The association between training load and physical development in professional male youth soccer players: a systematic review

Objectives 1) To evaluate current physical performance tests used within professional male youth soccer; 2) to understand the relationship of these tests performance in relation to specific measures of external and internal training load (TL) to conclude if there is a subsequent change in test performance. Methods Relevant literature was searched using five electronic databases (PubMed Medline, SPORTDiscus, Web of Science, CINAHL and Scopus), with additional articles identified by the authors. Articles relating to TL and physical development assessment within professional male youth soccer players were evaluated. Results Database searches yielded 5683 articles following removal of duplicates. After screening the titles, abstracts and full texts, 28 articles were identified. Both external TL (total distance, high speed distance, duration) and internal TL (rating of perceived exertion, training impulse) measures were found to be associated with improvements in physical test performance across both pre-season and in-season phases. Field-based testing was found to be sensitive to changes in physical performance for aerobic capacity, lower body power/strength and sprint performance. However, limited sensitivity to change was found when assessing player agility performance. Conclusion Future research in this area should look to enhance our understanding of the dose-response of TL with changes in fitness across different age groups in professional male youth soccer.


Introduction
Within professional male youth soccer, the training process is deemed central for a successful transition of players through playing pathways and development programmes. 1 For players to advance successfully they must be prepared for the next level of competition, whether that be from youth player to scholar, scholar to professional, or professional to first team regular, from a physical, mental, technical and sociological perspective. 2 Soccer coaches and support practitioners are responsible for both the design and delivery of player development programmes with the ultimate aim to produce a player worthy of a regular place within the senior team. This notion has led to significant investment across professional soccer clubs to enhance the resources available for both the recruitment and development of their youth players. 3 Whilst player development can be impacted by internal factors, such as genetics, 4 the structure of training programmes is within the control of the practitioner to appropriately manage in order to maximise the chance of success.
From a physical perspective, it is common for practitioners to monitor the training load (TL) undertaken by their players in order to understand both the external load and internal response of each individual player. 5 External load is defined as the physical work prescribed in the training plan, whereas internal load refers to the psychophysiological responses to the external load. 5 The quantification of external load is often monitored using micro-electro-mechanical (MEMS) devices containing a global positioning system (GPS) processor and inertial sensors to collect information of variables such as distances covered at different velocities, acceleration and deceleration efforts and estimated metabolic power. 6 The internal load is typically quantified using heart rate telemetry and subjective scales, such as the rating of perceived exertion (RPE) and wellness ratings. 6 Whilst assessment of the daily external load helps coaches understand whether the planned content matches that of the observed load, the internal load represents the stimulus for training induced adaptation. 7 In order to understand the long-term effects of accumulated TL, it is imperative that practitioners employ soccer-specific physical performance tests to understand how each player is responding to the development programme. Practitioners will often incorporate multiple testing points during an annual macrocycle to determine changes across different physical qualities. 8 Testing protocols are typically conducted within the field due to the practical limitations of laboratory-based testing across large numbers of players. In addition, it has been suggested that laboratory-based tests are not appropriate protocols for soccer, especially when assessing match stimulation protocols, and therefore their specific relation to soccer match performance may not be accurate. 9,10 Common tests used within youth soccer include the Yo-Yo intermittent test, 11 the 30-15 intermittent fitness test, 12 countermovement jump (CMJ), sprint testing (10m and 30m), repeated sprint ability test (RSA) 13 and agility testing. 14,15 The tests used in practice should be reliable, valid and sensitive to changes in fitness across longitudinal periods. 8 In order to determine the effectiveness of youth soccer player development programmes, it is important that practitioners understand the associations between measures of TL and physical performance tests. Whilst this association has recently been evaluated in senior professional soccer players, 13 there is limited information available to practitioners around which TL measures and physical tests are most suitable for youth player development. Therefore, the purpose of this systematic review was to: 1) To evaluate current physical performance tests used within professional male youth soccer; 2) to understand the relationship between test performance in relation to measures of external and internal training load (TL).

Literature search strategy
Articles were systematically identified via five electronic databases (PubMed, Medline, SPORTDiscus, Web of Science, CINAHL and Scopus) using the search strategy presented in Table 1. Terms within the search strategy (search strings 1, 2, 3 and 4) were connected with the 'OR' function. Three articles were identified from additional sources known to the authors. The search was conducted on 30 th April 2020 (registration no: 10.17605/ OSF.IO/DYP38), therefore any article published after this time was not included. The key terms and definitions used for the systematic review process are presented in Table 2.

Selection criteria
This review was conducted according to the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines 16 ( Figure 1). Articles were excluded from the initial search (title and abstract only) if (1) the participants were not male soccer players; (2) the article was not an original article in a peerreviewed scientific journal; (3) the article was not related to training load and fitness. Following this, from the remaining articles, the full texts were analysed, and articles were excluded if they fit the following criteria: (1) Participants were over the age of 23; (2) the participants were non-professional; (3) the article investigated other aspects of training (e.g. rehabilitation, nutrition); (4) the article did not report any training load measures; (5) the article did not report any physical development measures; (6) the article was unavailable in English.

Methodological quality assessment
Methodological quality was assessed using the risk of bias for included studies assessment tool and more specifically the NIH quality assessment tool for observational cohort and cross-sectional studies; 17 https://www.nhlbi.nih.gov/ health-topics/study-quality-assessment-tools). All the studies that were included were measured against the answers to fourteen questions (Table 3). Each item is scored as 1 = "Yes", and 0 = "No/unable to determine". Every article received a total score out of 14 which was then deemed 'poor', 'fair' or 'good' according to the assessment tool independently by two authors (J.R and J.M). These assessments were then used to provide an overall quality of the available evidence using the beforementioned ratings.

Data extraction and analysis
When searching each database, the lead author (JR) examined the article title, abstract and keywords in the first stage of screening according to the established inclusion and exclusion criteria. The process was cross-checked for accuracy by a second author (J.M.). The texts were examined to identify the terminologies employed in reference to the method used and for physical development by definition. The following data, where possible, were extracted from each article: •

Search findings and methodological quality
The electronic search yielded 9521 articles, with an additional 3 articles identified from an additional journal (Science and Medicine in Football). Following the removal of duplicates, 5683 articles remained for title and abstract screening. 5536 articles were removed based on initial study criteria ( Figure 1). 147 full-text articles were screened and 119 were removed based on the second level of criteria, leaving 28 articles included in the final review. Ratings from the NI quality assessment tool for each article are presented in Table 4, with scores ranging from 6-11. The overall quality of available evidence was rated as 'fair' by two authors (JR and JM) independently (see Table 3 for individual study ratings). Participant characteristics and study details are presented within Table 5.

Aerobic fitness
Eleven studies 18,20,23,24,26,29,34,35,37,39,42,43 used physical performance tests to detect changes in aerobic fitness across various testing points (Table 6). Four studies 26,29,39,42 used Yo-Yo intermittent testing, revealing improvements in test performance from the start of pre-season to the end of preseason 39,42 and late in-season, 29 with trivial changes in performance across a 5 day tournament. 26 Seven studies used both laboratory and field tests, including an incremental test derivative performed on a laboratory treadmill; 18,23,35,37 cycle ergometer 34 and fieldbased protocols. 20,43 Three studies found significant improvements in V˙O 2max from the start to the end of the pre-season phase. 23,34,35 Two studies 37,43 found significant improvements in lactate thresholds across a similar time period. Di Giminiani and Visca 20 found a ∼5.7% improvement in V˙O 2max over a two-season period. Akubat et al. 18 found no changes in aerobic fitness parameters across a 6-week pre-season phase.

Lower body power
Twelve studies assessed lower body power through CMJ performance using different technologies: a contact jump mat, 22,30,31,33,36,41,44,45 photocell system 19,29,42,45 and portable force plate 22,26 (Table 6). Six studies found a significant improvement in CMJ performance from the start of pre-season to the end of the in-season phase. 22,29,31,33,38,42,44 This improvement was also observed from the start to end of the pre-season phase. 29,42 Conversely, no significant change in CMJ performance was observed across a training microcycle, 26,45 mesocycle (4-9 weeks) 19,36 and full season. 30 In relation to strength-based measures, only two studies analysed seasonal physical changes. 22,44 The studies found small to moderate improvements in isometric mid-thigh pull (IMTP) peak force 22 and a significant increase in hamstring:quadtricep (H:Q) ratio peak torque values 41 from the start to the end of a season.

Sprint, agility and anaerobic capacity
The assessment of changes in sprint and agility capabilities was assessed in eight studies 19,22,33,[36][37][38]42,46 (Table 6). Three studies found significant improvements in 5 m, 36 15m 46 and 30m 38 sprint performance across the pre-season phase. Two studies 38,46 also found improvements in 10 m and 30 m sprint performance across a competitive season, with no change observed for 5 m performance. 46 No changes in agility performance were observed across both a full season 39 and 4-week training block. 19 However, Morris et al. 22 did observe significant improvements in arrowhead agility performance across a season across different peak height velocity (PHV) group players. Jastrzebski et al. 34 found significant improvements in anaerobic relative peak power at the end of the season compared to the start of pre-season using a 30-s Wingate test

External training load and changes in aerobic performance
Five studies 25,26,28,40,41 examined the association between external TL and changes in aerobic performance (Table 6). Two studies found significant correlations between total distance accumulated in both training and matches compared with improvements in aerobic capacity. 26, 40 Buchheit et al. 25 found moderate to large correlations between distance covered > 16.1 km.h −1 during matches and improvements in the VAM-Eval test across a 4-month period, particularly within the striker's group. There was no analysis directly related to the influence of match or training involvement. Significant correlations were observed across a 6-week training period between external load variables (total distance, distance covered > 17 km.h −1 and > 21 km.h −1 ) and improvements in maximal aerobic speed (MAS) scores. 41 Gil-Rey et al. 28 found large correlations between total training volume (i.e. duration) accumulated and improvements in the Université de Montreal Track Test

External training load and changes in neuromuscular performance
Two studies examined the association between external TL and changes in neuromuscular performance 28,45 (Table 6). Gil-Rey et al. 28 found trivial to small correlations between total training volume and both changes in CMJ and sprint (5 and 15 m) performance across a 9-week period. Similarly, Malone et al. 45 found no significant correlations between external TL variables (duration, total distance, average speed, distance covered > 19.8 km.h −1 ) and changes in CMJ performance across an in-season microcycle.

Internal training load and changes in aerobic performance
Four studies 18,27,32,36,47 examined the association between internal TL and changes in aerobic performance ( Table 6). Figueiredo et al. 47 found a large negative correlation between weekly mean internal TL variables (RPE and Edward's training impulse (TRIMP)) and improvements in Yo-Yo intermittent recovery test level 1 (IR1) performance across a pre-season phase. One study observed a significant large correlation between mean weekly individual TRIMP (iTRIMP) and improvements in velocity at 2 mmol.l −1 of blood lactate and during an incremental treadmill test across a 6-week period. 18 Los Arcos et al. 36 revealed moderate to large correlations between RPE derivatives (RPE respiratory (RPEres) and RPE muscular (RPEmus)) and improvements in submaximal running lactate values (12 km.h −1 , 13 km.h −1 and velocity at 3 mmol.l −1 of blood lactate) across a season. Conversely, Saidi et al. 32 found no significant correlations between RPE-based load and changes in Yo-Yo IR1 performance.

Internal training load measures and changes in neuromuscular performance
One study 36 evaluated the association between internal TL and changes in neuromuscular performance ( Table 6). The authors found moderate to large correlations between RPEres and Table 3. Questions within the quality assessment tool for observational cohort and cross-sectional studies as observed within NIH guidelines.

Question Details 1
Was the research question or objective in this paper clearly stated 2 Was the study population clearly specified and defined 3 Was the participation rate of eligible persons at least 50% 4 Were all the subjects selected or recruited from the same of similar populations (including the same time period). Were inclusion and exclusion criteria for being in the study prespecified and applied uniformly to all participants 5 Was a sample size justification, power description, or variance and effect estimates provided 6 For the analyses in this paper, were the exposure(s) of interest measured prior to the outcome(s) being measured 7 Was the timeframe sufficient so that one could reasonably expect to see an association between exposure and outcome if it existed 8 For exposures that can vary in amount of level, did the study examine different levels of the exposure related to the outcome (e.g. categories of exposure, or exposure measured as continuous variable)? 9 Were the exposure measures (independent variables) clearly defined, valid, reliable, and implemented consistently across all study participants 10 Was the exposure(s) assessed more than once over time 11 Were the outcome measures (dependent variables) clearly defined, valid, reliable, and implemented consistently across all study participants 12 Were the outcome assessors blinded to the exposure status of participants 13 Was loss to follow-up after baseline 20% or less 14 Were key potential confounding variables measured and adjusted statistically for their impact on the relationship between exposure(s) and outcome(s) improvements in 15 m sprint and CMJ with arm swing performance across a 9-week period. However, no correlations were found between internal TL (RPE, RPEres and RPEmus) variables and changes in CMJ or 15 m sprint performance. In addition, RPEmus was found to show no correlations with neuromuscular performance variables.

Discussion
The purpose of the present systematic review was to: 1) To evaluate current physical performance tests used within professional male youth soccer; 2) to understand the relationship of these tests performance in relation to specific measures of external and internal training load (TL) to conclude if there is a subsequent change in test performance. In terms of aerobic capacity evaluation, both field-based (e.g. Yo-Yo test derivatives) and laboratory-based (e.g. treadmill incremental test) tests were sensitive to changes across both pre-season and in-season phases. CMJ testing demonstrated limited sensitivity to detect acute changes in lower body power (e.g. across a microcycle), but with some evidence of sensitivity when evaluating across a full season. In addition, 10 m and 30 m sprint performance appeared sensitive to seasonal changes, whereas agility testing demonstrated limited sensitivity. In terms of the relationship between physical tests and TL measures, associations were found across tests with both external TL (total distance, high speed distance, accumulated duration) and internal TL (RPE, TRIMP) measures. Therefore, practitioners should be aware of some of the limitations in sensitivity of commonly used physical tests (e.g. CMJ and agility) when developing professional male youth soccer players. It is important that practitioners include a series of testing battery points during an annual macrocycle to determine changes within individual players across different physical qualities. 8 The present review revealed that the majority of physical tests were sensitive to change across the pre-season phase. Typically, the pre-season phase is designed to rebuild the fitness of players following the off-season phase. 48 The TL during the pre-season phase is higher than that observed during the in-season. 49 Therefore, it would seem logical that the phase with the highest TL (i.e. pre-season) would also demonstrate the most sensitivity in terms of physical test Table 4. Results of methodological quality assessment for included articles. Rating  Study  1  2  3  4  5  6  7  8  9  10  11  12  13  14  T otal   Akubat et al. 18  1  1  1  1  0  1  1  1  1  0  1  0  1  0  10 Fair Buchheit et al. 19 1 20 1 21 1

Question Number
Good Fitzpatrick et al. 22 1 Good Francioni et al. 23 1 Fair Gibson et al. 24 1 Fair Gil-Rey et al. 25 1 Good Hammami et al. 26 1 27 1 Fair Impellizzeri et al. 28 1 Fair Jastrzebski et al. 29 1 Fair Lopez-Segovia et al. 30 1 Fair Los Arcos et al. 31 1 Good Los Arcos et al. 32 1 Fair Malone et al. 33 1 Fair McMillian et al. 34 1 Fair Miranda et al. 35 1 Good Morris et al. 36 1 Poor Noon et al. 37 1 Fair Paul et al. 38 1 Good Perroni et al. 39 1 40 1 41 1 Fair Silwowski et al. 42 1 Poor Silwowski et al. 43  improvements. Changes in physical testing qualities were also observed from the start of pre-season to the end of the competitive season (e.g. Hammami et al.; 29 Morris et al. 36 ) These observed improvements are crucial for the physical development of players to successfully prepare them for the next level of competition within the development pathway. 2 It should be noted that limited data was available using regular in-season assessments of physical performance. This may be due to the reluctance of soccer coaches to incorporate physical fitness testing during this phase due to the competition demands. Of all the physical tests evaluated, it appeared that assessment of agility via a number of varied testing protocols provided limited results in terms of sensitivity to changes in TL. This may be due to the lower reliability and validity of field agility tests, possibly attributed to the complexity of the tests that incorporate both physical and perceptual aspects of speed. 50 For instance, in a recent review by Altmann et al., 50 the authors reported change of direction testing reliability ranging from 0.37 to 0.99 (ICC), suggesting large variance in the reliability of agility tests within soccer. Caution should also be made when using physical tests to detect acute changes in fitness, such as a weekly training microcycle, as this showed lower sensitivity compared to when assessing over a longitudinal period. 19,45 When assessing the long-term development of youth soccer players, with specific reference to enhancing physical qualities, it seems necessary to have a good understanding of the physiological and maturational characteristics that differ across the youth players' physical development pathway. 21 The articles within this review have limited information around the potential influence of maturation with the majority of studies not factoring this into their analysis models. As such, it would be interesting for future studies to analyse the effect of maturation as perhaps this may explain some of the physical improvements observed within youth soccer players throughout their long-term development programmes. For instance, previous research has suggested that factors such as maturation, hormones and musculoskeletal changes that incur when going through the development phase have direct impacts on the physical capabilities of youth players and that the training programmes which are implemented should account for these factors in order to optimally cater for development opportunities. 2 Overall, it should be concluded that both the effects of growth and maturation alongside the systematic training programmes are important in terms of physical performance development within junior athletes. The quantification of TL is now commonplace within professional male youth soccer, with practitioners routinely collecting both external and internal TL information on a daily basis. 6 The present review revealed that common GPS-derived measures of external TL (e.g. total distance, high speed distance) were associated with positive changes in fitness. Despite the plethora of GPS variables available to quantify external TL currently, the evidence around their association to changes in fitness in professional male youth soccer players was found lacking. Indeed, the beforementioned more 'traditional' external TL variables and their association with changes in fitness has also been recently observed in senior professional players. 13 In terms of internal TL variables, both RPE-based and heart rate TRIMP-based measures demonstrated the strongest association to changes in fitness. It would appear that both measures were associated with improvements in fitness across both the pre-season 47 and in-season 36,48 phases. Both RPEres and RPEmus were found to be associated with changes in fitness, 48 which could be a future direction for practitioners rather than using the standard RPE approach. Interestingly, there appeared to be more studies revealing associations with external TL variables compared to internal TL variables. One reason may be due to external TL data being 'easier' for practitioners to collect during training and matches. 6 Internal TL measures face practical issues, such as erroneous heart rate data 51 , which may dissuade practitioners from collecting and trusting this type of data, which is a limitation of the present literature available.

Limitations within the literature
Limitations of the present systematic review must be acknowledged. Only English language articles were considered, and only articles using professional male youth soccer players were included. It could be argued that the definition of a 'professional' youth soccer player can vary depending on the country of origin and standard of competition. In addition, limitations were found within consistency of the TL measures and physical fitness procedures used, thus limiting the ability to fully compare across articles. For example, there were differences observed in the threshold used for 'high speed distance' based variables (e.g. 17 and > 19.8 km.h −1 ). The limited available studies evaluating the association between internal TL and physical test performance must also be acknowledged and recommended for future study in this area.
Based on the present available literature, it's clear that further work is required in order to understand the association between training load and physical development in professional male youth soccer players. There were a number of studies that only used a short study duration to examine the response to load (e.g. 1 -6 weeks). Longitudinal data is required to understand the long-term athletic development in youth players across the different age groups. Of those that used longitudinal designs, often the physical testing points were too infrequent or the training load quantification was limited. With more teams now investing into the training process of youth players, it is hoped that future work in this area will comprehensively quantify the training load and physical response over longitudinal periods.

Conclusion
The present review was the first to provide a systematic evaluation of the association between training load and physical development using the current literature available within professional male youth soccer players. Both external TL (total distance, high speed distance, accumulated duration) and internal TL (RPE, TRIMP) measures were found to be associated with improvements in physical test performance across both pre-season and in-season phases. In addition, field-based testing was found to be sensitive to changes in physical qualities for aerobic capacity, lower body power and strength and sprint performance. However, limited sensitivity to change was found when assessing player agility performance. Future research in this area should look to enhance our understanding of the dose-response of TL (particularly internal TL) with changes in fitness across different age groups in professional male youth soccer. It would also be pertinent to examine whether enhancement of these physical quality improves a player's chance of making it as a senior professional player. It is also clear that there is very limited information which accurately monitors training load in conjunction with physical fitness changes in youth soccer and this should be explored in more detail in future work.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.