Key Points

  • Plyometric jump training is an effective method to improve measures of physical fitness (i.e. muscle strength and muscle power) and sport-specific performance (e.g. sport-specific time-trial speed) in-water sports athletes.

  • The results of this study are based on a total of 618 water sport athletes, from 26 articles of moderate-to-high methodological quality, low-to-moderate impact of heterogeneity, and very low-to-moderate certainty of evidence.

Introduction

All sports that are practiced on (e.g. rowing) or in (e.g. swimming; triathlon; water polo; synchronized swimming) water are considered water sports [1, 2]. Many water sports are part of the Olympic games [3, 4]. Water sports are physically demanding [5,6,7,8,9]. Therefore, to successfully cope with such high demands, water sport athletes (WSA) invest many hours of training per minute of competition [10]. For example, international medal-winning rowers spent ~ 1100–1200 h of training per year [11]; if these rowers practiced continuously for the 52 weeks of the year, this would represent an average of ~ 23 weekly hours of training. Therefore, the allocated time should be devoted towards optimal and time-efficient training activities [12]. Such training activities should consider the development of adequate physical fitness (PF) and sport-specific performance (SSP) components, usually involving highly developed muscular fitness [5,6,7,8,9] and aerobic endurance (e.g. cardiorespiratory fitness) [5]. In fact, muscle strength and power (e.g. jumping from the starting block and flip turns in swimmers) play a relevant role in competitive performance [13,14,15]. For example, up to large associations (r = 0.40 to 0.70) have been reported between 20- and 50-m front crawl swimming performance and mean propulsive power in jump squat [16], leg extension strength [17], and horizontal jump distance [18]. More experienced (older) rowers show greater power than younger rowers, and in junior rowers, greater power is shown in international-ranked rowers than in non-ranked rowers [19]. High lower limb muscle power was associated with effective free throws in-water polo [20]. Cycling peak power and vertical jump are positively associated with performance in sailing [21]. In addition to muscle strength and power, training activities should also consider the development of adequate body composition, as performance in-water sports relates to athlete’s body composition (e.g. faster swimmers and rowers usually have greater lean body mass or thigh perimeter) [8, 22, 23]. Moreover, strength and power-related training can also enhance WSA efficiency and competitive velocity [24].

In this regard, higher-intensity shorter-duration training approaches have been recommended over traditional low-intensity high-volume training methods [25,26,27,28]. Plyometric jump training (PJT), a high-intensity short-duration training method, may offer an adequate stimulus to improve PF and SSP [29,30,31,32,33,34,35,36], including (but not limited to) muscle strength [37], muscle power [38], and body composition [30, 32]. Training intensity seems to be a key feature of PJT programming [39, 40], with PJT exercises usually implicating a fast stretch–shortening cycle muscle action, allowing greater concentric work performance than an isolated concentric muscle action, stimulating a high rate of force development, and force absorption muscle capacities (i.e. eccentric force) [29, 41, 42]. In addition, PJT implicates muscle stimulus inducing neuro-mechanical adaptations [29] that may be reproduced in both isoinertial (e.g. land) and isokinetic (e.g. water) environments [43,44,45,46,47,48,49,50,51]. Indeed, PJT may improve WSA performance [31, 52, 53], targeting key muscles from ankle, knee, and hip joints that may aid during key competitive movements such as kicking in swimming [54], jumping from the start platform and flip turns, lower-limb extension during the stroke in rowing [5, 7], among others. Furthermore, PJT may be equally or even more effective to improve PF (e.g. vertical jump; endurance; bone mass) and SSP (e.g. sport-specific sprinting) in WSA compared to other modes of training-sports [55, 56]. Moreover, PJT may improve flexibility [57,58,59] which is of paramount importance for athletes in general and WSA in particular [54, 60,61,62,63,64]. Although previous intervention studies have evidenced the effectiveness of PJT on PF and SSP in WSA, most published studies have included relatively small samples (i.e. median n = 11), a common issue in the sport-science literature [65], casting doubts on the transferability of such findings into practice. Additionally, not all studies [66] agree with the beneficial effects of PJT on PF and SSP in WSA, reflecting inconsistency in the literature.

Previous works have been performed to solve controversy by systematically aggregating the literature related to strength and conditioning practices in WSA [52, 53, 67, 68]. However, most of these reviews did not consider including a meta-analysis, or addressed a restricted population (e.g. only swimmers; only rowers). Moreover, the aforementioned reviews focused on a myriad of strength and conditioning methods, or on very specific and isolated outcomes (e.g. swimming speed), precluding a more comprehensive view regarding the intervention effects. This leaves the question of the effects of single-mode PJT unanswered. Furthermore, when PJT studies were included, the number was low (n = 6 studies [53]; n = 4 studies [52]; n = 2 studies [68]). To account for the previous limitations (i.e. reduced sample size), a meta-analysis seems to be needed to help practitioners taking evidence-based informed decisions as to PJT implementation [69]. Additionally, a systematic review with meta-analysis may help to detect gaps and limitations in the PJT literature, providing valuable information for scientists and practitioners about future research avenues. However, to the authors’ knowledge, no review has attempted to meta-analyse the large amount of currently available literature regarding the potential effects of PJT on PF and SSP in WSA. Therefore, the primary aim of this systematic review with meta-analysis was to examine the effects of PJT, compared with active (i.e. standard sports training) or specific-active (i.e. alternative training intervention) control groups, on the PF (muscle strength, muscle power, body composition) and SSP (in-water vertical jump, in-water agility, time trial) outcomes in WSA, through a systematic review with meta-analysis of randomized and non-randomized controlled studies.

Methods

Procedures

A systematic review with meta-analysis was conducted following the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [70], and adapted a posteriori to new reporting guidelines (e.g. PRISMA 2020) [71] as such changes are expected as the field evolves (e.g. new databases; new concepts/terms). The most relevant adaptations are described in Additional file 1: Table S1.

Literature Search: Administration and Update

We considered recommendations from the two most comprehensive scoping reviews that previously examined PJT literature [42, 72]. Computerized literature searches were conducted in the electronic databases PubMed, Web of Science, and SCOPUS. The search strategy was conducted using (in different combinations) the Boolean operators AND/OR with the following keywords (all database fields used): “ballistic”, “complex”, “cycle”, “explosive”, “force”, “plyometric”, “shortening”, “stretch”, “training”, and “velocity”. Examples of combinations included: “ballistic” AND “training”; (“ballistic” OR “plyometric” OR “explosive”) AND “training”. Additionally, using the title database field, the following keywords were employed in the search: “jump”, “power”, and “training”. After an initial search in April 2017, an account was created by one of the authors (RRC) in each of the respective databases, through which the author received automatically generated email updates regarding the search terms used. The search was refined in May 2019 and August 2021, with updates received daily (if available). Studies were eligible for inclusion up to January 2022. The main advantage of this search approach is that it assumes that new knowledge will appear and allow improvements in sport/clinical decision-making. Indeed, the rate of PJT studies increased exponentially during the last years [42, 72], and we plan to update this systematic review every 5 years. The same author (RRC) conducted the initial search and removed duplicates. Thereafter, the search results were analysed according to the eligibility criteria (Table 1). The search strategy (code line) for each database and background of search history is described in Additional file 1: Table S1.

Table 1 Selection criteria used in the meta-analysis
Full size table

In selecting studies for inclusion, a review of all relevant titles was conducted before examination of the abstracts and full-text versions. Two authors (RRC and RKT) independently screened the titles, abstracts, and full-text versions of retrieved studies. During the search and review process, potential discrepancies between the same two authors regarding inclusion and exclusion criteria (e.g. type of control group, intervention adequacy) were resolved through consensus with a third author (APC). From selected articles to be included, reference lists were analysed to identify any additional relevant studies.

Inclusion and Exclusion Criteria

A PICOS (participants, intervention, comparators, outcomes, and study design) approach was used to rate studies for eligibility [70]. Table 1 indicates our inclusion/exclusion criteria.

Additionally, only full-text, peer-reviewed, original studies were considered for the present meta-analysis. Additional exclusion criteria are provided as Additional file 2: Table S2. Because of the potential difficulties of translating articles written in different languages and the fact that 99.6% of the jump training literature is published in English [72], only articles written in English, as well as Spanish, German, and Portuguese (i.e. the authors’ native languages), were considered for this meta-analysis.

Data Extraction

The extraction of dependent variables from the included studies considered previous recommendations regarding relevant PF and SSP attributes for WSA [5,6,7,8,9, 73,74,75,76,77,78]. Therefore, the effects of PJT (compared to controls) on either in/on land/water were identified through PF and SSP attributes. Although an extensive list of outcome data to be collected was considered a priori (e.g. maximum oxygen consumption [VO2max]; balance), the final list depended on the available number of studies reporting data for a given outcome. Therefore, measures of PF included maximal strength (i.e. one-repetition maximum [1RM] back squat), horizontal jump distance, squat jump height, countermovement jump height, body mass, fat mass, and thigh girth. Additionally, SSP measures included in-water vertical jump height, in-water agility velocity, and time-trial velocity. Jump, linear sprint, change-of-direction speed, and strength testing usually present very high test–retest reliability (with an intra-class correlation coefficient of > 0.9) [79,80,81,82], which is essential to ensure strong consistency between analysed studies within a meta-analysis [70]. Independently of the above, when reported, we extracted reliability measures from the included studies (e.g. intra-class correlation coefficient).

The means and standard deviation of dependent variables were extracted at pre- and post-PJT time points from included studies using Microsoft Excel (Microsoft Corporation, Redmond, WA, USA). When the required data were not clearly or completely reported, authors of the respective study were contacted for clarification. If no response was obtained from the authors (after two attempts) or the authors could not provide the requested data, the study outcome was excluded from the analysis. However, when data were displayed in a figure and no numerical data were provided by authors after being contacted, validated (r = 0.99, p < 0.001) [83] software (WebPlotDigitizer, version 4.5; https://apps.automeris.io/wpd/) was used to derive numerical data from figures by two independent authors (RKT and JA) and the Cronbach’s Alpha was then calculated. Two authors (RRC and RKT) performed data extraction independently, and any discrepancies between them (e.g. mean value for a given outcome, number of participants in a group) were resolved through consensus with a third author (APC).

Data extracted regarding PJT intervention characteristics included (i) the box height used during PJT exercises, (ii) whether the PJT was combined with another lower-limb training method, (iii) duration (number of weeks) of the PJT intervention; (iv) frequency of PJT sessions (sessions per week); (v) intensity of the PJT exercises; (vi) number of total jumps completed during the PJT intervention; (vii) progressive overload applied during the PJT intervention; (viii) recovery time between sets, repetitions, and training sessions; (ix) replacement of a given part of the standard sport training schedule with PJT exercises; (x) type of PJT exercises; and (xi) type of surface used during PJT. We also extracted data regarding participant’s sex, age (years), body mass (kg), height (m), previous experience with PJT, type and level of water sport practiced, and training period of the season.

Methodological Quality of the Included Studies

The Physiotherapy Evidence Database (PEDro) scale was used to assess the methodological quality of the included studies, which were rated from 0 (lowest quality) to 10 (highest quality). The validity and reliability of the PEDro scale have been established previously [84,85,86]. Moreover, the PEDro scale is probably the most frequently used in the PJT literature [42, 87, 88]. Its items mostly assess factors related to the risk of bias in studies. Accordingly, it helps to make comparisons between meta-analyses. Considering that it is not possible to satisfy all scale items in PJT interventions [89], as outlined in previous systematic reviews in the sub-field of PJT, the methodological quality of PJT studies was interpreted using the following convention [35, 87, 90]: ≤ 3 points was considered as “poor” quality, 4–5 points was considered as “moderate” quality, and 6–10 points was considered as “high” quality. If trials were already rated and listed in the PEDro database, the respective scores were adopted. Two authors (RRC and RKT) assessed the methodological quality for each included study independently, and any discrepancies between them were resolved via consensus with a third author (APC).

Summary Measures, Synthesis of Results, and Publication Bias

Although meta-analyses can be done with as few as two studies [91], because reduced sample sizes are common in the sport-science literature [92], including PJT studies [42, 65, 72, 93], meta-analysis was only conducted in the present case when ≥ 3 studies were available [76, 94]. Effect sizes (ES, i.e. Hedges’ g) for each PF and SSP attribute in the PJT and control/comparator groups were calculated using pre-training and post-training mean and standard deviation. Data were standardized using post-intervention standard deviation values. The random-effects model was used to account for differences between studies that might affect the PJT effect [95, 96]. The ES values are presented with 95% confidence intervals (95% CIs). Calculated ES were interpreted using the following scale: < 0.2 trivial, 0.2–0.6 small, > 0.6–1.2 moderate, > 1.2–2.0 large, > 2.0–4.0 very large, > 4.0 extremely large [97]. In studies including more than one intervention group, the sample size in the control group was proportionately divided to facilitate comparisons across multiple groups [98]. The level of heterogeneity was assessed using the I2 statistic, with values of < 25%, 25–75%, and > 75% representing low, moderate, and high levels of heterogeneity, respectively [99]. The risk of publication bias was explored for continuous variables (≥ 10 studies per outcome) [100,101,102] using the extended Egger’s test [102]. To adjust for publication bias, a sensitivity analysis was conducted using the trim and fill method [103], with L0 as the default estimator for the number of missing studies [104]. All analyses were carried out using the Comprehensive Meta-Analysis software (version 2, Biostat, Englewood, NJ, USA). Statistical significance was set at p ≤ 0.05.

Moderator Analyses

Using a random-effects model and independent computed single factor analysis, potential sources of heterogeneity likely to influence the effects of training were selected, including participants' sex, type of sport, programme duration (number of weeks), and total number of training sessions. When appropriate, subgroup analyses and single training factor analyses were divided using the median split technique [105,106,107]. The median was calculated if at least three studies provided data for a given moderator category. Of note, when two experimental groups (with the same information for a given moderator) were included in a study, only one of the groups was considered to avoid an augmented influence of the study on the median calculation. In addition, instead of using a global median value for a given moderator (e.g. median number of weeks, derived from all included studies), median values were calculated considering only those studies that provided data for the outcome being analysed.

Additional Analyses

Certainty of Evidence

Two authors (JA and RRC) judged the certainty of evidence using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) [108,109,110]. Evidence started at a high level of certainty (per outcome), but was downgraded based on the following criteria: (i) Risk of bias in studies: judgments were downgraded by one level if the average PEDro scores were moderate (< 6) or by two levels if they were poor (< 4); (ii) Indirectness: low risk of indirectness was attributed by default due to the specificity of populations, interventions, comparators and outcomes being guaranteed by the eligibility criteria; (iii) Risk of publication bias: downgraded by one level if there was suspected publication bias; (iv) Inconsistency: judgments were downgraded by one level when the impact of statistical heterogeneity (I2) was high (> 75%); (v) Imprecision: one level of downgrading occurred whenever < 800 participants were available for a comparison [111] and/or if there was no clear direction of the effects. In case both were observed, certainty was downgraded by two levels.

When the number of comparison trials was insufficient to perform meta-analysis, the evidence was automatically judged as very low certainty. Therefore, for the outcomes not included in the meta-analyses, the certainty of evidence should be considered very low.

Results

Study Selection

The search process in the databases identified 11,028 studies. Figure 1 provides a flowchart illustrating the study selection process.

Fig. 1
figure 1

Flow diagram of the search process

Full size image

Duplicate studies were removed (n = 6486). After study titles and abstracts were screened, 3,698 studies were removed and 844 full-text studies were screened. Thirty-one studies were included in qualitative synthesis and their reference lists were screened, with 2 records [112, 113] identified through the reference lists screening process. Thereafter, 7 full-text studies were excluded [114,115,116,117,118,119,120] (exclusion reasons in Additional file 3: Table S3). Finally, 26 were considered eligible for meta-analysis [56, 66, 112, 113, 121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137,138,139,140,141,142].

Methodological Appraisal of the Included Studies

According to the PEDro checklist, the median (i.e. nonparametric) score was 5.5, with 13 studies attaining high quality (6 points), and 13 studies were of moderate quality (4–5 points). Of note, no study scored above 6 points (Table 2). The two independent reviewers that performed the methodological appraisal of the included studies achieved a Spearman correlation (i.e. nonparametric data) agreement of 0.93.

Table 2 Methodological quality of the included studies using the PEDro rating scale
Full size table

Study Characteristics

The participant characteristics and the PJT programmes of the included studies are detailed in Tables 3 and 4, respectively.

Table 3 Participants’ characteristics from the included studies
Full size table
Table 4 Plyometric jump training programming variables
Full size table

The 26 included studies recruited swimmers, triathletes, rowers, artistic swimmers, and water polo athletes, for a total of 618 participants, with 330 participants in the intervention groups (31 groups) and 288 participants in the control groups (26 groups). Among the 26 control groups, 9 groups were specific-active controls (see Table 3; i.e. alternative training intervention controls), involved in a non-PJT intervention (e.g. resistance training), and the other 17 groups were active controls, involved in their regular sport-specific training schedules (see Table 3). Twelve studies included participants with a mean age of < 18 years old (Table 3). Regarding participants' sex, one study did not report the sex of the participants (n = 22 [4% of total participants]), nine studies reported a mixed sample of male and female participants (n = 202 [33% of total participants]), five studies involved females only (n = 114 [18% of total participants]), and 11 studies involved male participants (n = 280 [45% of total participants]) (Table 3).

The duration of the training programmes in the intervention and control groups ranged from 4 to 36 weeks (Table 4) and the frequency of weekly training sessions ranged from 2 to 4 (Table 4). Methods for reporting training intensity included maximal effort intensity, such as reactive strength index, vertical jump height, horizontal distance, power output (associated with a given external load), or minimal ground-contact time (Table 4). Training intensity was also reported as ground impact force (i.e. N), impact load rate (e.g. body mass/s), jumping rate (e.g. 1.8 jumps/s), rating of perceived exertion (e.g. 11–16, using a 20-point maximal scale), load (i.e. kg) for the loaded jumps, and percentage of one-repetition maximum. Some studies (n = 2) reported training intensity only qualitatively (e.g. low-moderate intensity), without further quantification, and some studies (n = 3) reported an individualized approach for the programming of PJT intensity. Nine studies did not provide any details regarding intensity (Table 4).

The testing protocols for each of the included PF and SSP outcomes in the meta-analysis are detailed in Table 5.

Table 5 Measurement protocols for studies outcomes included in meta-analysis
Full size table

Some studies involved results from a single trial [129, 130] (i.e. the studies used the same population). However, as the different publications reported different outcomes, both studies were included in the meta-analysis. In studies including more than one intervention group, the sample size in the control group was proportionately divided to facilitate comparisons across multiple groups [98]. This was also the case for studies involving four experimental groups [134] and two experimental groups [112, 121].

Results of the Meta-analysis

Sport-Specific Performance

Four studies provided data for in-water vertical jump performance height, involving 8 experimental and 4 control groups (pooled n = 126; specific-active control groups, n = 3). The results showed no effect for the PJT groups compared to the control groups (ES = 0.03; 95% CI = − 0.37 to 0.42; p = 0.899; Fig. 2; I2 = 12.2%).

Fig. 2
figure 2

Forest plot for changes regarding in-water vertical jump performance (e.g. cm) in participants after plyometric jump training (PJT) compared to controls. Forest plot values are shown effect sizes (Hedges’ g) with 95% confidence intervals (CI). Black squares: individual studies. Its size represents their relative weights. White rhomboid: summary value. Note: Letters (e.g. DG) at the end of a study (e.g. Martin et al. [134] DG) denotes that different experimental groups were included

Full size image

Three studies provided data for in-water agility performance (e.g. speed), involving 7 experimental and 3 control groups (pooled n = 105; specific-active control groups, n = 3). The results showed no effect for the PJT groups compared to the control groups (ES = − 0.15; 95% CI = − 0.56 to 0.26; p = 0.477; Fig. 3; I2 = 0.0%).

Fig. 3
figure 3

Forest plot for changes regarding in-water agility (e.g. agility time) in participants after plyometric jump training (PJT) compared to controls. Forest plot values are shown effect sizes (Hedges’ g) with 95% confidence intervals (CI). Black squares: individual studies. Its size represents their relative weights. White rhomboid: summary value. Note: Letters (e.g. DG) at the end of a study (e.g. Martin et al. [134] DG) denotes that different experimental groups were included

Full size image

Seventeen studies provided data for time trial performance (e.g. speed), involving 22 experimental and 17 control groups (pooled n = 438; specific-active control groups, n = 8). The results showed a significant small effect in favour of the PJT groups compared to the control groups (ES = 0.42; 95% CI = 0.13 to 0.72; p = 0.005; Fig. 4; I2 = 54.6%; Egger test = 0.051).

Fig. 4
figure 4

Forest plot for changes regarding in-water time trial performance (e.g. 25-m swimming speed) in participants after plyometric jump training (PJT) compared to controls. Forest plot values are shown effect sizes (Hedges’ g) with 95% confidence intervals (CI). Black squares: individual studies. Its size represents their relative weights. White rhomboid: summary value. Note: Letters (e.g. DG) at the end of a study (e.g. Martin et al. [134] DG) denotes that different experimental groups were included

Full size image

Physical Fitness

Four studies provided data for maximal strength performance (i.e. one-repetition maximum [1RM] back squat), involving 5 experimental and 4 control groups (pooled n = 97; specific-active control groups, n = 3). The results showed significant, moderate effect in favour of the PJT groups compared to the control groups (ES = 0.67; 95% CI = 0.03 to 1.31; p = 0.041; Fig. 5; I2 = 58.1%).

Fig. 5
figure 5

Forest plot for changes in maximal strength performance (i.e. squat one repetition maximum, as kg), in participants after plyometric jump training (PJT) compared to controls. Forest plot values are shown effect sizes (Hedges’ g) with 95% confidence intervals (CI). Black squares: individual studies. Its size represents their relative weights. White rhomboid: summary value. Note: Letters (e.g. DG) at the end of a study (e.g. Martin et al. [134] DG) denotes that different experimental groups were included

Full size image

Four studies provided data for horizontal jump displacement performance, involving 4 experimental and 4 control groups (pooled n = 105; specific-active control groups, n = 0). The results showed significant, large effect for the PJT groups compared to the control groups (ES = 1.47; 95% CI = 0.33 to 2.61; p = 0.011; Fig. 6; I2 = 84.9%).

Fig. 6
figure 6

Forest plot for changes in horizontal jump displacement performance (e.g. cm), in participants after plyometric jump training (PJT) compared to controls. Forest plot values are shown effect sizes (Hedges’ g) with 95% confidence intervals (CI). Black squares: individual studies. Its size represents their relative weights. White rhomboid: summary value

Full size image

Five studies provided data for squat jump performance height, involving 5 experimental and 5 control groups (pooled n = 91; specific-active control groups, n = 1). The results showed significant, moderate effect in favour of the PJT groups compared to the control groups (ES = 0.79; 95% CI = 0.38 to 1.20; p < 0.001; Fig. 7; I2 = 0.0%).

Fig. 7
figure 7

Forest plot for changes in squat jump performance (e.g. vertical height), in participants after plyometric jump training (PJT) compared to controls. Forest plot values are shown effect sizes (Hedges’ g) with 95% confidence intervals (CI). Black squares: individual studies. Its size represents their relative weights. White rhomboid: summary value

Full size image

Fifteen studies provided data for countermovement jump height, involving 20 experimental and 15 control groups (pooled n = 378; specific-active control groups, n = 8). The results showed significant, moderate effect in favour of the PJT groups compared to the control groups (ES = 0.89; 95% CI = 0.43 to 1.34; p < 0.001; Fig. 8; I2 = 75.9%; Egger test = 0.066).

Fig. 8
figure 8

Forest plot for changes in countermovement jump performance (e.g. vertical height), in participants after plyometric jump training (PJT) compared to controls. Forest plot values are shown effect sizes (Hedges’ g) with 95% confidence intervals (CI). Black squares: individual studies. Its size represents their relative weights. White rhomboid: summary value. Note: Letters (e.g. DG) at the end of a study (e.g. Martin et al. [134] DG) denotes that different experimental groups were included

Full size image

Seven studies provided data for body mass, involving 7 experimental and 7 control groups (pooled n = 158; specific-active control groups, n = 1). The results showed no effect for the PJT groups compared to the control groups (ES = 0.07; 95% CI = − 0.23 to 0.37; p = 0.657; Fig. 9; I2 = 0.0%). Similar results (figures not showed) were observed for fat mass and thigh girth (ES = 0.06 to 0.15; p = 0.452 to 0.841).

Fig. 9
figure 9

Forest plot for changes in body mass (i.e. kg), in participants after plyometric jump training (PJT) compared to controls. Forest plot values are shown effect sizes (Hedges’ g) with 95% confidence intervals (CI). Black squares: individual studies. Its size represents their relative weights. White rhomboid: summary value

Full size image

Moderator Analyses

Participants’ sex, type of sport, programme duration (number of weeks), and total number of training sessions were included as moderators. However, such moderators were available only for the analysis of time-trial performance and countermovement jump height, as the number of studies available for analysis of other outcomes was insufficient.

Regarding time-trial performance changes after PJT, no significant difference was noted between males (14 groups; ES = 0.53, 95% CI = 0.06 to 1.00; I2 = 63.3%) and females (3 groups; ES = 0.34, 95% CI = − 0.12 to 0.79; I2 = 0.0%; between-moderators categories p value = 0.560). Similarly, no significant difference was noted after < 8 weeks of PJT (10 groups; ES = 0.26, 95% CI =  − 0.10 to 0.61; I2 = 24.3%) compared to ≥ 8 weeks of PJT (12 groups; ES = 0.56, 95% CI = 0.10 to 1.01; I2 = 67.1%; between-moderators categories p value = 0.303). Further, no significant difference was noted between swimmers (12 groups; ES = 0.46, 95% CI = 0.02 to 0.91; I2 = 67.9%) and water polo athletes (9 groups; ES = 0.32, 95% CI = -0.06 to 0.71; I2 = 23.6%; between-moderators categories p value = 0.647). Furthermore, no significant difference was noted after < 18 PJT sessions (10 groups; ES = 0.65, 95% CI = 0.11 to 1.18; I2 = 69.6%) compared to ≥ 18 PJT sessions (12 groups; ES = 0.21, 95% CI =  − 0.08 to 0.51; I2 = 17.1%; between-moderators categories p value = 0.161).

Regarding CMJ height changes after PJT, no significant difference was noted between males (12 groups; ES = 0.53, 95% CI = 0.04 to 1.02; I2 = 63.9%) and females (3 groups; ES = 2.38, 95% CI = 0.09 to 4.66; I2 = 92.5%; between-moderators categories p value = 0.123). Further, no significant difference was noted between swimmers (10 groups; ES = 1.19, 95% CI = 0.51 to 1.86; I2 = 80.2%) compared to water polo athletes (9 groups; ES = 0.61, 95% CI =  − 0.08 to 1.30; I2 = 73.0%; between-moderators categories p value = 0.240). Furthermore, no significant difference was noticed after < 22 PJT sessions (10 groups; ES = 0.85, 95% CI = 0.17 to 1.53; I2 = 78.4%) and ≥ 22 PJT sessions (10 groups; ES = 0.94, 95% CI = 0.29 to 1.59; I2 = 75.6%; between-moderators categories p value = 0.849). However, a significant difference was reported after < 8 weeks of PJT (6 groups; ES = 0.39, 95% CI = − 0.01 to 0.80; I2 = 0.0%) compared to ≥ 8 weeks of PJT (14 groups; ES = 1.16, 95% CI = 0.54 to 1.79; I2 = 81.7%; between-moderators categories p value = 0.043).

According to the GRADE assessment (Table 6), for in-water vertical jump, in-water agility, time-trial performance, and squat jump, the certainty of evidence is considered low. For horizontal jump, countermovement jump, and body mass, the certainty of evidence is deemed very low. For maximal strength, the certainty of evidence is judged as moderate.

Table 6 Certainty of evidence for meta-analysed outcomes
Full size table

Discussion

Twenty-six moderate-to-high-quality studies were meta-analysed, involving 618 WSA of both sexes, aged between 10 and 26 years, and from different sports disciplines such as swimming, triathlon, rowing, artistic swimming, and water polo. The results of the meta-analyses showed that PJT interventions induced moderate-to-large improvements (ES = 0.67 to 1.47) in athletes' PF (1RM back squat, horizontal jump distance, squat jump height, and countermovement jump height), and a small improvement in SSP (i.e. time trial speed). Such improvements were noted even when comparison entailed specific-active control groups. Mostly, the level of heterogeneity in the above-mentioned results was low-to-moderate.

Regarding in-water vertical jump height and in-water agility speed (i.e. SSP measures), these did not improve after PJT when compared to control conditions (ES = 0.03 to − 0.15, respectively). Of note, most (for vertical jump) or all (for agility) of the control conditions involved specific-active control groups, meaning that the effect of PJT was compared to an analogous training intervention. Therefore, PJT seems at least equally effective as other training strategies, such as water-specific resistance training, for improving SSP in WSA. Additionally, data regarding in-water vertical jump and agility performance came from one single research group [112, 113, 134, 142], and although the results involved different experiments, all the aforementioned studies included highly trained-professional water polo athletes, meaning that their potential for adaptation is reduced compared with athletes of lower competitive level, or from different sports. Therefore, both in-water vertical jump and in-water agility may have greater chances of improvement following PJT in WSA training at non-professional levels and/or when the effects of PJT are compared to active controls performing their habitual training schedule. Nonetheless, according to GRADE, there is low confidence in these results given that the certainty of evidence for these two outcomes was rated as low. Therefore, future high-quality studies are needed to achieve more conclusive evidence, including WSA with different training experience (e.g. international vs. regional) and water sports of different competitive demands (e.g. mostly metabolic vs. mostly neuromuscular).

With respect to time trial performance tests, these involved swimming distances from 5 m up to 200 m, and rowing distances from 500 m up to 2500 m. Time trial performance improved in WSA after PJT compared to control conditions, involving eight specific-active control groups and nine active controls performing their habitual training schedule. The significant transference effect between PJT exercises and sport-specific performance in other sports such as soccer (e.g. kicking velocity) was previously reported [143,144,145]. Our analyses additionally suggest that land-based (i.e. isoinertial) PJT has a meaningful transference potential into water-based (i.e. isokinetic) muscle actions. Of note, the transferability of findings from individual PJT studies into WSA practice has been obscured given that most published individual studies have included relatively small sample sizes. Although the results of our meta-analysis contribute to solving previous limitations (i.e. reduced sample size), helping with evidence-based informed decisions regarding PJT implementation [69], it is recommended that future studies in this field include larger sample sizes, provide a full depiction of relevant characteristics of the participants (e.g. previous experience with PJT) and the PJT intervention (e.g. intensity). Such methodological research advances in the field may contribute to improving the confidence in current results.

Improvements in time-trial performance may be related to different physiological adaptations induced after PJT, such as increased motor unit firing rate, improved inter-muscular coordination, improved lower-limb stiffness, increased muscle fibre contraction velocity, and power and force generation capability [29, 146,147,148,149]. Such adaptations may transfer to kinetic and kinematic improvements such as greater rate of force development and maximal force generation capability [29, 37, 146,147,148]. These same physiological adaptations could also explain the increased maximal dynamic strength performance (i.e. 1RM back squat), horizontal jumping performance, and vertical jumping performance noted after PJT compared to control conditions (ES = 0.79 to 1.47). Relatedly, improved dynamic maximal strength and jumping performance may provide significant support to the time-trial performance enhancement noted for WSA. Indeed, up to large associations (r = 0.40 to 0.70) have been reported between 20- and 50-m front crawl swimming performance and mean propulsive power in jump squat [16], leg extension strength [17], and horizontal jump distance [18]. Moreover, maximal-intensity short-duration efforts in WSA competitions (e.g. swimmers’ start platform jump; rowers’ stroke) may be a key element of success [150, 151]. Lower-limb fast-force production capabilities (e.g. jumping) may aid performance of such efforts [152,153,154], for example, increasing the distance per stroke-kick [54, 129, 130]. Improved jumping ability may be associated with an enhanced in-water kicking propulsive force generation capacity [155]. In some WSA (e.g. swimmers), the high- and maximal-intensity short-duration movements (e.g. jump-start; flip turns) represent one-third of the total time-trial performance [156]. This may suggest a meaningful relevance for training interventions aimed at improving WSA high- and maximal-intensity short-duration neuromuscular performance, with PJT offering several advantages to achieve this aim.

Body composition may have a significant impact on WSA performance [8, 22, 23]. Although PJT may induce adaptations in body composition [30, 32], our meta-analyses revealed no effect of PJT compared to control conditions on WSA body mass, fat mass, and thigh girth. Increased power generation capabilities (e.g. greater jumping ability) and unchanged body mass may facilitate generating greater relative power (i.e. W/kg−1), an important determinant of performance in WSA [7,8,9]. Therefore, unchanged body mass may have indirectly contributed to improved time-trial performance, by allowing greater relative power during key movements involving maximal-intensity short-duration efforts (e.g. tumbling-turn in swimming). Regarding thigh girth, its unchanged value (compared to control conditions) may suggest a lack of a hypertrophy-related effect derived from PJT. Nonetheless, PJT may have a skeletal muscle hypertrophy effect [30, 32]. However, the studies that provided data for thigh girth in the current meta-analysis involved PJT interventions lasting only 4 up to 8 weeks, which probably was insufficient to detect any hypothetical hypertrophy effect [157]. Regarding body fat, unchanged values may be expected, due to the relatively low amount of total energy expenditure derived from traditional PJT sessions [30, 42, 72]. Nevertheless, the recent literature suggests that some PJT exercise variations (e.g. greater jumping rate, lower inter-repetition, and inter-set rest) may induce considerable cardiorespiratory responses [158, 159], providing potential for long-term body fat reductions.

Participants' sex (i.e. males vs. females), type of sport (i.e. swimmers vs. water polo athletes), programme duration (i.e. < 8 weeks vs. ≥ 8 weeks), total PJT sessions (i.e. < 18 sessions vs. ≥ 18 sessions), and training period (i.e. pre-season vs. in-season) had no significant moderating effects on time-trial performance changes after PJT. Similar findings were noted for CMJ height, as participants’ sex, type of sport, PJT programme duration, and training period had no significant moderating effects on CMJ changes after PJT. On the contrary, greater (p = 0.043) CMJ improvements were noted after ≥ 8 weeks of PJT compared to < 8 weeks of PJT. Previous studies [37, 160, 161] noted that the moderating effects of factors such as participants’ sex, type of sport, PJT duration, and total number of sessions can be outcome-specific. The sex of participants seems to affect jump performance changes after PJT, with no effect on strength or sprint performances [37, 160, 161]. The type of sport practiced seems to moderate sprint, but not strength or jump performances after PJT [37, 160, 161]. Total training duration seems to moderate changes in jump and sprint, but not strength [37, 160, 161]. Therefore, while time trial and CMJ height after PJT seem to be improved in WSA irrespective of participants’ sex or type of water sport, with a minimal effective duration of < 8 weeks and 18–22 total sessions, and with long-term intervention approaches probably increasing chances of significant improvements [162, 163], current novel findings should be considered cautiously. Indeed, due to a limited number of studies, moderator analyses for outcomes other than time trial and CMJ height were precluded in our meta-analysis. Regarding the period of the season, making a comparison with previous studies is difficult. This is because studies addressing the effects of isolated PJT interventions on athletes’ PF according to the season’s period are lacking. However, a previous systematic review found that multimodal interventions, particularly those involving jumping exercises among others, were similarly effective for injury prevention in youth team athletes, regardless of the period of the season [164]. Current findings related to WSA recommend regular implementation of PJT during the season, as commonly occurs in real settings in other sports [165,166,167].

Limitations

Despite our systematic review with meta-analysis making a novel and significant contribution to the existing literature and highlighting the benefits of PJT to improve measures of PF and SSP components in WSA, there are some limitations that should be mentioned and discussed. Firstly, a reduced number of studies were available for some outcomes such as in-water agility (i.e. three studies). Secondly, a reduced number of participants (median n = 11) were included in most studies. Thirdly, the risk of publication bias analysis was precluded (aside from time-trial performance and countermovement jump) as less than 10 studies were available for most comparisons. Fourthly, the descriptive information provided in some studies was sub-optimal. For example, most of the included studies did not report if the WSA had previous systematic experience with PJT. Further, the intensity of PJT interventions was not detailed in many studies and when reported in some, it was only partially described. Overall, all the included studies did not report one or more key descriptive characteristics of PJT intervention. Fifthly, though I2 showed low-to-moderate heterogeneity for most comparisons, two exceptions, horizontal jump distance and countermovement jump height, obtained a high heterogeneity (I2 = 75.9 to 84.9%). A subgroup and sensitivity analysis would be warranted in such cases. However, due to the extension of this work (e.g. ten outcomes being analysed and discussed; additional analyses; methodological quality assessment), these would be better approached in another publication, focused on secondary analysis. Finally and according to GRADE, the certainty of evidence ranged from very low-to-low for most outcomes, reducing the confidence in the presented estimates.

Practical Applications and Future Lines for Research

Aside from a greater number of studies needed in this field, future studies should conduct a priori sample size power analysis [65] to recruit a sufficient number of participants, therefore increasing the robustness of their statistical power. Moreover, large randomized-controlled trials should be encouraged in future efforts to address the effects of PJT on measures of PF and SSP components in WSA, providing a proper report of key moderator factors of PJT, such as PJT intensity. Indeed, a sub-analysis of the training intensity factor was precluded due to the lack of a standard method to quantify intensity and the wide variety of approaches used across the included studies. In line with this, to the authors’ knowledge, the effects of PJT intensity on WSA adaptations are currently unknown. Moreover, ~ 35% of the studies conducted in WSA did not provide clear details around the intensity of the applied jump drills. Therefore, the high heterogeneity of the methods used to quantify PJT intensity in addition to the high number of studies (35%) that did not sufficiently report on it precluded any further consideration of this training factor. Nonetheless, a discussion related to PJT intensity has been addressed elsewhere [39, 40, 42, 72, 168,169,170].

The surface type can affect acute and long-term responses to PJT [29, 41, 171]. Only one out of the 26 studies included in our meta-analysis reported a specific water-based PJT intervention, with the remaining studies using a land-based PJT approach. Compared to water-based, land-based PJT induced greater improvements in lower-limb power [172]. Two reviews of the literature suggested that water-based PJT is as effective as land-based PJT to improve sprint, strength, and jump performance [173, 174]. However, the aforementioned reviews did not focus on WSA. Future studies should clarify the effects of water-based versus land-based PJT on PF and SSP outcomes in WSA.

Similar to the surface type, only 4 out of 26 of the included studies in this meta-analyses used a tapering approach. This is considered an important programming variable for PJT [175] and competitive performance [176,177,178], particularly after interventions involving a large volume-load of training, commonly occurring for WSA [7,8,9,10,11]. Moreover, some PJT interventions involved up to ~ 28,000 total jumps. Future studies may analyse the effects of different tapering strategies on the PF and SSP outcomes in WSA.

In some competitive contexts, a greater emphasis may be provided on PJT. For example in swimmers, greater time-trial improvement may be expected after PJT for distances completed in shorter length pools. Shorter pools would mean that the time-trial distance covered due to the jump-start and flip-turn (i.e. jump-like action) movements represents a greater proportion of the total distance. Such actions are key elements of success in swimming [150, 151] and are expected to improve with PJT, as suggested by current meta-analyses.

Conclusions

PJT is a more effective method to improve measures of PF and SSP in WSA compared to control conditions involving traditional sport-specific training only as well as alternative training interventions. This conclusion is derived from 26 articles of moderate-to-high methodological quality, low-to-moderate heterogeneity for most outcomes, and very low-to-moderate certainty of evidence according to GRADE.