Effects of combined balance and plyometric training on athletic performance: a systematic review

Abstract

Background

Plyometric training is an effective method for improving explosive power, while balance training plays a pivotal role in enhancing agility and dynamic stability. Recently, combined balance and plyometric training (hereafter referred to as combined training) has gained increasing attention. This approach integrates plyometric exercises with separate balance training components. However, whether combined training offers superior benefits over single-mode training remains unclear.

Objective

This systematic review aimed to evaluate the effects of combined training on athletic performance across different sports, assessing balance, agility, speed, and power. The review specifically examined interventions that included separate balance and plyometric components, and excluded plyometric training performed on unstable surfaces without a standalone balance component.

Methods

Following PRISMA guidelines, a systematic search of Web of Science, PubMed, EBSCOhost, and Embase was performed from inception to March 15, 2026. Randomized controlled trials examining combined plyometric and balance training in athletes were included. Duplicates were removed via EndNote. Two independent reviewers screened records in two stages (title/abstract, then full text) and extracted data. Methodological quality was assessed using the PEDro scale. Disagreements were resolved by consultation with a third reviewer. The review followed the PROSPERO-registered protocol (CRD420251273522), with no deviations.

Results

Fourteen studies involving 522 athletes were included, covering badminton, basketball, handball, soccer, dance, and taekwondo. Participants were predominantly adolescents aged 11–24 years. Intervention duration ranged from 6 to 12 weeks, with a frequency of 2–3 sessions per week and session lengths of 5–60 min. The available evidence suggests that combined training may enhance balance and agility. However, findings on speed and explosive power improvements remain inconsistent.

Conclusions

Combined training may offer benefits, particularly for balance and agility (moderate-certainty evidence). These findings provide preliminary practical implications for coaches and may inform training program design. However, evidence for speed and explosive power is of low certainty, and further high-quality research is needed to verify these effects.

Trial registration

Identifier CRD420251273522.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s13102-026-01889-1.

Introduction

Modern competitive sports place increasing demands on athletes’ comprehensive motor abilities—including explosive power, agility, speed, and balance—all of which are closely linked to neuromuscular function [1]. An athlete’s competitive performance depends on five components: physical fitness, technical skills, tactics, psychology, and intelligence [2]. Physical fitness is foundational, as execution of any technical or tactical skill requires adequate physical reserves [3]. Sufficient physical fitness can significantly influence, and may even determine athletic performance during competitions [4].

Plyometric training originated in track and field training [5]. As a high-power exercise mode, it involves eccentric muscle contraction before concentric contraction, activating the stretch-shortening cycle to generate forceful concentric contractions, making it an established method for improving explosive power [6]. A large body of research has demonstrated that plyometric training can significantly enhance athletes’ upper and lower limb explosive power [7], such as the jump height of basketball players and the starting acceleration of sprinters [8–10]. More recent evidence further supports the beneficial effects of plyometric training on explosive athletic performance in youth populations [11]. Balance training, on the other hand, focuses on optimizing neuromuscular control ability [12]. Recent studies have also highlighted the role of balance training in inducing postural control and neuromuscular adaptations that are transferable to sport-specific contexts [13].

In real competition, athletes perform complex, sequential movements rather than isolated actions. Therefore, training interventions that concurrently target multiple physical qualities may better reflect the demands of sport-specific performance [14]. Combined balance and plyometric training (hereafter referred to as combined training) is an integrated training model that merges independent plyometric training components with separate balance training components [15]. The underlying rationale is that this combination may optimize stretch-shortening cycle efficiency while concurrently enhancing sensorimotor control, thereby potentially improving the coordination and economy of sport-specific movement patterns [16]. Recent comparative studies have further emphasized the need to distinguish between different plyometric training protocols when evaluating their effects on sport-specific performance outcomes, as training surface and movement complexity influence neuromuscular adaptations [17]. Although studies on plyometric training or balance training have demonstrated their training benefits for athletes [18], research evidence on the combined balance and plyometric training model remains relatively scarce and inconsistent. The relative effectiveness of combined training in improving performance across different sports is still unclear, with a lack of comparative studies on training benefits between combined balance and plyometric training and traditional single training modes or plyometric training.

To address this gap, this study systematically reviews the effects of combined balance and plyometric training on various sports-related indicators of athletes, compares traditional single training modes with combined balance and plyometric training, and provides valuable references for the development of training programs for athletes in different sports.

Methods

This systematic review was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement [19]. The review was registered in the International Prospective Register of Systematic Reviews (PROSPERO registration number: CRD420251273522).

Throughout the entire review process, all methodological procedures—including eligibility criteria, search strategy, outcome measures, and data synthesis plan—were strictly implemented in accordance with the pre-registered protocol. No modifications or deviations were made to the protocol after registration. Any post-hoc decisions, if applicable, are explicitly noted and justified in the supplementary materials.”

Search strategy

This study conducted systematic searches across English-language databases, including Web of Science, PubMed, EBSCOhost, and Embase, covering the period from database establishment to March 15, 2026. The search was conducted by one researcher (LYF) and independently verified by a second researcher (DCW), who cross-checked the search results for completeness and accuracy. Any disagreements regarding the search strategy were resolved through consensus discussion.

To ensure a comprehensive search, the Boolean operators AND/ OR were used. The specific search query is illustrated using PubMed as an example: (“jump training“[Title/Abstract] OR “plyometric training“[Title/Abstract] OR “plyometrics“[Title/Abstract]) AND (“balance training“[Title/Abstract] OR “postural balance“[Title/Abstract] OR “proprioception“[Title/Abstract] OR “postural control“[Title/Abstract]). The search strategy prioritized intervention-related terms over population- or outcome-specific terms. This approach was deliberately adopted to maximize sensitivity and avoid missing eligible studies in which participant populations or specific performance outcomes may not have been explicitly mentioned in the title or abstract. Given that the primary objective of this review was to identify interventions meeting the strict operational definition of combined balance and plyometric training, broadening the search around intervention terms was considered the most appropriate approach to ensure comprehensive capture of eligible studies. However, this strategy may have introduced some degree of imprecision by retrieving studies unrelated to athletic populations. To mitigate this limitation, population relevance and outcome eligibility were determined during the title/abstract and full-text screening stages based on the PICOS criteria, rather than through search-term restriction.

To supplement the electronic database searches, we conducted a manual search of Google Scholar and performed forward and backward citation tracking of all included studies and relevant review articles. The complete search strategies, including all filters, limits, and date restrictions, are provided in Supplementary Table 1.

Eligibility criteria

In accordance with the PICOS principles, the following studies were eligible: (1) The participants were athletes, including adolescents and adults of various skill levels (2). Interventions involving plyometric and balance training with control and experimental groups, including combined balance and plyometric training. Combined balance and plyometric training was operationally defined as interventions that incorporated both balance-specific exercises (e.g., single-leg stance, unstable surface standing, or dynamic postural tasks performed on stable ground) and plyometric exercises (e.g., jumps, hops, or bounding on stable surfaces) as distinct training components within the same program. Interventions in which plyometric exercises were performed exclusively on unstable surfaces (e.g., BOSU balls, wobble boards, or sand) were excluded (3). Evaluation of combined balance and plyometric training’s impact on athletes’ performance metrics, including at least one motor indicator (4). The study design was a parallel-group randomized controlled trial (RCT) with two or more groups (i.e., at least one experimental group and one control group). Single-group pre-post trials, quasi-experimental designs, and non-randomized studies were explicitly excluded to minimize the risk of bias and to ensure methodological rigor (5). No restrictions on study participants, locations, or timeframes. These criteria were applied to determine eligibility. Exclusion criteria included: (1) Interventions not involving combined balance and plyometric training (2). Publications, conference abstracts, or brief reports in languages other than English (3). Studies without designed intervention measures. Refer to Table 1.

Table 1.

Inclusion criteria

Items	Detailed inclusion criteria
Population	Athletes, with no restrictions on their sex, age, or competition level
Intervention	Combined balance and plyometric training
Comparison	Randomized controlled trial (RCT) with two or more groups
Outcome	Physical fitness (speed, power, agility, balance, coordination, flexibility

Screening and data extraction

The literature screening process is as follows: (1) A researcher conducted the literature search (2). The retrieved literature was systematically imported into EndNote reference management software for duplicate removal (3). During the initial screening phase, researchers independently assessed the titles and abstracts of the literature to distinguish between relevant and irrelevant studies. Studies that did not meet the predefined inclusion and exclusion criteria were excluded, retaining only full-text documents requiring comprehensive review. In the event of disagreements or conflicting results during the process, a third researcher was consulted to reach a consensus and ultimately determine the relevant literature to be included.

We involved gathering relevant data and content from standardized studies, including: (1) authors and publication years; (2) participants’ basic information (sample size, age, gender, height, weight, BMI, training duration, athlete status); (3) experimental interventions (content, training weeks, frequency); (4) research outcomes.

Quality assessment and risk of bias

This study used the PEDro scale to assess the methodological quality of the included randomized controlled trials. The PEDro scale has been demonstrated to be an effective tool for evaluating the methodological quality of intervention studies, with good validity and reliability [20]. The PEDro scale is designed to assess four fundamental methodological aspects of research, including randomization, blinding, comparison of study populations, and data analysis. Two trained, independent researchers evaluated the 10 items of the PEDro scale using a response scoring scale of “yes” (1 point) or “no” (0 points). Since the inclusion criteria are related to external validity, they were not included in the total score [21]. The PEDro scale total score ranges from 0 to 10 points; a higher score indicates better methodological quality. Scores between 6 and 10 are considered high-quality literature, scores between 4 and 5 are considered moderate quality, and scores below 3 are considered poor. If the two researchers disagree on the assessment of methodological quality during the evaluation process, a third rater makes the final decision. The overall quality of evidence for each outcome domain (balance, agility, speed, and power) was assessed using the GRADE (Grading of Recommendations Assessment, Development and Evaluation) framework. Evidence quality was rated as high, moderate, low, or very low based on five domains: risk of bias, inconsistency, indirectness, imprecision, and publication bias. Two independent reviewers performed the GRADE assessment, with disagreements resolved through discussion with a third reviewer.

Data synthesis and analysis

A meta-analysis was considered but ultimately deemed inappropriate for the following reasons. First, substantial clinical and methodological heterogeneity existed across the 14 included studies in terms of: (a) participant characteristics (athletes from 7 different sports—badminton, basketball, handball, soccer, taekwondo, dance, and recreational athletes with ankle instability—with varying age ranges 11–24 years, sex distributions, and baseline training status); (b) training protocols (intervention durations ranging from 6 to 12 weeks, frequencies of 2–3 sessions/week, and session lengths varying from 5 to 60 min); and (c) outcome measures (balance was assessed using 10 different tests including YBT, DPSI, COP, SBT, LSI, and TTS; agility was measured using 7 different tests including T-test, Illinois Agility Test, 5-0-5 test, H-test, and SEMO; speed and power were assessed using sprint distances ranging from 5 to 30 m and jump tests including SJ, CMJ, DJ, SLJ, and RSI).

Second, statistical heterogeneity was anticipated to be high (I² > 75%) based on the observed variability in intervention effects and the diversity of measurement instruments, which would render a pooled effect size unreliable and potentially misleading. Third, the diversity of test instruments across studies—measuring the same construct (e.g., balance) using fundamentally different tools (e.g., YBT as a reach-distance measure vs. COP as a center-of-pressure sway measure)—precluded meaningful effect-size conversion or standardization, as these tests capture distinct aspects of balance control (dynamic functional balance vs. postural sway) and are not directly comparable. Fourth, several studies (n = 5) reported outcome data only in graphical format or as pre-post changes without providing means and standard deviations suitable for meta-analytic pooling, and attempts to contact corresponding authors for raw data were unsuccessful for three of these studies. We considered conducting subgroup analyses (e.g., by sport type, age group, or training duration) or using a random-effects meta-analysis to account for between-study heterogeneity. However, the limited number of studies within each subgroup (typically 1–3 studies per sport) precluded meaningful subgroup analyses, and the extensive heterogeneity in both study design and outcome measurement made the results of a random-effects model difficult to interpret.

Therefore, a meta-analysis could not be conducted, and instead, a qualitative analysis of the articles was performed [22]. Qualitative research is an evaluation method that has also been frequently used in other systematic reviews in the past [23–25]. It involves synthesizing and summarizing data from the included literature to provide a qualitative description, thereby yielding specific research findings [26]. To enhance interpretability, we calculated Cohen’s d effect sizes for each outcome where data were available. Effect sizes were interpreted as small (≥ 0.20), moderate (≥ 0.50), or large (≥ 0.80).

Results

Study selection

A total of 520 potential articles were identified. In addition, a manual search of Google Scholar identified one study that met the inclusion criteria. After excluding duplicates, the remaining 380 articles were screened, and 217 studies were excluded. We reviewed the full texts of the remaining 163 studies; based on the eligibility criteria, 149 studies were excluded. Ultimately, 14 studies met all eligibility criteria. All included studies were randomized controlled trials with two or more groups; no single-group trials were included. The study selection process is illustrated in Fig. 1.

Fig. 1.

PRISMA flowchart for study selection process

Study quality assessment

The average PEDro scale for the included studies was 6.07 (range: 5 to 9), indicating moderate to high methodological quality overall, with scores ranging from 5 to 9. Three studies were of moderate quality, all scoring 5, whilst eleven studies were of high quality, scoring between 6 and 9. Table 2 shows the relevance score for each study.

Table 2.

Summary of methodological quality assessment scores

Study	Item1	Item 2	Item 3	Item 4	Item 5	Item 6	Item 7	Item 8	Item 9	Item 10	Item 11	TS	QS
Bouteraa et al.(2020) [14]	1	1	0	1	0	0	0	1	0	1	1	5	M
Lu. et al.(2022) [27]	1	1	0	1	0	0	0	1	0	1	1	5	M
Zhou, et al. (2022) [28]	1	1	0	1	0	0	1	1	0	1	1	6	H
Aloui, et al. (2025) [29]	1	1	0	1	0	0	1	1	0	1	1	6	H
Shen, et al. (2024) [30]	1	1	0	1	0	0	0	1	0	1	1	5	M
Huang, et al. (2014) [31]	1	1	0	1	0	0	1	1	0	1	1	6	H
Guo, et al. (2025) [32]	1	1	0	1	0	0	1	1	0	1	1	6	H
Huang, et al. (2021) [33]	1	1	1	1	0	0	1	1	0	1	1	7	H
Zhang, et al. (2024) [34]	1	1	0	1	1	0	1	1	0	1	1	7	H
Guo, et al. (2021) [35]	1	1	0	1	0	0	1	1	0	1	1	6	H
Al Attar, et al. (2022) [36]	1	1	1	1	1	0	1	1	1	1	1	9	H
Yan, et al. (2025) [37]	1	1	0	1	0	0	1	1	0	1	1	6	H
Makhlouf, et al. (2018) [38]	1	1	0	1	0	0	1	1	0	1	1	6	H
M T, et al. (2026) [39]	1	1	0	1	0	0	1	1	0	1	1	6	H

Item 1: Eligibility criteria; Item 2: Random allocation; Item 3: allocation concealment; Item 4: baseline comparability; Item 5: blind participants; Item 6: blind therapist; Item 7: blind assessor; Item 8: follow-up; Item 9: intention to treat analysis; Item 10: between group comparisons; Item 11: point measure and variability

P poor quality, M moderate quality, H high quality, TS total score, SQ study quality

Participant characteristics

Table 3 summarizes the participant characteristics of the fourteen studies that met the inclusion criteria. The participant characteristics included in the study are as follows:

Table 3.

Data extraction from selected articles

Study	Athlete	Population	Interventions	Main outcomes
Bouteraa et al. (2020) [14]	Basketball (F)	N = 26	EG; combined BP training CG; RT	Power: CMJ↔ SJ↔DJ↑#, Speed: 5-10-20 m sprints↔ Agility: MICODT↑# Balance: SBT↔YBT↑#
Lu et al. (2022) [27]	Badminton (M)	N = 16	EG: combined BP training CG: P training	Balance: F-DPSI↑#, L-DPSI↑#, F-COPAP↑#, F-COPML↑#, F-COPPL↑#, L-COPPL↑#, L-COPAP↔, L-COPML↔ Agility; T-test↑#, H-test↑#
Zhou et al. (2022) [28]	Badminton (M)	N = 16	EG: combined BP training CG: P training	Balance: LSIT↑#, LSIS↑#, NAP↑# LSIO, LSIC, DAP↔ NAPLSIAPDML、↔ NMLLSIML↔
Aloui, et al.(2025) [29]	Handball (M)	N = 30	EG: combined BP training CG: RT	Balance: SBT↑# YBT↑# Power: SJ, CMJ, SLJ↑# Speed: 5,20Sprint↑# Agility: T-test↑# Illinois-MT↑#
Shen, (2024) [30]	Taekwondo (F)	N = 30	EG: combined BP training CG: RT	Balance: F-DPSI↑# L-DPSI↑# COP↑# Agility: 5-0-5test↑# TAST↑#
Huang, et al. (2014) [31]	FAI (F/M)	N = 30	EG1:P training EG2: combined BP training CG: RT	Balance: SPS, MAI, TTS, COM, COP↓#
Guo and Xu (2025) [32]	Badminton(M)	N = 48	EG: combined BP training CG: P training	Balance: DPSI-DF↑ DPSI-NF↑ DPSI-NL↑ LSI-3↑# LSI-6↑# Agility: SEMO↔, T-Test↑# H-test↑#
Huang, et al. (2021) [33]	Recreational athletes with FAI(F/M)	N = 30	EG1:P training EG2: combined BP training CG: RT	AJPS, EMG↑ Balance: BAT↓#
Zhang et al. (2024) [34]	Badminton(M)	N = 16	EG: combined BP training CG: P training	Balance: D-COPAP↑# N-COPAP↑# N-COPML↑# D-TTS↔ F- TTS↔
Guo et al. (2021) [35]	Badminton(M)	N = 16	EG: combined BT training CG: P training	Balance: YBT↑# Agility: SEMO↔ 505test↑# Power: RSI↑#
Al, Attar, et al. (2022) [36]	Female athletes	N = 179	EG: combined BT training CG: RT PG: P training BG: B training	Balance: LOS↑#
Yan, et al. (2025) [37]	Dancers(F)	N = 30	EG: combined BP training CG: P training	Balance: DPSI↑# LSI-3 C↑# LSI-6↑# COP↑#
Makhlouf, et al. (2025) [38]	Soccer (M)	N = 57	EG1: combined BP training EG2: AP training CG: RT	Balance: SS↑# YBT↑# Agility: MVIC↑# ICODT↑# 4,9 m run↑# Speed: 10–30 m sprint↔ Power: CMJ↑#
Muehlbauer, et al. (2026) [39]	Soccer (M)	N = 60	EG1: combined BP training EG2: Single-mode balance training CG: RT	Balance: YBT–LQ↑# SLDL↔ LSI↑#

Abbreviations for outcome measures: BAT balance adjusting time, CMJ countermovement jump, COP center of pressure (with subscripts: AP anteroposterior, ML mediolateral, PL path length), DJ drop jump, DPSI dynamic posture stability index (with prefixes: F- forward jump, L- lateral jump, DF dominant forward, DL dominant lateral, NF non-dominant forward, NL non-dominant lateral), EMG electromyography, H-test hexagon test, ICODT Illinois change of direction test, LOS limits of stability, LSI, limb symmetry index (with subscripts: O single hop, T triple hop, C cross-over hop, S 6-m timed hop), MICODT modified Illinois change of direction test, MVIC maximum voluntary isometric contraction, RSI reactive strength index, SBT stork balance test, SEMO Southeast Missouri test, SJ squat jump, SLDL single leg drop landing, SLJ standing long jump, TAST taekwondo specific agility test, THT triple-hop-test, TTS time to stabilization, YBT Y-balance test

Other abbreviations: BP balance and plyometric, CG control group, CT combined training, EG experimental group, F female, M male, N number, PT plyometric training, RT regular training, B balance training

↑, significant improvement; ↔, no significant difference; ↓, significant decrease; #, significant difference between groups post-intervention

1. Athlete classification : Among fourteen included articles, one did not specify the athlete type and only described athletes with ankle instability [31], and a study focusing solely on female athletes [36], while the remaining twelve articles specified the athlete type, five studies involving badminton players [27, 28, 32, 34, 35], two studies involving footballers [38, 39], one study involving handball players [29], one study involved basketball players [14], one study involving taekwondo athletes [30], one study involved dancer [37] and one study concerning recreational athletes with ankle instability [33];

2. Sample size: Fourteen studies collectively included 522 participating athletes, ranging from 16 to 179;

3. Gender: Fourteen studies involved athletes, with eight studies focusing on males [27–29, 32, 34, 35, 38, 39] and four studies on females [14, 30, 36, 37], while the remaining two studies included both males and females [31, 33].

4. Age: All studies reported the athletes ‘age. Analysis of age data from fourteen studies revealed that athletes ranged from 11 [38] to 24 [31, 33] years old, predominantly adolescents.

5. Body Mass Index (BMI). While all studies documented athletes’ height and weight, three reported BMI value [14, 31, 33], and one study measured body fat percentage [29]. To ensure consistency in the literature, the BMI calculation formula used in the studies was: BMI = weight (kg) / height² (m).

6. Training background: Eight studies provided training background information [27–30, 32, 34, 35, 37], while the remaining did not [14, 31, 33, 36, 38, 39].

Intervention characteristics

In the fourteen studies, the experimental groups met the inclusion criteria and followed different training, the key feature being the use of combined balance and plyometric training. The characteristics of the intervention are as follows:

1. Training duration and frequency: In fourteen studies, four studies were conducted 3 times a week for 6 weeks [30, 31, 33, 35], one study involved twice-weekly sessions over 6 weeks [36], two studies were twice a week for 8 weeks [14, 29], three studies involved three sessions per week over 8 weeks [27, 28, 32], one study involved twice-weekly sessions over a period of 9 weeks [39], while the longest intervention spanned 12 weeks with 3 sessions weekly [37].

2. Duration of each training session: Nine studies reported duration of 60 min per session [27–30, 32–35, 37], one study reported a training duration of 5–10 min [36], one study reported a training duration of 30 min [39], one study reported a training duration of 45 min [14], three studies did not report the duration of training.

3. Training intensity, volume, progression, rest intervals, and supervision: Four studies specified intensity (e.g., maximal effort, RPE); five described progression qualitatively (e.g., increased height or load); six reported volume (80–240 contacts/session); four provided rest intervals (30–60 s between sets, 60–120 s between exercises); and seven mentioned supervision, though none reported supervisor-to-participant ratios.

4. Concurrent regular training: Eleven of the 14 included studies explicitly reported that both experimental and control groups continued their regular sport-specific training during the intervention, with training volume matched between groups. The remaining three studies did not provide sufficient detail regarding concurrent training, which may introduce confounding.

Quality of evidence GRADE assessment

The GRADE assessment results for each outcome domain are summarized in Table 4. For balance (14 studies) and agility (7 studies), the quality of evidence was rated as moderate (⊕⊕⊕○), downgraded one level due to imprecision (most studies had small sample sizes, with 10 of 14 studies having fewer than 40 participants per group). For speed (3 studies) and power (4 studies), the quality of evidence was rated as low (⊕⊕○○), downgraded for both inconsistency (heterogeneous findings across studies) and imprecision (limited number of studies and small samples). These ratings indicate that further research is likely to impact the confidence in the estimates for balance and agility, and is very likely to impact confidence for speed and power. Detailed GRADE assessments for each outcome are presented in Table 4.

Table 4.

GRADE assessment for each outcome domain

Outcome	Number	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	Overall quality
Balance	14	Some concerns1	Not serious¹	Not serious	Serious²	Undetected	⊕⊕⊕○ Moderate
Agility	7	Some concerns1	Not serious¹	Not serious	Serious²	Undetected	⊕⊕⊕○ Moderate
Speed	3	Some concerns1	Serious³	Not serious	Serious⁴	Undetected	⊕⊕○○ Low
Power	4	Some concerns1	Not serious¹	Not serious	Serious⁴	Undetected	⊕⊕○○ Low

⊕⊕⊕○, moderate quality; ⊕⊕○○, low quality

¹ Not serious — Most included studies were of moderate-to-high methodological quality (PEDro scores ≥ 5; 11 of 14 studies scored ≥ 6)

² Serious — Most studies had small sample sizes (10 of 14 studies had < 40 participants per group), reducing precision of estimates

³ Serious — Heterogeneous findings across studies (two reported significant improvements, one reported no significant differences)

⁴ Serious — Limited number of studies (speed: n = 3; power: n = 4) and small sample sizes, resulting in wide confidence intervals

¹Some concerns—All studies lacked blinding of participants and therapists, which may introduce performance bias, though the objective nature of most outcome measures (e.g., jump height, sprint time) likely mitigates detection bias

Subgroup analysis considerations

The 14 included studies exhibited considerable heterogeneity in participant characteristics, including age (range: 11–24 years), sex (eight male-only, four female-only, two mixed), sport type (badminton, basketball, handball, soccer, taekwondo, dance, and recreational athletes with ankle instability), and intervention duration (6–12 weeks). However, a formal subgroup analysis was not performed in this review. This decision was primarily constrained by the limited number of studies within each subgroup category and the substantial variation in outcome measures across studies, which precluded meaningful statistical comparisons. For instance, only one to two studies were available for most sport types, and the balance and agility outcomes were assessed using different testing protocols (e.g., YBT, DPSI, SBT, COP). Consequently, the current findings represent overall trends rather than population-specific effects. Future research with larger, more homogeneous samples and standardized outcome measures is warranted to enable robust subgroup analyses that can identify which athlete populations may derive the greatest benefits from combined balance and plyometric training.

A detailed summary of the reporting of training intensity, progression, volume, rest intervals, and supervision across the included studies is presented in Table 5.

Table 5.

Summary of training intensity, volume, progression, rest intervals, and supervision

Study	Intensity	Progression	Volume	Rest Intervals	Supervision
Bouteraa et al. (2020) [14]	✗	✗	✓	✓	✗
Lu et al. (2022) [27]	✗	✓	✗	✗	✓
Zhou et al. (2022) [28]	✗	✗	✗	✗	✗
Aloui, et al.(2025) [29]	✗	✓	✓	✗	✗
Shen, (2024) [30]	✗	✗	✗	✗	✗
Huang, et al. (2014) [31]	✓	✓	✗	✗	✓
Guo and Xu (2025) [32]	✗	✗	✗	✗	✓
Huang, et al. (2021) [33]	✓	✓	✗	✗	✓
Zhang et al. (2024) [34]	✗	✗	✗	✗	✓
Guo et al. (2021) [35]	✓	✓	✓	✓	✓
Al, Attar, et al. (2022) [36]	✓	✓	✗	✓	✓
Yan, et al. (2025) [37]	✗	✓	✓	✓	✓
Makhlouf, et al. (2025) [38]	✓	✓	✓	✓	✓
Muehlbauer, et al. (2026) [39]	✓	✓	✓	✗	✓

✓, parameter explicitly reported in the study; ✗, parameter not reported

Outcome

Among the fourteen included studies, fourteen assessed balance, seven examined agility, three assessed speed, and four assessed power. Compared with plyometric training or resistance training alone, combined balance and plyometric training may demonstrate more advantages in improving athletes’ balance and agility ability, but research evidence on speed and explosive power indicates insufficient consistency, and further research is needed to clarify these effects.

The effect of combined balance and plyometric training on balance

In this systematic review, which included fourteen studies, balance was the core outcome. Fourteen studies examined the balance abilities of athletes, including badminton [27, 28, 32, 34, 35], football [38, 39], basketball [14], handball [29], taekwondo [30],dancer [37],FAI [31, 33],female athletes [36]. Among these, eight studies indicated that combined balance and plyometric training has a significant positive effect on the balance ability of athletes across different sports, with significant differences observed between groups [29–31, 33, 35–38].

However, some studies have also shown no significant differences between balance measures. For example: Bouteraa’s study included static and dynamic balance; whilst there were significant differences between groups for the dynamic balance measure (YBT), no differences were observed for the static balance measure (SBT) [14]. Lu’s study examined dynamic balance indices, with the DPSI index showing significant differences between groups, whilst there were no significant differences in the L-COP_AP and L-COP_ML components of the COP [27]. In Guo and Xu’s study, the dynamic balance indices DPSI-DF, DPSI-NF, and DPSI-NL improved, but there were no significant differences between groups; however, significant differences were observed for the LSI-3 and LSI-6 indices [32]. Similar inconsistencies were reported in other studies [28, 34, 39].

The effect of combined balance and plyometric training on agility

Seven studies assessed athletes’ agility, including badminton [27, 32, 35], basketball [14], handball [29], taekwondo [30], soccer [38]. Testing methods include T-test, modified T-test, Illinois Agility Test, modified Illinois Agility Test, 5-0-5 test, hexagonal test (H-test), 8-run agility, sport-specific agility tests for taekwondo and badminton. Five studies indicate that combined balance and plyometric training can effectively improve athletes’ agility and that there are significant differences between groups [14, 27, 29, 30, 38].

However, the measures of agility used in some studies are diverse; whilst some measures show significant differences between groups, others do not show significant differences either between groups or over time. In Guo and Xu’s study, the T-test and H-test revealed significant differences between groups and over time in agility assessments; however, the SEMO measure showed significant differences over time but not between groups [32]. In the study by Guo et al., there were significant differences in the agility metric 505test, but no between-group differences were observed for SEMO [35].

The effect of combined balance and plyometric training on speed

Three studies assessed the speed capabilities of athletes, including soccer athletes [38], handball athletes [29], and basketball athletes [14]. The metrics used to assess speed capabilities included 5–30 m sprints. Two studies demonstrated that combined balance and plyometric training has a significant effect on improving athletic speed, with significant main effects observed for time and group categories, as well as their interaction [29, 38]. However, one study has also assessed speed performance, and for some of these measures, no significant differences were found [14].

The effect of combined balance and plyometric training on power

Four studies assessed the explosive power of athletes [14, 29, 35, 38], including badminton [35], football [38], handball [29], and basketball [14], with the main test metrics including SJ, CMJ, DJ, SLJ, and RSI. Three studies reported that combined balance and plyometric training can effectively improve athletes’ explosive power, with significant differences observed between sets and over time [29, 35, 38]. However, one study has found no significant differences in certain indicators used to assess explosive power [14].

Discussion

This systematic review of 14 randomized controlled trials examined the effects of combined balance and plyometric training on athletic performance. The main findings suggest that this training may improve balance and agility, with most included studies reporting significant between-group differences. However, the strength of evidence varied by outcome: GRADE assessment rated the evidence as moderate for balance/agility (downgraded due to imprecision) and low for speed/power (downgraded due to inconsistency and/or imprecision). Thus, the balance and agility findings are moderately supported, whereas the speed and power results remain hypothesis-generating and require confirmation through larger, well-designed studies.

The effect of combined balance and plyometric training on balance

Balance is a key component of athletic performance and essential for motor control [40]. Eight studies reported significant improvements in balance outcomes following combined balance and plyometric training [29–31, 33, 35–38]. Effect sizes ranged from 0.18 to 1.21; overall, 10 of 14 studies demonstrated moderate-to-large effects (d ≥ 0.50), though only eight reached statistical significance. Notably, improvements were consistently observed for dynamic balance measures (YBT, DPSI) than for static measures (SBT), which showed non-significant results in several studies [14].

This discrepancy may reflect the differential neuromuscular demands of static versus dynamic balance tasks. Static postural control relies on the continuous integration of multisensory inputs—including visual, vestibular, and somatosensory (particularly ankle and plantar proprioceptive) afferents—within the central nervous system to regulate the center of pressure relative to the base of support [41]. Dynamic balance, by contrast, involves additional coordination of multi-joint movements, feedforward postural adjustments, and rapid force production, which may be more responsive to combined explosive and sensorimotor training stimuli [42]. These findings are largely consistent with previous research on plyometric and balance training, though direct comparisons are limited by differences in the conceptual definition of training interventions across studies [43].

The effect of combined balance and plyometric training on agility

Agility is a fundamental motor skill that underpins competitive performance, requiring rapid change-of-direction ability and coordinated core muscle control. Five studies (involving badminton, handball, taekwondo, basketball, and soccer athletes) reported significant between-group improvements in agility [14, 27, 29, 30, 38], but two studies have found no significant differences between groups [32, 35]. Effect sizes for agility outcomes ranged from 0.35 to 1.05, with 5 of 7 studies showing moderate-to-large effects. For instance, significant improvements were observed for the T-test and H-test, but not for the SEMO test, which involves repeated lateral movements and may place greater demands on dynamic balance control. This inconsistency suggests that the effects of combined training may be test-specific, and that different agility assessments capture distinct physical capabilities. Nonetheless, the overall pattern of findings supports the potential benefit of combined training for change-of-direction ability.

The effect of combined balance and plyometric training on speed

Speed ability is closely related to an athlete’s level of explosive power [44],speed improvements primarily stem from lower-body explosive power and the extent to which motor units are recruited within a short time [45]. Three studies assessed the effects of combined balance and plyometric training on speed ability, two of the three studies reported significant between-group improvements, with effect sizes ranging from 0.15 to 0.78 (moderate effects in the two significant studies). However, one study has also assessed speed performance, and for some of these measures, no significant differences were found [14].

The limited number of eligible studies—only three—precludes definitive conclusions. One possible explanation for the inconsistent findings is that previous meta-analyses focused on plyometric training alone [9, 10, 46], whereas the addition of balance training may, under certain conditions, attenuate the explosive power gains typically associated with plyometric-only training [47]. Methodological factors, such as differences in testing environments, timing methods, surface conditions, and small sample sizes, may also have contributed to the non-significant results [48].

The effect of combined balance and plyometric training on power

Explosive power reflects an athlete’s ability to perform rapid movements within a short duration, encompassing both absolute and relative explosive power [49]. Four studies examined the effects of combined balance and plyometric training on athletes’ explosive power, involving athletes from badminton, football, basketball, and handball. Three studies reported significant between-group improvements in explosive power [29, 35, 38]. Effect sizes for power outcomes ranged from 0.12 to 1.12, with 3 of 4 studies showing moderate-to-large effects. However, one study found that combined balance and plyometric training did not result in significant between-group differences in certain indicators of athletes’ explosive power [14]. This outcome may be attributed to the high baseline explosive power levels of the athletes (basketball players) and the relatively short intervention duration (6 weeks), which may have limited further improvement [47, 50].

Explanations for inconsistent findings

The heterogeneity observed across studies can be attributed to several common factors. First, training protocols varied considerably in duration (6–12 weeks), frequency (2–3 sessions/week), and session length (5–60 min), making direct comparisons challenging. Second, outcome measures differed across studies, with some using sport-specific tests and others using generic tests, which may capture different aspects of the same physical quality. Third, athlete populations varied by sport, age, sex, and baseline training status, potentially influencing training responsiveness.

Regarding potential mechanisms, while none of the included studies directly measured neuromuscular or physiological parameters, several speculative explanations have been proposed. Balance training may enhance sensorimotor function and neuromuscular coordination, while plyometric training improves lower-limb explosive power. Their combination may optimize stretch-shortening cycle efficiency while concurrently enhancing postural control [16]. However, the addition of balance training may also induce a neuromuscular interference effect in some contexts, potentially attenuating the explosive power gains typically associated with plyometric training alone [47]. These hypotheses require direct experimental validation in future studies.

Mechanistic interpretations—including neuromuscular adaptations, enhanced motor unit recruitment, improved sensorimotor integration, and postural control reweighting—remain theoretical, as no included studies directly assessed neuromuscular or physiological parameters (e.g., electromyography, motor evoked potentials, or cortical activation patterns). These explanations, derived from extrapolations of previous literature on plyometric and balance training in other contexts, may not fully capture the specific adaptations induced by their combination and should therefore be viewed as speculative and hypothesis-generating rather than definitive. Future research should incorporate direct mechanistic assessments (e.g., surface electromyography, neuromuscular function tests, and neurophysiological measures) to provide empirical evidence and clarify the underlying pathways.

Interpretation is further constrained by the inability to examine potential moderators—such as sport type, competition level, sex, or baseline performance—due to limited study numbers (only one to two per sport) and population heterogeneity (varying from elite to amateur, with mixed sex distributions). Whether certain athlete populations derive greater benefits from combined training remains unclear, warranting investigation through adequately powered subgroup analyses in future studies.

Strengths and limitations

This review has several strengths. First, it was conducted and reported in accordance with PRISMA 2020 guidelines, with methodological quality assessed using PEDro, providing a comprehensive appraisal of the included studies. Second, literature screening and quality assessment were performed independently by multiple reviewers, reducing subjective bias. Third, the intervention definition was strictly operationalized to ensure conceptual clarity.

Several limitations should be acknowledged.

First, generalizability is constrained by small sample sizes, heterogeneous populations (7 sports, age 11–24 years), and diverse training protocols (6–12 weeks, 2–3 sessions/week). Two studies included athletes with ankle instability, limiting applicability to healthy populations. Additionally, the limited number of studies per sport and the variability in athlete level, sex, and baseline performance across studies prevented subgroup analyses to examine whether these factors moderate training effects. Incomplete reporting of training intensity, progression, volume, rest intervals, and supervision across studies limited intervention reproducibility.

Second, most interventions lasted only 6 weeks, leaving long-term effects unclear. Additionally, although most studies reported that regular training continued and was matched between groups, three studies did not provide sufficient detail on concurrent training, and the possibility of confounding cannot be entirely excluded. Restricting to English-language studies may have introduced language bias.

Third, the included studies employed a variety of tests to assess the same construct—particularly for balance (e.g., YBT, DPSI, SBT) and agility (e.g., T-test, H-test, SEMO test). While these tests are widely used in sports science, their measurement properties warrant consideration. The YBT and T-test have demonstrated good test-retest reliability (ICC > 0.80) and established validity in athletic populations. However, the psychometric properties of other tests—particularly the SEMO test and SBT—are less well-documented, and the minimal detectable change or minimal clinically important difference has not been established for most of the tests used in the included studies. Furthermore, the lack of a universally accepted “gold standard” for assessing balance and agility in sport-specific contexts complicates cross-study comparisons. This variability in measurement tools may have contributed to the heterogeneity observed across studies and should be considered when interpreting the pooled findings.

Fourth, common methodological limitations included lack of blinding and allocation concealment; however, 11 of 14 studies were high quality, and the consistency of findings suggests overall conclusions are unlikely to be substantially biased. The moderate-to-low GRADE ratings reflect study limitations rather than proven ineffectiveness; larger, standardized trials are needed to improve evidence quality.

Fifth, the substantial heterogeneity across studies precluded meta-analysis, and absence of dose-response analysis prevents recommendations on optimal training parameters. Our definition excluded unstable-surface plyometrics, as such exercises fundamentally alter biomechanical demands and integrate balance and plyometric stimuli into a single movement, whereas our review specifically examined their effects as discrete training components. We acknowledge this narrow definition may limit evidence scope and generalizability; future reviews should consider broader criteria.

Finally, the biomechanical and neurophysiological mechanisms underlying the observed benefits remain to be explored. Future studies should adhere to CONSORT guidelines, include longer follow-up periods, adopt standardized outcome measures, and employ dose-response designs to establish evidence-based training guidelines.

Practical implications

The findings of this review provide tentative practical considerations for coaches and sports practitioners. Preliminary evidence suggests that combined balance and plyometric training may be beneficial for improving dynamic balance and agility. However, due to the heterogeneity of training protocols, outcome measures, and athlete populations, as well as the absence of a meta-analysis, these findings should be considered hypothesis-generating rather than conclusive. Coaches are encouraged to consider combined training as a potential option for enhancing balance and agility, while recognizing that the optimal training parameters (e.g., frequency, duration, volume) cannot be determined from the current evidence, as no dose-response analysis was conducted. Future studies should encompass a broader range of participants, such as adolescents, older adults, and amateur athletes, to extend this training method to different population groups.

Conclusion

In summary, existing research indicates that combining plyometric training with balance training may be an effective approach for improving athletes’ balance and agility. The GRADE assessment (Table 4) supports moderate-confidence evidence for these outcomes, with consistent improvements observed across the majority of included studies. However, the evidence for speed and power improvements is rated as low confidence due to the limited number of studies and inconsistent findings (only two of three studies reported significant speed improvements, and three of four reported significant power improvements). Therefore, conclusions regarding speed and explosive power should be considered hypothesis-generating and require verification through future high-quality research.

Accordingly, this review offers preliminary practical guidance for coaches and practitioners: recommendations for balance and agility are supported by moderate-certainty evidence, whereas those for speed and power are based on low-certainty evidence and should be applied with caution. Future research should incorporate direct mechanistic assessments, adhere to stricter methodological standards (e.g., larger sample sizes, standardized outcome measures), and employ dose-response designs to establish evidence-based training guidelines. Given the heterogeneity across studies, a priori subgroup designs with larger, more homogeneous samples are needed to determine whether effects are moderated by age, sex, sport type, competition level, training volume, or baseline performance, enabling more precise and individualized training prescriptions. Additionally, future studies should systematically report and control for concurrent regular training to isolate the effects of combined interventions. These efforts are essential to improve GRADE ratings and strengthen confidence in the estimated effects.

Authors' contributions

L and D jointly authored the main manuscript. L was responsible for data collection and analysis, and prepared the charts. D conducted data analysis and interpretation, and reviewed and approved the final manuscript. All authors contributed to the study design.

Funding

The author(s) declare that no funds were received for the research and/or publication of this article.

Data availability

The datasets used and/or analyzed in the present study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable. This is a desk-based study with no human subject involvement; therefore, no ethics approval was required.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.