Journal Logo


Match-Related Fatigue in Soccer Players


Author Information
Medicine & Science in Sports & Exercise: November 2011 - Volume 43 - Issue 11 - p 2161-2170
doi: 10.1249/MSS.0b013e31821e9c5c
  • Free


Soccer is a team sport that requires prolonged high-intensity intermittent exercise (28). During the match, players change activity on average every 5 s and perform approximately 1300 actions, with 200 of these being completed at high intensity (28). Besides running, players perform other game-related demanding activities that require a high level of force production, such as direction changes, dribbling, tackling, and heading.

Fatigue can be defined as an acute impairment of performance that includes both an increase in the perceived effort to exert a desired force or power and/or any reduction in the ability to exert maximal muscle force or power (15). Recently, it has been suggested that players experience fatigue toward the end of the match as well as temporarily during the game (29). In fact, using computerized video analysis systems, it was reported that total distance (TD) covered and high-intensity activities are reduced after the most demanding 5-min period of the game (28) and are also reduced in the second half compared with the first (in particular in the last 15-min period of the game) (28,32). In addition, maximal strength, jumping ability, and sprint performance (4,6,18,22,36) are reduced immediately after compared with before a match. The time required for a full recovery of these qualities in male players may be very long (>72 h) (6,18); however, information regarding recovery time in high-level professional athletes is scarce. Match-related fatigue may also have a negative effect on passing precision (33), but again, information pertaining to top-level players is limited.

It has been demonstrated that neuromuscular mechanisms that contribute to performance impairment differ according to the contraction mode and the exercise intensity and duration (26). Fatigue can be classified as central when the origin is proximal and/or peripheral when the origin is distal to the neuromuscular junction (15). However, to date, no studies have quantified the relative contribution of central and peripheral factors in determining fatigue in soccer. Peripheral skeletal muscle function can be determined using electrical stimulations, and long-lasting fatigue may be evaluated using tetanic nerve stimulations at different frequencies (low- to high-frequency force ratio) (12). Nonetheless, tetanic nerve stimulations are very painful, and their usability is limited (37). Paired stimulations have been proposed as a surrogate of the stimulation trains (3), and recent research demonstrates the validity of this method to quantify muscle fatigue induced by eccentric exercise (37).

The physiological mechanisms responsible for the match-related fatigue, however, have not been completely clarified. Reduced muscle force, increased muscle soreness, and increased blood creatine kinase (CK) levels seem to suggest that muscle damage may be one factor that contributes to performance impairment after the match (6,18). However, it is not clear whether the strength loss is solely due to muscle damage. In fact, muscle fatigue may also originate at the spinal and/or supraspinal level (15). In recent research performed in an intermittent sport such as tennis (16), it was demonstrated that a combination of central and peripheral factors is responsible for the strength loss after a match. In addition, muscle fatigue may negatively influence not only maximal performance but also technical ability (33); fatigue induced by eccentric exercise has been reported to alter joint position sense and force modulation capacity (15). However, information about the effect of muscle fatigue on technical ability in professional soccer players is limited.

Therefore, the aims of the present study were 1) to determine the time course of different neuromuscular parameters (maximal voluntary contraction (MVC), maximal voluntary activation, quadriceps muscle contractile properties, sprint and short-passing ability), muscle soreness, and CK in professional players after a game and 2) to examine the central and peripheral contribution to match-related fatigue. We hypothesized that the time course of the parameters investigated in professional athletes is different from that previously reported for lower level players and, further, that match-related fatigue is determined by a combination of central and peripheral factors.



Twenty-two young male professional soccer players from an Italian Serie A team (two goalkeepers, eight defenders, eight midfielders, and four forwards; age = 19 ± 1 yr, body mass = 73.0 ± 7.0 kg, height = 181 ± 5 cm) were involved in the study. Players usually trained six times per week plus an official match. Two defenders and one midfielder were not able to complete all the evaluations during the study and were therefore excluded from the data analysis. The two goalkeepers were also excluded because their physical load during the game is different from the other field players. Written informed consent was obtained from all subjects before the beginning of the experiment. The study was approved by the Independent Institutional Review Board of MAPEI Sport Research Centre according to the Guidelines and Recommendations for European Ethics Committees by the European Forum for Good Clinical Practice and by the soccer clubs involved.

Experimental design.

A quasiexperimental control period design was used for the study, which was conducted during the competitive season. Before the commencement of the study, players were familiarized with the tests and procedures involved. During the control period, participants completed neuromuscular evaluation, the Loughborough Soccer Passing Test (LSPT), and a sprint test on two occasions separated by 1 wk (PRE1 and PRE2). The two test sections of the control period were used to determine test-retest reliability of all the variables with the intraclass correlation coefficient (ICC) (1,2) and SEM expressed as a CV. After the control period, the players performed a 90-min unofficial match arranged for the purpose of the study (two halves of 45-min duration with a 15-min interval between each half) on an artificial grass pitch. Forty minutes (POST), 24 h (POST24), and 48 h (POST48) after the game, the participants repeated the pregame evaluations in the same order as previously conducted. Players did not perform any physical activity in the 48 h before PRE1 and PRE2 and between PRE2 and POST48 with the exception of the match. The match and all testing procedures were performed in the afternoon with an ambient temperature of 16°C-19°C and air humidity of 40%-50%. To limit dietary influences on test results and physiological measures, players were asked to follow nutritional guidelines supplied by the researchers. A generic weekly nutritional plan was supplied to the subjects to ensure an adequate CHO intake (50%-60% of total energy intake). During the match, players were allowed to drink ad libitum.


The protocol used in each session is described in Figure 1. After a standardized warm-up consisting of 5 min of running at a self-selected pace (8-10 km·h−1), neuromuscular tests including electrical induced contraction (single and paired stimuli at 10 Hz and 100 Hz), knee extensor (KE) maximal voluntary isometric strength and activation, LSPT, and shuttle sprint test were performed.

Schematic representation of the protocol used during each test session. MVC, maximal voluntary contraction.

Measurement of muscle contractile function.

All muscle contractile measurements were conducted on the right KE with subjects seated in a leg extension machine. The knee and hip angles were fixed at 90° (0° = knee fully extended), and Velcro straps were used to minimize movement. Subjects were also instructed to grip the seat during the contraction to further stabilize the pelvis. The mechanical response was recorded using a load cell (AIP, Varese, Italy) connected to a data acquisition system (BIOPAC MP100; BIOPAC Systems, Inc., Santa Barbara, CA) at a sampling rate of 125 Hz. The load cell was routinely calibrated using International Organization for Standardization-certified weights. The measurements of KE muscle contractile function included MVC (ICC = 0.94, SEM as a CV of 6.1%), maximal voluntary activation (%VA) (ICC = 0.72, SEM as a CV of 4.2%), and electrical induced torque measurements. The participants performed three MVC, and the best result was used for the analysis (1-min rest between each trial). Quadriceps %VA was estimated by using the interpolated twitch technique. Two electrically evoked stimuli (10 ms apart) were superimposed to the isometric plateau. The ratio of the superimposed doublet over the size of the doublet in the relaxed muscle (control doublet) was then calculated to obtain %VA as follows:

The control doublet was a potentiated doublet because it was delivered 4 s after the end of the KE MVC (for the details of the measurement, see below).

The single (1 Hz) and paired electrical stimuli (10 Hz, 100-ms interval and 100 Hz, 10-ms interval) were produced using square wave pulses (200 μs) via a high-voltage (maximal voltage = 400 V) constant-current stimulator (Digitimer DS7AH; Hertfordshire, United Kingdom). The femoral nerve was stimulated by using a monopolar cathode ball electrode (0.5 cm in diameter) manually pressed into the femoral triangle by the experimenter. The anode was a 10 × 5-cm rectangular self-adhesive electrode (Compex, Ecublens, Switzerland) located in the gluteal fold opposite to the cathode. To limit the influence of the repositioning during the first test session (PRE1), the location of the electrodes was marked on the skin with a permanent marker. The amperage was progressively increased by 10 mA until there was no further increase both in peak twitch torque (the highest value of the KE twitch torque) and concomitant peak-to-peak M-wave amplitude (PPA). This intensity was further increased by 20% and subsequently maintained during the entire session (mean = 180 ± 42 mA, range = 120-252 mA). Before each test session, the adequate intensity was redetermined.

The following parameters were calculated from the mean torque response obtained for each type of stimulation (i.e., the average of the three single or paired stimulations): 1) the highest value of torque production (PT) (ICC from 0.97 to 0.98, SEM as a CV from 2.3% to 2.8%), 2) the time from the origin of the mechanical response to PT (CT) (ICC from 0.87 to 0.95, SEM as a CV from 3.4% to 4.1%), 3) maximal rate of torque development (MRTD) (i.e., maximal value of the first derivative of the torque signal) (ICC from 0.96 to 0.97, SEM as a CV from 3.3% to 3.6%), and 4) maximal rate of torque relaxation (MRTR) (i.e., the lowest value of the first derivative of the torque signal) (ICC from 0.93 to 0.95, SEM as a CV from 3.4% to 4.9%). The ratios of paired stimulation peak force at 10 over 100 Hz (10/100 Hz) were also calculated (ICC = 0.97, SEM as a CV of 2.0%).


EMG signals of the right vastus lateralis (VL) were recorded by using bipolar silver chloride surface electrodes during MVC and electrical stimulations. The electrodes were positioned over the middle of the muscle belly with an interelectrode distance of 20 mm. The position of the electrodes was marked on the skin during PRE1 to fix them in the same place during the other sessions. The reference electrode was attached to the patella. Low impedance between the two electrodes (<5 kΏ) was obtained by abrading the skin with emery paper and cleaning with alcohol. EMG signals were amplified (gain = 1200), filtered (bandwidth frequency = 10-500 Hz), and recorded (sampling frequency = 2000 Hz) using commercially available hardware (PocketEMG; BTS S.p.A., Garbagnate Milanese, Italy) and its dedicated software (Analyzer 1.10.427; BTS S.p.A.). PPA (ICC = 0.79, SEM as a CV of 12.6%) and peak-to-peak duration (ICC = 0.76, SEM as a CV of 10.9%) of the M-wave were determined during single twitches. In addition, the root mean square (RMS) value was calculated during the MVC trials during a 0.5-s period after the torque had reached a plateau and before the superimposed stimulation was evoked (ICC = 0.71, SEM as a CV of 13.5%). RMS values were also normalized to the M-wave amplitudes (e.g., RMS/PPA ratio) (ICC = 0.92, SEM as a CV of 10.1%). A reduction in the MVC RMS without reduction in PPA may be interpreted as a central activation failure (26).

Sprint test.

A test consisting of a 40-m (20 + 20 m) shuttle sprint was used (31). This test was designed to measure both speed and change in direction abilities. The athletes started from a line, sprinted for 20 m, touched a cone with their hand, and then returned to the starting line as fast as possible. The sprints' times were recorded using a photocells system (Microgate, Bolzano, Italy), and four indices were calculated: 1) time to complete entire distance (40 m, SPR40m) (ICC = 0.94, SEM as a CV of 0.7%), 2) time to complete acceleration phase (between 0 and 15 m, SPRacp) (ICC = 0.89, SEM as a CV of 1.2%), 3) time to complete direction change phase (between 15 and 25 m, SPRdcp) (ICC = 0.89, SEM as a CV of 1.2%), and 4) time to complete flying phase (between 25 and 40 m, SPRflp) (ICC = 0.89, SEM as a CV of 1.3%). After two submaximal trials, the best of three sprints (with 1-min rest between each sprint) was used for the final statistical analysis.


To obtain an objective measure of short-passing ability, the first version of the LSPT was used (1). This version of LSPT has been shown to be a reliable and valid test of passing ability in soccer players (1). The test consisted of 16 passes performed within a circuit of cones and grids in a clockwise direction, as quickly and as accurately as possible. Four colored target areas measuring 30 × 60 cm (two yellow and two blue) and one colored target area measuring 30 × 10 cm (purple) were situated on a purpose-built wooden target (30 × 250 cm). Subjects started with the ball in the central box, and after the operator starting signal, they were asked to dribble the ball into the passing area, to pass the ball against the first target (target A), to control the returned ball, to dribble the ball through the central box, and then to move to the next target (target B) in a clockwise direction. The ball had to be taken through the central box and out of it before playing the next pass, and for every pass, the ball had to be passed inside the passing area. Penalties were accrued according to the following criteria:

  •  5 s for completely missing the target or hitting the wrong target,
  •  3 s for missing the colored target areas but hitting the target (i.e., hitting an angle of the target),
  •  2 s for hitting the yellow target area,
  •  1 s for hitting the blue target area,
  •  2 s for passing the ball from outside the designated passing area,
  •  2 s for the ball touching any cone, and
  •  1 s for every second taken over the assigned 43 s to complete the test.

On the contrary, 1 s was subtracted from the overall performance time for every time the ball hit the 10-cm purple strip in the middle of the target area. Three indices of performance were calculated: 1) time necessary to complete the 16 passes and to return in the central box without the penalties accumulated (LSPT time) (ICC = 0.87, SEM as a CV of 2.3%); 2) penalties calculated from the errors committed by each player during the test execution (LSPT penalties) (ICC = 0.75, SEM as a CV 30.8%); and 3) total performance (LSPT total) consisting of the time necessary to complete the test after adjusting for penalties and bonus time (ICC = 0.85, SEM as a CV of 3.7%). The schematic representation of the LSPT circuit has been presented in a previous study (24). Two submaximal trials of LSPT were performed before the beginning of the test with the better of two attempts used for the final statistical analysis. The rating of perceived exertion (RPE) relative to LSPT was also collected using a printout of the Borg CR10 scale (9). All the athletes had been habituated to this scale before the start of the study and followed standardized instructions for rating perceived exertion (9).

Match-related physical performance and exercise intensity.

The match was monitored using a video-computerized semiautomatic match analysis image recognition system (Wisport, Genova, Italy). This method uses multiple cameras providing a simultaneous observation of the 22 players involved in the game. From the data obtained using the Wisport specific software, TD and high-intensity running distance (HIR, distance covered with a running speed higher than 15.0 km·h−1) covered during the match were chosen for further analysis because these have previously been shown to be suitable parameters of soccer-specific fatigue (28). During the game, HR was recorded every 5 s. After the game, each player was again asked to give an RPE using the Borg CR10 scale (9).

Perceived muscle soreness.

Perceived muscle soreness was assessed using a printed visual analog scale (VAS) with two extreme polar descriptors ("no pain" and "worst possible pain"). Subjects were asked to place a vertical marker intersecting with the 10-cm horizontal line at a distance between the two extreme descriptors that represented their current level of pain. Muscle soreness was assessed twice before the match (PRE1 and PRE2) (ICC = 0.65, SEM as a CV of 18.5%) and three times after (POST, POST24, and POST48).

Biochemical parameters.

Blood samples were collected into evacuated tubes (Vacutainer; BD Diagnostics, Franklin Lakes, NJ) without additive and centrifuged after clotting was completed to obtain serum. The serum was divided from the clot and stored at −20°C for no more than 5 d before assay. Total activity of CK was determined using a fully automated Aeroset platform (Abbott Diagnostics, Chicago, IL) with dedicated reagents. The method is based on a coupled enzymes reaction in the presence of N-acetyl-l-cysteine (N-acetyl-l-cysteine method).

Statistical analysis.

All data are presented as mean ± SD. A one-way repeated-measure ANOVA was used on each dependent variable to test the differences among the time points (PRE1, PRE2, POST, POST24, and POST48) along the duration of the study. Because sphericity assumption was violated for few variables, the multivariate technique was used for all the variables to limit the type I error (35). When a significant F value was found, a Fisher LSD post hoc test was applied. As a measure of effect size, the partial eta squared (pη2) was calculated for the ANOVA, and values of 0.01, 0.06, and above 0.15 were considered small, medium, and large, respectively (11). Differences between the first- and second-half exercise intensity (average and peak HR) or match performance (TD and HIR) were determined using Student's paired t-test. Pearson product-moment correlation coefficients were also calculated to determine relationships between POST versus PRE2 changes of central fatigue indicators (%VA and MVC RMS/PPA) or peripheral fatigue indicators (PT at 1, 10, and 100 Hz) and exercise intensity and physical performance during the match, changes in MVC peak torque, sprint performance, short-passing ability, and perceived muscle soreness. The level of statistical significance was set at P < 0.05.


Match performance and intensity.

The mean TD and HIR covered during the game by the players were 11,764 ± 1044 and 2664 ± 772 m, respectively. Significant reductions in TD (−10.3%, P < 0.001) and HIR (−14.9%, P = 0.017) were found between the first and second half. The mean and the peak HR recorded during the first half were 175 ± 9 and 192 ± 7 bpm, which correspond to 88.0% ± 4.1% and 96.3% ± 2.8%, respectively, of previously determined actual maximal HR. During the second half, the same parameters were significantly lower (HRmean = 164 ± 8 bpm, 82.5% ± 4.0% of HRmax, P < 0.001 and HRpeak = 190 ± 7 bpm, 95.1% ± 3.0% of HRmax, P = 0.036). The RPE rated by athletes was 7.5 ± 1.1 au.

MVC and maximal voluntary activation.

Statistical analysis revealed a significant time effect for KE MVC (P < 0.001, pη2 = 0.39) and for KE %VA (P < 0.001, pη2 = 0.45) (Fig. 2). The peak torque was ∼11% lower in the POST session compared with PRE1 and PRE2 (P < 0.001). The mean value was only partially recovered 24 h after the match (vs POST = +6.8%, P = 0.017; vs PRE1 = −5.7%, P < 0.001; and vs PRE2 = −5.0%, P < 0.001) but was similar to before values 48 h after the match (P > 0.233). Similarly, %VA decreased significantly from PRE1 and PRE2 to POST (−7.7% and −7.6%, respectively, all P < 0.001), remained lower at POST24 (−3.3%, vs both PRE1 and PRE2, P = 0.019 and P = 0.022, respectively) and returned similar to baseline within 48 h (P > 0.470).

Time course of MVC (bars in the graph) and maximal activation level (line and circles in the graph) of the KE. PRE1, first baseline evaluation; PRE2, second baseline evaluation; POST, evaluation 40 min after the match; POST24, evaluation 24 h after the match; POST48, evaluation 48 h after the match.

Both RMS and RMS/PPA recorded during MVC were significantly affected by the soccer match (P = 0.027, pη2 = 0.16 and P = 0.001, pη2 = 0.24, respectively) (Fig. 3). Immediately after the game, RMS was significantly lower compared with PRE1 (−12.1%, P = 0.003) and with PRE2 (−9.3%, P = 0.026). Twenty-four hours after the match, RMS remained lower compared with PRE1 (−8.6%, P = 0.033); however, the difference was not significant compared with PRE2 (−5.7%, P = 0.168). Forty-eight hours after, RMS was similar to PRE1 and PRE2 (P > 0.340). Similarly, RMS/PPA was reduced immediately after the match (vs PRE1 = −12.4%, P < 0.001; vs PRE2 = −11.2%, P = 0.001) but were not significantly different compared with baseline at POST24 and POST48 (P > 0.135). Similar results were obtained when RMS of the M-wave was used for the EMG data normalization (data not shown) (5).

Time course of the electromyography (EMG) of VL collected during MVC. RMS, bars in the graph; RMS/PPA, line and squares in the graph.

M-wave and quadriceps contractile properties.

The M-wave characteristics were not altered by the soccer match. Neither differences in PPA (8.76 ± 2.40, 8.64 ± 2.26, 8.84 ± 2.58, 8.70 ± 2.69, and 8.58 ± 2.48 mV at PRE1, PRE2, POST, POST24, and POST48, respectively, P = 0.956, pη2 = 0.01) nor differences in peak-to-peak duration (6.48 ± 1.14, 6.54 ± 1.26, 6.64 ± 1.18, 6.26 ± 0.80, and 6.17 ± 0.75 ms at PRE1, PRE2, POST, POST24, and POST48, respectively, P = 0.386, pη2 = 0.06) were significant.

Quadriceps muscle contractile properties before and after the game are reported in Table 1. Significant changes were observed for PT at 1 Hz (P < 0.001, pη2 = 0.33) and PT at 10 Hz (P < 0.001, pη2 = 0.53), whereas PT at 100 Hz was not affected by the game (P = 0.272, pη2 = 0.08). Immediately after the match, PT at 1 Hz was significantly lower compared with PRE1 (−7.9%, P < 0.001) and PRE2 (−7.9%, P < 0.001). No other significant differences were observed (P > 0.073). Also, PT at 10 Hz was significantly lower immediately after the game compared with PRE1 (−9.0%, P < 0.001) and with PRE2 (−9.1%, P < 0.001). PT at 10 Hz measured at POST24 was significantly higher than POST (+6.3%, P < 0.001) but remained lower than baseline values (vs PRE1 = −3.2%, P = 0.013; vs PRE2 = −3.4%, P = 0.010). Forty-eight hours after the match, PT at 10 Hz was similar to PRE1 and PRE2 (P > 0.446). As a consequence, the ratios of 10/100 Hz were significantly reduced by exercise (P < 0.001, pη2 = 0.24). The ratios immediately after the match were significantly lower (all P values < 0.001) than the ratios before the match (0.90 ± 0.12, 0.90 ± 0.11, and 0.85 ± 0.11, respectively, for PRE1, PRE2, and POST). Although the ratios remained reduced 24 h after the game (0.87 ± 0.11), the differences with baselines values were close to but not significant (vs PRE1, P = 0.074; vs PRE2, P = 0.105). Ratios at 48 h were similar to baselines values (0.90 ± 0.08; vs PRE1, P = 0.995; vs PRE2, P = 0.857).

Quadriceps contractile properties associated with single and paired stimulations before and after the soccer match.

All CT (1, 10, and 100 Hz) were unaffected by exercise (P > 0.323, pη2 < 0.08). Changes of MRTD measured at 10 Hz were significant (P < 0.001, pη2 = 0.41), whereas at 1 and 100 Hz, the differences were close to significant (P = 0.086, pη2 = 0.12 and P = 0.075, pη2 = 0.12, respectively). MRTD at 10 Hz during the POST session was significantly reduced compared with PRE1 (−11.0%, P < 0.001) and with PRE2 (−10.3%, P < 0.001). At 24 h after the game, MRTD was partially recovered (vs POST = +6.3%, P = 0.003) but remained lower compared with PRE1 (−5.4%, P = 0.005) and with PRE2 (−4.7%, P = 0.014). Forty-eight hours after the game, MRTD was similar to PRE1 and PRE2 (P > 0.768). Significant differences were found for MRTR at 1 Hz (P = 0.044, pη2 = 0.14) and at 10 Hz (P < 0.001, pη2 = 0.41). However, no significant changes were observed for MRTR at 100 Hz (P = 0.257, pη2 = 0.08). MRTR at 1 and 10 Hz were significantly lower immediately after the match compared with PRE1 (−7.7%, P = 0.015 and −10.3%, P < 0.001, respectively) and with PRE2 (−8.5%, P = 0.007 and −9.9%, P < 0.001, respectively). Both MRTR at 1 and 10 Hz measured during POST24 and POST48 sessions were not significantly different from baselines values (P > 0.069).


Sprint performance was significantly affected by the soccer match (Table 2). In fact, SPR40m, SPRacp, SPRdcp, and SPRflp times were significantly elevated after the game (P < 0.001, pη2 = 0.44; P = 0.008, pη2 = 0.19; P < 0.001, pη2 = 0.29; and P = 0.003, pη2 = 0.22, respectively). In particular, SPR40m time at POST was significantly higher compared with PRE1 (+2.6%, P < 0.001) and with PRE2 (+2.9%, P < 0.001). SPR40m time at POST24 was lower than POST (−1.6%, P = 0.001) but remained higher compared with PRE1 (+0.9%, P = 0.049) and with PRE2 (+1.2%, P = 0.010). Twenty-four hours after the game, SPR40m time was similar to baselines values (P > 0.219). SPRacp time and SPRdcp time were significantly higher immediately after the match compared with PRE1 (+2.3%, P = 0.003 and +3.9%, P < 0.001, respectively) and with PRE2 (+2.7%, P = 0.001 and +4.4%, P < 0.001, respectively). However, both SPRacp and SPRdcp times at POST24 were similar to prematch values (P > 0.105). SPRflp time was higher after the match until 24 h (between 1.1% and 1.6%, P < 0.038) but returned similar to before values 48 h after the match (P > 0.580).

Sprint and short-passing ability performances (determined using LSPT) before and after a soccer match.


The players' short-passing ability was not affected by the game (Table 2). In fact, differences in LSPT total (P = 0.442, pη2 = 0.06), LSPT time (P = 0.172, pη2 = 0.09), and LSPT penalties (P = 0.389, pη2 = 0.06) were not significant. Only RPE (P < 0.001, pη2 = 0.51) associated with LSPT was elevated after the match compared with PRE1 (+49.7%, P < 0.001) and with PRE2 (+60.7, P < 0.001). However, at POST24 and at POST48, RPE was similar to baselines values (P > 0.422).

Perceived muscle soreness and biochemical parameters.

Perceived muscle soreness measured using the VAS was significantly affected by exercise (P < 0.001, pη2 = 0.59), and it remained elevated until 24 h (P < 0.046) when compared with baseline values (PRE1 = 1.12 ± 0.72 arbitrary unit (au), PRE2 = 1.09 ± 1.12 au, POST = 3.82 ± 1.65 au, POST24 = 1.79 ± 1.32 au, and POST48 = 1.71 ± 1.08 au). Plasma CK were significantly increased (P < 0.001, pη2 = 0.62) until 48 h after the match (P < 0.001) when compared with before values (PRE2 = 310 ± 226 U·L−1, POST = 652 ± 319 U·L−1, POST24 = 695 ± 375 U·L−1, and POST48 = 506 ± 258 U·L−1).

Relationships with central and peripheral fatigue indicators.

The relationships between the changes of central or peripheral fatigue indicators immediately after the match and the selected variables are reported in Table 3. Among the physical match performance indicators, only TD was significantly related to %VA changes (r = −0.51, P < 0.05). RPE of the match was significantly related to PT at 10-Hz decrement (r = −0.50, P < 0.05) and PT at 100-Hz decrement (r = −0.50, P < 0.05). MVC peak torque decrement was significantly associated with decrement in %VA (r = 0.86, P < 0.001) and decrement of RMS/PPA (r = 0.61, P < 0.01) but not with peripheral indicators (r < 0.27, P > 0.05). Central fatigue indicators were significantly related also to sprint performance changes. In fact, the relationships between %VA decrement and SPR40m, SPRacp, and SPRdcp were significant (r = −0.63, −0.55, and −0.52, respectively, all P < 0.05) as well as the relationships between MVC RMS/PPA decrement and SPR40m and SPRacp changes (r = −0.49 and r = −0.53, respectively, P < 0.05). LSPT total time changes were not significantly related to selected central or peripheral fatigue indicators. Postmatch perceived muscle soreness was significantly associated to peripheral fatigue indicators (r = −0.77, −0.65, and −0.57, for PT at 1-, 10-, and 100-Hz decrements, respectively, all P < 0.05) but not to the central fatigue ones.

Pearson product-moment correlation coefficients between changes of central and peripheral fatigue indicators and match performance, match intensity, changes of MVC, sprint performance, short-passing ability, and muscle soreness (POST vs PRE2 changes).


The main finding of the present study is that a soccer match performed by high-level professional players induces a reduction in MVC and sprinting ability. On the other hand, despite the increased RPE, short-passing ability is preserved. In this population, 48 h of recovery was found to be adequate to ensure a complete recovery of all the variables investigated. In addition, match-related fatigue is determined by a combination of central and peripheral factors. In fact, %VA and RMS/PPA recorded during MVC evaluation were reduced after the match as well as quadriceps contractile properties determined using passive electrical stimulations (in particular, paired stimulations delivered at 10 Hz). The highest level of central fatigue was shown in players with the greatest MVC and sprinting decline after the match, whereas athletes with a large increase of muscle soreness experienced a greater degree of peripheral fatigue.

Match performance and intensity.

Match performance and intensity were similar to those reported for high-level professional players. In fact, the mean and peak HR (∼85% and ∼95% of HRmax, respectively) were close to or slightly higher than values reported by several authors (6,13,17,18). Similarly, TD and HIR distances covered during the game were comparable to match distances previously reported for top-level players (28,32). The significant reduction of the physical performance (TD and HIR) between the first and the second halves and the high RPE suggest that players experienced match-related fatigue during this "friendly" game to a similar extent compared with "real" competition (28,32).

Fatigue after the match.

The MVC and sprinting ability of the professional players involved in the study were reduced (∼11% and ∼3%, respectively) immediately after the game, whereas short-passing ability remained stable. The MVC decrement is similar to the 9%-10% strength loss reported by other authors in male players after a match (6,18,36) and is considered moderate when compared with those (20%-35%) observed after prolonged running, cycling, or skiing exercises (26). Differences in exercise duration, exercise intensity, muscle contraction mode, and nature of the effort (continuous vs intermittent exercise) may explain the discrepancy between studies. The sprinting ability decrement was similar to the results obtained by Ispirlidis et al. (18) (∼2%) and by Krustrup et al. (22) (∼3%) using 20- and 30-m sprint tests but lower than that reported by Ascensão et al. (6) (∼7%). Because the exercise intensity of the matches among studies was similar, the different changes in sprinting performance might be ascribed to the soccer players' own characteristics. In addition, all three phases of the shuttle sprint test (acceleration, direction change, and flying phase) were significantly affected after the match. Contrary to a previous study (33), short-passing ability, evaluated using LSPT, was not reduced, and only RPE was elevated. It might be speculated that, in the previous study, the lower competitive level of the young players entailed a higher degree of match-related fatigue that significantly worsened short-passing ability. However, this hypothesis should be supported with future work that measures neuromuscular fatigue in soccer players of different levels. Moreover, the timing of assessment after the game (immediately after the match in the previous study vs 40 min after the match in the present study) may be another factor that may have influenced the different results.

During MVC evaluation, additional measurements were performed to determine the presence of central fatigue. Raw RMS of the VL EMG signal was significantly reduced during the POST session; however, this does not necessarily imply a reduction in central neural drive (26). In fact, EMG signal normalization using M-wave amplitude is useful to avoid the influence of any alterations in muscle excitability. The M-wave was found to be unchanged during the present study; however, the ratio RMS/PPA was significantly reduced after the game, suggesting that the reduced ability to produce voluntary force can be attributed in part to central factors. To confirm these results, %VA was measured using the interpolated-twitch method. The ∼8% drop of %VA further suggests that central neural drive has been altered after the match and may have influenced not only MVC but also other maximal performances like sprinting ability. As reported by Gandevia (15), central fatigue can originate from the spinal and/or supraspinal level; unfortunately, in this study, it was not possible to determine the exact origin. However, several biochemical alterations occurring after a soccer match (e.g., elevated free fatty acids, increased blood ammonia levels, interleukin-1 and -6 concentrations) (18,22) are a likely indirect evidence of a possible impairment of the supraspinal neural drive (25). In addition, the altered muscle pH and K+ during the game might stimulate group III and IV muscle afferents, which may inhibit motoneurons at the spinal level (15). Furthermore, a fusimotor system may be involved in a decreased %VA because it has been demonstrated that passive muscle stretching reduces stretch reflex (7).

Using passive electrical stimulations, it was possible to determine peripheral neuromuscular function before and after a soccer match. The unchanged M-wave characteristics seem to suggest that neuromuscular transmission-propagation of the action potential along the sarcolemma is preserved (14). The results of the present study are similar to other researches that investigated the effect of specific fatigue on quadriceps M-wave duration and intensity in other intermittent sports (tennis and squash) (16). Although neuromuscular characteristics of plantar flexors were not measured in the present study, they warrant future investigation because they seem more sensitive to fatigue than the quadriceps muscle (8). Moreover, M-wave characteristics of plantar flexors have been reported to be significantly affected by a tennis match (16). Quadriceps mechanical responses to single and paired stimulations at 10 Hz (low frequency) were negatively affected by the match. In fact, PT was reduced by 8%-9%, whereas MRTD and MRTR were reduced by 9%-10%. On the contrary, no significant differences were found using paired stimulations at 100 Hz (high frequency). As a consequence, the ratio 10/100 Hz was significantly reduced demonstrating the presence of low-frequency fatigue (long-lasting fatigue). These altered quadriceps contractile properties suggest that, in addition to central fatigue, a soccer match also induces peripheral fatigue. The presence of long-lasting fatigue in combination with increased muscle soreness, increased CK values, and decreased MVC suggests that muscle damage may occur during a game (38). In fact, long-lasting fatigue may be caused by both impaired excitation-contraction coupling and structural damage (2). It is well known that long-lasting fatigue is determined by a reduction in Ca2+ release, a decrease in Ca2+ binding to troponin, and/or structural muscle damage and/or redistribution (19).

Central and peripheral contribution to match-related fatigue.

Correlational analyses were performed with the purpose of assessing the potential influence of central and peripheral fatigue indicators on match performance and intensity, MVC, sprint performance, LSPT, and muscle soreness. The only significant moderate correlation among physical match performance and fatigue indicators was the relationship between %VA and TD. This suggests that, in a soccer match played by professional athletes, the total work done or the high-intensity distances covered per se have a limited effect on fatigue development. However, in the present study, absolute distances covered by the players have been considered. It would have been interesting to normalize distances run by the players for the individual fitness level because of its influence on soccer performance (20,21,31). Significant moderate relationships were found between RPE of the match and peripheral fatigue indicators (paired stimuli at 10 and 100 Hz), highlighting the link between increased fatigue perception and reduced quadriceps contractile properties. In fact, reduced neuromuscular peripheral function during the game might lead athletes to increase central motor drive to maintain a consistent performance over time. Consequently, the increased motor drive could lead to an increase in perception of exertion (34).

The relationships (from moderate to strong) found between impaired MVC or sprint performance decrement and central fatigue indicators are of interest. In fact, similar to concepts associated with fatigue after prolonged running exercise (27), our data support the notion that the ability to produce maximal performance after a soccer match is more influenced by central mechanisms related to fatigue rather than peripheral muscle contractile properties.

No significant correlations were found between changes in short-passing ability and central or peripheral fatigue indicators. However, as suggested above, it is possible that short-passing ability may be negatively influenced only when higher levels of fatigue are reached, although future studies are needed to better clarify this issue.

Finally, in this study, the reduced peripheral muscle function was associated with elevated muscle soreness (correlations between moderate and strong). Indeed, muscle damage and/or muscle inflammation are the most likely candidates to explain this relationship (10).

Recovery after the match.

In the present study, MVC, sprint performance, and muscle soreness (significantly affected by the match) returned to baseline 48 h after the game, and a similar pattern was followed by central and peripheral fatigue indicators. In fact, 1 d after the game, %VA, RMS of VL EMG, and quadriceps contractile properties determined using paired stimulations at 10 Hz (low-frequency stimulation indicator of long-lasting fatigue) were still reduced, but they were fully recovered by 48 h after the match. Interestingly, quadriceps contractile properties determined using paired stimulations at 100 Hz (indicator of high-frequency fatigue) were stable throughout the duration of the study. Other studies have reported that soccer match-induced fatigue persists for several hours after the game (6,18); however, the novelty of the present study consists in showing that match-related fatigue is determined by a combination of central and peripheral factors both immediately after the game and within 2 d of recovery. In this study, it was possible to determine the origin of fatigue; however, the physiological mechanisms were not specifically investigated. We might presume that the most likely candidates explaining this phenomenon are muscle damage, muscle glycogen depletion, general inflammation related to exercise, and muscle soreness per se, which are known to occur after a soccer match (23,26,30); however, future research is needed to better clarify this aspect.

The high-level professional players involved in this study showed a shorter recovery time compared with results reported by other authors (48 vs 72 h) (6,18). Because the exercise intensity of the matches among studies seems similar, differences in soccer players' performance and fitness levels may be the key factors to explain the slightly lower level of fatigue after the match and the faster recovery observed in the present study. Therefore, on the basis of the present results, it seems that high-level professional soccer players are physically trained to sustain two matches within 3 d. This is particularly relevant when we consider that high-level teams, engaging both in national and international competitions, play very often every 3-4 d. However, at the moment, we do not know, and it cannot be excluded, whether a cumulative effect on fatigue can occur when several matches are played in few days. Therefore, future research will aim to assess the time course of central and peripheral fatigue factors after several matches.


In conclusion, a soccer match performed by high-level professional players induced MVC and sprinting ability reductions, whereas short-passing ability was preserved. The variables negatively affected by the game returned to baseline after a 48-h recovery period. Match-related fatigue is determined by a combination of central and peripheral factors both immediately after the match and within hours of recovery. Central fatigue seems to be the main cause of the decline in MVC and sprinting ability, whereas peripheral fatigue seems to be more related to increased muscle soreness and therefore seems very likely linked to muscle damage and inflammation.

The research did not receive funding from the National Institutes of Health, Wellcome Trust, Howard Hughes Medical Institute, or any other source requiring deposit.

The authors report no conflict of interest.

The authors thank Mr. Fulvio Pea and Mr. Roberto Sassi for their valuable support and Mr. Paolo Menaspà, Mr. Domenico Carlomagno, Mr. Marco Riggio, Miss Lara Tonetti, Mr. Andrea Azzalin, and Mr. Maurizio Fanchini for their technical contribution during the data collection. The authors also thank Dr. Maria Rosa Graziani and Dr. Alberto Dolci of Cedal laboratory for the biochemical analysis and Mr. Matteo Campodonico and Mr. PierMaria Saltamacchia of Wisport for the match analysis. The authors also thank Unione Calcio Sampdoria and all the athletes involved in the study for their contribution. They also thank Miss Laura Garvican for her kind English revision.

The results of the present study do not constitute endorsement by American College of Sports Medicine.


1. Ali A, Eldred J, Hirst M, et al. The reliability and validity of the modified Loughborough Soccer Passing Test. J Sports Sci. 2003;21(4):258.
2. Allen DG. Eccentric muscle damage: mechanisms of early reduction of force. Acta Physiol Scand. 2001;171(3):311-9.
3. Amann M, Romer LM, Pegelow DF, Jacques AJ, Hess CJ, Dempsey JA. Effects of arterial oxygen content on peripheral locomotor muscle fatigue. J Appl Physiol. 2006;101(1):119-27.
4. Andersson H, Raastad T, Nilsson J, Paulsen G, Garthe I, Kadi F. Neuromuscular fatigue and recovery in elite female soccer: effects of active recovery. Med Sci Sports Exerc. 2008;40(2):372-80.
5. Arabadzhiev TI, Dimitrov VG, Dimitrova NA, Dimitrov GV. Interpretation of EMG integral or RMS and estimates of "neuromuscular efficiency" can be misleading in fatiguing contraction. J Electromyogr Kinesiol. 2010;20(2):223-32.
6. Ascensão A, Rebelo A, Oliveira E, Marques F, Pereira L, Magalhães J. Biochemical impact of a soccer match-analysis of oxidative stress and muscle damage markers throughout recovery. Clin Biochem. 2008;41(10-11):841-51.
7. Avela J, Kyröläinen H, Komi PV. Altered reflex sensitivity after repeated and prolonged passive muscle stretching. J Appl Physiol. 1999;86(4):1283-91.
8. Behm DG, St-Pierre DM. Effects of fatigue duration and muscle type on voluntary and evoked contractile properties. J Appl Physiol. 1997;82(5):1654-61.
9. Borg G. Borg's Perceived Exertion and Pain Scales. Champaign (IL): Human Kinetics; 1998. p. 39-43.
10. Cheung K, Hume P, Maxwell L. Delayed onset muscle soreness: treatment strategies and performance factors. Sports Med. 2003;33(2):145-64.
11. Cohen J. Statistical Power Analysis for the Behavioral Sciences. Mahwah (NJ): Lawrence Erlbaum; 1988. p. 16-8.
12. Edwards RH, Hill DK, Jones DA, Merton PA. Fatigue of long duration in human skeletal muscle after exercise. J Physiol. 1977;272(3):769-78.
13. Ekblom B. Applied physiology of soccer. Sports Med. 1986;3(1):50-60.
14. Fitts RH. Cellular mechanisms of muscle fatigue. Physiol Rev. 1994;74(1):49-94.
15. Gandevia SC. Spinal and supraspinal factors in human muscle fatigue. Physiol Rev. 2001;81(4):1725-89.
16. Girard O, Millet GP. Neuromuscular fatigue in racquet sports. Phys Med Rehabil Clin N Am. 2009;20(1):161-73, ix.
17. Helgerud J, Engen LC, Wisloff U, Hoff J. Aerobic endurance training improves soccer performance. Med Sci Sports Exerc. 2001;33(11):1925-31.
18. Ispirlidis I, Fatouros IG, Jamurtas AZ, et al. Time-course of changes in inflammatory and performance responses following a soccer game. Clin J Sport Med. 2008;18(5):423-31.
19. Jones DA. High-and low-frequency fatigue revisited. Acta Physiol Scand. 1996;156(3):265-70.
20. Krustrup P, Mohr M, Amstrup T, et al. The Yo-Yo intermittent recovery test: physiological response, reliability, and validity. Med Sci Sports Exerc. 2003;35(4):697-705.
21. Krustrup P, Mohr M, Nybo L, Jensen JM, Nielsen JJ, Bangsbo J. The Yo-Yo IR2 test: physiological response, reliability, and application to elite soccer. Med Sci Sports Exerc. 2006;38(9):1666-73.
22. Krustrup P, Mohr M, Steensberg A, Bencke J, Kjaer M, Bangsbo J. Muscle and blood metabolites during a soccer game: implications for sprint performance. Med Sci Sports Exerc. 2006;38(6):1165-74.
23. Lamb GD. Mechanisms of excitation-contraction uncoupling relevant to activity-induced muscle fatigue. Appl Physiol Nutr Metab. 2009;34(3):368-72.
24. Lyons M, Al-Nakeeb Y, Nevill A. Performance of soccer passing skills under moderate and high-intensity localized muscle fatigue. J Strength Cond Res. 2006;20(1):197-202.
25. Meeusen R, Watson P, Hasegawa H, Roelands B, Piacentini MF. Central fatigue: the serotonin hypothesis and beyond. Sports Med. 2006;36(10):881-909.
26. Millet GY, Lepers R. Alterations of neuromuscular function after prolonged running, cycling and skiing exercises. Sports Med. 2004;34(2):105-16.
27. Millet GY, Martin V, Lattier G, Ballay Y. Mechanisms contributing to knee extensor strength loss after prolonged running exercise. J Appl Physiol. 2003;94(1):193-8.
28. Mohr M, Krustrup P, Bangsbo J. Match performance of high-standard soccer players with special reference to development of fatigue. J Sports Sci. 2003;21(7):519-28.
29. Mohr M, Krustrup P, Bangsbo J. Fatigue in soccer: a brief review. J Sports Sci. 2005;23(6):593-9.
30. Nicol C, Avela J, Komi PV. The stretch-shortening cycle: a model to study naturally occurring neuromuscular fatigue. Sports Med. 2006;36(11):977-99.
31. Rampinini E, Bishop D, Marcora SM, Ferrari Bravo D, Sassi R, Impellizzeri FM. Validity of simple field tests as indicators of match-related physical performance in top-level professional soccer players. Int J Sports Med. 2007;28(3):228-35.
32. Rampinini E, Coutts AJ, Castagna C, Sassi R, Impellizzeri FM. Variation in top level soccer match performance. Int J Sports Med. 2007;28(12):1018-24.
33. Rampinini E, Impellizzeri FM, Castagna C, Azzalin A, Ferrari Bravo D, Wisløff U. Effect of match-related fatigue on short-passing ability in young soccer players. Med Sci Sports Exerc. 2008;40(5):934-42.
34. Taylor JL, Gandevia SC. A comparison of central aspects of fatigue in submaximal and maximal voluntary contractions. J Appl Physiol. 2008;104(2):542-50.
35. Thomas JR, Nelson JK. Research Methods in Physical Activity. 5th ed. Champaign (IL): Human Kinetics; 2005. p. 169.
36. Thorlund JB, Aagaard P, Madsen K. Rapid muscle force capacity changes after soccer match play. Int J Sports Med. 2009;30(4):273-8.
37. Verges S, Maffiuletti NA, Kerherve H, Decorte N, Wuyam B, Millet GY. Comparison of electrical and magnetic stimulations to assess quadriceps muscle function. J Appl Physiol. 2009;106(2):701-10.
38. Warren GL, Lowe DA, Armstrong RB. Measurement tools used in the study of eccentric contraction-induced injury. Sports Med. 1999;27(1):43-59.


©2011The American College of Sports Medicine