Thirty-four participants of all positions from a single male soccer team in the National First Division, with a playing experience of 4.4 ± 0.8 years, were participated. Prior to participation, participants were deemed fit to participate in the study, without orthopedic or other conditions which would preclude them from undertaking the plyometric and change-of-direction training. Subjects were randomly allocated to an experimental group (EG; n = 17, age: 14.6 ± 0.5 years, body mass: 60.5 ± 7.1 kg, height: 1.64 ± 0.08 m, body fat: 11.3 ± 1.4%) and a control group (CG; n = 17, age: 14.6 ± 0.4 years, body mass: 61.0 ± 3.9 kg, height: 1.67 ± 0.05 m, body fat: 11.8 ± 1.4%). There was no baseline difference between the groups in anthropometric characteristics (p ≥ 0.05). Participants were without any problems regarding health status and physical condition as they had completed a 6-week preseason training period of 5-6 sessions each week.

Both groups undertook five training sessions per week (∼90 min each session), plus a competitive match one time per week. Conditioning training was completed two times per week, with the first session developing aerobic fitness through small-sided games while the second session targeted anaerobic fitness with resistance (40-60% 1 RM). The controls maintained their normal training schedule throughout the 2 months of the intervention, whereas the EG replaced the technical-tactical part of their standard regimen by combined plyometric and short sprint training. Participants also engaged in the 60-min school physical education sessions weekly (Table 1).

The University Institutional Review Committee for the ethical human experimentation approved the interventional study (pre- and post-design, CG vs. EG), which was in line with current national laws and regulations (reference number: KS000002020; date of approval: November 10, 2020). Participants (and their guardians, in the case of minors) gave written informed consent. Two familiarization sessions preceded the testing period by 2 weeks, which was 2 months into the competitive season.

The training group undertook four sessions or workshops, biweekly (Figure 1). Workshops commenced with plyometric exercises (i.e., hurdle jumps, lateral hurdle jumps, bouncy strides, and single-leg hop jumps) and ended with a short sprint of 10-15 m (Table 2). The total number of ground contacts per session was gradually increased throughout the intervention (from 72 to 144 ground contacts per session), as well as the number of sets (from 3 to 6 sets) for each workshop. The number of contacts per set was maintained at 6 contacts. A 90-s rest interval was planned between each set of exercises to allow sufficient recovery time (Negra et al., 2016). Jump training protocols were supervised by a qualified instructor. The training protocol was based on previously published recommendations for training volume and intensity from the study by Bedoya et al. (2015). Furthermore, the sprint training protocol was based on previously published recommendations for sprint distances (Sáez de Villarreal et al., 2015; Kargarfard et al., 2020).

Tests were conducted at least 3 days after the last competitive match and 5-9 days after the previous training session. Testing took place on a tartan surface integrated into the weekly training schedule of players and a standardized warm-up preceded each test. Tests were conducted on three separate days in the following order:

The 10 and 30 m sprint performance, S90°, SBF, SJ, CMJA, FJT, Y balance test, and the RSSA have all been previously described in detail (Sporis et al., 2010; Hammami et al., 2020; Negra et al., 2020) so are not repeated in this study to avoid self-plagiarism and for brevity. Anthropometric characteristics (body mass and body fat percentage) were evaluated barefoot in the morning after an overnight fast, with bioelectrical impedance analysis (BIA) (BC-602, Tanita Co., Tokyo, Japan; Sitko et al., 2020). Each test was explained and demonstrated by one of the trained assessors before the athlete started practice trials. The same assessor supervised the same test station during pre- and posttest to ensure consistency. A 90-s rest interval was planned between each set of exercises to allow sufficient recovery time (Negra et al., 2016).

All statistical analysis was conducted using SPSS version 28.0 for Windows (IBM, Armonk, NY, United States). Normal distribution was tested using the Shapiro–Wilk test. The results showed that the majority of parameters showed normal distribution (76%). However, 24% (9/37) of the other parameters are not normally distributed (body fat: p = 0.023, SJ 1: p = 0.017, SJ 2: p = 0.006, CMJA 1: p = 0.002, FJT 1: p = 0.001, FJT 2: p = 0.039, repeated sprint ability (RSA) fatigue index 1: p = 0.016, anterior direction of left leg 2: p = 0.010, and posteromedial direction of left leg 2: p = 0.034). Homogeneity of variance was determined using the Levene’s test. Notably, 38% (6/16) of variance tests indicated a missing of homogeneity of variance (SJ: p = 0.012, sprint 10 m: p = 0.038, RSA best time: p = 0.032, RSA mean time: p = 0.042, RSA fatigue index: p < 0.001, and posteromedial direction of left leg: p = 0.027).

A sample size calculation (nQuery Advisor 4.0; Statistical Solutions, Saugus, MA, USA) provided a sample size with n = 7 in each group. The calculation is based on a mean difference of 0.4 s in 10 m sprint using a two-sided t-test with the α level of 0.05 and a pooled SD of 0.3 s, with a statistical power of 0.8 (Bortz, 1999).

Independent samples t-tests examined baseline intergroup differences. Paired sample t-tests tested for pre-to-post performance changes within each group. Training effects were assessed by mixed factorial 2-way (group × time) ANOVA with repeated measures. Subsequently, Tukey’s post-hoc tests examined pairwise differences. Alpha levels are reported as exact p-values as suggested by the American Statistical Association (Hurlbert et al., 2019). Effect sizes are reported as Cohen’s d and are interpreted as trivial (< 0.20), small (≥ 0.20–0.49), moderate (≥ 0.50–0.79), and large (≥ 0.80) (Cohen, 1988). Percentage changes were calculated as ((post-training value − pre-training value)/pre-training value) × 100. Reliability was assessed using intraclass correlation coefficients (ICCs) (Vincent, 1995) and the coefficients of variation (CVs) over consecutive pairs of intra-participant trials (Schabort et al., 1998). Vertical and horizontal jump parameters, sprint parameters, change-of-direction ability, and balance performance had an ICC > 0.80 and a CV < 5% (Table 3). Data are reported as mean ± SD.

There were interaction effects (group × time) for vertical and horizontal jump performances, with the EG outperforming the CG in terms of improvement from pre- to post-intervention (SJ: Δ19%, p < 0.01, d = 0.71; CMJA: Δ20%, p < 0.05, d = 0.86; FJT: Δ15%, p < 0.05, d = 0.74; Table 4).

There was a training effect (group × time) on sprint performance, with the EG improving more than CG (10 m sprint: Δ−11%, p < 0.01, d = 0.71; 30 m sprint: Δ−9%, p < 0.05, d = 0.64; Table 4).

There was a training effect for S90° and SBF with the EG improving more than CG (S90°: Δ−9%, p < 0.01, d = 0.76; SBF: Δ−8%, p < 0.05, d = 0.66; Table 4).

For the ability to perform RSS, the RSSA test also showed group × time interactions for most of its parameters (RSSA best and RSSA mean), with time decreases of Δ−8% (p < 0.01, d = 0.72) and Δ−8% (p < 0.01, d = 0.71; Table 5).

There was a training effect on Y balance test performance with the EG improving more for balance measured on both legs (anterior: Δ11%, p = 0.05, d = 0.50; posteromedial: Δ10%, p < 0.05, d = 0.59; posterolateral: Δ16%, p < 0.05, d = 0.54 for right support leg). For the left support leg, we found similar results (anterior: Δ11%, p < 0.05, d = 0.53; posteromedial: Δ10%, p < 0.05, d = 0.66; posterolateral: Δ15%, p < 0.05, d = 0.54; Table 5).

The ability to jump is a distinctive feature of success in adult and young soccer players (Söhnlein et al., 2014; Ramirez-Campillo et al., 2018b). Therefore, developing the ability of a soccer player to generate vertical power quickly is likely to be advantageous for the competition (Oliver et al., 2021). The results of this study suggest the intervention effects on vertical and horizontal jump performances (Table 4). Our findings are confirmed by Hammami et al. (2020) who observed improvements in vertical and horizontal jump performances following combined plyometric and short sprint training in male U17 soccer players. In addition, Sáez de Villarreal et al. (2015) reported significant increases in vertical jump performance following combined plyometric-speed training and technical drills work, in male youth soccer players (U15). Jump performance augmentation could be the result of neuromuscular adaptations such as greater neural drive of agonist muscles, changes in mechanical stiffness of tendons, size changes and/or the architecture of muscles, and changes in the mechanics of single fibers (Maffiuletti et al., 2002; de Villarreal et al., 2009; Thomas et al., 2009). Other potential mechanisms include (i) changes in muscle activation strategies (or intermuscular coordination) during vertical jumps, especially during the preparatory phase jump; and (ii) changes in the stretch reflex excitability (Bishop and Spencer, 2004; de Villarreal et al., 2009).

Stølen et al. (2005) reported that 96% of sprint bouts during soccer match play are shorter than 30 m, with 49% being shorter than 10 m. The present results showed significant intervention effects in the EG compared to the CG for 10 and 30 m sprint performance. This corroborates findings reported by Sáez de Villarreal et al. (2015), who examined the efficacy of combined plyometric-speed training and technical drills work in youth male soccer players (U15) and noted improved sprint performance over 5 and 10 m. Moreover, Kargarfard et al. (2020) investigated the effects of combined plyometric and short sprint exercises in youth male under 19 (U19) soccer players. They reported improved sprint performance over 30 m. The improved sprint performance observed in this study may be useful for soccer players as it would correspond to an advantage in duels sprints, which can allow players to reach the ball before the opponent (Stølen et al., 2005).

During soccer match play, players perform many accelerations and sprints at maximum speed (Haugen et al., 2014), with and without changes in direction (Bloomfield et al., 2007; Barnes et al., 2014). The present investigation demonstrated the intervention effects on the change-of-direction ability. There are few studies evaluating the effectiveness of combined plyometric and short sprint training program on change-of-direction ability, in soccer players (Sáez de Villarreal et al., 2015; Hammami et al., 2020; Kargarfard et al., 2020). This study corroborates the results obtained by Sáez de Villarreal et al. (2015), who studied the effects of combined plyometric-speed training and technical drills work in youth male soccer players (U15). They noted improvements in change-of-direction ability. Indeed, Kargarfard et al. (2020) observed significant increases in change-of-direction ability, following combined plyometric and short sprint training, in youth male soccer players (U19). Improvements in the change-of-direction ability necessitate rapid development of strength, eccentric force production in the lower limbs, and a rapid change from eccentric muscle to concentric contraction in the knee extensors. Plyometric training can enhance these factors (Miller et al., 2006; Sheppard and Young, 2006).

The present study demonstrated intervention effects on all RSSA parameters except the fatigue index. Our results are in line with those of Hammami et al. (2020) who reported improvement in all repeated change-of-direction ability parameters except the fatigue index in male U17 soccer players following combined plyometric and short sprint training. In contrast, our results contradict those reported by Kargarfard et al. (2020) who reported no improvements in RSA following combined plyometric and short sprint training in male U19 soccer players. Improvements in RSA may correspond to increased running efficiency (Balsalobre-Fernández et al., 2016) or increased tendon stiffness (Markovic and Mikulic, 2010), which allows a faster transfer of force from the contracting muscles, reducing reaction times, and improving change-of-direction ability (Markovic and Mikulic, 2010; Balsalobre-Fernández et al., 2016). Although maximum sprint speed is one of the performance composites of anaerobic performance in repeated sprints (Girard et al., 2011), other factors determining inter-sprint recovery and maintenance of performance are required to enhance RSA (da Silva et al., 2010; Girard et al., 2011).

Chimera et al. (2004) reported that incorporating plyometric exercises into athlete training reduces the risk of injury by improving the functional stability of the leg joints. Our study seems to be the first to examine the effects of the combined plyometric and short sprint training program on balance performance, in male U15 soccer players. Our results of improved balance disagree with those obtained by Hammami et al. (2020) who examined the efficacy of combined plyometric and short sprints, in male U17 soccer players and reported no improvements in static and dynamic balance performance. Differences in the combined plyometric and short sprint training program (duration, intensity, frequency, and type of exercise) and in methodology (age group and competitive level of the study population; duration of intervention) could contribute to these discrepancies. Improvements in balance reduced the risk of lower limb injury in soccer players (Lloyd, 2001). In the context of increased athletic performance (Zech et al., 2010), data from the present investigation emphasize the importance of plyometric training as an effective injury-reduction strategy in young athletes. A systematic review (Hübscher et al., 2010) showed evidence of the effectiveness of balance training in reducing the incidence of certain types of sports injuries among adolescent and young athletes during pivoting sports. Recently, Rivera et al. (2017) reported that balance training programs were effective in reducing the incidence rates of ankle sprains, including those with and those without a history of ankle sprains, in soccer players. In addition, the increased balance may be related to an improvement in lower extremity muscle co-contraction (Lloyd, 2001) or the changes in proprioception and neuromuscular control (Michailidis et al., 2019).