sports medicine Volume 40 Issue 4 April 2010

Sports Med 2010; 40 (4): 271-283 0112-1642/10/0004-0271/$49.95/0

CURRENT OPINION

ª 2010 Adis Data Information BV. All rights reserved.

Expert Performance in Sport and the Dynamics of Talent Development Elissa Phillips,1,2,3 Keith Davids,2 Ian Renshaw2 and Marc Portus3 1 Biomechanics and Performance Analysis, Australian Institute of Sport, Canberra, Australian Capital Territory, Australia 2 School of Human Movement Studies, Queensland University of Technology, Brisbane, Queensland, Australia 3 Sport Science Sport Medicine Unit, Cricket Australia Centre of Excellence, Brisbane, Queensland, Australia

Abstract

Research on expertise, talent identification and development has tended to be mono-disciplinary, typically adopting genocentric or environmentalist positions, with an overriding focus on operational issues. In this paper, the validity of dualist positions on sport expertise is evaluated. It is argued that, to advance understanding of expertise and talent development, a shift towards a multidisciplinary and integrative science focus is necessary, along with the development of a comprehensive multidisciplinary theoretical rationale. Here we elucidate dynamical systems theory as a multidisciplinary theoretical rationale for capturing how multiple interacting constraints can shape the development of expert performers. This approach suggests that talent development programmes should eschew the notion of common optimal performance models, emphasize the individual nature of pathways to expertise, and identify the range of interacting constraints that impinge on performance potential of individual athletes, rather than evaluating current performance on physical tests referenced to group norms.

Research on expertise and talent identification and development has been dominated by opposing genocentric or environmentalist positions.[1] Recently, Dunwoody[2] has challenged psychologists to look beyond the individual, suggesting researchers have neglected the role of the environment and focused on organism structure and process in isolation. Although there is a need to move away from the cognitive psychology asymmetric-organismic focus,[2,3] the tendency to overemphasize the role of environmental constraints on expertise acquisition also needs to be eschewed. For example, Ericsson and colleagues[4,5] emphasized an environmental perspective, ad-

vocating that expertise is acquired as performers specialize at an early age and engage in deliberate practice.[4] Other investigators have associated the hereditary nature of physiological, anthropometric and psychological characteristics with elite performance in sports.[6] For example, early anthropometric research on Olympic athletes advocated a close relationship between physical characteristics and specific Olympic events.[7] This line of evidence has been somewhat over-interpreted, leading to the questionable practice of anthropometric profiling of adolescents to identify potential for early specialization in a sport. Consequently, anthropometric profiling of physical

272

dispositions has tended to skew the rationale for talent identification models in sport, despite a lack of supportive evidence and the unstable nature of anthropometric and physical parameters during adolescence.[8,9] Recently, there have been proposals to integrate polar perspectives on sports performance and expertise into a multidisciplinary approach, to enhance understanding of the athleteenvironment relationship as exemplifying a complex, dynamical system. In this article we identify limitations of existing research on expert performance and talent development, and provide evidence-based arguments for adopting a multidisciplinary research perspective for the comprehensive study of sport expertise. To support our argument we review and evaluate extant literature on expertise in psychology and sport science. The search reference terms used included ‘talent development’, ‘talent identification’ and ‘sport expertise’ in PubMed, SportDiscus and Ovid databases. A major reason for adopting polarized, monodisciplinary positions on the acquisition of sport expertise has been the absence of a powerful multidisciplinary theory to act as an integrative conceptual framework. Although some models have advocated a multidisciplinary approach to talent identification and development, such as Simonton’s[10] model of talent as a multiplicative, dynamic process and the Differentiated Model of Giftedness and Talent,[11] these approaches tend to be operational and propositional in nature. These models have not provided a detailed, explanatory theoretical rationale underpinning a dynamic and multidimensional basis for expertise and how it may support the process of identifying and developing talent. One suitable theoretical approach to the study of performance, expertise and talent development in sport conceptualizes the athlete as a complex, dynamical system. In such systems there is great potential for interactions between system components and the environment, often leading to rich and unique patterns of behaviour. To date, studies of complex system behaviour in sport include match play and behaviour during competitions,[12,13] decision-making,[14,15] motor learning,[16,17] coordination,[18,19] human gait and ª 2010 Adis Data Information BV. All rights reserved.

Phillips et al.

injury,[20,21] and medicine in sports.[22] In this article, relevant theoretical concepts and insights from dynamical systems theory, complex systems theory, ecological psychology and evolutionary sciences are identified and their implications for the acquisition of sports expertise and talent development are discussed. 1. Traditional Expertise Approaches As noted, much research has focused on the role of environmental constraints in expertise acquisition, including participation in play and practice activities and the role of family and environmental contexts.[23,24] For example, the expert performance approach of Ericsson and Smith[25] explored the contribution to expertise of specific practice environments. Ericsson and colleagues’[5] deliberate practice approach has highlighted the importance of structured activities involving goal-directed skill learning, which requires effort and concentration. It was estimated that experts spend typically ~10 years or 10 000 hours in deliberate practice to attain exceptional performance.[26] The unidimensional nature of the deliberate practice approach led Starkes[27] to label it a ‘very environmentalist’ theory of expertise acquisition, while others (e.g. Arau´jo[28]) have proposed it has an organismic bias. Additionally, researchers studying deliberate practice in sport have encountered some incongruities with the tenets of the main theory. For example, in contrast to previous findings, many athletes tend to find appropriate practice and training enjoyable and motivating across all development stages.[29] Moreover, early specialization has not been found to be essential for acquisition of expert sport skills in adulthood.[30-32] Time spent in sport-specific training does discriminate between experts and non-experts in some sports, although the relationship between practice and performance is non-linear.[23] Further, retrospective analysis of historical data on time spent in practice was unable to clarify differences between experts and novices based on microstructure of practice activities.[33] For example, Weissensteiner et al.[34] examined the practice histories of under 15, under 20 and adult cohorts Sports Med 2010; 40 (4)

Expert Performance and Talent Development

of skilled and less skilled cricket batsmen and found that hours spent in practice explained only a small proportion of variance in the development of anticipation skills. A particular methodological concern in this approach is the proliferation of ‘weaker’ data from a plethora of cross-sectional studies, relative to the lack of longitudinal research. More evidence on putative benefits of deliberate practice is needed, following cohorts of young athletes through developmental pathways. To summarize, although numerous hours of training are necessary for success in sport at the elite level, research suggests that attainment of an expert level is not always accomplished by engaging in deliberate practice alone. In another polarized approach to elite sports performance, some research has investigated the genetic makeup of individual athletes. Molecular testing has attempted to identify single gene variants deemed ‘responsible’ for performance in specific sports (e.g. a gene responsible for physical power or endurance capacity). This ‘single gene as magic bullet’ philosophy has led to claims that elite performers are born to succeed. For example, there have been attempts to identify sprinters and endurance runners on the basis of differing alleles (i.e. forms) of a single gene known as a-actinin-3.[35,36] Besides obvious ethical issues, reports of gene-profiling and gene-transfer technology raise more general theoretical and practical questions about the nature of genetic and environmental constraints on skill acquisition and performance.[37] Research on human behaviour has yet to reveal that the function of a single gene variant can be inferred from identifying performance phenotypic variance (e.g. power athletes and endurance athletes), supporting the notion of non-specificity of genetic constraints.[38] 1.1 Methods for Assessing Expert Performance in Traditional Research

Empirical expertise research has typically been conducted with two main methods involving: (i) quantitative analyses, which include implementation of multiple test batteries on developing athletes; and (ii) qualitative examination of developmental histories of past or present elite ª 2010 Adis Data Information BV. All rights reserved.

273

athletes, with methods such as interviews, questionnaires or self-reported, retrospective recall of practice histories. 1.2 Outcomes of Research with Quantitative Methodologies

Attempts to distinguish athletic expertise based on skills testing have been unsuccessful or have been confounded by the overriding importance of maturational factors on observed performance.[9] The over-use of ‘closed-skill drill tasks’ to elucidate athletic potential has been a major weakness in extant research because they lack task representativeness, resulting in a failure to identify skilled players from lesser skilled individuals. For example, Gabbett et al.[39] attempted to assess skills of passing (digging), setting, serving and spiking as part of a test battery to discriminate between junior volleyball players of different abilities. The only test that revealed differences between selected and non-selected players was passing. This may have been because skills in volleyball (digging, setting, spiking and blocking) are not performed in stable environments but are dynamic and characterized by marked temporal and spatial variability dependent on preceding play (e.g. quality of contact of the ball by team-mates and opponents). In other work, Hoare and Warr[40] assessed talented athletes (non-football players) aged 15–19 years to identify and develop female football players. Initial screening required coaches to grade players by assessing skill performance such as ball juggling, dribbling (around cones) and passing. Players were grouped by performance test scores for playing small-sided games before progressing to full-size games for coach observation and selection. They used an innovative quasi-applied research model to identify and develop talent in women’s soccer, and highlighted the need to assess athletes over a longer time period than typical in sport talent assessment, suggesting that 2–3 months is necessary for coaches to make a valid assessment of athlete potential. They also proposed that components of performance require different weightings, since the importance of speed and acceleration was underestimated in their tests. Finally, the need to develop an objective test of game sense was suggested, as Sports Med 2010; 40 (4)

274

some athletes performed well on physical tests, but weakly in terms of general tactical understanding.[40] Other quantitative approaches in empirical talent identification research have attempted to examine sport-specific, relevant performance characteristics, through anthropometric, physical, technical, tactical and psychological testing.[40-44] In the main, these tests have been limited in utility for identifying talented performers in sport. For example, test batteries focusing on discrete performance measures have provided limited information of athletic potential and adaptability in different performance environments.[8,41] Questions on the data of these studies, and the efficacy of traditional talent identification practices, arise due to constraints on individuals like growth, maturation, development and training.[42] To exemplify, Elferink-Gemser et al.[43] employed a test battery to examine 30 elite and 35 sub-elite youth hockey players over three competitive seasons. Their research established differences in technical and tactical variables between skill levels at 14 years as well as variations in endurance capacity in the subsequent 2 years. However, the task demands of the technical test did not provide a valid representation of the competitive performance setting (cf. the importance of representative experimental design in studies of human behaviour).[44] The use of closed skill proficiency tests, without opposition and in a static environment, and physical performance measures considered in isolation need to be carefully evaluated, because they have a low correlation with the specific demands on an individual during a dynamic competitive situation. Several studies have highlighted the importance of using more sportsspecific assessments of tactical and technical competence rather than generic physical measures such as strength or endurance tests.[40,45] Skill-based differences in perceptual-cognitive skills such as anticipation and decision making have been successfully identified in a number of studies (see Williams and Ward[46]). Although a number of generalized visual training programmes are predicated on the idea that a superior visual hardware system provides perceptual advantages to skilled performers, research evidence does not support this view. Rather, it has been ª 2010 Adis Data Information BV. All rights reserved.

Phillips et al.

argued that expertise advantages emerge as a result of sport-specific practice and experience.[46-48] Despite the large volume of studies examining the role of perceptual-cognitive expertise in sport, there has been some confusion over whether research has actually demonstrated the true extent of expertise. Van der Kamp et al.[49] criticised the lack of precision in perceptual-cognitive research, highlighted the magnitude of errors reported in many occlusion-based studies of anticipation skill and suggested a weakness in a popular experimental methodology. In a popular approach, liquid crystal spectacles permit information sources available to participants to be manipulated by limiting vision during key moments of performance.[50] It was argued that high levels of error observed in these studies may have been due to the failure of the methodology to capture the complementary efforts of the ventral and dorsal cortical visual systems in regulating perception and action. This is important since, as Williams and Ericsson[47] indicated, the first stage in determining expertise is that ‘‘performance be observed in situ in an attempt to capture the essence of expertise in the domain of interest and to design representative tasks that allow component skills to be faithfully reproduced in the laboratory’’ (p. 286). Van der Kamp et al.[49] also argued that a weakness of previous work was requiring participants to undertake non-representative actions such as pressing buttons, marking crosses on images of court surfaces and verbally reporting findings. Another common error is the provision of video footage that does not replicate the view of performers in action, thereby limiting understanding of the true extent of the expert advantage. Some researchers, such as Mu¨ller and Abernethy,[50] have developed and incorporated what they termed an ‘‘ecologically valid’’ batting test to measure a batsman’s perceptual-action coupling and found skilled players utilized prospective ball flight information to a greater extent than less skilled players. 1.3 Outcomes of Research Using Qualitative Methods

Qualitative studies on the development histories of elite athletes have been somewhat limited in Sports Med 2010; 40 (4)

Expert Performance and Talent Development

number, examining physical, psychological, environmental and social factors that shape performance.[51,52] They have established the importance of environmental constraints, including support from family or friends and the opportunity to participate in residency development programmes. A range of individual constraints – including mental preparation, focus and commitment, and clear goal setting – have been found to contribute to the development of expertise.[8] Durand-Bush and Salmela[52] examined qualitatively the development and maintenance of expert performance in Olympic and world championship competitions. Their data revealed that athletes did not all follow the same pathway to become world and Olympic champions, highlighting how individuals can take different routes to expert status, use various resources and strategies, and be innovative and creative as they develop and maintain their expertise in sport. However, some common factors were observed, including: (i) all athletes underwent stages described as the sampling, specializing, investment and maintenance years;[53] (ii) high levels of self confidence and motivation; and (iii) high levels of creativity and innovation during the maintenance years, a continuous drive to learn and improve, and a strong work ethic. 2. Dynamics Systems Theory and Constraints on the Acquisition of Expertise in Sport In contrast to more traditional approaches to studying expertise, a potentially valuable conceptual framework for modelling the acquisition of expertise exists in understanding performers as complex, dynamical systems. Complexity sciences have been used to study and explain the rich patterns formed in complex systems such as animal collectives, weather systems, the human brain and movements in team sports, where patterns emerge from seemly random component trajectories.[54-56] From this description an emerging expert can be viewed as a complex system, composed of many degrees of freedom on many system levels. The potential for interaction between system components provides the platform ª 2010 Adis Data Information BV. All rights reserved.

275

for rich patterns of behaviour to emerge as individuals interact with dynamically changing environments. This new perspective reveals that adaptation in performance occurs as the result of system trade-offs between specificity and diversity of behaviours.[57] Dynamical systems theory emphasizes the influence of interacting constraints on performance and provides a framework showing how expertise can be achieved in diverse ways as individual performers attempt to satisfy the unique constraints on them.[22] Some current models of talent development are harmonious with these theoretical ideas, although their tenets are not necessarily predicated conceptually on these insights. For example, Simonton[10] proposed that talent emerges from multidisciplinary, multiplicative and dynamic processes and is likely to operate as an intricate system beyond the scope of the polar naturenurture debate. He pioneered mathematical equations to operationalize the potential components that contribute to talent development. These components were weighted by relevance and included reference to genetic dispositions (e.g. height or endurance capacity), environmental (e.g. social and familial support) and developmental constraints. Subsequently, the model was described as emergenic and epigenetic, comprising components that interact and change with time.[10] The emergenic aspect proposed that potential talent consists of multiple components, including all physical, physiological, cognitive and dispositional traits that facilitate the manifestation of superior expertise in a specific domain. Beyond individual differences, Simonton’s[10] model captured the dynamics of epigenetics. Epigenetics were seen in the diverse components that make up talent, which slowly appear and differentiate over time in an individual and ultimately depend on underlying neurological, muscular, cultural, skeletal, social, psychological, physiological and environmental variables.[58] However, they emerge gradually during the course of long-term interactions between the internally developing organism and appropriate environmental constraints.[10] This system is complicated, as it includes the evolutionary interaction of components, and any examination Sports Med 2010; 40 (4)

Phillips et al.

276

of talent development that utilizes this model needs to be holistic, impartial and sophisticated.[10] Although this mathematical model is a useful starting point for sport scientists, since it attempts to operationalize key concepts such as multidisciplinary talent development practices and multiple developmental trajectories towards potential expertise,[8] it lacks theoretical power because it is not conceptualized within a theoretical framework. This weakness could be mediated by including a dynamical systems theory as a viable rationale for talent development as an emergenic and epigenetic process. Furthermore, on a practical level, adopting Simonton’s[10] model may be extremely difficult, as its efficacy is predicated on identifying all components that contribute to expertise in any one specific sporting domain. Identification of every component is essential because the model is proposed to be multiplicative, and any score of zero for any specific factor signifies that expertise cannot be achieved. Vaeyens et al.[9] provided an insightful model of talent identification, capturing the dynamic nature of talent and its development and focusing on potential for development and inclusion rather than early identification. Such a model can be strengthened with a dynamical system theoretical approach also, and in the next section we elucidate key ideas in this framework. 2.1 The Role of Neurobiological Degeneracy in Expertise Acquisition

Athletes considered as complex, neurobiological systems are composed of many interacting parts and levels, which self-organize under constraints.[59] These systems have been conceptualized recently as pleiotropic and degenerate with the ability to adapt to different task and environmental demands.[60] Pleitropy provides neurobiological systems with a variety of alternative performance solutions.[61] Neurobiological degeneracy refers to the ability of structurally different components to be coordinated together to achieve the same behavioural goal.[57] At the level of gene networks, degeneracy promotes evolutionary fitness by ensuring that genetic diversity ª 2010 Adis Data Information BV. All rights reserved.

supports functional adaptation to variable environments.[62] The degenerate relationship between system components and system output in developing experts implies that there are many different pathways to achieving expert performance. Genetic diversity is responsible for a portion of training or performance response differences between individuals, and when there is a favourable interaction with important environmental constraints, performance benefits may be observed. Given differences in genetic contributions, performance variations are more likely to assert themselves under intensive practice regimens. The characteristics of pleiotropy and degeneracy in athletes highlight the need for a systemsoriented, multidimensional framework.[24] Expertise attainment in a particular sport depends on many additional constraints outside the cognitive domain, including but not limited to genetics, social and physical environment, opportunity, encouragement and the effect of these variables on physical and psychological traits.[63] Monodisciplinary approaches to the acquisition of expertise fail to capture the complementary nature of the relationship between individual, task and environmental constraints.[8,24,64] For these reasons, Davids and Baker[24] highlighted the need for an interactionist explanatory framework to examine performance attainment in sport. As we note in the following sections of this paper, this theoretical approach provides a viable platform for explaining the dynamic relationship between an individual’s genetic disposition and the environment and the acquisition of expertise in sport through variable pathways and processes. 2.2 Constraints on Acquiring Expertise

In sport, the expression of expertise is limited or shaped by interacting constraints at many system levels. The concept of constraints (boundaries that constrain the interactions of system components) in human movement science was readdressed by Newell,[65] who classified them into organismic, task and environmental constraints. Constraints in the expertise acquisition context can be conceived of as the numerous variables that form each individual expert’s Sports Med 2010; 40 (4)

Expert Performance and Talent Development

developmental trajectory, and it is important to identify the range of constraints on the acquisition of expertise. Given that an individual is born with distinguishing physical characteristics (with a degree of genetic influence), expertise research is concerned with how environmental constraints interact to affect the development of skill and the expression of genotypes. Performance emerges from the intrinsic dynamics of experts, the preferred behavioural tendencies that arise from the interaction of environmental, task and organismic constraints (including development, experience, genes and learning of each performer).[66] Kelso[66] proposed that intrinsic dynamics reflect the organisational tendencies of an individual. Thelen[67] referred to intrinsic dynamics as ‘‘the preferred states of the system given its current architecture and previous history of activity’’ (p. 76). The intrinsic dynamics of each individual are unique and shaped by many constraints, including experience, learning, development, morphology and genes, which interact to shape performance and the acquisition of expertise in sport.[68] In developing experts, these contributions can lead to significant variations in performance solutions. The manner in which each developing expert attempts to satisfy constraints during performance and learning is determined by the matching of his/her intrinsic dynamics with the specific task dynamics.[69] By knowing an individual’s intrinsic dynamics, one can specify what in the performance repertoire actually changes due to environmental, learned or intentional influences.[70] Due to variations in intrinsic dynamics, expert performers are able to generate different types of functional performance solutions. This idea fits with data in cricket fast bowling reported by Pyne et al.,[71] who examined the relationship between junior and senior high performance athletes’ anthropometric and strength characteristics and bowling speed. They observed differences in the variables and strength of correlation of predictors of peak ball speed between age groups, and suggested growth and biological maturation largely accounted for greater peak ball speed in seniors. This line of thinking could be followed in ª 2010 Adis Data Information BV. All rights reserved.

277

a comprehensive performance analysis including technical/coordination variants under a variety of conditions, such as different delivery types, to gain an understanding of the relationship between the intrinsic dynamics and performance solutions of individuals in developmental groups. Therefore, understanding the nature of each individual’s intrinsic dynamics is central to understanding how expert performance develops in sport. Due to different relationships between an individual’s intrinsic dynamics and a set of task dynamics, each athlete may harness system variability in a different way, implying that expertise may develop in quite unique ways. As individuals progress towards a state of expertise and explore different performance solutions, their intrinsic dynamics will alter and diversify. If the behavioural requirements of a task provide a close match with the pre-existing intrinsic dynamics of an individual learner, the rate of expertise acquisition is likely to be enhanced. The matching of intrinsic and task dynamics may explain praecocial behaviour in sport when some athletes can perform incredibly well from an early age and forms the basis of talent transfer in sport. In contrast, acquiring task dynamics that are dissimilar to those of a previously learned task (e.g. tennis and squash movement pattern dynamics) may lead to a longer process of learning because the specific learner’s intrinsic dynamics may need to be significantly re-shaped.[72] These ideas in dynamical systems theory have important implications for understanding the development and maintenance of expert performance. How the intrinsic dynamics of developing experts are continually shaped by genetic and environmental constraints needs to be understood. The effects of environmental constraints on phenotypic gene expression suggest that athletes with what may be perceived as less favourable genetic dispositions may still achieve expert levels of performance given an appropriate skill acquisition environment.[73] Alternatively, genetically gifted athletes may fail to achieve expert status without an environment that allows rich interactions for acquiring and practising skills (for a discussion of rich learning environments see ideas of Hammond and Bateman[74]). Rich Sports Med 2010; 40 (4)

278

learning environments provide many opportunities to engage in continuous interactions with skilled peers, mentors, skilled coaches and supportive parents as athletes acquire movement skills and knowledge in specific sports. In Australia, Abernethy[75] discussed the disproportionate success of athletes from rural locations as the ‘Wagga-effect’. Wagga Wagga is a rural town with a population of 46 735 people,[76] where children typically play many different sports, often engaging in senior or adult competition prior to specialization. It has produced an inordinately high number of international and professional athletes, more than any other urban conurbation in Australia. Coˆte´ et al.[77] suggested that smaller cities (defined as having populations <100 000 people) provide advantageous opportunities for talent development, because the populace is large enough to support the need for facilities, different sporting codes, club networks and competition infrastructure, and allows early exposure to adult competition. Coˆte´ et al.[77] referred to the ‘birthplace effect’, where the quality and quantity of play and practice interactions afforded by the physical environment of smaller cities is favourable for talent development. These data highlight the need to undertake more research examining appropriate environmental and cultural constraints underpinning development of expertise. The range of unique constraints on athlete performance and behaviour have made predictions of expert performance difficult, due to the dynamic nature of sport performance contexts.[24] To summarize so far, it seems unlikely that a singular common optimal pathway to performance expertise exists, because of the degenerate neurobiological system characterizing each individual performer and the effect of interactions between environmental and personal constraints on the intrinsic dynamics of each learner. Expert performers are able to generate different types of functional performance solutions, depending on differences in their intrinsic dynamics. Future research needs to provide a more comprehensive examination of influential constraints, including technical/coordination variants under a variety of tasks’ demands. For example, in cricket fast bowling, investigating different delivery types ª 2010 Adis Data Information BV. All rights reserved.

Phillips et al.

would allow greater insight into the individual dynamics and movement solutions of such a group (e.g. varying ball speed or type and line and/or length in cricket fast bowling).

3. A Complex Systems Approach to Talent Development Conceptualizing expertise acquisition as a complex system has several implications for talent development programmes. First, the aim of talent development is to aid individuals in gaining the expertise needed to satisfy the unique constraints impinging on them in specific performance domains. This can be achieved by measuring intrinsic dynamics of each individual athlete, identifying key constraints on him/her and facilitating development by manipulating these constraints to encourage exploration of movement solutions.[67,68,78,79] Second, talent development programmes can harness existing system nonlinearities by developing strategies to induce phase transitions in individual performers. This aim might be met by understanding how to force individuals into the meta-stable region of the perceptual-motor landscape of practice where a strategy of co-adaptation can underpin the emergence of creative behaviours (typical of the profile of expert performers in line with the findings of Durand-Bush and Salmela[52]). Meta-stable regions of a performance landscape are areas where the system is poised in a state of dynamic stability that allows it to rapidly undergo phase transitions to new, more functional states of organization as constraints change. If this strategy were aligned with observations of critical developmental periods in young athletes, abrupt phase transitions in expertise and performance may be induced. An implication of this approach for coaches is to incorporate tasks that promote adaptability and creativity in performers. In cricket, this aim could be tested through the implementation of carefully designing games that require batters to probe the boundaries of their skill set and force co-adaptive behaviours in learners as long as the design of game tasks are based on grounded principles of the game.[72] Sports Med 2010; 40 (4)

Expert Performance and Talent Development

Third, these ideas signify the nonlinear nature of expertise acquisition. Traditional models of talent development need to be adjusted to consider the different rates of development of potentially talented athletes.[8] Sub-system behaviours continually shape an individual athlete’s intrinsic dynamics. Because of variations in each athlete’s intrinsic dynamics, individual rates of skill acquisition are likely to progress at different time scales.[16] Talent development models need to take into account the different rates of learning, and growth and maturation processes experienced by individuals on their pathway to expertise (not least, it must take into account the relative age effect to maximize the available talent).[80] These different rates of learning can be influenced by key constraints that act as ‘rate limiters’, causing systems to find new functional performance solutions.[78] Rate limiters can be defined as system controllers, i.e. components or sub-systems which limit the development of an individual in sport.[79] For example, children’s strength in key muscle groups may act as a rate limiter that inhibits them from demonstrating the skills that they have acquired already through practice and experience in sport. An important task is to identify the rate-limiting constraints that are acting on an expert system in order to manipulate them and facilitate transitions to a new performance level. 4. The Role of Meta-Stability in Acquiring Expertise Two important characteristics of complex systems mentioned already include meta-stability and co-adaptation. These features are important for talent identification and development but need to be conceptualized in the sport performance domain. Emerging expert athletes exhibit complexity and meta-stability due to the potential for interaction between their sub-systems. In this way, the whole dynamical landscape of expertise can suddenly change with small variations in responses to constraints impinging on the developing athlete. Critical periods have been identified as brief windows of time and space during which a complex system’s organization is ª 2010 Adis Data Information BV. All rights reserved.

279

most open to modification from external and internal constraints. It has been argued that motor learning can be enhanced when developing athletes are located within these critical periods.[81] The challenge for coaches and sport scientists is to identify when individual athletes enter metastable regions and/or critical periods while developing their expertise levels so that performance paradigm shifts in expertise and skill acquisition might be triggered, as we exemplify in this section. For the individual to utilize technical and cognitive variability and find new performance solutions and cope with instabilities (meta-stability) requires complementary cognitive attributes (e.g. confidence, sacrifice, dedication and perseverance). By ensuring exposure to meta-stable regions of the performance landscape during development, experts can discover new modes of behaviour to satisfy interacting perceptual, affective and task constraints. These new modes of performance are likely to emerge as novel solutions to performance problems as developing athletes co-adapt their responses to challenging constraints imposed by opponents, coaches or performance environments. Co-adaptation is an evolutionary strategy that has implications for the way that constraints can influence the process of talent development by forcing the developing expert to find new functional performance solutions. It may represent a useful expertise development strategy for practitioners, as different performers and/or technologies attempt to pressurize individuals to seek unique performance solutions. Furthermore, the process occurs naturally as a sport develops (i.e. through technique changes, rule changes or equipment change, or as athletes pose each other new performance problems). New solutions to performance problems emerge as talented individuals learn how to assemble creative movement solutions during practice. In cricket, fast bowlers can make the ball ‘swing’, which is the curved trajectory of the ball’s flight path, thereby making the task of hitting the ball more difficult for the batter. A cricket ball will swing when asymmetrical aerodynamic forces are acting on the ball, conventionally achieved by angling the ball’s seam obliquely to its direction Sports Med 2010; 40 (4)

Phillips et al.

280

of travel and caring for the ball so that one side is shiny and smooth and the other dull or rough. A more modern variant of this flight characteristic is ‘reverse swing’, which originally gained prominence in Pakistan in the 1970s, with the success of players such as Sarfraz Nawaz, Imran Khan, Wasim Akram and Waqar Younis. These players experimented with ball care and grips and found they could achieve reverse swing by focusing less on keeping the ball smooth and shiny on one side and bowling the ball with the rough side forward (rotating the ball 180 from the traditional position for swing bowling).[82] The prominence of this technique in Pakistan has been attributed to the specific environmental conditions such as hard dry grounds resulting in balls accruing greater wear and tear at a faster rate.[82] Reverse swing fast bowling is now considered one of the most potent forms of attacking bowling in cricket, seldom mastered but often aspired to in elite level fast bowling. In other sports, performance paradigm shifts have been created by: (i) equipment changes (e.g. the change from bamboo to fibreglass pole vaults leading to the World record increasing from around 4.5 m to 6.14 m in a vault by Sergey Bubka in 1993); (ii) changes to playing surfaces (such as international hockey matches on artificial turf instead of grass), which may explain the ebb and flow of current performance standings and international rankings between countries; and (iii) rule changes, such as the turnover law in rugby union or the distance of the three-point line in Olympic basketball versus the NBA (National Basketball Association). In elite sport, the drive for success means that performers are being challenged constantly to co-adapt to succeed. Through co-adaptation, players need to add new skills or strategies in the off-season in order to continue to challenge opponents with new problems.[83] Essentially, players have to constantly reinvent themselves or demonstrate an ability to adapt to the strategies developed by opponents. A good example of this adaptation was observed in 2008 in cricket when the use of ‘switch hitting’ by Kevin Pieterson (a current English test batsman) was first formalized. He developed a strategy of changing his ª 2010 Adis Data Information BV. All rights reserved.

stance (from his typically right-handed position) as a bowler was in the process of delivering a ball, by ‘jumping’ into a left-handed stance in order to overcome the restrictive field placings of opponents in one-day cricket.[84] The idea of sub-systems co-adapting to constraints imposed by other sub-systems of the body is influential in a dynamic systems analysis of motor development across the lifespan.[67] For example, during children’s motor development, dynamic systems analysts emphasize the idea that specific behaviours may not have yet appeared in developing children, because specific sub-systems act as system ‘rate limiters’ and are ‘lying in wait’ for another critical sub-system to reach a critical level (e.g. changes in the muscle to fat ratio in infants to enable upright postural control). In the performance of specific sports, various subsystems could be critical to the performance development in athletes, such as strength, speed, mobility or game understanding as a result of numerous experiences. 5. Conclusions Historically, expertise research has typically focused either on the role of nature (genes) or nurture (environment) as mechanisms for understanding how experts emerge in performance domains such as sport. This dualist approach has failed to emulate the complementary nature of the relationship between individual and environmental constraints. Although numerous hours of training are needed at the elite level, attainment of an expert level of skill is not accomplished by hours of deliberate practice alone. Similarly, the importance of an individual’s genetic make-up has been accentuated with biased interpretation of genomic studies. In recent years, sport performance research has encompassed a move toward multidimensional models of performance and learning in sport, with significant implications for understanding processes of expertise and talent development. Dynamical systems theory and the complexity sciences might provide the basis of an interactionist perspective on expertise acquisition in sports. Dynamical systems theory is an appropriate functional framework for expert performance Sports Med 2010; 40 (4)

Expert Performance and Talent Development

research because it can be used to consider developing athletes as nonlinear, complex neurobiological systems. It avoids the organismic asymmetry that can be observed in traditional models of expertise acquisition and talent development, addressing questions that other frameworks do not have the language and tools to pose. Several concepts with important implications for expertise and talent development researchers include self-organization under constraints, emergence, meta-stability, creativity, degeneracy and system stabilities and instabilities over different timescales. Within this overarching theoretical framework it has been argued that the same performance outcomes can be achieved in diverse ways as individual performers attempt to satisfy the unique constraints imposed on them.[22] Genetic diversity may be responsible for a small part of training or performance response differences between individuals, and only when there is a favourable interaction with important environmental constraints are performance benefits observed. Phenotypic expression of behaviour might be best understood at the level of individual interactions with key environmental and task constraints. Given differences in genetic contributions, performance variations are more likely to assert themselves under intensive practice regimens. Common optimal pathways to performance expertise are not expected, because of neurobiological degeneracy characterizing each individual performer and the effect of interactions between environmental and personal constraints on the intrinsic dynamics of each learner. The acquisition of expertise is domain specific and involves adaptation to performance environments through satisfying unique constraints that impinge on each developing expert. Expertise acquisition emphasizes the changing nature of the performer-environment relationship through development, and gaining experience through training, practice, coaching and competing. A comprehensive examination of expertise involves identifying the intrinsic dynamics of each individual and the specific rate limiters and constraints that shape their behaviour. Each individual athlete comes to a performance context with a particular set of intrinsic dynamics ª 2010 Adis Data Information BV. All rights reserved.

281

that has already been shaped by genes, development and early experiences. Individualized pathways to expert performance are expected because of the uniqueness of these dynamics constraints. Acknowledgements No sources of funding were used to assist in the preparation of this review. The authors have no conflicts of interest that are directly relevant to the content of this review.

References 1. Howe MJA, Davidson JW, Sloboda JA. Innate talents: reality or myth? Behav Brain Sci 1998; 21 (3): 399-407 2. Dunwoody PT. The neglect of the environment by cognitive psychology. J Theoret Philosoph Psychol 2006; 26: 139-53 3. Yarrow K, Brown P, Krakauer JW, et al. Inside the brain of an elite athlete: the neural processes that support high achievement in sports. Nat Rev Neurosci 2009; 10: 585-96 4. Starkes JL, Deakin JM, Allard F, et al. The road to excellence: the acquisition of expert performance in the arts and sciences, sports and games. In: Ericsson KA, editor. Deliberate practice in sports: what is it anyway? Mahwah (NJ): Lawrence Erlbaum Associates, 1996: 81-106 5. Ericsson KA, Krampe RT, Tesch-Romer C. The role of deliberate practice in the acquisition of expert performance. Psychol Rev 1993; 100 (3): 363-406 6. Rankinen T, Bray MS, Hagberg JM, et al. The human gene map for performance and health-related fitness phenotypes: the 2005 update. Med Sci Sports Exerc 2006; 38 (11): 1863-88 7. Carter JEL, Heath B. Somatotyping: development and applications. Cambridge: Cambridge University Press, 1990 8. Abbott A, Button C, Pepping G-J, et al. Unnatural selection: talent identification and development in sport. Nonlin Dynam Psychol Life Sci 2005 Jan; 9 (1): 61-88 9. Vaeyens R, Lenoir M, Williams AM, et al. Talent identification and development programmes in sport: current models and future directions. Sports Med 2008; 38 (9): 703-14 10. Simonton DK. Talent and its development: an emergenic and epigenetic model. Psychol Rev 1999; 106 (3): 435-57 11. Gagne´ F. Transforming gifts into talents: the DMGT as a developmental theory. High Ability Studies 2004; 15 (2): 119-47 12. McGarry T, Anderson DI, Wallace SA, et al. Sport competition as a dynamical self-organizing system. J Sports Sci 2002; 20: 771-81 13. Chow JY, Davids K, Button C, et al. Nonlinear pedagogy: a constraints-led framework for understanding emergence of game play and movement skills. Nonlin Dynam Psychol Life Sci 2006; 10 (1): 71-103 14. Arau´jo D, Davids K, Serpa S. An ecological approach to expertise effects in decision-making in a simulated sailing regatta. Psychol Sport Exerc 2005; 6 (6): 671-92 15. Arau´jo D, Davids K, Hristovski R. The ecological dynamics of decision making in sport. Psychol Sport Exerc 2006; 7 (6): 653-76

Sports Med 2010; 40 (4)

282

16. Liu Y-T, Mayer-Kress G, Newell KM. Qualitative and quantitative change in the dynamics of motor learning. J Exp Psychol 2006; 32 (2): 380-93 17. Newell KM, Liu Y-T, Mayer-Kress G. Time scales in motor learning and development. Psychol Rev 2001; 108 (1): 57-82 18. Davids K, Renshaw I, Glazier P. Movement models from sports reveal fundamental insights into coordination processes. Exerc Sport Sci Rev 2005; 33 (1): 36-42 19. Jaitner T, Mendoza L, Schollhorn WI. Analysis of the long jump technique in the transition from approach to takeoff based on time-continuous kinematic data. Eur J Sport Sci 2001; 1 (5): 1-12 20. Van Emmerik REA, Hamill J, McDermott WJ. Variability and coordinative function in human gait. Quest 2005; 57: 102-23 21. Hamill J, van Emmerik REA, Heiderscheit BC, et al. A dynamical systems approach to lower extremity running injuries. Clin Biomech 1999; 14 (5): 297-308 22. Davids K, Glazier P, Araujo D, et al. Movement systems as dynamical systems: the functional role of variability and its implications for sports medicine. Sports Med 2003; 33 (4): 245-60 23. Baker J, Coˆte´ J, Deakin J. Expertise in ultra-endurance triathletes early sport involvement, training structure, and the theory of deliberate practice. J Appl Sport Psychol 2005; 17 (1): 64-78 24. Davids K, Baker J. Genes, environment and sport performance: why the nature-nurture dualism is no longer relevant. Sports Med 2007; 37 (11): 961-80 25. Ericsson KA, Smith J. Toward a general theory of expertise: prospects and limits. New York: Cambridge University Press, 1991 26. Hodges NJ, Starkes JL. Wrestling with the nature of expertise: a sport specific test of Ericsson, Krampe, and Tesch-Ro¨mers (1993) theory of deliberate practice. Int J Sport Psychol 1996; 27: 400-24 27. Starkes J. The road to expertise: is practice the only determinant? Int J Sport Psychol 2000 10; 31 (4): 431-51 28. Arau´jo D. Promoting ecologies where performers exhibit expert interactions. Int J Sport Psychol 2007; 38: 73-7 29. Helsen WF, Starkes JL, Hodges NJ. Team sports and the theory of deliberate practice. J Sport Exerc Psychol 1998; 20: 12-34 30. Baker J, Coˆte´ J, Abernethy B. Sport-specific practice and the development of expert decision-making in team ball sports. J Appl Sport Psychol 2003; 15 (1): 12-25 31. Coˆte´ J. The influence of the family in the development of talent in sports. Sport Psychol 1999; 13: 395-417 32. Soberlak P, Coˆte´ J. The developmental activities of elite ice hockey players. J Appl Sport Psychol 2003; 15 (1): 41-9 33. Young BW, Salmella JH. Diary analyses of practice and recovery activities for elite and intermediate middle distance runners. In: Papaioannou A, Goudas M, Theodorakis Y, editors. International Society of Sport Psychology 10th World Congress of Sport Psychology; 2001 May 28-Jun 2; Skiathos Island: Christodoulidi Publications, 2001: 19-21 34. Weissensteiner J, Abernethy B, Farrow D, et al. The development of anticipation: a cross-sectional examination of the practice experiences contributing to skill in cricket batting. J Sport Exerc Psychol 2008; 30: 1-23

ª 2010 Adis Data Information BV. All rights reserved.

Phillips et al.

35. Yang N, MacArthur DG, Gulbin JP, et al. ACTN3 genotype is associated with human elite athletic performance. Am J Hum Genet 2003; 73 (3): 627-31 36. Coghlan A. Elite athletes born to run. New Scientist 2003; 30: 4-5 37. Stubbe JH, Boomsma DL, De Geus EJC. Sport participation during adolescence: a shift from environmental to genetic factors. Med Sci Sports Exerc 2005; 37: 563-70 38. Greenspan RJ. The flexible genome. Nature Genet 2001; 2: 383-7 39. Gabbett T, Georgieff B, Domrow N. The use of physiological, anthropometric, and skill data to predict selection in a talent-identified junior volleyball squad. J Sports Sci 2007; 25 (12): 1337-44 40. Hoare DG, Warr CR. Talent identification and women’s soccer: an Australian experience. J Sports Sci 2000; 18: 751-8 41. Morris T. Psychological characteristics and talent identification in soccer. J Sports Sci 2000; 18 (9): 715-26 42. Reilly T, Williams AM, Nevill A, et al. A multidisciplinary approach to talent identification in soccer. J Sports Sci 2000; 18: 695-702 43. Elferink-Gemser MT, Visscher C, Lemmink KAPM, et al. Multidimensional performance characteristics and standard of performance in talented youth field hockey players: a longitudinal study. J Sports Sci 2007; 25 (4): 481-9 44. Araujo D, Davids K, Passos P. Ecological validity, representative design, and correspondence between experimental task constraints and behavioral setting: comment on Rogers, Kadar, and Costall (2005). Ecolog Psychol 2007; 19 (1): 69-78 45. Falk B, Lidor R, Lander Y, et al. Talent identification and early development of elite water-polo players: a 2-year follow-up study. J Sports Sci 2004; 22 (4): 347-55 46. Williams AM, Ward P. Perceptual-cognitive expertise in sport: exploring new horizons. In: Tenenbaum G, Eklund R, editors. Handbook of sport psychology. 3rd ed. New York: John Wiley & Sons, 2007: 203-23 47. Williams AM, Ericsson KA. Perceptual-cognitive expertise in sport: some considerations when applying the expert performance approach. Hum Move Sci 2005; 24: 283-307 48. Abernethy B, Wood J. Do generalized visual training programmes for sport really work? An experimental investigation. J Sports Sci 2001; 19: 203-22 49. van der Kamp J, Rivas F, van Doorn H, et al. Ventral and dorsal system contribution to visual anticipation in fast ball sports. Int J Sport Psychol 2008; 39: 100-30 50. Mu¨ller S, Abernethy B. Batting with occluded vision: an in situ examination of the information pick-up and interceptive skills of high- and low-skilled cricket batsmen. J Sci Med Sport 2006; 9: 446-58 51. Gould D, Dieffenbach K, Moffett A. Psychological characteristics and their development in Olympics champions. J Appl Sport Psychol 2002; 14 (3): 172-204 52. Durand-Bush N, Salmela J. The development and maintenance of expert athletic performance: perceptions of World and Olympic champions. J Appl Sport Psychol 2002; 14 (3): 154-71 53. Coˆte´ J, Hay J. Children’s involvement in sport: a developmental perspective. In: Silva JM, Stevens D, editors.

Sports Med 2010; 40 (4)

Expert Performance and Talent Development

54.

55. 56. 57.

58.

59. 60.

61.

62.

63. 64.

65.

66. 67. 68.

69.

70.

71.

Psychological foundations of sport. 2nd ed. Boston (MA): Merrill, 2002: 484-502 Kauffman SA. The origins of order: self-organization and selection in evolution. New York: Oxford University Press, 1993 Bak P, Chialvo DR. Adaptive learning by extremal dynamics and negative feedback. Physical Rev 2001; 63 (3): 1912-24 Sumpter DJT. The principles of collective animal behaviour. Philosoph Transact Royal Soc 2006; 361: 5-22 Edelman GM, Gally JA. Degeneracy and complexity in biological systems. Proc Natl Acad Sci U S A 2001; 98 (24): 13763-8 Obler LK, Fein D. The exceptional brain: neuropsychology of talent and special abilities. New York: Guilford Press, 1988 Scho¨ellhorn WI. Coordination dynamics and its consequences on sport. Int J Comput Sci Sport 2003; 2: 40-6 Chow JY, Davids K, Button C, et al. Coordination changes in a discrete multi-articular action as a function of practice. Act Psychol 2008; 127 (1): 163-76 Davids K, Arau´jo D, Button C, et al. Degenerate brains, indeterminate behavior and representative tasks: implications for experimental design in sport psychology research. In: Tenenbaum G, Eklund R, editors. Handbook of sport psychology. 3rd ed. New York: Wiley, 2007 Kaerns M, Elston TC, Blake W, et al. Stochasticity in gene expression: from theories to phenotypes. Nature Rev Genet 2005; 6: 451-64 Wolstencroft E. Talent identification and development: an academic review. Edinburgh: sportscotland, 2002 Beek PJ, Jacobs DM, Daffertshofer A, et al. Expert performance in sport: views from the joint perspectives of ecological psychology and dynamical systems theory. In: Starkes JL, Ericsson KA, editors. Expert performance in sport. Champaign (IL): Human Kinetics 2003: 321-44 Newell KM. Constraints on the development of coordination. In: Wade MG, Whiting HTA, editors. Motor skill acquisition in children: aspects of coordination and control. Boston (MA): Martinies NIJHOS, 1986: 341-60 Kelso JAS. Anticipatory dynamical systems, intrinsic pattern dynamics and skill learning. Hum Move Sci 1991; 10: 93-111 Thelen E. Motor development: a new synthesis. Am Psychol 1995; 50 (2): 79-95 Davids K, Button C, Bennett S. Dynamics of skill acquisition: a constraints-led approach. Champaign (IL): Human Kinetics, 2008 Corbetta D, Veriejken B. Understanding development and learning of motor coordination in sport: the contribution of dynamical systems theory. Int J Sport Psychol 1999; 30 (44): 507-30 Zanone PG, Kelso JAS. Learning and transfer as dynamical paradigms for behavioral change. Adv Psychol 1992; 87 (3): 563-82 Pyne DB, Duthie G, Saunders PU, et al. Anthropometric and strength correlates of fast bowling speed in junior

ª 2010 Adis Data Information BV. All rights reserved.

283

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83. 84.

and senior cricketers. J Strength Cond Res 2006; 20 (3): 620-6 Renshaw I, Davids K, Shuttleworth R, et al. Insights from ecological psychology and dynamical systems theory can underpin a philosophy of coaching. Int J Sport Psychol 2009; 40 (4): 540-602 Baker J, Horton S. A review of primary and secondary influences on sport expertise. High Ability Studies 2004; 15 (2): 211-28 Hammond KR, Bateman RA. Sport psychology as an instance of ecological psychology. Int J Sport Psychol 2009 01; 40 (1): 38-49 Abernethy, B. Sport expertise: From theory to practice. In: Farrow D, editor. Proceedings of the Applied Sport Expertise and Learning Workshop 2005 [CD1]; Canberra (ACT): Australian Institute of Sport, 2005 Australian Bureau of Statistics. Wagga Wagga (Urban Centre/Locality): 2006 Census QuickStats [online]. Available from URL: http://www.censusdata.abs.gov.au/ABS Navigation/prenav/LocationSearch [Accessed 2009 May 10] Coˆte´ J, Macdonald DJ, Baker J, et al. When ‘‘where’’ is more important than ‘‘when’’: birthplace and birthdate effects on the achievement of sporting expertise. J Sports Sci 2006; 24 (10): 1065-73 Handford C, Davids K, Bennett S, et al. Skill acquisition in sport: some applications of an evolving practice ecology. J Sports Sci 1997; 15 (6): 621-40 Thelen E, Smith LB. A dynamics systems approach to the development of perception and action. Cambridge (MA): MIT Press, 1994 Cobley S, Baker J, Wattie N, et al. Annual age-grouping and athlete development: a meta-analytical review of relative age effects in sport. Sports Med 2009; 39 (3): 235-56 Anderson DI. Do critical periods determine when to initiate sport skill learning? In: Smoll FL, Smith RE, editors. Children and youth in sports: a biopsychosocial perspective. 2nd ed. Indianapolis (IN): Brown & Benchmark, 2002: 105-48 Mehta RD. Cricket ball aerodynamics: myth versus science. In: Subic AJ, Hakke SJ, editors. The engineering of sport research, development and innovation. Oxford: Blackwell Science, 2000: 153-67 Cowdrey MC. Tackle cricket. London: Stanley Paul, 1974 Hughes S. Kevin Pietersen ‘switch hits’ break new ground. Telegraph Media Group [online]. Available from URL: http:// www.telegraph.co.uk/sport/cricket/international/england/ 2303325/Kevin-Pietersen-switch-hits-break-new-ground.html [Accessed 2008 Nov 12]

Correspondence: Elissa Phillips, Biomechanics and Performance Analysis, Australian Institute of Sport, PO Box 176, Belconnen, ACT 2616, Australia. E-mail: [email protected]

Sports Med 2010; 40 (4)

Sports Med 2010; 40 (4): 285-302 0112-1642/10/0004-0285/$49.95/0

REVIEW ARTICLE

ª 2010 Adis Data Information BV. All rights reserved.

The Maximal Accumulated Oxygen Deficit Method A Valid and Reliable Measure of Anaerobic Capacity? Dionne A. Noordhof,1 Jos J. de Koning1,2 and Carl Foster2 1 Faculty of Human Movement Sciences, Research Institute MOVE, VU University, Amsterdam, the Netherlands 2 Department of Exercise and Sports Science, University of Wisconsin-La Crosse, La Crosse, Wisconsin, USA

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Anaerobic Energy Production: the Time Course of the Different Components. . . . . . . . . . . . . . . . . . . 2. The Construction of a Linear Relationship Between Exercise Intensity and Oxygen Uptake . . . . . . . . 2.1 Exercise Modality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1 Running . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2 Cycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.3 Running versus Cycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 The Duration of Submaximal Exercise Bouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 The Effect of the Number of Submaximal Exercise Bouts on the Linear Power Output-Oxygen Uptake Relationship . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. The Influence of Supramaximal Exercise Protocol on the Computed Maximal Accumulated Oxygen Deficit (MAOD) Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. The Validity of the MAOD Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. The Reliability of the MAOD Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

285 287 288 288 288 289 290 290 291 293 294 296 299

The maximal accumulated oxygen deficit (MAOD) method has been extensively, but unfortunately not very methodically, used; the procedure used to determine the MAOD varies considerably. Therefore, this review evaluates the effect of different numbers and durations of submaximal exercise bouts . on the linear power output (PO)-oxygen uptake (VO2) relationship and thus the MAOD. Changing the number and duration of the submaximal exercise bouts substantially influences the calculated MAOD when relatively long submaximal exercise bouts are used and no fixed value of the y-intercept is forced into. the linear regression line. This is most likely due to non-linearity intensities above the lactate threshold of the PO-VO2 relationship for exercise . is probably caused by the (LT). Non-linearity of the PO-VO2 relationship . development of a slow component in VO2 during submaximal exercise at intensities above the LT. Thus, it is important to standardize the number, duration and intensity of submaximal exercise bouts necessary to establish

Noordhof et al.

286

. the PO-VO2 relationship. Beyond changing the number and duration of the submaximal exercise bouts, the effect of different supramaximal exercise bouts on the calculated MAOD has been investigated. While it has become clear that different exercise protocols result in relatively similar values of the MAOD, a closer look at individual data suggests that it may be important to choose an exercise protocol that is representative of the athlete’s event. The validity of the MAOD method was studied by different authors comparing the MAOD with metabolic measurements of anaerobic adenosine triphosphate (ATP) production. The main limitation with the metabolic measurements of anaerobic ATP production from muscle biopsy data is that the active muscle mass is unknown, which makes it hard to accurately study the validity of the MAOD method. From the studies that evaluated the reliability of the MAOD method it is clear that the MAOD method may not be a reliable measure of anaerobic capacity. From these findings it can be concluded that the MAOD method may have limitations as a valid and reliable measure of anaerobic capacity and needs to be further improved. We suggest the use of 10 · 4 minute submaximal exercise bouts and . a fixed value of the y-intercept for the construction of the linear PO-VO2 relationship, after which the MAOD can be determined during a supramaximal exercise protocol. specific for the athlete’s event. This method will lead to a more robust PO-VO2 relationship and will therefore result in more valid and reliable results.

A number of studies have been performed trying to define a practical measure of anaerobic capacity. Anaerobic capacity is defined as the maximal amount of adenosine triphosphate (ATP) that can be resynthesized by anaerobic metabolism; that is, mainly phosphocreatine (PCr) hydrolysis and glycolysis. Over the last 20 years, the gold standard to determine anaerobic capacity has been the maximal accumulated oxygen deficit (MAOD). Krogh and Lindhard[1] introduced the term ‘oxygen deficit’ in 1920, Hermansen[2] reintroduced the term in 1969 and Medbø et al.[3] suggested in 1988 that the MAOD was a quantitative expression of anaerobic capacity. A number of studies published since the original report of Medbø et al.[3] have accepted the MAOD as the best practical measure of anaerobic capacity.[4-49] In the original protocol used by Medbø et al.,[3] a linear relationship . between treadmill speed and oxygen uptake ( VO 2) or between power output . (PO) and VO2[50,51] is established for each individual. This relationship is extrapolated . to supramaximal exercise intensities and the VO2 demand corresponding to supramaximal workloads is ª 2010 Adis Data Information BV. All rights reserved.

. predicted. When this VO2 demand is multiplied by the duration of the exercise bout, the accu. mulated VO2 demand is estimated. The MAOD . is calculated by subtracting the accumulated VO2, measured during the . exercise bout, from the estimated accumulated VO2 demand. Medbø et al.[3] and Medbø and Tabata[50,51] used at least ten submaximal exercise bouts with a duration of 10 minutes for . the determination of the individual linear PO-VO2 relationship. In the literature, the number, duration and intensity of the submaximal exercise bouts has varied considerably. One goal of this review is to evaluate the effect of different numbers and durations. of submaximal exercise bouts on the linear PO-VO2 relationship. In the original study of Medbø et al.[3] and in more recent publications,[40-42,44,46,47,52-56] a supramaximal constant intensity work bout to fatigue was used to determine the MAOD. If anaerobic capacity is a well defined individual entity,[3] then it can be hypothesized that if every protocol ends when the subject is totally exhausted, the anaerobic capacity would be the same for different Sports Med 2010; 40 (4)

The MAOD Method

exercise protocols. More recently, supramaximal all-out protocols,[37-39,44,45,52] ramp exercise protocols,[22] intermittent exercise protocols[10,11] and differently paced supramaximal cycling protocols[6,49] have been used to determine the MAOD. Therefore, the second goal of this review is to compare the effect of different exercise protocols on the computed value of anaerobic capacity. There is also a lack of clear evidence about the validity of the MAOD method. Medbø et al.[3] concluded that the MAOD is a quantitative expression of anaerobic capacity and Medbø and Tabata[51] found a strong relationship between muscle characteristics and oxygen deficit, from which they concluded that the MAOD is indeed a valid measure of anaerobic capacity. However, Bangsbo[57,58] questioned the methodology of the MAOD method as described by Medbø et al.[3] Thus, the last goal of this review is to evaluate the validity and the reliability of the MAOD method. 1. Anaerobic Energy Production: the Time Course of the Different Components In order to perform exercise, ATP has to be continuously re-synthesized by aerobic and anaerobic ATP-forming processes. During steady-state submaximal exercise ATP re-synthesis is largely accomplished by aerobic processes. However, when exercise is performed at high intensities or during the onset of even submaximal exercise, the ATP-turnover rate exceeds the momentary aerobically attributable energy production and relies on both aerobic and anaerobic processes.[50,51] Several studies used maximal intermittent cycling exercise to study anaerobic ATP re-synthesis.[59-62] Short-term supramaximal sprint-type exercise is highly suitable to study anaerobic energy production, because the POs elicited by . the. cyclists are far in excess of the PO at VO2max (PVO2max). To meet the high ATP demand of supramaximal exercise, substrate (e.g. adenosine diphosphate [ADP]) phosphorylation from PCr hydrolysis and glycolysis is necessary to overcome the oxygen deficit at the onset of exercise.[63] ª 2010 Adis Data Information BV. All rights reserved.

287

The upper limit of glycolysis and pyruvate production is determined by glycogen phosphorylase (Phos) which catalyzes the rate-limiting step in glycogenolysis.[59] Pyruvate dehydrogenase (PDH), which catalyzes the oxidative decarboxylation of pyruvate to acetyl-CoA, controls the entry of pyruvate into the tricarboxylic acid cycle.[59,60,63] Therefore, the time course of the activation of Phos and PDH has been extensively studied in order to get more insight into anaerobic energy production.[59,60,62,63] Subjects in the study of Parolin et al.[59] had to perform three 30-second bouts of maximal isokinetic cycling, separated by 4-minute rest periods, in order to study the metabolic responses during sprint-type exercise. Muscle biopsy data revealed that the primary source of ATP resynthesis during the initial 6 seconds of supramaximal exercise was substrate phosphorylation from PCr hydrolysis. The active form of Phos increased significantly during the first 6 seconds and continued to be elevated during the remaining exercise time, which resulted in significant pyruvate production. During the last 15 seconds of the exercise bout, the PCr stores became depleted and glycogenolysis became inhibited, probably by the high [H+], which diminished the effect of substrate level phosphorylation to ATP re-synthesis. The active form of PDH became fully active throughout the second part of the sprint bout, most likely stimulated by the enhanced pyruvate production and high [H+],[62] resulting in the oxidation of pyruvate and minimal further lactate production. During the initial seconds of the third supramaximal exercise bout, phosphorylation from PCr hydrolysis was less than during bout one, but still contributed a large amount to ATP re-synthesis. However, inhibition of Phos excluded a large contribution of substrate level phosphorylation to ATP re-synthesis. PDH was completely activated before the start of the third sprint which resulted in close agreement between pyruvate production and oxidation. Thus, the inability to maintain a high level of substrate phosphorylation from PCr hydrolysis and glycolysis during each supramaximal exercise bout and during successive exercise bouts was accounted for by an increased level of oxidative Sports Med 2010; 40 (4)

Noordhof et al.

288

phosphorylation.[59] It has to be mentioned that besides substrate level phosphorylation, the myokinase reaction also contributes to ATP resynthesis; however, the resultant amount of free adenosine monophosphate (AMP) is relatively small[59] and therefore contributes only slightly to ATP re-synthesis. The time course of substrate phosphorylation from PCr hydrolysis and glycolysis is relatively the same during supramaximal constant PO protocols. Gonza´lez-Alonso et al.[64] studied heat production of the quadriceps muscles and the Musculus tensor fasciae latae during intense dynamic knee-extensor exercise to compute the energy liberation of the anaerobic and aerobic energy producing systems. During the initial 30 seconds of the 180 seconds of knee-extensor exercise, the contribution of the aerobic system was only 32%, which increased to >82% after 60 seconds. The increase in oxidative phosphorylation was accompanied by a significant increase (107%) in heat production, without any significant change in PO. The most likely explanation for this finding is that substrate phosphorylation from PCr hydrolysis and glycolysis results in lesser heat liberation per ATP produced than oxidative phosphorylation. These results are confirmed by the study of Krustrup et al.,[65] who found a decrease in mechanical efficiency from 56 – 5% to 32 – 3% during low intensity knee-extensor exercise and a decrease from 47 – 4% to 36 – 3% during a 90-second high intensity bout. The studies of Gonza´lez-Alonso et al.[64] and Krustrup et al.[65] supported the hypothesis that the mechanical efficiency is higher for anaerobic ATP re-synthesis than for aerobic ATP-resynthesis. This important finding of a higher anaerobic mechanical efficiency than aerobic mechanical efficiency has to be taken into account when the MAOD values expressed in mL O2 eq (oxygen equivalent) per kg body mass (mL O2 eq kg-1) are converted to concentrations of the different anaerobic components in mmol kg-1. Thus, as long as we speak about the anaerobic capacity in oxygen equivalents or in mechanical work performed, the difference in anaerobic and aerobic mechanical efficiency is of no hindrance.



ª 2010 Adis Data Information BV. All rights reserved.



2. The Construction of a Linear Relationship Between Exercise Intensity and Oxygen Uptake 2.1 Exercise Modality

In the original study of Medbø et al.,[3] the chosen exercise modality was treadmill running. Besides treadmill running, Medbø and Tabata[50,51] also used bicycle exercise to determine anaerobic capacity with the MAOD method. Since these early studies of Medbø et al.[3,66] and Medbø and Tabata[50,51] the MAOD has also been determined for supramaximal rowing,[19,53] arm cranking,[42] swimming[24,25,42] and one-legged, dynamic, knee-extensor exercise.[67] We will only discuss the concerns that come along with the estimation of anaerobic capacity during running and bicycling exercise, as those exercise modalities are most often used. 2.1.1 Running

Several studies investigated the effect of different treadmill inclinations on the MAOD attained during running, from which it can be concluded that uphill running results in significantly higher anaerobic capacities than horizontal running.[20,68,69] Olesen[20] found that the MAOD increased from 39.5 – 10.7 and 56.9 – 8.0 mL O2 eq kg-1 at a 1% grade to 69.4 – 8.3 and 99.8 – 7.1 mL O2 eq kg-1 at a 20% grade in anaerobically untrained (+76%) and anaerobically trained (+75%) runners, respectively. The results of the study of Craig and Morgan[69] supported the findings of Oleson[20] but they did not find a correlation between the treadmill belt inclination and the MAOD attained during treadmill running (r = -0.15), which suggests that certain subjects show a higher running economy during horizontal running while other subjects show a higher running economy during uphill running. The finding that the percentage inclination and the MAOD are unrelated[69] is in contrast with the finding of Olesen.[20] The oxygen deficit reached a maximum value at a 15% gradient and a further increase to a 20% gradient did not result in an additional increase in oxygen deficit, which suggests the attainment of the anaerobic capacity for running.[20]





Sports Med 2010; 40 (4)

The MAOD Method

Medbø et al.[3] concluded that the MAOD is heavily dependent on the active muscle mass, which suggests that during uphill running a larger muscle mass is activated than during horizontal running. Sloniger et al.[70] used exercise-induced contrast shifts in magnetic resonance images to determine the pattern and intensity of muscle activation during treadmill running. A significant 9% increase in the active muscle mass was found during uphill running as compared with horizontal running, which was achieved by an altered pattern of muscle activation, i.e. the percentage of muscle volume increased significantly (p £ 0.004) for the Musculus vastus group [i.e. the M. vastus lateralis, M. vastus medialis and the M. vastus intermedius] (23%) and the Musculus soleus (14%), and decreased significantly (p £ 0.004) for the Musculus rectus femoris (29%), the Musculus gracilis (18%) and the Musculus semitendinosus (17%) from horizontal to uphill running.[70] However, they only found a moderately strong correlation between the increase in MAOD and the percentage increase in muscle activation.[68] It can be concluded that the percentage active muscle mass only explains a small part of the increase in MAOD from horizontal to uphill running and that other factors must be responsible for the difference.[20,57,68,70] Thus, for the determination of anaerobic capacity during treadmill running it is important to make a carefully considered decision about the treadmill inclination, because the MAOD at a 15% gradient possibly reflects the anaerobic capacity of a different set of muscles to those used during horizontal running.[21] Besides a higher MAOD found during uphill running compared with level running, there is another concern related . to the inclination of the treadmill, namely the VO2 slow component (SC), which is the late (i.e. secondary) increase in . VO2.[71] Pringle et al.[72] studied the oxygen kinetics during horizontal and uphill (10% grade) . running and found that the amplitude of the VO2 SC was significantly higher (40%) during uphill running compared with horizontal running. Based on this finding, it can be concluded that uphill running is less suitable for the determination of anaerobic capacity than horizontal running. ª 2010 Adis Data Information BV. All rights reserved.

289

2.1.2 Cycling

Just as treadmill inclination was a confounding factor for the determination of anaerobic capacity during running, pedalling rate could influence the estimation of anaerobic capacity during cycling. Several studies used a fixed ped. alling frequency to establish the linear PO-VO2 relationship and eventually used this relationship to determine the MAOD during all-out exercise, throughout which pedalling frequency was free to vary,[5,35,38,39,44,45,52] without having studied the effect of pedalling frequency on MAOD. The effect of pedalling frequency on the MAOD was investigated by Woolford et al.,[6] . who established three different PO-VO2 relationships, based on submaximal exercise tests conducted on different cycle ergometers, which required pedalling frequencies in the range of 90–100 revolutions per minute (rpm), 120–130 rpm and 90–130 rpm. A higher pedalling frequency resulted in a higher oxygen demand of cycling and a higher resultant MAOD. The MAOD derived during cycling at pedalling frequencies between 90 and 100 rpm was approximately 30% lower than during cycling at frequencies between 120 and 130 rpm or at frequencies between 90 and 130 rpm.[6] Zoladz et al.[73] investigated the effect of ped. alling frequency on VO2 kinetics and. found a gradual, non-significant increase in VO2 when pedalling frequency increased from 60 to 100 rpm and a significant increase in oxygen uptake when pedalling frequency was further augmented to 120 rpm. Besides an effect of pedalling frequency . on the PO-VO2 relationship and on anaerobic capacity, there . is also an effect of pedalling frequency on VO2 SC.[74] Pringle et al.[74] found . a significant increase in the amplitude of the VO2 SC when pedaling frequency increased from 35 rpm . to 115 rpm, even when the amplitude of the VO2 SC . was expressed relatively. to the end exercise VO2. The increase in the VO2 SC was probably due to a change in motor unit recruitment patterns. . Thus, for the use of a linear PO-VO2 relationship it is important to choose a fixed pedalling frequency in the range of 60–100 rpm and to maintain this pedalling frequency during the Sports Med 2010; 40 (4)

Noordhof et al.

290

exhausting exercise bout used to determine the MAOD. 2.1.3 Running versus Cycling

Billat et al.[75,76] published two studies on the . SC, from effect of exercise modality on the VO 2 . which it became clear that VO2max did not differ between running .and cycling exercise for triathletes but that the VO2 SC was significantly higher during cycling compared with running (20.9 – 2 vs 268.8 – 24 mL min-1). Based on these results it. can be concluded that for the use of a linear POVO2 relationship horizontal running exercise . is preferred above cycling exercise, because the VO2 SC is less. during running compared with cycling and the VO2 SC is less during horizontal running compared with uphill running.



2.2 The Duration of Submaximal Exercise Bouts

Medbø et al.[3] suggested that submaximal exercise bouts of 10-minutes .duration are required to construct . the linear PO-VO2 relationship from which the VO2 demand of supramaximal exercise bouts can be predicted. The required duration of the submaximal exercise bouts, as proposed by Medbø et al.,[3] has been questioned by others.[57,58] Bangsbo[57,58] stated that the duration of the submaximal exercise bouts influences the . measured VO2 values. Whipp and Wasserman[77] and Barstow and Mole´[78] found that at .low cycling intensities (W) (50, 75 and 100 W) VO2 reached a steady-state value within 3 minutes. . VO2 continued to increase after 3 minutes and did not reach a steady-state value within 6 minutes for higher exercise intensities (125, 150, and 175 W). . The question that rises is: does the SC of the VO2 kinetics substantially affect the computation of the MAOD? . Bangsbo[57] constructed two . linear PO-VO2 relationships, one based on VO2 measurements after 4–6 minutes of submaximal running and one . based on VO2 measurements after 8–10 minutes of . submaximal running. The difference in the VO . 2 demand, estimated from the two linear POVO2 relationships, resulted in a difference in MAOD of 10.5 mL O2 eq kg-1 (21%), with the longer submaximal bouts yielding a higher pre-



ª 2010 Adis Data Information BV. All rights reserved.

. dicted VO2 demand and thus computed MAOD. Thus, the duration of the submaximal exercise . bouts can substantially influence the PO-VO2 relationship and the calculated MAOD.[57,58] Based on these results, Bangsbo[57,58] concluded that it is hard to validate the use of a linear relationship between exercise intensity (i.e. running . speed) and VO2 for the determination of anaerobic capacity. Buck and Mc Naughton[46] studied the effect of submaximal exercise bout durations of 2–4 minutes, 4–6 minutes, 6–8 minutes, and the proposed . 8–10 minutes[3] on the estimated VO2 demand, and thus the calculated MAOD. Increasing durations of the submaximal cycling bouts. resulted in increasing values of the estimated VO2 demand and, accordingly MAOD. The linear . PO-VO2 relationship based on submaximal cycling bout durations of 2–4 minutes resulted in a 25.8 – 8.7% lower MAOD than the linear relation. ship based on the VO2 during minutes 8–10. The results of Buck and Mc Naughton[46] supported the findings of Bangsbo[57] that the duration of the submaximal . exercise bouts indeed influences the measured VO2 and thus the calculated MAOD. However, Buck and Mc Naughton[46] drew a different conclusion from their results. They concluded that 10-minute bout durations are ne. cessary to establish a linear PO-VO2 relationship, because otherwise the calculated MAOD is too low. [32] also compared . the Maxwell . and Nimmo average VO2 during minutes 4–6 with the VO2 measured during minutes 8–10 of submaximal exercise bouts. Significant (p < 0.001). differences were found between the average VO2 of 4–6 minutes versus 8–10 minutes (53.8 – 3.7 vs 55.8 – 3.9 mL kg-1 min-1.), which resulted in a different treadmill speed-VO2 relationship, and thus different MAOD values. Although Maxwell and Nimmo[32] . found significant differences in the average VO2 between 4–6 minutes and 8–10 minutes, no significant differences were found between 8–9 minutes and 9–10 minutes or between 7–10 minutes and 8–10 minutes. Based on these results they recommended the use of 8–10 minutes as an appropriate interval to define . the linear treadmill speed-VO2 relationship.

 

Sports Med 2010; 40 (4)

The MAOD Method

In summary, the conclusions drawn by Buck and Mc Naughton[46] and by Maxwell and Nimmo[32] are in conflict with the conclusion drawn by Bangsbo.[57] In the study of Buck and Mc Naughton[46] and in the study of Maxwell and Nimmo[32] they did not . consider the possibility that the steady-state VO2 measured during minutes 8–10 of the highest submaximal exercise bouts could be the asymptotic value . that was. attained after a late increase in VO2 (the VO2 [78,79] SC). Green and Dawson[79] found a large SC . in VO2 between minutes 3–6 for high submaximal POs and a smaller amplitude between minutes [78] even found that the 6–10. Barstow and . Mole´ amplitude of the VO2 SC increased with increas. ing PO. This SC causes the steady-state VO2, . if attained, to be larger .than the VO2 demand predicted from the PO-VO2 relationship based on the lower submaximal exercise intensities.[78] Thus, we can conclude that the relationship be. tween treadmill speed and/or PO and VO2 may be non-linear for exercise intensities above lactate threshold (LT) and that the degree of non-linearity increases with increasing exercise intensity. 2.3 The Effect of the Number of Submaximal Exercise Bouts on the Linear Power Output-Oxygen Uptake Relationship

Medbø et al.[3] used an iterative procedure to determine the minimum number of submaximal exercise bouts necessary to construct a linear . PO-VO2 relationship. The iterative procedure . consisted of three steps: (i) a linear PO-VO2 relationship based on the first two submaximal exercise bouts was established; (ii) the MAOD of the supramaximal constant intensity exercise bout was determined; and (iii) the next submaximal exercise intensity and the corresponding . VO2 was added to the linear regression line. These three steps were repeated until the MAOD started to converge. At least ten submaximal exercise bouts were necessary for the MAOD to converge to a stable value.[3] Despite this early proposed number of submaximal exercise bouts, different numbers of submaximal exercise. bouts have been used in the literature. PO-VO2 relationships have been based on ª 2010 Adis Data Information BV. All rights reserved.

291

two,[16,32] three,[13,15,35,36,41,56,80] four,[5,19,23,26,28,29,66,67] five,[21,31,40,42,44,71] six,[7-9,11,12,24,25,30,37,38] or eight[69] submaximal exercise bouts. Buck and Mc Naughton[47] systematically studied the effect of different numbers of sub. maximal exercise bouts on the linear PO-VO2 relationship, established for bicycle exercise, .and thus the calculated MAOD. Steady-state VO2 values were determined during 10 · 10-minute submaximal .exercise bouts at intensities between 30 and 90%. VO2max and were used to construct a linear PO-VO2 relationship.[47] Data points were . sequentially removed from the linear PO-VO2 relationship based on 10 · 10-minute submaximal exercise bouts, starting with the lowest PO, the highest PO or the most central PO. This systematic removal of data points resulted in 17 . different linear PO-VO2 relationships, which were compared with the linear relationship based on 10 · 10-minute submaximal exercise bouts, as proposed by Medbø et al.[3] Sequentially removing the lowest or highest submaximal exercise bout from the linear relationship resulted in an increasingly larger difference in MAOD when compared with the MAOD . calculated from the 10 · 10-minute linear PO-VO2 relationship. Removing the most central PO from the linear relationship predictably resulted in smaller differences in MAOD compared with the other methods for removing data points.[47] It was concluded that the effect of changing the number of . submaximal exercise bouts on the linear POVO2 relationship and thus MAOD is dependent on the intensity of the bouts used to establish the linear relationship. Removing the highest submaximal exercise intensities had the most pronounced effect on MAOD and removing the most central exercise intensities had the least pronounced effect. That the effect of removing the lowest or the highest submaximal exercise bouts was not the same. implies that the relationship between PO and VO2 is not truly linear. The outcome of a reduced number of submaximal exercise bouts, with and without the use of a fixed value of the y-intercept, on MAOD was studied by Russell .et al.[14] The accuracy of the estimations of the VO2 demand and the MAOD was evaluated with 95% confidence intervals Sports Med 2010; 40 (4)

Noordhof et al.

292

(CIs), because they questioned the use of the Pearson correlation coefficient, as large differences in MAOD were found despite correlation coefficients ‡0.99 in the study of Buck and McNaughton.[47] Submaximal exercise bouts at intensities of 50, 60, 70, 80 and 90% of the average PO during a 2000 m rowing ergometer test, with a duration of 5 to 7 minutes, and a forced y-inter-1 min-1 were used to estabcept of 5.1 mL kg . lish a linear PO. VO2 relationship (5+Y). A second linear PO-VO2 relationship was established using the same submaximal exercise bouts but -1 without a forced y-intercept of 5.1 mL kg . -1 min (5-Y). The third and last linear PO-VO2 relationship required only two submaximal bouts, . at intensities between 75–82% and 85–92% of VO2peak, and a fixed value for the y-intercept (MED). The third regression line is named ‘procedure 3’ in the study of Medbø et al.,[3] and is suggested as a less time consuming alternative procedure from which the resultant MAOD differed by only 2 mL O2 eq kg-1 from the MAOD estimated with the 10-point regression line. However, the use of ‘procedure 3’ and thus the use of a fixed y-intercept of 5.1 mL kg-1 min-1, as suggested in the study of Medbø et al.,[3] has only been confirmed for treadmill exercise at an inclination of 10.5%.[3] The three different PO. VO2 relationships in the study of Russell et al.[14] did not result in statistically . significant different values for the estimated VO2 demand and, accordingly, the MAOD. However, the size of the 95% CI was significantly smaller when using the regression line based on the 5+Y method (12.88 – 2.93) than when using the 5-Y method (27.63 – 14.47). The length of the 95% CI was of similar magnitude for the methods 5+Y and MED (16.89 – 10.07). Thus, Russell et al.[14] concluded that the three different regression lines . resulted in similar VO2 demand and MAOD. values, but that the precision of estimating the VO2 demand and thus the MAOD improved when a y-intercept of 5.1 mL kg-1 min-1 was added to the regression line (5+Y method). Bickham et al.[81] evaluated the effect of different methodological improvements on the anaerobic capacity attained during treadmill . running at a grade of 1%. The first PO-VO2 re-

 







 

 

ª 2010 Adis Data Information BV. All rights reserved.

lationship was based on 10 · 4-minute submaximal bouts and a forced y-intercept, the second . PO-VO2 relationship was based on 4 · 4-minute submaximal bouts and a forced y-intercept, and the third relationship was based on 2 submaximal bouts and a forced y-intercept, as suggested by Medbø et al.[3] Five of the ten submaximal exercise intensities used in the first linear regression line were below LT and five of them were above LT. When only four submaximal bouts were used to establish a linear regression line, the chosen submaximal .intensities averaged 69%, 76%, 86% and 93% VO2max. Average submaximal intensities . of 83% and 93% were used for the linear PO-VO2 relationship that was based on two submaximal bouts and a y-intercept. The three different regression lines did not result in significant differences in MAOD and in the accompanying 95% CIs, which means that reducing the number of submaximal bouts is acceptable when a fixed value . of the y-intercept is used and submaximal VO2 data is carefully collected. [82] doubted the validity of Green and Dawson . relationship for the prediction using the POVO 2 . of the VO2 demand of maximal and supramaximal cycling exercise. Therefore, they compared . the linear PO-VO2 relationship based on exercise intensities below LT with the linear relationship based on exercise intensities above LT. The use of submaximal exercise intensities above LT resulted in a 14% and 6% greater slope of the regression line for cyclists and untrained men, respectively, compared with the regression line based on exercise intensities below LT. This difference was due to the disproportionally higher . VO2 attained during exercise above LT. Based on these results, .Green and Dawson[82] concluded that the PO-VO2 relationship is non-linear for exercise intensities above LT, and that the degree of non-linearity increases with increasing fitness of the subjects. The results of Russell et al.[14] and Bickham et al.[81] are not consistent with the findings of Buck and McNaughton[47] and Green and Dawson,[82] that changing the number of submaximal exercise . bouts influences the estimated VO2 demand and thus the MAOD. This may be due to the use of only 4-minute duration exercise bouts for the Sports Med 2010; 40 (4)

The MAOD Method

293

submaximal bouts in the study of Bickham et al.,[81] which is much shorter than the 10-minute duration of the submaximal exercise bouts in the study of Buck and McNaughton,[47] as well as in the original studies of Medbø.[3,50,51] However, Green and Dawson[82] also used 4-minute sub. maximal bouts to establish the linear PO-VO2 relationship, but the difference between the study of Bickham et al.[81] and the study of Green and and Dawson[82] comDawson[82] is that Green . pared the linear PO-VO2 relationship based on exercise intensities below LT with the linear relationship based on exercise intensities above LT and that Bickham et al.[81] only used exercise intensities above LT for the regression lines that were based on two or four submaximal bouts. The other difference is that Russell et al.[14] and Bickham et al.[81] used a fixed value for the y-intercept, which influences the characteristics of the regression line. The y-intercept was forced to be 5.1 mL kg-1 min-1[3] in the study of Russell et al.[14] and the mean y-intercept was 5.5 – 0.4 mL kg-1 min-1 in the study of Bickham et al.,[81] where the mean y-intercept was 6.1 – 2.3 mL kg-1 min-1 in the study of Buck and McNaughton.[47] In the study of Green and Dawson,[82] three different y-intercept values were found for data below, at or above LT, which were 8.8, 7.2 and 2.6 mL kg-1 min-1 in well trained cyclists, and 9.5, 8.7 and 8.0 mL kg-1 min-1 in untrained men. These data suggest that there is large variation in the y-intercept . when this value is not determined from resting VO2 data or set at a fixed value of 5.1 mL kg-1 min-1 as has been reported by Medbø et al.[3] The changes in y-intercept for the data below, at or above LT were also accompanied by changes in the slope of the regression line, and even small changes. resulted in large differences in the estimated VO2 demand due to extrapolation of . the regression line. The robustness of the PO-VO2 relationship increased with the use of a fixed. y-intercept, the power of predicting the VO2 demand improved by approximately 45% in the study of Russell et al.[14] when a fixed value of the y-intercept was forced in the linear regression line. Although, it has to be mentioned that even with the use of a fixed y-intercept there was still a

   

 

    

ª 2010 Adis Data Information BV. All rights reserved.

 

95% CI of 12.88 – 2.93 mL kg-1. min-1 with a mean value for the estimated VO2 demand of 71.28 – 9.49 mL kg-1 min-1.[14] Based on the results of the described studies, we can conclude that there is an effect of changing the number . of submaximal exercise bouts on the linear PO-VO2 relationship and thus MAOD. This effect is .most likely due to the non-linearity of the PO-VO2 relationship for exercise intensities above LT. The effect of reducing the number of submaximal exercise bouts is reduced when a fixed value of the y-intercept has been used to establish the linear regression line. Thus, . the robustness of the VO2 demand prediction increases with the use of a fixed y-intercept, since values at the extremes of the regression line can . profoundly affect the extrapolated VO2 demand.

 

3. The Influence of Supramaximal Exercise Protocol on the Computed Maximal Accumulated Oxygen Deficit (MAOD) Values In the original protocol of Medbø et al.,[3] a supramaximal constant intensity exercise bout was used to determine the MAOD. In more recent publications, different exercise protocols have been used to determine the MAOD. In this section we will discuss the effect of different supramaximal exercise protocols on the computed MAOD values. Craig et al.[52] compared the MAOD .of a supramaximal constant intensity (115% VO2max) exercise test with the anaerobic capacity attained during 70, 120 and 300 seconds of supramaximal all-out cycling. Twelve of their eighteen subjects were high performance track endurance cyclists and the other six were sprint cyclists. Track endurance cyclists attained the highest MAOD values during the 300-second all-out protocol, and the sprint cyclists reached their highest MAOD values during the 70-second all-out protocol. This suggests that it is important to determine the MAOD during an exercise protocol that is specific for the athlete’s event. In a study of Gastin et al.,[37] untrained and endurance trained subjects performed supramaximal constant intensity and supramaximal Sports Med 2010; 40 (4)

Noordhof et al.

294

all-out exercise protocols for the determination of anaerobic capacity. They found no significant differences between the mean MAOD values of the different exercise protocols. However, closer analyses of the individual data showed that individual subjects may perform better on a specific protocol. Part two of their study showed that untrained subjects achieved a higher MAOD during the supramaximal all-out protocol and that endurance-trained subjects attained a higher MAOD during the supramaximal constant intensity protocol.[37] The effect of a continuous ramp exercise protocol on the computed value of the anaerobic capacity was investigated by Pouilly and Busso.[22] No significant difference was found between the MAOD attained during the supramaximal constant intensity protocol and during two ramp exercise protocols. However, the variability in MAOD between the two ramp exercise protocols was very high (54.5%).[22] Tabata et al.[11] even showed that a high intensity intermittent exercise protocol could be used to determine anaerobic capacity. The MAOD of intermittent exercise was calculated by . subtracting the difference between the VO2 de. mand at rest . and the VO2 during the rest periods from the VO2 deficit of each bout of exercise. The attainment of anaerobic capacity depends on the duration of the exercise intervals and the duration of the resting periods.[11] In summary, we can say that different supramaximal exercise protocols (i.e., constant, all-out, ramp and intermittent) can be used to determine the MAOD with broadly comparable results. Nevertheless, it is probably important to choose an exercise protocol that is specific for the athlete’s event. 4. The Validity of the MAOD Method In order to be able to use the MAOD method for determining anaerobic capacity we have to be sure that the method is valid. Is the MAOD method correct in measuring what it is designed to measure, that is, anaerobic capacity? Bangsbo et al.[67] evaluated the validity of the MAOD method by comparing the MAOD atª 2010 Adis Data Information BV. All rights reserved.

tained by untrained males during a supramaximal constant intensity one-legged knee-extension protocol, during which only the quadriceps muscles were active, with the amount of anaerobic energy release estimated from metabolic measurements. The changes in ATP, creatine phosphate, inosine monophosphate and lactate concentration ([La-]) were determined from muscle biopsy samples taken from .the active muscles. Further, lactate efflux and VO2 of the active muscles were estimated from measurements of blood flow and the femoral arterialvenous difference for lactate and oxygen. In order to avoid blood flow from the lower leg from influencing the measurements, an occlusion cuff was placed just below the knee. The anaerobic ATP production of the active muscle mass was calculated from these metabolic measurements and was. related to the leg and pulmonary (wholebody) VO2 deficit of the exhaustive work bout. Bangsbo et al.[67] reported that the MAOD was closely related to the amount of anaerobically attributable energy in a single muscle group during one-legged exercise. Based on these results it was concluded that the MAOD is a quantitative expression of anaerobic capacity when exercising with a single muscle group.[67] Green et al.[83] assessed the validity of the MAOD method in well trained male cyclists by inspecting the relationship between MAOD, the amount of anaerobic ATP produced, and measures of the anaerobic potential of muscles. Muscle biopsy samples were taken from the M. vastus lateralis and analysed for ATP, PCr, creatine, [La-] and ADP. Equation 1 was used to determine the amount of anaerobic ATP produced from differences in pre- and post-exercise concentrations of muscle metabolites:[83] anATPm ¼ 1:5D½La  þ D½PCr þ ð2D½ATP D½ADPÞ (Eq: 1Þ No relationship was found between the MAOD of well trained male cyclists and the amount of anaerobically attributable ATP produced, in vitro b (muscle buffer value) or enzyme activities. However, there were significant relations between the amount of anaerobic ATP produced and in vitro b or enzyme activities, from which Sports Med 2010; 40 (4)

The MAOD Method

Green et al.[83] concluded that their measures of the muscle anaerobic ATP content are valid.[83] The results of the study of Green et al.[83] were not in agreement with the results of Medbø and Tabata[51] who found a close relationship between rates of anaerobic ATP turnover in the active muscles and the anaerobic energy production in the whole body, as measured by the MAOD method. Muscle metabolite concentrations (ATP, PCr and [La-]) were determined from muscle biopsy samples taken from the M. vastus lateralis after termination of the exhausting cycling bout, as in the study of Green et al.[83] Equation 2 was used to calculate the muscle anaerobic ATP production (anATPm) in the study of Medbø and Tabata:[51] anATPm ¼ 1:5D½La  þ D½PCr þ ½ATP (Eq: 2Þ However, there were some differences between both studies that could have resulted in different outcomes. Green et al.[83] used an extended version of the equation used by Medbø and Tabata,[51] (equation 1). Neither equation accounts for the amount of lactate released to the blood, which accounts for somewhere between 5 and 38% of the total anaerobic ATP production according to Bangsbo.[58] Thus, the estimated amount of anaerobic ATP produced is underestimated by both Green et al.[83] and Medbø and Tabata.[51] Further, it is unknown if the measured muscle metabolite concentrations are representative of the concentration in the whole muscle or of the whole muscle group[84] and the main limitation of the validation attempts of Medbø and Tabata[51] and Green et al.[83] is that the active muscle mass is unknown during whole body exercise. The amount of energy release estimated from muscle biopsy data and the MAOD were significantly correlated when the working muscle mass was estimated to be 25% of the body mass.[51] Medbø et al.[3] also used this percentage in the discussion of the metabolic components of the MAOD during running exercise. However, the estimation of the active muscle mass from total body mass is not based on direct measures and does not take into account differences in fat percentage between men and women and between different ª 2010 Adis Data Information BV. All rights reserved.

295

subjects. Additionally, the percentage of active muscle mass can also vary between subjects. The mean active muscle mass determined from magnetic resonance images in the study of Sloniger et al.[68] was 5.6 kg for horizontal running and 6.1 kg for uphill running with a mean body mass of 59.7 – 8.2 kg, which corresponds to 9.4 and 10.2% of the body mass of physically active female students. A lower percentage of active muscle mass, 10.2% instead of 25%, results in a much higher concentration of anaerobic ATP produced per unit of active muscle mass, 141 mmol ATP/kg instead of 63.5 mmol ATP/kg, in the same subjects.[68] These uncertainties about the active muscle mass result in significantly different MAOD values. Besides that, there is also some doubt about the ratio between high energy phosphate produced per oxygen molecule consumed (P/O), used for the conversion of oxygen deficit in mL O2 eq kg-1 to moles. P/O ratios between 2.5 and 3.0 mmol ATP (mmol O2)-1 have been used in the literature,[64,65] but the results of the study of Krustrup et al.[65] suggest that the additional recruitment of type II fibres during intense exercise may result in a lower P/O ratio, which would disprove the use of a fixed value for the P/O ratio. Further research is necessary to investigate the relationship between muscle fibre type and the P/O ratio. Another difference between studies was that Green et al.[83] determined the MAOD during a supramaximal constant intensity exercise bout of 173 – 24 seconds in duration, and Medbø and Tabata[51] established the MAOD during exhausting exercise of ~30 seconds, 1 minute and 2–3 minutes. Based on the difference in number of exhausting exercise bouts, Green et al.[83] reported that the large range of values in the study of Medbø and Tabata[51] caused a spuriously high correlation coefficient. Bangsbo[58] determined, from the data of Medbø and Tabata,[51] the relationship between the MAOD and muscle anaerobic ATP production for the three different exercise durations (~30 seconds, 1 minute and 2–3 minutes), which resulted in non-significant correlation coefficients. In addition, a 2-fold range in MAOD was found for the same muscle anaerobic ATP production.



Sports Med 2010; 40 (4)

Noordhof et al.

296

From the studies described, we must conclude that the most precise assessment of the anaerobic energy yield has been performed by Bangsbo et al.[67] during intense dynamic knee-extension exercise. From the results of this study it could be concluded that in a small muscle mass the MAOD method is a valid method to determine anaerobic capacity. 5. The Reliability of the MAOD Method Anaerobic capacity is an important determinant of high-intensity exercise. Thus, it would be helpful for coaches, athletes, and sports scientists if anaerobic capacity could be measured accurately. The MAOD method not only has to be valid but it also has to be reliable. Coaches need to be sure that test-retest differences are real differences in anaerobic capacity and are not due to random error. Doherty et al.[56] evaluated the reproducibility of the MAOD and the exercise time to exhaustion, their results are summarized in table I. Based on intraclass correlation coefficients and sample coefficients of variation they concluded that the MAOD method is a reliable measure. However, they also determined the 95% limits of agreement for the MAOD, which measures the absolute reliability. Doherty et al.[56] found that 95% of the agreement ratios had to be between 0.80 and 1.28, which means that a subject with a MAOD of 70.0 mL O2 eq kg-1 the first time, could have a MAOD between 70.0 · 0.80 = 56.0 mL O2 eq kg-1 and 70.0 · 1.28 = 89.6 mL O2 eq kg-1 during the second supramaximal exercise bout. So, based on the 95% limits of agreement they concluded that the MAOD method was not a reliable measure. In the study of Bickham et al.,[81] the effect of different methodological improvements on the . precision of the estimated VO2 demand and on the MAOD was investigated, . using 95% CIs (table I). The 95% CI of the VO2 demand, estimated from the regression line based on ten submaximal bouts and a fixed y-intercept (11-point) was 7.8 – 2.7 mL O2 kg-1 and was not significantly different between the 11-point, 5-point (four submaximal bouts and a y-intercept) and





ª 2010 Adis Data Information BV. All rights reserved.



3-point (two submaximal bouts and a y-intercept) regression line. When additional error data from the metabolic system was included, the width of . the 95% CI of the estimated VO2 demand became even larger, namely 13.9 – 2.8 mL O2 kg-1 for the 11-point regression line, which shows that half of the possible error is related to the regression line. It also became clear that a relatively small 95% CI . (7.8 mL O2 kg-1) for the estimated VO2 demand could result in a large 95% CI for the MAOD (22.4 mL O2 eq kg-1), the increase in the 95% CI will be larger when the length of the supramaximal exercise bout is longer. The study of Bickham et al.[81] showed that the 95% CI of the MAOD is relatively large, which suggests that estimating anaerobic capacity with the MAOD method is not very accurate. Although the precision of a measurement influences the reliability, the 95% CI is not a complete measure of relia[81] did not bility. Unfortunately, Bickham et al. . determine the reproducibility of the VO2 demand and MAOD. Jacobs et al.[85] determined the MAOD during two trials in order to evaluate the reliability of the MAOD method. The test-retest correlation coefficient was 0.97, from which they concluded that the MAOD method is a reliable method. This conclusion is supported by the study of Weber and Schneider,[8] who investigated the reliability of the MAOD method at POs calculated . to require 110% and 120% VO2peak with the use of intraclass correlation coefficients. The intraclass correlation coefficients for the supramaximal exercise tests were 0.95 and 0.97, respectively. This study showed that the MAOD method resulted in highly repeatable values.[8] However, the conclusions of Jacobs et al.[85] and Weber and Schneider[8] did not support the findings of Doherty et al.[56] In a letter to the editor, Doherty and Smith[55] questioned the outcomes of the study of Weber and Schneider.[8] Both Weber and Schneider[8] and Jacobs et al.[85] used a traditional measure of reliability, namely intraclass correlation coefficients, which has some limitations.[55] One of the limitations is that the correlation coefficient is dependent on the range of values that are attained by the subjects. With a very heterogeneous



Sports Med 2010; 40 (4)

Reference (y)

Subjects (n)

Ergometer

Number of submaximal tests

Length of submaximal tests (min)

Protocol

MAOD (mL O2 eq kg-1)

CIa O2 demand (mL O2 eq kg-1)

Muscle biopsies

Bangsbo et al.[53] (1993)

Distance runners (14) Soccer players (150) Oarsmen (5) Cyclists (3)

Treadmill (5% gradient) Rowing ergometer Cycle ergometer

6–10

6

Constant PO protocol (120–140% of the PO at . VO2max)

Soccer players (running) 49.5 – 3.0 Runners 51.9 – 3.8 Oarsmen running) 47.3 – 6.3 Oarsmen (rowing) 64.1 – 4.4 Cyclists 56.5 – 8.4

Soccer players (running) [62.2, 68.8] Runners [75.6, 80.6] Oarsmen (running) [66.3, 71.7] Oarsmen (rowing) [63.1, 69.7] Cyclists [62.9, 84.9] (mL O2 eq min-1 kg-1)

Runners, Musculus gastrocnemius Cyclists, Musculus vastus lateralis (fibre type distribution, capillaries, buffer capacity and enzyme activities)

11-point regression 43.3 – 5.3 5-point regression 44.2 – 7.2 3-point regression 42.6 – 8.2

11-point regression 7.8 – 2.7 [95% CI 1.96 SEp] 5-point regression 7.7 – 3.5 3-point regression 6.7 – 2.9 With inclusion of error data: 13.9 – 2.8, 13.8 – 3.6, and 12.8 – 2.7 (mL O2 eq min-1 kg-1)

Bickham et al.[81] (2002)

Trained male distance runners (7)

Treadmill (1% gradient)

10 and a fixed y-intercept, 4 and a fixed y-intercept, and 2 and a fixed y-intercept

4

Constant intensity protocol . (110% VO2max)





 

The MAOD Method

ª 2010 Adis Data Information BV. All rights reserved.

Table I. Quality criteria of studies challenging the maximal accumulated oxygen deficit (MAOD) method



 

Craig and Morgan[69] (1998)

Male distance runners (9)

Treadmill (1% and 10.5% gradient)

8

6

(1%, linear) 45.0 – 6.9 (10.5%, linear) 63.2 – 10.6 (1%, curvilinear) 59.3 – 10.1 (10.5%, curvilinear) 93.6 – 19.7

(1%, linear) [65.4, 80.2] (10.5%, linear) [76.8, 94.8] (1%, curvilinear) [69.8, 87.8] (10.5%, curvilinear) [83.8, 120.6] (mL O2 eq min-1 kg-1)

 

Continued next page

297

Sports Med 2010; 40 (4)

Constant intensity protocol (2 trials: 1% and 10.5% gradient)

Reference (y)

Subjects (n)

Ergometer

Number of submaximal tests

Length of submaximal tests (min)

Protocol

MAOD (mL O2 eq kg-1)

CIa O2 demand (mL O2 eq kg-1)

Doherty et al.[56] (2000)

Physically active male students (15)

Treadmill (10.5% gradient)

3 and a fixed y-intercept

6

Constant intensity protocol (3 trials: 125% . VO2max)

Trial 1, 69.0 – 13.1 Trial 2, 71.4 – 12.5 Trial 3, 70.4 – 15.0

Agreement ratios MAOD 0.8, 1.28 (log-transformed)

Green and Dawson[82] (1995)

Male cyclists (10), untrained men (9)

Cycle ergometer

6 cyclists 5 untrained men

4

Green and Dawson[79] (1996)

Male cyclists (7)

Cycle ergometer

5–8

4–15; CON4, CON10, DISCON4, DISCON6, DISCON10, DISCON15

Green et al.[83] (1996)

Well trained male cyclists (10) [2 pursuitists, 1 sprinter, 7 endurance cyclists]

Cycle ergometer

¨ zyener O et al.[71] (2003)

Male subjects (10)

Cycle ergometer

5

4

10

Muscle biopsies



0.04, CON4 0.06, DISCON4 CI between 0.06 and 0.07 mL O2 eq min-1 W-1 for DISCON6, DISCON10, and DISCON15

 

Constant PO trial (432 – 41 W)

55.2 – 10.3

Three constant supra-lactate threshold work rates (moderate, heavy, very heavy and severe)

Moderate, 0.81 – 0.2 L Heavy, -1.70 – 1.6 L Very heavy, -1.43 – 2.3 L Severe, 1.69 – 0.4 L



[164.9, 286.9] (mL kg-1)

M. vastus lateralis 202.7 – 46.9 mmol ATP kg-1 dw r = -0.38 (p >0.05)



The reported confidence intervals (CIs) of the studies of Bangsbo et al.[53] (1993), Craig and Morgan[69] (1998) and Green et al.[83] (1996) are values calculated from the published standard deviations of the energy demands (mean – 1.96 SD). Bickham et al.[81] (2002), Doherty et al.[56] (2000) and Green and Dawson[79] (1996) did report the length of the 95% CIs.



ATP = adenosine triphosphate; CI = confidence.interval; CON = continuous; DISCON = discontinuous; dw = dry weight; O2 eq = oxygen equivalent; PO = power output; r = correlation coefficient; SEp = standard error of prediction; VO2max = maximum oxygen uptake.

Noordhof et al.

Sports Med 2010; 40 (4)

a

8–12



298

ª 2010 Adis Data Information BV. All rights reserved.

Table I. Contd

The MAOD Method

group of subjects (i.e. males and females, individuals of varying fitness levels) it is easy to find a strong correlation coefficient.[86] Atkinson and Nevill[86] suggested the use of the standard error of measurement when heteroscedasticity is absent and the use of the coefficient of variation (CV) and the limits of agreement method when heteroscedasticity is present. These measures all evaluate the absolute reliability and are not influenced by the range of measurements. In summary, most studies[8,85] used traditional measures of reliability, which are not sensitive enough for assessing the reliability.[86] Adequate measures of reliability are the CV, the limits of agreement method and the standard error of measurement.[86] Doherty et al.[56] used the limits of agreement method (i.e. 95% CI), which resulted in the conclusion that the MAOD method is not a reliable measure of anaerobic capacity. However, the study of Doherty et al.[56] is one of the first studies which assessed the reliability of the MAOD method with an adequate measure of reliability and further research is therefore necessary to study the reliability of the MAOD method. 6. Conclusions The purpose of this review was to evaluate the procedures and outcomes of studies that used the MAOD method to determine anaerobic capacity, as proposed by Medbø et al.[3] From the reviewed literature it can be concluded that the most widely used approach to determine anaerobic capacity (the MAOD method), although clearly reasonable from a conceptual standpoint, is probably not (as currently used) a fully defensible method to determine anaerobic capacity. Further research is necessary to determine how to approach the problem of measuring anaerobic capacity. Clearly the most profitable areas of immediate concern are standardization of the number, duration and intensity of submaximal . exercise bouts necessary to define the . PO-VO2 relationship, whether forcing the PO-VO2 relationship into a linear analysis mode is the correct approach to allowing computation of an extrapolated value . for VO2 demand, and the relative PO (e.g. duraª 2010 Adis Data Information BV. All rights reserved.

299

tion) and character (square wave, all-out, free range) of the supramaximal exercise bout. We suggest the use of the following methodology for the determination of anaerobic capacity with the MAOD method. Relatively short submaximal exercise bouts are .preferred for the construction of the linear PO-VO2 relationship, . as VO2 shows a secondary increase after 3 minutes of exercise at the highest submaximal exercise intensities, which will lead to non-linearity . of the PO. VO2 relationship. To construct a robust PO-VO2 relationship, about ten submaximal exercise tests at intensities evenly distributed be. tween 30 and 90% VO2max are necessary, as is the use of . a fixed y-intercept, determined from resting VO2 data. Furthermore, it is advisable to use horizontal running or cycling exercise at a fixed and relatively low pedalling frequency as this will . also minimize the VO2 SC. Linear regression . analysis can be used to determine the PO-VO2 relationship, as it has been shown that a curvilinear fit did not improve the relationship.[69] A carefully considered decision has to be made about the supramaximal exercise protocol used to determine the MAOD, it is probably best to use a protocol and exercise modality specific for the athlete’s chosen event. The last recommendation is to evaluate the reliability of the MAOD method with the use of measures of absolute reliability, when heteroscedasticity is present the CV or the limits of agreement method (ratios) should be used and when heteroscedasticity is absent the standard error of measurement or the limits of agreement method (absolute values) should be reported.[86] Thus far, the MAOD method is the most widely used and probably the best noninvasive method to determine anaerobic capacity. Acknowledgements No sources of funding were used to assist in the preparation of this review. The authors have no conflicts of interest that are directly relevant to the content of this review

References 1. Krogh A, Lindhard J. The changes in respiration at the transition from work to rest. J Physiol 1920 May 18; 53 (6): 431-9

Sports Med 2010; 40 (4)

300

2. Hermansen L. Anaerobic energy release. Med Sci Sports Exerc 1969; 1 (1): 32-5 3. Medbø JI, Mohn AC, Tabata I, et al. Anaerobic capacity determined by maximal accumulated O2 deficit. J Appl Physiol 1988 Jan; 64 (1): 50-60 4. Withers RT, Sherman WM, Clark DG, et al. Muscle metabolism during 30, 60, and 90 s of maximal cycling on an airbraked ergometer. Eur J Appl Physiol 1991; 63: 354-62 5. Withers RT, Van der Ploeg G, Finn JP. Oxygen deficits incurred during 45, 60, 75 and 90-s maximal cycling on an airbraked ergometer. Eur J Appl Physiol Occup Physiol 1993; 67 (2): 185-91 6. Woolford SM, Withers RT, Craig NP, et al. Effect of pedal cadence on the accumulated oxygen deficit, maximal aerobic power and blood lactate transition thresholds of high-performance junior endurance cyclists. Eur J Appl Physiol Occup Physiol 1999 Sep; 80 (4): 285-91 7. Weber CL, Schneider DA. Increases in maximal accumulated oxygen deficit after high-intensity interval training are not gender dependent. J Appl Physiol 2002 May; 92 (5): 1795-801 8. Weber CL, Schneider DA. Reliability of MAOD measured at 110% and 120% of peak oxygen uptake for cycling. Med Sci Sports Exerc 2001 Jun; 33 (6): 1056-9 9. Weber CL, Schneider DA. Maximal accumulated oxygen deficit expressed relative to the active muscle mass for cycling in untrained male and female subjects. Eur J Appl Physiol 2000 Jul; 82 (4): 255-61 10. Vuorimaa T, Vasankari T, Rusko H. Comparison of physiological strain and muscular performance of athletes during two intermittent running exercises at the velocity associated with VO2max. Int J Sports Med 2000 Feb; 21 (2): 96-101 11. Tabata I, Irisawa K, Kouzaki M, et al. Metabolic profile of high intensity intermittent exercises. Med Sci Sports Exerc 1997 Mar; 29 (3): 390-5 12. Tabata I, Nishimura K, Kouzaki M, et al. Effects of moderate-intensity endurance and high-intensity intermittent training on anaerobic capacity and VO2max. Med Sci Sports Exerc 1996 Oct; 28 (10): 1327-30 13. Scott CB, Roby FB, Lohman TG, et al. The maximally accumulated oxygen deficit as an indicator of anaerobic capacity. Med Sci Sports Exerc 1991 May; 23 (5): 618-24 14. Russell AP, Le Rossignol P, Lo SK. The precision of estimating the total energy demand: implications for the determination of the accumulated oxygen deficit [online]. JEP online 2000; 3 (2). Available from URL: http:// faculty.css.edu/tboone2/asep/JEPRussell.html [Accessed 2010 Jan 12] 15. Roberts AD, Clark SA, Townsend NE, et al. Changes in performance, maximal oxygen uptake and maximal accumulated oxygen deficit after 5, 10 and 15 days of live high/train low altitude exposure. Eur J Appl Physiol 2003 Jan; 88 (4-5): 390-5 16. Ravier G, Dugue B, Grappe F, et al. Maximal accumulated oxygen deficit and blood responses of ammonia, lactate and pH after anaerobic test: a comparison between international and national elite karate athletes. Int J Sports Med 2006 Oct; 27 (10): 810-7 17. Ramsbottom R, Nevill AM, Nevill ME, et al. Accumulated oxygen deficit and short-distance running performance. J Sports Sci 1994 Oct; 12 (5): 447-53

ª 2010 Adis Data Information BV. All rights reserved.

Noordhof et al.

18. Ramsbottom R, Nevill AM, Seager RD, et al. Effect of training on accumulated oxygen deficit and shuttle run performance. J Sports Med Phys Fitness 2001 Sep; 41 (3): 281-90 19. Pripstein LP, Rhodes EC, McKenzie DC, et al. Aerobic and anaerobic energy during a 2-km race simulation in female rowers. Eur J Appl Physiol Occup Physiol 1999 May; 79 (6): 491-4 20. Olesen HL. Accumulated oxygen deficit increases with inclination of uphill running. J Appl Physiol 1992 Sep; 73 (3): 1130-4 21. Olesen HL, Raabo E, Bangsbo J, et al. Maximal oxygen deficit of sprint and middle distance runners. Eur J Appl Physiol Occup Physiol 1994; 69: 140-6 22. Pouilly JP, Busso T. Accumulated oxygen deficit during ramp exercise. Int J Sports Med 2008 Jan; 29 (1): 16-20 23. Pizza FX, Naglieri TA, Holtz RW, et al. Maximal accumulated oxygen deficit of resistance-trained men. Can J Appl Physiol 1996 Oct; 21 (5): 391-402 24. Ogita F, Hara M, Tabata I. Anaerobic capacity and maximal oxygen uptake during arm stroke, leg kicking and whole body swimming. Acta Physiol Scand 1996 Aug; 157 (4): 435-41 25. Ogita F, Onodera T, Tabata I. Effect of hand paddles on anaerobic energy release during supramaximal swimming. Med Sci Sports Exerc 1999 May; 31 (5): 729-35 26. Ogawa T, Hayashi K, Ichinose M, et al. Metabolic response during intermittent graded sprint running in moderate hypobaric hypoxia in competitive middle-distance runners. Eur J Appl Physiol 2007 Jan; 99 (1): 39-46 27. Nevill AM, Ramsbottom R, Nevill ME, et al. The relative contributions of anaerobic and aerobic energy supply during track 100-, 400- and 800-m performance. J Sports Med Phys Fitness 2008 Jun; 48 (2): 138-42 28. Naughton GA, Carlson JS. Anaerobic capacity assessment in male and female children with all-out isokinetic cycling exercise. Aust J Sci Med Sport 1995 Dec; 27 (4): 83-7 29. Naughton GA, Carlson JS, Buttifant DC, et al. Accumulated oxygen deficit measurements during and after highintensity exercise in trained male and female adolescents. Eur J Appl Physiol Occup Physiol 1997; 76 (6): 525-31 30. Minahan C, Chia M, Inbar O. Does power indicate capacity? 30-s Wingate anaerobic test vs. maximal accumulated O2 deficit. Int J Sports Med 2007 Oct; 28 (10): 836-43 31. Minahan C, Wood C. Strength training improves supramaximal cycling but not anaerobic capacity. Eur J Appl Physiol 2008 Apr; 102 (6): 659-66 32. Maxwell NS, Nimmo MA. Anaerobic capacity: a maximal anaerobic running test versus the maximal accumulated oxygen deficit. Can J Appl Physiol 1996 Feb; 21 (1): 35-47 33. Marth PD, Woods RR, Hill DW. Influence of time of day on anaerobic capacity. Percept Mot Skills 1998 Apr; 86 (2): 592-4 34. Levine BD, Stray-Gundersen J. ‘‘Living high-training low’’: effect of moderate-altitude acclimatization with low-altitude training on performance. J Appl Physiol 1997 Jul; 83 (1): 102-12 35. Glaister M, Stone MH, Stewart AM, et al. Aerobic and anaerobic correlates of multiple sprint cycling performance. J Strength Cond Res 2006 Nov; 20 (4): 792-8

Sports Med 2010; 40 (4)

The MAOD Method

36. Glaister M, Stone MH, Stewart AM, et al. The influence of endurance training on multiple sprint cycling performance. J Strength Cond Res 2007 May; 21 (2): 606-12 37. Gastin PB, Costill DL, Lawson DL, et al. Accumulated oxygen deficit during supramaximal all-out and constant intensity exercise. Med Sci Sports Exerc 1995 Feb; 27 (2): 255-63 38. Gastin PB, Lawson DL. Influence of training status on maximal accumulated oxygen deficit during all-out cycle exercise. Eur J Appl Physiol Occup Physiol 1994; 69 (4): 321-30 39. Gastin PB, Lawson DL. Variable resistance all-out test to generate accumulated oxygen deficit and predict anaerobic capacity. Eur J Appl Physiol Occup Physiol 1994; 69 (4): 331-6 40. Gardner A, Osborne M, D’Auria S, et al. A comparison of two methods for the calculation of accumulated oxygen deficit. J Sports Sci 2003 Mar; 21 (3): 155-62 41. Friedmann B, Frese F, Menold E, et al. Effects of acute moderate hypoxia on anaerobic capacity in endurancetrained runners. Eur J Appl Physiol 2007 Sep; 101 (1): 67-73 42. Faina M, Billat V, Squadrone R, et al. Anaerobic contribution to the time to exhaustion at the minimal exercise intensity at which maximal oxygen uptake occurs in elite cyclists, kayakists and swimmers. Eur J Appl Physiol Occup Physiol 1997; 76 (1): 13-20 43. Craig NP, Norton KI, Bourdon PC, et al. Aerobic and anaerobic indices contributing to track endurance cycling performance. Eur J Appl Physiol Occup Physiol 1993; 67 (2): 150-8 44. Calbet JA, Chavarren J, Dorado C. Fractional use of anaerobic capacity during a 30- and a 45-s Wingate test. Eur J Appl Physiol Occup Physiol 1997; 76 (4): 308-13 45. Calbet JA, De Paz JA, Garatachea N, et al. Anaerobic energy provision does not limit Wingate exercise performance in endurance-trained cyclists. J Appl Physiol 2003 Feb; 94 (2): 668-76 46. Buck D, Mc Naughton L. Maximal accumulated oxygen deficit must be calculated using 10-min time periods. Med Sci Sports Exerc 1999 Sep; 31 (9): 1346-9 47. Buck D, Mc Naughton LR. Changing the number of submaximal exercise bouts effects calculation of MAOD. Int J Sports Med 1999 Jan; 20 (1): 28-33 48. Aisbett B, Le Rossignol P. Estimating the total energy demand for supra-maximal exercise using the VO2-power regression from an incremental exercise test. J Sci Med Sport 2003 Sep; 6 (3): 343-7 49. Aisbett B, Le Rossignol P, Sparrow WA. The influence of pacing during 6-minute supra-maximal cycle ergometer performance. J Sci Med Sport 2003 Jun; 6 (2): 187-98 50. Medbø JI, Tabata I. Relative importance of aerobic and anaerobic energy release during short-lasting exhausting bicycle exercise. J Appl Physiol 1989 Nov; 67 (5): 1881-6 51. Medbø JI, Tabata I. Anaerobic energy release in working muscle during 30 s to 3 min of exhausting bicycling. J Appl Physiol 1993 Oct; 75 (4): 1654-60 52. Craig NP, Norton KI, Conyers RA, et al. Influence of test duration and event specificity on maximal accumulated oxygen deficit of high performance track cyclists. Int J Sports Med 1995 Nov; 16 (8): 534-40

ª 2010 Adis Data Information BV. All rights reserved.

301

53. Bangsbo J, Michalsik L, Petersen A. Accumulated O2 deficit during intense exercise and muscle characteristics of elite athletes. Int J Sports Med 1993 May; 14 (4): 207-13 54. Bosquet L, Duchene A, Delhors PR, et al. A comparison of methods to determine maximal accumulated oxygen deficit in running. J Sports Sci 2008 Apr; 26 (6): 663-70 55. Doherty M, Smith PM. The reliability of cycling maximal accumulated oxygen deficit (MAOD) and time to exhaustion (T(lim)) in untrained subjects. Med Sci Sports Exerc 2001 Oct; 33 (10): 1794-5 56. Doherty M, Smith PM, Schroder K. Reproducibility of the maximum accumulated oxygen deficit and run time to exhaustion during short-distance running. J Sports Sci 2000 May; 18 (5): 331-8 57. Bangsbo J. Is the O2 deficit an accurate quantitative measure of the anaerobic energy production during intense exercise? J Appl Physiol 1992 Sep; 73 (3): 1207-9 58. Bangsbo J. Oxygen deficit: a measure of the anaerobic energy production during intense exercise? Can J Appl Physiol 1996 Oct; 21 (5): 350-63; discussion 64-9 59. Parolin ML, Chesley A, Matsos MP, et al. Regulation of skeletal muscle glycogen phosphorylase and PDH during maximal intermittent exercise. Am J Physiol 1999 Nov; 277 (5 Pt 1): E890-900 60. Parolin ML, Spriet LL, Hultman E, et al. Regulation of glycogen phosphorylase and PDH during exercise in human skeletal muscle during hypoxia. Am J Physiol Endocrinol Metab 2000 Mar; 278 (3): E522-34 61. Spriet LL, Lindinger MI, McKelvie RS, et al. Muscle glycogenolysis and H+ concentration during maximal intermittent cycling. J Appl Physiol 1989 Jan; 66 (1): 8-13 62. Putman CT, Jones NL, Lands LC, et al. Skeletal muscle pyruvate dehydrogenase activity during maximal exercise in humans. Am J Physiol 1995 Sep; 269 (3 Pt 1): E458-68 63. Greenhaff PL, Campbell-O’Sullivan SP, ConstantinTeodosiu D, et al. An acetyl group deficit limits mitochondrial ATP production at the onset of exercise. Biochem Soc Trans 2002 Apr; 30 (2): 275-80 64. Gonzalez-Alonso J, Quistorff B, Krustrup P, et al. Heat production in human skeletal muscle at the onset of intense dynamic exercise. J Physiol 2000 Apr 15; 524 (Pt 2): 603-15 65. Krustrup P, Ferguson RA, Kjaer M, et al. ATP and heat production in human skeletal muscle during dynamic exercise: higher efficiency of anaerobic than aerobic ATP resynthesis. J Physiol 2003; 15: 255-69 66. Medbø JI, Burgers S. Effect of training on the anaerobic capacity. Med Sci Sports Exerc 1990 Aug; 22 (4): 501-7 67. Bangsbo J, Gollnick PD, Graham TE, et al. Anaerobic energy production and O2 deficit-debt relationship during exhaustive exercise in humans. J Physiol 1990 Mar; 422: 539-59 68. Sloniger MA, Cureton KJ, Prior BM, et al. Anaerobic capacity and muscle activation during horizontal and uphill running. J Appl Physiol 1997 Jul; 83 (1): 262-9 69. Craig IS, Morgan DW. Relationship between 800-m running performance and accumulated oxygen deficit in middle-distance runners. Med Sci Sports Exerc 1998 Nov; 30 (11): 1631-6 70. Sloniger MA, Cureton KJ, Prior BM, et al. Lower extremity muscle activation during horizontal and uphill running. J Appl Physiol 1997 Dec; 83 (6): 2073-9

Sports Med 2010; 40 (4)

302

71. O¨zyener F, Rossiter HB, Ward SA, et al. Negative accumulated oxygen deficit during heavy and very heavy intensity cycle ergometry in humans. Eur J Appl Physiol 2003 Sep; 90 (1-2): 185-90 72. Pringle JS, Carter H, Doust JH, et al. Oxygen uptake kinetics during horizontal and uphill treadmill running in humans. Eur J Appl Physiol 2002 Nov; 88 (1-2): 163-9 73. Zoladz JA, Duda K, Majerczak J. VO2/power output relationship and the slow component of oxygen uptake kinetics during cycling at different pedaling rates: relationship to venous lactate accumulation and blood acid-base balance. Physiol Res 1998; 47 (6): 427-38 74. Pringle JS, Doust JH, Carter H, et al. Effect of pedal rate on primary and slow-component oxygen uptake responses during heavy-cycle exercise. J Appl Physiol 2003 Apr; 94 (4): 1501-7 75. Billat VL, Mille-Hamard L, Petit B, et al. The role of cadence on the VO2 slow component in cycling and running in triathletes. Int J Sports Med 1999 Oct; 20 (7): 429-37 76. Billat VL, Richard R, Binsse VM, et al. The V(O2) slow component for severe exercise depends on type of exercise and is not correlated with time to fatigue. J Appl Physiol 1998 Dec; 85 (6): 2118-24 77. Whipp BJ, Wasserman K. Oxygen uptake kinetics for various intensities of constant-load work. J Appl Physiol 1972; 33 (3): 351-6 78. Barstow TJ, Mole´ PA. Linear and nonlinear characteristics of oxygen uptake kinetics during heavy exercise. J Appl Physiol 1991 Dec; 71 (6): 2099-106

ª 2010 Adis Data Information BV. All rights reserved.

Noordhof et al.

79. Green S, Dawson BT. Methodological effects on the VO2power regression and the accumulated O2 deficit. Med Sci Sports Exerc 1996 Mar; 28 (3): 392-7 80. Doherty M. The effects of caffeine on the maximal accumulated oxygen deficit and short-term running performance. Int J Sport Nutr 1998 Jun; 8 (2): 95-104 81. Bickham D, Le Rossignol P, Gibbons C, et al. Re-assessing accumulated oxygen deficit in middle-distance runners. J Sci Med Sport 2002; 5 (4): 372-82 82. Green S, Dawson BT. The oxygen uptake-power regression in cyclists and untrained men: implications for the accumulated oxygen deficit. Eur J Appl Physiol Occup Physiol 1995; 70 (4): 351-9 83. Green S, Dawson BT, Goodman C, et al. Anaerobic ATP production and accumulated O2 deficit in cyclists. Med Sci Sports Exerc 1996 Mar; 28 (3): 315-21 84. Medbø JI. Is the maximal accumulated oxygen deficit an adequate measure of the anaerobic capacity? Can J Appl Physiol 1996 Oct; 21 (5): 370-83; discussion 84-8 85. Jacobs I, Bleue S, Goodman J. Creatine ingestion increases anaerobic capacity and maximum accumulated oxygen deficit. Can J Appl Physiol 1997 Jun; 22 (3): 231-43 86. Atkinson G, Nevill AM. Statistical methods for assessing measurement error (reliability) in variables relevant to sports medicine. Sports Med 1998; 26 (4): 217-38

Correspondence: Dionne A. Noordhof, MSc, Van der Boechorststraat 7, 1081 BT Amsterdam, the Netherlands. E-mail: [email protected]

Sports Med 2010; 40 (4)

Sports Med 2010; 40 (4): 303-326 0112-1642/10/0004-0303/$49.95/0

REVIEW ARTICLE

ª 2010 Adis Data Information BV. All rights reserved.

A Biomechanical Evaluation of Resistance Fundamental Concepts for Training and Sports Performance David M. Frost,1,2 John Cronin1,3 and Robert U. Newton1 1 School of Exercise, Biomedical and Health Sciences, Edith Cowan University, Joondalup, Western Australia, Australia 2 Department of Kinesiology, University of Waterloo, Waterloo, Ontario, Canada 3 School of Sport and Recreation, Auckland University of Technology, Auckland, New Zealand

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Constant Resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Concentric Only . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1.1 Kinematics and Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1.2 Muscle Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Stretch-Shortening Cycle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.1 Kinematics and Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.2 Muscle Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Ballistic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.1 Kinematics and Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.2 Muscle Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Accommodating Resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Hydraulic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1 Kinematics and Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2 Muscle Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Isokinetic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 Kinematics and Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2 Muscle Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Variable Resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Cam- or Lever-Based System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.1 Kinematics and Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.2 Muscle Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Band and Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1 Kinematics and Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2 Muscle Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Pneumatic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 Kinematics and Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.2 Muscle Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

303 305 306 306 310 311 311 313 313 314 315 315 316 316 316 316 317 318 318 319 319 319 319 319 320 321 321 322 323

Newton’s second law of motion describes the acceleration of an object as being directly proportional to the magnitude of the net force, in the same direction as the net force and inversely proportional to its mass (a = F/m). With respect to linear motion, mass is also a numerical representation of an

Frost et al.

304

object’s inertia, or its resistance to change in its state of motion and directly proportional to the magnitude of an object’s momentum at any given velocity. To change an object’s momentum, thereby increasing or decreasing its velocity, a proportional impulse must be generated. All motion is governed by these relationships, independent of the exercise being performed or the movement type being used; however, the degree to which this governance affects the associated kinematics, kinetics and muscle activity is dependent on the resistance type. Researchers have suggested that to facilitate the greatest improvements to athletic performance, the resistance-training programme employed by an athlete must be adapted to meet the specific demands of their sport. Therefore, it is conceivable that one mechanical stimulus, or resistance type, may not be appropriate for all applications. Although an excellent means of increasing maximal strength and the rate of force development, freeweight or mass-based training may not be the most conducive means to elicit velocity-specific adaptations. Attempts have been made to combat the inherent flaws of free weights, via accommodating and variable resistancetraining devices; however, such approaches are not without problems that are specific to their mechanics. More recently, pneumatic-resistance devices (variable) have been introduced as a mechanical stimulus whereby the body mass of the athlete represents the only inertia that must be overcome to initiate movement, thus potentially affording the opportunity to develop velocity-specific power. However, there is no empirical evidence to support such a contention. Future research should place further emphasis on understanding the mechanical advantages/disadvantages inherent to the resistance types being used during training, so as to elicit the greatest improvements in athletic performance.

Dynamic strength training can be classified into the following three different categories: (i) constant, isoinertial or free-weight resistance; (ii) accommodating resistance; and (iii) variable resistance, based on the nature in which the resistance is imposed/applied on the contracting musculature,[1] thereby resulting in significant differences in the associated kinetics, kinematics and muscle activity.[2-5] However, to date, the majority of research into resistance training is conducted with the aim of examining specific performance parameters, while failing to address or even understand the biomechanical determinants underlying the resistance or muscular effort being used. An increased understanding of the biomechanical properties that govern each resistance type, and/or how they can be manipulated, will provide the researcher, clinician and practitioner with a much greater appreciation of the benefits and limitations associated with each resistance-training mode. With this in mind, we will ª 2010 Adis Data Information BV. All rights reserved.

briefly introduce some mechanical formulae that are important for understanding the mechanics of loading muscle. Newton’s second law of motion (law of acceleration) dictates that the acceleration of an object, as produced by a force, is directly proportional to the magnitude of the net force, in the same direction as the net force and inversely proportional to its mass. For example, an object with a smaller mass requires less force to elicit the same acceleration (see equation 1): Fnet ¼ m  a

(Eq: 1Þ

where Fnet is the sum of all external forces acting on the object (N), m is the mass of the object (kg) and a is the acceleration of the object (m/sec2). For example, see equation 2: X Fnet ¼ Fext ¼ Fapp þ Fload (Eq: 2Þ where Fext is equal to Fnet (N), Fapp is the force being applied by the athlete (N) and Fload is Sports Med 2010; 40 (4)

Biomechanics of Resistance

305

resistive force of the load (N). The acceleration of an object can also be defined as a first-order derivative of the velocity time curve, or the change in velocity over a specific period of time (see equation 3): a ¼ ðv2  v1 Þ=ðt2  t1 Þ (Eq: 3Þ where v2 is the velocity (m/sec) of the object at time, t2 (seconds) and v1 is the velocity (m/sec) of the object at time, t1 (seconds). Therefore, equation 1 can be expressed as follows (see equation 4): Fnet ¼ m  ðv2  v1 Þ=ðt2  t1 Þ (Eq: 4Þ By rearranging and expanding equation 4 we get equation 5: Fnet  ðt2  t1 Þ ¼ ðm  v2 Þ  ðm  v1 Þ (Eq: 5Þ Newton’s first law of motion (law of inertia) describes how an object at rest will remain at rest and an object in motion will remain in motion with the same magnitude and direction of velocity unless acted on by an unbalanced force. An object’s resistance to a change in its state of motion, defined as its inertia, is directly proportional to its mass (mass of an object is a numerical value of its inertia). Once an object acquires a velocity, it can also be characterized as having momentum, which is the product of its mass and velocity (see equation 6): p¼mv (Eq: 6Þ where p is the momentum of the object (kg m/sec). By comparing equation 6 with equation 5, it is evident that the change in the momentum of an object is directly proportional to the magnitude and duration of the net force on the object (impulse); however, for resistance-training applications in which the net force does not remain constant (as a result of changing mechanical and physiological advantage) during the muscular action, equation 5 must be integrated with respect to time to provide a representation of the change in momentum that occurs over the entire muscular effort (see equation 7): Z t2 Z t2 Fnet  dt ¼ m  dv (Eq: 7Þ



t1

t1

R t2 where t1 Fnet  dt is the integral of the forcetime curve (n sec) between time t2 and time t1,



ª 2010 Adis Data Information BV. All rights reserved.

R t2 and t1 dv is the integral of the velocity (m/sec) between time t2 and time t1. Newton’s third law (law of action/reaction) affirms that for every action, there is an equal and opposite reaction, implying that each time an athlete applies a force against an object, an equivalent force in the opposite direction is exerted on the athlete by the object (see equation 8): (Eq: 8Þ Fapp ¼ Freact where Fapp is the sum of the net force and the resistive force of the load and Freact is the force exerted by the object on the athlete. The product of this applied force and the object’s resultant movement velocity (v) is defined as the power output (P) of the system, as measured in watts (W) for a specific instant in time (see equation 9). (Eq: 9Þ P ¼ Fapp  v In the subsequent sections of this review, these kinetic and kinematic formulae will be discussed with respect to each resistance type. Understanding the biomechanical differences between these resistance types will allow researchers and strength coaches alike to utilize the resistancetraining mode that is most conducive to their specific exercise or training goal. Although not the focus of this review, it is important to note that the calculation and reporting of some variables (e.g. power) will depend on the inclusion or exclusion of bodyweight. An athlete’s performance is often dictated by their ability to generate sufficient force to move a load at a given velocity, and for the purpose of understanding various resistance types, it must be realized that bodyweight is a load (bodyweight = m · g) and must be included in all kinematic and kinetic calculations. Because there remains to be consistency in the literature in regards to the reporting of such variables for whole-body movements (e.g. squats), the discussion in this review will be focused on the bench-press movement (limited bodyweight involved), unless otherwise stated. 1. Constant Resistance Dynamic training with constant resistance, or isoinertial training,[6,7] is characterized by exercises in which the entire resistive force is Sports Med 2010; 40 (4)

Frost et al.

306

dependent on the mass of the object being lifted (force = mass · [gravity + acceleration]), and is currently the most widely used method to enhance strength and power capacity for sport. The use of constant resistance, or free weight, allows the strength coach/practitioner to prescribe sport-specific, multiple-joint efforts that involve both concentric and eccentric muscular actions, at various velocities. However, to increase the movement velocity, and thus the acceleration, a proportional increase in the net force is required when using the same load (equation 1). As a result, an athlete’s ability to move quickly might be limited by his/her rate of force development, potentially resulting in the development of force-specific and not velocity-specific power. Furthermore, when using constant resistance, athletes may not be able to elicit maximal activation of the primary musculature throughout the entire range of motion[8] as a result of mechanical advantages or disadvantages at specific joint angles. This can result in significant reductions in movement velocity and power output during the early and/or later stages of the concentric phase.[9] Researchers have investigated the kinematics and kinetics of various non-ballistic (concentric-only, stretch-shortening cycle [SSC]) and ballistic constant resistance efforts with the intent of identifying possible mechanical benefits that may lead to improved transference to sport performance. A concentric-only, nonballistic effort differs from SSC and ballistic movements in that the concentric phase begins without any pre-load from the eccentric phase and it ends without projecting the load into free space, respectively. Given that each of these efforts has different mechanical characteristics, each should be investigated separately to determine their role as a stimulus for strength and power adaptation. 1.1 Concentric Only

The initiation of the concentric (ascending) phase during constant resistance movements requires that the applied force be greater than the weight of the load, thus resulting in a net force ª 2010 Adis Data Information BV. All rights reserved.

and a proportional acceleration in the positive direction (equation 2). 1.1.1 Kinematics and Kinetics Velocity

Mean concentric velocity is a function of the displacement and duration of the concentric action. Therefore, the mean movement velocity should increase proportionally with a decrease in movement duration, assuming that the displacement (dictated in the bench press by limb length) is unchanged. Cronin et al.[2] have shown that an increase in the mean velocity is associated with a decrease in both the duration of the concentric phase and the magnitude of the load lifted, although such changes are not proportional (table I). Using the mean velocities and the durations of the concentric phase reported by Cronin et al.[2] for loads of 30–80% of the one repetition maximum (1RM), the calculated displacement was found to vary by as much as 0.169 metres. Because of the explosive nature of the efforts performed, such a difference in the displacement is likely due to greater protraction of the scapula with lighter loads. Similar to the mean concentric velocity, greater peak concentric velocities are produced with lighter loads. However, Cronin et al.[2] reported that a 100% increase in mean velocity necessitated a 37.5% reduction in the load, whereas a 50% reduction in load was required to attain an equivalent increase in the peak velocity (tables I and II). Conversely, the time to peak velocity, a key variable of interest among researchers,[9,12,13] was greater for heavier loads, even when expressed as a percentage of the total duration of the concentric phase (table I). Its position marks the point at which acceleration is equal to zero (equation 3), and the applied force is equivalent to the load (equation 2) and, thus, the separation of the acceleration and deceleration phases of the movement. These results highlight one of the major disadvantages of using constant resistance for high-velocity training: lighter loads are required to increase the movement velocity (1.49 vs 0.68 m/sec for 30% and 80% 1RM loads, respectively) but this is associated with a decrease in the duration of Sports Med 2010; 40 (4)

Biomechanics of Resistance

307

Table I. A comparison of power and velocity across loads[2,10] Load

Power

% 1RM

MP

PP

% PP/MP

TPP

% TTP/DC

MV

Velocity PV

TPV

% TPV/DC

80

232.2

515.1

221.8

0.784

63.6

NA

NA

NA

NA

80

207.8

440.3

211.9

0.717

61.2

NA

NA

NA

NA

80

222.0

478.2

215.4

NA

NA

0.33

0.68

0.928

82.9

70

266.8

542.4

203.3

NA

NA

0.36

0.86

0.774

80.3

60

280.8

549.3

196.0

NA

NA

0.55

1.01

0.620

75.8

60

314.6

572.7

182.0

0.595

71.3

NA

NA

NA

NA

60

237.6

501.8

211.2

0.564

67.3

NA

NA

NA

NA

50

271.1

552.1

203.7

NA

NA

0.64

1.16

0.528

69.4

40

290.8

613.6

211.0

0.452

65.7

NA

NA

NA

NA

40

201.1

445.1

221.3

0.434

63.0

NA

NA

NA

NA

40

250.0

532.0

212.8

NA

NA

0.73

1.35

0.459

67.2

30

211.1

467.4

221.4

NA

NA

0.82

1.49

0.401

63.8

1RM = one repetition maximum; DC = duration of concentric phase (sec); MP = mean power (W); MV = mean velocity (m/sec); NA = not reported/could not calculate; PP = peak power (W); PV = peak velocity; TPP = time to PP (sec); TPV = time to PV (sec).

the acceleration phase during the concentric effort; 63.8% and 82.9% of the duration of the concentric phase for 30% and 80% 1RM loads, respectively.[2] Acceleration and Force

Distinct start- and endpoints of the concentric phase, at which the movement velocity is zero, result in a mean acceleration of zero for all nonballistic movements (equation 3). Furthermore, since force is directly proportional to the acceleration of the load, the mean net force will be equal to zero and the mean applied force will be equivalent to the weight of the load (equation 1), regardless of mass. This is illustrated in table II; mean forces are equal to the load lifted despite differences in strength or percentage of the 1RM being used. The magnitude of the peak force produced is directly proportional to the peak acceleration and the mass (inertia) of the object (equation 1); thus, to achieve greater concentric accelerations, peak force must be equal to a higher percentage of the object weight (table II). However, the peak force required with a specific load to elicit an acceleration or movement velocity similar to athletic performance may not be achievable. Therefore, a lighter load must be lifted to produce the desired acceleration (table II), possibly reduª 2010 Adis Data Information BV. All rights reserved.

cing the training specificity; with constant resistance, concentric-only efforts, extremely light loads may not provide an appropriate training stimulus even if the acceleration is high at the onset of the movement; peak force (acceleration) has been shown to occur at the onset of the concentric action,[9,11] which in turn leads to more variation in the force curve, greater changes in momentum[2,12,14] and, thus, may be one of the reasons for an extended deceleration phase. The displacement-limited nature of these efforts requires that the movement velocity be zero at the end of the concentric phase; thus, larger accelerations will need to be matched by larger decelerations or an extended deceleration phase. Therefore, concentric-only, non-ballistic movements may be an appropriate stimulus to improve an athlete’s rate of force development; however, caution must be exercised since this will occur in concert with a reduction in velocity towards the end range of movement.[13] Power

Power output, as defined in equation 9, is a function of the applied force and movement velocity of an object. Researchers and practitioners alike have regarded improvements in power output as being important in improving athletic performance.[15-18] Since mean acceleration Sports Med 2010; 40 (4)

Frost et al.

308

Table II. A comparison of force and acceleration across loads[2,3,9,11] Load % 1RM

W

M

Force MF

PF

% PF/W

% PPF/TCD

Acceleration MA PA

100

1668

170.0

NA

NA

NA

NA

NA

2.08

2.27

100

2349

239.5

NA

2621

111.6

2.46

NA

1.91

1.14

100

1814

184.9

NA

2004

110.5

2.60

NA

2.21

1.00

100

995

101.4

NA

1349

135.6

2.49

NA

3.54

1.76

90

1370

139.7

NA

2003

146.2

NA

NA

NA

1.30

81

1381

140.8

NA

NA

112.1

4.38

NA

NA

NA

80

677

69.0

678

826

122.0

NA

0.00

1.93

1.12

75

1138

116.0

NA

2097

184.3

NA

NA

NA

0.86

70

593

60.4

589

766

129.3

NA

0.06

2.43

0.96

60

508

51.8

509

661

130.0

NA

0.03

2.96

0.82

50

423

43.2

429

584

138.0

NA

0.14

3.78

0.76

40

339

34.5

340

506

149.4

NA

-0.03

4.63

0.68

30

254

25.9

258

414

163.0

NA

0.15

5.87

0.63

DC

1RM = one repetition maximum; DC = duration of concentric phase (sec); M = mass of load (kg); MA = mean acceleration (m/sec2); MF = mean force (N); NA = not reported/could not calculate; PA = peak acceleration (m/sec2); PF = peak force (N); PPF = position of peak force (m); TCD = total concentric displacement (m); W = weight of free weight load (N).

ª 2010 Adis Data Information BV. All rights reserved.

These values were subsequently used to determine the respective contributions of force and velocity for each power output, expressed as a percentage of the maximum (table III). Maximum mean power was produced with approximately one-half (45.4%) of the maximum force and one-quarter (25.0%) of the maximum velocity for a non-ballistic, concentric-only bench press (table III). Such an analysis might provide greater insight into the components that contribute to power output, which may in turn be used to monitor and programme for power development in a more systematic and sportspecific fashion (e.g. monitoring the effects of (a)

(b) b'

b' Force

a'

Force

and mean net force for non-ballistic movements are both zero, the mean concentric power becomes a product of the magnitude of the resistance (force) and the mean movement velocity over the concentric phase. However, identifying the degree to which each variable (force and velocity) contributes may prove to be of greater theoretical and practical significance. Most researchers continue to report an optimal load that disregards separate force and velocity contributions,[2,10,12,15-30] which could limit the transference to athletic performance. Although it has been stated that maximum power is developed at one-third of the maximum speed and one-quarter of the maximum force,[24,31] figure 1 depicts the potential effect of increasing either the maximal strength or peak velocity on the power output of muscle; both require an alteration in the force capability, but are the product of different training stimuli. Cronin et al.[2] reported minimum and maximum mean power outputs for the concentric phase of the bench press of 211 W (30% 1RM) and 281 W (60% 1RM), respectively (table II). Extrapolating their force and velocity data from tables I and II (figure 2) using a second-order polynomial, maximum force and velocity were estimated at 1120 N and 2.20 m/sec, respectively.

a'

b a

a Velocity

b

Velocity

Fig. 1. The effect of increasing: (a) maximal strength; and (b) contraction velocity of muscle on power output (dashed line). a = before; b = after.

Sports Med 2010; 40 (4)

Biomechanics of Resistance

309

1400

Mean Peak Peak trendline Mean trendline

1200 Force (N)

1000 800 600 400 200 0 −0.25 −200

0.25

0.75

1.25

1.75

2.25

Velocity (m/sec)

Fig. 2. Extrapolated force and velocity curves (reproduced from Cronin et al.,[2] with permission from Lippincott Williams & Wilkins).

different training methods on either the force or velocity capabilities of muscle, with the aim of determining whether an athlete is force or velocity deficient). In this example, maximum mean power during constant resistance exercise appeared to have a greater dependence on the applied force than the movement velocity (table III). In such cases, it may be advantageous to attempt to increase the movement velocity contribution to power output and, thus, the transference to sport performance. However, as previously outlined, loading with constant resistance necessitates a proportional increase in force to attain this velocity, which in turn limits the possibility of achieving such training goals. Peak power must be identified via an analysis of the force-velocity curve for each of the loads, which makes a similar extrapolation of the data impossible. Table I shows similar peak power outputs for all loads between 30% and 80% 1RM, despite reports stating that a load of 40–60% 1RM is optimal for producing maximum power.[2,10]

Furthermore, as reported by Cronin et al.,[10] time to peak power is reduced with a subsequent decrease in the load, although there is little difference when it is expressed as a percentage of the total concentric duration (table I). However, as can be seen in table I, the point of the concentric phase at which peak velocity occurs is reduced with a decrease in load. Therefore, the contributions from force and velocity are changing despite similar power outputs. If velocity-specific power is of greater importance to improving sport specificity, then future research should separate the components of power output to better understand the adaptations to training with various loads. Concentric Phases

Lander et al.[3] has proposed that concentric action of a constant resistance, non-ballistic movement actually involves two acceleration and deceleration phases when using loads between 75% and 100% of the 1RM. The first and true acceleration phase is defined as the initial portion of the ascending phase that describes the period of maximum acceleration, whereas the late stages described by the period of maximum deceleration is defined as the second and true deceleration phase. The middle portion of the ascent is composed of the first deceleration phase, or the sticking region, and is defined as the portion of the effort where the applied force falls below the weight of the load. The second acceleration phase, or the maximum strength region, is defined as the period where the applied force becomes greater than the load for the second time (figure 3).

Table III. Force and velocity contributions to concentric-only and stretch-shortening cycle (SSC) mean power [reproduced from Cronin et al.,[2] with permission from Lippincott Williams & Wilkins] Load

Concentric only

% 1RM

MP (W)

% MF/Fmax

% MV/Vmax

MP (W)

SSC % MF/Fmax

% MV/Vmax

80

222.0

60.5

15.0

261.3

60.6

17.7

70

266.8

52.6

16.4

316.3

53.5

24.1

60

281.8

45.4

25.0

315.4

45.8

28.2

50

271.1

38.3

29.1

312.8

38.1

34.1

40

250.0

30.3

33.2

283.1

30.4

37.7

30

211.1

23.0

37.3

237.0

23.4

41.4

1RM = one repetition maximum; Fmax = maximum force (N); MF = mean force (N); MP = mean power (W); MV = mean velocity (m/sec); Vmax = maximum velocity (m/sec).

ª 2010 Adis Data Information BV. All rights reserved.

Sports Med 2010; 40 (4)

Frost et al.

310

1500

AP

SR

MSR

DP

1450 100% 1RM

1400 Force (N)

Load weight 1350 1300 1250

81% 1RM

1200 AP

1150

DP

1100 0

0.5

1.0 Time (sec)

1.5

2.0

Fig. 3. Force-time curve for 81% and 100% one repetition maximum (1RM) loads (reproduced from Elliott et al.,[9] from Lippincott, Williams & Wilkins). AP =acceleration phase; DP = deceleration phase; MSR = maximum strength region; SR = sticking region.

Using athletes with similar bench-press experience, Elliott et al.[9] found that a 100% 1RM load produced the same four phases as reported by Lander et al.;[3] however, an 81% load was described as having one acceleration and deceleration phase (figure 3). The researchers suggested that the differences between the two studies could have resulted from the sequencing of the lifts: the heavier load was lifted first in the study by Lander et al.,[3] resulting in muscular fatigue and, thus, a reduced ability to accelerate the load through the sticking region. Although Elliott et al.[9] reported that only 48.3% of the concentric phase was spent accelerating with a load of 81% of the 1RM, combining the duration of the acceleration phase and maximum strength region and the sticking region and deceleration phase for the 100% 1RM condition resulted in similar acceleration and deceleration durations (47.9% and 52.1%, respectively). Expressed as the point of the concentric phase at which peak velocity occurred, Cronin et al.[2] reported acceleration phases lasting between 63.8% and 82.9% of the duration of the concentric effort for loads of 30–80% 1RM, respectively (table II). However, force and velocity profiles were not described, making it difficult to determine whether the curves were similar to the 81% or 100% 1RM load as shown in figure 3. Changing the load lifted or the way a repetition is performed may alter the point at which peak velocity occurs and, thus, ª 2010 Adis Data Information BV. All rights reserved.

the duration of the acceleration phase; however, in the absence of the force and velocity data for the entire concentric movement, conclusions may be misrepresentative. Regardless of the attempts made to increase the length of the acceleration phase during a nonballistic movement, mean force will always equal the weight of the load. This implies that if a higher peak force is achieved early in the concentric phase, a subsequent decrease in force will be seen during the later stages of the movement or, conversely, if the time to peak velocity is delayed and the subsequent acceleration phase is longer, there will be a greater decline in force and a drastic deceleration at the conclusion of the concentric phase. This highlights another disadvantage of using non-ballistic, constant resistance movements; the peak forces required to elicit high-movement velocities will cause greater variation in the force-displacement curves, via accentuated accelerations and decelerations, which will lead to changes in the muscle activity and might increase the risk of injury.[13] 1.1.2 Muscle Activity

Elliott et al.[9] has reported large sustained increases of the agonist musculature during the acceleration phase of a concentric effort. However, once the load begins to decelerate, muscle activity subsides and subsequent reductions in agonist activity are seen. Using 60% and 80% Sports Med 2010; 40 (4)

Biomechanics of Resistance

1RM loads, McCaw and Friday[5] found that peak triceps brachii activity occurred at 70% and 60% of the total concentric duration, peak anterior deltoid activity was produced at approximately 60% and 75%, although this dropped dramatically after an additional 10% duration, and peak pectoralis major activity occurred at 17% and 10%, although this was halved after completing only 50% of the movement. An inability to maintain a greater percentage of the maximum activity throughout the entire range of motion may inhibit an athlete’s ability to maximize any number of kinetic and kinematic variables (e.g. power, force and velocity); however, such a contention requires further investigation. Researchers have reported antagonist activity;[5,9,32,33] however, additional research is required to investigate the electromyographic profiles, particularly during high-velocity movements. A stronger antagonist will allow for a shorter deceleration phase, thus enabling the agonists to develop higher movement velocities,[34] although, as stated above, this may increase the risk of injury.[13] 1.2 Stretch-Shortening Cycle

Muscle actions in which the concentric action is immediately preceded by an eccentric action (pre-stretch) are referred to as SSC movements.[35] Use of the SSC alters the profiles of several kinetic, kinematic and electromyographic variables because of the inertial properties of mass and the physiological characteristics of the muscle. Researchers have shown that a prestretch can increase mean velocity and power output by as much as 16% and 6%, respectively,[2,21,36-38] although the mechanisms responsible for the augmentation to the concentric action remain unclear. Potential mechanisms include the utilization of elastic strain energy stored in the series elastic component, reflexively induced neural input and a higher active muscle state prior to the onset of the concentric muscle action; however, a detailed explanation of each is beyond the scope of this review and the authors refer those interested to several papers on this topic.[10,12,35,36,38-43] ª 2010 Adis Data Information BV. All rights reserved.

311

1.2.1 Kinematics and Kinetics Velocity

Similar to concentric-only non-ballistic movements, the mean velocity of non-ballistic SSC movements is equal to the change in displacement divided by the duration of the displacement. However, it has been consistently shown that the use of a pre-stretch can augment the velocity curve by decreasing the duration of the concentric phase.[2,36-38] Cronin et al.[2] reported that mean velocity could be increased by as much as 47% with a 70% 1RM load and by an average of 20% across loads of 30–80% 1RM (table IV). Similarly, using ballistic efforts, Newton et al.[12] presented an average increase of 30%, with the largest increase being produced with the greatest load tested (90% 1RM). Conversely, the use of an SSC has been unable to elicit any change in the peak velocity or the time to peak velocity, expressed as a percentage of the total concentric duration, at any load tested.[2,12] As can be observed in equation 4, the change in movement velocity over a specific duration is proportional to the net force being applied and inversely proportional to the mass of the load being lifted. Therefore, an increase in mean velocity requires that a greater force be generated over the same time or that the same force be applied over a longer time. However, changing the peak velocity necessitates the production of greater accelerations over the same period of time, or the same acceleration over a longer period of time. If it is not possible to extend the duration, as is the case during movements with a fixed range of motion (e.g. limb length), larger accelerations and, thus, larger forces are needed. This may be difficult to accomplish with constant resistance. Acceleration and Force

For non-ballistic SSC movements, mean acceleration and, thus, mean force are zero since there is no change in the velocity between the start and end of the muscular effort. Conversely, peak acceleration and peak force have been shown to increase by as much as 69.5% and Sports Med 2010; 40 (4)

Frost et al.

312

Table IV. A comparison of the velocity and power between a concentric-only and stretch-shortening cycle (SSC) bench press [reproduced from Cronin et al.,[2] with permission from Lippincott Williams & Wilkins] Load

Concentric only

% 1RM

MV

PV

% TPV/DC

MP

PP

MV

SSC PV

% TPV/DC

MP

PP

80

0.33

0.68

82.9

222.0

478.2

0.39

0.65

75.4

261.3

468.9

70

0.36

0.86

80.3

266.8

542.4

0.53

0.83

76.2

316.3

542.3

60

0.55

1.01

75.8

280.8

549.3

0.62

0.99

73.8

315.4

550.6

50

0.64

1.16

69.4

271.1

552.1

0.75

1.21

68.7

312.8

557.9

40

0.73

1.35

67.2

250.0

532.0

0.83

1.37

66.2

283.1

536.3

30

0.82

1.49

63.8

211.1

467.4

0.91

1.52

56.7

237.0

463.2

1RM = one repetition maximum; DC = duration of concentric phase (sec); MP = mean power (W); MV = mean velocity (m/sec); PP = peak power (W); PV = peak velocity (m/sec); TPV = time to PV (sec).

19.0% at loads of 70% and 60% 1RM, respectively (table V). An SSC movement requires sufficient force to reduce the eccentric velocity of the load to zero prior to beginning the concentric action (figure 4). As can be observed in figure 4, the change in momentum is directly proportional to the change in velocity and the mass of the load. This change in momentum is also proportional to the force that is causing such a change, and the duration over which the change takes place (see equation 5). Compared with a non-ballistic, concentriconly movement that requires no reduction in eccentric velocity, the sum of the external forces (equation 2) may be greater, thus changing the external demands placed on the athlete. Consequently, higher accelerations may be produced during the initial portion of the concentric action, although this is not always the case. Without a subsequent increase in the peak velocity or duration of the acceleration phase, it is difficult to speculate as to whether the SSC action alone is advantageous to athletic perfor-

mance when the movement duration exceeds ~200 ms.[10,12] Perhaps, the benefits of such a movement stem from an ability to utilize the increased force quickly, via muscle tendon interactions in terms of elastic and contractile tension, thus allowing for a greater peak velocity to be achieved. In any case, further research is required to investigate the effects of an increased peak force and, thus, variation in the force-displacement curve, on the kinetic, kinematic and electromyographic profiles during the later stages of the movement. Power

Non-ballistic, SSC movements are characterized by mean forces that are identical to the magnitude of the resistance being lifted (equation 1) and mean velocities that are significantly greater than their concentric-only equivalents.[2] Since power is the product of force and velocity, mean power should be increased to the same degree as the mean velocity. Cronin et al.[2] reported an average increase in mean power of 14.8%

Table V. A comparison of the acceleration and force between a concentric-only and stretch-shortening cycle (SSC) bench press[2,44] Load

Concentric only

% 1RM

MA

SSC PA

MF

PF

MA

PA

MF

PF

80

0.00

1.93

677.8

826.3

0.02

3.19

679.0

906.5

70

-0.06

2.43

588.8

766.4

0.11

4.12

599.1

876.6

60

0.03

2.96

509.4

660.9

0.11

4.97

513.4

786.4

50

0.14

3.78

429.2

583.8

0.09

6.33

427.2

692.7

40

0.02

4.63

339.5

505.9

0.05

7.11

340.5

596.1

30

0.15

5.87

257.9

414.1

0.31

8.54

262.1

491.8

1RM = one repetition maximum; MA = mean acceleration (m/sec2); MF = mean force (N); PA = peak acceleration (m/sec2); PF = peak force (N).

ª 2010 Adis Data Information BV. All rights reserved.

Sports Med 2010; 40 (4)

Biomechanics of Resistance

t1 = t s v1 = v m/sec

313

CO Fs = ma

SSC Fs = ma

t1 = t s v1 = v m/sec m

a

a

vtrans = −vecc t2 = ttrans

m

m

t0 = 0 s v0 = 0 m/sec a0 = 0 m/sec2

t0 = 0 s v0 = 0 m/sec a0 = a m/sec2 Fg = mg Ftotal = mg + ma

Fg = mg Change in momentum = m(v0 − vtrans) = Fsttrans a0 = Fs/m m/sec2 Ftotal = mg + ma0

Fig. 4. Force contributions during concentric-only (CO) and stretch-shortening cycle (SSC) movements. The total force (Ftotal) at time zero (t0), which represents the transition between the eccentric and concentric phase, is the sum of the system weight (Fg = mg) and the force due to acceleration (Fs = ma). The total force during SSC movements will be greater than the CO equivalent if the acceleration (a0) at this timepoint is greater than zero. The velocity at a specific instant in time (ttrans), prior to the completion of the eccentric phase, is represented by vtrans.

(velocity was 20%) across loads of 30–80% 1RM, compared with a concentric-only movement. The increase in mean power production for a SSC movement is a function of the greater relative contribution from velocity (table III). However, upon further examination of the power profiles for SSC actions, significant increases are limited to the initial 200 ms of a movement,[10,12,42,43] and any benefit of the pre-stretch has been reported lost if the concentric duration exceeds 0.37 seconds.[10,41,42] Cronin et al.[2,10] have provided further support for these claims by finding no difference in the peak power or time to peak power, expressed as a percentage of the concentric duration, between SSC and concentric-only conditions. However, in the absence of any acute or long-term training studies investigating the force and velocity contributions to power output, any conclusions pertaining to the effect of an SSC action on power development, or sport-specific applications may be misleading and/ or misrepresentative. 1.2.2 Muscle Activity

Although the electromyographic (EMG) profiles of the agonist musculature have been shown to parallel the force profiles for SSC actions,[13] there remains to be sufficient evidence to make ª 2010 Adis Data Information BV. All rights reserved.

definitive claims regarding this relationship. If the mechanism responsible for the augmentation to the concentric action stems from an ability to utilize stored elastic energy, then there should not be any change in the muscle activity at the beginning of the concentric phase. However, if the SSC enhancement is the result of a higher active muscle state, then increases in agonist activity could be expected during the initial duration of the concentric phase. Newton et al.[12] found that pectoralis major and triceps brachii activity were significantly higher during the first 50 ms of the concentric phase for an SSC movement, although no differences were seen when expressed as mean or peak values of the entire repetition. Consequently, additional research is required to verify the underlying mechanism(s), identify the role of the agonist and antagonist musculature throughout an SSC action and improve the transference of these training principles to sportspecific applications. 1.3 Ballistic

In the strength-training literature, ballistic denotes accelerative, of high velocity and with projection into free space.[45] Ballistic efforts have been introduced as a potentially superior form of Sports Med 2010; 40 (4)

Frost et al.

314

constant resistance training because the athlete is no longer limited by having to decelerate the load at the end of the concentric phase,[12] consequently resulting in higher mean velocity, force and muscle activity. This type of training may also provide a stimulus more representative of athletic movements in which objects are thrown or the body is propelled in a certain direction as quickly as possible.[46] 1.3.1 Kinematics and Kinetics Velocity

The mean velocity produced with a ballistic effort is equal to the change in displacement between the onset of the concentric action and the point at which the load is projected, divided by the duration of this phase. Furthermore, it should exceed the mean velocity of a non-ballistic movement performed with the same load. The point at which the load loses contact with the athlete, he/she is no longer able to apply any muscular force and, therefore, any subsequent change in velocity is a result of the force of gravity and should not be included in any kinetic or kinematic calculations. Comparing ballistic and non-ballistic techniques at a load of 45% 1RM, Newton et al.[13] found that projecting the load at the end of the concentric phase increased the mean velocity by 27.3%. Conversely, Cronin et al.[2] reported similar mean velocities between ballistic and non-ballistic efforts for loads ranging from 30% to 80% 1RM; however, the velocity data were calculated by differentiating the entire displacement curve, including the duration of the movement during which the load had been released (the barbell was no longer in contact with the hands but it was still travelling in an upwards direction), making it a non-ballistic analysis. Ballistic movements have been regarded as a superior form of power training because they allow for a greater portion of the concentric phase to be spent accelerating; 96% compared with 60% of total displacement with a 45% 1RM load,[13] which resulted in a 36.5% increase in peak velocity.[13] However, additional research is required to fully understand the transference of such a stimulus to sport performance. ª 2010 Adis Data Information BV. All rights reserved.

Acceleration and Force

The kinetics and kinematics of ballistic efforts are not constrained by a predetermined displacement or a final concentric velocity of zero. Consequently, mean acceleration and mean force will both be greater than their non-ballistic equivalent, contrary to some reports.[2] Comparing a ballistic and non-ballistic bench press, Newton et al.[13] found that projecting the load at the end of the concentric phase could increase the mean force by 35%. This finding, in combination with the increases to mean and peak velocity, led the researchers to conclude that ballistic training provided superior loading conditions for the neuromuscular system and a more sport-specific training stimulus. However, projecting the load failed to increase the peak acceleration, peak force or the time to peak force, even when utilizing an intermediate load of 45% 1RM. This implies that the maximum rate of force development or the ability to produce maximal force in minimal time,[24] and the force achieved early in the concentric phase (e.g. 30 ms) or the ability to increase force as rapidly as possible once the effort has begun,[24] was not changed. Participants in the study simply extended the duration of the acceleration phase, thereby increasing the peak velocity, without changing the ‘explosiveness’ of the movement. Extending the duration of the acceleration phase simply requires that an athlete provide a force greater than the weight of the load over a longer period of time. Increasing the explosiveness of a movement demands that a greater percentage of the absolute maximum force is produced during the initial portion of the concentric effort. With regards to athletic ability, it is important that the practitioner and researcher understand the demands of the sport and the training status of the athlete in question so that they are able to provide the stimulus most conducive to improving performance. Power

The ability to increase the mean velocity and mean force via ballistic efforts will result in greater mean power (equation 9), provided that the kinetic and kinematic analysis is completed from the point at which the load is projected and Sports Med 2010; 40 (4)

Biomechanics of Resistance

does not include data from thereafter. Using a 45% 1RM load, Newton et al.[13] found that mean power could be increased by 70% in comparison to a non-ballistic equivalent, whereas Cronin et al.[2] failed to report any significant differences for loads of 30–80% 1RM. However, the data from Cronin et al.[2] were analysed from the onset of the concentric phase to the point at which the load achieved a velocity of zero, making it a non-ballistic analysis and thereby masking any differences that may have been produced. Compared with a non-ballistic movement, similar increases in peak power output have been reported,[2] whether performed with or without an SSC action (13.3% and 9.2% for SSC and concentric-only, respectively). Newton et al.[13] reported an increase of 67%, although it was not made clear whether the enhancement was due to a greater contribution from force or velocity. Ballistic training may provide a means of increasing the specificity of training; however, there may be disadvantages associated with a protocol that simply emphasizes the projection of the load (e.g. reduced rate of force development). An in-depth analysis of the force and velocity profiles of ballistic movement is required to fully understand their potential, or limitations, and identify possible means to increase the degree of specificity. 1.3.2 Muscle Activity

Ballistic movements have been described as having a particular neural pattern of motor unit firing known as the tri-phasic or ‘ABC’ pattern.[47,48] This pattern is characterized by a large burst of activity by the agonist musculature followed by a brief period of concomitant inactivity of the agonist and activity of the antagonist musculature, concluding finally with a second burst of agonist activity.[47-49] Behm[48] contended that while both the initial agonist and antagonist muscle activity are pre-programmed and immutable, the agonist has the responsibility of driving the movement while the antagonist protects the integrity of the joint and provides greater accuracy. Behm[48] also suggested that the pre-movement silent period of the agonist may help to synchronize the motor units for the subsequent second burst of activity by the agonist ª 2010 Adis Data Information BV. All rights reserved.

315

musculature, which is largely controlled by sensory feedback. Because the net force produced during a movement is a trade off between the force of the agonist muscles and the counteracting force of the antagonist muscles, the interaction between these bursts of myoelectrical activity warrants further research. Although more prevalent in trained individuals,[48] the interfering effect of the co-contraction between agonist and antagonist muscles in ballistic movements is reduced with strength training. Therefore, a more efficient control of the ABC pattern may benefit the power athlete. Newton et al.[13] compared the electromyographic profiles between a non-ballistic and ballistic SSC movement, but did not address the notion of an ABC pattern. The researchers reported increases in the peak and mean EMG for each muscle investigated; pectoralis major (70–84% 1RM and 69–70% 1RM), deltoid (70–93% 1RM and 56–68% 1RM), triceps brachii (59–85% 1RM and 57–68%) and biceps brachii (65–83% 1RM and 61–67% 1RM). The most significant differences were seen during the later portion of the concentric phase, which is reflected by the changes in the force-displacement curve. Although increasing an individual’s muscle activity during the concentric phase of movement may elicit performance improvements, the complexity of intermuscular and intramuscular coordination make a direct correlation highly unlikely. 2. Accommodating Resistance Accommodating resistance training allows for the development of maximal tension throughout the complete range of motion rather than at a particular point.[50] The use of accommodating resistance allows athletes to develop maximal force at various velocities, while being unaffected by the inertial properties of the load. That is, if velocity remains constant for the entire muscular effort, acceleration will be zero, meaning that the force produced is not dependent on the mass of the load. This reportedly results in increased force and motor unit recruitment at training Sports Med 2010; 40 (4)

Frost et al.

316

velocities similar to those achieved with a constant resistance equivalent.[4] However, many researchers question the validity and efficiency of these devices because they are biomechanically different from natural movements, and the degrees of freedom, or permissible movement directions, are reduced from six to one.[50] Currently, there are two types of devices used to accomplish these goals: hydraulics and isokinetics. The kinetics, kinematics and muscle activity associated with both these devices will form the basis of the discussion in this next section. 2.1 Hydraulic

Hydraulic devices do not offer a true type of accommodating resistance, in the sense that they do not provide a resistance equal to the force that is being applied.[51] Instead, hydraulic devices provide a resistance that is proportional to the applied force and accommodate for the movement velocity[50] by allowing the user to adjust the opening diameter of a valve that controls the speed at which hydraulic fluid flows through the system;[52,53] the greater the velocity, the greater the resistance.[50] 2.1.1 Kinematics and Kinetics

There are a limited number of researchers who have used hydraulic equipment in their studies[51,53-59] and even fewer who have reported the associated kinetics and kinematics.[53,55] Most studies using hydraulics have tested with isokinetic devices[54,57-59] and simply used these devices as a means of training. In one of the only studies reporting kinetics, Hortobagyi et al.[53] compared slow (0.037 m/sec) and fast (0.126 m/sec) hydraulic bench-press movements with a 1RM free-weight effort. For the slow and fast conditions, respectively, the mean force was found to be 8.3% greater and 62.6% lower, relative to the free-weight condition. However, the researchers failed to report either peak forces or accelerations, which would have allowed for better comparisons with free weights and an improved insight into the sportspecific training effects of hydraulic resistance. ª 2010 Adis Data Information BV. All rights reserved.

Burke et al.[55] found that forces equivalent to 171% and 231% of the mean force could be produced with slow (2.5-second) and fast (0.8-second) concentric efforts, respectively. Comparing these results with similar free-weight efforts is difficult because the total displacement was not reported and the subjects in the study by Burke et al.[55] performed the concentric action from a seated position. Cronin et al.[2] found that for a similar concentric duration (0.76 seconds), the peak force produced was only equivalent to 130% of the mean value, suggesting that there are much greater variations in the magnitude of the force-displacement curves for hydraulic resistance. Telle and Gorman[56] have advocated the combination of free weights and hydraulics as a superior strength-training mode, rationalizing that the inherent disadvantages associated with either type of resistance will compensate for the other. Hydraulic devices provide a resistance that is negligible at the onset of the concentric phase, gradually increases over the duration of the effort and is reduced to zero for the eccentric phase. Despite these claims, researchers and strength coaches alike do not regard hydraulic resistance as a superior stimulus to free weights as it lacks the training specificity thought to be necessary to elicit improvements in athletic performance. 2.1.2 Muscle Activity

At the time of writing this review, the authors were unable to locate any research that had investigated the acute or long-term effect of hydraulic-resistance exercise on muscle activity. Since hydraulics are a form of accommodating resistance, it could be hypothesized that maximum agonist activity is produced; however, this is assuming maximum voluntary efforts, which may not always be the case. 2.2 Isokinetic

The concept of isokinetic exercise was introduced in 1967 by Hislop and Perrine[4] as an alternative to isotonic (constant resistance) and isometric exercise. It was proposed that by keeping a constant velocity, and thus zero acceleration, Sports Med 2010; 40 (4)

Biomechanics of Resistance

they could provide a more suitable mechanical means of obtaining a maximal concentric effort throughout a range of motion compared with isotonic training, while still allowing for work to be done (work = force · displacement), in contrast to isometric training. Although some degree of generality may exist between single-joint isokinetic efforts and athletic movement, it is thought that being limited to machine use, the lack of mechanical feedback and the absence of an eccentric component on some of the earlier systems do not provide a training stimulus conducive to high degrees of sport specificity; therefore, some researchers have began to challenge the effectiveness of isokinetic dynamometry.[8,50,60,61] However, isokinetics do offer a unique method of altering the kinetics and kinematics associated with resistance training and have assisted in improving our understanding of velocity specificity;[62-64] thus, they are still regarded as an effective means of high-velocity training. 2.2.1 Kinematics and Kinetics

The majority of studies that have used isokinetic devices as a means of training or testing have attempted to identify a velocity-specific response or an increase in torque over a specific training period.[41,62-68] Few researchers have actually attempted a comparison of the kinetics between isokinetics and other resistance-training modes.[3]

317

centric phase,[3,65,69,70] and can limit the constant velocity duration to 16% of the exercise if the movement velocity is high enough.[65] Although an increase in velocity is associated with a subsequent reduction in force, the magnitude of such a reduction is much less than for constant resistance efforts. A 100% increase in the mean angular velocity of a knee extension only elicited a 10.5% reduction in force, or torque[62] compared with a 37.5% reduction during constant resistance exercise (37.5%).[2] In order to reduce the torque by 37%, researchers have had to increase the angular velocity by 400% (38%),[65] 500% (42%)[62] and 1000% (37%).[66] Furthermore, the group of subjects who required a 1000% increase in velocity were able to attenuate their reduction in force to just 30% after becoming familiarized with the isokinetic device.[66] Acceleration and Force

The load acting in an isokinetic exercise is the result of the mechanical process of energy absorption.[4] Compared with constant resistance exercise in which part of the energy is dissipated with accelerations, isokinetic acceleration is, theoretically, always zero, allowing for complete energy absorption and a resisting force that is proportional to the magnitude of the input.[4] As such, the mean force for an isokinetic movement should be greater than its free-weight equivalent, at a comparable movement velocity. Comparing the force-time profiles of equivalent isokinetic

Velocity

ª 2010 Adis Data Information BV. All rights reserved.

60 Angular velocity (°)

Isokinetic devices are characterized by their ability to control the movement velocity and keep the acceleration at zero. However, all movement must start from a position of rest, or zero velocity, and terminate in a similar fashion, implying that exercises completed within a specific range of motion will involve some form of an acceleration and deceleration phase.[69,70] Chen et al.[69] has stated that the duration of an isokinetic movement spent at constant velocity and the accuracy of measurements are limited by acceleration, oscillation and deceleration phases of the exercising limb (figure 5). As the movement velocity is increased, the three non-constant velocity phases represent a greater relative portion of the con-

AP

OP

CV

DP

50 40 30 20 10 0 0

0.5

1.0

1.5 2.0 Time (sec)

2.5

3.0

Fig. 5. An example isokinetic angular velocity-time curve for a movement performed at 30/sec (reproduced from Chen et al.,[69] with permission from the Orthopaedic and Sports Physical Therapy Sections of the American Physical Therapy Association). AP = acceleration phase; CV = constant velocity; DP = deceleration phase; OP = oscillation phase.

Sports Med 2010; 40 (4)

Frost et al.

318

2000

Isokinetic

Force (N)

1600

Load

1200

Free weight

800 SR 400

AP

MSR OP

DP

0 0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Time (sec) Fig. 6. Slow isokinetic (0.45–0.84 rad/sec) and heavy free-weight (90% one repetition maximum) bench press force-time profiles (reproduced from Lander et al.,[3] with permission from Lippincott Williams & Wilkins). AP = acceleration phase; DP = deceleration phase; MSR = maximum strength region; OP = oscillation phase; SR = sticking region.

and free-weight resistance exercises, Lander et al.[3] reported 5% and 6% greater mean forces during the isokinetic condition when testing at velocities equivalent to a 90% (0.45–0.84 rad/sec) and 75% (0.79–1.61 rad/sec) 1RM lift, respectively. Furthermore, isokinetic mean forces during the sticking and maximum-strength regions (oscillation phase) were equivalent to 130.2% and 116.2% of the 90% 1RM free-weight movement and 133.3% and 126.6% of the 75% 1RM movement (figure 6). Although there were no significant differences between the peak forces seen during the acceleration phase, peak forces were consistently lower for the isokinetic condition. Peak force is directly proportional to the peak acceleration achieved in the initial stage of the concentric phase. Because isokinetics seek to limit the duration of the acceleration phase, the maximum force that an athlete is able to produce may be compromised using this form of dynamometry; however, there is no evidence to substantiate this contention. Power

Referring to equation 9, the power output of a muscle is the product of its velocity of shortening and force production.[71] Kanehisa and Miyashita[68] and Caiozzo et al.[62] both reported maximum mean powers at an angular velocity of 4.19 rad/sec. However, it is difficult to assess this claim in terms of sport-specific power ª 2010 Adis Data Information BV. All rights reserved.

development because there is an important difference between the power produced during coordinated multiple-joint efforts and the power generated during a single joint movement,[71] as is traditionally the case with isokinetics. Furthermore, controlling the movement velocity will encourage the development of force-specific power development, resulting in an upward shift of the force-velocity curve (figure 1a) and potentially less transference to sport. 2.2.2 Muscle Activity

Isokinetics have been designed as a way to maximize the force and, thus, muscle activity at a specific movement velocity.[4] This type of dynamometry has received a great deal of advocacy as a means of testing pre- and post-measures of strength by allowing an individual to maximally contract throughout the range of motion.[64,68,72-74] However, upon further examination of the EMG-torque relationship during a single repetition of isokinetic knee flexion, Onishi et al.[75] found that all agonists do not achieve maximum levels of activity during maximal efforts, and that there are varying degrees of synergistic involvement depending on the knee-joint angle. Furthermore, dissimilar movement patterns, associated with isokinetic dynamometry, have relegated the importance of whole-body coordination and control and, thus, may not provide the most conducive means to improve athletic performance. 3. Variable Resistance From as early as 1900, attempts have been made to develop innovative training devices that could combat the changing mechanical advantage and inertial properties associated with free weight or constant resistance.[76] At constant velocity, the resistance offered to the moving body segment remains constant; however, the resistance applied to the muscle varies with a change in joint angle.[4,77] Conversely, the acceleration of a load (equation 1), or a change in its velocity, necessitates an increase in peak force,[2,12] variation of the applied force[12,14] and an increase in the duration of the deceleration Sports Med 2010; 40 (4)

Biomechanics of Resistance

phase.[3,9] As much as 67% of the concentric effort may be spent decelerating with a load of 19.4% 1RM.[14] Variable resistance devices alter the resistive force during a movement to match changes in the joint leverage,[50] and provide compensatory accelerations[76,78] with the intent of reducing the effect of these limitations. 3.1 Cam- or Lever-Based System

Arthur Jones and Harold Zinkin, founders of Nautilus and Universal, respectively, began their careers with a similar hypothesis: a muscle works at maximum capacity during a very small portion of a dynamic repetition. Therefore, to facilitate maximum muscular involvement, the resistance must be varied. Both inventors developed a cam- or lever-based system that would vary the resistance to coincide with the changing leverage of the body during movement.[8,79] Unfortunately, individual differences in limb length, point of muscle attachment, velocity of movement and maximum-force development affect the torque-angle curve of the machines, rendering them ineffective for their intents and purposes.[8,67,80] 3.1.1 Kinematics and Kinetics

The inherent flaws associated with cam-based variable resistance have limited the amount of available research on these devices. Those who have used them in their studies have done so with the intent of examining their effectiveness[67,80] or as a means of increasing strength,[57,58] and not for the purpose of a kinetic or kinematic comparison. Although found to be equally effective as hydraulics and surgical tubing to increase strength,[57,58] researchers have shown that there is little similarity between machine-resistive torque and human torque capabilities.[80] The torque-angle curve for the Universal bicep curl has been shown to be more representative of the maximum torque capabilities of the muscle than its free-weight counterpart;[79] however, these data were presented in a review on the device and not in a research study. Harman[80] showed that for the same exercise, machineª 2010 Adis Data Information BV. All rights reserved.

319

resistive torque varied up to 12% below maximum during the exercise movement, while human torque capabilities were as much as 59% below maximum at the extremes in the range of motion. Given the limitations of this type of equipment and the dissimilar resistance profiles to open- and closed-chain athletic movements, it should be deemed ineffective for the development of athletic performance and be reserved for recreational use. 3.1.2 Muscle Activity

To the authors’ knowledge, there has been no research to investigate the EMG profiles of a cam-based variable resistance effort. As stated in the previous section,[9] the inertial properties of a constant resistance load cause muscle activity to fluctuate throughout the repetition; therefore, it was speculated that cam-based devices would provide a resistance that resulted in near maximum muscular involvement throughout the range of motion. However, such a contention is theoretically based, and now receives little to no support from researchers and practitioners alike. 3.2 Band and Chain

Bands and chains are used as a means of overcoming the mechanical disadvantages associated with specific joint angles[76,78,81-83] and increasing the degree of sport specificity.[84] Reviews by Simmons[76,78] have resulted in an increased advocacy among powerlifters and strength coaches alike, despite claims that they are not an appropriate mode for resistance training as they work in opposition to the way human muscles contract and produce force.[84] Often reviewed simultaneously,[78,81] different inertial properties between the two training modalities necessitate separate discussions of the associated kinetics and kinematics. 3.2.1 Kinematics and Kinetics

The addition of bands and chains to a constant resistance movement will alter the kinematic and kinetic profiles of the respective lift; however, the manner in which changes and, thus, the training effect occur is governed by the inertial properties of the additional resistance. Sports Med 2010; 40 (4)

Frost et al.

320

The acceleration of any object is proportional to the force causing such an acceleration and inversely proportional to its mass or inertia (equation 1). Therefore, using bands, which have a minimal mass, will allow for greater accelerations to be produced compared with a constant resistance movement with an equivalent load. However, the load being lifted and, thus, the force required to elicit movement will increase proportionally with the displacement or stretch of the band (see equation 10): Fband ¼ k  l

(Eq: 10Þ

where k is the stretch coefficient of the band and l is the distance that the band isP being stretched (m). Using equation 2, Fnet ¼ Fext ¼ Fapp þ Fload ; Fload becomes equation 11: Fload ¼ Fband þ Fmass

(Eq: 11Þ

where Fmass is the gravitational force of the load (N). Therefore, it becomes increasingly difficult to maintain high-movement velocities and accelerations towards the end of the concentric phase. However, the benefit of such a modality stems from these characteristics. A load that increases proportionally with the concentric displacement necessitates sufficient acceleration in the early stages of the effort to overcome the elastic recoil of the bands and still complete the movement. Because the band resistance is primarily a function of its elastic properties and not the force of gravity, there is also opportunity for much greater eccentric accelerations to be achieved. Greater eccentric accelerations will result in proportional increases in eccentric forces and a greater involvement of the SSC when applicable. Behm[85] has suggested that the benefits of band resistance are most apparent when combined with free weights so that the inherent shortcomings of either resistance type compensate for one another. Used alone, the bands do not provide sufficient overloading to the initial portion of the movement, thereby potentially reducing any opportunity to increase the rate of force development. However, used in combination with free weights, the bands may benefit the end range of motion by controlling the associated momentum and providing needed additional resisª 2010 Adis Data Information BV. All rights reserved.

tance, which allows for the lift to be completed with greater effort later in the movement. Contrary to bands, chains are an additional system mass. The magnitude of this mass is proportional to the percentage of the concentric phase completed, assuming that the chains are evenly distributed on the floor prior to the initiation of the concentric effort. Equation 11 becomes the sum of Fmass and Fchain where Fchain at position z is (equation 12): Fchain ¼ mz  g (Eq: 12Þ where mz is the mass of the chain (kg) at position z (m) and g is gravity (m/sec2). However, mz can be stated as (see equation 13): mz ¼ ðmchain  zÞ=ðd  cÞ (Eq: 13Þ where mchain is the total mass of the chain (kg), d is the length of the concentric displacement (m) and c is the portion of the concentric displacement completed prior to the addition of any links in the chain (m). An increase in system mass also means that greater peak forces are required to achieve specific accelerations and movement velocities (equation 1 and 5). These forces, applied in ‘x’ amount of time, result in a change in momentum of the system, which is counteracted to some degree by the inertia of additional mass as subsequent chain links leave the ground. Similar to bands, chains compensate for mechanical advantages at the end range of motion of a squat or bench press[76] by increasing the resistance and limiting the effects of momentum. As a result, the rate of force development, the time to peak force and time to peak power may be improved; however, this contention has yet to be researched. 3.2.2 Muscle Activity

Despite their emergent popularity, a limited number of studies have examined the muscle activity associated with bands or chains.[81,86] Cronin et al.[86] compared the concentric and eccentric EMG of the vastus lateralis for two supine squats: band and no band conditions, with an equivalent mean load. The only significant differences observed between the two conditions were in the last 40% of eccentric displacement Sports Med 2010; 40 (4)

Biomechanics of Resistance

where the band condition was greater. Similarly, Ebben and Jensen[81] examined quadriceps and hamstring EMG over the eccentric and concentric phase of a back squat over three conditions: squat with barbell and weight; squat with barbell, weights and chains to replace 10% of the load; and squat with barbell, weights and elastics to replace 10% of the load. No significant differences in the mean EMG were found between any of the conditions; however, presenting data as a mean may have masked any variations that did exist through the range of motion. 3.3 Pneumatic

The term ‘pneumatic’ denotes relating to or using air. Analogous to the resistive forces provided by band tension, pneumatic devices offer loads that are not dependent on an object’s mass. Conversely, they are a function of the air pressure being produced and the area through which this pressure is exerted (see equation 14): (Eq: 14Þ P ¼ Fpneumatic =A where P is the air pressure (Pa), Fpneumatic is the resultant force (N) and A is the area through which the air is compressed (m2). As a result, pneumatic loads will, theoretically, exhibit less variation in the muscular force required to complete an exercise stroke, thereby reducing the risk of strain and injury to the operator.[87] The first pneumatic exercising device was developed in 1978,[88] with the aim of duplicating the superior operative characteristics of ‘weight type’ exercising machines while avoiding the undesirable characteristics.[88] Dennis Keiser, founder of Keiser’s pneumatic technology, introduced a device that would not permit the operator to employ inertia to his/her advantage once the movement was initiated, but to exert a force during the return stroke, allowing for greater movement velocity to be achieved with an equivalent load.[88] Limiting the mass of the load will augment the kinetic, kinematic and electromyographic profiles of a movement, potentially resulting in dissimilar training adaptations; however, this remains to be investigated. The numerous researchers that have implemented pneumatic ª 2010 Adis Data Information BV. All rights reserved.

321

technology in their respective protocols have done so as a means of training for the elderly[89-95] and not to assess the biomechanical characteristics of the resistance type. Therefore, the purpose of this section is to characterize pneumatic resistance in terms of Newton’s laws of motion and provide the theoretical framework for future experimental research. 3.3.1 Kinematics and Kinetics Velocity

As stated previously (see section 1), velocity is a function of the duration over which a specific displacement takes place. If the movement is nonballistic in nature, then both the initial and final velocities of the eccentric and concentric phase will be zero, meaning that the only way to change the mean velocity, given a fixed range of motion, is via changes in the duration of the movement. Therefore, for mean velocity to be greater during a pneumatic effort, the duration of that effort needs to be reduced. Peak velocity is a measure of the maximum instantaneous velocity at any point in time during the movement. Because it also represents a zero crossing on the acceleration curve (point that separates acceleration and deceleration phases), larger accelerations over a greater duration of the dynamic effort will elicit higher peak velocities. If pneumatic resistance does provide a means of increasing such acceleration, then there should be a proportional increase in the peak velocity. Acceleration and Force

The proposed advantages of pneumatic technology stem from Newton’s second law, which outlines the relationship between force, acceleration and mass. Referring to equation 1, the acceleration of an object is directly proportional to the sum of the net forces acting on the object and inversely proportional to its mass. Pneumatic devices utilize air pressure as a means of resistance, thereby reducing the mass component of the load to near zero. The magnitude of this mass component will vary in accordance with the body segments involved in producing such movement, but remain significantly lower than the freeweight equivalent, with greater differences being Sports Med 2010; 40 (4)

Frost et al.

322

Fa = 1081 N (applied force)

Fa = 1081 N (applied force)

FW

Pneumatic a = 1 m/sec2

a = 20 m/sec2

Fnet = 100 N

m = 100 kg

Fpneumatic = 932 N

Fr = 981 N (resistive force)

m = 5 kg

Fr = 981 N (resistive force)

Fig. 7. A theoretical kinematic and kinetic comparison of freeweight (FW) and pneumatic resistance using equivalent loads.

seen at higher loads. Consequently, an athlete will be able to achieve greater accelerations with a pneumatic load of the same weight if an identical force is applied (figure 7). Newton’s first law implicitly states that an object will maintain its state of motion unless acted on by an unbalanced force. Resistance to change in this present state, defined as inertia, will depend on the mass of the object, with larger masses characterized by greater resistance. Furthermore, in equation 5, we note that specific changes in the movement velocity necessitate proportional increases in either the magnitude of the applied force or the duration of the applied force for a given mass. Considering sport-specific velocity and power development, the only viable option is to increase the magnitude of the applied force, which consequently limits the load that can be lifted (table I). Because an equivalent pneumatic load uses less mass, less force will be required to produce the same desired velocities, or conversely, the same force will produce greater velocities via greater acceleration (table VI), possibly improving the transference of highvelocity resistance training to sport.

Power

A negligible mass component and near zero momentum makes projecting a pneumatic load near impossible, which results in a mean force that is always equal to the weight of load. Therefore, any increase to mean power must be from a proportional increase in the mean velocity. There are two ways to increase mean velocity: increase the duration of the phase spent accelerating; or increase the magnitude of the acceleration. Pneumatic resistance provides a training stimulus conducive to increased accelerations by reducing the inertia, momentum and peak force requirements of a movement, which may possibly lead to greater mean power. Furthermore, less mass implies less variation in the force curve, and the potential for peak power to be produced with a greater contribution from the movement velocity than the force (a rightward shift of the force-velocity curve; figure 1b). Peak power is an instantaneous measure of the optimal force-velocity combination at a specific point during the movement; however, to increase sport specificity, the velocity component should reflect the speeds at which movements are performed in athletic events. Pneumatic resistance may enable athletes to increase the peak power by producing greater movement velocities without necessitating proportional increases in the force; however, this contention needs validation from experimental research. 3.3.2 Muscle Activity

Newton et al.[12] has shown that during a bench press, the muscle activity of the agonists follow a similar pattern to the force-displacement curve; dramatic reductions in EMG activity are observed during the deceleration portion of the

Table VI. Force-velocity comparison of free-weight (FW) [reproduced from Lander et al.,[3] with permission from Lippincott Williams & Wilkins] and pneumatic resistance using the impulse-momentum theory Resistance

Fload

M

Vo

PV

MF

% MF/Fload

TPV

FW

1370

139.7

0.00

0.54

1704

1.24

0.225

Pneumatic

1370

10.0

0.00

0.54

1394

1.02

0.225

Pneumatic

1370

10.0

0.00

7.52

1704

1.24

0.225

Equation 5: Fnet  ðt2  t1 Þ ¼ ðm  v2 Þ  ðm  v1 Þ Fload = resistive force (N); M = mass of load (kg); MF = mean force between Vo and TPV (N); PV = peak velocity (m/sec); TPV = time to PV (sec); Vo = velocity at start of concentric phase (m/sec).

ª 2010 Adis Data Information BV. All rights reserved.

Sports Med 2010; 40 (4)

Biomechanics of Resistance

movement. Because pneumatic resistance will theoretically limit the variation in force, this resistance mode may also allow an athlete to maintain a more consistent muscular effort throughout a specified range of motion. An increase in agonist activity may result in a longer acceleration phase, thus allowing for greater movement velocities to be produced during the later stages of the concentric phase. However, as was stated for the isokinetic dynamometry, greater muscle activity does not necessarily equate to superior performance. Thus, it should be noted that increasing muscle activity alone is not sufficient evidence to validate a device as superior for performance development. 4. Conclusions Mechanical disadvantages inherent to constant resistance movements may limit the degree to which velocity-specific power output can be produced and, therefore, transference to athletic performance may be compromised. Dramatic variations in the kinetics, kinematics and muscle activity have been reported for both concentriconly and SSC movements because of the inertial properties of the load. Ballistics can reduce the magnitude of these variations via increases in mean force and mean and peak velocity; however, they fail to increase the rate of force development. Consequently, alternative resistance-training strategies continue to be investigated. Isokinetics, hydraulics and variable-resistance devices using cams and levers have offered possible solutions to the constant resistance problems by controlling the movement velocity and/or system load, but have done so by introducing limitations of their own. Movements are restricted to machines, thereby reducing the permissible movement directions from six to one, and have negated the importance of an acceleration component to training, likely decreasing the transference of gym-based strength and power gains to sporting activity. The addition of bands to constant resistance efforts may afford a promising solution, via compensatory accelerations, a variable resistance and a reduced mass. Compared with a constant resistance effort with an equivalent load, bands ª 2010 Adis Data Information BV. All rights reserved.

323

allow for greater accelerations and, thus, velocities to be produced at the onset of the concentric effort, and prevent a reduction in force, velocity, power and muscle activity during the later stages of the movement. As a result, transference to sport may be improved, although this contention is yet to be shown through research; most evidence reports no acute differences compared with similar free-weight movements. Possible characteristics, such as a reduced load at the onset of the concentric phase or large variations in the load-displacement curve, may result in adverse training effects with extended use. Pneumatic resistance offers similar mechanical advantages to both constant resistance and band training: a reduced mass may allow for an increased acceleration and movement velocity at a given load; force, velocity, power and muscle activity should be maintained throughout the concentric phase; and the same load will theoretically be maintained for the duration of the movement. As a result, pneumatics may offer a resistance-training stimulus more conducive to sport-specific velocity and power development; however, this hypothesis remains to be investigated. Alternatively, perhaps it is naive to assume that one resistance type offers a mechanical stimulus conducive to improving all aspects of athletic performance. Sport is multifaceted and necessitates that a methodical strategy be implemented in order to achieve the best results. Understanding the biomechanical advantages and disadvantages associated with each resistance type may simply provide researchers and practitioners with a greater appreciation for the complex nature of performance enhancement, thereby prompting the evolution of hybrid or mixed-method training strategies. Acknowledgements No sources of funding were used to assist in the preparation of this review. The authors have no conflicts of interest that are directly relevant to the content of this review

References 1. Pipes TV. Variable resistance versus constant resistance strength training in adult males. Eur J Appl Physiol 1978; 39 (1): 27-35

Sports Med 2010; 40 (4)

324

2. Cronin JB, McNair PJ, Marshall RN. Force-velocity analysis of strength-training techniques and load: implications for training strategy and research. J Strength Cond Res 2003; 17 (1): 148-55 3. Lander JE, Bates BT, Sawhill JA, et al. A comparison between free-weight and isokinetic bench pressing. Med Sci Sports Exerc 1985; 17 (3): 344-53 4. Hislop HJ, Perrine JJ. The isokinetic concept of exercise. Phys Ther 1967; 47 (2): 114-7 5. McCaw ST, Friday JJ. A comparison of muscle activity between a free weight and machine bench press. J Strength Cond Res 1994; 8 (4): 259-64 6. Abernethy PJ, Jurimae J. Cross-sectional and longitudinal uses of isoinertial, isometric, and isokinetic dynamometry. Med Sci Sports Exerc 1996; 28 (9): 1180-7 7. Jacobs I, Bell DG, Pope J. Comparison of isokinetic and isoinertial lifting tests as predictors of maximal lifting capacity. Eur J Appl Physiol 1988; 57 (2): 146-53 8. Foran B. Advantages and disadvantages of isokinetics, variable resistance and free weights. Natl Strength Cond Assoc J 1985; 7 (1): 24-5 9. Elliott BC, Wilson GJ, Kerr GK. A biomechanical analysis of the sticking region in the bench press. Med Sci Sports Exerc 1989; 21 (4): 450-62 10. Cronin JB, McNair PJ, Marshall RN. The role of maximal strength and load on initial power production. Med Sci Sports Exerc 2000; 32 (10): 1763-9 11. McLaughlin TM, Madsen NH. Bench press techniques of elite heavyweight powerlifters. Natl Strength Cond Assoc J 1984; 6 (4): 44, 62-5 12. Newton RU, Murphy AJ, Humphries BJ, et al. Influence of load and stretch shortening cycle on the kinematics, kinetics and muscle activation that occurs during explosive upper-body movements. Eur J Appl Physiol 1997; 75 (4): 333-42 13. Newton RU, Kraemer WJ, Hakkinen K, et al. Kinematics, kinetics, and muscle activation during explosive upper body movements. J Appl Biomech 1996; 12 (1): 31-43 14. Alamasbakk B, Hoff J. Coordination, the determinant of velocity specificity? J Appl Physiol 1996; 80 (5): 2046-52 15. Baker D. Differences in strength and power among juniorhigh, senior-high, college aged, and elite professional rugby league players. J Strength Cond Res 2002; 16 (4): 581-5 16. Baker D. Comparison of upper-body strength and power between professional and college-aged rugby league players. J Strength Cond Res 2001; 15 (1): 30-5 17. Baker D, Nance S. The relationship between strength and power in professional rugby league players. J Strength Cond Res 1999; 13 (3): 224-9 18. Baker D. A series of studies on the training of high intensity muscle power in rugby league football players. J Strength Cond Res 2001; 15 (2): 198-209 19. Baker D, Nance S, Moore M. The load that maximizes the average mechanical power output during explosive bench press throws in highly trained athletes. J Strength Cond Res 2001; 15 (1): 20-4 20. Baker D, Nance S, Moore M. The load that maximizes the average mechanical power output during jump squats in power-trained athletes. J Strength Cond Res 2001; 15 (1): 92-7

ª 2010 Adis Data Information BV. All rights reserved.

Frost et al.

21. Cronin JB, Henderson ME. Maximal strength and power assessment in novice weight trainers. J Strength Cond Res 2004; 18 (1): 48-52 22. Siegal JA, Gilders RM, Staron RS, et al. Human muscle power output during upper and lower body exercises. J Strength Cond Res 2002; 16 (2): 173-8 23. Izquierdo M, Hakkinen K, Gonzalez-Badillo JJ, et al. Effects of long term training specificity on maximal strength and power of the upper and lower extremities in athletes from different sports. Eur J Appl Physiol 2002; 87: 264-71 24. Siff MC. Supertraining. 6th ed. Denver (CO): Supertraining Institute, 2003 25. Stone MH, O’Bryant HS, McCoy L, et al. Power and maximum strength relationships during performance of dynamic and static weighted jumps. J Strength Cond Res 2003; 17 (1): 140-7 26. Rahmani A, Viale F, Dalleau G, et al. Force/velocity and power/velocity relationships in squat exercise. Eur J Appl Physiol 2001; 84 (3): 227-32 27. Wilson GJ, Newton RU, Murphy AJ, et al. The optimal training load for the development of dynamic athletic performance. Med Sci Sports Exerc 1993; 25 (11): 1279-86 28. McBride JM, Triplett-McBride T, Davie A, et al. The effect of heavy versus light load jump squats on the development of strength, power, and speed. J Strength Cond Res 2002; 16 (1): 75-82 29. Haff GG, Stone MH, O’Bryant HS, et al. Force-time dependent characteristics of dynamic and isometric muscle actions. J Strength Cond Res 1997; 11 (4): 269-72 30. Carlock JM, Smith SL, Hartman MJ, et al. The relationship between vertical jump power estimates and weightlifting ability: a field test approach. J Strength Cond Res 2004; 18 (3): 534-9 31. Hill AV. The heat of shortening and the dynamic constants of muscle. Proc R Soc Biol 1938; 126: 136-95 32. Barnett CB, Kippers V, Turner P. Effects of variations of the bench press exercise on the EMG activity of five shoulder muscles. J Strength Cond Res 1995; 9 (4): 222-7 33. Lehman GJ. The influence of grip width and forearm pronation/ supination on upper-body myoelectric activity during the flat bench press. J Strength Cond Res 2005; 19 (3): 587-91 34. Jaric S, Ropret R, Kukolj M, et al. Role of agonist and antagonist muscle strength in performance of rapid movements. Eur J Appl Physiol 1995; 71 (5): 464-8 35. van Ingen Schenau GJ, Bobbert MF, de Haan A. Does elastic energy enhance work and efficiency in the stretch shortening cycle? J Appl Biomech 1997; 13 (4): 389-415 36. Wilson GJ, Wood GA, Elliott BC. Optimal stiffness of series elastic component in a stretch shorten cycle activity. J Appl Physiol 1991; 70 (2): 825-33 37. Bosco C, Komi PV. Potentiation of the mechanical behavior of the human skeletal muscle through prestretching. Acta Physiol Scand 1979; 106 (4): 467-72 38. Bosco C, Vitasalo JT, Komi PV, et al. The combined effect of elastic energy and myoelectrical potentiation during stretch shortening cycle exercise. Acta Physiol Scand 1982; 114 (4): 557-65 39. Bobbert MF, Gerritsen GM, Litjens MCA, et al. Why is countermovement jump height greater than squat jump height? Med Sci Sports Exerc 1996; 28 (11): 1402-12

Sports Med 2010; 40 (4)

Biomechanics of Resistance

40. Ettema GJC. Muscle efficiency: the controversial role of elasticity and mechanical energy conversion in stretch shortening cycles. Eur J Appl Physiol 2001; 85 (5): 457-65 41. Walshe AD, Wilson GJ, Ettema GJC. Stretch shorten cycle compared with isometric preload: contributions to enhanced muscular performance. J Appl Physiol 1998; 84 (1): 97-106 42. Wilson GJ, Elliott BC, Wood GA. Stretch shorten cycle performance enhancement through flexibility training. Med Sci Sports Exerc 1992; 24 (1): 116-23 43. Wilson GJ, Elliott BC, Wood GA. The effect on performance of imposing a delay during a stretch shorten cycle movement. Med Sci Sports Exerc 1991; 23 (3): 364-70 44. Cronin J, McNair PJ, Marshall R. Developing explosive power: a comparison of technique and training. J Sci Med Sport 2001; 4 (1): 59-70 45. Newton RU, Kraemer WJ. Developing explosive muscular power: implications for a mixed methods training strategy. Strength Cond 1994; 16 (5): 20-31 46. Vossen JF, Kramer JF, Burke DG, et al. Comparison of dynamic push-up training and plyometric push-up training on upper body power and strength. J Strength Cond Res 2000; 14 (3): 248-53 47. Baker D, Newton RU. Acute effect on power output of alternating an agonist and antagonist muscle exercise during complex training. J Strength Cond Res 2005; 19 (1): 202-5 48. Behm DG. Neuromuscular implications and applications of resistance training. J Strength Cond Res 1995; 9 (4): 264-74 49. Behm DG, Sale DG. Velocity specificity of resistance training. Sports Med 1993; 15 (6): 374-88 50. Zatsiorsky VM. Science and practice of strength training. Champaign (IL): Human Kinetics, 1995 51. Holt L, Pelham TW. The double acting concentric dynamometer. Natl Strength Cond Assoc J 1992; 14 (6): 35-8 52. Pereira MIR, Gomes PSC. Movement velocity in resistance training. Sports Med 2003; 33 (6): 427-38 53. Hortobagyi T, Katch FI, LaChance PF. Interrelationships among various measures of upper body strength assessed by different contraction modes: evidence for a general strength component. Eur J Appl Physiol 1989; 58 (7): 749-55 54. Aagaard P, Simonsen EB, Trolle M, et al. Specificity of training velocity and training load on gains in isokinetic knee joint strength. Acta Physiol Scand 1996; 156 (2): 123-9 55. Burke DG, Pelham TW, Holt LE. The influence of varied resistance and speed of concentric antagonistic contractions on subsequent concentric agonistic efforts. J Strength Cond Res 1999; 13 (3): 193-7 56. Telle JR, Gorman IJ. Combining free weights with hydraulics. Natl Strength Cond Assoc J 1985; 7 (1): 66-8 57. Behm DG. An analysis of intermediate speed resistance exercises for velocity specific gains. J Appl Sport Sci Res 1991; 5 (1): 1-5 58. Hunter GR, Culpepper MI. Joint angle specificity of fixed mass versus hydraulic resistance knee flexion training. J Strength Cond Res 1995; 9 (1): 13-6 59. Peterson SR, Bagnall M, Wenger HA, et al. The influence of velocity specific resistance training on the in vivo torque velocity relationship and the cross sectional area of quadriceps femoris. J Orthop Sports Phys Ther 1989; 10 (11): 456-62

ª 2010 Adis Data Information BV. All rights reserved.

325

60. Cronin JB, McNair PJ, Marshall RN. Is velocity-specific strength training important in improving functional performance? J Sports Med Phys Fitness 2002 Sep; 42 (3): 267-73 61. Abernethy PJ, Wilson G, Logan P. Strength and power assessment: issues, controversies and challenges. Sports Med 1995; 19 (6): 401-17 62. Caiozzo VJ, Perrine JJ, Edgerton VR. Training induced alterations of the in vivo force velocity relationship of human muscle. J Appl Physiol 1981; 51 (3): 750-4 63. Lesmes GR, Costill DL, Coyle EF, et al. Muscle strength and power changes during maximal isokinetic training. Med Sci Sports 1978; 10 (4): 266-9 64. Kanehisa H, Miyashita M. Specificity of velocity in strength training. Eur J Appl Physiol 1983; 52 (1): 104-6 65. Brown LE, Whitehurst M. The effect of short term isokinetic training on force and rate of velocity development. J Strength Cond Res 2003; 17 (1): 88-94 66. Rutherford OM, Purcell C, Newham DJ. The human forcevelocity relationship; activity in the knee flexor and extensor muscles before and after eccentric practice. Eur J Appl Physiol 2001; 84 (1-2): 133-40 67. Pizzimenti MA. Mechanical analysis of the Nautilus leg curl machine. Can J Sport Sci 1992; 17 (1): 41-8 68. Kanehisa H, Miyashita M. Effect of isometric and isokinetic muscle training on static strength and dynamic power. Eur J Appl Physiol 1983; 50 (3): 365-71 69. Chen WL, Su FC, Chou YL. Significance of acceleration period in a dynamic strength testing study. J Orthop Sports Phys Ther 1994; 19 (6): 324-30 70. Brown LE, Whitehurst M, Gilbert R, et al. The effect of velocity and gender on load range during knee extension and flexion exercise on an isokinetic device. J Orthop Sports Phys Ther 1995; 21 (2): 107-12 71. Jones DA, Rutherford OM, Parker DF. Physiological changes in skeletal muscle as a result of strength training. Q J Exp Physiol 1989; 74 (3): 233-56 72. Behm DG, Sale DG. Intended rather than actual movement velocity determines velocity specific training response. J Appl Physiol 1993; 74 (1): 359-68 73. Moss BM, Refsnes PE, Abildgaard A, et al. Effects of maximal effort strength training with different loads on dynamic strength, cross-sectional area, load-power and load-velocity relationships. Eur J Appl Physiol 1997; 75 (3): 193-9 74. Toji H, Kaneko M. Effect of multiple load training on the forcevelocity relationship. J Strength Cond Res 2004; 18 (4): 792-5 75. Onishi H, Yagi R, Oyama M, et al. EMG-angle relationship of the hamstring muscles during maximum knee flexion. J Electromyogr Kinsiol 2002; 12 (5): 399-406 76. Simmons LP. Chain reactions: accommodating leverages. Powerlifting USA 1996; 19 (12): 2-3 77. Ariel G. Variable resistance versus standard resistance training. Scholastic Coach 1976 Dec; 46 (5): 68, 69, 74 78. Simmons LP. Bands and chains. Powerlifting USA 1999; 22 (6): 26-7 79. Smith F. Strength training modes: dynamic variable resistance and the universal system. Natl Strength Cond Assoc J 1982; 4 (4): 14-9 80. Harman E. Resistive torque analysis of five Nautilus exercise machines. Med Sci Sports Exerc 1983; 7: 248-61

Sports Med 2010; 40 (4)

326

81. Ebben WP, Jensen RL. Electromyographic and kinetic analysis of traditional, chain, and elastic band squats. J Strength Cond Res 2002; 16 (4): 547-50 82. Berning JM, Coker CA, Adams KJ. Using chains for strength and conditioning. Strength Cond J 2004; 26 (5): 80-4 83. Wallace BJ, Winchester JB, McGuigan MR. Effects of elastic bands on force and power characteristics during the back squat exercise. J Strength Cond Res 2006; 20 (2): 268-72 84. Findley BW. Training with rubber bands. Strength Cond J 2004; 26 (6): 68-9 85. Behm DG. Surgical tubing for sport and velocity specific training. Strength Cond J 1988; 10 (4): 66-70 86. Cronin JB, McNair PJ, Marshall RN. The effects of bungy weight training on muscle function and functional performance. J Sports Sci 2003; 21 (1): 59-71 87. Keiser DL, inventor, Kintron Incorporated, assignee. Exercising device including linkage for control of muscular exertion required through exercising stroke. US Patent 4, 227, 689. 1980 88. Keiser DL, inventor, Keiser, assignee. Pneumatic exercising device. US Patent 4, 257, 593. 1981 89. Schroeder ET, Vallejo AF, Zheng L, et al. Six week improvements in muscle mass and strength during androgen therapy in older men. J Gerontol 2005; 60A (12): 1586-92

ª 2010 Adis Data Information BV. All rights reserved.

Frost et al.

90. Paulus DC, Reiser RFI, Troxell WO. Pneumatic strength assessment device: design and isometric measurement. Biomed Sci Instrum 2004; 40: 277-82 91. Orr R, de Vos NJ, Singh NA, et al. Power training improves balance in healthy older adults. J Gerontol 2006; 61A (1): 78-85 92. Joszi AC, Campbell WW, Joseph L, et al. Changes in power with resistance training in older and younger men and women. J Gerontol 1999; 54A (11): M591-6 93. Herman S, Kiely DK, Leveille S, et al. Upper and lower limb muscle power relationships in mobility limited older adults. J Gerontol 2005; 60A (4): 476-80 94. Fielding RA, LeBrasseur NK, Cuoco A, et al. High velocity resistance training increases skeletal muscle peak power in older women. J Am Geriatr Soc 2002; 50 (4): 655-62 95. de Vos NJ, Singh NA, Ross DA, et al. Optimal load for increasing muscle power during explosive resistance training in older adults. J Gerontol 2005; 60A (5): 638-47

Correspondence: David Frost, Department of Kinesiology, University of Waterloo, 200 University Avenue West, Waterloo, ON N2L 3G1, Canada. E-mail: [email protected]

Sports Med 2010; 40 (4)

REVIEW ARTICLE

Sports Med 2010; 40 (4): 327-346 0112-1642/10/0004-0327/$49.95/0

ª 2010 Adis Data Information BV. All rights reserved.

Intravenous versus Oral Rehydration in Athletes Simon Piet van Rosendal,1 Mark Andrew Osborne,2 Robert Gordon Fassett,3 Bill Lancashire4 and Jeff Scott Coombes1 1 2 3 4

School of Human Movement Studies, University of Queensland, Brisbane, Queensland, Australia Queensland Academy of Sport, Brisbane, Queensland, Australia Renal Medicine, Royal Brisbane and Women’s Hospital, Brisbane, Queensland, Australia Port Macquarie Base Hospital, Port Macquarie, New South Wales, Australia

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Overview of Intravenous Fluid and Exercise Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Physiological Responses to Intravenous Fluid Infusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Cardiovascular Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1 Plasma Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2 Heart Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Thermoregulatory Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 Core Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2 Skin/Forearm Blood Flow and Skin Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.3 Sweat Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Endocrine Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1 Fluid-Regulating Hormones: Antidiuretic Hormone and Aldosterone. . . . . . . . . . . . . . . . . . 2.3.2 Stress Hormones: Adrenocorticotropic Hormone and Cortisol . . . . . . . . . . . . . . . . . . . . . . . 2.3.3 Catecholamines: Noradrenaline and Adrenaline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Subjective Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.1 Thirst Sensation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.2 Thermal Sensation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.3 Rating of Perceived Exertion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Exercise Performance Following Intravenous versus Oral Rehydration . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Effects of Intravenous Fluids on Recovery from Exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Areas for Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 Volume of Fluid Replenished . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Effects of Using Partial Intravenous and Partial Oral Rehydration . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 Effect of Intravenous Fluid on Acid-Base Physiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4 Effect of Intravenous Fluid on Respiratory Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Adverse Effects from Intravenous Rehydration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

327 329 333 334 334 335 335 338 338 339 339 339 339 340 341 341 341 342 342 342 343 343 343 343 344 344 344

Fluid is typically administered via intravenous (IV) infusion to athletes who develop clinical symptoms of heat illness, based on the perception that dehydration is a primary factor contributing to the condition. However, other athletes also voluntarily rehydrate with IV fluid as opposed to, or in

van Rosendal et al.

328

conjunction with, oral rehydration. The voluntary use of IV fluids to accelerate rehydration in dehydrated, though otherwise healthy athletes, has recently been banned by the World Anti-Doping Agency. However, the technique remains appealing to many athletes. Given that it now violates the Anti-Doping Code, it is important to determine whether potential benefits of using this technique outweigh the risks involved. Several studies have shown that rehydration is more rapid with IV fluid. However, the benefits are generally transient and only small differences to markers of hydration status are seen when comparing IV and oral rehydration. Furthermore, several studies have shown improvements in cardiovascular function and thermoregulation with IV fluid, while others have indicated that oral fluid is superior. Subsequent exercise performance has not been improved to a greater extent with one technique over the other. The paucity of definitive findings is probably related to the small number of studies investigating these variables and the vast differences in the designs of studies that have been conducted. The major limitation of IV rehydration is that it bypasses oropharyngeal stimulation, which has an influence on factors such as thirst sensation, antidiuretic hormone (arginine vasopressin) release, cutaneous vasodilation and mean arterial pressure. Further research is necessary to determine the relative benefits of oral and IV rehydration for athletes.

It is well accepted that moderate to severe dehydration detrimentally affects endurance exercise performance,[1,2] and also probably attenuates strength, power and high-intensity endurance (30 seconds to 2 minutes effort).[3] Traditionally, a 2% bodyweight reduction has been recognized as a physiological ‘threshold’ sufficient to consistently reduce endurance performance.[4] However, the magnitude of this threshold has recently been questioned,[1] based on multiple studies suggesting that athletes who performed the best in marathon and iron-man triathlon races were frequently also the most ‘dehydrated’.[5,6] Although the measurement of ‘dehydration’ in these studies is ambiguous, they raise the contentious issues at the centre of this debate. What level of dehydration constitutes a physiological perturbation sufficient to reduce performance? And how much fluid is required to stay on the euhydration side of this threshold, thereby maintaining performance? These questions are important because they influence the rehydration behaviours of athletes. Many athletes have recorded sweat rates of 2–3 L/h or greater,[7] and therefore may require ª 2010 Adis Data Information BV. All rights reserved.

very large volumes of fluid for adequate rehydration to stay above the ‘dehydration threshold’. The most logical method for replacing fluid deficits is via oral fluid ingestion. However, there are athletes for whom oral rehydration is difficult, leading to the use of intravenous (IV) fluids in athletes in two contexts. The first is as a medical intervention for collapsed athletes who show clinical symptoms and signs of heat illness and are generally unable to consume fluids orally. Under these circumstances it is typically assumed that dehydration is a primary cause of the condition and that rapidly administering fluids is imperative to reverse the symptoms and signs.[8] As discussed by Casa et al.,[9] the World AntiDoping Agency (WADA) permits IV infusions in these circumstances as a legitimate medical treatment.[10] Several papers have been published in recent years discussing issues associated with IV use in collapsed athletes, such as the importance of measuring serum sodium before providing large volumes of IV fluid to eliminate the risk of iatrogenic hyponatraemia.[8,11-14] This is a large topic worthy of a review in itself and is not the focus of the present article. Sports Med 2010; 40 (4)

Intravenous versus Oral Rehydration

The second and more contentious scenario of IV fluid infusion is as an ergogenic aid to rapidly restore the body’s fluid balance in ‘healthy’ athletes. Two groups of athletes are typically represented in this situation: athletes from predominantly individual sports such as triathlon and cycling who voluntarily rehydrate via this technique; and athletes from team sports for whom IV rehydration is included as a component of recovery practices administered by the sports club. The voluntary use of IV rehydration is often based on the perception that IV fluids are more effective at replenishing lost body water than oral fluid, leading to a shortened recovery time.[15] Indeed, IV fluid more rapidly enhances circulating blood volume because it bypasses the absorption delays associated with oral rehydration.[16-18] However, for many of these athletes IV fluid is also considered to be a logistically beneficial rehydration technique when large volumes are required with little time available for rehydration. For example, in a sport such as American football, athletes weighing upwards of 135 kg undertake pre-season training sessions of 2–3 hours, twice a day, with approximately 4 hours recovery between sessions. With average sweat rates of 2–3 L/h, these athletes may be required to replenish more than 6 L after each training session, or 10–12 L in a 10-hour stretch. It is therefore desirable for these groups of athletes and their athletic trainers or physicians to consider IV rehydration. However, while IV fluid is allowed as a medical intervention,[9] WADA has recently added IV infusions to their list of prohibited ergogenic methods.[10] Therefore, an athlete can only legally receive IV fluids for rehydration when it is medically warranted.[9] The present review focuses on the use of IV fluids as an ergogenic rehydration technique in healthy athletes. The primary aim is to establish whether IV rehydration may be physiologically advantageous compared with oral rehydration for athletes who are dehydrated and otherwise healthy (e.g. not experiencing heat-associated illnesses). Further articles comparing the use of IV fluids with differing compositions (hypotonic and isotonic) are also discussed. When discussing ª 2010 Adis Data Information BV. All rights reserved.

329

specific studies within this review, unless otherwise indicated, oral fluid refers to 0.45% sodium chloride (NaCl) solution ingested orally, and hypotonic and isotonic IV indicate 0.45% and 0.9% NaCl solutions given intravenously. 1. Overview of Intravenous Fluid and Exercise Research Throughout this review, the terms ‘study’ and ‘article’ refer to a specific publication in a journal, while the terms ‘trial’, ‘experiment’ and ‘investigation’ refer to the physical process of testing subjects. The MEDLINE database was searched using PubMed for articles published before 12 January 2009 that investigated the use of IV fluid for rehydration in athletes. The search found 3 articles investigating the admission of IV fluids during exercise and 11 articles reporting the use of IV fluids when rehydrating following exercise-induced dehydration (table I). A number of issues need to be raised before discussing the findings. Firstly, data in the 11 published rehydration articles originate from only 5 trials. One of these was from a trial administering IV fluid following a marathon race and therefore contained little scientific intervention.[22] The remaining ten articles are from four laboratory based experiments.[15-18,23-28] To assist the reader understanding the protocols used in these trials, figures 1–4 have been constructed to illustrate the differences between the conditions (e.g. what rehydrating solution was administered) and the timepoints at which outcomes were assessed. As can be seen from these figures, the five articles by Castellani et al.,[15,24] Kenefick et al.[26,28] and Riebe et al.[23] report different aspects of the same major experiment (figure 1), as do those by Casa et al.[17,25] and Maresh et al. (figure 2).[16] Thus, more research in the area is still needed. Secondly, many of the discrepant findings reported between studies may be related, at least in part, to differences in study designs and timepoints at which primary outcome measures were assessed. As can be seen in figures 1–4, the typical design of IV versus oral rehydration studies has been to: (i) dehydrate subjects via exercise in the heat; (ii) rehydrate subjects with IV or oral fluid; Sports Med 2010; 40 (4)

330

ª 2010 Adis Data Information BV. All rights reserved.

Table I. Summary of protocols of studies investigating intravenous (IV) rehydration in relation to exercise Study

Subjects

Treatments

Dehydration protocol

Dehydration Rehydration protocol % BW

Rehydration Exercise % BW environment

Exercise protocol

Isotonic IV No fluid

No pre-exercise dehydration

NA

1.28 L of 0.9% NaCl infused as rapidly as possible (69 mL/min) IV = 24C

NR

24C, 26% RH

Cycling at 73% . VO2max until exhaustion 50 min cycling at . 60% VO2max; IV infused between min 20–50

IV during exercise 9 males, healthy fit adults

Nose et al.[20]

6 males, healthy adults Isotonic IV, warm environment No fluid, warm environment Isotonic IV, cool environment No fluid, cool environment

NA 30 min passive equilibrium period preexercise Cycled at 60% . VO2max for 50 min then further 40 min passive rest

IV infusion at 29 mL/kg BW during min 20–50 of exercise

NR

22C, <30% RH or 30C, <30% RH

Hamilton et al.[21] experiment I

10 fit, heat-acc

No pre-exercise dehydration

NA

NR

22C

Cycling at 70% . VO2max for 120 min

IV (18% glucose) No pre-exercise No fluid dehydration

NA

Oral fluid given in equal volumes at 0, 20, 40, 60, 80 and 100 min of exercise Average total volume 2.34 L IV infused from 8 to 118 min of exercise at rate adjusted to maintain blood glucose level at 10.1 mmol/L Average total volume 1224 mL

NR

22C

Cycling at 70% . VO2max for 120 min

Isotonic IV Rotterdam (100 mL – control) marathon Hypotonic IV with 2.5% glucose, (2.5 L)

NR

NR

5.8C, 74% RH

(Recovery from marathon)

Hypotonic IV Oral No fluid

-4

Isotonic IV (100 mL) infused over 15 min (acting as control) Hypotonic IV with 2.5% glucose (2.5 L) infused over 50 min 15 min passive equilibrium period at 25.5C; rehydration over 45 min; oral = 4C, equal volumes every 5 min for 45 min; IV = 25C at 56 mL/kg/min; total fluid intake (IV and oral) ~25 mL/kg pre-dehydration BW; exercise began 75 min after rehydration finished

2.5

36C, 47% RH, Treadmill walking . air flow = 2.3 m/sec at ~50% VO2max for up to 90 min, 3–5% grade

Hamilton 8 fit, heat-acc et al.[21] experiment II

IV during rehydration 66 male marathon Polak runners et al.[22]

Riebe et al.[23]

8 males, non-acc, . VO2 = 58 mL/kg/min

Oral No fluid

Alternating cycle and treadmill Total exercise duration 2–4 ha

Sports Med 2010; 40 (4)

Continued next page

van Rosendal et al.

Deschamps et al.[19]

Subjects

Treatments

Dehydration protocol

Dehydration Rehydration protocol % BW

Rehydration Exercise % BW environment

Exercise protocol

Castellani et al.[15]

8 males, non-acc, . VO2 = 58 mL/kg/min

Hypotonic IV Oral No fluid

Alternating cycle and treadmill Total exercise duration 2–4 ha

-4

15 min passive equilibrium period at 25.5C; rehydration over 45 min; oral = 4C, equal volumes every 5 min for 45 min; IV = 25C at 56 mL/kg/min; total fluid intake (IV and oral) ~25 mL/kg pre-dehydration BW; exercise began 75 min after rehydration finished

2.5

36C, 47% RH, Treadmill walking . air flow = 2.3 m/sec at ~50% VO2max for up to 90 min, 3–5% grade

Castellani et al.[24]

8 males, non-acc, . VO2 = 58 mL/kg/min

Isotonic IV Hypotonic IV No fluid

Alternating cycle and treadmill Total exercise duration 2–4 ha

-4

15 min passive equilibrium period at 25.5C; rehydration over 45 min; IV = 25C at 56 mL/kg/min; total fluid intake ~25 mL/kg predehydration BW; exercise began 75 min after rehydration finished

2.5

36C, 47% RH, 45 min treadmill air flow = 2.3 m/sec at 53–54% . VO2max

Casa et al.[17]

8 males, trained, . non-acc, VO2 = 61.4 mL/kg/min

Hypotonic IV Oral No fluid

Cycling at moderate intensity for 2 h following fluidreduced meals

-4 (overnight)

Rehydration followed 110 min passive sitting in 22C environment; rehydration over 20 min period, 2% BW; oral = 10C, equal volumes every 2 min; IV = 22C at 67.1 mL/min; exercise performance test began immediately after rehydration

2

37C, 48% RH, Cycling at 74% . air flow = 2.3 m/sec VO2peak until exhaustion

Casa et al.[25]

8 males, trained, . non-acc, VO2 = 61.4 mL/kg/min

Hypotonic IV Oral No fluid

Cycling at moderate intensity for 2 h following fluidreduced meals

-4 (overnight)

Rehydration followed 110 min passive sitting in 22C environment; rehydration over 20 min period, 2% BW; oral = 10C, equal volumes every 2 min; IV = 22C at 67.1 mL/min; exercise performance test began immediately after rehydration

2

37C, 48% RH, Cycling at 74% . air flow = 2.3 m/sec VO2peak to exhaustion

Continued next page

331

Sports Med 2010; 40 (4)

Study

Intravenous versus Oral Rehydration

ª 2010 Adis Data Information BV. All rights reserved.

Table I. Contd

332

ª 2010 Adis Data Information BV. All rights reserved.

Table I. Contd Subjects

Treatments

Dehydration protocol

Dehydration Rehydration protocol % BW

Rehydration Exercise % BW environment

Exercise protocol

Kenefick et al.[26]

8 males, non-acc, . VO2 = 58 mL/kg/min

Isotonic IV Hypotonic IV Oral No fluid

Alternating cycle and treadmill Total exercise duration 2–4 ha

-4.4 to -5.1

15 min passive equilibrium period at 25.5C; rehydration over 45 min; oral = 4C, equal volumes every 5 min for 45 min; IV = 25C at 56 mL/kg/min; total fluid intake (IV and oral) ~25 mL/kg pre-dehydration BW; exercise began 75 min after rehydration finished

2–3

NA

NA

Maresh et al.[16]

8 males, trained, . non-acc, VO2 = 61.4 mL/kg/min

Hypotonic IV Oral No fluid

Cycling at moderate intensity for 2 h following fluidreduced meals

-4 (overnight)

Rehydration followed 110 min passive sitting in 22C environment; rehydration over 20 min period, 2% BW; oral = 10C, equal volumes every 2 min; IV = 22C at 67.1 mL/min; exercise performance test began immediately after rehydration

2

37C, 48% RH, Cycling at 74% . air flow = 2.3 m/sec VO2peak until exhaustion

Echegaray et al.[27]

8 males, healthy, . VO2 = 57.3 mL/kg/min

Isotonic IV Hypotonic IV No fluid

Alternating cycle and treadmill Total exercise duration 2–4 hb

-4

20 min passive equilibrium period at 22–25C; rehydration over 45 min; IV = 25C at 56 mL/kg/min; total fluid intake ~25 mL/kg pre-dehydration BW; exercise began 75 min after rehydration finished

2.4

36Cc

Kenefick et al.[18]

8 males, non-acc, . VO2 = 63.7 mL/kg/min

Hypotonic IV Oral No fluid

75 min treadmill walking at 50% . VO2max

-2.8

Subjects remained in 37C environment; rehydration over 20 min: oral = 15C, equal volumes every 4 min; IV = 15C at 85.5 mL/kg/min; total fluid intake = 100% BW lost during dehydration; exercise began immediately after rehydration finished

2.8

37C, 42% RH, Up to 75 min air flow = 6.1 m/sec treadmill walking . at 50% VO2max (2.4 m/sec, 2–3% grade, air flow = 6.1 m/sec)

Treadmill walking . at ~50% VO2max for up to 90 min, 5–10% grade

Continued next page

van Rosendal et al.

Sports Med 2010; 40 (4)

Study

BW = bodyweight; heat-acc = heat-acclimated; hypotonic IV = 0.45% NaCl; isotonic IV = 0.9% . NaCl; NA = not applicable; NaCl . = sodium chloride; non-acc = non-acclimated; . NR = not reported; oral = oral fluid (0.45% NaCl); RH = relative humidity; VO2 = oxygen uptake; VO2max = maximal oxygen uptake; VO2peak = peak oxygen uptake.

20 min passive equilibrium . – alternating cycle and treadmill in intervals of 25 min with 5 min breaks between. Urine included as weight loss. Total exercise duration was 2–4 h (mean 186 min). Mean VO2max = 50.5%. Temperature 33C, wind speed 2.3 m/sec, humidity NR. b

ª 2010 Adis Data Information BV. All rights reserved.

c Humidity and air flow NR.

20 min passive equilibrium – alternating cycle and treadmill in intervals of 25 min with 5 min breaks between. Urine included as weight loss. Total exercise duration was . 179–190 min. Mean VO2max = 50.8%. Temperature 33C, 47.6% RH, wind speed 2.3 m/sec.

333

a

36C, 47% RH, Treadmill walking . air flow = 2.3 m/sec at ~50% VO2max for up to 90 min, 3–5% grade -4.4 to -5.1 Alternating cycle and treadmill Total exercise duration 2–4 ha Isotonic IV Hypotonic IV Oral No fluid 8 males, non-acc, . VO2 = 58 mL/kg/min Kenefick et al.[28]

2.5 15 min passive equilibrium period at 25.5C; rehydration over 45 min; oral = 4C, equal volumes every 5 min for 45 min; IV = 25C at 56 mL/kg/min; total fluid intake (IV and oral) ~25 mL/kg predehydration BW; exercise began 75 min after rehydration finished

Dehydration Rehydration protocol % BW Dehydration protocol Treatments Subjects Study

Table I. Contd

Rehydration Exercise % BW environment

Exercise protocol

Intravenous versus Oral Rehydration

and (iii) complete a further exercise bout or performance test. However, the vertical lines in figures 1–4 display the timepoints at which specific variables were reported in each article and the vast difference between them is apparent. Further critical factors associated with study design include: (i) IV fluid being administered during exercise, which is unrealistic in a practical application;[19-21] (ii) approximately 1100 mL more fluid given in oral than in IV trials;[21] (iii) 18% glucose being given in the IV trial but not in the oral trial;[21] (iv) no comparisons being made with oral rehydration;[19,20] (v) subjects being only partially rehydrated;[15-17,23,26-28] and (vi) IV and oral fluids being administered at different temperatures.[15-17,23-26,28] In relation to subsequent exercise, in two studies exercise was performed in a mild environment[20,21] and eight articles report using set exercise duration as opposed to more specific performance tests (e.g. time trial).[15,18,20,21,23,24,27,28] Finally, in one experiment reported over a number of articles, subjects were sent home overnight dehydrated by -4% bodyweight and asked to consume foods low in fluids. They then returned the next morning (12 hours postprandial), undertook a further 2-hour rest period to fulfil the criteria for a further trial and were only halfrehydrated before completing the performance test (figure 2).[16,17,25] While it is common for athletes to inadequately rehydrate following exercise, and therefore present in a dehydrated state the following morning, it is unlikely that athletes would typically forego any attempt at rehydrating before subsequent exercise.

2. Physiological Responses to Intravenous Fluid Infusion The mode of rehydration has important implications for many physiological functions. IV rehydration bypasses oropharyngeal stimulation, which has an influence on factors such as thirst sensation,[29] antidiuretic hormone (ADH; arginine vasopressin) release,[29] cutaneous vasodilation[30] and mean arterial pressure.[30] The following sections discuss the findings of studies comparing IV fluid infusion with oral fluid Sports Med 2010; 40 (4)

van Rosendal et al.

334

Dehydration ~180 min

Rest Rehydration 15 min 45 min

Rest 55 min

Equilibration 20 min

Performance test 90 min

Experiment outline PrD

PoD

PrD

PoD

PoR

PrE

PoE

Castellani et al.[15] PrE 15 30 45 60 75 PoE

Riebe et al.[23] PoD

PoR 15

35

55

PrE

20

60

40

80

Castellani et al.[24] PrD

PoD

PrD

PoD

PrD

PoD

PrE 15

45

Kenefick et al.[26] 15

35

55

Kenefick et al.[28] PoE

PrE [15,24]

[23]

Fig. 1. Experiment outline and overview of the studies by Castellani et al., Riebe et al. and Kenefick et al. dration; PoE = post-exercise; PoR = post-rehydration; PrD = pre-dehydration; PrE = pre-exercise.

ingestion on physiological parameters associated with hydration and exercise performance. 2.1 Cardiovascular Parameters

Rapidly restoring the body’s fluid balance to a euhydrated state should quickly reverse the effects of dehydration. However, whether IV or oral fluid will most effectively restore whole body fluid homeostasis is unclear. 2.1.1 Plasma Volume

The maintenance of plasma volume is important for both thermoregulatory and cardioDehydration ~180 min

Experiment outline PrD

Equilibration 20 min PoD

[26,28]

PoD = post-dehy-

vascular function. A goal of rehydrating both during and after exercise is therefore to assist the preservation of plasma volume. IV fluid enters the vasculature immediately, thus bypassing the time constraints associated with oral fluid absorption. Several studies have shown that plasma volume is better maintained with IV fluid infused during exercise, compared with oral fluid.[19-21] Furthermore, whether following sustained (i.e. overnight [figure 2])[17] or acute[18] exercise-induced dehydration, plasma volume is consistently restored more rapidly and completely when rehydrating with IV compared with oral fluid. Rest 110 min

Rehydration Performance 20 min test ~35 min

PoD

PrE

PoE

PoD

PrE

PoE

PoD

PrE

PoE

PrE

PoE

Day 1

PrR

Day 2

Casa et al.[17] PrD

PoD

Casa et al.[25]

Maresh et al.[16]

Fig. 2. Experiment outline and overview of the studies by Casa et al.[17,25] and Maresh et al.[16] PoD = post-dehydration; PoE = post-exercise; PrD = pre-dehydration; PrE = pre-exercise; PrR = pre-rehydration.

ª 2010 Adis Data Information BV. All rights reserved.

Sports Med 2010; 40 (4)

Intravenous versus Oral Rehydration

Dehydration ~180 min

335

Rest Rehydration 20 min 45 min

Rest 55 min

Equilibration 20 min

Performance test 90 min

Experiment outline PrD

PoD

PrD

PoD

PoR

PrE

PoE

Echegaray et al.[27] PrE 15 30 45 60 75 PoE

Fig. 3. Experiment outline and overview of the study by Echegaray et al. dration; PrD = pre-dehydration; PrE = pre-exercise.

Despite this, the rapid increase of fluid in the vasculature is only transient and appears to equilibrate within 35 minutes during passive rehydration[26] or 5–25 minutes when exercise begins soon after rehydration is completed.[16-18] This trend is similar for both hypotonic and isotonic IV fluid.[26] Thus, IV rehydration may offer an advantage when little time is available for rehydration. However, when periods longer than approximately 1 hour are available for fluid absorption and equilibration, few differences between IV and oral rehydration are observed. 2.1.2 Heart Rate

The initial increase in plasma volume with IV fluid given during exercise and in rehydration will assist in the maintenance of stroke volume and therefore cardiac output. This should allow a concomitant reduction in heart rate and attenuation of cardiovascular drift by preventing a displacement of blood from the central circulation to peripheral veins.[31] All three studies infusing IV fluid during exercise showed a reduction in heart rate,[19-21] although in one the reduction was only observed in a warm environment (30C, <30% relative humidity [RH]) and not in a cool environment (22C, <30% RH) [table II].[20] The data reported by Hamilton et al.[21] may also be confounded by the fact that the IV fluid contained glucose whereas the oral solution did not.

PoD = post-dehydration; PoE = post-exercise; PoR = post-rehy-

In total, seven articles have reported heart rate during exercise following IV rehydration (table II). One of these observed lower heart rates after 30–75 minutes of exercise following IV compared with oral fluid.[15] However, there was no difference between hypotonic and isotonic IV fluids[24] and no difference in maximal heart rate between conditions.[28] Additionally, there were no differences in plasma volume between IV and oral rehydration to account for the reduced heart rate.[15] The lower heart rate in the IV trials may have been related to sympathetic nervous activity because plasma noradrenaline level was increased in the oral condition.[15] It should be noted that this study employed a 45-minute rehydration period in which subjects were rehydrated with 50% of fluid lost, followed by 75 minutes rest before exercise. Interestingly, two other studies where exercise began immediately post-rehydration (figures 2 and 4), when the biggest difference between rehydration modes would be expected, reported no significant difference in heart rate during subsequent exercise.[17,18] 2.2 Thermoregulatory Parameters

The acute expansion of blood volume acts to restore central venous pressure and in conjunction with a lower heart rate, also allows a progressive increase in peripheral blood flow during Performance test 75 min

Rehydration 30 min

Dehydration 75 min

Experiment outline

[27]

PrD

PoD

PrE

PrD

PoD

PrE

PoE

Kenefick et al.[18] 25

38

PoE

Fig. 4. Experiment outline and overview of the study by Kenefick et al.[18] PoD = post-dehydration; PoE = post-exercise; PrD = predehydration; PrE = pre-exercise.

ª 2010 Adis Data Information BV. All rights reserved.

Sports Med 2010; 40 (4)

336

ª 2010 Adis Data Information BV. All rights reserved.

Table II. Cardiovascular and thermoregulatory effects during performance tests following intravenous (IV) rehydration Study

Treatments

PV

HR

SV

Sweat rate

Rectal temp

Skin temp

. VO2

[Lactate]

FBF

Subjective sensations

Performance

IV during exercise Deschamps et al.[19]

Isotonic IV No fluid

› with IV

fl with IV

NM

-

fl with IV

-

. - (VO2 at exhaustion)

-

NM

NM

-

Nose et al.[20]

Isotonic IV, warm environment No fluid, warm environment Isotonic IV, cool environment No fluid, cool environment

› with IV in both warm and cool environments

fl ~6 bpm between min 35 and 50 of exercise after IV in warm environment; - cool environment

NM

- in warm and cool environments

Oesophageal temp fl with IV infusion from 45 to 50 min in warm environment; - cool environment

NM

NM

NM

› with IV in both warm and cool conditions

NM

NM

Hamilton et al.[21] experiments 1 and II

Oral No fluid IV (18% glucose)

IV prevented the fl in BV seen during exercise with oral fluid

fl with IV

-

NM

NM

NM

fl with IV

NM

NM

NM

NM

IV during rehydration Hypotonic IV Oral No fluid

- (tended to be higher for initial 35 min of rehydration period in IV trial)

-

NM

NM

-

-

. (%VO2max)

-

NM

Central RPE fl in oral vs IV and no fluid Thirst sensation fl in oral vs IV and no fluid

NM

Castellani et al.[15]

Hypotonic IV Oral No fluid

-

During exercise, › in oral vs IV from 45 min onwards

NM

- (whole body and local chest)

-

-

. (%VO2max)

NM

NM

NM

- endurance time to exhaustion IV vs oral (oral tended to be longer than IV 77.4 vs 84.2 min)

Castellani et al.[24]

Isotonic IV Hypotonic IV No fluid

- between IV trials

- between IV trials

NM

NM

- between IV trials

NM

NM

NM

NM

NM

NM

Casa et al.,[17] Maresh et al.[16]

Hypotonic IV Oral No fluid

› restoration with IV over oral and no fluid

- IV vs oral (IV tended to be › )

SV or CO

- (IV vs oral)

fl with oral over IV and no fluid for 1st 12 min of exercise

fl with oral over IV and no fluid for 1st 12 min of exercise; fl in oral vs IV post-exercise

- (IV vs oral)

fl with oral vs IV and no fluid; fl with IV vs no fluid at min 15 of exercise

- (IV vs oral)

Central, local and overall RPE all › in no fluid vs IV and oral at min 5 and 15 of exercise Central RPE fl during oral vs IV at min 5 and 15 of exercise

- endurance time to exhaustion IV vs oral, but both › over no fluid (oral tended to be longer than IV; 34.9 vs 29.5 min; p = 0.07)

Continued next page

van Rosendal et al.

Sports Med 2010; 40 (4)

Riebe et al.[23]

Study

Treatments

PV

HR

SV

Sweat rate

Rectal temp

Skin temp

. VO2

[Lactate]

FBF

Subjective sensations

Performance

Overall RPE fl during oral vs IV at minute 15 of exercise Thirst ratings fl during oral vs IV and no fluid; and during IV vs no fluid at min 0, 5 and 15 Thermal sensation fl during oral and IV vs no fluid at min 5; and during IV vs oral and no fluid at min 15 Isotonic IV Hypotonic IV Oral No fluid

IV › more than oral for first 15 min

NM

NM

NM

NM

NM

NM

NM

NM

NM

NM

Echegaray et al.[27]

Isotonic IV Hypotonic IV No fluid

NM

- (between IV trials; only maximum reported)

NM

- (between IV trials)

- (between IV trials)

NM

NM (- RER)

(between IV trials)

NM

NM

- endurance time to exhaustion (between IV trials)

Kenefick et al.[18]

Hypotonic IV Oral No fluid

› restoration with IV; - IV vs oral and no fluid after 25 min of exercise

- (IV vs oral; both fl vs no fluid)

NM

- (IV vs oral)

fl following rehydration with IV vs oral and no fluid; - IV vs oral after 25 min of exercise

fl with IV and oral vs no fluid after 25 min of exercise

-

NM

NM

- RPE - thermal sensation IV vs oral; both fl vs no fluid up to 25 min of exercise fl thirst sensation oral vs IV; fl IV vs no fluid (oral < IV < no fluid)

- endurance time to exhaustion IV vs oral, but both › over no fluid (IV 72.6; oral 70.6; no fluid 38.7 min)

Kenefick et al.[28]

Isotonic IV Hypotonic IV Oral No fluid

- (between experimental conditions; all › vs no fluid)

- (between experimental conditions; all fl vs no fluid pre-exercise; only maximum reported)

NM

NM

- (between experimental conditions; all fl vs no fluid pre-exercise)

- (between experimental conditions; all fl vs no fluid pre-exercise)

NM

NM

NM

NM

- endurance time between experimental conditions, all › over no fluid (oral NS › vs IV: oral 84 min; hypotonic IV 77 min; isotonic IV 76 min)

BV = blood volume; CO = cardiac output; FBF = forearm blood flow; HR = heart rate; hypotonic IV = 0.45% NaCI; isotonic IV = 0.9% NaCI; [Lactate] = plasma lactate concentration; NaCI = sodium chloride; NM = not measured; NS = nonsignificant; oral = oral .fluid (0.45% NaCl); PV = plasma volume; RER = respiratory exchange ratio; RPE = rating of perceived . exertion; SV = stroke volume; temp = temperature; VO2 = oxygen uptake; VO2max = maximal oxygen uptake; › indicates increased; fl indicates decreased; - indicates not significantly different.

337

Sports Med 2010; 40 (4)

Kenefick et al.[26]

Intravenous versus Oral Rehydration

ª 2010 Adis Data Information BV. All rights reserved.

Table II. Contd

van Rosendal et al.

338

continuous body heating.[32-34] It could therefore be expected that acute plasma volume expansion with IV fluid may translate to thermoregulatory benefits. As the following sections discuss, this finding is yet to be consistently witnessed. 2.2.1 Core Temperature

Three trials reported over a number of articles have monitored core temperature during exercise following IV and oral rehydration (table II).[15-18,23,28] Of these, one showed reduced rectal temperature pre-exercise that persisted for the first 12 minutes of exercise in the oral fluid condition;[16,17] one showed reduced rectal temperature pre-exercise that had equilibrated after 25 minutes of exercise in the hypotonic IV fluid condition;[18] and one showed no difference between conditions.[15,23,28] Again, differences in study design may explain these disparate results. The first of these studies showing reduced temperature in the oral compared with IV trials had a fundamental difference between experimental conditions whereby the oral fluids were given at 10C while hypotonic IV fluid was infused at 22C. The 12C difference between these fluids would be sufficient to elicit the 0.6C difference witnessed between conditions.[17,18] In the second experiment, rectal temperature was 1C lower immediately after rapidly rehydrating with hypotonic IV (15C) compared with oral (15C) and no-fluid conditions.[18] The authors state that the large volume of fluid entering the vasculature at 15C may have reduced core temperature in the IV trials and they used equations by Kay and Marino[35] to predict a reduction of 0.7C. The additional 0.3C difference between conditions may have been a consequence of the more rapid restoration of plasma volume in the IV trials causing earlier re-establishment of skin blood flow and sweating responses, and a momentary enhancement of thermoregulation.[18] However, given that the bodyweight changes following rehydration were not different between conditions, when corrected for urine loss, it appears as though sweat rate was not higher during the rehydration period in the IV trials. The third study gave oral fluid at 4C and IV fluid at 25C and found no differences in core temperature between hypotonic IV, isotonic IV ª 2010 Adis Data Information BV. All rights reserved.

and oral fluids immediately before or during ensuing exercise.[15,23,28] However, these measurements were taken 120 minutes after the start of rehydration, allowing ample time for the equilibration of core temperature (figure 1). 2.2.2 Skin/Forearm Blood Flow and Skin Temperature

Forearm blood flow is reduced during exercise as a consequence of a reduction in central venous pressure or cardiac filling pressure.[36] These changes occur subsequent to a reduction in venous return due to blood pooling in the extremities or loss of blood volume accompanying dehydration.[32,37] By rapidly replenishing plasma volume and maintaining central blood pressure, IV fluid should theoretically increase skin blood flow (as evidenced by an in increased forearm blood flow). One could subsequently expect a reduction of heat storage by increasing heat transfer from the body core to the periphery and increasing the evaporation of sweat from the skin surface.[31] One study showed that forearm blood flow increased 16% from 30 to 105 minutes of exercise with IV infusion during exercise.[20] However, it also indicated that fluid replacement increases skin blood flow by a mechanism other than increasing blood volume because the use of blood volume expanders resulted in similar forearm blood flow during exercise to that with no fluid replacement at all.[31] This is further supported by the only study that has investigated forearm blood flow during exercise after rehydration with hypotonic IV and oral fluids.[16,17] Forearm blood flow was not increased more in the IV trials even at the timepoints where plasma volume was higher. In contrast, Nose et al.[20] found a direct relationship between plasma volume and forearm blood flow whereby the greater the increase in plasma volume, the greater the increase in forearm blood flow. It may be that the difference in plasma volume between rehydration techniques is not large enough to translate into benefits, especially after sufficient time has passed to allow equilibration of the fluids. The expected effect of IV fluid in reducing skin temperature is yet to be reported. When exercise began immediately following oral or hypotonic IV rehydration, Maresh et al.[16] and Casa Sports Med 2010; 40 (4)

Intravenous versus Oral Rehydration

et al.[17] showed a significant reduction in skin temperature during the first 12 minutes of exercise, which tended to stay lower throughout exercise following oral rehydration. That skin temperature was lower in the oral trials may be related to the fact that the fluids were administered at different temperatures (oral 10C, IV 25C) immediately before exercise. Other studies have found no difference in skin temperature following hypotonic IV compared with oral rehydration when fluids were given at the same temperature,[18] or when exercise began 120 minutes after the start of rehydration.[15,23] Additionally, Deschamps et al.[19] found no effect on skin temperature when isotonic IV fluid was infused during exercise. 2.2.3 Sweat Rate

Two studies found no change in sweat rate with IV saline infusion during exercise.[19,20] A further three found no difference in sweat rate or change in bodyweight during exercise after rehydration with IV or oral fluid.[15,17,18] These results are most likely a consequence of the relatively short exercise durations in one study[17] and the fact that fluid compartments equilibrate relatively quickly so that only small differences exist between different rehydration techniques when similar fluid volumes are given. 2.3 Endocrine Parameters

The disturbance to fluid homeostasis associated with dehydration and rehydration causes dramatic changes to hormones that control body fluid volume regulation. The physiological stress associated with these changes, and the physical stress of intense exercise itself, also cause significant changes to stress hormones such as cortisol. A number of hormones associated with fluid retention, physiological stress and preparation for physical activity have therefore been investigated when rehydrating via these different modes, and during subsequent exercise (table III). 2.3.1 Fluid-Regulating Hormones: Antidiuretic Hormone and Aldosterone

Reduced plasma osmolality,[38,39] oropharyngeal stimulation[40] and blood volume[41] are ª 2010 Adis Data Information BV. All rights reserved.

339

known to be the primary stimuli for ADH release. These factors are influenced by the mode of rehydration and composition of fluids. The act of drinking has important implications because it quickly eliminates thirst[42,43] and prevents ADH secretion,[43,44] even before changes to plasma volume and osmolality are noticed.[43] Indeed, Kenefick et al.[26] report that, while not statistically significant, ADH concentration decreased more rapidly and remained lower following oral rehydration compared with both hypotonic and isotonic IV conditions. These changes occurred despite the fact that plasma volume was initially restored more rapidly in the IV trials.[26] Kenefick et al.[26] also measured aldosterone at the same timepoints as ADH in the abovementioned study. Aldosterone tended to be higher in the oral trial than in both the IV trials. Plasma volume increased significantly more in both the IV trials compared with the oral trial over the first 15 minutes.[26] These findings are in agreement with Nose et al.,[28] who showed that changes in blood volume are a major influence on aldosterone secretion.[45] As reported in their later article, there were no differences in plasma volume between the treatments after 75 minutes, and subsequently, there were no differences in aldosterone after this longer duration. 2.3.2 Stress Hormones: Adrenocorticotropic Hormone and Cortisol

Both adrenocorticotropic hormone (ACTH), a precursor to cortisol release, and cortisol itself increase in response to stress. Two studies have investigated these hormones in response to oral and IV (isotonic and hypotonic) rehydration.[15,24,25] In the first, exercise commenced 2 hours after rehydration began (figure 1).[15,24] ACTH and cortisol levels both dropped to a greater extent and more rapidly following rehydration with hypotonic IV compared with oral fluid.[15] This indicates a more rapid decline in physiological stress when rehydrating with IV fluid. However, after 15 minutes of exercise, the level of these hormones had dropped in the oral trials so they were similar to the IV trials and they remained similar after 45 minutes of exercise.[15] Interestingly, there was also a difference when Sports Med 2010; 40 (4)

van Rosendal et al.

340

Table III. Studies investigating the influence of fluid-regulating and stress hormones during rehydration with intravenous (IV) fluid Study

Treatments

Hormone

Findings

Castellani et al.[15]

Hypotonic IV Oral No fluid

Noradrenaline

› less during exercise with IV and oral vs no fluid; › less with IV vs oral after 45 min of exercise NSD

Castellani et al.[24]

Isotonic IV Hypotonic IV No fluid

Adrenaline ACTH

D in plasma ACTH was fl in IV vs oral post-rehydration (ACTH fl more in IV vs oral); NSD at 15 and 45 min of exercise

Cortisol

D in plasma cortisol was fl in IV vs oral post-rehydration (cortisol fl more in IV vs oral); NSD at 15 and 45 min of exercise

Noradrenaline

NSD between isotonic and hypotonic IV › less during exercise with both IV conditions compared with no fluid D in plasma ACTH was fl in hypotonic vs isotonic IV and no fluid (ACTH fl more in hypotonic vs isotonic and no fluid) at min 0 and 15; NSD at 45 min of exercise

ACTH Cortisol

D in plasma cortisol was fl in hypotonic vs isotonic IV and no fluid at min 0, 15 and 45 Isotonic IV was fl vs no fluid at min 0 and 15 (cortisol fl more in hypotonic IV vs isotonic IV vs no fluid)

Casa et al.[25]

Hypotonic IV Oral No fluid

Noradrenaline ACTH Cortisol

NSD IV vs oral NSD IV vs oral NSD IV vs oral

Kenefick et al.[26]

Isotonic IV Hypotonic IV Oral No fluid

ADH

NSD (DADH fell (although ADH fl to a greater extent [not significant] following oral rehydration over either isotonic or hypotonic IV in first 15 min despite a lower % DPV. May be due to stimulation of oropharyngeal reflex) NSD (although D aldosterone tended to fall less [ › aldosterone] in oral compared with either isotonic or hypotonic IV)

Echegaray et al.[27]

Isotonic IV Hypotonic IV No fluid

Noradrenaline Adrenaline

NSD between IV trials NSD between IV trials

Kenefick et al.[28]

Isotonic IV Hypotonic IV Oral No fluid

ADH Aldosterone

NSD between any treatments NSD between any treatments

Aldosterone

% DPV = percentage change in plasma volume; ACTH = adrenocorticotropic hormone; ADH = antidiuretic hormone (arginine vasopressin); hypotonic IV = 0.45% NaCI; isotonic IV = 0.9% NaCI; NaCI = sodium chloride; NSD = no significant difference between treatments; oral = oral fluid (0.45% NaCl); › indicates increased; fl indicates decreased; D indicates change.

comparing hypotonic and isotonic IV solutions. Both ACTH and cortisol were lower following IV rehydration with hypotonic versus isotonic saline, with the difference persisting throughout the 45-minute exercise period.[24] In contrast, when exercise began immediately after rehydration, Casa et al.[25] showed no differences in ACTH and cortisol between hypotonic IV and oral solutions. However, this article only reported the hormones pre- and postexercise to exhaustion.[25] Although differences pre-exercise may have been expected, it is not surprising that ACTH and cortisol levels would be similar at the point of exhaustion, irrespective of rehydration condition. ª 2010 Adis Data Information BV. All rights reserved.

2.3.3 Catecholamines: Noradrenaline and Adrenaline

Similar to the response seen for stress hormones, the mode of rehydration has been shown to significantly affect noradrenaline level. Noradrenaline was higher following 45 minutes of exercise in the oral compared with hypotonic IV trial when exercise began 120 minutes after the start of rehydration,[15] but not when exercise began immediately after rehydration.[25] The first finding indicates greater sympathetic nervous system activity following oral rehydration, which may have contributed to the higher heart rate in the oral condition in that study, as previously mentioned.[15] Again, the second study may have missed transient Sports Med 2010; 40 (4)

Intravenous versus Oral Rehydration

changes to hormone levels since measurements were only taken post-dehydration, and pre- and post-exercise (figure 2).[17] It also appears that changing plasma volume had a bigger influence on noradrenaline release than altering plasma sodium concentration. Noradrenaline was significantly higher in the no-fluid trial when plasma volume was reduced, than in either of the rehydration conditions.[15,24] However, when plasma volume was restored to a similar level, but with altered plasma sodium concentration (hypotonic vs isotonic IV rehydration), no differences were reported for noradrenaline.[24,27] Furthermore, there were no significant differences for adrenaline level following rehydration or during exercise after rehydration with hypotonic IV or oral fluid,[15] or between hypotonic and isotonic IV fluid.[27] 2.4 Subjective Parameters

The subjective response to exercise influences the intensity that an individual is able to maintain. When sensations of thirst, thermal stress and perceived exertion are lower, an individual is typically able to maintain higher relative exercise intensities without feeling as if they are working harder. These parameters are influenced by both the level of dehydration and mode of rehydration, as discussed in the following sections. 2.4.1 Thirst Sensation

It has recently been proposed that thirst sensation may act in a feed-forward manner to involuntarily slow the athlete and prevent dehydration from becoming physiologically significant.[1] When exercising, it has therefore been suggested that drinking ad libitum according to thirst is optimal, while performance will be degraded if no fluid is consumed during prolonged aerobic exercise in the heat.[46] It is known that ‘voluntary dehydration’ normally occurs with exercise heat stress and has the potential to degrade performance and recovery if little attempt at fluid replacement is made.[47-49] Thirst sensation may therefore be an important regulatory factor influencing the level to which an individual voluntarily dehydrates during exercise.[1] Furthermore, thirst not only ª 2010 Adis Data Information BV. All rights reserved.

341

affects the subjective response to exercise, but also influences factors such as the release of fluidregulating hormones. Thirst sensation remains consistently higher following IV compared with oral rehydration because of the lack of stimulation of oropharyngeal and gastric mechanisms.[16,18,23] Maresh et al.[16] found that thirst sensation was significantly correlated with overall rating of perceived exertion (RPE) in both the no-fluid and IV trials and with central RPE in the oral trials. Thus, thirst may influence RPE and contribute to exercise feeling more strenuous following IV rehydration for a given intensity. Because an individual maintains the sensation of thirst following IV rehydration, they must be cautious not to over-drink water on top of the IV fluid they have received, otherwise they may increase the risk of hyponatraemia.[13,50] It should also be noted here that some athletes are unable to dictate rehydration based on thirst because they are unable to stop the practice or leave the game to drink. Such athletes must anticipate fluid losses and rehydrate accordingly rather than relying solely on thirst. 2.4.2 Thermal Sensation

When exercise began immediately after rehydration, thermal sensation was lower in the IV trial than both the oral and no-fluid trials.[16] This is most likely due to the peripheral displacement of the large volume of IV fluid that was administered. After 15 minutes of exercise, thermal sensation was lower in the oral trial than either of the IV (hypotonic and isotonic) or the nofluid trials.[16] Both rectal and skin temperatures were also lower in the oral trial at this timepoint, indicating an overall reduction to thermal strain. Contrasting results were found by Kenefick et al.,[18] who also exercised subjects immediately after rehydration but found no effect on thermal sensation when exercising in similar environmental conditions. This was despite the fact that rectal temperature was 1C lower in the IV condition at the start of exercise. The fact that skin temperature was similar indicates that it may have a greater influence on thermal sensation than does rectal temperature. Maresh et al.[16] Sports Med 2010; 40 (4)

van Rosendal et al.

342

reported a strong correlation between thermal sensations and skin temperature (r = 0.83) after oral rehydration, although Kenefick et al.[18] reported a weak correlation (r = 0.33) between the same variables. Additionally, thermal sensations were highly correlated in all three experimental conditions with local, central and overall RPE indicating that increased thermal strain detrimentally affects exercise performance.[16] 2.4.3 Rating of Perceived Exertion

Three studies have monitored RPE during exercise after IV or oral rehydration (tables I and II). In two of these, overall and central RPE were lower at various timepoints during exercise following oral rehydration than after IV rehydration.[16,23] Thirst sensation was also generally lower following oral rehydration at the same timepoints.[16,23] Interestingly, each of these trials only half rehydrated the subjects. In the other study, the subjects were rehydrated with 100% of the fluid lost during dehydration and even though thirst sensation was elevated in the IV trials, there was no increase in RPE.[18] 3. Exercise Performance Following Intravenous versus Oral Rehydration Several investigations have studied exercise performance after rehydrating with IV or oral fluid (tables I and II).[15-17,19,25] One of these used an impractical study design (figure 2) that included a short-duration (~30 minutes) exercise test to exhaustion (a protocol that has been associated with poor reliability and a coefficient of variation in individuals of up to 27%[51-53]).[17] The trial showed a nonsignificant trend (p = 0.07) to exercise longer following oral rehydration (34.9 minutes) than IV infusion (29.5 minutes). However, because of the substantial thirst associated with being dehydrated to -4% bodyweight overnight, exercising the following day, and having no oral fluid intake, it is not surprising that exercise duration was shorter in the IV trials. Each of the other studies used protocols where an exercise duration target was set.[15,18,27,28] Comparing the mean exercise duration between ª 2010 Adis Data Information BV. All rights reserved.

conditions is therefore confounded by the fact that subjects were stopped once they reached the target duration as opposed to reaching exhaustion. It is therefore erroneous to assign meaning to differences in exercise duration under these conditions. Future research should include performance tests such as time trials or time to complete a certain amount of work, as these are considered more reliable indicators of performance.[51] Lactate response during exercise following rehydration with IV fluid has also been investigated in three studies.[17,23,27] One found no difference between hypotonic and isotonic IV trials,[27] while another found no difference between hypotonic IV and oral trials.[23] However, Casa et al.[17] found reduced lactate levels during and after exercise (although difference post-exercise was not significant) after oral compared with hypotonic IV rehydration. Higher lactate levels in the IV trials may indicate an earlier switch to anaerobic glycolysis or may have resulted from reduced oxygen perfusing the muscle due to greater cardiorespiratory strain (seen through differences in expired ventilation and respiratory rate) or from muscle dynamics being affected by the lower core and skin temperatures in the oral trial.[17] 4. Effects of Intravenous Fluids on Recovery from Exercise To date only one study has investigated whether IV fluid aids recovery. Polak et al.[22] gave athletes 2.5 L of IV fluid (2.5% glucose in 0.45% NaCl) or a sham infusion (100 mL of IV fluid) following the Rotterdam marathon. Athletes who received 2.5 L tended to have longer recovery periods, longer-lasting muscle pain, increased pain levels, increased stiffness and increased fatigue or tiredness on at least 1 of the 10 recovery days, compared with those who received the sham infusion.[22] However, the athletes receiving IV fluid also had a faster average completion time and were more likely to have completed a personal best time. They were also required to sit stationary for a much longer period after the race while the fluid was administered. It is thus impossible to distinguish whether these effects on recovery were a result of the rehydration Sports Med 2010; 40 (4)

Intravenous versus Oral Rehydration

treatment or the fact that the subjects who received IV fluid had induced greater muscle damage during the performance, or greater stiffness due to immobilization.[22] 5. Areas for Future Research The small number of studies specifically investigating IV rehydration in athletes has left several important questions unanswered. Further research is required before definitive guidelines can be provided to athletes. The following points have a substantial practical application. 5.1 Volume of Fluid Replenished

To date, the majority of articles reporting IV rehydration are from investigations that have dehydrated subjects by 4% bodyweight, before rehydrating with half of the fluid lost (table I).[15-17,23-26,28] Such a scenario is likely to be practical when large athletes are losing large volumes of fluid, with limited time available for rehydration (such as in the previously used example of American football, where the athletes are very large and fluid losses over the first two quarters are often substantial; it would not be practical or likely that these large fluid losses could be replaced with IV rehydration over a half-time break). Thus, while full fluid and electrolyte replacement should be a goal,[54] in many real-world scenarios it may not be practical. However, many other athletes do attempt to completely rehydrate in between exercise bouts. To date, one study has rehydrated subjects with 100% of the fluid lost during dehydration.[18] Because sweating and insensible water loss remain elevated for a period after exercise has ceased, it is generally recommended that 150% of fluid deficits be replaced.[55] Thus, to assess the effects of IV rehydration in those attempting to completely rehydrate, volumes of this size should be investigated in future research. 5.2 Effects of Using Partial Intravenous and Partial Oral Rehydration

Unless they are severely nauseated, vomiting or unconscious, it is very unlikely that any athlete ª 2010 Adis Data Information BV. All rights reserved.

343

would undertake IV rehydration without drinking any fluid. IV rehydration appears to more rapidly replenish plasma volume and may therefore provide a hydration benefit if subsequent exercise commences soon after the preceding dehydrating exercise bout. However, IV fluid has the major limitation of providing no oropharyngeal stimulation.[9] Thus, athletes still feel thirsty after IV rehydration, resulting in numerous physiological and performance implications. It may be that a combination of IV and oral rehydration could provide an optimal fluid intake protocol. This could be especially true in situations where athletes dehydrate by >4% bodyweight. A trained, heat-acclimated athlete can sweat in excess of 2 L/h.[7] When considering that 150–200% of a fluid deficit must be consumed to fully restore fluid balance,[7] an athlete may have to consume 3–4 L of fluid for each 2 hours of exercise. Although this volume can be reduced to 125–150% when rehydrating between exercise sessions, trying to replenish fluid volumes of this magnitude in a short time via oral ingestion only can be challenging.[7] For athletes completing multiple training sessions in a single day, IV rehydration may be a more practical method for fluid replenishment.[7] 5.3 Effect of Intravenous Fluid on Acid-Base Physiology

Post-infusion acidosis is a well known adverse effect when administering large volumes of IV saline (e.g. during medical procedures).[56,57] Several approaches to acid-base physiology have been used to describe this condition. The Henderson-Hasselbalch approach indicates that IV saline infusion may cause acidosis in two stages: (i) dilution of both plasma carbon dioxide and bicarbonate due to infused fluid; and (ii) ongoing carbon dioxide production will increase plasma carbon dioxide while bicarbonate remains lower in the short term.[57] The Stewart approach indicates that saline infusion will cause a fall in the strong ion difference resulting in a drop in pH.[57] It has been proposed that including weak acids in the infused saline (e.g. lactate, acetate and gluconate) may reduce the acidosis because these are converted to bicarbonate in the Sports Med 2010; 40 (4)

van Rosendal et al.

344

liver.[58] Components of acid-base balance (such as plasma levels of lactate, bicarbonate, carbon dioxide and electrolytes) are altered in the dehydrated athlete, which will influence the potential to induce acidosis with saline infusion. To our knowledge, no studies have yet investigated acidosis when repeatedly administering IV fluids during rehydration in dehydrated athletes. Therefore, this is an area that requires further research. 5.4 Effect of Intravenous Fluid on Respiratory Function

A final proposed area of future research of high importance for athletes is the effect of IV fluid on respiratory function. Several authors have shown reduced respiratory function after rapid infusion of IV saline in euhydrated individuals,[59,60] possibly resulting from airway mucosal oedema.[60] Furthermore, during subsequent exercise, maximal oxygen uptake, anaerobic threshold, maximal heart rate and maximal power output were all reduced.[59] These reductions in maximal exercise performance appeared unrelated to ventilation limitations.[59] Robertson et al.[59] proposed that peripheral oedema in the exercising muscle may have either increased intramuscular pressure causing a reduction to the perfusion of the muscle bed, or acted to increase the diffusion distance between capillaries and exercising muscle, thereby reducing the oxygen extraction capabilities of the muscle to an extent that was physiologically relevant. Bicarbonate was reduced to a greater extent than haemoglobin, indicating that more fluid went into the extravascular space, which would increase the diffusion distance between the capillaries and myocytes.[59] These factors will be less likely to occur when rehydrating with IV fluid from a state of dehydration, but need to be investigated none the less. 6. Adverse Effects from Intravenous Rehydration None of the studies rehydrating with a volume of IV fluid £100% of the bodyweight lost during dehydration have reported any adverse effects. In the situation of small volumes of isotonic or hypotonic IV fluid being provided to dehydrated individuals, no adverse effects are expected. ª 2010 Adis Data Information BV. All rights reserved.

However, a major cause of hyponatraemia in athletes is over-hydration with hypotonic fluids during exercise (i.e. drinking water and losing sodium in the sweat).[61-63] In fact, when so much fluid is consumed that weight gain during exercise exceeds 2.5% bodyweight, there is a very dangerous negative linear relationship between bodyweight and serum sodium level.[50,64] Should too much IV fluid be administered to such an athlete, one potentially serious adverse effect is the development of iatrogenic hyponatraemia.[8] Thus, care needs to be taken if any attempt is to be made at administering large volumes of IV fluid, especially to individuals who have maintained hydration during exercise by ingesting water without adequate electrolyte replacement. 7. Conclusion The issue of using IV fluid to rehydrate athletes who are dehydrated and otherwise healthy is a contentious one. The current research specifically comparing IV and oral rehydration suggests that IV rehydration may provide a means of more rapidly replenishing the body’s fluid stores compared with oral rehydration. Despite this, studies investigating IV versus oral rehydration have shown that benefits appear to be transient and typically fail to translate into meaningful cardiovascular, thermoregulatory or performance benefits during subsequent exercise. One factor that may have contributed to the paucity of definitive findings is the vast difference in the design of studies investigating these variables. Furthermore, no studies have actually used a protocol that rehydrates athletes according to the American College of Sports Medicine’s general guidelines of 150% of fluid lost during dehydration.[54] A significant amount of future research in the field is therefore required before any recommendations regarding the voluntary use of IV fluids in athletes after exercise can be made. The present literature suggests that IV fluids should be administered to athletes who are experiencing severe dehydration only after a medical practitioner has ascertained that dehydration and not hyponatraemia is causing the symptoms. Especially in view of the fact that WADA has banned Sports Med 2010; 40 (4)

Intravenous versus Oral Rehydration

IV fluid use unless medically warranted, all other athletes are advised to continue rehydrating with oral fluids based on current evidence.[9] Acknowledgements The authors acknowledge the University of Queensland Graduate School Research Travel Grant for financial assistance. The authors have no conflicts of interest that are directly relevant to the contents of this review.

References 1. Sawka MN, Noakes TD. Does dehydration impair exercise performance? Med Sci Sports Exerc 2007 Aug; 39 (8): 1209-17 2. Cheuvront SN, Carter 3rd R, Castellani JW, et al. Hypohydration impairs endurance exercise performance in temperate but not cold air. J Appl Physiol 2005 Nov; 99 (5): 1972-6 3. Judelson DA, Maresh CM, Anderson JM, et al. Hydration and muscular performance: does fluid balance affect strength, power and high-intensity endurance? Sports Med 2007; 37 (10): 907-21 4. Cheuvront SN, Carter 3rd R, Sawka MN. Fluid balance and endurance exercise performance. Curr Sports Med Rep 2003 Aug; 2 (4): 202-8 5. Sharwood K, Collins M, Goedecke J, et al. Weight changes, sodium levels, and performance in the South African ironman triathlon. Clin J Sport Med 2002 Nov; 12 (6): 391-9 6. Sharwood KA, Collins M, Goedecke JH, et al. Weight changes, medical complications, and performance during an ironman triathlon. Br J Sports Med 2004 Dec; 38 (6): 718-24 7. Shirreffs SM, Armstrong LE, Cheuvront SN. Fluid and electrolyte needs for preparation and recovery from training and competition. J Sports Sci 2004 Jan; 22 (1): 57-63 8. Noakes TD. Perpetuating ignorance: intravenous fluid therapy in sport. Br J Sports Med 1999 Oct; 33 (5): 296-7 9. Casa D, Ganio MS, Lopez RM, et al. Intravenous versus oral rehydration: physiological, performance, and legal considerations. Curr Sports Med Rep 2008; 7 (4): S41-9 10. WADA. The World Anti-Doping Code: the 2007 Prohibited List-International Standard. 2007 [online]. Available from URL: http://www.wada-ama.org/en/prohibitedlist.ch2 [Accessed 2009 May 20] 11. Speedy DB, Noakes TD, Holtzhausen LM. Exerciseassociated collapse: postural hypotension, or something deadlier? Phys Sportsmed 2003; 31 (3): 23-9 12. Holtzhausen LM, Noakes TD. Collapsed ultraendurance athlete: proposed mechanisms and an approach to management. Clin J Sport Med 1997 Oct; 7 (4): 292-301 13. Noakes T. Hyponatremia in distance athletes: pulling the IV on the ‘dehydration myth’. Phys Sportsmed 2000; 28 (9): 71-6 14. Pyne S. Intravenous fluids post marathon: when and why? Sports Med 2007; 37 (4-5): 434-6 15. Castellani JW, Maresh CM, Armstrong LE, et al. Intravenous vs. oral rehydration: effects on subsequent exercise-heat stress. J Appl Physiol 1997 Mar; 82 (3): 799-806

ª 2010 Adis Data Information BV. All rights reserved.

345

16. Maresh CM, Herrera-Soto JA, Armstrong LE, et al. Perceptual responses in the heat after brief intravenous versus oral rehydration. Med Sci Sports Exerc 2001 Jun; 33 (6): 1039-45 17. Casa DJ, Maresh CM, Armstrong LE, et al. Intravenous versus oral rehydration during a brief period: responses to subsequent exercise in the heat. Med Sci Sports Exerc 2000 Jan; 32 (1): 124-33 18. Kenefick RW, O’Moore KM, Mahood NV, et al. Rapid IV versus oral rehydration: responses to subsequent exercise heat stress. Med Sci Sports Exerc 2006 Dec; 38 (12): 2125-31 19. Deschamps A, Levy RD, Cosio MG, et al. Effect of saline infusion on body temperature and endurance during heavy exercise. J Appl Physiol 1989 Jun; 66 (6): 2799-804 20. Nose H, Mack GW, Shi XR, et al. Effect of saline infusion during exercise on thermal and circulatory regulations. J Appl Physiol 1990 Aug; 69 (2): 609-16 21. Hamilton MT, Gonzalez-Alonso J, Montain SJ, et al. Fluid replacement and glucose infusion during exercise prevent cardiovascular drift. J Appl Physiol 1991 Sep; 71 (3): 871-7 22. Polak AA, van Linge B, Rutten FL, et al. Effect of intravenous fluid administration on recovery after running a marathon. Br J Sports Med 1993 Sep; 27 (3): 205-8 23. Riebe D, Maresh CM, Armstrong LE, et al. Effects of oral and intravenous rehydration on ratings of perceived exertion and thirst. Med Sci Sports Exerc 1997 Jan; 29 (1): 117-24 24. Castellani JW, Maresh CM, Armstrong LE, et al. Endocrine responses during exercise-heat stress: effects of prior isotonic and hypotonic intravenous rehydration. Eur J Appl Physiol Occup Physiol 1998 Feb; 77 (3): 242-8 25. Casa DJ, Maresh CM, Armstrong LE, et al. Intravenous versus oral rehydration during a brief period: stress hormone responses to subsequent exhaustive exercise in the heat. Int J Sport Nutr Exerc Metab 2000 Dec; 10 (4): 361-74 26. Kenefick RW, Maresh CM, Armstrong LE, et al. Plasma vasopressin and aldosterone responses to oral and intravenous saline rehydration. J Appl Physiol 2000 Dec; 89 (6): 2117-22 27. Echegaray M, Armstrong LE, Maresh CM, et al. Blood glucose responses to carbohydrate feeding prior to exercise in the heat: effects of hypohydration and rehydration. Int J Sport Nutr Exerc Metab 2001 Mar; 11 (1): 72-83 28. Kenefick RW, Maresh CM, Armstrong LE, et al. Rehydration with fluid of varying tonicities: effects on fluid regulatory hormones and exercise performance in the heat. J Appl Physiol 2007 May; 102 (5): 1899-905 29. Figaro MK, Mack GW. Regulation of fluid intake in dehydrated humans: role of oropharyngeal stimulation. Am J Physiol 1997 Jun; 272 (6 Pt 2): R1740-6 30. Kamijo Y, Okumoto T, Takeno Y, et al. Transient cutaneous vasodilatation and hypotension after drinking in dehydrated and exercising men. J Physiol 2005 Oct 15; 568 (Pt 2): 689-98 31. Montain SJ, Coyle EF. Fluid ingestion during exercise increases skin blood flow independent of increases in blood volume. J Appl Physiol 1992 Sep; 73 (3): 903-10

Sports Med 2010; 40 (4)

346

32. Fortney SM, Nadel ER, Wenger CB, et al. Effect of acute alterations of blood volume on circulatory performance in humans. J Appl Physiol 1981 Feb; 50 (2): 292-8 33. Ekblom B, Wilson G, Astrand PO. Central circulation during exercise after venesection and reinfusion of red blood cells. J Appl Physiol 1976 Mar; 40 (3): 379-83 34. Williams MH, Goodwin AR, Perkins R, et al. Effect of blood reinjection upon endurance capacity and heart rate. Med Sci Sports 1973 Fall; 5 (3): 181-6 35. Kay D, Marino FE. Fluid ingestion and exercise hyperthermia: implications for performance, thermoregulation, metabolism and the development of fatigue. J Sports Sci 2000 Feb; 18 (2): 71-82 36. Mack G, Nose H, Nadel ER. Role of cardiopulmonary baroreflexes during dynamic exercise. J Appl Physiol 1988 Oct; 65 (4): 1827-32 37. Nadel ER, Fortney SM, Wenger CB. Effect of hydration state of circulatory and thermal regulations. J Appl Physiol 1980 Oct; 49 (4): 715-21 38. Moses AM, Miller M, Streeten DH. Quantitative influence of blood volume expansion on the osmotic threshold for vasopressin release. J Clin Endocrinol Metab 1967 May; 27 (5): 655-62 39. Moses AM, Miller M. Osmotic threshold for vasopressin release as determined by saline infusion and by dehydration. Neuroendocrinology 1971; 7 (4): 219-26 40. Seckl JR, Williams TD, Lightman SL. Oral hypertonic saline causes transient fall of vasopressin in humans. Am J Physiol 1986 Aug; 251 (2 Pt 2): R214-7 41. Robertson GL, Athar S. The interaction of blood osmolality and blood volume in regulating plasma vasopressin in man. J Clin Endocrinol Metab 1976 Apr; 42 (4): 613-20 42. Rolls BJ, Wood RJ, Rolls ET, et al. Thirst following water deprivation in humans. Am J Physiol 1980 Nov; 239 (5): R476-82 43. Thompson CJ, Burd JM, Baylis PH. Acute suppression of plasma vasopressin and thirst after drinking in hypernatremic humans. Am J Physiol 1987 Jun; 252 (6 Pt 2): R1138-42 44. Geelen G, Keil LC, Kravik SE, et al. Inhibition of plasma vasopressin after drinking in dehydrated humans. Am J Physiol 1984 Dec; 247 (6 Pt 2): R968-71 45. Nose H, Mack GW, Shi XR, et al. Involvement of sodium retention hormones during rehydration in humans. J Appl Physiol 1988 Jul; 65 (1): 332-6 46. Noakes TD. Drinking guidelines for exercise: what evidence is there that athletes should drink ‘as much as tolerable’, ‘to replace the weight lost during exercise’ or ‘ad libitum’? J Sports Sci 2007 May; 25 (7): 781-96 47. Rivera-Brown AM, Ramirez-Marrero FA, Wilk B, et al. Voluntary drinking and hydration in trained, heat-acclimatized girls exercising in a hot and humid climate. Eur J Appl Physiol 2008 May; 103 (1): 109-16 48. Zetou E, Giatsis G, Mountaki F, et al. Body weight changes and voluntary fluid intakes of beach volleyball players during an official tournament. J Sci Med Sport 2008 Apr; 11 (2): 139-45

ª 2010 Adis Data Information BV. All rights reserved.

van Rosendal et al.

49. Noakes TD. Fluid replacement during exercise. Exerc Sport Sci Rev 1993; 21: 297-330 50. Noakes T. Hyponatremia in distance runners: fluid and sodium balance during exercise. Curr Sports Med Rep 2002 Aug; 1 (4): 197-207 51. Jeukendrup A, Saris WH, Brouns F, et al. A new validated endurance performance test. Med Sci Sports Exerc 1996 Feb; 28 (2): 266-70 52. Palmer GS, Dennis SC, Noakes TD, et al. Assessment of the reproducibility of performance testing on an air-braked cycle ergometer. Int J Sports Med 1996 May; 17 (4): 293-8 53. McLellan TM, Cheung SS, Jacobs I. Variability of time to exhaustion during submaximal exercise. Can J Appl Physiol 1995 Mar; 20 (1): 39-51 54. Sawka MN, Burke LM, Eichner ER, et al. American College of Sports Medicine position stand: exercise and fluid replacement. Med Sci Sports Exerc 2007 Feb; 39 (2): 377-90 55. Shirreffs SM, Taylor AJ, Leiper JB, et al. Post-exercise rehydration in man: effects of volume consumed and drink sodium content. Med Sci Sports Exerc 1996 Oct; 28 (10): 1260-71 56. Story DA. Hyperchloraemic acidosis: another misnomer? Crit Care Resusc 2004 Sep; 6 (3): 188-92 57. Story DA. Intravenous fluid administration and controversies in acid-base. Crit Care Resusc 1999 Jun; 1 (2): 156 58. Kirkendol PL, Starrs J, Gonzalez FM. The effects of acetate, lactate, succinate and gluconate on plasma pH and electrolytes in dogs. Trans Am Soc Artific Intern Organs 1980; 26: 323-7 59. Robertson HT, Pellegrino R, Pini D, et al. Exercise response after rapid intravenous infusion of saline in healthy humans. J Appl Physiol 2004 Aug; 97 (2): 697-703 60. Pellegrino R, Dellaca R, Macklem PT, et al. Effects of rapid saline infusion on lung mechanics and airway responsiveness in humans. J Appl Physiol 2003 Aug; 95 (2): 728-34 61. Speedy DB, Faris JG, Hamlin M, et al. Hyponatremia and weight changes in an ultradistance triathlon. Clin J Sport Med 1997 Jul; 7 (3): 180-4 62. Speedy DB, Noakes TD, Rogers IR, et al. Hyponatremia in ultradistance triathletes. Med Sci Sports Exerc 1999 Jun; 31 (6): 809-15 63. Noakes TD, Sharwood K, Speedy D, et al. Three independent biological mechanisms cause exercise-associated hyponatremia: evidence from 2,135 weighed competitive athletic performances. Proc Natl Acad Sci U S A 2005 Dec 20; 102 (51): 18550-5 64. Irving RA, Noakes TD, Buck R, et al. Evaluation of renal function and fluid homeostasis during recovery from exercise-induced hyponatremia. J Appl Physiol 1991 Jan; 70 (1): 342-8

Correspondence: Mr Simon Piet van Rosendal, School of Human Movement Studies, The University of Queensland, St Lucia Campus, Brisbane, QLD, 4072, Australia. E-mail: [email protected]

Sports Med 2010; 40 (4)

Sports Med 2010; 40 (4): 347-360 0112-1642/10/0004-0347/$49.95/0

REVIEW ARTICLE

ª 2010 Adis Data Information BV. All rights reserved.

Match Analysis and the Physiological Demands of Australian Football Adrian J. Gray and David G. Jenkins School of Human Movement Studies, The University of Queensland, Brisbane, Queensland, Australia

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Description of Australian Football. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Game Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Early Time-Motion Analysis Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Changes in Game Speed and Structure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Modern Australian Football. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Global Positioning Systems Match Analysis and Game Demands . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Heart Rate Responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Physiological and Anthropometric Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Mass, Stature and Body Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Maximal Oxygen Uptake. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Muscular Strength and Power. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4 Speed and Repeated Sprint Ability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Practical Applications for Conditioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

347 348 350 350 351 353 353 355 356 356 356 356 358 358 359

Australian Football, the most popular football code in Australia, is a contact sport played by two teams of 18 players who contest play over four 20-minute quarters; the object of the game is to score the most points through goal kicking. Sixteen professional senior sides compete against each other in the Australian Football League (AFL) and, similar to other football codes, game demands at the elite level in the AFL have changed considerably in recent years. Early time-motion analysis studies highlighted the long periods of time players spent in low intensity activities (standing and walking). While recent studies utilizing global positioning systems (GPS) technology are somewhat in agreement with earlier findings, available evidence suggests that the game is getting faster. For example, ‘playing on’ after a mark (a feature of the game where players who catch the ball on the full from a kick longer than 15 m are awarded a free kick) is now much quicker. Indeed, rule changes in recent years have increased the flow and speed of the game; there has been a reduction in the time taken for umpires to restart play, and for players to kick-in (after the opposition kicks a behind) or take a set shot at goal. Nomadic players (a broad term for midfielders and ruckmen because they follow play over the entire playing field) cover slightly greater distances (12 310 m) than both forwards (11 920 m) and backs (11 880 m) in a game. Compared with players in other positions, midfielders are consistently found to

Gray & Jenkins

348

spend the most time at higher intensities (running and sprint efforts with movement velocities >4.44 m/sec), complete more high intensity efforts (~98 per game), sustain them for longer and have shorter recovery periods between high intensity exercise bouts (~90 seconds on average). ‘Ruckmen’ have similar but less intense running profiles, while forwards and backs generally have less game involvement but have a more intermittent running profile (longer recovery periods with shorter duration high intensity exercise bouts and less time spent in constant pace running). Endurance fitness remains very important for players at the elite level of competition, as does upper and lower body strength and power. In addition, given the increasing speed at which Australian Football is now played, repeated sprint ability of players is arguably more important now than it was in previous years. There are no significant differences in these measures between playing position. Similarly, speed over 10–40 m does not appear to differ between playing position. Establishing the reliability of distance and velocity-derived GPS data in highly specific game-related activities is needed; once achieved, GPS data have the potential to accurately inform coaches of the position-specific demands on their players and to drive the development of training practices that reflect the changing demands of the game.

An intense national competition coupled with the push by administrators to maintain high spectator appeal for the game has led to significant changes in the way Australian Football is now played. The impact of these changes on player demands (e.g. injury risk, playing career longevity) and on training methods has been of particular interest to sports scientists.[1] The primary aim of this article is to review the current game demands, including movement patterns and game activities in Australian Football. In addition, the physiological attributes of players is evaluated. Findings in the literature should be used by conditioning staff to better prepare players in the current competition. Injuries in Australian Football have been comprehensively addressed by others and are therefore not included in this review. The interested reader is referred to the 2007 Australian Football League (AFL) injury report[2] and an earlier review article.[3] Indeed, a recent search in 2009 of the term ‘Australian Football’ in PubMed returned 193 articles, of which 47% related to some aspect of injury, 19% were not related to Australian Football specifically and 8% discussed various physiological testing methods. Aspects of game analysis (5%), fitness attributes (5%), performance/ ergogenic aids (6%), skill acquisition/decision ª 2010 Adis Data Information BV. All rights reserved.

making (5%) and sociology (5%), made up the remaining body of available literature. Given that much of the published literature is outdated and of questionable relevance to the modern game, a review of recent research utilizing current technologies is timely. Collectively, a number of recent AFL research reports describe the current changes in game structure and game demand. These reports are central to the review as they arguably best reflect the demands being placed on modern elite players. Reference is made to earlier publications for historical perspectives. 1. Description of Australian Football Australian Football is currently played in over 30 structured leagues throughout 30 countries worldwide. The premier competition, the AFL, is played in all seven of Australia’s states and territories and is currently the most popular football code in Australia. Undoubtedly, its strongest following is in the southern state of Victoria where it originated in 1858. Australian Football is a contact, invasion game that generally involves less collisions and tackling than rugby union and rugby league. The objectives and game structure of Australian Football are similar to those of football (soccer), and have been described as a Sports Med 2010; 40 (4)

Game Demands of Australian Football

349

running game combining athleticism with speed, and requiring skilful foot and hand passing.[4] In modern Australian Football two teams contest play over four 20- to 30-minute periods with the objective of scoring more points than the opposing team to win. A team is composed of 22 players, with 18 players allowed on the field at any one time. The remaining four players make up an interchange bench and these players can be rotated onto the field as frequently as the coach deems necessary. Field dimensions and tradiB

G

tional playing positions are detailed in figure 1. Positions can be grouped into forwards (offensive), nomadic players and backs (defensive). Forwards and backs each include two generally taller, more stationary key position players (centre half forward/back and full forward/back) and four smaller, more mobile players (half forward/back flanks and forward/back pockets). The forwards’ primary role is to mark the ball (catching it on the full from a kick longer than 15 m) in the best position possible to then take a free kick at the G

B

6.4 m 6.4 m

6.4 m

9m Full back Back pocket

Back pocket

50 m line Centre half-back

3m

Wing

15 m

Wing

Centre

Length 135−185 m Boundary line

Followers rover

10 m 15 m

Half-back flank

50 m

Boundary line

Half-back flank

50 m Centre half-forward

Half-forward flank

Forward pocket

Half-forward flank

Forward pocket Full forward

6.4 m

B

6.4 m

6.4 m

G G B Width 110−155 m

Fig. 1. Australian Football playing field and traditional playing positions.[5] B = ‘behind’ posts; G = ‘goal’ posts.

ª 2010 Adis Data Information BV. All rights reserved.

Sports Med 2010; 40 (4)

Gray & Jenkins

350

goals. The backs aim to defend against the opposition forwards, ‘spoiling’ their marking attempts. The small forwards and backs generally work alongside the key position players, either picking up the loose ball from the ground and kicking for goal (small forwards) or regaining possession and clearing the ball from their goal area (small backs). The traditional playing positions of centre, wings and followers are collectively termed midfielders. The midfielders and the ruckman are all nomadic players who link their offensive and defensive team mates, moving the ball up or down field. For the purpose of this review, midfielders will collectively refer to centres, wings and followers, whilst nomadic players will collectively refer to centres, wings, followers and ruckmen.1 However, the ruckman (usually one of the tallest players) has the additional role of contesting the restart of play, where the ball is bounced or thrown up by an umpire. Two opposing ruckmen jump to ‘tip off’ the ball to one of their surrounding team mates who aims to ‘clear’ the ball by kicking or hand passing to others in more space. Permitted contact is limited to: tackling the player in possession of the ball (between the shoulders and knees); using the body or arm to push, bump or block opposition players within 5 m of the football (shepherding); and contact, which is incidental to a legitimate marking contest. Players are penalized for charging (e.g. using unnecessary force), tripping, striking and bumping a player above the shoulders. Points are awarded for kicking the ball through two centre goal posts (a ‘goal’, 6 points) or within the outside posts (a ‘behind’, 1 point). Each quarter can include about 10 extra minutes of stoppage time, which brings the total time of a game to approximately 120 minutes. 2. Game Evolution 2.1 Early Time-Motion Analysis Research

Match analysis was first used in Australian Football in 1963 by Nettleton and Sandstrom[6]

who, in their discussion of skill and conditioning, commented ‘‘the content of training will depend upon intelligent analysis of the individual performance of players during a large number of games.’’ The few published articles that have since analysed match demands in Australian Football have focused on recording and classifying the time spent and/or distance travelled by players in various movement patterns. Movement descriptors such as ‘stationary’, ‘standing’, ‘walking’, ‘jogging’, ‘low intensity’, ‘running’, ‘medium pace’, ‘sprinting’ and ‘high intensity’ have been used in previous studies.[6-9] Early investigations established that much of the game was spent by players in low intensity activities. Estimates of time spent in low intensity activities ranged from 60% to 90% of the game depending on playing position and the method of analysis employed (table I).[7,8,10] Nomadic players such as ruckmen and the other midfielders spent more time engaged in higher intensities of exercise[7,8] and covered greater distances than forwards or backs.[7,8] The total distance covered by players in a game was first reported as a team average of 9591 m.[7] Later studies estimated that midfielders and backs covered over 10 000 m.[8] Hahn et al.[9] were the first to report on the intermittent exercise demands on players, finding that two midfielders completed 110 and 145 sprints in a game, over average distances of 12 m and 19 m, respectively. McKenna et al.[10] recorded 108.5 high intensity efforts during a game; 65% of these lasted <4 seconds in duration, 83% were <6 seconds and only 5% lasted >10 seconds. Described as ‘‘patchy summaries of individual players,’’[1] the findings from early time-motion studies in Australian Football are now outdated and bear little relevance to the modern game. The game is now considerably faster, played over a reduced duration (quarters reduced from 25- to 20-minute quarters in 1994) and governed by many different rules. Furthermore, small sample sizes (e.g. n = 2), data collection from only fractions of a full match (e.g. halves or quarters) and dated methods of analysis, limit the value of many

1 Rovers or ruck rovers are now known as ruckmen. Followers typically ‘follow play’ across the field and are one of the midfield positions.

ª 2010 Adis Data Information BV. All rights reserved.

Sports Med 2010; 40 (4)

Game Demands of Australian Football

351

Table I. Major findings of time-motion analysis in Australian Football Study (year)

Jaques and Pavia[7] (1974)

Collection period per game/s

1/4

No. of players

Position

Movement pattern (mean % of time in activity) low intensity

high intensity

stationary

walk

jog

run

14

61

21

5 5

5

Ruckmen

3

Followers

9

58

28

3

Centre and wings

17

60

16

7

4

Forwards

11

71

15

4

12

67

18

5

Backs

1

Midfielder

1

SB

4

Midfielders

Pyke and Smith[8] (1975)

1/2

McKenna et al.[10] (1988)

1

Schokman et al.[11] (2003)

16

5

Dawson et al.[12] (2004)

4–6

2–3

3

62

38

76 8.8

sprint

24

44.5

40.9

22

43

24

4.1

FF/FB

23.2

53.7

18.7

3.4

1

2–3

Midfielders

10.8

48.2

34.7

5.6

0.7

2–3

Ruckmen

13.9

52.4

25.4

4.7

1

2–3

SF/SB

16.5

52.4

25.4

4.7

1

2–3

CHF/CHB

11.1

49.3

34.4

4.6

0.6

10

1

CHB = centre half back; CHF = centre half forward; FB = full back; FF = full forward; SB = small back; SF = small forward.

of the early findings when attempting to understand the game demands of modern Australian Football. 2.2 Changes in Game Speed and Structure

The game structure and speed of Australian Football have changed significantly over the last 4 decades; video analysis of selected games between 1961 and 1997 found that ball velocity (whilst ‘in play’) increased from approximately 3.3 m/sec in 1961 up to approximately 6.6 m/sec in 1997. However, the fraction of time that the ball is in play has decreased.[1] In 1961, the ball was in play for 81 minutes (69.2%), with 36 minutes (30.8%) of stoppage time. In 1997, the ball was in play for 59 minutes (49.3%) and stoppage time was calculated to be 61 minutes (50.7%). Australian Football now comprises more frequent shorter periods of play (at higher intensities) interspersed by more frequent and longer stoppage periods. Changes in game tactics have paralleled rule changes over the past 40 years. During the late 1960s, coaches began challenging the traditional ‘kick and mark’ type game and introduced the concept of ‘possession’ football, moving away ª 2010 Adis Data Information BV. All rights reserved.

from traditional positions and getting players to move in groups.[4] This style of play still remains effective today and is likely to be the most influential coaching strategy on increasing game demands.[4] Additionally, rule changes such as disallowing deliberate kicks out of bounds (1968) and the introduction of the centre square (1973) have tended to reduce congestion around the ruck and improve the flow of the game.[4] Increasing professionalism, the start of a national competition and the greater size and skill of current players have also contributed to changes in game demands.[1] Whilst Norton et al.[1] reported an increase in the mean number of total bounce downs per game (reflecting stoppage time) from 11 in 1961 to 30 in 1997, analysis of more recent game statistics revealed that this increased further to 60 per game in the following 6 years (2003).[13] Over the same period there was a reduction in long kicks (-23%) and contested marks per game (-27%), with corresponding increases in short kicks (22%) and uncontested marks per game (17%).[13] Thus, the trend toward increasing intermittent play had continued. Characteristic elements of the game that appealed to spectators and supporters (high flying contested marks from Sports Med 2010; 40 (4)

Gray & Jenkins

352

long kicks and flowing play) were fading, and the AFL had growing concerns. The response of administrators has included the progressive introduction of rule changes from the 2004 season onwards (see table II). Although no further research on such game trends has been published in peer-reviewed journals, a number of AFL-commissioned projects investigating game structure, player game demands (using global positioning systems [GPS] Table II. Major rule changes for the Australian Football League (AFL) since the mid-1960s[4,14] Year

Major rule changes

1969

Players free kicked for kicking the ball out of bounds on the full

1973

Centre square established where only four players a side are allowed for centre bouncesa

1976

Two field umpires used in matches for the first time

1978

Introduction of two interchange players (unlimited interchanges)

1988

Players awarded a free kick are obliged to kick the ball rather than playing on

1990

Players who are awarded free kicks again are given option of playing on

1994

Quarters are reduced from 25 min to 20 min each (plus stoppages) The introduction of three interchange players (unlimited interchanges)

1998

The introduction of four interchange players (unlimited interchanges)

2005

Players are penalized for holding up another player after a markb or free kick Reduced time is allowed after marks and free kicks are awarded

2006

Players are able to kick-inc without having to wait for the goal umpire’s flag to be waved A limit on time for players to line up for set shots on goal Umpires are instructed to perform quicker throw-ins,d ball upse and centre bounces

a

To commence play, the field umpire bounces the ball in the centre circle.

b

When the ball is caught on the full from a kick longer than 15 m. The player is awarded a free kick.

c

To recommence play after a behind is scored, a player kicks the ball into play from the goal square.

d

To recommence play if the ball goes out of bounds, the boundary umpire throws the ball backwards over their head into play from the boundary line.

e

To recommence play at a neutral contest after a stoppage within the field of play, the field umpire bounces the ball or throws it upwards.

ª 2010 Adis Data Information BV. All rights reserved.

technology) and trends in game statistics (to monitor the impact of interventions) from recent seasons are available.[13-18] Collectively, these reports provide an assessment of the influence recent rule changes in Australian Football have had on game demands. However, the use of the GPSbased research[15-18] to identify seasonal changes in game demand is limited. For each season, game files from all positional groups were pooled; however, since 2005 the relative proportions of forwards, nomadic players and backs have changed markedly. In 2005, approximately 44% of all game files were for nomadic players, while in 2008 this increased to approximately 82%. Thus, values from the 2007 and 2008 seasons generally reflect the game demands of nomadic players only. Any apparent trends over the last four seasons should therefore be considered with this in mind; in light of this bias, we have chosen to exclude these data from this section of the review. Recently, introduced rule changes (summarized in table II) appear to have reduced total stoppage time.[14] This has resulted from improvements in non-clearance rates from ruck contests, a reduction in the time taken for umpires to restart play (i.e. ball-ups, boundary throw-ins, centre bounce-downs), a reduction in time taken to kick-in (after the opposition kicks a behind) or take a set shot at goal, and faster ‘playing on’ after a mark.[14] These are essentially a reversal of trends found over the last 4 decades. The improved flow of the game was anticipated to reduce players’ on-field recovery and their subsequent ability to maintain high game speeds. This was thought to be beneficial as it could potentially reduce the severity of collision injuries and extend playing careers. However, coaches have increased their utilization of the interchange bench, and since 2000 (see table III) there has been a large reduction in the percentage of players remaining on the field for entire matches (66% in 2000 reducing to 20% in 2005).[14] Consequently, game speed during play has increased since 2002.[14] Given that the mean speed of players is highest in the first few minutes of a quarter and that players who are interchanged more frequently perform more high intensity efforts,[20] it is reasonable to conclude that there is a Sports Med 2010; 40 (4)

Game Demands of Australian Football

353

Table III. Key game statistics relating to game flow and speed[14,19] Season (year)

Stoppages

Time in play (%)

Ball speed (m/sec)

Interchanges

no.

non-clearance rate (%)

duration (sec)

2001

90

18.5

30.5

47.5

6.4

20

2002

94

23.4

27

49

5.8

24

2003

91

22.8

29

51

5.8

22

2004

88

23.4

28

50

5.9

30

2005

79

18.3

33.2

51

6.2

36

2006

78

19.0

21.6

61

6.5

46

2007

80

20.4

22

60

6.8

57

2008

83

22.0

relationship between game speed and interchange utilization. This coaching strategy aimed at maintaining fresh, involved players in the game seems to have been employed to combat the greater demand imposed by the recent rule changes. The possible effect of limiting the number of interchanges per game is currently under review. 3. Modern Australian Football 3.1 Global Positioning Systems Match Analysis and Game Demands

Team use of GPS technology in AFL competition games was first permitted during the 2005 season; clubs were allowed to monitor five players during ten games of the season. In 2009, ten players per game could be monitored. Since 2005, many clubs have submitted their data to AFL researchers, who have subsequently assessed the seasonal game demands of players.[15-18] Despite the popularity of GPS technology in the AFL, few studies have examined the accuracy and reliability of GPS techniques in the measurement of distance and velocity in field sports.[21-24] Nonetheless, GPS technology has the potential to provide coaches and support staff in AFL clubs with detailed objective information on the movement demands of their players. Wisbey and co-authors[15-18] used GPS data to derive an exertion index to represent the total physical game load on players in recent seasons. The exertion index is thought to encompass all aspects of a player’s movement profile, summing ª 2010 Adis Data Information BV. All rights reserved.

81

the short high intensity efforts with low and moderate intensity jogging and running. Although it is difficult to derive detailed meaning from reported values, the simplicity of a single number to describe the overall movement demand in a game or training setting is appealing to coaches. The authors report that nomadic players (midfielders and ruckmen) consistently maintain higher indices than both forwards and backs. Similarly, variables reflecting playing intensity (average speed and exertion index/minute) are presently significantly higher for nomadic players than for forwards and backs. No consistent differences between forwards and backs were found by the authors in these variables.[18] When all playing positions are pooled, the mean total distances covered by players during a game have been found to range from 12 030 m to 12 510 m.[15-18] Consistent with data from previous GPS reports, 2008 data confirmed that nomadic players covered a slightly greater distance (12 310 m) than both forwards (11 920 m) and backs (11 880 m). These values vary markedly from earlier estimates. Previous time-motion analysis studies used ground markings and cues[6,7,9] or stride length and frequency[8,12] to estimate the distances travelled. Dawson et al.[12] used step distance and step frequency from video footage to estimate distances travelled by selected players during the 2000 season. The reported mean distances travelled were 13 614 m, 15 393 m, 16 005 m, 16 278 m and 16 976 m for full forwards/backs, ruckmen, centre half forwards/backs, small forwards/backs and midfielders, respectively. Similarly, analysis of elite Sports Med 2010; 40 (4)

Gray & Jenkins

354

AFL players in 2002–3 (using a drawing tablet attached to a laptop) established mean total distance covered by players to be 16 600 m.[25] The variation between the findings of Wisbey and coauthors,[15-18] Dawson et al.[12] and Burgess and Naughton[25] should be interpreted with caution, given the differences in methods employed and the changes in game speed and structure. Whilst total distance travelled in a game is of interest and can be used to estimate the volume of work a player has completed, some have suggested that it is the frequency and duration of moderate and high intensity activities that more accurately describe the game demands imposed on players, as these activities place considerable strain on the anaerobic energy systems.[26] Sprints, sustained efforts, accelerations, decelerations and work-recovery ratios collectively help define the more demanding activities expected of players during a match. Wisbey and co-authors[15-18] have used speed zones in much the same way as movement descriptors (e.g. jog/run) to describe the intensity of play from data collected using GPS tracking. For comparative purposes, standing and walking »0–1.66 m/sec, jogging »1.94–3.88 m/sec, running »4.16–5.55 m/sec and sprinting »6.11–10 m/sec have been used. Analysis of data from the 2007 season revealed the time spent in each zone during games by forwards, nomadic players and backs[17] (table IV). It is clear that much of the current game is spent by players recovering from high intensity exercise, and that this reflects the increased speed at which the game is now played.

No significant differences were found by the authors between forwards and backs across the speed zones, suggesting relatively similar movement demands for these positions.[15-18] However, nomadic players spent significantly less time (~11 minutes) moving <1.66 m/sec (i.e. standing and walking) and significantly more time running at speeds between 4.44–5 m/sec and >5 m/sec than forwards and backs (p < 0.05). Backs spent the least amount of time running at speeds >6.94 m/sec (i.e. sprinting) in a game (24.90 – 16.31 seconds) – significantly less than nomadic players (33.45 – 17.01 seconds). These more recent findings are in agreement with those of Dawson et al.[12] relating to games played several seasons earlier. From their research, full forward/back positions were noted as being the most different from the others and were considered ‘specialist’ positions.[12] Full forwards/ backs spent an estimated 23.2% of the game standing, but also covered more distance sprinting and spent a longer percentage of time sprinting when compared with other positions who were found to be more ‘mobile’ (10.8–16.5% of the game spent standing). Midfielders spent the least amount of time standing (10.8%) and the most amount of time fast running (5.6%). For all positions, fast running and sprinting (i.e. high intensity movement patterns) accounted for only 4.4–6.3% of game time.[12] The positional data from the study by Dawson et al.[12] are summarized in table I. GPS tracking analyses reveal that players in all positions make more moderate accelerations (~250) than decelerations (~235) in a game, but

Table IV. Mean time (–SD) spent in speed zones by forwards, nomadic players and backs per game in season 2007 (reproduced from Wisbey et al.,[17] with permission) Speed zone (m/sec)

Forwards (n = 39)

Nomadic (n = 493)

Backs (n = 29)

<1.66

68:48 – 10:29 (min)

57:22 – 11:24 (min)

69:22 – 12:44 (min)

<2.22

80:02 – 11:29 (min)

67:00 – 12:09 (min)

79:26 – 13:35 (min)

2.22–2.78

6:31 – 1:29 (min)

7:23 – 1:37 (min)

7:20 – 1:30 (min)

2.78–3.33

6:40 – 1:35 (min)

8:04 – 1:42 (min)

6:55 – 1:21 (min)

3.33–3.89

5:26 – 1:19 (min)

6:54 – 1:39 (min)

5:31 – 1:11 (min)

3.89–4.44

3:51 – 0:55 (min)

5:07 – 1:20 (min)

4:04 – 0:59 (min)

4.44–5

2:22 – 0:38 (min)

3:17 – 0:53 (min)

2:34 – 0:53 (min)

4:36 – 1:14 (min)

5:49 – 1:25 (min)

4:19 – 1:14 (min)

31.72 – 16.04 (sec)

33.45 – 17.01 (sec)

24.90 – 16.31 (sec)

>5 >6.94

ª 2010 Adis Data Information BV. All rights reserved.

Sports Med 2010; 40 (4)

Game Demands of Australian Football

more rapid decelerations (~15) than accelerations (~12).[17] Significantly more moderate changes in velocity occur than rapid changes, with ruckmen performing significantly less rapid accelerations and decelerations than all other positions.[16] Analysis of the number of ‘surges’ or running efforts at higher speeds has shown that nomadic players move into the higher speed zones (>4.44 m/sec and >5 m/sec) significantly more frequently than forwards and backs.[17,18] Nomadic players totalled 98.5 – 26.1 surges >5 m/sec compared with 87.1 – 22.9 and 81.0 – 20.7 for forwards and backs, respectively.[18] Dawson et al.[12] had previously reported that high intensity efforts numbered between 158 (full forward/backs) and 208 (midfielders) per game, with approximately 80–85% of sprint efforts lasting <3 seconds. All were <6 seconds in duration. This is consistent with other field sports, where sprint efforts typically last 2–3 seconds, and players cover distances of 10–20 m.[27] Data from the 2008 AFL season show the longest efforts >5.55 m/sec averaged 11–13 seconds for all positions.[18] Time spent by players in steady-state running (>2.2 m/sec) per game was significantly greater in nomadic players (24 min 40 sec) than both forwards (21 min 10 sec) and backs (21 min 28 sec).[18] Considering players average 100 minutes of on-field playing time,[18] steady-state running represents a relatively minor component of the game. The longest continuous periods of time spent moving at velocities <1.38 m/sec (recovery periods) averaged 90 seconds, 92 seconds and 114 seconds for nomadic positions, forwards and backs, respectively.[17] Dawson et al.[12] had previously reported recovery between high intensity efforts to be relatively brief, with almost half of the recovery periods between high intensity efforts generally lasting <20 seconds. However, it was not uncommon for some low intensity work periods to be >2 minutes in duration.[12] Where ‘work’ was considered the total time spent >2.22 m/sec and ‘recovery’ was the total time below, Wisbey and Montgomery[15] found the average work-recovery ratio in the 2005 season to be 1 : 2.6, 1 : 2.3 and 1 : 2 for forwards, defenders and nomadic players, respectively (more recent data are not available). These ratios may not reflect ª 2010 Adis Data Information BV. All rights reserved.

355

current game demands, given the recent changes in game structure. Further analysis of GPS data is likely to reveal the frequency and lapsed time between work periods and short, moderate and long recovery periods. Such information would aid in the design of game-specific conditioning drills. 3.2 Heart Rate Responses

Despite the ease with which heart rate data can now be collected during exercise, only two studies have reported the heart rate response during Australian Football games. In the 1970s, Pyke and Smith[8] collected heart rate data on two positions, a ‘small’ back and a midfielder, over one quarter of a game. Mean heart rates were 160 bpm and 178 bpm for the back and midfielder, respectively. It was found that during intense periods of play, the heart rate of the back reached 170–180 bpm (»94% maximum heart rate), dropping below 140 bpm during lengthy recovery periods. In contrast, the heart rate of the midfielder did not fall below 150 bpm and remained between 170 bpm and 185 bpm for most of the quarter. The authors concluded that the running demands of ‘nomadic’ players such as midfielders and ruckmen mean they have little opportunity to recover, taxing their cardiovascular system to a greater extent.[8] In the only other study to report heart rates, two midfielders were monitored during an entire game by Hahn and colleagues in 1979.[9] The heart rates of the players ranged from 140 to 180 bpm with averages of 164 bpm and 159 bpm reported. Updated analysis of the heart rate responses of players during Australian Football would be of value. Such information may be used in the development and evaluation of game-specific training drills where GPS technology is unavailable. Heart rate monitoring has been used to quantify exercise intensity[28] in continuous aerobic exercise, and training load[29,30] in intermittent exercise bouts such as those performed in Australian Football. Currently, GPS is used in preference to heart rate monitoring as it is able to provide a more accurate description of the movement patterns during play. Heart rate monitoring Sports Med 2010; 40 (4)

Gray & Jenkins

356

in intermittent sports is perhaps more appropriately used to monitor weekly or monthly training loads. 4. Physiological and Anthropometric Characteristics 4.1 Mass, Stature and Body Composition

The average height of current elite Australian Football players is ~1.87 m,[31,32] with sub-elite and elite juniors slightly shorter (table V).[33,36] Young et al.[32] found that height was not significantly different between midfielders, forwards and backs. However, other studies have found substantial positional differences in height, with small forwards and small midfielders shorter than the mean height, whilst key position forwards/ backs and ruckmen are all taller.[35] A recent study investigating the relationship between anthropometric/fitness measures and playing performance found height to be the only significantly different variable between players who achieved more than six hit outs (hitting the ball clear of a ruck contest such as a bounce down) per game compared with those who achieved less.[36] This confirms the importance of having the tallest players (ruckmen) competing in ruck contests, at least at the elite junior level. On average, body mass has been found to range between 85 and 90 kg at the elite level in the AFL with some key position players weighing more than 100 kg; draftees and elite juniors are slightly lighter and have body masses ranging from 74 to 82 kg.[32-35] This alone suggests greater mass (presumably muscle mass) is necessary to play Australian Football at the elite level and confirms the importance of strength and size. Elite juniors identified as taking significantly more marks over an eight-game period were found to be significantly heavier.[36] However, at an elite level, body mass has not been found to differentiate between ‘successful’ and ‘unsuccessful’ players in the AFL,[32] nor between young players at the national draft camp.[34] Body mass appears to be more advantageous for competitors at the sub-elite level. It has been stated that as an elite junior, the biggest physical ª 2010 Adis Data Information BV. All rights reserved.

change needed to play in the AFL is an increase in body mass with proportionate increases in strength.[33] Interestingly, long term increases in body mass are evident especially at the juniorelite level. The average body mass of a state under-18-years squad in 1999 was approximately 75 kg, while in 2005 a larger cohort of players at the same level weighed approximately 80 kg. Skinfold data reveal that AFL players have low levels of body fat and that these have not changed with time. However, no consistent relationship has been found between estimates of body fat and playing position,[35] nor is level of body fat able to differentiate between ‘successful’ or ‘unsuccessful’ players.[32,34] Average ‘sum of seven’ and ‘sum of eight’ skinfold data from several studies are summarized in table V. 4.2 Maximal Oxygen Uptake

Several studies have reported on maximal aerobic power values of Australian Football players at the elite and junior elite levels of competition.[31-36] Estimated from the 20 m shuttle run test, these studies found that the maximal oxygen uptake . (VO2max) values of players ranged from 55 to 65 mL/kg/min with little variation existing between junior elite and elite level players. This is consistent with elite players in other running-based sports such as soccer (57.8 – 6.5 mL/kg/min)[38] and field hockey (61.8 – 1.8),[39] but greater than professional rugby league and rugby union players (~50 mL/kg/min).[26,40] In the few studies that have considered playing position, no positional differences were identified;[32] however, a high . VO2max has been found to be a defining quality of successful draftees to AFL clubs.[34] Physiological profiles of sub-elite or state league Australian Football players are not available for comparison. Further examination of current players from a variety of playing levels assessed in a laboratory setting may be of interest to those developing junior athletes or for talent identification purposes. 4.3 Muscular Strength and Power

Bumping, tackling and ‘wrestling’ opposition players when contesting a mark or loose ball on Sports Med 2010; 40 (4)

Study Buttifant[31]

Keogh[33]

Pyne et al.[34]

Pyne et al.[35]

Young et al.[32]

Young and Pryor[36]

Year of data collection

1993–7

1999

1999–2001

1999–2004

2004

2005

No. of players

18

29

283

495

38

177–200

Level of competition

AFL (elite)

State U/18

AFL draft camp

AFL draft camp

AFL (elite)

State U/18

Age (y)

22.7 [1993]–26.8 [1997]

15.9

17–18

17–18

22.6 – 2.9

16–18

Positions

All

All

All

All

Nomadic

Forwards

Backs

All

Anthropometry height (cm)

186

180.2 – 7.2

186 – 6.62

187 – 6.6

188 – 9

186 – 10

187 – 5

183.9 – 6.9

body mass (kg)

85.7 [1993]–90.5 [1997]

74.6 – 8.3

80.5 – 7.82

81 – 7.6

86.8 – 8.9

87 – 7.75

87.7 – 7.5

79.8 – 8.3

skinfolds (mm)

55.4 [1993]–63.6 [1997]b

56 – 3.4b

55.3 – 12.6b

47 – 7.8c

59.7 – 16c

53.3 – 12c

57.8 – 3.4

57.8 – 3.4

61.6 – 3.5

57.8 – 5.1

61.1 – 3.5

747 – 123

656 – 128

743 – 142

1.9 – 0.06

1.93 – 0.1

1.88 – 0.07

Game Demands of Australian Football

ª 2010 Adis Data Information BV. All rights reserved.

Table V. Physiological characteristics of Australian Football playersa

Endurance capacity

. 20 m shuttle run (VO2max; mL/kg/min)

53.6 [1993]–59.2 [1997]

57.7

yo-yo IR2d test (m)

57.3 – 3.5

Speed and acceleration 1.09 – 0.06

5 m (s) 1.83 [1993]–1.66 [1997]

1.81 – 0.07

20 m (s)

3.06 [1993]–2.92 [1997]

3.04 – 0.09

40 m (s)

5.37 [1993]–5.40 [1997]

10 m (s)

1.12 – 0.05 3.04 – 0.09

3.13 – 0.09 3.48 – 0.11

Flying 30 m split (s)

3.6 – 0.15

3.49 – 0.06

Strength 97.9 – 11.9

93 – 10.4

98.8 – 8.4

3RM leg press (kg)

399 – 48

348 – 64

410 – 44

3RM chin ups (kg)

109 – 10

103 – 12

106 – 5

3RM bench press (kg)

63.8 – 10.7

Leg power concentric squat jump (W/kg)

68.3 – 6.2

70.5 – 7.4

68.1 – 7.1

countermovement squat jump (W/kg)

70.2 – 8.8

65.1 – 9.4

69.1 – 7.8

62.3 – 3.7

58.4 – 5.1

63.4 – 2.4

Jumping ability countermovement jump (cm)

55.2 – 7.9

60.1 – 5.76 322 – 11.1

absolute running vertical jump (cm)

322 – 10.9

Flexibility 10.7 – 6.4

Performance and anthropometric values are expressed as mean – SD.

b

Sum of seven skinfolds.

8.8 – 7.5

c Sum of eight skinfolds. d

There are two yo-yo protocols: level 1 and level 2. Level 2 is reported here and has higher initial running speeds.[37]

. AFL = Australian Football League; IR2 = intermittent recovery level 2; RM = repetition maximum; U/18 = under 18 y age group; VO2max = maximal oxygen uptake.

357

Sports Med 2010; 40 (4)

sit and reach test (cm) a

Gray & Jenkins

358

the ground requires high levels of upper and lower body strength and power. Although not statistically significant, the data of Young et al.[32] suggest elite forwards and backs possess slightly greater upper body strength than elite midfielders (table V). Mean 3 repetition maximum (3RM) bench press values for players at the highest level (~95–100 kg) are far superior to junior elite players (~63 kg), reinforcing the significant gains in size and strength junior athletes must attain to be competitive in the AFL. Young et al.,[32] measured leg strength by 3RM leg press, again finding higher values for forwards and backs (~400–410 kg) than for midfielders (~350 kg). Measures of leg power and jumping ability assessed using the vertical jump are higher for Australian Football players (~62 cm) than for rugby league players (~55 cm)[26] but comparable to rugby union backs (~60 cm).[40] Pyne et al.[34] identified that junior players drafted to AFL clubs who went on to debut in the AFL had better jumping and agility qualities than those who did not. These fitness components reflect the skill demands required of the various positional groups. 4.4 Speed and Repeated Sprint Ability

The ability to outrun defenders and ‘find space’ or chase down opponents is integral to Australian Football; good acceleration and sprint ability are required by players in all positions, particularly with the speed of the game having increased in recent years. Similar to aerobic capacity, speed over 10–40 m does not appear to vary greatly between positions. Mean times over these distances for junior elite and elite players are summarized in table V. Recent estimates of maximal running speeds determined from ‘flying’ split times following 20 m of acceleration have been found to be 8.6 and 8.7 m/sec for two elite AFL teams.[41] In a practical sense, these players are capable of accelerating to speeds >80% of their maximal velocities in just a few seconds; a desirable quality that influences team selection at the elite level.[32] Young et al.,[32] also found the ability to recover from intermittent high intensity exercise to be a defining quality of ‘better’ players in elite ª 2010 Adis Data Information BV. All rights reserved.

Australian Football. Again, given that the game is becoming faster, repeated sprint ability is becoming increasingly important for players. Surprisingly, there has been only one published article describing the repeated sprint abilities of Australian Football players. Using a 6 · 30 m sprint protocol (on a 20-second cycle), Pyne et al.[42] reported the mean total time of all six sprints to be 25.83 – 0.6 seconds for 60 junior elite players attending the AFL national draft camp. A performance decrement of 4–6% was determined. Future research should assess repeated sprint ability at both elite and sub-elite levels of Australian Football competition. Knowledge of the degree to which repeated sprint ability varies relative to playing position would also be of value. In addition to the acknowledged importance of aerobic fitness, strength, power and speed for players who play Australian Football at a competitive level, there is evidence to suggest that reduced flexibility may predispose athletes to soft tissue injury.[43] The importance of maintaining flexibility should therefore not be overlooked.

5. Practical Applications for Conditioning Given that recent GPS research reports describe a high intensity intermittent movement profile for all playing positions, development of repeated sprint ability should be a primary focus for gamespecific player preparation. However, the specific nature of repeated high intensity exercise training may need to differ across playing positions. For example, nomadic players complete a greater number of high intensity efforts during a game, sprint for longer and have shorter recovery periods between bouts compared with players in other positions. Although steady-state running represents less than a quarter of total game demands, aerobic or endurance capacity will remain an extremely important component of fitness for all AFL players; recovery from high intensity exercise (e.g. removal of lactic acid and the resynthesis of creatine phosphate) is recognized as being strongly related to peripheral oxidative capacity.[44] This may be developed through various exercise modes. Sports Med 2010; 40 (4)

Game Demands of Australian Football

The relatively short duration of sprint efforts (2–3 seconds) suggests players seldom reach their maximal speed in games. Thus, development of acceleration may be of greater benefit than a focus on improving maximal speed during game-specific training. Nonetheless, players with greater speed have a distinct advantage when longer efforts are required. The importance of developing maximal speed should therefore not be underestimated. Recovery periods could be manipulated to alternate between developing ‘pure’ speed/acceleration (e.g. long recovery with a focus on technique) and repeated sprint ability (e.g. short recovery periods between bouts mimicking game demands). Future match analysis research should aim to couple preceding movement velocities with the frequency and type of accelerations/decelerations performed. This may help assess the value of standing starts, ‘flying starts’ and ‘varied pace’ acceleration drills. 6. Conclusions As is the case in most team sports, the available literature relating to Australian Football clearly reveals significant differences in game demands between different positions. The preparation of players for competition needs to reflect these positional differences and also take into general consideration the acknowledged increase in the speed of the game. It is reasonable to conclude that repeated sprint ability is more important for elite AFL players now than it was 20 or 30 years ago. Nonetheless, endurance (aerobic) fitness will remain a central determinant of the ability of the player to compete effectively in the AFL, not only given the large accumulated distances covered by all players during a game but also because of the need for players to recover rapidly between bouts of repeated high intensity exercise. GPS technology is providing coaches with objective and detailed data relating to specific movement demands of players. However, the accuracy and reliability of distance and velocityderived GPS data in highly specific game-related activities need to be established. Once this is achieved, GPS data that describe changes in veª 2010 Adis Data Information BV. All rights reserved.

359

locity, repeated sprint intensities and subsequent recovery periods can be used to develop positionspecific training programmes for players and to monitor workloads both in competition and in training. Assessment of the relationships between the workload and various overuse injuries can follow. Furthermore, the potential of GPS data to differentiate movement patterns between elite and non-elite competition may aid in the development of junior players. Acknowledgements No sources of funding were used to assist in the preparation of this review. The authors have no conflicts of interest that are directly relevant to the content of this review.

References 1. Norton KI, Craig NP, Olds TS. The evolution of Australian Football. J Sci Med Sport 1999; 2 (4): 389-404 2. Orchard J, Seward H. AFL injury report: season 2007. Melbourne (VIC): Australian Football League, 2008 3. Hoskins WT, Pollard H. Injuries in Australian rules football: a review of the literature. Australas Chiropr Osteopathy 2003; 11 (2): 49-56 4. Stewart B. Boom time football, 1946-1975. In: Hess R, Stewart B, editors. More than a game: the real story of Australian rules football. Carlton (VIC): Melbourne University Press, 1998: 165-99 5. AFL. Laws of Australian Football 2008. Melbourne (VIC): Australian Football League, 2008 6. Nettleton B, Sandstrom ER. Skill and conditioning in Australian Rules Football. Aust J Phys Educ 1963; 29: 17-30 7. Jaques TD, Pavia GR. An analysis of the movement patterns of players in an Australian Rules league football match. Aust J Sports Med 1974; 5 (10): 10-21 8. Pyke FS, Smith RS. Football: the scientific way. Perth (WA): The University of Western Australia Press, 1975 9. Hahn A, Taylor N, Hunt B, et al. Physiological relationships between training activities and match play in Australian Football rovers. Sports Coach 1979; 3 (3): 3-8 10. McKenna MJ, Patrick JD, Sandstrom ER. Computer-video analysis of activity patterns in Australian Rules Football. In: Reilly T, editor. Science and football. London: Routledge, 1988 11. Schokman P, Sparrow WA, Le Rossignol P. The use of a movement and skills analysis system to measure player performances in Australian Football: an intervention case study. Melbourne (VIC): Australian Football League, 2003 12. Dawson B, Hopkinson R, Applby B, et al. Player movement patterns and game activities in the Australian Football League. J Sci Med Sport 2004; 7 (3): 278-91 13. Anderson A. AFL annual report: football operations. Melbourne (VIC): Australian Football League, 2005

Sports Med 2010; 40 (4)

360

14. Norton KI. Laws of Australian Football discussion paper. Melbourne (VIC): Australian Football League, 2007 15. Wisbey B, Montgomery P. Quantifying AFL player game demands using GPS tracking. Florey (ACT): FitSense Australia, 2005: 16-7 16. Wisbey B, Montgomery P. Quantifying changes in AFL game demands using GPS tracking. Florey (ACT): FitSense Australia, 2007 17. Wisbey B, Montgomery P, Pyne DB. Quantifying changes in AFL player game demands using GPS tracking: 2007 AFL season. Florey (ACT): FitSense Australia, 2008 18. Wisbey B, Rattray B, Pyne DB. Quantifying changes in AFL player game demands using GPS tracking: 2008 AFL season. Florey (ACT): FitSense Australia, 2008 19. Champion Data Game Statistics. Melbourne (VIC): Australian Football League, 2009 20. Norton K. Modeling the effect of a restriction on the number of interchanges in the AFL. Melbourne (VIC): Australian Football League, 2007 21. Edgecomb SJ, Norton KI. Comparison of global positioning and computer-based tracking systems for measuring player movement distance during Australian Football. J Sci Med Sport 2006; 9 (1): 25-32 22. Macleod H, Morris J, Nevill A, et al. The validity of a nondifferential global positioning system for assessing player movement patterns in field hockey. J Sports Sci 2009; 27 (2): 121-8 23. Townshend AD, Worringham CJ, Stewart IB. Assessment of speed and position during human locomotion using nondifferential GPS. Med Sci Sports Exerc 2008; 40 (1): 124-32 24. Witte TH, Wilson AM. Accuracy of non-differential GPS for the determination of speed over ground. J Biomech 2004; 37 (12): 1891-8 25. Burgess D, Naughton G. Quantifying the gap between under 18 and senior AFL football [abstract]. J Sci Med Sport 2003; 6 (4): 525 26. Gabbett TJ, King T, Jenkins D. Applied physiology of rugby league. Sports Med 2008; 38 (2): 119-38 27. Spencer M, Bishop D, Dawson B, et al. Physiological and metabolic responses of repeated-sprint activities. Sports Med 2005; 35 (12): 1025-44 28. Karvonen J, Vuorimaa T. Heart rate and exercise intensity during sports activities: practical application. Sports Med 1988; 5 (5): 303-12 29. Hopkins WG. Quantification of training. Sports Med 1991; 12 (3): 161-83 30. Stagno KM, Thatcher R, van Someren KA. A modified TRIMP to quantify the in season training load of team sport players. J Sports Sci 2007; 25 (6): 629-34

ª 2010 Adis Data Information BV. All rights reserved.

Gray & Jenkins

31. Buttifant D. Physiological and performance characteristics of Australian Football League players. J Sports Sci 1999; 17: 808-9 32. Young WB, Newton RU, Doyle TLA, et al. Physiological and anthropometric characteristics of starters and nonstarters and playing positions in elite Australian Rules Football: a case study. J Sci Med Sport 2005; 8 (3): 333-45 33. Keogh J. The use of physical fitness scores and anthropometric data to predict selection in an elite under 18 Australian Rules football team. J Sci Med Sport 1999; 2 (2): 125-33 34. Pyne DB, Gardner AS, Sheehan K, et al. Fitness testing and career progression in AFL football. J Sci Med Sport 2005; 8 (3): 321-32 35. Pyne DB, Gardner AS, Sheehan K, et al. Positional differences in fitness and anthropometric characteristics in Australian Football. J Sci Med Sport 2006; 9 (1-2): 143-50 36. Young WB, Pryor L. Relationship between pre-season anthropometric and fitness measures and indicators of playing performance in elite junior Australian Rules Football. J Sci Med Sport 2007; 10 (1): 110-8 37. Krustrup P, Mohr M, Amstrup T, et al. The Yo-Yo intermittent recovery test: physiological response, reliability and validity. Med Sci Sports Exerc 2003; 35 (4): 697-705 38. Williams C, Reid RM, Coutts R. Observations on the aerobic power of university rugby players and professional soccer players. Br J Sports Med 1973; 7: 390-1 39. Boyle PM, Mahoney CA, Wallace W. The competitive demands of elite male field hockey. J Sports Med Phys Fitness 1994; 34: 235-41 40. Duthie G, Pyne DB, Hooper S. Applied physiology and game analysis of rugby union. Sports Med 2003; 33 (13): 973-91 41. Young WB, Russell A, Burge P, et al. The use of sprint tests for assessment of speed qualities of elite Australian Rules footballers. Int J Sport Phys and Perf 2008; 3 (2): 199-206 42. Pyne DB, Saunders PU, Montgomery PG, et al. Relationships between repeated sprint testing, speed, and endurance. J Strength Cond Res 2008; 22 (5): 1633-7 43. Bradley PS, Portas MD. The relationship between preseason range of motion and muscle strain injury in elite soccer players. J Strength Cond Res 2007; 21 (4): 1155-60 44. Jansson E, Dudley GA, Norman B, et al. Relationship of recovery from intense exercise to the oxidative potential of skeletal muscle. Acta Phys Scand 1990; 139 (1-2): 147-52

Correspondence: Mr Adrian J. Gray, School of Human Movement Studies, The University of Queensland, St Lucia Campus, Brisbane, QLD, 4072, Australia. E-mail: [email protected]

Sports Med 2010; 40 (4)