Effects and dose–response relationships of resistance training on physical performance in youth athletes: a systematic review and meta-analysis

Melanie Lesinski; Olaf Prieske; Urs Granacher

doi:10.1136/bjsports-2015-095497

Article Text

XML

Review

Effects and dose–response relationships of resistance training on physical performance in youth athletes: a systematic review and meta-analysis

Melanie Lesinski,
Olaf Prieske,
Urs Granacher

Division of Training and Movement Sciences, University of Potsdam, Research Focus Cognition Sciences, Potsdam, Germany

Correspondence to Melanie Lesinski, University of Potsdam, Research Focus Cognition Sciences, Division of Training and Movement Sciences, Am Neuen Palais 10, Building 12, Potsdam D-14469, Germany; mlesinsk{at}uni-potsdam.de

Abstract

Objectives To quantify age, sex, sport and training type-specific effects of resistance training on physical performance, and to characterise dose–response relationships of resistance training parameters that could maximise gains in physical performance in youth athletes.

Design Systematic review and meta-analysis of intervention studies.

Data sources Studies were identified by systematic literature search in the databases PubMed and Web of Science (1985–2015). Weighted mean standardised mean differences (SMD_wm) were calculated using random-effects models.

Eligibility criteria for selecting studies Only studies with an active control group were included if these investigated the effects of resistance training in youth athletes (6–18 years) and tested at least one physical performance measure.

Results 43 studies met the inclusion criteria. Our analyses revealed moderate effects of resistance training on muscle strength and vertical jump performance (SMD_wm 0.8–1.09), and small effects on linear sprint, agility and sport-specific performance (SMD_wm 0.58–0.75). Effects were moderated by sex and resistance training type. Independently computed dose–response relationships for resistance training parameters revealed that a training period of >23 weeks, 5 sets/exercise, 6–8 repetitions/set, a training intensity of 80–89% of 1 repetition maximum (RM), and 3–4 min rest between sets were most effective to improve muscle strength (SMD_wm 2.09–3.40).

Summary/conclusions Resistance training is an effective method to enhance muscle strength and jump performance in youth athletes, moderated by sex and resistance training type. Dose–response relationships for key training parameters indicate that youth coaches should primarily implement resistance training programmes with fewer repetitions and higher intensities to improve physical performance measures of youth athletes.

Strength
Weight lifting
Children
Adolescent
Physical fitness

This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/

https://doi.org/10.1136/bjsports-2015-095497

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Resistance training (RT) is a safe and effective way to improve proxies of physical performance in healthy children and adolescents when appropriately prescribed and supervised.1–4 Several meta-analyses have shown that RT has the potential to improve muscle strength and motor skills (eg, jump performance) in children and adolescents.1 ,5–7 However, youth athletes have different training capacities, adherence, physical demands of activities, physical conditions and injury risks compared with their non-athlete peers; so the generalisability of previous research on youth athletes is uncertain.8–10

To the best of our knowledge, there is only one meta-analysis available that examined the effects of RT on one specific proxy of physical performance (ie, jump performance) and in one age group (ie, youth aged 13–18 years).11 It is reasonable to hypothesise that factors such as age, sex and sport may influence the effects of RT. Therefore, a systematic review with meta-analysis is needed to aggregate findings from the literature in terms of age, sex and sport-specific effects of RT on additional physical performance measures (eg, muscle strength, linear sprint performance, agility, sport-specific performance) in youth athletes.

There is also little evidence-based information available regarding how to appropriately prescribe exercise to optimise training effects and avoid overprescription or underprescription of RT in youth athletes.12 The available guidelines for RT prescription are primarily based on expert opinion, and usually transfer study findings from the general population (ie, healthy untrained children and adolescents) to youth athletes. This is important because the optimal dose to elicit a desired effect is likely to be different for trained and untrained youth.13

Therefore, the objectives of this systematic literature review and meta-analysis were (1) to analyse the effectiveness of RT on proxies of physical performance in youth athletes by considering potential moderator variables, including age, sex, sport and the type of RT, and (2) to characterise dose–response relationships of RT parameters (eg, training period, training frequency) by quantitative analyses of intervention studies in youth athletes. We hypothesised that (1) RT would have a positive effect on proxies of physical performance in youth athletes, and (2) the effects would be moderated by age, sex, sport and RT type.

Methods

Our meta-analysis was conducted in accordance with the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).14

Literature search

We performed a computerised systematic literature search in the databases PubMed and Web of Science.

The following Boolean search syntax was used: (‘strength training’ OR ‘resistance training’ OR ‘weight training’ OR ‘power training’ OR ‘plyometric training’ OR ‘complex training’ OR ‘weight-bearing exercise’) AND (athlete OR elite OR trained OR sport) AND (children OR adolescent OR youth OR puberty OR kids OR teens OR girls OR boys). The search was limited to: full-text availability, publication dates: 01/01/1975 to 07/31/2015, ages: 6–13; 13–18 years, and languages: English, German. The reference list of each included study and relevant review article1 ,4–6 ,11 ,15–19 was screened for title to identify any additional suitable studies for inclusion in our review.

Selection criteria

Based on the defined inclusion and exclusion criteria (table 1), two independent reviewers (ML and OP) screened potentially relevant articles by analysing titles, abstracts and full texts of the respective articles to elucidate their eligibility. In case ML and OP did not reach an agreement concerning inclusion of an article, UG was contacted.

View this table:

Table 1

Selection criteria

Coding of studies

Each study was coded for certain variables listed in table 2. Our analyses focused on different outcome categories. If studies reported multiple variables within one of these outcome categories, only one representative outcome variable was included in the analyses. The variable with the highest priority for each outcome is mentioned in table 2.

View this table:

Table 2

Study coding

If a study solely used other tests, we included those tests in our quantitative analyses that were most similar with regard to the ones described above in terms of their temporal/ spatial structure.

Further, we coded RT according to the following training parameters: training period, training frequency, and training volume (ie, number of sets per exercise, number of repetitions per set), training intensity, temporal distribution of muscle action modes per repetition, and rest (ie, rest between sets and repetitions). Training parameters were categorised according to common classifications of RT protocols.21 If a study reported exercise progression over the training period, the mean number of sets per exercise, repetitions per sets, rest between sets and training intensity were computed.

To obtain sufficient statistical power to calculate dose–response relationships, we summarised RT types as conventional RT (ie, machine based, free weights, combined machine based and free weights, functional training) and plyometric training (ie, jumping). As it is not possible to classify complex training as either conventional RT nor plyometric training,22 we excluded these studies23–27 from dose–response analyses. Our dose–response analyses were computed independent of age, sex and sport.

Assessment of risk of bias

The Physiotherapy Evidence Database (PEDro) scale was used to quantify the risk of bias in eligible studies and to provide information on the general methodological quality of studies. The PEDro scale rates internal study validity and the presence of statistical replicable information on a scale from 0 (high risk of bias) to 10 (low risk of bias) with ≥6 representing a cut-off score for studies with low risk of bias.28

Statistical analyses

To determine the effectiveness of RT on proxies of physical performance and to establish dose–response relationships of RT in youth athletes, we computed between-subject standardised mean differences (SMD=(mean postvalue intervention group−mean postvalue control group)/pooled standard deviation). We adjusted the SMD for the respective sample size by using the term (1−(3/(4N-9))).29 Our meta-analysis on categoric variables was computed using Review Manager V.5.3.4 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2008). Included studies were weighted according to the magnitude of the respective SE using a random-effects model.

At least two RT intervention groups had to be included to calculate weighted mean SMDs, hereafter refered to as SMD_wm, for each performance category.30 We used Review Manager for subgroup analyses: computing a weight for each subgroup, aggregating SMD_wm values of specific subgroups, comparing subgroup effect sizes with respect to differences in intervention effects across subgroups.31 To improve readability, we reported positive SMDs if superiority of RT compared with active control was found. Heterogeneity was assessed using I² and χ² statistics.

Owing to a low number of studies in each physical performance outcome category that completely reported information on the applied RT parameters, metaregression was precluded.30 According to a scale for determining the magnitude of effect sizes in strength training research for individuals who have been consistently training for 1–5 years,32 we interpreted SMD_wm as: trivial (<0.35); small (0.35–0.79); moderate (0.80–1.50); large (≥1.50). The level of significance was set at p<0.05.

Results

Study characteristics

A total of 576 potentially relevant studies were identified in the electronic database search (figure 1). Finally, 43 studies remained for the quantitative analyses. A total of 1558 youth athletes participated, and of these, 891 received RT in 62 RT intervention groups. The sample size of the RT intervention groups ranged from 5 to 54 participants (table 3).

View this table:

Table 3

Included studies examining the effects of resistance training in youth athletes

Figure 1

Flow chart illustrating the different phases of the search and study selection.

There were 13 studies (21 RT intervention groups) that included children, and 29 studies (36 RT intervention groups) that included adolescents. In terms of biological maturation, only 15 studies reported Tanner stages. Three (5 RT intervention groups) of those studies examined prepubertal and 12 (15 RT intervention groups) postpubertal/pubertal youth athletes. Thirty studies (44 RT intervention groups) included boys only, whereas 4 studies (4 RT intervention groups) included girls only.

Youth athletes were recruited from team sports (soccer (20 studies; 34 RT intervention groups), basketball (9 studies; 11 RT intervention groups), baseball (3 studies; 5 RT intervention groups), handball (3 studies; 3 RT intervention groups), tennis (2 studies; 3 RT intervention groups), volleyball (1 study; 1 RT intervention group)), and strength-dominated sports (swimming (3 studies; 3 RT intervention groups), track and field (1 study, 1 RT intervention group)). No included study investigated youth athletes recruited from martial arts or technical/acrobatic sports.

Regarding the type of RT, 4 studies performed RT using machines, 4 studies using free weights, 4 studies using both machines and free weights, 5 studies performed functional RT, 5 studies performed complex training, and 19 studies applied plyometric training. Classification of studies was not always feasible due to missing information or group heterogeneity.

The RT interventions lasted between 4 and 80 weeks, with training frequencies ranging from 1 to 3 sessions per week, 1–8 sets per exercise, 4–15 repetitions per set, and 20–220 s of rest between sets. Training intensity ranged from 35% to 88% of the 1 repetition maximum (RM). Training parameters (eg, temporal distribution of muscle action modes per repetition, and rest in-between repetitions) which have gained attention in the literature71 were not quantified due to insufficient data.

A median PEDro score of 4 (95% CI 4 to 5) was detected and only 4 out of 43 studies reached the predetermined cut-off value of ≥6, which can be interpreted as an overall high risk of bias of the included studies (table 3).

Effectiveness of RT

Table 4 shows the overall as well as age, sex, sport and training type-specific effects of RT on measures of muscle strength, vertical jump and linear sprint performance, agility and sport-specific performance.

View this table:

Table 4

Overall as well as age, sex, sport and training type-specific effects of resistance training in youth athletes

There were moderate effects of RT on measures of muscle strength (SMD_wm=1.09; I²=81%; χ²=114.24; df=22; p<0.001; figure 2) and vertical jump performance (SMD_wm=0.80; I²=67%; χ²=137.47; df=46; p<0.001; figure 3), while there were small effects for linear sprint performance (SMD_wm=0.58; I²=41%; χ²=55.74; df=33; p<0.01; figure 4), agility (SMD_wm=0.68; I²=50%; χ²=48.19; df=24; p<0.01; figure 5) and sport-specific performance (SMD_wm=0.75; I²=62%; χ²=67.81; df=26; p<0.001; figure 6). By considering only the four studies with high quality (ie, low risk of bias), RT had moderate effects on measures of muscle strength (SMD=1.07; 1 study), vertical jump (SMD_wm=0.89; 3 studies) and linear sprint performance (SMD_wm=1.19; 2 studies); small effects on agility (SMD=0.28; 1 study); and large effects on sport-specific performance (SMD_wm=1.73; 2 studies).

Figure 2

Effects of resistance training (experimental) versus active control on measures of muscle strength (IV, inverse variance).

Figure 3

Effects of resistance training (experimental) versus active control on measures of vertical jump performance (IV, inverse variance).

Figure 4

Effects of resistance training (experimental) versus active control on measures of linear sprint performance (IV, inverse variance).

Figure 5

Effects of resistance training (experimental) versus active control on agility (IV, inverse variance).

Figure 6

Effects of resistance training (experimental) versus active control on proxies of sport-specific performance (IV, inverse variance).

There was no statistically significant effect of chronological and/or biological age on any proxy of physical performance. However, a tendency (p=0.05) towards larger RT effects were found for proxies of sport-specific performance in adolescents (SMD_wm=1.03) compared with children (SMD_wm=0.50; table 4). Subgroup analyses indicated that RT produced significantly larger effects (p<0.05) on proxies of sport-specific performance in girls (SMD_wm=1.81) compared with boys (SMD_wm=0.72; table 4). Given that most included studies (n=38) examined participants competing in team sports, our subgroup analyses regarding the moderator variable ‘sport’ is limited and did not show any significant subgroup differences (table 4). Subgroup analyses demonstrated that different training types of RT produced significantly different gains in muscle strength (p<0.001), agility (p<0.05) and sport-specific performance (p<0.05). Free weight RT showed the largest effects on muscle strength and agility, while for sport-specific performance, complex training produced the largest effects (table 4).

Dose–response relationships of RT

Training period

There was a significant difference for the effects of conventional RT on measures of muscle strength (p<0.001), vertical jump height (p<0.05) and agility (p<0.001; figure 7). The dose–response curves indicated that long lasting conventional RT (>23 training weeks) resulted in more pronounced improvements in measures of muscle strength (SMD_wm=3.40) and agility (SMD_wm=1.31), as compared with shorter training periods (<23 weeks). In terms of vertical jump height, a training period of 9–12 weeks appeared to be the most effective (SMD_wm=1.20).

Figure 7

Dose–response relationships of the parameter ‘training period’ on measures of muscle strength, vertical jump and linear sprint performance, agility, and sport-specific performance. Each filled grey circle illustrates between-subject SMD per single study with active control. Filled black triangles represent weighted mean SMD of all studies. NA, not applicable; SGA, subgroup analyses; SMD, standardised mean difference.

Training frequency

There were no significant differences between the observed training frequencies (ie, 1, 2, 3 times per week) for RT as well as plyometric training (figure 8).

Figure 8

Dose–response relationships of the parameter ‘training frequency’ on measures of muscle strength, vertical jump and linear sprint performance, agility, and sport-specific performance. Each filled grey circle illustrates between-subject SMD per single study with active control. Filled black triangles represent weighted mean SMD of all studies. NA, not applicable; SGA, subgroup analyses; SMD, standardised mean difference.

Training intensity

There was a significant difference with regard to the effects of conventional RT on measures of muscle strength (p<0.01; figure 9). High-intensity conventional RT (ie, 80–89% of 1 RM) resulted in more pronounced improvements in muscle strength (SMD_wm=2.52) compared with lower training intensities (ie, 30–39%, 40–49%, 50–59%, 60–69%, 70–79% of the 1 RM).

Figure 9

Dose–response relationships of the parameter ‘training intensity’ on measures of muscle strength, vertical jump and linear sprint performance, agility, and sport-specific performance. Each filled grey circle illustrates between-subject SMD per single study with active control. Filled black triangles represent weighted mean SMD of all studies. NA, not applicable; SGA, subgroup analyses; SMD, standardised mean difference; RM, repetition maximum.

Training volume (number of sets per exercise)

There was a significant difference with regard to the effects of conventional RT on muscle strength (p<0.01), and a tendency towards significance for measures of vertical jump performance (p=0.06; figure 10). Five sets per exercise resulted in more pronounced improvements in muscle strength (SMD_wm=2.76) compared with fewer sets. Three sets per exercise tended to be more effective in improving vertical jump performance (SMD_wm=1.19), as compared with four or five sets per exercise.

Figure 10

Dose–response relationships of the parameter ‘sets per exercise’ on measures of muscle strength, vertical jump and linear sprint performance, agility, and sport-specific performance. Each filled grey circle illustrates between-subject SMD per single study with active control. Filled black triangles represent weighted mean SMD of all studies. NA, not applicable; SGA, subgroup analyses; SMD, standardised mean difference.

For plyometric training, there was a tendency towards larger training-related effects on measures of muscle strength (p=0.09), linear sprint performance (p=0.07), as well as sport-specific performance (p=0.05) depending on the number of sets per exercise. Four sets per exercise revealed the largest effects for measures of muscle strength (SMD_wm=0.79) and sport-specific performance (SMD_wm=1.84), while three or four sets appear to be most effective for improving linear sprint performance (SMD_wm=0.95).

Training volume (number of repetitions per set)

There was a significant difference in terms of the effects of conventional RT on measures of muscle strength (p<0.05; figure 11). Six to eight repetitions per set produced the largest effects on muscle strength (SMD_wm=2.42). For plyometric training, there was a tendency towards significance for proxies of sport-specific performance (p=0.05). Six to 8 repetitions per set were less effective (SMD_wm=0.15), while 3–5 and 9–12 repetitions per set produced similar effects (SMD_wm=0.89 and 0.93).

Figure 11

Dose–response relationships of the parameter ‘repetitions per set’ on measures of muscle strength, vertical jump and linear sprint performance, agility, and sport-specific performance. Each filled grey circle illustrates between-subject SMD per single study with active control. Filled black triangles represent weighted mean SMD of all studies. NA, not applicable; SGA, subgroup analyses; SMD, standardised mean difference.

Rest between sets

There was a significant difference for the effects of conventional RT on measures of muscle strength (p<0.05; figure 12). Three to 4 min of rest between sets resulted in more pronounced improvements in measures of muscle strength (SMD_wm=2.09), as compared with shorter durations of rest.

Figure 12

Dose–response relationships of the parameter ‘rest between sets’ on measures of muscle strength, vertical jump and linear sprint performance, agility, and sport-specific performance. Each filled grey circle illustrates between-subject SMD per single study with active control. Filled black triangles represent weighted mean SMD of all studies. NA, not applicable; SGA, subgroup analyses; SMD, standardised mean difference.

Discussion

This systematic review with meta-analysis examined the general effects as well as the age, sex, sport and training type-specific impact of RT on proxies of physical performance in healthy young athletes. In addition, dose–response relationships of RT parameters were independently computed. The main findings were: (1) RT has moderate effects on muscle strength as well as on vertical jump performance, and small effects on linear sprint, agility and sport-specific performance in young athletes, (2) the effects of RT were moderated by the variables sex and RT type, (3) most effective conventional RT programmes to improve measures of muscle strength in healthy young athletes comprised training periods of more than 23 weeks, 5 sets per exercise, 6–8 repetition per set, a training intensity of 80–89% of the 1 RM, and 3–4 min of rest between sets.

Effects of RT on physical performance in youth athletes

In general, RT is an effective way to improve proxies of physical performance in youth athletes, and our findings support recently published literature.4 ,17 ,72 ,73 We found that the main effects of RT on measures of muscle strength and vertical jump performance were moderate in magnitude, with small effects for secondary outcomes, including linear sprint performance, agility and sport-specific performance (eg, throwing velocity). The lower RT effects on secondary outcomes might be explained by the complex nature of these qualities, with various determinants contributing to the performance level. For instance, agility depends on perceptual factors and decision-making as well as on changes in direction of speed, which is again influenced by movement technique, leg muscle quality and straight sprinting speed.74 Thus, muscle strength appears to be only one of several factors contributing to agility.

We recommend the incorporation of RT as an important part of youth athletes’ regular training routine to enhance muscle strength and jump performance.

How age, sex, sport and training type moderate RT effects

Age-specific effects of RT in youth athletes

Biological maturity is related to chronological age, and has a major impact on physical performance in youth athletes.75 However, unlike age, growth and maturation are not linear factors.76 ,77 There is often a discrepancy between chronological age and biological maturity among youth athletes.4 ,16 ,78

We found no significant differences in effect sizes for any proxy of physical performance between prepubertal and postpubertal athletes. Similarly, we did not find significant differences for the effects of RT on any physical performance measure with respect to the moderator variable ‘chronological age’ (table 4). Merely, a tendency (p=0.05) towards higher sport-specific performance gains following RT in adolescents, compared with children, was identified.

Although a minimum age has been defined at which children are mentally and physically ready to comply with coaching instructions,4 our subgroup analyses regarding biological and chronological age suggest that youth athletes may benefit to the same extent from RT, irrespective of age. However, it is important to note that most studies did not report the biological maturity status of the participants. Therefore, more research is needed to elucidate biological age-specific RT effects on physical performance in youth athletes and to verify our preliminary findings.

Sex-specific effects of RT in youth athletes

Previous research on the effects of RT on proxies of physical performance in youth athletes has primarily focused on boys. However, findings from male youth athletes can only partially be transferred to female youth athletes because the physiology of boys and girls (eg, hormonal status during puberty) varies. We found that male and female youth athletes show similar RT-related gains in muscle strength and vertical jump performance, but girls had significantly larger training-induced improvements in sport-specific performance (SMD_wm=1.81) compared with boys (SMD_wm=0.72). This suggests preliminary evidence that the RT trainability of female adolescent athletes may be at least similar or even higher compared with males. Given that girls’ and boys’ physiology changes differently with age and maturation,76 ,77 sex-specific effects of RT in youth athletes should be investigated with respect to biological maturity. Owing to an insufficient number of studies that examined female youth athletes and reported their biological maturity status, we were not able to include ‘biological maturity’ as a moderator variable in our subgroup analyses. We consider our sex-specific findings preliminary because these are based on five studies only investigating female youth athletes. More research is needed to elucidate sex-specific RT effects on physical performance in youth athletes and to verify our preliminary findings.

Sport-specific effects of RT in youth athletes

The effects of RT in elite adult athletes may be specifically moderated by the respective athlete profile of the sport performed.79 ,80 Whether this is also the case in youth athletes remains unresolved. Given that most included studies (n=38) investigated young athletes competing in team sports, our analyses with regard to the moderator variable ‘sport’ was limited and did not reveal any significant differences between sports disciplines (table 4). Therefore, further research has to be conducted to examine if youth athletes respond differently to RT programmes as per the sport practiced.

Training type-specific effects of RT in youth athletes

Various types of RT have been reported (eg, machine-based RT, free weight RT and functional RT). Each of these types has specific benefits and limitations.20 ,73 Machine-based RT may represent a safe environment for young athletes when supervision cannot be ensured, whereas supervised RT using free weights allows full range of motion that better mimics sports-specific movements.20 ,73 We found that RT programmes using free weights were most effective to enhance muscular strength and agility. In addition, complex training produced the largest effect sizes if the goal was to improve sport-specific performance. Therefore, the choice of RT types should be variable and based on the exercise goal (eg, enhancing muscle strength or sport-specific performance).

Dose–response relationships of RT in youth athletes

Planning and designing RT programmes is a complex process that requires sophisticated manipulation of different training parameters. Owing to a lack of evidence-based information on dose–response relationships following RT in youth athletes, it is quite common for established and effective RT protocols for healthy untrained children and adolescents to be transferred to youth athletes. However, this may hinder to fully recruit the adaptative potential of young athletes because the optimal dose to elicit the desired effect appears to be different in trained compared with untrained youth.13 Owing to the observed limitations regarding female youth athletes and biological maturation status in the present meta-analysis, the dose–response relationships of RT in youth athletes were determined irrespective of sex and maturity.

In general, the specific configuration of RT parameters determines the underlying training stimulus and thus, the desired physiological adaptations. However, significant effects were predominantly identified for conventional RT parameters for measures of muscle strength. Therefore, it appears that gains in muscular strength may be more sensitive to the applied training parameters of the conventional RT programmes, as compared with the secondary performance outcomes (eg, linear sprint performance, agility, sport-specific performance).