International Journal of Sports Physiology and Performance, 2021, 16, 1270-1280

https://doi.org/10.1123/ijspp.2020-0380
© 2021 Human Kinetics, Inc. ORIGINAL INVESTIGATION

Recovery Kinetics Following Small-Sided Games in Competitive

Soccer Players: Does Player Density Size Matter?
Konstantinos Papanikolaou, Panagiotis Tsimeas, Angeliki Anagnostou, Alexandros Varypatis,
Christos Mourikis, Theofanis Tzatzakis, Dimitrios Draganidis, Dimitrios Batsilas,
Theodoros Mersinias, Georgios Loules, Athanasios Poulios, Chariklia K. Deli, Alexios Batrakoulis,
Athanasios Chatzinikolaou, Magni Mohr, Athanasios Z. Jamurtas, and Ioannis G. Fatouros

Purpose: To examine the recovery kinetics of exercise-induced muscle damage (EIMD), neuromuscular fatigue, and performance
following small-sided games (SSGs) of different densities in soccer. Methods: Ten male players randomly completed 3 trials: a
control trial (no SSGs), 4v4 SSGs (62.5 m2/player), and 8v8 SSGs (284.4 m2/player). External and internal load were monitored
using GPS technology, heart-rate monitors, and rating of perceived exertion. Delayed-onset muscle soreness (DOMS), creatine
kinase (CK), isokinetic strength, countermovement jump (CMJ), and sprint were determined at baseline, as well as at 24, 48, and 72
hours post-SSGs. Neuromuscular fatigue was assessed at baseline and at 1, 2, and 3 hours post-SSGs. Results: DOMS increased
(P < .05) in 4v4 for 72 hours and in 8v8 for 24 hours with that of knee flexors being more pronounced than that of extensors. CK
increased (P < .05) in 4v4 for 72 hours and in 8v8 for 24 hours. Neuromuscular fatigue increased (P < .05) in 4v4 for 2 hours and in
8v8 for 3 hours. Strength declined (P < .05) in 4v4 for 48 hours and in 8v8 for 72 hours. CMJ decreased (P < .05) in 4v4 for 24 hours
and in 8v8 for 48 hours. Sprint decreased (P < .05) for 48 hours in 4v4 and for 72 hours in 8v8. Conclusions: SSGs are associated
with a prolonged rise of EIMD and induce short-term neuromuscular fatigue and slow recovery kinetics of strength, jump, and
sprinting performance. The time for complete recovery is longer for SSGs of lower density.

Keywords: association football, fatigue, high-intensity intermittent drills, muscle damage, muscle fatigue, player-to-pitch area

Soccer is an intermittent sport characterized by abrupt high- EIMD causes performance deterioration due to the loss of muscle
intensity actions interspersed with periods of low- to moderate-intensity fiber integrity as a result of sarcomeres overstretching, the devel-
efforts.1 Although soccer primarily engages oxidative metabolism, opment of edema, delayed-onset muscle soreness (DOMS), and
nonoxidative energy pathways are highly activated due to the substan- aseptic inflammation.8 Very intense soccer ball drills (eg, speed-
tial number of explosive movements (eg, accelerations, decelerations, endurance training) induce considerable EIMD, fatigue, and
sprints).2 Therefore, high-intensity intermittent ball drills are fre- marked performance decline.9 Although recent reports have shown
quently incorporated during a training microcycle. Small-sided compromised neuromuscular function and performance following
games (SSGs) are popular training protocols utilized to develop various SSGs, these studies used limited performance and EIMD
players’ capacity to perform high-intensity actions and to improve markers over a limited recovery time frame (∼24 h).10,11
technical and tactical abilities, using soccer-specific movement pat- To our knowledge, the recovery kinetics of EIMD, neuro-
terns replicating the demands of match play.3 Factors like pitch area, muscular fatigue, and performance markers following SSGs have
the number of players, and rules may influence the intensity and not been investigated. Information on the amount of time required
physiological adaptations.4 SSGs with a small number of players for complete performance restoration following SSGs would
(≤4v4) appear to raise the physiological loading, as evidenced by the determine the optimal frequency of SSG use during a training
higher heart rate (HR) and blood lactate (BLa) values,5 whereas larger microcycle. Thus, this investigation aimed to establish the recov-
pitch areas (>40 × 30 m) or lower densities (≥175 m2/player) induce a ery kinetics of EIMD, neuromuscular fatigue, and performance
greater external load.6 following SSGs of different density size. We hypothesized that
The SSGs incorporate repeated, high-intensity actions with a low-density SSGs would need a more prolonged recovery due to
distinct eccentric element, and they have been associated with the the greater distance covered, with actions incorporating eccentric
onset of exercise-induced muscle damage (EIMD) and fatigue.7 contractions.

Papanikolaou, Tsimeas, Anagnostou, Varypatis, Tzatzakis, Draganidis, Batsilas, Methods

Mersinias, Loules, Poulios, Deli, Batrakoulis, Jamurtas, and Fatouros are with the
Dept of Physical Education and Sport Science, and Jamurtas and Fatouros, also the Participants
Dept of Nutrition and Dietetics, University of Thessaly, Trikala, Greece. Mourikis is A power analysis (G*Power, version 3.1.9.2, Düsseldorf Univer-
with Olympiakos Piraeus F.C., Soccer Performance Lab, Piraeus, Greece. Chatzi-
nikolaou is with the Dept of Physical Education and Sport Science, Democritus
sity, Düsseldorf, Germany) indicated a sample size of 8 to 10
University of Thrace, Komotini, Greece. Mohr is with the Dept of Sports Science
participants for a repeated-measures design. Accordingly, 10
and Clinical Biomechanics, SDU Sport and Health Sciences Cluster (SHSC), players (Table 1) participated, using the following criteria: partici-
University of Southern Denmark, Odense, Denmark, and the Faculty of Health, pation at a competitive level (≥5 training sessions per week and
Center of Health Science, University of the Faroe Islands, Tórshavn, Faroe Islands. ≥1 match per week) for ≥4 years, no recent history of illness
Fatouros ([email protected]) is corresponding author. or musculoskeletal injury, and abstinence from ergogenic
1270
 Small-Sided Games and Recovery 1271

supplements and medications. The procedures were in accordance maximal voluntary isometric contraction (MVIC) of knee exten-
with the Declaration of Helsinki and its later amendments. The study sors (KE) and knee flexors (KF) of the dominant (DL, the limb
was approved by the University of Thessaly Institutional Ethics used to kick the ball) and nondominant limb (NDL) was deter-
Committee and registered at ClinicalTrials.gov (NCT03970707). mined at the baseline and every hour for 3 hours post-SSG as a
measure of neuromuscular fatigue. DOMS, creatine kinase (CK)
activity, concentric/eccentric isokinetic strength, countermove-
Study Overview ment jump (CMJ), and 30-m sprint time were evaluated at the
This study was performed 1 week after the completion of the baseline and daily for 72 hours of recovery. BLa and the rating of
competitive season, while players were still familiar with vigorous perceived exertion (RPE) were measured at the baseline and post-
training to minimize detraining. A randomized, crossover, repeated- SSG. Field locomotor activity during SSGs and HR were moni-
measures design was adapted with 3 trials (Figure 1): (1) control tored continuously. Between trials, a 7-day washout period of
trial (no training), (2) 4v4 (20 × 25 m, 62.5 m2/player) trial, and daily light training was applied.
(3) 8v8 (70 × 65 m, 284.4 m2/player) trial. Initially, the partici-
pants were familiarized with experimental procedures and had Small-Sided Games
their anthropometric profile and performance measured. The
The SSGs of small and large pitch areas were performed on a grass
surface under similar environmental conditions (20°C–24°C and
40%–50% humidity). During the SSGs, the participants consumed
Table 1 Participants’ Baseline Characteristics only water ad libitum. Before the SSGs, a standard breakfast and
meal were consumed.12 4v4 (6 sets of 4 min each, 180-s rest,
Age, y 21.7 (2.1)
62.5 m2 area/player, and high-density SSGs) and 8v8 (3 sets of
Body mass, kg 78.6 (5.3) 8 min each, 90-s rest, 284.4 m2 area/player, and low-density SSGs)
Height, m 1.81 (0.1) had the same total exercise duration (24 min). During each session,
BMI, kg/m2 24.0 (1.9) 5 players participated, while substitute players were used to
Body fat, % 20.7 (5.9) account for the total number of players. This procedure was
Body fat, kg 14.5 (5.4) repeated twice for both SSGs (4v4 and 8v8) to include all 10
players. The players were allocated to each team in a balanced
Lean body mass, kg 58.8 (6.2)
manner. The main aim of the SSG trials was ball possession, and
VO2max, mL/kg/min 56.8 (3.4) throughout each repetition, the players were verbally encouraged
Yo-Yo IE2, m 2196.0 (246.1) to perform at high intensity. No goalkeepers were included, and
Yo-Yo IR2, m 1272.0 (141.1) the players were limited to 2 ball touches. The players were matched
Creative speed test, s 16.7 (0.6) considering their individual technical skill level. A passive recovery
Short dribbling test, s 12.1 (0.5) was used between repetitions. A standard 20-min warm-up (shuttle
running, dynamic stretching, neuromuscular coordination exercises,
Abbreviations: BMI, body mass index; VO2max, maximal oxygen consumption; and low-intensity agility ball drills) and a 10-min cooldown (mainly
Yo-Yo IE2, Yo-Yo intermittent endurance level 2 test; Yo-Yo IR2, Yo-Yo
intermittent recovery level 2 test. Note: Data are presented as means (SDs). static stretching exercises) were applied.

Figure 1 — The experimental flowchart. CMJ indicates countermovement jump; DOMS, delayed-onset muscle soreness; DXA, dual-energy X-ray
absorptiometry; IE, intermittent endurance; IR, intermittent recovery; MVIC, maximal voluntary isometric contraction; VO2max, maximal oxygen
consumption.

IJSPP Vol. 16, No. 9, 2021

1272 Papanikolaou et al

Measurements
Descriptives. Body mass and height (day 1) were measured Results
on a beam balance with a stadiometer, using standard procedures, The participants’ baseline characteristics (Table 1) for all variables
as described.7 Body composition (day 1) was assessed using dual- were comparable among trials. The effect sizes and confidence
energy X-ray absorptiometry, as previously described.13 VO2max intervals for both time-dependent and between-group differences
(day 2) was determined via open-circuit spirometry, using an are shown in Table 2.
automated online pulmonary gas exchange analyzer (Vmax Encore
29, BEBJO296, Yorba Linda, CA) via breath-by-breath analysis
during a graded exercise test on a treadmill, as described.13 Soccer-
Performance and Physiological Responses
specific endurance performance was assessed with the Yo-Yo During SSGs
intermittent endurance level 2 test (day 4) and the Yo-Yo intermit- The 2 SSG sessions were comparable in respect to total distance, mean
tent recovery level 2 test (day 6), as described.12,14 Participants’ speed, intense accelerations, and decelerations. In 8v8, the players had
level of technical performance (day 7) was determined using higher HIR (P = .000; Table 3), HSR (P = .018; Table 3), and maxi-
creative speed and short dribbling tests, as previously described.9 mum speed (P = .001; Table 3) than 4v4. The mean HR (P = .024),
Field GPS Activity and Physiological Responses. During SSGs, peak HR (P = .047), and time at >90% HRmax (P = .027) were greater in
external load and internal load were monitored using a wearable GPS 4v4 compared with 8v8. Postexercise RPE was higher in 8v8 compared
system and HR telemetry (10 Hz; 200-Hz triaxial accelerometry; Polar with 4v4 (P = .000; Table 3). BLa increased immediately postexercise
Team Pro; Polar Electro, Kempele, Finland), respectively. Activity was in both sessions compared with the baseline (4v4: P = .000; 8v8:
classified as the total distance high-intensity running (HIR, distance P = .003), with a greater increase in 4v4 (P = .001) than 8v8 (Figure 2).
covered at speeds 14–21 km/h), high-speed running (HSR, distance
covered at speeds >21 km/h), number of intense accelerations (>2 m/ Muscle Damage
s2), and number of intense decelerations (>2 m/s2). The mean and
maximum speed and HR were recorded and were expressed as a In 4v4, DOMS (Figure 3) of KE of both limbs increased postexer-
percentage of maximum speed (achieved during the baseline 30-m cise (DL: P = .001; NDL: P = .001), at 24 hours (DL: P = .003;
sprint test) and maximum HR (achieved during the baseline VO2max NDL: P = .031), at 48 hours (DL: P = .008; NDL: P = .045), and
test), respectively. RPE was monitored using the 10-point scale. normalized at 72 hours. DOMS of KF of both limbs increased
postexercise (DL: P = .001; NDL: P = .000), at 24 hours (DL:
Performance. The MVIC and concentric and eccentric peak tor- P = .000; NDL: P = .000), and 48 hours (DL: P = .000; NDL:
que of KE and KF at 60°/s were measured on an isokinetic P = .000), and remained elevated at 72 hours (DL: P = .004;
dynamometer (Cybex Norm, Ronkonkoma, New York, NY) using NDL: P = .000).
a 5-repetition protocol, as previously described.15 CMJ was measured In 8v8, DOMS of KE of DL increased postexercise (P = .017)
on a contact platform (Chronojump; Bosco-System, Barcelona, and at 24 hours (P = .04), and normalized thereafter, while in NDL
Spain), as described.16 The 30-m sprint time was measured using increased only postexercise (P = .035). DOMS of KF of both limbs
infrared, single beam photocells (Chronojump, Bosco-System), with increased postexercise (DL: P = .006; NDL: P = .002) and at
a sampling rate of 100/s, as described elsewhere.17 DOMS of the KE 24 hours (DL: P = .007; NDL: P = .018), and normalized thereafter.
and KF was evaluated by palpation, as previously described.18 In 4v4, CK (Figure 4) increased throughout recovery (24 h:
Blood Sampling. After overnight fasting, blood samples (∼12 mL) P = .001; 48 h: P = .000; 72 h: P = .000). In 8v8, CK increased at
were collected (8:00–9:00 AM) from an antecubital vein by veni- 24 hours (P = .011) and normalized thereafter. CK in 4v4 were
puncture with the participants in a supine position. With commercially higher than 8v8 at 48 hours (P = .004) and 72 hours (P = .009). CK
available kits (P. Zafiropoulos S.A., Athens, Greece), serum was demonstrated a greater change in 4v4 than 8v8 at 48 and 72 hours.
separated by centrifugation (1370g at 4°C for 10 min) from blood
samples that were collected into tubes containing coagulation factor to Neuromuscular Fatigue
measure (in duplicate) CK activity using an automated Clinical
Chemistry Analyzer Z1145 (P. Zafiropoulos S.A.). The BLa concen- In 4v4, MVIC (Figure 5) of KE of both limbs decreased only at
tration was measured in capillary blood using a hand-portable 1 hour (DL: P = .000; NDL: P = .000) and normalized thereafter,
automated analyzer (Lactate Plus; Nova Biomedical, Waltham, MA), whereas MVIC of KF of both limbs decreased at 1 hour (DL:
as previously described.19 P = .000; NDL: P = .000) and at 2 hours (DL: P = .000; NDL:
P = .000), and normalized thereafter.
In 8v8, MVIC of KE of both limbs decreased at 1 hour (DL:
Statistical Analysis P = .000; NDL: P = .000), at 2 hours (DL: P = .000; NDL:
P = .000), and at 3 hours (DL: P = .003; NDL: P = .034). MVIC
The data are shown as means (SDs). The Shapiro–Wilk test was of KF decreased in both limbs at 1 hour (DL: P = .000; NDL:
utilized to verify data normality. Between- and within-group differ- P = .000), at 2 hours (DL: P = .000; NDL: P = .000), and at 3 hours
ences were analyzed using a 2-way (condition vs time) repeated- only in NDL (P = .000). No differences were observed between
measures analysis of variance, with planned contrasts on time points groups at any time point.
and a Bonferroni test for the post hoc analysis. Statistical significance
was accepted at P < .05. The Hedge g method (corrected for bias) was Isokinetic Strength
utilized to estimate effect sizes (ESs) and CIs. ESs were translated as
minimal, small, medium-, and large-sized for values 0.00 to 0.19, In 4v4, the concentric strength of KE decreased only in DL at
0.20 to 0.49, 0.50 to 0.79, and ≥0.8 respectively, according to the 24 hours (P = .013), whereas in KF, it decreased in both limbs at
Cohen d criteria. The IBM SPSS Statistics for Windows was used for 24 hours (DL: P = .000; NDL: P = .000) and recovered thereafter
analyses (version 20; IBM Corp, Armonk, NY). (Figure 6). The eccentric strength of KE and KF of both limbs
IJSPP Vol. 16, No. 9, 2021
Table 2 P, ES, and CI for All Significant Differences Observed
Variable Time-dependent differences Between-groups differences
Blood lactate
4v4, baseline vs post: P = .000; −3.24; −4.57 to −1.91 Post, 4v4 vs control: P = .000; −3.07; −4.36 to −1.78
8v8, baseline vs post: P = .003; −1.70; −2.73 to −0.68 Post, 8v8 vs control: P = .014; −1.72; −2.75 to −0.69
Post, 4v4 vs 8v8: P = .002; 1.36; 0.38 to 2.33
DOMS, KE, and DL
4v4, baseline vs post: P = .001; −1.36; −2.33 to −0.39 Post, 4v4 vs control: P = .008; −1.38; −2.35 to −0.40
4v4, baseline vs 24 h: P = .003; −1.70; −2.73 to −0.68 24 h, 4v4 vs control: P = .020; −1.70; −2.73 to −0.68
4v4, baseline vs 48 h: P = .008; −1.15; −2.10 to −0.20 48 h, 4v4 vs control: P = .038; −3.07; −2.12 to −0.22
8v8, baseline vs post: P = .017; −1.27; −2.23 to −0.31
8v8, baseline vs 24 h: P = .043; −0.79; −1.70 to 0.12
DOMS, KE, and NDL
4v4, baseline vs post: P = .001; −1.40; −2.38 to −0.42 Post, 4v4 vs control: P = .011; −1.40; −2.38 to −0.42
4v4, baseline vs 24 h: P = .031; −0.99; −1.92 to −0.06
4v4, baseline vs 48 h: P = .045; −1.27; −2.23 to −0.31
8v8, baseline vs post: P = .035; −1.27; −2.23 to −0.31
DOMS, KF, and DL
4v4, baseline vs post: P = .001; −1.63; −2.64 to −0.62 Post, 4v4 vs control: P = .008; −1.56; −2.57 to −0.56
4v4, baseline vs 24 h: P = .000; −1.56; −2.57 to −0.56 24 h, 4v4 vs control: P = .005; −1.65; −2.67 to −0.64
4v4, baseline vs 48 h: P = .000; −1.49; −2.48 to −0.50 48 h, 4v4 vs control: P = .005; −1.49; −2.48 to −0.50
4v4, baseline vs 72 h: P = .004; −1.27; −2.23 to −0.31 72 h, 4v4 vs control: P = .024; −1.29; −2.25 to −0.33
8v8, baseline vs post: P = .006; −1.43; −2.41 to −0.45 Post, 8v8 vs control: P = .030; −1.43; −2.41 to −0.45
8v8, baseline vs 24 h: P = .007; −1.40; −2.38 to −0.42 24 h, 8v8 vs control: P = .035; −1.40; −2.38 to −0.42
DOMS, KF, and NDL
4v4, baseline vs post: P = .000; −1.68; −2.70 to −0.66 Post, 4v4 vs control: P = .003; −1.70; −2.72 to −0.67
4v4, baseline vs 24 h: P = .000; −1.68; −2.70 to −0.66 24 h, 4v4 vs control: P = .002; −1.70; −2.72 to −0.67
4v4, baseline vs 48 h: P = .000; −2.12; −3.21 to −1.02 Post, 8v8 vs control: P = .015; −1.65; −2.67 to −0.64
4v4, baseline vs 72 h: P = .000; −1.96; −3.02 to −0.89
8v8, baseline vs post: P = .002; −1.63; −2.64 to −0.62
8v8, baseline vs 24 h: P = .018; −1.40; −2.38 to −0.42
Creatine kinase
4v4, baseline vs 24 h: P = .001; −1.43; −2.41 to −0.44 24 h, 4v4 vs control: P = .017; −1.39; −2.36 to −0.41
4v4, baseline vs 48 h: P = .000; −2.06; −3.14 to −0.97 48 h, 4v4 vs control: P = .000; −2.20; −3.31 to −1.09
4v4, baseline vs 72 h: P = .000; −1.46; −2.44 to −0.47 72 h, 4v4 vs control: P = .001; −1.52; −2.52 to −0.53
8v8, baseline vs 24 h: P = .011; −1.18; −2.14 to −0.23 48 h, 4v4 vs 8v8: P = .004; 1.25; 0.30 to 2.21
72 h, 4v4 vs 8v8: P = .009; 1.17; 0.22 to 2.12
MVIC, KE, and DL
4v4, baseline vs 1 h: P = .000; 0.54; −0.35 to 1.43
8v8, baseline vs 1 h: P = .000; 0.90; −0.02 to 1.82
8v8, baseline vs 2 h: P = .000; 0.50; −0.39 to 1.39
8v8, baseline vs 3 h: P = .003; 0.17; −0.71 to 1.05
MVIC, KE, and NDL
4v4, baseline vs 1 h: P = .000; 0.47; −0.42 to 1.36
8v8, baseline vs 1 h: P = .000; 0.77; −0.13 to 1.68
8v8, baseline vs 2 h: P = .000; 0.39; −0.49 to 1.27
8v8, baseline vs 3 h: P = .034; 0.10; −0.77 to 0.98
MVIC, KF, and DL
4v4, baseline vs 1 h: P = .000; 0.66; −0.24 to 1.56
4v4, baseline vs 2 h: P = .000; 0.43; −0.46 to 1.31
(continued)

IJSPP Vol. 16, No. 9, 2021 1273

Table 2 (continued)
Variable Time-dependent differences Between-groups differences
8v8, baseline vs 1 h: P = .000; 0.85; −0.07 to 1.76
8v8, baseline vs 2 h: P = .000; 0.73; −0.18 to 1.63
MVIC, KF, and NDL
4v4, baseline vs 1 h: P = .000; 0.66; −0.24 to 1.56
4v4, baseline vs 2 h: P = .000; 0.43; −0.46 to 1.32
8v8, baseline vs 1 h: P = .000; 0.90; −0.02 to 1.82
8v8, baseline vs 2 h: P = .000; 0.75; −0.16 to 1.66
8v8, baseline vs 3 h: P = .000; 0.59; −0.31 to 1.48
Concentric strength, KE, and DL
4v4, baseline vs 24 h: P = .013; 0.69; −0.21 to 1.59 24 h, 8v8 vs control: P = .014; 1.32; 0.35 to 2.29
8v8, baseline vs 24 h: P = .000; 1.23; 0.28 to 2.19
8v8, baseline vs 48 h: P = .000; 0.83; −0.08 to 1.75
Concentric strength, KE, and NDL
8v8, baseline vs 24 h: P = .000; 1.03; 0.09 to 1.96 24 h, 8v8 vs control: P = .036; 1.08; 0.15 to 2.02
8v8, baseline vs 48 h: P = .000; 0.70; −0.20 to 1.61
Concentric strength, KF, and DL
4v4, baseline vs 24 h: P = .000; 0.37; −0.51 to 1.25
8v8, baseline vs 24 h: P = .000; 0.23; −0.65 to 1.11
Concentric strength, KF, and NDL
4v4, baseline vs 24 h: P = .000; 0.39; −0.49 to 1.27
8v8, baseline vs 24 h: P = .000; 0.42; −0.47 to 1.30
Eccentric strength, KE, and DL
4v4, baseline vs 24 h: P = .000; 0.69; −0.21 to 1.60 24 h, 8v8 vs control: P = .018; 1.20; 0.25 to 2.16
4v4, baseline vs 48 h: P = .000; 0.37; −0.51 to 1.26
8v8, baseline vs 24 h: P = .000; 1.35; 0.38 to 2.32
8v8, baseline vs 48 h: P = .000; 0.94; 0.01 to 1.86
8v8, baseline vs 72 h: P = .000; 0.17; −0.71 to 1.05
Eccentric strength, KE, and NDL
4v4, baseline vs 24 h: P = .000; 0.77; −0.13 to 1.68 24 h, 8v8 vs control: P = .015; 1.25; 0.30 to 2.21
4v4, baseline vs 48 h: P = .000; 0.37; −0.51 to 1.26
8v8, baseline vs 24 h: P = .000; 1.28; 0.32 to 2.24
8v8, baseline vs 48 h: P = .000; 0.80; −0.11 to 1.71
8v8, baseline vs 72 h: P = .001; 0.14; −0.74 to 1.02
Eccentric strength, KF, and DL
4v4, baseline vs 24 h: P = .000; 0.85; −0.06 to 1.77 24 h, 8v8 vs control: P = .010; 1.51; 0.52 to 2.50
4v4, baseline vs 48 h: P = .000; 0.40; −0.49 to 1.28
8v8, baseline vs 24 h: P = .000; 1.60; 0.59 to 2.61
8v8, baseline vs 48 h: P = .000; 1.04; 0.11 to 1.98
8v8, baseline vs 72 h: P = .001; 0.23; −0.65 to 1.11
Eccentric strength, KF, and NDL
4v4, baseline vs 24 h: P = .000; 0.87; −0.05 to 1.78 24 h, 8v8 vs control: P = .010; 1.40; 0.42 to 2.38
4v4, baseline vs 48 h: P = .000; 0.42; −0.47 to 1.31
8v8, baseline vs 24 h: P = .000; 1.34; 0.37 to 2.31
8v8, baseline vs 48 h: P = .000; 0.86; −0.05 to 1.78
8v8, baseline vs 72 h: P = .002; 0.17; −0.71 to 1.05
Countermovement jump
4v4, baseline vs 24 h: P = .001; 0.28; −0.60 to 1.16
8v8, baseline vs post: P = .041; 0.39; −0.50 to 1.27
8v8, baseline vs 24 h: P = .000; 0.54; −0.35 to 1.44
(continued)

1274 IJSPP Vol. 16, No. 9, 2021

 Small-Sided Games and Recovery 1275

Table 2 (continued)
Variable Time-dependent differences Between-groups differences
8v8, baseline vs 48 h: P = .016; 0.33; −0.55 to 1.22
30-m sprint time
4v4, baseline vs 24 h: P = .000; −1.06; −2.00 to −0.13 24 h, 4v4 vs control: P = .049; −1.11; −2.05 to −0.17
4v4, baseline vs 48 h: P = .005; −0.33; −1.21 to 0.55 24 h, 8v8 vs control: P = .011; −1.42; −2.40 to −0.44
8v8, baseline vs 24 h: P = .000; −1.40; −2.38 to −0.42
8v8, baseline vs 48 h: P = .000; −0.69; −1.59 to 0.21
8v8, baseline vs 72 h: P = .034; −0.16; −1.04 to 0.71
Abbreviations: CI, confidence interval; DL, dominant limb; DOMS, delayed-onset muscle soreness; ES, effect size; KE, knee extensors; KF, knee flexors; MVIC, maximal
voluntary isometric contraction; NDL, nondominant limb. Note: Data are presented as P; ES; 95% CI.

Table 3 Field GPS Activity and Physiological Responses During SSGs

Variable 4v4 8v8 ES; 95% CI
Total distance, m 2829.1 (413.4) 2975.4 (505.8)
High-intensity running, m 37.5 (25.9) 160.5 (78.1)* −2.01; −3.08 to −0.93
High-speed running, m 0.9 (0.9) 19.8 (20.9)* −1.21; −2.17 to 0.26
Mean speed, %max. speed 23.5 (3.1) 24.2 (3.7)
Maximum speed, % 59.8 (5.1) 73.4 (7.8)* −1.49; −2.48 to −0.50
Intense Acc counts, >2 m/s2 25.0 (3.1) 27.2 (6.7)
Intense Dec counts, >2 m/s2 7.8 (2.4) 7.5 (3.0)
Mean HR, %max. HR 88.7 (6.6) 81.5 (8.7)* 0.88; −0.03 to 1.80
Maximum HR, % 93.8 (6.1) 88.6 (8.2)* 0.68; −0.22 to 1.58
Time at >90% of HRmax, s 14.6 (8.3) 7.7 (8.2)* 0.79; −0.12 to 1.70
RPE, score 8.7 (1.2) 6.2 (1.1)* 1.98; 0.91 to 3.05
Abbreviations: %max., %maximum; Acc, accelerations; CI, confidence interval; Dec, decelerations; ES, effect size; HR, heart rate; HRmax, maximum HR; RPE, rating of
perceived exertion; SSGs, small-sided games. Note: Data are presented as means (SDs).
*Significant difference between SSGs.

P = .000; NDL: P = .000), and normalized thereafter. The concen-

tric strength of KF in both limbs decreased only at 24 hours (DL:
P = .000; NDL: P = .000) and recovered thereafter. The eccentric
strength of KE and KF of both limbs decreased at 24 hours (KE/
DL: P = .000; KE/NDL: P = .000; KF/DL: P = .000; KF/NDL:
P = .000), 48 hours (KE/DL: P = .000; KE/NDL: P = .000; KF/
DL: P = .000; KF/NDL: P = .000), and 72 hours (KE/DL: P = .000;
KE/NDL: P = .001; KF/DL: P = .001; KF/NDL: P = .002). No
differences were observed between the SSG trials.

Thirty-Meter Sprint Time

The 30-m sprint time in both SSG groups increased at 24 hours
(4v4: P = .000; 8v8: P = .000), at 48 hours (4v4: P = .005; 8v8:
P = .000), and at 72 hours only in the 8v8 group (P = .034)
Figure 2 — Time-course changes in blood lactate measurements. (Figure 7). No differences were observed between the SSG groups.
*Significant difference with baseline in 4v4, P < .05. #Significant
difference with baseline in 8v8, P < .05. †Significant difference between
4v4 and control within time point, P < .05. +Significant difference between Countermovement Jump
8v8 and control within time point, P < .05. ‡Significant difference between 4v4 In 4v4, CMJ decreased only at 24 hours (P = .001). In 8v8, CMJ
and 8v8 within time point, P < .05.
decreased immediately postexercise (P = .041), at 24 hours
(P = .000), and at 48 hours (P = .016), and normalized thereafter
demonstrated a similar decrement pattern at 24 hours (KE/DL: (Figure 7). No differences were detected between the SSG trials.
P = .000; KE/NDL: P = .000; KF/DL: P = .000; KF/NDL: P =
.0007) and 48 hours (KE/DL: P = .000; KE/NDL: P = .000; KF/DL: Discussion
P = .000; KF/NDL: P = .000) and recovered at 72 hours.
In 8v8, the concentric strength of KE of both limbs decreased In this study, we examined the effects of 2 different SSGs on the
at 24 hours (DL: P = .000; NDL: P = .000) and at 48 hours (DL: recovery kinetics of EIMD, neuromuscular fatigue, and performance
IJSPP Vol. 16, No. 9, 2021
1276 Papanikolaou et al

Figure 3 — Time-course changes in DOMS measurements. DOMS indicates delayed-onset muscle soreness. *Significant difference with baseline in
4v4, P < .05. #Significant difference with baseline in 8v8, P < .05. †Significant difference between 4v4 and control within time point, P < .05. +Significant
difference between 8v8 and control within time point, P < .05.

compared to 8-a-side may be attributed to a higher ball possession

that is usually observed with higher densities.3,20 Running with the
ball is associated with a higher energy expenditure than running
without, thereby increasing the intensity in the drills.21 Moreover,
longer, less intense exercise durations than those recorded in 8v8
are normally associated with lower BLa concentrations.22 As such,
higher density SSGs are associated with higher intensity, despite that
the total distance covered was comparable. As expected for SSGs of
lower density,11 HIR and HSR were greater in 8v8. Intense accel-
erations and decelerations were comparable among SSGs, but lower
than those reported previously,23 a fact that could be attributed to the
specific SSG formats (player density and rules limitations) applied
here. It appears that these 2 SSG sessions replicated the workload
that athletes are usually exposed to during typical training based on
total distance covered, average intensity, and mean RPE values.9 It is
Figure 4 — Time-course changes in creatine kinase measurements.
*Significant difference with baseline in 4v4, P < .05. #Significant dif- noteworthy that no previous studies examined the recovery kinetics
ference with baseline in 8v8, P < .05. †Significant difference between 4v4 of SSGs within a 72-hour time frame.
and control within time point, P < .05. +Significant difference between 4v4 The first principal finding was that SSGs induce prolonged
and 8v8 within time point, P < .05. EIMD independent of their density, as previously shown for 4v4
and intensified speed-endurance protocols.9,10 The CK and DOMS
rise were of lower magnitude, but of a duration similar to that seen
in competitive male soccer players. Our results indicate that, following competitive soccer matches,15,19 with the latter being
although 4v4 induces a greater EIMD response than 8v8, the latter surprising, considering that only 24 minutes of efficient training
is associated with a more prolonged fatigue and slower recovery time were performed. Intense actions that incorporate a strong
kinetics of performance markers. eccentric component (ie, intense running, accelerations, and decel-
It appears that the washout period was effective, since the erations) probably contributed to this rise of EIMD markers.8
baseline values of all dependent variables were comparable Strenuous short-term intermittent exercise results in an efflux of
between the trials. Although both SSGs elevated the internal CK into the bloodstream due to the mechanical disruption of
load markedly, the 4v4 was characterized by a higher mean HR myofiber, as evidenced by Z-line fragmentation, and second, as
and post-SSG lactate values, an observation also reported by part of a Ca2+-mediated inflammatory proteolysis that promotes
others.20 The higher exercise intensity recorded in 4-a-side secondary damage.24 The prolonged rise of CK in 4v4 may be
IJSPP Vol. 16, No. 9, 2021
 Small-Sided Games and Recovery 1277

Figure 5 — Time-course changes in MVIC measurements. MVIC indicates maximal voluntary isometric contraction. *Significant difference with
baseline in 4v4, P < .05. #Significant difference with baseline in 8v8, P < .05.

related to a reduced clearance rate through the lymphatic transport Concentric peak torque of both KE and KF decreased by ∼5%
system, as high-intensity exercise and passive recovery may delay to 15% for 24 to 48 hours post-SSGs. In contrast, the eccentric peak
CK removal.25 Once more, DOMS of KF was more pronounced torque of both KE and KF decreased for as long as 72 hours, with
and prolonged than that of KE.9,15 the latter demonstrating a greater reduction than the former. These
The SSGs induced an increase in neuromuscular fatigue of data suggest that KF eccentric strength may be a more sensitive
both KE and KF for as long as 3 hours post-SSGs, irrespective marker of postexercise performance recovery following SSG
of density. The neuromuscular fatigue increase of KE (∼5%–8%) training, as it was also shown for recovery following a soccer
was lower compared with that observed following a full soccer match.15 Interestingly, no differences were noted between DL and
match,26,27 but similar to that induced by speed-endurance train- NDL, which is in contrast with findings suggesting that the DL is
ing.9 In line with others,11 the SSGs resulted in a more pronounced more affected by high-intensity soccer activity.15 Similarly, CMJ
(∼7%–12%) reduction of KF’s neuromuscular fatigue compared was impaired for 24 to 48 hours, with peak reductions (∼5%–7%)
with KE, probably due to their predominant eccentric-type con- seen 24 hours post-SSGs in both trials. A 24-hour decline was also
traction during soccer-specific movements, such as shooting, observed previously, following a high-density SSG.10 Sprint per-
jumping, and tackling.19 These impairments may be attributed formance, on the other hand, exhibited a more persistent decline
to KF involvement in knee joint stability during intense decelera- (∼3%–5%) for as long as 72 hours, which coincides with observa-
tions that ultimately provokes fatigue development. Reports sug- tions for other intense soccer training protocols and match play.10,16
gest that central fatigue is responsible for performance decline a These results suggest that high-intensity training contents may
few hours postexercise (≤24 h).27 Molecular mechanisms respon- not be appropriate following demanding SSGs in a single training
sible for the evolution of central fatigue immediately postexercise session or even after 1 or 2 days. Central fatigue may, at least partly,
include, but are not limited to, protein catabolism pathways, explain this short-term performance deterioration due to changes in
disturbances in calcium homeostasis in the sarcoplasmic reticulum, neuron motor control, reflected in the reduced capacity of the
and down-regulated Na+-K+ ATPase activity.28 Evidence suggests central nervous system to activate skeletal muscle, which in turn,
that high-intensity, interval-based, speed-endurance protocols may compromises maximum power output during intense actions, such
also induce peripheral fatigue that attenuates muscle performance as sprints and jumps.27 However, it has been postulated that the
for a longer period of time (≥24 h), mainly due to the onset of days after (>24 h) soccer match play and/or intense training, fatigue
EIMD and inflammation within the muscle tissue.9 is predominantly of a peripheral type and interconnected with the
IJSPP Vol. 16, No. 9, 2021
Figure 6 — Time-course changes in isokinetic strength measurements. *Significant difference with baseline in 4v4, P < .05. #Significant difference
with baseline in 8v8, P < .05. †Significant difference between 8v8 and control within time point, P < .05.

1278 IJSPP Vol. 16, No. 9, 2021

 Small-Sided Games and Recovery 1279

Although it did not reach statistical significance, the average speed

in 8v8 was 6 m/min higher than that in 4v4 (123.9 vs 117.8 m/min).
To this point, it must be noted that the physiological responses
and recovery kinetics observed in this study are limited to the SSG
formats applied. Moreover, despite the fact that the participants’
fitness status and technical skills were similar to those reported
for elite soccer players, the results of the present study should be
interpreted with caution with respect to the concept of elite soccer
performance. It must also be stated that the results of this study should
be taken into consideration in the context of youth players, as they may
exhibit different time frames of recovery in response to SSGs.31 A
major limitation of this investigation is the application of different time
frames for the SSGs (6 × 4 min in 4v4 vs 3 × 8 min in 8v8). Both
internal load and external load may have been significantly influenced
by the number of rest periods. Nevertheless, the total exercise duration
was comparable between the SSGs, allowing for the allocation of our
results to the applied densities. Future studies need to elucidate the
impact of SGG formats and components on performance and fatigue
recovery using well-controlled experimental designs.

Practical Applications
Acute SSG training of high and low density is associated with
prolonged performance recovery (24–48 h and 72 h following high-
and low-density SSGs, respectively), inducing fatigue and EIMD in
male soccer players. SSGs promote acute neuromuscular fatigue for
up to 3 hours, especially in KF, suggesting that, if multiple training
Figure 7 — Time-course changes in CMJ and 30-m sprint time mea- sessions are scheduled on the same day, high-intensity contents
surements. CMJ indicates countermovement jump. *Significant difference should not be performed during this period. The SSGs of lower
with baseline in 4v4, P < .05. #Significant difference with baseline in 8v8, densities should be performed no more than 3 times a week if a coach
P < .05. †Significant difference between 4v4 and control within time point,
wants the players to be fully recovered at the beginning of a training
P < .05. +Significant difference between 8v8 and control within time point,
P < .05. session or if a coach actively wants to avoid an accumulation of
fatigue. On the other hand, SSGs of higher densities could be
performed more frequently and at least 2 days prior to a game.
evolution of EIMD, inflammation, and the skeletal muscle healing These observations should be taken into consideration when design-
process.27 Intense soccer-specific actions depend upon rapid force ing and applying SSG training protocols during an in-season
generation capacity, yet are sensitive to intramuscular metabolic microcycle, as optimal recovery is essential for performance adapta-
disturbances (ie, ion homeostasis, acidosis, interstitial K+), which tions and injury prevention. Alternating intense SSGs or reducing
can reduce sarcolemmal excitability.29 Other factors, such as their volume could be an alternative strategy to shorten recovery.
metabolic by-products (ie, lactate) and energy substrate concentra-
tion in the skeletal muscle (ie, glycogen), have also been proposed
to affect performance and recovery.28 The increased post-SSG BLa Conclusions
concentration observed in our study provides indirect evidence of Collectively, SSG training results in significant performance
elevated glycogen utilization. Although not measured here, muscle decline, neuromuscular fatigue development, and EIMD for as
glycogen depletion has been proposed as a major contributing long as 24 to 72 hours. Higher densities seem to induce a more
factor toward the development of fatigue during and following a intense and prolonged EIMD response, while lower densities are
soccer match.30 In 8v8, the degree of muscle glycogen depletion is associated with a more prolonged performance deterioration.
likely to be higher (and may reach the proposed critical level),
especially in the type II fibers, and cause slow recovery based on
full game studies.28 The onset of EIMD may have contributed to Acknowledgments
these prolonged recovery kinetics.
A second major finding is that density size does matter. The authors would like to thank all the players for their participation and
Although 4v4 was more intense (based on internal load monitor- commitment to the study. This study was supported by departmental funding.
ing), 8v8 induced more prolonged performance recovery kinetics
(72 h for strength and speed, and 48 h for CMJ), in contrast to 4v4,
where performance deterioration was more short-lived and reached References
baseline values at earlier time points (48 h for strength and speed,
and 24 h for CMJ). Strength, CMJ, and speed decline were more 1. Bangsbo J, Mohr M, Krustrup P. Physical and metabolic demands
prolonged following 8v8, further corroborating previous research of training and match-play in the elite football player. J Sports Sci.
linking larger pitch areas and lower player density with sustained 2006;24(7):665–674. PubMed ID: 16766496 doi:10.1080/026404
post-SSG fatigue due to increased distances covered with HSR.11 10500482529

IJSPP Vol. 16, No. 9, 2021

1280 Papanikolaou et al

2. Mohr M, Krustrup P, Bangsbo J. Match performance of high-standard Clin J Sport Med. 2008;18(5):423–431. PubMed ID: 18806550
soccer players with special reference to development of fatigue. doi:10.1097/JSM.0b013e3181818e0b
J Sports Sci. 2003;21(7):519–528. PubMed ID: 12848386 doi:10. 17. Panagoulis C, Chatzinikolaou A, Avloniti A, et al. In-season integrative
1080/0264041031000071182 neuromuscular strength training improves performance of early-
3. Rampinini E, Impellizzeri FM, Castagna C, et al. Factors influencing adolescent soccer athletes. J Strength Cond Res. 2020;34(2):516–
physiological responses to small-sided soccer games. J Sports Sci. 526. PubMed ID: 30431535 doi:10.1519/JSC.0000000000002938
2007;25(6):659–666. PubMed ID: 17454533 doi:10.1080/026404 18. Chatzinikolaou A, Fatouros IG, Gourgoulis V, et al. Time course of
10600811858 changes in performance and inflammatory responses after acute
4. Casamichana D, Román-Quintana JS, Castellano J, Calleja-González plyometric exercise. J Strength Cond Res. 2010;24(5):1389–1398.
J. Influence of the type of marking and the number of players on PubMed ID: 20386477 doi:10.1519/JSC.0b013e3181d1d318
physiological and physical demands during sided games in soccer. J 19. Mohr M, Draganidis D, Chatzinikolaou A, et al. Muscle damage,
Hum Kinet. 2015;47(1):259–268. PubMed ID: 26557209 doi:10. inflammatory, immune and performance responses to three football
1515/hukin-2015-0081 games in 1 week in competitive male players. Eur J Appl Physiol.
5. Aguiar MV, Botelho GM, Gonçalves BS, Sampaio JE. Physiological 2016;116(1):179–193. PubMed ID: 26377004 doi:10.1007/s00421-
responses and activity profiles of football small-sided games. J 015-3245-2
Strength Cond Res. 2013;27(5):1287–1294. PubMed ID: 22820211 20. Arslan E, Alemdaroglu U, Koklu Y, Hazir T, Muniroglu S, Karakoc
doi:10.1519/JSC.0b013e318267a35c B. Effects of passive and active rest on physiological responses and
6. Casamichana D, Castellano J. Time-motion, heart rate, perceptual and time motion characteristics in different small sided soccer games. J
motor behaviour demands in small-sides soccer games: effects Hum Kinet. 2017;60(1):123–132. PubMed ID: 29339992 doi:10.
of pitch size. J Sports Sci. 2010;28(14):1615–1623. PubMed ID: 1515/hukin-2017-0095
21077005 doi:10.1080/02640414.2010.521168 21. Reilly T, Ball D. The net physiological cost of dribbling a soccer ball.
7. Fatouros IG, Chatzinikolaou A, Douroudos II, et al. Time-course of Res Q Exerc Sport. 1984;55(3):267–271. doi:10.1080/02701367.
changes in oxidative stress and antioxidant status responses following 1984.10609363
a soccer game. J Strength Cond Res. 2010;24(12):3278–3286. 22. Ferguson BS, Rogatzki MJ, Goodwin ML, Kane DA, Rightmire Z,
PubMed ID: 19996787 doi:10.1519/JSC.0b013e3181b60444 Gladden LB. Lactate metabolism: historical context, prior misinterpre-
8. Fatouros IG, Jamurtas AZ. Insights into the molecular etiology of tations, and current understanding. Eur J Appl Physiol. 2018;118(4):
exercise-induced inflammation: opportunities for optimizing perfor- 691–728. PubMed ID: 29322250 doi:10.1007/s00421-017-3795-6
mance. J Inflamm Res. 2016;9:175–186. PubMed ID: 27799809 23. Castagna C, D’Ottavio S, Cappelli S, Araújo Póvoas SC. The effects
doi:10.2147/JIR.S114635 of long sprint ability–oriented small-sided games using different
9. Tzatzakis T, Papanikolaou K, Draganidis D, et al. Recovery kinetics ratios of players to pitch area on internal and external load in soccer
after speed-endurance training in male soccer players. Int J Sports players. Int J Sports Physiol Perform. 2019;14(9): 1265–1272.
Physiol Perform. 2020;15(3): 395–408. doi:10.1123/ijspp.2018- doi:10.1123/ijspp.2018-0645
0984 24. Fridén J, Lieber RL. Eccentric exercise-induced injuries to contractile
10. Sparkes W, Turner A, Weston M, Russell M, Johnston M, Kilduff L. and cytoskeletal muscle fibre components. Acta Physiol Scand.
Neuromuscular, biochemical, endocrine, and mood responses to 2001;171(3):321–326. PubMed ID: 11412144 doi:10.1046/j.1365-
small-sided games’ training in professional soccer. J Strength Cond 201x.2001.00834.x
Res. 2018;32(9):2569–2576. PubMed ID: 29351166 doi:10.1519/ 25. Brancaccio P, Maffulli N, Limongelli FM. Creatine kinase monitor-
JSC.0000000000002424 ing in sport medicine. Br Med Bull. 2007;81(1):209–230. PubMed
11. Madison G, Patterson SD, Read P, Howe L, Waldron M. Effects ID: 17569697 doi:10.1093/bmb/ldm014
of small-sided game variation on changes in hamstring strength. 26. Fransson D, Vigh-Larsen JF, Fatouros IG, Krustrup P, Mohr M.
J Strength Cond Res. 2019;33(3):839–845. PubMed ID: 30789860 Fatigue responses in various muscle groups in well-trained competitive
doi:10.1519/JSC.0000000000002955 male players after a simulated soccer game. J Hum Kinet. 2018;61(1):
12. Krustrup P, Mohr M, Nybo L, Jensen JM, Nielsen JJ, Bangsbo J. The 85–97. PubMed ID: 29599862 doi:10.1515/hukin-2017-0129
Yo-Yo IR2 test: physiological response, reliability, and application to 27. Thomas K, Dent J, Howatson G, Goodall S. Etiology and recovery of
elite soccer. Med Sci Sports Exerc. 2006;38(9):1666–1673. PubMed neuromuscular fatigue after simulated soccer match play. Med Sci
ID: 16960529 doi:10.1249/01.mss.0000227538.20799.08 Sports Exerc. 2017;49(5):955–964. PubMed ID: 28060035 doi:10.
13. Poulios A, Fatouros IG, Mohr M, et al. Post-game high protein intake 1249/MSS.0000000000001196
may improve recovery of football-specific performance during a 28. Mohr M, Krustrup P, Bangsbo J. Fatigue in soccer: a brief review.
congested game fixture: results from the PRO-FOOTBALL study. J Sports Sci. 2005;23(6):593–599. PubMed ID: 16195008 doi:10.
Nutrients. 2018;10(4):494. doi:10.3390/nu10040494 1080/02640410400021286
14. Papanikolaou K, Chatzinikolaou A, Pontidis T, et al. The Yo-yo 29. Hostrup M, Bangsbo J. Limitations in intense exercise performance of
intermittent endurance level 2 test: reliability of performance scores, athletes—effect of speed endurance training on ion handling and
physiological responses and overload characteristics in competitive fatigue development. J Physiol. 2017;595(9):2897–2913. PubMed
soccer, basketball and volleyball players. J Hum Kinet. 2019;67(1): ID: 27673449 doi:10.1113/JP273218
223–233. PubMed ID: 31523320 doi:10.2478/hukin-2018-0091 30. Bangsbo J, Iaia FM, Krustrup P. Metabolic response and fatigue in
15. Draganidis D, Chatzinikoloau A, Avloniti A, et al. Recovery kinetics soccer. Int J Sports Physiol Perform. 2007;2(2):111–127. PubMed
of knee flexor and extensor strength after a football match. PLoS One. ID: 19124899 doi:10.1123/ijspp.2.2.111
2015;10(6):e0128072. PubMed ID: 26043222 doi:10.1371/journal. 31. Deli CK, Fatouros IG, Paschalis P, et al. A Comparison of exercise-
pone.0128072 induced muscle damage following maximal eccentric contractions in
16. Ispirlidis I, Fatouros IG, Jamurtas AZ, et al. Time-course of changes men and boys. Ped Exerc Sci. 2017;29(3):316–325. doi:10.1123/pes.
in inflammatory and performance responses following a soccer game. 2016-0185

IJSPP Vol. 16, No. 9, 2021

Copyright of International Journal of Sports Physiology & Performance is the property of
Human Kinetics Publishers, Inc. and its content may not be copied or emailed to multiple
sites or posted to a listserv without the copyright holder's express written permission.
However, users may print, download, or email articles for individual use.