MATERIALS AND METHODS

Subjects

Thirty four male soccer players volunteered to participate in the study and provided written informed consent in accordance with the Declaration of Helsinki. The university ethics committee approved the study protocol. The subjects could withdraw from the study at any time. Their physical characteristics were as follows (mean (SD)): age 17.5 (1.1) years; height 177.8 (6.7) cm; weight 70.5 (6.4) kg; body mass index 22.5 (1.4) kg/m²; percentage body fat 11.8 (2.0). Percentage body fat was calculated according to the formula of Siri based on four skinfold measurements (biceps, triceps, subscapularis, and suprailiac).¹⁷ Twenty two of the subjects were members of the Tunisian national under 19 team, and the remaining 12 belonged to one of the best three ranked teams in Tunisia during the last 10 years. The latter composed the elite players of the club and lived in a special “centre of excellence” included in the club’s infrastructure. Seven of these 12 players were recruited from Senegal three months before the experiment. The subjects were informed about the test protocols, without being informed about the aim of the study. They participated in the national soccer championship as regular players, and some of them were selected for the senior team of their respective clubs. At the time of the experiment, their average weekly training programmes included 8–10 training sessions a week (each session lasting about 90 minutes), mainly soccer training and very rarely on track running or muscular strength training.

The experiment was performed mid-season—that is, three to five months after the beginning of the competitive season. The cohort studied was comprised four goalkeepers, 12 defenders, 12 midfield players, and six forwards.

Test protocols

Owing to injuries and the busy programme of some of the players, not all the subjects were able to perform all the tests. Each player was tested separately, instructed, and verbally encouraged to give maximal effort on all the tests. The number of players who took part in the different tests is presented in the tables and figures.

Laboratory testing

Laboratory testing was performed between 2 and 5 pm (mean (SD) ambient temperature 21(1)°C) on two days, separated by one week. The subjects were wearing shorts and running shoes. They had been asked to abstain from exercise the day before the tests and not to drink caffeine-containing beverages on the day of the tests.

Day 1

On entering the laboratory, subjects lay on a bench and rested for 15 minutes. The lowest heart rate during that period was recorded. They then ran on a flat treadmill (Ergo XELG 90; Woodway, Weil, Germany) for three minutes at 9 km/h. The speed was then increased by 1 km/h every minute until exhaustion, which occurred within 10–15 minutes for all subjects. The following criteria were met by all players when Vo₂max was tested: (a) a levelling off of Vo₂ despite an increase in treadmill speed; (b) a respiratory gas exchange ratio higher than 1.1; (c) blood lactate higher than 6 mmol/l. The highest heart rate attained at exhaustion was recorded. Cardiorespiratory variables were determined using a breath by breath system (ZAN 680; Messgeräte, Oberthulba, Germany) allowing continuous measurement of heart rate, oxygen uptake, and lung ventilation. Before each test, the gas analysers were calibrated with gasses of known concentration, and the ventilatory membrane was calibrated with a 1 litre syringe. Heart rate was determined from a six lead electrocardiograph with 12 derivations. Heart rate and respiratory data were recorded once every 30 seconds, with the values averaged over the last 10 respiratory cycles on a sliding technique basis.¹⁸ The lowest and highest velocity associated with Vo₂max (vVo₂max and vpeakVo₂max respectively) were established as described by Billat and Koralsztein¹⁹ and Paavolaïnen et al,²⁰ and respiratory compensation threshold (Th2_vent) as described by Beaver et al.²¹ If Vo₂max is expressed as ml/kg/min, the work capacity of light subjects is overestimated and that of heavy subjects is underestimated.¹⁵ Consequently, Bergh et al,²² Wisløff et al,¹⁵ Helgerud et al,³ and Hoff et al² have expressed oxygen uptake as ml/kg^0.75/min for comparisons among athletes and soccer players of different body mass. Indeed, Vo₂max does not increase in direct proportion to body mass.²² Dimensional scaling of geometrically similar subjects suggests that the cross section area of the aorta will increase in proportion to the square of the height (L²) whereas body mass is dependent on body volume, which varies according to L³. Consequently, Vo₂max, which is primarily limited by maximal cardiac output, should be proportional to body mass (m_b) raised to the power of 0.67 (m_b^0.67). This dimensional scaling approach was supported by Bergh et al,²² who found that Vo₂max relative to body mass raised to the power of 0.75 was indicative of performance capacity in running. Despite the fact that the theoretically correct m_b^0.67 should be used, most published papers on soccer players have used the ml/kg^0.75/min expression. Wisløff et al¹⁵ suggested that expressing Vo₂max in relation to m_b^0.67 or m_b^0.75 may not be critical as long as the unit approximates the theoretical value and not the traditional m_b. Dimensional scaling should ideally be based on fat-free mass, because fat has very low metabolic activity.²³ In the present study the statistical outcome of using fat-free body mass or body mass was similar (data not shown). We chose to use dimensional scaling based on body mass so that we could directly compare our results with previous studies.^3,13,15,24

Day 2

The vertical jump was performed from a fixed semisquat position with the hands fixed at the hips using a force platform (9281 C; Bioware, Kistler, Switzerland). Each player performed three jumps with two minutes of rest in between, and the best jump was selected for analysis. Peak jumping force (F_peak), peak jumping velocity (v_peak), peak jumping anaerobic power (W_an), and the peak height of the jump (H_peak) were recorded. After this procedure, the subject also performed a free counter movement jump protocol. The statistical results were similar using data from these two tests, and for clarity we chose to present only the former in this paper.

As for the Vo₂max expression, dimensional scaling must also be considered when evaluating strength measures.¹⁵ In two geometrically similar and quantitatively identical individuals, one may expect all linear dimensions (L) to be proportional. The length of the arms, the legs, and the individual muscles will have a ratio of L:1, the cross sectional area L²:1, and the volume ratio L³:1. As muscular strength is directly proportional to the muscle cross sectional area, and body mass (m_b) varies directly with body volume, whole body muscular strength measures will vary in proportion to m_b^0.67. The peak force is therefore expressed as N/kg^0.67.

Field testing

All field tests were performed on one day, one week after the end of the laboratory testing. The sprint tests were performed in the morning and the Bangsbo test in the afternoon. The tests were performed on a natural grass soccer pitch, and the subjects were wearing soccer kit. During the non-raining field test days, wind speed did not exceed 8 knots (4.11 m/s) and air temperature ranged from 23°C to 26°C.

30 m sprint

The subjects performed 20 minutes of individual warm up including several accelerations knowing that they had to choose which foot they had to put on the starting line for the sprint standing position start. They then performed three 30 m sprints including 10 m lap timing with three minutes of recovery in between. Speeds were measured with infrared photoelectric cells (Matsport timing BTS, Seyssinet, France) positioned at exactly 10 and 30 m from the starting line at a height of 1 m. The 20 m (from the 10 m to the 30 m line) performance was then calculated. The stopwatch starting pedal was positioned behind the starting line. The subject had to start from a standing position placing his forward foot just behind the starting line and his rear foot on the pedal after having positioned the pedal according to his natural starting position. The timing started as soon as the foot of the player left the pedal. Before testing, each subject performed a submaximal sprint to familiarise himself with the test procedure. The sprint resulting in the best 30 m performance was selected for analysis.

The Bangsbo test

This soccer specific endurance test was proposed by Bangsbo and Lindquist⁷ and further described by Bangsbo.^8,10 Figure 1 illustrates the test track. Briefly, the test lasts 16.5 minutes, during which players alternate between 40 bouts of high intensity exercise each lasting 15 seconds and 40 bouts of low intensity exercise each lasting 10 seconds. The work-rest periods were dictated by one long sound signal for the beginning and two short sound signals (whistle) at the end of each sprint. During the high intensity periods, the subjects follow an outlined circuit around the penalty area of a soccer field. They run 40 m forward, 8.25 m backwards, 95.25 m forward and through a 120° angle slalom, 8.25 m sideways while facing away from the centre of the circuit, and 8.25 m sideways while facing the centre of the circuit. During the low intensity periods, players jog to the centre of the circuit and back to the last cone marked position they reached at the end of the previous high intensity period. The test performance is the distance covered during the 40 periods of high intensity running. Before each test, after 10 minutes of standardised warm up, the players had 10 minutes to familiarise themselves with the circuit by light jogging. They then had to try the test pace—that is, the 15 second high intensity and 10 second light intensity bouts of exercise—for four minutes. This allowed them to find a personal running pace to try to perform the best performance over the 16.5 minute test duration. After the warm up and the familiarisation, they rested for 10 minutes before performing the test. During the Bangsbo test, heart rate was monitored with a heart rate monitor (Polar S-610; Polar Electro, Kempele, Finland). Five second intervals were selected (0.2 Hz).

Blood sampling and determination of blood lactate concentration

Blood samples were collected 3.5 minutes after Vo₂max and after the Bangsbo tests. The 20 μl samples of capillary blood were withdrawn from an earlobe with Microzym micropipettes. They were stored in tubes containing 180 μl of a haemolytic solution to ensure preservation of the samples at room temperature. Blood lactate concentration was subsequently measured by an enzymatic method (Microzym L; Setric Génie Industriel, Toulouse, France).

Statistical analysis

Values are expressed as mean (SD). For comparison between groups, we used a two way analysis of variance. The Scheffé statistic was used to calculate post hoc p values of possible difference between groups. A Pearson correlation matrix was performed between the variables of the field and laboratory tests. A stepwise linear regression analysis was used when appropriate. Statistical significance was fixed at p<0.05.

RESULTS

No positional differences for any of the physiological variables were observed and the mean data for all players are therefore presented.

Laboratory testing

During treadmill testing, the oxygen uptake levelled off despite increased running speed in all players—that is, the true Vo₂max was determined (table 1). Table 2 presents vertical jump height variables.

Field testing

The mean (SD) time for the 30 m sprint test was 1.87 (0.10) seconds with a 10 m lap time of 4.38 (0.18) seconds (n  =  30). Table 3 presents data from the Bangsbo field test.

Correlations between laboratory and field testing

There was no significant correlation between the Bangsbo test and Vo₂max. However, there was a positive correlation between vVo₂max, vpeakVo₂max, and distance covered in the Bangsbo test (fig 2).

There was no significant correlation between vertical jump height and sprint performance (table 4). However, the peak jumping velocity and force correlated positively with 20 m and 30 m sprinting performances (table 4), and 46% of the 30 m sprinting performance was explained by the jumping force (N/kg) and jumping velocity.

DISCUSSION

This study shows that the performance in the Bangsbo test did not correlate significantly with Vo₂max in young soccer players. This was surprising as an average running intensity of 95% of maximal heart rate normally corresponds to Vo₂max when running is continuous.²³ The nature of the Bangsbo test is intermittent running, and the heart rate achieved is probably not a good picture of the actual exercise intensity, as it normally takes one minute of continuous running to reach an exercise intensity corresponding to 95% of maximal heart rate.³ Therefore the exercise during the 15 seconds of high intensity running probably corresponds to an anaerobic exercise intensity beyond Vo₂max. This is confirmed by the high concentration of blood lactate measured (table 3). However, the soccer players in this study represent a homogeneous group with regard to Vo₂max, and therefore a correlation between the performance in the Bangsbo test and Vo₂max may be found in a more heterogeneous group of players. Heart rate values during the present Bangsbo test are similar to previous reports,¹ confirming high heart rate responses to the test. Furthermore, the venous blood lactate concentrations are in accordance with those of Mujika et al,¹² but higher than reported by Bangsbo and Lindquist⁷ and Bangsbo.¹ The senior Tunisian squad (n  =  25) that participated in the 2002 African Cup and World Championships had a mean blood lactate concentration of 14.84 (1.31) mmol/l (range 12.6–17) and covered a distance of 1879.1 (123.7) m in the Bangsbo test with a heart rate of 186.5 (9.9) beats/min.

The fact that the Bangsbo test correlates with the lowest and highest running speed associated with Vo₂max is surprising because normally there is a close relation between Vo₂max and running speed on an inclined treadmill.²³ Testing of Vo₂max normally requires an inclined treadmill,²³ and therefore perhaps the true Vo₂max was not reached in our study using a flat treadmill. However, all the criteria for reaching Vo₂max were achieved by all subjects, and we therefore conclude that the true Vo₂max was reached. Thus the relation between running speed and Vo₂max may be different when a flat treadmill is used rather than an inclined treadmill. Running technique probably plays a more prominent role when the former is used. As it has been shown in several studies^3,25 that Vo₂max is a very important variable of match performance of soccer players, tests other than the Bangsbo test are required to either estimate or directly measure it. Helgerud et al³ showed that improving Vo₂max by 11% induces a 20% (∼1800 m) increase in the distance covered in a match, 23% increase in involvement with the ball, and 100% increase in the number of sprints performed. The importance of increasing the performance in the Bangsbo test on a player’s match performance is still uncertain. A complementary study to determine the effect of a training induced increase in the Bangsbo test on match performance (distance covered, number of sprints, involvement with the ball), aerobic function (Vo₂max), and aerobic performance (vVo₂max, time to exhaustion at vVo₂max, and running economy) would extend our knowledge of this often used test. Although the mean data show that 55% of the performance in the Bangsbo test can be explained by vVo₂max, the precision when individual values are estimated is not very accurate—that is, the standard error of the y estimate is quite large (fig 2A). For example, if a player has a vVo₂max of 18 km/h, fig 2A shows that the estimated distance covered in the Bangsbo test is 1700–1950 m; the corresponding data for 19 km/h and 17 km/h are 1750–1950 m and 1700–1800 m respectively. So on an individual basis this is not very impressive. Therefore, before this test is used in a standard test battery, more studies to determine its practical use should be performed. The optimum soccer specific test has yet to be defined, and, for the time being, the most useful tests for evaluating strength and endurance training are laboratory tests. However, the Bangsbo test is useful for looking at other aspects of aerobic endurance performance besides the traditional Vo₂max, running economy, and anaerobic threshold—for example, vVo₂max and vpeakVo₂max. On the other hand, the usefulness of increasing vVo₂max and vpeakVo₂max without increasing Vo₂max has yet to be defined.

The values obtained for Vo₂max and anaerobic threshold in this study are of the same order as found for elite junior players in Norway.³ With the present knowledge of the importance of a high Vo₂max on match performance, further emphasis should be placed on increasing that. There is no reason for elite junior players to have a lower Vo₂max than elite senior players, approaching 70 ml/kg/min (corresponding to 205 ml/kg^0.75/min for a 75 kg player).¹⁵ Effective exercise training regimens, with or without the ball, have been presented in detail elsewhere,^2,3 and include periods of four minutes at an exercise intensity corresponding to 90–95% of maximal heart rate. Such a training regimen could be expected to increase Vo₂max by 0.5% per training session.³

As a soccer player jumps on average 15.5 times, with about nine headers a match,¹ and performs a sprint about every 90 seconds, each lasting two to four seconds,^1,26 jumping and sprinting performance obviously influence the outcome of a game. The 30 m sprint time did not significantly correlate with vertical jump height, but rather with peak velocity during the jump. This is not surprising as it has been shown that it is the level of maximal strength and the rate of force development that influence both jump height and sprint performance.^24,27 This is in accordance with previous studies in Norwegian senior elite soccer players¹³ and emphasises the fact that muscular force and power are relevant variables with respect to sprinting performance. Surprisingly, the 10 m lap time did not correlate with any of the vertical jump variables, which contrasts with previous reports.¹³ It must be emphasised that subjects in the latter study were adult elite players and that half of the population studied performed regular high intensity strength training, which was not the case for the players in this study. As suggested by Wisløff et al,¹³ when subjects with different training regimens are studied, clearly the correlation between sprinting and jumping performance may be different. The role of age could not be ruled out either, with possible different physiological capacities¹⁷ and/or lower level of acceleration skills. The jumping performance in the present study was in the normal range for soccer players.^3,15 The sprint results fit in with observations in elite junior players in a previous study.³ The 10 m lap time could give important information, as indicated by the substantial differences within the 30 m test, with some of the subjects having similar 30 m times but notably different 10 m times. In this context, it must be emphasised that the 10 m performance is a relevant test variable in modern soccer. Cometti et al¹⁶ have shown that French professional and amateur soccer players had similar 30 m sprint performances but the professionals had significantly lower 10 m lap times.

Conclusions

In young soccer players, the Bangsbo test and 30 m sprint test correlate with vVo₂max and peak jumping velocity respectively. The Bangsbo test does not give a good estimate of Vo₂max, and further studies are necessary to determine its practical use in a test battery for soccer players.