concen- trations of the control group during the two trials. They did, however, decrease towards the end of the 30-kin run on both occasions (P<0.001; Fig. 3). In contrast, the blood glucose concentrations of the CHO group de- creased during trial 1 ( P < 0.001), but not during trial 2. Compared with trial 1, the blood glucose concentrations of the CHO group during trial 2 were higher at 20 km, 25 km and at 30 km ( P < 0.05). Blood lactate concentra- tions for the two groups during the two trials are shown in Fig. 4. Trend analysis revealed a decrease ( P < 0.05) in the values of the control group during trial 1. The CHO group had higher blood lactate concentrations after 30 km during trial 2, compared to the values obtained after the same distance during trial 1 (P<0.05) . Both groups increased their plasma FFA and glycerol concen- trations during trial 1 and trial 2 (P<0.001) (Figs. 5, 6).

Control Group

; ','0 , ; ' = ' 0 ' 2 ' 5 3'0 Distance (kin)

21

CHO Group

' ; , '0 ,'9 ;o '2'5 3'° Distance (km)

Fig. 2. Mean (SEM) respiratory exchange ratios (R) during trial 1 (--El--; T1) and trial 2 (--~--; T2) for the control group and the carbohydrate group (CHO) during the two 30-kin treadmill time trials. * P<0.05

Control Group

f r T I i 1 I 5 10 1 5 20 25 30

Distance (km)

CHO Group 6.0 A

5.5

i 5.0

4.51 o O 4,0

_~ 3.5

3 . 0 5 T ~ 10T 151 , 2[0 , 215 310 Distance (km)

Fig. 3. Mean (SEM) b lood glucose concentrations during tr ia l 1 (--El--; T1) and trial 2 (--~--; T2) for the control group and the carbohydrate group (CHO) during the two 30-kin treadmill time trials. * P<0.05

The plasma catecholamine concentrations of the two groups before and immediately after trial 1 and trial 2 are shown in Table 3. Plasma noradrenaline concentra- tions of the control group increased during both trials by similar amounts (P < 0.001). The plasma noradrenal- ine concentrations of the CHO group also increased in both trials by similar amounts, but the values achieved were not identical to those obtained for the control

22

E g

o

Control Group

0 0 110 115 f 210 ' 215 310

Distance (kin)

5

4

g 3 w

..l

0

CHO Group

I , I ~ I , 2tO , 2f5 ~ 310 5 10 15 Dls iance (km)

Fig. 4. Mean (SEM) blood lactate concentrations during trial 1 (--E]--; T1) and trial 2 ( - - , - - ; T2) for the control group and the carbohydrate group (CHO) during the two 30-km treadmill time trials. * P<0.05

0.8

0.7

O 0.6 E 0.5

O 0.4 eJ O 0.3 o t~ 0 . 2

0.1 [

0 .0

Control Group

Trial 1 Trial 2

[] Pre ° [ ] Po=

0.8

• ~ 0.7

E o.6 g - - 0.5 o o~ 0 . 4

O 0 .3

0 .2

O. 0.1

CHO Group

~ [] Pre [ ] POSt

Trial 1 Trial 2

Fig. 6. Mean (SEM) plasma glycerol concentrations before and after trial 1 and trial 2 for the control group and the carbohydrate group (CHO) during the two 30-km treadmill time trials

1.2 A "m 1.0

E 0.8 g

0.6 u.

0.4

a. 0.2

0.0

Control Group

Trial 1 Trial 2

CHO Group 1.2

10

i 0.8 v < 0.6

i .4

[ 0.2

0.0 Trial 1 Trial2

Fig. 5. Mean (SEM) plasma free fatty acid (FFA) concentrations before and after trial 1 and trial 2 for the control group and the carbohydrate group (CHO) during the two 30-km treadmill time trials

group. The plasma adrenaline concentrat ions o f the control group increased during trials 1 and 2 ( P < 0.01) and there were no differences between trials. However , the increase in plasma adrenaline concentra t ion o f the C H O group during trial 2 was only 63°7o of the change recorded for trial 1 ( P < 0 . 0 5 ) .

Body masses of the runners in the C H O group de- creased by 1.99 (0.09) kg during trial 1 and by 2.07

Table 3. Plasma catecholamine concentrations for the control and the CHO group before and after the two 30-km time trials; values are means (SEM)

Noradrenaline Adrenaline (nmol.1-1) (nmol.1-1)

Group Trial 1 Trial 2 Trial 1 Trial 2

Control

CHO

Pre- 2.02 2.21 0.37 0.37 (0.22) (0.29) (0.08) (0.05)

Post- 9.84 9.22 3.77 3.08 (1.34) (0.92) (0.68) (0.61)

Pre- 2.19 2.06 0.25 0.22 (0.12) (0.13) (0.04) (0.04)

Post- 15.54 15.79 5.43 1.92' (1.79) (0.90) (0.04) (0.19)

* Denotes significantly lower (P< 0.05) than trial 1 value

(0.11) kg during trial 2. The runners in the control group had a decrease in body mass of 1.92 (0.1) kg during trial 1 and 1.94 (0.08) kg during trial 2. The fluid intake for the control and C H O groups during trial 1 was 353 (151) ml and 349 (63) ml, respectively and during trial 2 the values were 298 (70) ml and 370 (84) ml, respectively. There were no significant differences between trials or between groups. The mean plasma volume changes for the C H O group were - 7 . 0 (1.2)°70 during trial 1, and - 5 . 9 (1.1)o70 during trial 2. For the control group, the mean plasma volume changes were +0 .8 (2.0)°7o and - 4 . 7 (2.4)°7o on trial 1 and trial 2 respectively. These changes in plasma volumes were not significantly differ- ent.

23

Discussion

The main finding of this study was that the subjects ran faster during the last 5 km of the 30-km time trial after 7 days on a high carbohydrate diet than they did after consuming their normal mixed diets. The control group showed no such improvement in performance, even though their energy intake was the same as the CHO group. Although there was no significant improvement in the performance time for the CHO group as a whole, it is worth noting that 8 of the 9 subjects completed the 30-kin run in a shorter time during the CHO trial. Fur- thermore, an analysis of the results of the male runners showed that the 6 men in the CHO group had signifi- cantly better overall performance times during trial 2. They improved their times by 3.6 min (2.8%) after car- bohydrate loading, whereas the men in the control group did not improve their performance times during trial 2.

These improvements were not as great, however, as those reported by Karlsson and Saltin (1971) in their study on the influence of carbohydrate loading on run- ning performance during cross-country races over 30 km. Their subjects improved their performance times by 5.4% after carbohydrate loading. This improvement is twice as great as that achieved by the men in the pres- ent study. These differences may be explained in two ways. The first is that Karlsson and Saltin (1971) had two groups of runners in their study who were clearly of different ability. One group was made up of experienced runners with high ~rO2max values and the other consisted of active physical education students who had only modest lkO2max values. The reductions in running times for the 30-km run were 5 min (3.2%) and 12 min (7.6%) for the experienced and less experienced runners respec- tively. The second and more important explanation for the differences in performances reported by the two studies is that the subjects in the earlier study ran over a cross-country course, whereas the subjects in the present study completed the 30-km run on a level treadmill. The demanding nature of the undulating cross-country course is reflected by the performance times of the run- ners. The ~rO2max values of their subjects were 67 .7ml 'kg - l 'min -1, but their average performance time was 143.0 rain. In contrast the CHO group, in the present study, with an average lkO2max value of only 58 .7ml 'kg - l 'min -~, covered the 30-kin distance in 137.6min. Uphill running uses more muscle glycogen than running at the same speeds on the level (Costill et al. 1974). Therefore, there was probably a greater reduc- tion in the muscle glycogen stores of the runners who completed the 30-km cross country races than occurred in the runners who completed the 30-km treadmill time trials. Carbohydrate loading would, therefore, have been more important in preparation for the 30-km races over the cross-country course than in the present study.

The overall rate of carbohydrate oxidation by the CHO group during trial 1 was approximately 2.2 g.min -~ and 2.5 g.min -1 during trial 2. The total amounts of carbohydrate oxidized during trial 1 and during trial 2 were 297 g and 330 g, respectively. The

carbohydrate intake of the CHO group during the 7-day recovery period was approximately 3476g, which is 1138 g more than they would have consumed had they not undertaken carbohydrate loading (Table 2). After replacing the 297 g carbohydrate used during trial 1, the excess carbohydrate intake amounted to about 840 g. This approaches the predicted saturation level for carbo- hydrate storage beyond which additional carbohydrate intake is converted into fat (Acheson et al. 1988). The additional 33 g carbohydrate used in trial 2 represents, therefore, only a small fraction of the carbohydrate con- sumed over the 7-day recovery period. A similar result was obtained in a recent study using the same exercise procedures to assess the influence of consuming carbo- hydrate/electrolyte solutions on endurance performance during 30-km treadmill time trials (Williams et al. 1990). We found that when our highly trained runners drank water during these simulated endurance races they slowed down during the last 10 km. However, when they drank the carbohydrate/electrolyte solutions, through- out the time trials, they were able to sustain their opti- mum running speeds for the whole 30 km. They con- sumed 1 1 of the 5% carbohydrate/electrolyte solutions and oxidized approximately 30g carbohydrate more than during the water trial. Carbohydrate loading does improve endurance performance but the changes in run- ning times are relatively modest compared with improv- ements in endurance capacity. For example, the 5.4% improvement in running times for a 30-kin cross country race (Karlsson and Saltin 1971) contrasts quite markedly with an improvement in running time to exhaustion of 26°70 (Brewer et al. 1988). There is not a strong relation- ship between initial muscle glycogen concentration and endurance performance as there appears to be with en- durance capacity (Bergstrom et al. 1967). There are fac- tors other than the carbohydrate stores of the competi- tor which influence endurance performance. To sustain high running speeds, an athlete requires a high J~rO2max and to be sufficiently well trained to utilize a large % 1202max for a long time. An adequate supply of mus- cle glycogen is, nevertheless, an essential prerequisite for endurance competitions.

An additional important finding in this study was the lower plasma adrenaline concentrations of the runners in the CHO group during trial 2. They ran the last 5 km of the 30-km run faster after carbohydrate loading. Therefore, it might have been expected that they would have had higher, rather than lower, adrenaline concen- trations because of the greater physiological stress of running at faster speeds. The explanation for this change is probably associated with the consequences of carbohydrate loading, in general, and blood glucose concentrations in particular. Previous studies have shown that a reduction in blood glucose concentrations, to hypoglycaemic levels, increases the concentration of plasma adrenaline. Conversely, restoring low blood glu- cose concentrations to normal resting levels, by glucose infusion, reduces plasma adrenaline concentration (Gal- bo et al. 1979). During trial 1, the blood glucose concen- trations of both the control and the CHO groups de- creased below pre-exercise values towards the end of the

24

30 km. This was also the response of the control group during trial 2. Thus their blood glucose concentrations and their plasma adrenaline concentrations were similar in both trials. In contrast, the CHO group maintained their normal blood glucose concentrations for the whole of trial 2 and had lower adrenaline concentrations than in trial 1. These results are consistent with the proposed relationship between plasma adrenaline and blood glu- cose concentrations (Galbo et al. 1979). There were no differences in the plasma noradrenaline concentrations between trials for both groups. However, the mean no- radrenaline values for the CHO group after both trials appeared to be higher, though not significantly so, than those of the control group. The only explanation we can offer is that during trial 1 the CHO group ran for a slightly longer time than the control group. During trial 2 they ran the last 5 km faster than in trial 1 and so the contribution of the neuro-transmitter, noradrenaline, to cardiovascular control was probably somewhat greater.

Even though carbohydrate loading improved the en- durance performance of the runners, trend analysis showed a significant decrease in their speeds as they ap- proached 30 kin. This raises the question as to what oth- er factors contributed to the onset of fatigue in this group. Savard and colleagues (1987) suggested that myocardial fatigue may contribute to fatigue, during prolonged exercise, as a result of the competing de- mands on the cardiovascular system for oxygen trans- port and thermoregulation. The contribution of inade- quate thermoregulation to the onset of fatigue cannot be assessed in the present study because we did not include measurements of body temperature. Nevertheless, it is worth noting that the subjects drank very little, even though water was freely available throughout the time trials. Their average water intake was between 300 and 370 ml and some runners did not drink at all during either trial. The runners lost approximately 3°7o of their body mass during each of the trials which was almost entirely the result of sweating. The low fluid intake may have contributed to dehydration and possibly the accu- mulation of an intolerable heat load towards the end of the time trials (Sawka et al. 1985). Although dehydra- tion may not have a detrimental influence on muscle me- tabolism (Nielsen et al. 1990), it will, nevertheless, lead to either a decrease in skin blood flow or be a threat to blood pressure. The consequence of either of these events occurring during prolonged exercise would con- tribute to fatigue, irrespective of the status of the body's glycogen stores.

In summary, the results of this study confirm the benefits of carbohydrate loading as a means of improv- ing endurance-running performance. Furthermore, they show that supplementing a normal diet with confection- ery products is an effective way of increasing carbohy- drate intake in preparation for endurance competi- tions.

Acknowledgements. The authors wish to acknowledge their grati- tude to their colleague Steve Brooks for the catecholamine ana- lyses and to Dr. Maureen Edmondson, of Mars Confectionery UK, for her support throughout this study.

References

Acheson K J, Schutz Y, Bessard T, Anantharaman K, Flatt J-P, Jequier E (1988) Glycogen storage capacity and de novo lipo- genesis during massive carbohydrate overfeeding in man. Am J Clin Nutr 48 : 240-247

Astrand PO (1967) Diet and athletic performance. Fed Proc 26: 1772-1777

Bergstrom JB, Hermansen L, Hultman E, Saltin B (1967) Diet, muscle glycogen and physical performance. Acta Physiol Scand 71 : 140-150

Brewer J, Williams C, Patton A (1988) The influence of high car- bohydrate diets on endurance running performance. Eur J Appl Physiol 57 : 698-706

Brooks S, Burrin J, Cheetham ME, Hall GM, Yeo T, Williams C (1988) The responses of the catecholamines and B-endorphin to maximal exercise in man. Eur J Appl Physiol 57:230-234

Cohen L, Holliday M (1982) Statistics for Social Sciences. Harper and Row, London

Costill DL, Jansson E, Gollnick PD, Saltin B (1974) Glycogen uti- lization in leg muscles of men during level and uphill running. Acta Physiol Scand 91:475-481

Costill DL, Sherman WM, Fink W J, Maresh C, Witten M, Miller JM (1981) The role of dietary carbohydrates in muscle glycog- en resynthesis after strenuous running. Am J Clin Nutr 34:1831-1836

Coyle EF (1991) Timing and method of increased carbohydrate intake to cope with heavy training, competition and recovery. J Sports Sci 9 [Suppl] : 29-52

Dill DB, Costill DL (1974) Calculation of percentage change in volumes of blood, plasma and red cells in dehydration. J Appl Physiol 37 : 247-248

Galbo H, Hoist J J, Christensen NJ (1979) The effects of different diets and of insulin on the hormonal response to prolonged exercise. Acta Physiol Scand 107:19-32

Hultman E (1967) Studies on muscle metabolism of glycogen and active phosphate in man with special reference to diet. Scand J Clin Lab Invest [Suppl] 94:19

Karlsson J, Saltin B (1971) Diet, muscle glycogen and endurance. J Appl Physiol 31 : 203-206

Maughan RJ (1982) A simple rapid method for the determination of glucose, lactate, pyruvate, alanine, 3-hydroxybutyrate and acetoacetate on a single 20 ~tl blood sample. Clin Chim Acta 122: 231-240

Nielsen B, Savard G, Richter EA, Hargreaves M, Saltin B (1990) Muscle blood flow and muscle metabolism during exercise and heat stress. J Appl Physiol 69:1040-1046

Paul AA, Southgate DAT (1978) The composition of foods. HMSO, London

Roberts KM, Noble EG, Hayden DB, Taylor AW (1987) Simple and complex carbohydrate-rich diets and muscle glycogen con- tent of marathon runners. Eur J Appl Physiol 57 : 70-74

Savard G, Kiens B, Saltin B (1987) Central cardiovascular factors as limits to endurance; with a note on the distinction between maximal oxygen uptake and endurance fitness. In: Macleod D, Maughan R, Nimmo M, Reilly T, Williams C (eds) Exercise, benefits, limits and adaptations. Spon, London, pp 162-177

Sawka MN, Young A J, Francesconi RP, Muza SR, Pandolf KB (1985) Thermoregulatory and blood responses during exercise at graded hypohydration levels. J Appl Physiol 59:1394-1401

Sherman WM, Costill DL, Fink WJ, Miller J (1981) Effect of exercise-diet manipulation on muscle glycogen and its subse- quent utilization during performance. Int J Sports Med 2:114- 118

Williams C, Nute MG, Broadbank L, Vinall S (1990) Influence of fluid intake on endurance running performance: a comparison between water, glucose and fructose solutions. Eur J Appl Physiol 60 : 112-119