sports medicine Volume 39 Issue 3 March 2009

Sports Med 2009; 39 (3): 167-177 0112-1642/09/0003-0167/$49.95/0 LEADING ARTICLE ª 2009 Adis Data Information BV. All ...

Author: adisonline.com

10 downloads 846 Views 960KB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Sports Med 2009; 39 (3): 167-177 0112-1642/09/0003-0167/$49.95/0

LEADING ARTICLE

ª 2009 Adis Data Information BV. All rights reserved.

Physical Activity in Culturally and Linguistically Diverse Migrant Groups to Western Society A Review of Barriers, Enablers and Experiences Cristina M. Caperchione,1 Gregory S. Kolt2 and W. Kerry Mummery1 1 Institute for Health and Social Science Research, CQUniversity Australia, Rockhampton, Queensland, Australia 2 School of Biomedical and Health Sciences, University of Western Sydney, Sydney, New South Wales, Australia

Abstract

A close examination of epidemiological data reveals burdens of disease particular to culturally and linguistically diverse (CALD) migrants, as these individuals adjust to both culture and modernization gaps. Despite the increased risk of hypertension, diabetes mellitus, overweight/obesity and cardiovascular disease, individuals from CALD groups are less likely to be proactive in accessing healthcare or undertaking preventative measures to ensure optimal health outcomes. The purpose of this paper is to review literature that outlines the barriers, challenges and enablers of physical activity in CALD groups who have recently migrated to Western society, and to identify key strategies to increase physical activity participation for these individuals. Electronic and manual literature searches were used to identify 57 publications that met the inclusion criteria. Findings from the review indicate that migration to Western societies has a detrimental effect on the health status and health behaviours of CALD groups as they assimilate to their new surroundings, explore different cultures and customs, and embrace a new way of life. In particular, there is evidence that physical inactivity is common in migrant CALD groups, and is a key contributing risk factor to chronic disease for these individuals. Challenges and barriers that limit physical activity participation in CALD groups include: cultural and religious beliefs, issues with social relationships, socioeconomic challenges, environmental barriers, and perceptions of health and injury. Strategies that may assist with overcoming these challenges and barriers consist of the need for cultural sensitivity, the provision of education sessions addressing health behaviours, encouraging participation of individuals from the same culture, exploration of employment situational variables, and the implementation of ‘Health Action Zones’ in CALD communities. This information will inform and support the development of culturally appropriate programmes designed to positively influence the physical activity behaviours of individuals from CALD populations.

168

Rising levels of immigration have been shaped by changing government policies, allowing for a diverse flow of immigrants from varied origins to traditional countries of immigration. In countries such as the US, UK, Canada, Australia and New Zealand, the numbers of foreign-born individuals has approximately doubled over the past 40 years, with the largest proportion originating from developing countries.[1] For example, the population of Australia increased by 134 600 persons in 2005–6 as a result of net overseas migration, representing 51% of the total annual population growth.[2] Individuals from the UK (24%), New Zealand (9%), Italy (5%), China and Vietnam (4% each), Greece, Germany, the Philippines and India (approximately 3% each) accounted for the majority of the overseas migration into Australia.[3] Additionally, Australia has witnessed a rapid increase in migration from countries such as Sudan, Afghanistan, Somalia, Bangladesh and Iraq.[4] A large proportion of migrants to traditional immigration countries have been born in countries recently affected by war and political unrest. Many of them have been subject to traumatic events such as prolonged periods of deprivation, the loss of family and friends in violent circumstances, or a perilous escape from their homeland.[1,4,5] Under such circumstances, many of these culturally and linguistically diverse (CALD) groups, a common term used to describe non-Anglo migrant groups,[6,7] are exempted from meeting certain health requirements that govern migration. As a result, many migrants arrive in the host countries with suboptimal or poor health, posing a public health challenge.[8] Close examination of epidemiological data reveals particular burdens of disease in CALD communities throughout many traditional immigration countries.[9-11] It has been suggested that migration from a developing to an industrialized/Westernized country has a detrimental impact on chronic disease risk factors as individuals from CALD groups adjust to both culture and modernization gaps.[5,12] Moreover, there is a consensus that among Western countries, significant racial and ethnic disparities exist in terms of morbidity and mortality, highlighting ª 2009 Adis Data Information BV. All rights reserved.

Caperchione et al.

higher rates of risk factors for a number of chronic diseases in such populations.[13,14] Of particular concern is the increased prevalence of hypertension, diabetes mellitus and overweight/ obesity in CALD populations, all of which are predominant risk factors for cardiovascular disease (CVD).[13] For example, Asian men in the UK appear to be more prone to coronary heart disease than others, and both men and women of South Asian origin have 30–40% higher coronary disease mortality rates than UK-born individuals.[15,16] Similar findings have been reported in New Zealand, where there is a high prevalence of risk factors for lifestyle diseases including CVD and diabetes in older Asian Indians[17] and older Tongan adults[18] living in urban settings compared with individuals born in New Zealand. Despite the increased risk of hypertension, diabetes, overweight/obesity and CVD, individuals from CALD groups are less likely to be proactive in accessing health care or undertaking preventative measures to ensure optimal health outcomes.[14,19] Although it has been well documented that engaging in preventive measures such as regular physical activity is associated with reduced risk of CVD and other chronic diseases,[20] it is evident that individuals from CALD backgrounds are less likely than others to participate in such activities.[10,17,21,22] For many CALD individuals there are several constraints on activity participation beyond personal motivation. Language barriers, socioeconomic factors, psychological trauma relating to migration and alternative health-seeking behaviours are just a few of the constraints that are likely to have a detrimental impact on health in these populations.[14,23,24] In an attempt to limit these constraints and positively influence the physical activity behaviours of CALD individuals, it is necessary to carefully consider cultural diversity whilst developing and planning health promotion (e.g. physical activity), resources and programmes. The limited nature of research in this area is evident.[9,25,26] The challenge, therefore, is to understand more about the influence of the migration process on the physical activity behaviours of CALD groups. Such information will support the development of culturally appropriate programmes designed to Sports Med 2009; 39 (3)

Physical Activity in Migrant Groups to Western Society

positively influence the physical activity behaviours of individuals from CALD populations.[24,27] The purpose of this paper is to review literature that outlines the barriers, challenges and enablers of physical activity for people in CALD groups who have recently migrated to Western society, and to make recommendations and propose strategies that may positively affect physical activity participation in these populations. 1. Methods For the purpose of this review, individuals from CALD groups refers to individuals from non-English-speaking backgrounds (those born overseas, those who speak a language other than English at home, or those who have little proficiency in speaking English) and individuals who identify with or have a social orientation towards a non-English speaking culture.[6,7] Indigenous populations are not clearly defined by these variables and thus were not included in this review. Moreover, the physical activity behaviours of indigenous populations differ to those of CALD groups and are reported elsewhere in the research literature.[28-30] Our search strategy included an exploration of publications from a number of electronic databases as well as manual searching of reference lists and relevant texts. The electronic databases used included Scopus, MEDLINE, ProQuest, CINAHL, ScienceDirect, PsycINFO and Google Scholar. These electronic databases were explored using the search terms ‘physical activity’, ‘physical activity barriers’, ‘cultural diversity’, ‘acculturation’, ‘immigration’, ‘migration’, ‘cultural sensitivity’, ‘ethnic groups’, ‘ethnicity’, ‘refugees’ and ‘westernisation’. Specifically, the terms ‘physical activity’ and ‘physical activity behaviours’ were combined with each of the other terms listed above. These search terms were chosen based on previous research literature associated with migration and sociocultural issues. Publications were selected for the review if they met the following criteria: (i) must be in the English language; (ii) must include CALD adults (aged ‡18 years); (iii) must refer to the migration/immigration process; and (iv) must ª 2009 Adis Data Information BV. All rights reserved.

169

discuss physical activity behaviours, barriers or enablers. From the initial search strategy, publications were excluded if they did not meet the inclusion criteria, as determined by the title and abstract. After this initial assessment, 120 publications (i.e. articles, articles from reference lists, relevant texts) potentially met the criteria. These publications were further screened and a total of 57 publications were identified as meeting the inclusion criteria, and thus included in this review. 2. Results Findings from the review identified key issues associated with the physical activity behaviours of those who belong to CALD groups. These key issues include, but are not limited to, the migration process and health status of these groups, barriers to physical activity as a risk factor for disease in these particular groups, and specialized strategies for increasing physical activity behaviours in CALD groups. Each of these key issues is discussed below. 2.1 Migration Process and Health Status of Culturally and Linguistically Diverse (CALD) Groups

Literature pertaining to the health status of these groups at the time of migration is inconsistent, claiming that while many migrants are in good physical and mental health when entering a new host country,[8,10,31,32] others are considered to be in relatively poor health.[5,33-35] Many researchers have suggested that many migrants are in good health upon arrival due to the lifestyle behaviours in which they are traditionally engaged (e.g. work that is physically demanding, walking for transport, eating foods high in fibre and low in fat) in their country of origin.[22,36,37] Other researchers have argued that migrants, particularly those entering as humanitarian refugees, are highly likely to be in poor health as a consequence of past deprivation and prolonged periods of suboptimal diet while living in their country of origin.[5,34] Furthermore, the poor physical health of these humanitarian Sports Med 2009; 39 (3)

Caperchione et al.

170

refugees is accompanied by stress, anxiety and overall poor mental health as a result of being subject to traumatic events such as physical and psychological violation and torture.[26,38-40] Although the literature is equivocal on the issue of health status of migrants entering a new host country, the literature favours a ‘healthy immigrant effect’.[8,31,41-44] A ‘healthy immigrant effect’ exists where migrants from non-Western countries are in generally very good health on arrival to a Western country; however, this condition erodes with increased time since migration, and is associated with what is commonly referred to as ‘acculturation’.[10,45] Evenson et al.[46] defined acculturation as changes in cultural patterns when groups of individuals from different cultures come into first-hand continuous contact with each other. Acculturation, which is part of the migration process, is often associated with the adoption of detrimental Western behaviours such as the consumption of a high-fat, calorie-dense diet, smoking, alcohol intake and a more sedentary lifestyle.[47-50] The adoption of these behaviours by migrant populations serve as risk factors for a number of chronic diseases including hypertension, diabetes, cardiovascular disease, obesity and poor mental health.[27,37,51-54] In a recent review, Steffen et al.[51] concluded that immigrants to the US and Europe from Africa, Asia, Latin America and Polynesia have consistently shown higher blood pressure with increasing levels of acculturation to Western society, whilst Nakanishi et al.[55] described the impact of lifestyle Westernization on diabetes mellitus in Japanese people migrating to America. Whilst all of these risk factors are important, for the purpose of this paper, discussion is limited to physical inactivity as a risk factor of disease. Poor diet/nutrition, smoking and excess alcohol intake have been discussed elsewhere.[56-60] 2.2 Barriers to Physical Activity as a Risk Factor for Disease in CALD Groups

There are a number of contributing factors inherent in the decreasing levels of physical activity in CALD groups. Specific to individuals ª 2009 Adis Data Information BV. All rights reserved.

from CALD groups, the literature has indicated that many of these barriers fit into common themes such as cultural and religious issues, issues of social relationships, socioeconomic challenges, environmental barriers and perceptions of health and injury (see table I). 2.2.1 Cultural and Religious Barriers

Cultural and religious barriers vary amongst different CALD groups, as some are quite generic and common to many individuals, while others are unique to the individual and require culturalspecific attention.[26] For example, the Muslim community exemplifies the need for cultural sensitivity as many of their religious practices and traditions have a particular influence on their physical activity behaviours. Guerin et al.[61] and Rogerson and Emes[26] reported that Muslim men and women arriving from Arabic-speaking countries (e.g. Morocco, Mauritania, Algeria, Tunisia, Libya, Sudan, Egypt and Somalia) traditionally pray five times a day and observe the month of Ramadan. Times of prayer can be a constraint, as all activity must stop and organizing a time for regular physical activity has to accommodate this religious activity. Also, during the month of Ramadan, Muslims fast from first light until sundown, abstaining from food and drink. As it is essential that individuals are properly hydrated and have energy stores to perform physical activity, fasting for Ramadan also becomes a constraint in terms of maintaining a correct energy balance for these particular CALD groups. Sex appropriateness was another common cultural barrier cited in the literature. Many CALD groups specified that physical activity for women is not encouraged, as these women are often expected to spend their time at home looking after their immediate and extended families.[24,64] Many individuals of Muslim faith interpret scriptures of the Quran as prohibiting physical activity participation for women,[65] while others allow participation to occur only if it does not conflict with their family responsibilities, if they engage in female-only programmes that ensure that men will not see them or be active Sports Med 2009; 39 (3)

Physical Activity in Migrant Groups to Western Society

171

Table I. Summary of barriers to physical activity (PA) in culturally and linguistically diverse groups Study

Participant characteristics

Barriers to PA

Kolt et al.[18]

24 Tongans living in Auckland, New Zealand Sex: 12 male, 12 female Age: 60–79 y

Lack of education about benefits of PA Lack of motivation Safety of physical environment Lack of facilities and transport Family commitments Physical and health limitations Cultural barriers

Kalavar et al.[22]

10 Asian Indian migrants in the US Sex: five male, five female Age: 66–79 y

Health problems Fear of injury Physical infrastructure Weather Incompatibility of PA and old age Inexperience with PA in the past Lack of interest and motivation Lifestyle in the US Lack of time due to family and domestic commitments Not having someone to be active with

Belza et al.[23]

71 migrant adults from seven cultures (American Indian/Alaskan Native, African American, Filipino, Chinese, Latino, Korean, Vietnamese) Sex: 29 male, 42 female Age: 52–85 y

Lack of motivation associated with low self-esteem to be PA Feeling disconnected and culturally isolated Weather Neighbourhood safety, crime Programme costs Lack of access and transport Family and work obligations Not having someone to be active with Lack of culture-specific activities

Lawton et al.[24]

32 South Asian migrants living in the UK (23 Pakistani, 9 Indian) Sex: 15 male, 17 female Age: 40–79 y

Lack of time due to family obligations Unfamiliarity of neighbourhood Fear and shame Lack of culturally sensitive facilities Weather Religious fatalism Negative perceptions of PA and disease

Rogerson and Emes[26]

Qualitative review paper Example of Muslim migrants Sex: not reported Age: older adults

Lack of affordable transport Lack of access to programmes Use of technical language used delivery activity instructions Poor health/pain Weather Cultural insensitivity

Barnes and Almasy[36]

31 adult migrants living in the US (11 Bosnians, 10 Cubans, 10 Iranians) Sex: 50/50 male/female Age: 19–71 y All spoke first language at home 90.3% were married 58% employed in low paying jobs

Lack of access to PA facilities Unfamiliarity with the physical environment Neighbourhood safety

Evenson et al.[46]

671 Latin migrant women living in US Age: 20–50 y Poor English language levels Low household income Low levels of employment

Language barriers Religious fatalism Lack of transportation due to not having a drivers license

Continued next page

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (3)

Caperchione et al.

172

Table I. Contd Study

Participant characteristics

Barriers to PA

Guerin et al.[61]

27 Somali women living in Hamilton, New Zealand Age: 17–67 y 75% had children Majority unemployed and on government benefit plans

Lack of childcare Cultural/religious insensitivity Lack of transport Safety Language barriers Lack of financial stability

Amesty[62]

Qualitative review paper Hispanics living in inner city Philadelphia, PA, USA

Lack of knowledge of benefits of PA associated with low education levels Not having someone to exercise with Cultural/social isolation Safety and crime Lack of access to PA facilities Poor health

Hayes et al.[63]

653 South Asian migrants living in UK (Indian, Pakistani, Bangladeshi) Sex: 320 men, 333 female Age: 25–75 y Low socioeconomic status

Cultural attitudes and values Perceptions of illness associated with being PA Modesty and avoidance of mixed sex activity Fear of racism Lack of finances Lack of transport

with them, and if they are appropriately dressed at all times.[5,24,53,61] Language acculturation can also be a barrier to physical activity, as many individuals migrate with an inability to read or write in their own language. This problem is exacerbated with the need to become literate in the language of the host country.[23,50] In two recent studies looking at physical activity programmes for refugee Somali women in New Zealand, participants found it difficult to follow activity instructions from an instructor or from programme manuals, even though interpreters were made available as much as possible.[53,61] Although subtle, religious fatalism has also been recognized as a common cultural and religious barrier to physical activity for some CALD groups. In a study of South Asian migrants living in Britain, Lawton et al.[24] reported that fatalistic notions of health, illness and death that were pre-ordained by Allah/God acted as a barrier, as participants did not perceive that physical activity could help reduce the risk of disease or death, but rather they believed that their religious fate was in the hands of Allah/God. Evenson et al.[46] reported a similar tendency in ª 2009 Adis Data Information BV. All rights reserved.

Latin migrants in North Carolina, who believe that prayer alone can help them stay healthy, that their health is in God’s hands and their future is out of their control; therefore, they see no need to engage in preventive health measures such as physical activity.

2.2.2 Social Relationships

The literature suggests the lack of social support and the prevalence of isolation is high amongst individuals from CALD groups and has a detrimental effect on their physical activity behaviours.[33,35,50] Belza et al.[23] examined the physical activity behaviours of individuals from multiple cultures (including Korean, Latino, Filipino and Vietnamese migrants), and reported a common theme in which migrants from these multiple cultures expressed feelings of isolation. Amesty,[62] who investigated barriers to physical activity in Hispanic migrants living in the US, found that migration to a different country, separation from family and friends, and loss of social capital had a profound influence on their physical activity behaviours. Sports Med 2009; 39 (3)

Physical Activity in Migrant Groups to Western Society

2.2.3 Socioeconomic Challenges

Socioeconomic factors, such as low education and literacy levels, poverty status and lack of access in general, play an adverse role in the physical activity behaviour of migrants from CALD groups.[18,27,48,63] For example, it has been proposed that migrants and refugees coming to Western countries as a consequence of economic hardships, war and displacement are usually quite poor and as a result are forced to move to residential areas that are deteriorating, and into houses that have been worn out and abandoned by others.[62] As a consequence, incidence of disease tends to rise, education and literacy levels are minimal and there is a lack of general resources (particularly with transportation), all of which affect participation in physical activity.[26,33,35]

2.2.4 Environmental Barriers

Safety was most often recognized as a major barrier to physical activity participation amongst women from CALD groups. Amesty[62] and Eyler et al.[66] stated that areas of high crime and violence found in low socioeconomic neighbourhoods where CALD migrants commonly resided was a reason why many women were not active. Lawton et al.[24] also suggested that women from CALD groups are fearful of being physically active due to lack of familiarity with their local neighbourhood, leaving them feeling vulnerable when they leave their house, which is compounded by their difficulties with communicating in English. A change in climate may also act as a barrier to physical activity, significantly affecting migrants who originally arrive from countries with very warm and dry climates to countries with varied weather and climate conditions.[26] For example, migrants from Vietnam living in the Seattle area reported the cold weather as being very problematic when attempting to exercise, as they perceived that the cold weather made it hard for them to breathe.[23] Additionally, Asian Indians living in New Jersey viewed snow as a barrier to activity because of a fear of falling and subsequent injury.[22] ª 2009 Adis Data Information BV. All rights reserved.

173

2.2.5 Perceptions of Health and Injury

Another commonly reported barrier to activity was migrants’ perceptions of ill health and injury associated with being physically active.[26,63] Rather than seeing sweating, increased heart rate and breathlessness as ‘normal’ byproducts of physical activity, individuals of Pakistani and Indian origin who recently migrated to the UK perceived these as illness states and something to be avoided.[24] Kalavar et al.[22] also noted concerns of injury from falling and overexertion while participating in physical activities for Asian Indian adults. 2.3 Strategies for Increasing Physical Activity Levels in CALD Groups

The research literature has proposed a number of strategies that may assist with overcoming the challenges and barriers faced by many CALD migrants. One particular approach highlighted by many was the need for cultural sensitivity on the part of the health professional.[31,53,67] It is imperative that health professionals who work with CALD groups acknowledge cultural diversity, display a caring attitude, and place the individual at the centre of programme development in an attempt to respond to their specific needs and deliver a programme that encourages participation and respects the culture of the participant.[26,68] For example, Guerin et al.[61] suggested that adapting activity times to work around prayer and Ramadan and blocking windows in a manner so that men are not able to see Muslim women perform activities are just a few responses to cultural competence. Furthermore, culture-specific tailoring of physical activity messages, resources (social and environmental structures), and programmes must also be considered for accommodating the specific needs of the different CALD groups. The beliefs and perceptions of the benefits of physical activity on disease vary amongst many CALD migrants. While some acknowledge the influence that physical activity has on the risk of disease, others hold fatalistic beliefs and do not fully understand the benefits of engaging in healthy behaviours such as physical activity.[46,54] Sports Med 2009; 39 (3)

174

Providing detailed yet simple education sessions that utilize tools such as curriculum-based books and DVDs (nutrition and physical activity information), as well as interpreters, has been suggested as a strategy to assist CALD groups with learning about the benefits of preventive health measures.[5,33] Additionally, health professionals should consider partnering these sessions with learning centres/organizations who teach English to these migrants.[23] Consistently, the literature has addressed the importance of social support and social networks to migrant groups when they first arrive to their new host country.[5,22,46] Designing programmes that promote and encourage the participation of individuals from the same culture will provide migrants with the support of others who are familiar with their culture-specific traditions and customs, who have had similar migration experiences, and who are also moving through the acculturation process.[5,36] Investigation of other mediums for physical activity, such as occupational physical activity, is also supported by the literature.[31,69] Wolin et al.[50] found that leisure time physical activity was low in a multiethnic group of working class migrants, but reported that these individuals were more likely to do active work and thus they may be meeting activity recommendations through non-leisure, occupational activity. Future research should explore employment situational variables for promoting physical activity in migrant groups, as many of them are employed in labour-intensive jobs.[45] In addition to the above considerations and strategies, there is a general consensus that the provision of resources, in particular childcare and transportation, are essential strategies to overcoming the barriers of physical activity participation in CALD migrant groups.[5,26,61] Childcare is one of the most cited reasons for not attending community centre classes and fitness centres,[61] and the implementation of a childminding scheme within the community has been identified as a particularly effective strategy to assist CALD women to attend exercise classes.[61] Furthermore, the identification of existing schemes and respite programmes that are in place to assist ª 2009 Adis Data Information BV. All rights reserved.

Caperchione et al.

new migrants with childcare and other factors such as food shopping and transportation has also been highlighted.[27,70] In addition to providing transportation through community services, the literature describes an alternative strategy for overcoming transportation barriers. Guerin et al.[61] described the notion of bringing the programmes to the CALD groups. As most migrants reside in the same geographic area in their new host country,[62] developing programmes in these areas, which the participants can walk to, would help to alleviate transportation difficulties. The further development of ‘health action zones’, which have been introduced in the UK, provides economical resources to deprived communities, allowing these communities to offer healthcare and health resources to migrants without having to leave their residential areas.[33] 3. Conclusions and Future Directions The literature has suggested that the migration to Western societies has a detrimental effect on the health status and health behaviours of CALD groups as they assimilate to their new surroundings, explore different cultures and customs, and embrace a new way of life. In particular, there is evidence that physical inactivity is common in migrant CALD groups, and is a key contributing risk factor to chronic disease for these individuals.[10,17,71] Clearly, there are particular barriers and challenges associated with the decreased physical activity levels, yet research has also recognized possible strategies for overcoming these barriers and challenges.[8,23,26,62] Despite the significance of this literature, there are a number of gaps that do exist and should be considered when undertaking further research with migrant CALD groups. Although it appears that migration from a non-Western society to a Western society has an effect on the health status and activity levels of these migrant CALD groups, few studies have directly examined the association between the stages of the migration process and its effect on the physical activity behaviours of CALD populations. In particular, a search of the literature Sports Med 2009; 39 (3)

Physical Activity in Migrant Groups to Western Society

failed to identify any studies addressing the activity levels of CALD populations in their country of origin and how this may or may not affect their level of physical activity once they have migrated to their new host country. Given that the migration process and the cultural changes associated with it appear to influence health status and preventative health behaviours, the assessment of this process and the cultural changes associated with it may be worthy areas for future research. Another limitation is that the majority of the literature has been cross-sectional in nature. Limited data are available on how migration to Western society influences the health status and physical activity behaviours of CALD populations over time, making it difficult to infer causation. Future studies would benefit from controlled intervention trials and longitudinal designs that carefully assess the prevalence of lifestyle disease risk factors associated with the migration process and the effects that physical activity has on the incidence of these risk factors and/or chronic disease. There may also be some measurement issues associated with assessing the physical activity perceptions and behaviours of migrant CALD groups. CALD groups represent a variety of nationalities and ethno-cultural groups, and as such, the meanings and perceptions of physical activity, and physical activity measures used in research, are likely to vary. In consideration of this, valid and reliable instruments for assessing physical activity must be specifically developed and tested to suit each cultural group. Furthermore, given that many of these cultures do not read or write in their own language or in the English language, alternative delivery modes of self-report physical activity need to be developed and tested. Objective measures of physical activity should be considered in these groups to help overcome some of these barriers. Keeping in mind that perceptions of physical activity do differ between cultures, it is necessary that the research examines the educational components of health promotion programmes, paying close attention to the exchange of information pertaining to the physiological and psychological ª 2009 Adis Data Information BV. All rights reserved.

175

benefits of physical activity. The educational component should occur in the initial stages of the migration process and should address the lifestyle behaviour changes (e.g. physical activity and diet) resulting from cultural changes, and the direct effect that these behaviours have on chronic disease. Lastly, the bulk of the literature is based on data from the US, with some studies also originating in the UK. Although this research has made a valuable contribution to the literature pertaining to the lifestyle behaviours of migrant CALD groups, further research needs to be undertaken with CALD populations who have migrated to other Western societies such as Australia and New Zealand. As there may be differences when migrating to these different Western societies, it would be important to see some comparative literature describing these different experiences. Acknowledgements This research was funded by a research grant from the Office for Women, Department of Families, Community Services and Indigenous Affairs made to CM Caperchione and KW Mummery. The authors have no conflicts of interest that are directly relevant to the contents of this article.

References 1. United Nations Department of Economic and Social Affairs. World economic and social survey: trends and policies in the world economy. New York: United Nations Publications, 2001 2. Commonwealth of Australia. Population flows: immigration aspects. Canberra (ACT): Department of Immigration and Multicultural Affairs, 2007 3. Australian Bureau of Statistics. 2007 Australian yearbook (cat. no. 1301.0). Canberra (ACT): Commonwealth of Australia, 2007 4. Australian Bureau of Statistics. 2006 Census of population and housing: media releases and fact sheets. Canberra (ACT): Commonwealth of Australia, 2006 5. Burns C, Webster K, Crotty P, et al. Easing the transition: food and nutrition issues of new arrivals. Health Prom J Australia 2000; 10 (3): 230-6 6. Health Department of Western Australia. Cultural competency standard and self-assessment audit tool manual. Perth (WA): Government of Western Australia, 2005 7. Commonwealth Interdepartmental Committee on Multicultural Affairs. The guide: implementing the standards

Sports Med 2009; 39 (3)

Caperchione et al.

176

8.

9.

10. 11.

12.

13.

14.

15. 16.

17.

18.

19.

20.

21.

22.

23.

24.

for statistics on cultural and language diversity. Canberra (ACT): Department of Immigration and Multicultural Affairs, 2001 Carrington K, McIntosh A, Walmsley J. The social costs and benefits of migration into Australia. Canberra (ACT): Commonwealth of Australia, 2007 Adily A, Ward J. Improving health among culturally diverse subgroups: an exploration of trade-offs and viewpoints among a regional population health workforce. Health Promot J Austr 2005 Dec; 16 (3): 207-12 Perez CE. Health status and health behaviour among immigrants. Health Reports 2002; 13 (Suppl): 89-100 Antecol H, Bedard K. Unhealthy assimilation: why do immigrants converge to American health status levels? Demography 2006; 43 (2): 337-60 Abate N, Chandalia M. Ethnicity and type 2 diabetes: focus on Asian Indians. J Diabetes Complications 2001 Nov-Dec; 15 (6): 320-7 Johnson PA, Fulp RS. Racial and ethnic disparities in coronary heart disease in women: prevention, treatment, and needed interventions. Womens Health Issues 2002 Sep-Oct; 12 (5): 252-71 Davidson PM, Daly J, Hancock K, et al. Australian women and heart disease: trends, epidemiological perspectives and the need for a culturally competent research agenda. Contemp Nurse 2003 Dec-2004 Feb; 16 (1-2): 62-73 Chaturvedi N. Ethnic differences in cardiovascular disease. Heart 2003 Jun; 89 (6): 681-6 Dhanjal TS, Lal M, Haynes R, et al. A comparison of cardiovascular risk factors among Indo-Asian and Caucasian patients admitted with acute myocardial infarction in Kuala Lumpur, Malaysia and Birmingham, England. Int J Clin Pract 2001 Dec; 55 (10): 665-8 Kolt G, Schofield G, Rush E, et al. Body fatness, physical activity, and nutritional behaviours in Asian Indian immigrants to New Zealand. Asia Pac J Clin Nutr 2007 Kolt G, Paterson J, Cheung Y. Barriers to physical activity participation in older Tongan adults living in New Zealand. Australasian J Ageing 2006; 25 (3): 119-27 Gettleman L, Winkleby MA. Using focus groups to develop a heart disease prevention program for ethnically diverse, low-income women. J Community Health 2000 Dec; 25 (6): 439-53 Bauman A. Updating the evidence that physical activity is good for health: an epidemiological review 2000-2003. J Sci Med Sport 2004; 7 (1 Suppl.): 6-19 Australian Bureau of Statistics. Migrants and participation in sport and physical activity 2006. Canberra (ACT): National Centre for Culture and Recreation Statistics, 2006 Kalavar J, Kolt G, Giles L, et al. Physical activity in older Asian Indians living in the United States: barriers and motives. Activ Adapt Aging 2004; 29 (1): 47-67 Belza B, Walwick J, Shiu-Thornton S, et al. Older adult perspectives on physical activity and exercise: voices from multiple cultures. Prev Chronic Dis 2004 Oct; 1 (4): A09 Lawton J, Ahmad N, Hanna L, et al. ‘I can’t do any serious exercise’: barriers to physical activity amongst people of

ª 2009 Adis Data Information BV. All rights reserved.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34. 35.

36. 37. 38.

39.

40.

41. 42.

Pakistani and Indian origin with type 2 diabetes. Health Educ Res 2006 Feb; 21 (1): 43-54 Kreuter MW, Lukwago S, Bucholtz D, et al. Achieving cultural appropriateness in health promotion programs: targeted and tailored approaches. Health Educ Behav 2002; 30 (2): 133-46 Rogerson M, Emes C. Physical activity older immigrants and cultural competence: a guide for fitness practitioners. Activ Adapt Aging 2006; 30 (4): 15-27 Dawson AJ, Sundquist J, Johansson SE. The influence of ethnicity and length of time since immigration on physical activity. Ethn Health 2005 Nov; 10 (4): 293-309 Gracey M, Bridge E, Martin D, et al. An Aboriginal-driven program to prevent, control and manage nutrition-related ‘‘lifestyle’’ diseases including diabetes. Asia Pac J Clin Nutr 2006; 15 (2): 178-88 Rowley KG, Daniel M, Skinner K, et al. Effectiveness of a community-directed ‘healthy lifestyle’ program in a remote Australian aboriginal community. Aust N Z J Public Health 2000 Apr; 24 (2): 136-44 Thompson SJ, Gifford SM, Thorpe L. The social and cultural context of risk and prevention: food and physical activity in an urban Aboriginal community. Health Educ Behav 2000 Dec; 27 (6): 725-43 Lindstrom M, Sundquist J. Immigration and leisure-time physical inactivity: a population-based study. Ethn Health 2001 May; 6 (2): 77-85 Berrigan D, Dodd K, Troiano RP, et al. Physical activity and acculturation among adult Hispanics in the United States. Res Q Exerc Sport 2006 Jun; 77 (2): 147-57 Hjelm K, Sundquist J, Apelquist J. The influence of socioeconomic status and life style on self-reported health in diabetics and non-diabetics: a comparison of foreign-born and Swedish-born individuals. Prim Health Care Res and Develop 2002; 3: 249-59 Ackerman LK. Health problems of refugees. J Am Board Fam Pract 1997 Sep-Oct; 10 (5): 337-48 Sundquist J, Iglesias E, Isacsson A. Migration and health: a study of Latin American refugees, their exile in Sweden and repatriation. Scand J Prim Health Care 1995 Jun; 13 (2): 135-40 Barnes DM, Almasy N. Refugees’ perceptions of healthy behaviors. J Immigr Health 2005 Jul; 7 (3): 185-93 Brown PJ. Culture and the evolution of obesity. Human Nature 1991; 2 (1): 31-57 Richardson S, Miller-Lewis L, Ngo P, et al. The settlement experiences of new migrants: a comparison of wave one of LSIA 1 and LSIA 2. Canberra (ACT): DIMIA, 2002 Vanden Heuvel A, Wooden M. New settlers have their say: how immigrants fare over the early years of settlement. Canberra (ACT): DIMIA, 1999 Choudhry UK. Uprooting and resettlement experiences of South Asian immigrant women. West J Nurs Res 2001 Jun; 23 (4): 376-93 Australian Institute of Health and Welfare. Australia’s Health 2004. Canberra (ACT): AIHW, 2004 Williams R. Health and length of residence among south Asians in Glasgow: a study controlling for age. J Public Health Med 1993 Mar; 15 (1): 52-60

Sports Med 2009; 39 (3)

Physical Activity in Migrant Groups to Western Society

43. Singh GK, Miller BA. Health life expectancy, and mortality patterns among immigrant populations in the United States. Can J Public Health 2004 May-Jun; 95 (3): I14-21 44. Fennelly K. The ‘‘healthy migrant’’ effect. Minn Med 2007 Mar; 90 (3): 51-3 45. Tremblay MS, Bryan SN, Perez CE, et al. Physical activity and immigrant status: evidence from the Canadian Community Health Survey. Can J Public Health 2006 Jul-Aug; 97 (4): 277-82 46. Evenson KR, Sarmiento OL, Ayala GX. Acculturation and physical activity among North Carolina Latina immigrants. Soc Sci Med 2004 Dec; 59 (12): 2509-22 47. Abraido-Lanza AF, Chao MT, Florez KR. Do healthy behaviors decline with greater acculturation? Implications for the Latino mortality paradox. Soc Sci Med 2005 Sep; 61 (6): 1243-55 48. Jaber LA, Brown MB, Hammad A, et al. Lack of acculturation is a risk factor for diabetes in Arab immigrants in the US. Diabetes Care 2003 Jul; 26 (7): 2010-4 49. Hosper K, Nierkens V, Nicolaou M, et al. Behavioural risk factors in two generations of non-Western migrants: do trends converge towards the host population? Eur J Epidemiol 2007; 22 (3): 163-72 50. Wolin KY, Colditz G, Stoddard AM, et al. Acculturation and physical activity in a working class multiethnic population. Prev Med 2006 Apr; 42 (4): 266-72 51. Steffen PR, Smith TB, Larson M, et al. Acculturation to Western society as a risk factor for high blood pressure: a meta-analytic review. Psychosom Med 2006 May-Jun; 68 (3): 386-97 52. Schulz LO, Bennett PH, Ravussin E, et al. Effects of traditional and western environments on prevalence of type 2 diabetes in Pima Indians in Mexico and the U.S. Diabetes Care 2006 Aug; 29 (8): 1866-71 53. Guerin PB, Elmi FH, Corrigan C. Body composition and cardiorespiratory fitness among refugee Somali women living in New Zealand. J Immigr Minor Health 2007 Jul; 9 (3): 191-6 54. Reijneveld SA, Westhoff MH, Hopman-Rock M. Promotion of health and physical activity improves the mental health of elderly immigrants: results of a group randomised controlled trial among Turkish immigrants in the Netherlands aged 45 and over. J Epidemiol Community Health 2003 Jun; 57 (6): 405-11 55. Nakanishi S, Okubo M, Yoneda M, et al. A comparison between Japanese-Americans living in Hawaii and Los Angeles and native Japanese: the impact of lifestyle westernization on diabetes mellitus. Biomed Pharmacother 2004 Dec; 58 (10): 571-7 56. Gutmann MC. Ethnicity alcohol, and acculturation. Soc Sci Med 1999 Jan; 48 (2): 173-84 57. Loury S, Kulbok P. Correlates of alcohol and tobacco use among Mexican immigrants in rural North Carolina. Fam Community Health 2007 Jul-Sep; 30 (3): 247-56 58. Weber TR. The influence of acculturation on attitudes toward alcohol and alcohol use within the Punjabi

ª 2009 Adis Data Information BV. All rights reserved.

177

59.

60.

61.

62. 63.

64.

65.

66.

67.

68.

69.

70.

71.

community: an exploratory analysis. Subst Use Misuse 1996 Sep-Oct; 31 (11-12): 1715-32 Yang EJ, Chung HK, Kim WY, et al. Chronic diseases and dietary changes in relation to Korean Americans’ length of residence in the United States. J Am Diet Assoc 2007 Jun; 107 (6): 942-50 Lee WP, Lingard J, Bermingham M. Change in diet and body mass index in Taiwanese women with length of residence in Australia. Asia Pac J Clin Nutr 2007; 16 (1): 56-65 Guerin PB, Diiriye RO, Corrigan C, et al. Physical activity programs for refugee Somali women: working out in a new country. Women Health 2003; 38 (1): 83-99 Amesty SC. Barriers to physical activity in the Hispanic community. J Public Health Policy 2003; 24 (1): 41-58 Hayes L, White M, Unwin N, et al. Patterns of physical activity and relationship with risk markers for cardiovascular disease and diabetes in Indian, Pakistani, Bangladeshi and European adults in a UK population. J Public Health Med 2002 Sep; 24 (3): 170-8 Walseth K, Fasting K. Islam’s view on physical activity and sport: Egyptian women interpreting Islam. Int Rev Socio Sport 2003; 38 (1): 45-60 De Knop P, Theeboom M, Wittock H, et al. Implications of Islam on Muslim girls’ sport participation in western Europe: literature review and policy recommendations for sport promotion. Sport Educ Society 1996; 1: 147-64 Eyler AA, Baker E, Cromer L, et al. Physical activity and minority women: a qualitative study. Health Educ Behav 1998 Oct; 25 (5): 640-52 Cooper RS, Rotimi CN, Kaufman JS, et al. Prevalence of NIDDM among populations of the African diaspora. Diabetes Care 1997 Mar; 20 (3): 343-8 Morioka-Douglas N, Sacks T, Yeo G. Issues in caring for Afghan American elders: insights from literature and a focus group. J Cross Cult Gerontol 2004 Mar; 19 (1): 27-40 Eyler AA, Brownson RC, Donatelle RJ, et al. Physical activity social support and middle- and older-aged minority women: results from a US survey. Soc Sci Med 1999 Sep; 49 (6): 781-9 Stewart MJ, Neufeld A, Harrison MJ, et al. Immigrant women family caregivers in Canada: implications for policies and programmes in health and social sectors. Health Soc Care Community 2006 Jul; 14 (4): 329-40 Mahajan D, Bermingham MA. Risk factors for coronary heart disease in two similar Indian population groups, one residing in India, and the other in Sydney, Australia. Eur J Clin Nutr 2004 May; 58 (5): 751-60

Correspondence: Dr Cristina M. Caperchione, Institute for Health and Social Science Research, CQUniversity Australia, Bruce Highway, Rockhampton, QLD 4702, Australia. E-mail: [email protected]

Sports Med 2009; 39 (3)

Sports Med 2009; 39 (3): 179-206 0112-1642/09/0003-0179/$49.95/0

REVIEW ARTICLE

ª 2009 Adis Data Information BV. All rights reserved.

Physiological Differences Between Cycling and Running Lessons from Triathletes Gregoire P. Millet,1 V.E. Vleck2 and D.J. Bentley3 1 ISSEP, University of Lausanne, Lausanne, Switzerland 2 School of Biosciences, University of Westminster, London, United Kingdom 3 Health and Exercise, School of Medical Science, University of New South Wales, Sydney, Australia

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Maximal Aerobic Power and the Anaerobic Threshold: Comparison between Cycling and Running . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Trained Cyclists and Runners. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1.1 Maximal Aerobic Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1.2 Submaximal Intensity/Anaerobic Threshold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Triathletes Performing Cycle Ergometry and Treadmill Running. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.1 Maximal Aerobic Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.2 Anaerobic Threshold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.3 Sex Differences. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.4 Heart Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.5 Running Economy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.6 Delta Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Physiological Mechanisms Associated with Differences between Cycling and Running. . . . . . . . . . . 2.1 Ventilatory Responses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Central and Peripheral Blood Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Skeletal Muscle Oxidative Capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Central and Peripheral Innervation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.1 Muscle Recruitment Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.2 Pedalling Frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.3 Neuromuscular Fatigue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

179 181 181 181 181 185 185 185 191 191 192 194 194 194 195 196 197 197 198 199 200

The purpose of this review was to provide a synopsis of the literature concerning the physiological differences between cycling and running. By comparing physiological variables such as maximal oxygen consumption . (VO2max), anaerobic threshold (AT), heart rate, economy or delta efficiency measured in cycling and running in triathletes, runners or cyclists, this review aims to identify the effects of exercise modality on the underlying mechanisms (ventilatory responses, blood flow, muscle oxidative capacity, peripheral innervation and neuromuscular fatigue) of adaptation. The majority of studies . on treadmill whereas cyclists indicate that runners achieve a higher VO 2max . can achieve a VO2max value in cycle ergometry similar to that in treadmill

Millet et al.

180

. running. Hence, VO2max is specific to the exercise modality. In addition, the muscles adapt specifically to a given exercise task over a period of time, resulting in an improvement in submaximal physiological variables such as . the ventilatory threshold, in some cases without a change in VO2max. However, this effect is probably larger in cycling than in running. At the same time, skill influencing motor unit recruitment patterns is an important influence on the anaerobic threshold in cycling. Furthermore, it is likely that there is more physiological training transfer from running to . cycling than vice versa. In triathletes, there is generally no difference in VO2max measured in cycle ergometry and treadmill running. The data concerning the anaerobic threshold in cycling and running in triathletes are conflicting. This is likely to be due to a combination of actual training load and prior training history in each discipline. . The mechanisms surrounding the differences in the AT together with VO2max in cycling and running are not largely understood but . are probably due to the relative adaptation of cardiac output influencing VO2max and also the recruitment of muscle mass in combination with the oxidative capacity of this mass influencing the AT. Several other physiological differences between cycling and running are addressed: heart rate is different between the two activities both for maximal and submaximal intensities. The delta efficiency is higher in running. Ventilation is more impaired in cycling than in running. It has also been shown that pedalling cadence affects the metabolic responses during cycling but also during a subsequent running bout. However, the optimal cadence is still debated. Central fatigue and decrease in maximal strength are more important after prolonged exercise in running than in cycling.

Exercise physiologists working with multisport athletes such as triathletes or duathletes who undergo both cycle and run training often use incremental exercise tests to monitor maximal oxygen . uptake (VO2max), the anaerobic threshold (AT) and related physiological variables for both cycling . and running. ‘Performance VO2max’ (i.e. how long a given rate of aerobic and anaerobic metabolism can be sustained) is determined by the interaction . between VO2max and lactate threshold (LT), whereas efficiency determines how much speed or power (i.e. ‘performance velocity’) can be achieved for a given amount of energy consumption.[1] However, these physiological variables measured in either cycling or running may adapt indifferently as a consequence of cross training in cycling and running;[2-13] cross training is defined as ‘combined alternative training modes within a sport specific regimen’. The physiological adaptations to cross training in previously untrained athletes have been reviewed.[2,3] ª 2009 Adis Data Information BV. All rights reserved.

It is also possible that the results of such physiological tests in cycling and running may be influenced by the athlete’s original training background. However, no in-depth analysis has been carried out examining the differences. in selected physiological parameters such as VO2max and the AT, measured in cycling and running in athletes participating in either cycling or running or in triathletes actively participating in both cycling and running. This review aims to provide a synopsis of the literature concerning the mechanisms associated with physiological differences between cycling and running. This article is not intended to review performance in triathlons or the physiological characteristics of triathletes. The comparison between triathletes, runners and cyclists is shown to identify the training acclimation triggered by mode of exercise.. By comparing physiological variables such as VO2max, AT, heart rate, economy or delta efficiency measured in cycling Sports Med 2009; 39 (3)

Physiological Differences Between Cycling and Running

and running in triathletes, runners or cyclists, we aim to identify the effects of exercise modality on the underlying mechanisms (ventilatory responses, blood flow, muscle oxidative capacity, peripheral innervation, neuromuscular fatigue) of adaptation. 1. Maximal Aerobic Power and the Anaerobic Threshold: Comparison between Cycling and Running 1.1 Trained Cyclists and Runners 1.1.1 Maximal Aerobic Power

. Whole body VO2max typically reflects a heightened capacity for cellular oxygen utiliza[14-16] The accepted tion within a specific activity. . hypothesis concerning VO2max is that the greater the muscle mass used in . an incremental task, the higher the measured VO2max.[17-19] For. example, Gleser et al.[17] found an ~10% higher VO2max for arm cranking in combination with lower limb muscle contractions compared with arm cranking alone. . Other authors have reported no difference in VO2max when upper limb activity is compared with lower limb and upper limb work.[14,20] It is generally accepted that in exercise situations involving a greater muscle mass, as in running, the . higher the VO2max value is observed in untrained subjects. Table I shows the studies that have reported maximal oxygen uptake for cycling and running in cyclists and runners.[5-7,14,18,21-36] However, direct comparison between studies is difficult since it is known that the incremental test protocol design (e.g. starting work rate, increments and duration of each stage) or the analysis performed on the data (e.g. exhaustion criteria, mathematical modeling) can greatly influence . the values of physiological variables such as VO2max or maximal aerobic power so obtained.[37,38] Pechar et al.[5] concluded that athletes with previous experience in cycling may exhibit . VO2max values that are either equal to or approach those obtained in treadmill . running. Various studies that have compared VO2max in cycling and treadmill running in trained cyclists and runners support their premise.[30,39] Stromme et. al.[39] reported a significantly higher (5.6%) VO2max in cycling compared with treadmill ª 2009 Adis Data Information BV. All rights reserved.

181

running in male elite cyclists. Ricci and Leger[40] . also found a higher VO2max in cycle ergometry when compared with treadmill running (62.4 – 8.1 vs 54.7 – 8.1 mL/kg/min). Bouckaert . et al.[34] later compared VO2max in cyclists and runners completing incremental treadmill and cycling activity. These authors reported the . VO2max was 14% higher in treadmill running compared with bicycle ergometry in runners and 11% higher on the bicycle ergometer than on the treadmill in cyclists. Moreira-da-Costa et al.[32] . found that VO2max was highest in the exercise mode that the athletes had trained. exclusively in. These authors reported a higher VO2max value in treadmill exercise for runners and cycle ergometry in cyclists, respectively. Other research groups . have found no significant difference in the VO2max obtained in an incremental cycle test when compared with an incremental running Thus well test in trained cyclists.[6,7,27,41,42] . trained cyclists can exhibit a VO2max similar to that observed in treadmill running. Therefore, a training history and its accompanying adapta. tions in cycling may elicit a VO2max value that is similar to that for treadmill running despite treadmill running potentially requiring a greater muscle mass. In only one study have trained . cyclists been shown to exhibit a higher VO2max value in treadmill running than in cycle ergo[8] a surprising result considering that metry; . the VO2max of the subjects was well developed (i.e. in excess of 70 mL/kg/min) for both such exercise modes. 1.1.2 Submaximal Intensity//Anaerobic Threshold

It is generally described that submaximal running exercise induces a higher oxygen uptake and, probably, energy expenditure than cycling at the same intensity.[43-45] However, since the postexercise oxygen consumption (EPOC) is similar in cycling and running[45] but the lactate concentration was shown to be higher after cycling at the same submaximal intensity,[29] one can speculate that the metabolic demands, i.e. the relative contribution of the glycolytic and oxidative processes, are different. In a recent experiment, Scott et al.[36] reported that during a short bout at the same intensity (1 minute at 250 W; treadmill Sports Med 2009; 39 (3)

PE

5M

R&C

5 M, 3 F

S

5M

None

6M

. Relative VO2max run (mL/kg/min)a

. Absolute VO2max run (L/min)a

30 – 5.2

76.4 – 4.7

4.5

4.7

[14]

High/national

23

69

4.6***

5.0***

[21] NDP

2–3 sessions/wk

22

69.0

4.0*

4.3*

25

72.1

3.2*

3.5*

R

10 M

High/national

23

67.4

4.6*

4.8*

R

14 M

Active

14–68

67.3

3.4*

3.7*

R

7M

2 mo running

43

73.7

3.4

3.4

S

4M

1 well trained

19–21

3.8 – 0.6

4.1 – 0.7

V

9M

Well trained

24–34

4.6 – 0.5*

4.8 – 0.4*

S

23 M

Trained

20.4 – 1.8

77.3 – 12.7

V

8M

29.6 – 7.6

77.5 – 4.8

V

3 M, 1 F

Pre C

19.1 – 3.9

20 M

38.5 – 4.3

73.3 – 9.9

Post C Pre R

19.6 – 1.4

20 M

72.3 – 6.2

Post R None

22.4 – 3.9

20 M

None

None

30 M

Low

6M

High

V

18 M 6F

[22]

3.3 – 0.5

[23] SF

3.4 – 0.5

3.8 – 0.7

[24] NDP

5.2*

4.8 – 1.4*

[18]

3.5 – 0.5

4.0 – 0.5

[5]

3.7 – 0.4#

4.1 – 0.5#

3.5 – 0.3

3.5 – 0.5

3.8 – 0.3#

3.7 – 0.4#

3.5 – 0.4

3.9 – 0.5

3.4 – 0.3

3.9 – 0.4

3.0 – 0.5

42.7 – 4.9

21.2 – 1.6

71.6 – 8.5

48.1 – 5.1

3.4 – 0.4

54.8 – 4.9

3.9 – 0.5

[25]

22.5 – 2.6

75.5 – 9

48.8 – 5.4

3.7 – 4.0

52.9 – 4.7

4.0 – 0.5

[26]

73.4 – 2

65 – 3.4

4.6 – 0.2

62.9 – 1.7

4.6 – 0.2

[27] SF

3.3 – 0.1***

[28] SF

1.9 – 0.2*

[29] NDP

2.7 – 0.1***

19–24 19.4 – 1.9

57.9 – 6.61

32.1 – 6.1*

1.8 – 0.3*

34.9 – 4.2*

Continued next page

Millet et al.

Sports Med 2009; 39 (3)

50 M

None

73.7 – 9.5

8 wk post

PE

C

Referenceb

182

ª 2009 Adis Data Information BV. All rights reserved.

Table I. Studies (n = 21) that have assessed maximal oxygen uptake for cycling and running in cyclists and runners . . Sport n/sex Level/details Age (y) Mass (kg) Absolute VO2max Relative VO2max a bike (mL/kg/min) bike (L/min)a

Sport

n/sex

Level/details

C

10 M

Steady training

R

10 M

None

10 M

Non-athletes

Age (y)

26 – 6

Mass (kg)

. Relative VO2max bike (mL/kg/min)a

. Absolute VO2max bike (L/min)a

. Relative VO2max run (mL/kg/min)a

. Absolute VO2max run (L/min)a

69.2 – 7

65.9 – 6.5

4.5 – 0.4

62.8 – 3.5

4.2 – 0.3*

68.4 – 8.5

61.7 – 6.2*

4.3

68.1 – 6.3*

4.6 – 0.4

72.8 – 11.8

40.0 – 5.5*

2.9 – 0.4*

45.1 – 3.9*

3.2 – 0.4*

Referenceb [6] H

[30] SF, DP, BLA, HR

R

12 M

National

22.7 – 4.0

61.7 – 8.7

61.9 – 4.9*

4.1 – 0.5*

57.3 – 4.5*

3.8 – 0.6*

C

10 M

National

20.6 – 2.1

67.3 – 9.2

56.5 – 5.8*

3.6 – 0.34*

64.3 – 7.3*

4.0 – 0.3*

12 M

No details

23 – 5

75 – 9

60 – 6*

66 – 8*

[32] BLA, DP, H90

R

10 M

C

9M

2 y specific training

22.7 – 4.0

61.7 – 8.7

3.5 – 0.4*

4.0 – 0.3*

20.6 – 2.1

67.3 – 9.2

4.2 – 0.6*

3.9 – 0.6*

46.7 – 1.5

V

[33]

56.9 – 1.8

R

9M

23.5 – 2.3

62.4 – 5.3

67.7 – 5.4*

4.2 – 0.4*

77.1 – 4.1*

4.8 – 0.4*

C

8M

22.3 – 2.7

68.1 – 5.4

75.5 – 8.2*

5.1 – 0.5*

68 – 4.1*

4.6 – 0.4*

PE

9M

PE students

22.3 – 1.2

69.8 – 5.8

53.5 – 3.2

3.7 – 0.3

54 – 4.6

3.8 – 0.5

R

7M

Low-middle

28.1 – 3.6

70.3 – 7.8

50.1 – 8.5*

59.6 – 8.3*

C

7M

24.3 – 4.6

70.9 – 11.2

55.1 – 7.2*

59.5 – 8.2* 2.9 – 0.7*

6 F, 6 M 13 M,1 F

b

28.6 – 9.6

77.5 – 14.2

57 – 12.9

[7]

3.1 – 0.8* 59.3 – 13.7

[34] H

[35] DP

[36]

Values are reported as mean – SD. Where the original study reported the standard error of the mean, the standard deviation was calculated using the formula SEMOn, where SEM is the standard error of the mean and n is the sample size. . Criteria for VO2max are presented as a superscript in this column.

BLA = blood lactate concentration > 8 mM; C = cycle trained; DP = defined plateau; F = females; H = highest averaged . value reached within last stage; H90 = HR ‡90% of age predicted maximum; H95 = HR ‡95% of age predicted maximum; HR = heart rate; M = males; NDP = .non-defined plateau in VO2 despite increase in speed or work rate; PE = physical education students; R = run trained; S = students; SF = subjective/volitional exhaustion; V = various; VO2max = maximal oxygen consumption. * p < 0.05: differences between running vs cycling, # p < 0.05: differences in the same sex, within subgroups, *** p < 0.05 cyclists < runners.

183

Sports Med 2009; 39 (3)

a

Active

[32]

Physiological Differences Between Cycling and Running

ª 2009 Adis Data Information BV. All rights reserved.

Table I. Contd

184

speed was calculated to elicit 250 W based on subject weight), total energy expenditure was similar between cycling and uphill running but the extent of aerobic and anaerobic energy transfer was different: the glycolytic component was 28% in cycling and 17% in running. If confirmed, this different metabolic contribution would directly influence the determination of the critical anaerobic threshold. The physiological mechanisms surrounding the AT are complex and may be influenced by both neurological and peripheral muscle factors.[46] In addition, there is no biochemical evidence to support the fact that lactate production causes acidosis.[47] It is beyond the scope of this article to present and discuss the numerous (19 invasive and 15 non-invasive) techniques that could be used to determine the AT (for reviews see Bentley et al.[37] and Loat and Rhodes[48]). Training of specific muscle groups may facilitate O2 transport and improve the metabolic profile of the specific muscles . involved in the training thereby affecting the VO2max as well as the AT. The AT can be expressed at an absolute work rate or as a percentage (%) of a maximal value. Also, the methods that are used to measure and define the AT are diverse and often confuse comparison of any variables associated with the concept.[49] Information concerning the AT measured in trained cyclists or runners completing incremental treadmill running and cycle ergometry tests seems . to indicate that, similar to VO2max, training background may influence the observed response during incremental exercise in cycling or running. Davis et al.[26]. investigated the comparability of the AT and VO2max obtained from cycling and treadmill exercise (walking and running) in college students. No significant difference was found between cycling and treadmill exercise .modes for the AT (63.8 – 9.0 and 58.6 – 5.8% VO2max for cycling and treadmill exercise, respectively). However, the subjects in this study were not trained in either cycling or running. Withers et al.[6] later showed the AT .was similar when expressed as a percentage of VO2max in both the runners and the cyclists when obtained from an incremental cycle or treadmill running test. Coyle et al.[41] were also able to compare the LT ª 2009 Adis Data Information BV. All rights reserved.

Millet et al.

measured in well trained cyclists completing both incremental cycle ergometer and treadmill run. ning tests, and found no difference in VO2max in cycling and running. However, they . found that subjects with the highest LT (%VO2max) in cycling had a similar LT in treadmill running. In contrast, the subjects with the lowest LT possessed an LT in treadmill running that closely matched the subjects in the high LT group. These data indicate that in well trained cyclists the LT may not necessarily be different when measured in running but this is probably due to specific adaptations in cycling rather than to any lack of adaptation to running.[50] However, more importantly, the conclusion drawn from this study is that cycling skill, muscle recruitment patterns and coordination in part influences the LT in cycling but not in running. Mazzeo and Marshall[42] compared a group of trained cyclists completing both incremental cycling and running tests. These authors confirmed . that the cyclists obtained similar VO2max values in both tests,. whereas in the distance runners . treadmill VO2max was higher than .cycle VO2max. LT and VT occurred at a lower % VO2max in both the cyclists and runners in the exercise mode in which they had not trained in extensively. However, the difference in the LT and VT was larger in the runners than in the cyclists. In addition, these authors found that the inflection point in plasma catecholamines also shifted in a similar manner to the blood lactate concentration between exercise modes, indicating the muscle metabolic response also changed with the different exercise modes. They suggested that the runners were not familiar with cycling, which may explain the differences in the LT and VT in cycling in trained runners. In line with this suggestion, Bouckaert et al.[34] found the blood lactate response to incremental exercise was markedly different during incremental exercise in trained runners and cyclists . performing exercise testing with the absolute VO2 corresponding to the onset of blood lactate accumulation (OBLA) lower in the exercise mode that they were not accustomed. However, the effect was much greater in the runners performing the incremental cycle testing, indicating the Sports Med 2009; 39 (3)

Physiological Differences Between Cycling and Running

nature of the exercise was more unfamiliar than the cyclists performing running exercise. In another study, Hassmen[51] examined the effect of specialized training upon both physiological performance and perceptual responses to incremental cycling or running exercise. It was shown that the OBLA was much lower in the exercise mode that the subjects were not training in. In addition, the perception of effort and discomfort was far greater for the runners completing incremental cycle exercise at each submaximal workload. Therefore, these findings suggest that the perception of discomfort and peripheral fatigue is much greater in runners completing incremental cycle exercise. Although in this study the HR corresponding to the set lactate inflection points was not directly measured, the data indicate that calculation of HR corresponding to the OBLA would have been far higher in cycling relative to running in the runners, but similar in cyclists completing the two exercise tests. 1.2 Triathletes Performing Cycle Ergometry and Treadmill Running 1.2.1 Maximal Aerobic Power

Table II shows the studies that have reported maximal oxygen uptake and peak work load or power for cycling and running in triathletes.[8-10,12,52-89] Kohrt et al.[53] and O’Toole et al.[54] were among the first groups of researchers to compare . VO2max of triathletes measured in both cycle ergometry and treadmill running. Kohrt et al.[53] assessed 13 triathletes in preparation for an . Ironman triathlon. They found that VO2max was significantly lower in cycle ergometry as compared with treadmill running (57.9 – 5.7 vs 60.5 – 5.6 mL/kg/min). In . contrast, O’Toole et al.[54] reported similar VO2max values in treadmill running and cycling. Therefore, these data . are inconclusive with regard to differences in VO2max between cycling and running in triathletes. In light of the data . showing trained cyclists are able to achieve a VO2max in cycling comparable with running,[5,30,39] it is possible that these subjects had adapted more to cycle .training, effectively reducing the difference in VO2max between the exercise modes. However, the subjects recruited ª 2009 Adis Data Information BV. All rights reserved.

185

for each investigation were of very mixed. ability level (as evidenced by a relatively low VO2max), which could also have confounded the results. However, others have reported similar values for . VO2max in cycling and running in short distance specialist triathletes.[9,68,72] For example, Sleivert and Wenger[68].reported similar values for cycling and running VO2max in 18 triathletes of mixed ability. In another study, Miura et al.[80] examined two groups of triathletes whom they characterized as either ‘superior’ or ‘slower’ . level. They found no significant difference in VO2max in cycling and running . in both groups. Therefore, the differences in VO2max between exercise modes may not be due to . ability level. However, Schabort et al.[81] found VO2max to be significantly higher in treadmill running then cycle ergometry (68.9 – 7.4 vs 65.6 – 6.3 mL/kg/min) in national level triathletes. Most studies have also shown . that VO2max is similar in cycling and running for triathletes of a wide range of competitive level In performing incremental tests.[9,60,67,68,72,77] . only one study has a significantly higher VO2max value been reported in cycling than in treadmill running (65.4 – 4.2 vs 62.1 – 6.3 mL/kg/min) in well trained triathletes.[78] This seems to be an exceptional result compared with the body of scientific evidence. 1.2.2 Anaerobic Threshold

Despite there still being controversy over the validity of the AT and LT determined via different procedures (invasive vs non-invasive, lactic vs ventilatory) [for reviews see Bentley et al.[37] and Loat and Rhodes[48]], a number of studies in triathletes .have extended on initial studies by comparing both VO2max and a measure of the AT in cycling and running in triathletes.[9,10,68,90,91] Table III shows the studies with ventilatory or anaerobic threshold data in cycling and running for cyclists, runners and triathletes.[6,9,10,26,30,31,52,58,68,72,73,78,80,87,89,90,92-97] Kohrt et al.[62] conducted a 6- to 8-month longitudinal investigation of 14 moderately trained triathletes in training for a long distance . triathlon. The researchers quantified VO . 2max and the LT in both cycling and running. VO2max remained relatively constant in both cycling and running until the latter stages of the training Sports Med 2009; 39 (3)

. Relative VO2max run (mL/kg/min)a

T

56.3

57.6

[52] [53] H

. Absolute VO2max run (L/min)a

9M

Experienced

LDT

13 M

Competitive

29.5 – 4.8

69.8 – 5.6

57.9 – 5.6*

60.5 – 5.7*

LDT

8M

SA

30.5 – 8.8

74.7 – 10

66.7 – 10.1

68.8 – 10.4

5.1 – 0.9

6F

WC

31.3 – 5.6

60.3 – 4.6

61.6 – 7

65.9 – 8.1

3.9 – 0.4

5F

WC subgroup

67.0 – 7.7

61.0 – 8.5

1F

SA subgroup

60.6

64.6

6M

SA subgroup

66.1 – 9.2

63.9 – 9.2

2M

WC subgroup

77.0 – 10.0

75.1 – 10.0

8M

Highly trained

LDT

7F LDT

30.5 – 8.8

74.7 – 10.0

66.7 – 10.1

68.8 – 10.4

31.3 – 5.6

58.8 – 5.7

64.0 – 8.9

68.1 – 9.4

56.9

61.0

64.3

67.2

8F 10 M

Not clear

LDT

11 M

Top 200 finishers

SDT

10 M

None given

LDT

9M

SDT LDT

31.4 – 5.9

74.5 – 7.6

Referenceb

[54] H, SF

[55] SF

[56]

4.7 – 0.3

4.8 – 0.3

[57]

4.6 – 0.5

4.9 – 0.8

[12]

Varied

64.3 – 8.5

68.1 – 11.9

[58]

14 M

Competitive

43.6 – 8.1

49.7 – 7.5

[59]

11 M

Not clear

31.4 – 1.8

74.5 – 2.3

63.2 – 1.7

4.81

65.3 – 1.3

186

ª 2009 Adis Data Information BV. All rights reserved.

Table II. Studies (n = 42) that have assessed maximal oxygen uptake for cycling and running in triathletes . . Sport n/sex Level/details Age (y) Mass (kg) Absolute VO2max Relative VO2max a bike (mL/kg/min/) bike (L/min)a

4.7 – 0.1

[60] H, DP, R > 1.0 S

SDT

4M

*

4.7 – 0.4

‘Elite’

4.8 – 0.4

[61] [62] H

LDT

8M 6F

(I, Feb)

29.4 – 5.1

M 55.3–56.4 F 69.9–71.3

(II, 6–8 wk post I)

10 M

57.4 – 1.4

55.5 – 1.5*

57.8 – 1.5

*

57.2 – 1.5

53.4 – 1.5

(III, 6–8 wk post II)

54.2 – 1.5

(IV, 6–8 wk post III)

56.0 – 1.3*

Highly trained

27.6 – 6.3

72 – 5.4

70.3 – 6*

58.4 – 1.4 5.1*

75.4 – 7.3*

5.4 – 0.6*

[10] H, SF

Continued next page

Millet et al.

Sports Med 2009; 39 (3)

SDT

*

. Relative VO2max run (mL/kg/min)a

. Absolute VO2max run (L/min)a

Referenceb

Level/details

SDT

10 M

Competitive

62.9 – 3.8

67.0 – 4.2

[63]

LDT

10 M

Not clear

60.8 – 1.4

61.6 – 1.1

[64]

SDT

7F

Recreational

48.2 – 3.8

Not clear

56.5 – 8.5

SDT

7M

Mass (kg)

. Absolute VO2max bike (L/min)a

n/sex

16 M

Age (y)

. Relative VO2max bike (mL/kg/min/)a

Sport

2.9 – 0.3

3.1 – 0.2

69.9 – 5.5 MW

60.5 – 6.2

HW

51.9 – 3.9

55.6 – 4.1

[66]

HW

7M

Competitive

T

7M

Horizontal TR

24 – 3

75 – 10

66.4 – 1

T

18 M

Recreational

27.7 – 1.3

76.2 – 2.1

60.1 – 1.5

4.6 – 0.1

63.7 – 1.6

4.8 – 0.1

28.3 – 2.3

59.3 – 2.1

51.1 – 2

3.0 – 0.1

51.4 – 1.3

3.1 – 0.1

7F

[65]

62.0 – 8.4

MW

Competitive

50.7 – 2.6

[67]

66.1 – 7.9

[68] H, NDP

SDT

9M

57.9 – 4.5

59.3 – 6.9

[69]

T

8M

67.1 – 2.6

68.1 – 5.4

[70]

SDT

14 M

58.5 – 6.8

61.3 – 6.6

[71]

SDT

10 M

T

T

Amateur

27.4 – 5.7

78.4 – 8

61.3 – 10.1

4.75 – 0.5

63.3 – 8.9

7F

54.3 – 3.6

57.3 – 3.6

7F

63.2 – 3.9

65.4 – 2.9

6F

10 wk R

20.3 – 0.9

58.2 – 3.3

4.9 – 0.2

[72] H, SF, DP

Physiological Differences Between Cycling and Running

ª 2009 Adis Data Information BV. All rights reserved.

Table II. Contd

[73]

2.3 – 0.1

2.6 – 0.1

[74] NDP, R>1.15

6F

10 wk C

20.5 – 1.0

61.6 – 3.6

2.3 – 0.1

2.5 – 0.1

6F

10 wk C + R

21.3 – 0.6

62.4 – 3.0

2.5 – 0.2*

2.6 – 0.2

SDT

5M

Competitive

SDT

6M

SDT

17 M

60.8 – 3.0

[75]

64.3 – 4.7 4.53 – 0.1

26.5 – 8.2

62.8 – 5.1

61.1 – 8.1

4.5 – 0.1

[76] [77] NDP,

63.8 – 8.1

R<1.05

SDT

7M

Competitive

20.8 – 2.9

69.7 – 4.5

65.4 – 4.2

[78] H, DP

62.1 – 6.3

HRm, I

SDT

T

70 – 4.8

Competitive

[79]

71.7 – 4.9 #

#

8M

Superior

27.3 – 6.8

65.4 – 5.8

67.8 – 6.1

69.7 – 5.4

8M

Lower

26 – 10.3

60.8 – 3.2

54.9 – 3.8#

59.3 – 7.1#

21.3 – 1.6

65.7 – 5.6

64.6 – 2.6*

4M 2F

4.2 – 0.6*

66.9 – 3.7*

[80] H, NDP

4.4 – 0.4*

[8]

Continued next page

187

Sports Med 2009; 39 (3)

SDT

9M

Sport

n/sex

Level/details

Age (y)

Mass (kg)

. Relative VO2max bike (mL/kg/min/)a

. Absolute VO2max bike (L/min)a

. Relative VO2max run (mL/kg/min)a

. Absolute VO2max run (L/min)a

C

6M

24.3 – 7.5

72.5 – 3.7

71.2 – 3.9*

5.2 – 0.5*

75.3 – 3.8*

5.3 – 0.4*

R

4M 2F

21.0 – 2.4

64.8 – 13.8

61.7 – 5*

4.0 – 1.1*

68.4 – 4.1*

4.5 – 1.1*

All

14 M 2F

22.2 – 5

67.7 – 9.1

65.8 – 4.8*

4.4 – 0.1*

70.2 – 4.3*

4.7 – 0.8*

SDT

29 M

Competitive

20.9 – 2.6

68 – 7.8

69.1 – 7.2

4.7

70.2 – 6.2

4.8 – 0.4

6M

Elite

21.8 – 2.4

69.9 – 7.3

75.9 – 5.2

5.3

78.5 – 3.6

5.5 – 0.3

5M

National squad

23 – 4

72.1 – 4.7

69.9 – 4.5*

5.0 – 0.4

74.7 – 5.3*

5.3 – 0.5

25 – 7

59.3 – 5.8

61.3 – 4.6

3.6 – 0.4

63.2 – 3.6

3.7 – 0.3

65.6 – 6.3*

4.3 – 0.8

68.9 – 7.4*

4.5 – 1

SDT

5F 5M 5F

Referenceb

[9]

[81] H

[82]

8M

Competitive

21.7 – 1.7

71.4 – 2.2

64.7 – 2.4

64.2 – 2.1

5M

Elite senior

25.4 – 0.8

72.2 – 3.4

75.7 – 2.3

76.3 – 3.2

12 M

Good level

70.7 – 3.8

61.0 – 6.2

12 M

Middle level

67.7 – 6.4

56.9 – 5.5

SDT

13 M

University team

67.2 – 1.6

68.8 – 1.8

[84]

T

10 M

67.1 – 1.6

68.7 – 2.6

[85]

SDT

13 M

67.2 – 1.6

68.8 – 1.8

[86]

SDT

8M

University team

24.0 – 3.0

71.1 – 6.5

68.7 – 3.2

69.9 – 5.5

[87] SF

T

4M 4F

Preparatory training

22 – 2

60.7 – 10

60.9 – 6.7*

64.8 – 5.8*

[88] NDP, R > 1.05

Specific training

61.9 – 6.4*

66.1 – 6.9*

Pre-competition

62.8 – 7.2*

67.1 – 5.9*

SDT

SDT

SDT

b

71.7 – 1.8

Competitive

28.9 – 7.4

73.3 – 6.0

67.6 – 3.6

4.9 – 0.4

68.9 – 4.6

[83]

5.0 – 0.5

[89] SF

Values are reported as mean – SD. Where the original study reported the standard error of the mean, the standard deviation was calculated using the formula SEMOn, where SEM is the standard error of the mean and n is the sample size. . Criteria for VO2max are presented as a superscript in this column.

Feb = February; C = cycle trained; DP = defined plateau; F = females; H = highest averaged value reached within last stage; HRm = HR > age predicted HRmax; HW = heavy weight; . LDT = long distance triathletes; M = males; MW = medium weight; NDP = non-defined plateau in VO2 despite increase in speed or work rate; R = run trained; R > 1.05 RER > 1.05, R. > 1.15 RER > 1.15; S = students; SA = serious amateur; SDT = short distance triathletes; Sept = September; SF = subjective/volitional exhaustion; T = triathletes; TR = treadmill; VO2max = maximal oxygen consumption; WC = world class. * p < 0.05: differences between running vs cycling, # p < 0.05: differences in the same sex, within subgroups.

Millet et al.

Sports Med 2009; 39 (3)

a

8M

23.1 – 1.2

188

ª 2009 Adis Data Information BV. All rights reserved.

Table II. Contd

Physiological Differences Between Cycling and Running

189

Table III. Studies (n = 21) showing ventilatory (VT)/anaerobic threshold-related data for cycling and running in cyclists, runners and triathletes . . . . Sport Performance n/sex VT run Referencea VT bike VT VO2 run VT VO2 bike VT VO2 run VT VO2 bike . . level (%VO2max) (L/min) (L/min) (mL/kg/min) (mL/kg/min) (%VO2max) 63.8 – 9

58.6 – 5.8

[26]1,4,6

46.8 – 3.2*

66.3 – 6.9

74.3 – 6.1

[6]

37.8 – 5.3*

52.7 – 6.2*

61.2 – 4.9

77.3 – 2.6

48 – 5

57 – 5

79 – 7

85 – 5

44.3

45.7

2.4 – 0.3

S

Previously untrained

30 M

C

Steady training

10 M

3.0 – 0.5*

3.2 – 2.7*

43.7 – 6.2*

10 M

2.6 – 0.3*

3.6 – 0.4*

R

12 M T

Experienced

T

9M 10 M

*

[52]

*

3.9

4.42

85

R

2 y training

C

10 M

2.6 – 0.3

9M

3.3 – 0.4

T

Highly trained

10 M

3.0 – 0.5

T

Highly trained

10 F

2.2 – 0.1

T

Elite

T

Amateur

T

Competitive

T

Competitive

T

Superior

T

Lower

T

All

T T T

T a

3.1 – 4.3 *

3.9 – 0.3

*

2.8 – 0.1

46.9 – 4.3

*

37.7 – 1.9

80.8 – 6.9

66.8 – 3.7

71.9 – 6.6

[10]5

47.2 – 2

62.7 – 2.1*

74.0 – 2.0*

[90]

74.8 – 1.9*

85.0 – 2.1*

[68]5

*

*

81.4 – 1.3

85.0 – 1.3

57.8 – 5.3

84.6 – 5.0

SwBK

4.1 – 0.7

53.5 – 3.5

84.6 – 2.5

RBK

4.0 – 0.2

63.5 – 3.5

7M 4.5 – 0.2*

52.2 – 3.2* 42.5 – 6.5

7M

57.7 – 2.7* 46.4 – 6.3

8M

48.7 – 3.8#

50.9 – 4.8#

8M

#

#

39.7 – 2.9

87.0 – 7.0 [73]

86.2

85.0 – 1.3*

91.1 – 1.0*

[72]6

65.0 – 9.9

74.7 – 10.1

[78]

72.5 – 0.4

84.9 – 0.6

[94] [80]

40.4 – 4.8

3.0 – 0.6*

2.6 – 0.4*

45.1 – 8.2

46.7 – 4.1

65.3

Elite

6M

3.0 – 0.6

2.8 – 0.3

49 – 10.9

50.9 – 4.3

64.5

64.8

Well trained

8M

69.9 – 3.3

70.1 – 3.4

Well trained

9M

[87]5 [95]5

88.9 – 0.2

#

(Precompetitive)

7M 1F

3.7 – 0.2

55.8 – 2.8

(Competitive)

7M 1F

3.7 – 0.2

55.4 – 3.3

88.6 – 0.2#

(Postcompetitive)

7M 1F

3.3 – 0.2#,

49.0 – 4.1

79.0 – 0.2

8M

(LT) 3.8 – 0.4*

(LT) 4.4 – 0.5*

[9]2

66.4

67.0 – 3.6 #

[93]3

71.8

29 M

#,

[30]5

53.9 – 3.8

4.0 – 0.5

4.0 – 0.2*

*

79.78 – 4.3

6M

10 M

[92]

71.0 – 2.4 *

79.78 – 4.9 *

18 M Well trained

T

T

74.8 – 5.7

7F

T T

3.2 – 0.3

*

[58]8

90

71.0 – 3.5 *

[31]

[96]5,7

[89]

VT determination by method of: 1 = abrupt FEO2 increase; 2 = V slope; 3 = VT equivalent; 4 = abrupt R increase; 5 = nonlinear increase equivalent in O2 determined by a computerized algorithm; 6 = non-linear increment in minute VE to exercise time identified by computerized analysis, checked from VT equivalents; 7 = determined visually from VE time-course curve; 8 = subjectively determined via VT breakpoint.

C = cycling; F = female; LT = lactate threshold modified Dmax method; M = male; R = running; RBK = run background; S = student; SwBK = ˙ O2max = maximal oxygen consumption; * p < 0.05: differences between running swim background; T = triathlete; V˙O2 = oxygen consumption; V vs cycling, # or p < 0.05: differences in the same sex, within subgroups.

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (3)

190

period, possibly reflecting an increase in training . intensity at that time. However, VO2max together with the LT in cycling was consistently lower than that obtained for treadmill running. These findings suggest that adaptation of the LT is specific to the muscles involved in training in a particular exercise mode. It may also be that the subjects’ training background was more extensive in cycling than running. This study also indicates the nature of training in either exercise mode may influence adaptation in cycling or running. In a more recent longitudinal study[98] over one season in trained Olympic distance triathletes, the . relative stability of VO2max and the larger change in VT under the influence of specific training has been confirmed. However, Albrecht et al.[52] found . no difference between the VT (expressed as % VO2max) obtained in cycling (78.8%) or running (79.3%). In accordance with this, Kreider[12] showed no significant difference in the VT in triathletes completing incremental tests in cycling and treadmill running. The study methodology and the training background of the subjects that these studies involved, however, were not clear. Interestingly, these authors found that the exercise intensity sustained during the cycle and running stages of a short distance triathlon was similar. In single sport endurance competitions it is generally thought that the AT reflects the ability to sustain a set percentage of maximum capacity.[99] Kreider’s data,[12] collected for a triathlon event, imply otherwise. Despite the VT of the athletes occurring at a different exercise intensity within isolated incremental running and cycling . tests (90% vs 85% of VO2max), the exercise intensity that they sustained during a race was similar for both exercise modes. However, De Vito et al.[93] showed the VT in running to be lower after prior cycle exercise. These results and those reported by Zhou et al.[72] suggest that the cycle stage of a triathlon influences the ability to sustain a set percentage of maximal capacity during the subsequent running stage. This has implications for training prescription on the basis of incremental tests that have been performed in isolation. . Hue et al.[9] examined VO2max and the VT in triathletes competing in short distance events. ª 2009 Adis Data Information BV. All rights reserved.

Millet et al.

. They found that the VO2max (75.9 – 5.2 vs 78.5 – 3.6 mL/kg/min) values of international standard triathletes were similar in cycling and running but higher than in competitive triathletes (69.1 – 7.2 vs 70.2 – 6.3 . mL/kg/min). Similarly, VT (expressed as % VO2max) was also similar in cycling and running in the two groups. Miura et al.[80] also reported VT measured in cycling and running to be similar, in absolute terms, in two groups of triathletes who varied in short distance triathlon race time. Schneider et al.[10] were able to . confirm these findings and found that whilst VO2max was significantly higher in running when compared with cycle exercise (75.4 – 7.3 vs 70.3 – 6.0 mL/kg/min), the VT was not significantly different between cycling. and running when expressed as an . absolute VO2 value but did differ relative to VO2max (66.8 – 3.7 vs 71.9 – 6.6%). In contrast, some reports do show differences in the AT between cycling and running.[67,72,90] It is possible that the volume or intensity of training in either cycling or running may influence the AT in either exercise mode. However, Schneider and Pollack[90] quantified training volume and intensity in cycling and running completed by ‘elite’ triathletes and found no significant differences in these variables . despite a significant difference in the VT (%VO2max) in cycling versus running (74.0 – 2.0 vs 62.7 – 2.1%). On the other hand, Medelli et al.[67] quantified both the ‘aerobic’ (corresponding to VT) and ‘anaerobic’ threshold in both treadmill running and cycling in triathletes that they reported as ‘elite’. These authors found that . the AT occurred at a significantly greater % VO2max (88.8%) during inclined (1.5% slope) treadmill running as compared with cycle ergometry (83.3%). However, the aerobic threshold was not significantly different between inclined running and cycle ergometry. Since Sloniger et al.[100] reported a greater activation of the vastus and soleus in uphill than in level running, inclined treadmill running may elicit different recruitment patterns in the calf and quadriceps muscle groups during progressive incremental exercise. This may partly explain the results observed by Medelli et al.[67] In another study, Zhou et al.[72] found no. significant differences in treadmill or ergometry VO2max. However, the VT Sports Med 2009; 39 (3)

Physiological Differences Between Cycling and Running

was significantly higher in running compared with . cycling (91.1 – 1.0 vs 85.0 – 1.3% VO2max). . There are also limited data available comparing VO2max and the AT in cycling and running exercise modes [101] found for ‘duathletes’. In one study, . Bolognesi a significant difference in VO2max measured in cycle ergometry (66.3 – 9.0 mL/kg/min) and treadmill running (71.4 – 10.3 mL/kg/min) in eight male duathletes. In this study a significant difference was . also observed between the VT (expressed as % VO2max) in cycling (68.8 – 3.7%) and treadmill running (73.9 – 6.6%).

191

differences between males and females are probably induced by different training characteristics. The hypothesized differences in training between elite male and female triathletes is likely influenced by the characteristics of the competition at this level: the cycling bout (where the number and density of athletes is lower than in males [102]) has a stronger influence on the overall race result in females. From the limited literature available, it appears that in general males and females exhibit the . same differences between running and cycling VO2max and AT.

1.2.3 Sex Differences

. There are limited reports comparing VO2max and the AT obtained within cycling and running between males and females. Sleivert and Wenger[68] found both male and. female triathletes to exhibit no differences in VO2max between running and cycling. O’Toole et al.[54] also found no . significant differences in cycling and running VO2max in male (68.8 – 10.4 vs 66.7 – 10.1 mL/kg/min) or female (65.9 – 8.1 vs 61.6 – 7.0 mL/kg/min) triathletes. Therefore, the difference . in VO2max between cycling and running does not seem to vary appreciably between males and females. Similarly, Sleivert and Wenger[68] reported that the AT was not significantly different in cycling and running in male athletes. However, there was a significant difference in the VT between cycling and running in female athletes. Millet and Bentley[97] have also investigated if the differences in performance between elite junior and elite senior triathletes were due to the same physiological differences in men and women. Irrespective of sex, there were no differences in . VO2max (74.7 – 5.7 vs 74.3 – 4.4 and 60.1 – 1.8 vs 61.0 – 5.0 mL/kg/min) and cycling economy (72.5 – 4.5 vs 73.8 – 4.3 and 75.6 – 4.5 vs 79.8 – 9.8 W/L/min) between junior and senior triathletes. However, the difference in performance between juniors and seniors was due to different reasons in male and female triathletes: senior males had a higher VT than junior males whereas VT was similar in female junior and senior triathletes. In female triathletes, senior triathletes had a higher PPO and a lower increase in the energy cost of running after cycling. These ª 2009 Adis Data Information BV. All rights reserved.

1.2.4 Heart Rate

Maximal heart rate (HRmax) is generally reported to be slightly (~5%) higher when obtained from an incremental treadmill test as compared with an incremental cycle test in untrained subjects.[7,23,103] In addition, the relationship . between HR and exercise intensity or VO2 is exercise dependent[14,104,105] and is influenced by training mode, postural position[106] or laboratory environment.[107] In triathletes, the HRmax observed in cycling is often lower by 6–10 beats/min than that obtained during running.[62,78,79,103] Longitudinal investigations have demonstrated HRmax to remain relatively stable over the course of a season,[98] with higher values (~5 beats/min) observed in running than in cycling.[62] In contrast, there is also evidence suggesting that HRmax is similar between cycling and running modes.[53,67,72,88,99] Although this appears to hold for males, differences were observed for this variable in females by some authors.[55] Schneider and Pollack,[90] however, found no such significant differences between cycling and running HRmax in elite female triathletes. The HR corresponding to the AT is used to prescribe submaximal exercise training loads.[105,108] The data concerning triathletes indicate that the HR corresponding to certain inflection points associated with the AT is always higher in running than cycling, both when expressed in absolute terms and relative to HRmax.[10,72,78,79,90,103] Schneider et al.[10] reported a significant difference in the HR corresponding to the VT in cycling and running (145.0 – 9.0 vs 156.0 – 8.0) in ‘highly trained’ Sports Med 2009; 39 (3)

Millet et al.

192

triathletes. This corresponded to 80.9 – 3.4 versus 85.4 – 4.1% HRmax. In another study by the same research group and conducted on elite female triathletes,[90] a higher HR was recorded at the VT in running than in cycling (164.7 – 4.0 vs 148.2 – 3.4) and this difference was also evident when HR was expressed as a percent of HRmax (87.3 – 1.6 vs 79.7 – 1.5%). Similarly, Roecker et al.[103] found a difference of 20 beats/min between HR determined at the LT on cycling ergometer (149.9 – 18.0 beats/min) and treadmill (169.6 – 15.7 beats/min). However, recreational subjects (-22 beats/min) and cyclists (-14 beats/ min) exhibited lower differences than triathletes and runners. Additionally, the differences were not influenced by sex. There is some evidence that HR may not differ between cycling and running when it is determined from a submaximal inflection point. Medelli et al.[67] reported HR values corresponding to the ‘aerobic’ but not those corresponding to the ‘anaerobic’ threshold in well trained triathletes to be different in cycling compared with running. In another study, Bolognesi[101] reported no significant difference in the HR corresponding to the VT in cycling and running (152.0 – 8.0 vs 158.0 – 9.0) in duathletes. Similarly, Hue et al.[9] found a non-significant difference in the HR corresponding to the VT in elite triathletes. However, in both these two latter studies the mean difference between cycling and running was ~7 beats/min, which is quite large and practically relevant in terms of training presthat cription. Basset and Boulay[8] have reported . the relationship between HR and % VO2max did not differ when calculated either from a treadmill or from a cycle ergometer test. These authors also showed that HR was similar between running and cycle ergometer tests throughout the training year and concluded that triathletes could use a single mode of testing for prescribing their training HR in running and cycling throughout the year.[88] Zhou et al.[72] showed that the HR corresponding to the VT was significantly higher in running (174.6 – 4.5) as compared with cycling (166.4 – 7.6). However, these authors found that the HR measured in a triathlon race was similar to the HR at the VT in cycling but much lower in ª 2009 Adis Data Information BV. All rights reserved.

running. Other studies have also shown a decrease in the HRmax and the HR corresponding to the VT during an incremental running test performed after submaximal cycling.[48] Hue et al.[78] have also demonstrated that the HR during a 10 km run after 40 km of cycling is higher when compared with the same run without cycling. Therefore, even though the HR corresponding to the AT or HRmax may be similar in running compared with cycling (in exercise tests performed in isolation), the HR corresponding to the AT determined from an incremental running test may be different to that observed in a race situation, especially in running. At the elite level, because of the stochastic pace, there is no demand to control the exercise intensity for the run in an Olympic distance triathlon via HR. Within Ironman, the potential use of HR for controlling the running pace might be of interest, at least at the beginning of the marathon. However, to our knowledge there is no published protocol for determining HR for this purpose. Furthermore, the effect of prior cycling on HR during running should be considered when prescribing HR during running training on its own. 1.2.5 Running Economy

. Running economy can be defined by the VO2 (in mL O2/kg/min) of running at a certain speed, and is usually expressed by the energy cost (EC) of running a distance of 1 km (in mL/kg/km) . calculated as VO2 divided by the velocity. It is known that training-induced, genetic, physiological and anthropometric factors influence economy (for reviews see Foster and Lucia[109] and Saunders et al.[110]). EC has been reported in triathletes within both the conditions of isolated running and ‘triathlon running’.[11,58,60,78,97,111-117] It is generally reported that in trained triathletes, EC measured at the end of an Olympic distance triathlon is higher by ~10% when compared with an isolated run, e.g. 224 versus 204 mL/kg/km,[115] 224 versus 207 mL/kg/km.[111] It has also been reported that the extent of any change in EC subsequent to an exhaustive cycling bout is influenced by athlete performance level, event distance, sex and age. The effect of a fatiguing cycling bout on the Sports Med 2009; 39 (3)

Physiological Differences Between Cycling and Running

subsequent running energy cost was different between elite (-3.7 – 4.8%, when compared to an isolated run) and middle-level (2.3 – 4.6%) triathletes.[116] Elite long-distance triathletes had slightly (but not significantly) lower EC values than short-distance triathletes (163.8 vs 172.9 and 163.0 vs 177.4 mL/kg/km during an isolated and a ‘triathlon’ run, respectively).[11] Surprisingly, no difference has been observed in EC between elite junior and senior triathletes, whether male or female, during an isolated run and a ‘triathlon’ run (173–185 mL/kg/km).[97] However, the increase in EC subsequent to cycling was higher in juniors than in seniors in females (5.8 vs -1.6%) but not in males (3.1 vs 2.6%).[97] The mechanisms underlying the deterioration in economy in the ‘triathlon run’ when compared with an isolated run are various: both reported changes. in the ventilatory pattern[79] leading to a higher VO2 of the respiratory muscles,[116,118] and neuromuscular alterations reducing the efficiency of the stretch-shortening-cycle[113,116,119] have been proposed. Some metabolic factors such as shift in circulating fluids, hypovolaemia and increase in body temperature have also been suggested.[111,114,115] Of interest are the studies of Hausswirth et al.[112-114] comparing EC at the end of a short-distance triathlon and at the end of a marathon of similar duration: EC was more increased during the marathon (+11.7%) than during the triathlon (+3.2%) running when compared with a 45-minute isolated run. The differences are due mainly to a higher decrease in bodyweight related to fluid losses, a larger increase in core temperature during the long run and significant mechanical alterations during the long run when compared with the running part of a triathlon. Interestingly, recent values of EC in worldlevel distance runners have been reported:[120-122] Jones[120] showed a continuous decrease in EC of Paula Radcliffe, the current world record holder for the women’s marathon between 1992 (~205 mL/kg/km) and 2003 (~175 mL/kg/km) corresponding to a 15% improvement, whereas . VO2max (~70 mL/kg/min) and body mass (~54 kg) remained unchanged over the period. Jones reported also that her EC was more recently measured at 165 mL/kg/km. Billat et al.[123,124] ª 2009 Adis Data Information BV. All rights reserved.

193

reported higher values in elite female Portuguese and French (196 – 17 mL/kg/km)[124] or Kenyan (208 – 17 mL/kg/km)[123] distance runners. Overall, this compares favourably with values obtained for elite female triathletes: Millet and Bentley[97] reported in nine elite females (including one long-distance world champion, second at the Hawaii Ironman and five European medallists) an average . value of 176.4 mL/kg/km, whereas the average VO2max was 61.0 mL/kg/min for a body mass of 60.3 kg. In males, Lucia et al.[121,122] reported a value of 150–153 mL/kg/km in Zersenay Tadese, the current long cross-country and half-marathon world . champion for a VO2max of 83 mL/min/kg. The EC of Tadese is lower (the lowest reported so far) than previously reported values in elite runners: 180 mL/kg/km for Steve Scott;[125] 203–214 mL/ kg/km in elite French and Portuguese[124] or Kenyan[123] runners; ~190–192 mL/kg/km in elite East-African runners;[121,126] ~211 mL/kg/km in elite Spanish runners.[121] So, similar to females, with the exception of Tadese, running economy in male distance runners does not appear to be better than the ones reported in elite triathletes: 174 – 9 and 164 – 8 mL/kg/km for short-distance and long-distance triathletes, respectively.[11] However, further investigation with Elite Ironman triathletes is required to confirm . such partial results. Since EC is calculated as VO2 divided by the running velocity, it is unclear how this later parameter influences the comparison between elite triathletes and elite runners who have higher absolute training and competition velocities and therefore a biased lower EC. To our knowledge, there are no values of EC measured at the same absolute or relative (percentage of velocity at . VO2max) speed in the two groups. Overall, from these data, it appears that the main difference in running performance between elite runners and triathletes comes mainly from a higher body mass in triathletes (affecting propor. tional VO2max) rather than from differences in running economy. Since mean lower leg thickness and calf mass have been shown to be related to running economy,[127] one may speculate that the higher body mass in triathletes comes mainly from the upper body muscles more and – probably – Sports Med 2009; 39 (3)

Millet et al.

194

from the higher skinfold thicknesses that are associated with swimming. 1.2.6 Delta Efficiency

Delta efficiency represents the ratio in the changes of external mechanical work to energy expenditure respectively within an incremental stage test.[128,129] Delta efficiency is consistently reported to be higher in running than in cycling;[129-131] however, the mechanisms underlying such a difference are not clear. It is speculated that it comes mainly from the storage-recoil of elastic energy in the series of elastic components of the knee extensors that exists in running[132,133] but not in cycling.[130,131] The acceleration-deceleration of the limbs (internal mechanical power) is higher in running than in cycling. Similarly the metabolic cost of running has been shown to be mainly dependent on the cost of generating peak force during the stance phase and inversely proportional to the contact time.[134,135] The increase in contribution of muscles not directly involved in the force production (arms and trunk) might be higher in cycling than in running and this may therefore influence the delta efficiency. Finally, it cannot be excluded that this later mechanism is related to the described differences in ventilatory pattern between cycling and running (see section 2.1). By comparing delta efficiency between cycling and running in a protocol that excluded several confounding factors such as differences in the metabolic power, extra external load and cycle (pedalling or stride) frequency, Bijker et al.[129] confirmed that delta efficiency is lower in cycling (25.7 – 1.3%) than in running (45.5%). They did not provide strong evidence about the respective contributions of the different mechanisms that they discussed. Efficiency has important consequences in terms of physiological testing (section 1.2), development of fatigue (section 2.4.3) or training content. Efficiency determines the ‘performance velocity’[1] and is therefore of high interest for the coaches. Secondly, the knowledge on the physiological determinants of running and cycling efficiency . is relatively lacking in comparison to VO2max and the lactate threshold. So this area is also of ª 2009 Adis Data Information BV. All rights reserved.

the highest interest for the scientists investigating cross-training adaptations. 2. Physiological Mechanisms Associated with Differences between Cycling and Running . The VO2max is thought to be influenced by a combination of factors of central and peripheral origin, and their respective contributions have been the subject of considerable debate in recent years.[99,136-139] The reader is directed to these reviews for more comprehensive appreciation of . this area. In contrast to the concept of VO2max, there is surprisingly little research examining the mechanisms associated with the AT. Also, minimal research into the. factors surrounding differences in the AT or VO2max between cycling and running has taken place. 2.1 Ventilatory Responses

Differences in ventilatory responses to exercise (exercise-induced arterial hypoxaemia, O2 diffusion capacity, ventilatory fatigue, pulmonary mechanics) have been reported between running and cycling in the literature. It is well documented that there is a drop in partial pressure of oxygen in arterial blood (A-a) DO2 that (PaO2) associated with a widening . begins at around 60–70% VO2max during incremental exercise both in running and in cycling. EAIH is associated with a decrease to 10 mmHg in PaO2 and can be indirectly diagnosed by a decrease in pulse oximetry saturation in oxygen (SpO2) below 90% (for a review, see Prefaut et al.[140]). Exercise-induced arterial hypoxaemia (EAIH) is associated with relative hypoventilation and therefore might occur more often in cycling than in running.[140] In addition, in multisport athletes, EAIH is exercise dependent and influenced by the order of the running and cycling bout.[141] Arterial saturation . in oxygen (SaO2) does not directly influence VO . 2max. However it has been hypothesized that VO2max decreases by ~1% for each 1% decrease in SaO2.[142] However, it is unlikely that the observed differences in SaO2 between cycling and running are linked to the differences Sports Med 2009; 39 (3)

Physiological Differences Between Cycling and Running

. in VO2max. Green et al.[143] compared oxygen desaturation of cyclists and runners in incremental exercise testing. These authors found no differences for this variable at maximal exertion. However, this study did not compare the different athlete groups in both cycling and running tasks. Galy et al.[141] showed that in trained triathletes the desaturation was higher in running (SpO2 = 93.0–93.5%; PaO2 = 86.6–88.7 mmHg) than in cycling (SpO2 = 94.8–95.4%; PaO2 = 91.4–93.7 mmHg), irrespective of the order of bouts (running followed by cycling or vice versa). In addition, to the higher desaturation during running, a higher decrease in pulmonary diffusing capacity was reported after cycling.[96] This was explained by several factors, such as the crouched position on the bicycle (in turn inducing higher intrathoracic pressure), a decrease in thorax volume due to the ‘triathlete’ position on the handlebars and a lower efficiency of the peripheral muscular pump. These latter factors would limit the venous return to the heart and would therefore induce a low pulmonary blood volume after cycling. These different ventilatory/haemodynamic factors were concomitant to different blood rheological responses: cycling was associated with an important decrease in blood volume and running with an increase in the erythrocyte rigidity.[144] The incidence and relationships between these different mechanisms are still unclear, but this group[9,78,79,96,141,144-148] provided convincing arguments that pulmonary diffusion is different between cycling and running. Smith et al.[149] found no difference in O2 saturation between the two exercise modes in combination. However, it has also been shown that at maximum exertion there is a lower O2 saturation in treadmill running compared with cycle ergometry.[150] In this study, a lower pulmonary ventilation during treadmill running was associated with higher breathing frequency and no change in tidal volume, indicating breathing mechanics were not altered by the different exercise modalities. These results were not confirmed by Boussana et al.[147] They reported that ventilation was more impaired in cycling than running, therefore inducing a greater decrease in ventilatory kinetics.[141,144-146,148] Boussana ª 2009 Adis Data Information BV. All rights reserved.

195

et al.[146] also reported that the ventilatory fatigue was higher in recreational triathletes when compared with their elite counterparts and that the order of successive submaximal bouts of cycling and running influenced the kinetics of the respiratory fatigue that was experienced.[145] After a cycling or a cycle-run bout, respiratory fatigue was significant, whereas after a running exercise, the signs of fatigue (i.e. decrease in maximal inspiratory pressure or the time to exhaustion that a respiratory load can be sustained) were not apparent. These results show that specific ventilatory adaptations occur as a result of the order of the cycle and run bouts during a triathlon event and that these may be partly compensated for by training. Only a few studies have compared the O2 concentration of arterial blood during maximal exercise testing in cycling and running.[22,141,143,149,150] Hermansen et al.[22] found no difference in the arteriovenous difference of O2 between running and cycling. Hopkins et al.[151] found trained female cyclists to exhibit higher pulmonary ventilation at . maximal exertion, despite no difference in VO2max, during an incremental cycle test compared with a running test. It was suggested that differences in ventilation were associated with changes in pulmonary mechanics between cycling and running. The difference in mechanics was thought to be associated with differences in entrainment of the muscles of the diaphragm between the two exercise modes.[147,148,152] Indeed, another study has shown that entrainment of these muscles is higher in cycling than in running in triathletes.[153] Therefore, the degree of adaptation of pulmonary mechanics in response to combined cycling and running training may affect breathing mechanics during incremental cycle or. running exercise, thereby influencing observed VO2max. To summarize, a conclusive set of studies have shown that ventilatory pattern is more altered in cycling than in running. 2.2 Central and Peripheral Blood Flow

There are some studies in untrained subjects that have demonstrated that an increase in Sports Med 2009; 39 (3)

196

. VO2max with endurance training is associated predominantly with an increase in maximum cardiac output (CO) induced by increase in stroke volume (SV), as compared with increases in arteriovenous (a-v) O2 difference.[154,155] The data presented by Hermansen et al.[22] suggest that CO influenced by SV has .an important influence on cycling or running VO2max. These researchers showed SV to be higher during treadmill running than during cycling. This result reflected the dif. ferences in VO2max that were observed between the two exercise modes. Faulkner et al.[24] also measured SV during maximal running and . cycling exercise. They found that a lower VO2max was associated with a lower SV. Furthermore, a-v differences in arterial O2 concentration were similar between the two exercise modes. Therefore, both these studies provide evidence that a lower . VO2max in cycling is associated with a lower SV influencing CO. However, .an older study has suggested that the lower VO2max in cycling is thought to be due to a lower a-v O2 difference together with a lower maximal CO.[5] The lower CO observed in cycling compared with running could be due to a reduced rate of cardiac filling influenced . by limited venous return thereby influencing VO2max.[21] The reduced venous return may be due in part to peripheral muscle blood flow. Some evidence suggests that peripheral blood flow is different in the lower extremities during cycling as compared with running.[28,156-158] It has been suggested that ‘‘factors influencing venous return to the heart ‘drive’ the circulation during exercise’’.[158] Extravascular compression expels blood from the venous vasculature and impedes inflow of blood into the arterial vasculatures. This mechanism, called the ‘muscle pump’, which facilitates venous return to the heart and perfusion of skeletal muscle (in addition to suction at ventricular level or during muscle relaxation, vasodilator chemicals and decrease in peripheral resistances), occurs to a greater extent during locomotory muscle rhythmic contractions than during twitch or isometric contractions, and has therefore been reported both in cycling and running. The efficiency of this ‘muscle pump’ that is assumed to increase the local muscle blood flow is influenced ª 2009 Adis Data Information BV. All rights reserved.

Millet et al.

by the type of activity performed.[156] To our knowledge, there has been no direct comparison of muscle pump efficiency between cycling and running. However, there are several mechanisms suggesting that the muscle pump efficiency is greater in running than in cycling: first, there is a direct mechanical coupling between contraction frequency and muscle blood flow and therefore muscle pump is directly influenced by the strides frequency;[159] secondly, its efficiency is greater in erect position; finally the type of contraction during running (stretch-shortening cycle) induces some pro-inflammatory processes that per se increase the muscle blood flow.[157] In addition, Matsui et al.[28] found that total lower limb blood flow was significantly lower immediately postexercise after cycle exercise than after running exercise. However, the measurements that they took were indirect. The adaptation of blood flow in the calf and quadriceps muscle groups to training in cycling and running is a potentially interesting area of research in triathletes. It has also been suggested that maximum CO is influenced by coronary blood flow or is mediated by the CNS.[160,161] Indeed, incremental exercise to exhaustion at altitude does not induce skeletal muscle acidosis or even a maximal cardiac output relative to sea level conditions.[33] Hence, fatigue at maximal exertion and in turn . VO2max may be influenced by blood flow to the heart or central neural innervation. However, to our knowledge, it is not known whether this differs in cycling compared with running. 2.3 Skeletal Muscle Oxidative Capacity

The peripheral muscle mitochondria are the site of cellular respiration and electron transport. An increase in mitochondria content and enzyme activity with endurance training is thought to result in an increase in the potential for cellular O2 uptake.[162] Endurance athletes typically possess a greater number of slow twitch (ST) fibres than non-endurance trained athletes. ST fibres are known to be higher in mitochondria and oxidative enzymes, which could be .associated with an increased whole body VO2max or AT.[41,163,164] In contrast, an increase in skeletal Sports Med 2009; 39 (3)

Physiological Differences Between Cycling and Running

muscle mitochondria may occur . without the same corresponding increase in VO2max,[165] suggesting that muscle . oxidative capacity is not a factor related to VO2max. However, a number of studies have reported significant correlations between a. number of skeletal muscle characteristics, VO2max and the AT.[163,166,167] Other investigators have shown a significant relationship between % ST fibres and mechanical efficiency during cycle exercise.[168,169] A few research groups have shown that muscle buffering capacity has a positive influence on endurance rowing, cycling and running performance.[127,164,170] Whilst this data should not be viewed as evidence of ‘cause and effect’, it provides evidence . that skeletal muscle has some influence on the VO2max and the AT together with endurance performance. However, whether these findings are replicated in both cycling and running modes in triathletes or even single sport athletes when muscle is analysed from both quadriceps and the calf muscle groups is not clear. In line with this, there are limited data available regarding the skeletal muscle characteristics of triathletes and how . they may impact on cycle and run AT and VO2max. Only one study has compared the oxidative capacity of skeletal muscle and the AT in both cycling and running.[31] These . untrained . authors found that in subjects the VO2max and OBLA (% VO2max) was higher in running than cycling. The % ST fibre and oxidative enzyme capacity (determined in the vastus lateralis and gastrocnemius muscles) was . not related to either the OBLA or VO2max in cycling or running. However, subject numbers and the training status of the subjects limit the validity of the results from this study. Flynn et al.[57] also obtained muscle tissue from the gastrocnemius, vastus lateralis and posterior deltoid muscles of 11 triathletes and four normally active controls. Muscle fibre type, respiratory capacity and citrate synthase (CS) activity were examined in the samples. There was no significant difference in the % ST fibres (59 – 4.0 vs 63 – 3.3%) in the gastrocnemius and vastus lateralis muscles. The respiratory capacity also did not differ between the gastrocnemius and vastus lateralis muscles. However, CS activity of the vastus lateralis and ª 2009 Adis Data Information BV. All rights reserved.

197

the gastrocnemius were significantly different. Therefore, adaptation of oxidative enzyme activity in these muscles may occur independently. The significance of this in triathletes and performance in cycling and running is not known. Further studies are required to determine the influence of muscle fibre type and enzyme capacity in the calf and quadriceps on physiological variables measured in cycling and running. 2.4 Central and Peripheral Innervation 2.4.1 Muscle Recruitment Patterns

Running and cycling activity is performed by muscle contraction of the lower limbs. The main muscle groups that are involved in cycling and running are the quadriceps and plantar flexors, respectively.[171] An exception to this is during uphill running when the recruitment of the quadriceps muscle is increased.[100] Some researchers have suggested that any observed differences in the AT for cycling and running are a reflection of differences in muscle recruitment during exercise[46] between such exercise modes. Coyle et al.[41] have . stated that skill level may influence the LT (% VO2max) measured in cycling but not in running exercise. The same research group has shown that both the LT and performance level in cycling is influenced by differences in force application to the bicycle crank system.[163] Whether this was associated with modified recruitment patterns of the quadriceps or even calf muscles is not known. However, the authors suggested that the better cyclists were able to generate more pedalling force, at a lower metabolic cost, due to recruitment of the hip flexor muscles. Therefore, the different involvement of the different muscle groups in conjunction with specific training adaptations induced by a combination of cycling and running programmes may be influential on the AT in triathletes or single sport athletes. Marcinik et al.[172] found that a short-term strength training programme resulted in an improvement in the LT in cycling regardless of any cycle training. It was concluded that the strength training resulted in an improvement in muscle recruitment patterns during exercise. This could have influenced the pattern Sports Med 2009; 39 (3)

Millet et al.

198

of muscle recruitment during incremental exercise thereby resulting in a change in the LT. However, the results of Marcinik et al.[172] together with the conclusions drawn by Coyle et al.[163] seem to indicate that the AT is influenced to some extent by the muscle recruitment of the lower limbs. However, longitudinal studies involving different running and cycling training interventions are required in this area. In elite triathletes, it was shown that the mechanical alterations during the first minutes of the running bout subsequent to cycling were minimal[118] and lower than in their recreational counterparts. These mechanical changes were brief and disappeared in less than 6 minutes.[119] However, these alterations induced a higher increase in running energy cost in non-elite than in elite triathletes.[116] Recently it was confirmed that running mechanics are only slightly modified in elite triathletes when compared with an isolated run: Chapman et al.[173] reported that leg kinematics (as measured via 3-D analysis) were not modified in running after cycling at moderate intensity whereas tibialis anterior muscle activity was modified only in 35% of the group and was not associated with any fatigue variables. Therefore, even in elite triathletes, leg muscle activity during running can be influenced by cycling but this is presumably of little influence on race performance. 2.4.2 Pedalling Frequency

It has been reported that differences in physiological variables measured in cycling and running could be due to a greater perception of difficulty in cycling as compared with running.[51,174] The greater perception of effort observed in cycling by some researchers may be in part due to the interaction between optimal pedalling frequency and muscular strength.[172] It has been reported that cycling requires a considerable muscular strength component to performance in the activity.[25,172] The relative volume of training performed in running and cycling respectively may affect these responses in cycling. In terms of performance in cycling, these processes may be an important component to exercise adaptation during exercise. ª 2009 Adis Data Information BV. All rights reserved.

Classically, it is described that the ‘energetically optimal cadence’ (EOC; ~50–75 rpm) does not match the preferred or ‘freely chosen cadence’ (FCC; ~80–100 rpm), although these two cadences are influenced by the performance level and skills of the cyclists.[175] FCC seems to be influenced by perceptual feedback related to objective or subjective neuromuscular fatigue, the decrease in joint loads or in force on cranks.[176,177] It is generally reported that trained cyclists have a higher FCC than untrained subjects,[178,179] but the differences between elite cyclists and either cyclists of lower ability or runners are not always observed.[179,180] At the same time, FCC seems to decrease during prolonged exercise.[181,182] This may indicate that cycling cadence may influence performance when other physiological variables are similar within a group of trained subjects. Furthermore, the cadence selected at the start of a cycling trial together with the reduction of this variable during the trial may be related to the training completed in either running or cycling in triathletes. However, Marsh et al.[179] found no significant differences in preferred cycling cadence during an incremental exercise test between trained runners and cyclists. In this study,[179] delta efficiency did not differ between the athletic groups. These data suggest that cycling cadence may not be influential when primarily running training is performed. This possibly indicates that pedalling cadence is not affected by a training history in either cycling or running. In contrast, during prolonged exercise (2 hours), the preferred cadence is relatively stable (83 rpm) in triathletes.[183] However, the time course of changes in cadence during prolonged exercise has not been compared in athletes specializing in running, cycling or triathlon training. Most of the studies have shown that performance and the stride patterns during running after cycling are greatly influenced by the pedalling frequency during cycling.[87,89,95,182,184] However, the conclusions drawn from such studies are equivocal: Gottschall and Palmer[184] reported that by using a high cadence (FCC +20%) during a 30-minute cycle time trial, the subsequent 3 km running performance was Sports Med 2009; 39 (3)

Physiological Differences Between Cycling and Running

1 minute faster than by cycling at slow cadence (FCC -20%). This was due to an increased stride frequency whereas stride length was unchanged. Bernard et al.[95] reported that pedalling cadence (60, 80 or 100 rpm) has a short-term effect since during the first 500 m of a subsequent run, stride rate and running velocity were significantly higher after cycling at 80 or 100 rpm than at 60 rpm. Interestingly, the low cadence induced a deteriorated economy during the first part of the running bout. However, pedalling cadence did not influence overall subsequent 3 km performance. The same group reported contradictory results, since Vercruyssen et al.[87,89] recommended the use of low pedalling cadences. By comparing low (EOC; 75 rpm), medium (FCC; 81 rpm) and high pedalling cadence (‘mechanical optimal cadence’ [MOC], 90 rpm), they showed oxygen consumption during a subsequent 15-minute treadmill run to be increased only in the MOC (+4.1%) and FCC (+3.6%) conditions when compared with an isolated run. Recently by comparing the metabolic responses and time. to exhaustion during a running test at 85% of VO2max following three conditions of 30minute cycling with the last 10 minutes performed at different pedalling cadence (FCC = 94 rpm; FCC -20% = 74 rpm and FCC +20% = 109 rpm), Vercruyssen et al.[89] confirmed that the low cadence (FCC -20%) induces a lower energy expenditure during cycling, leading to an increased time to exhaustion during running. It is unclear how these later results are relevant in ‘real triathlon’, where running performance might be limited by other factors than metabolic ones. In addition, the stochastic nature of the cycling bout in triathlon is now well described[185-187] and there is an acceleration in the final portion of the cycling bout to enter the transition area in a good position. How the change in speed affects the change in pedalling cadence and how the FCC influences the subsequent run is still under debate. Whether differences in muscle contraction frequency influenced by different volume and intensity during cycling and running training affect the physiological adaptation in these exercise modes is also not known. ª 2009 Adis Data Information BV. All rights reserved.

199

2.4.3 Neuromuscular Fatigue

Since the type of muscle contraction and potential muscle damage are different between cycling and running, the neuromuscular fatigue induced by prolonged exercise probably originates from different sites (central, i.e. spinal and supraspinal, vs peripheral) and leads to a different level of strength loss (for a review see Millet and Lepers[188]). In running, the decrease in isometric strength loss is proportional to the duration of the exercise,[188-190] whereas it is less obvious in cycling.[191,192] The decrease in concentric strength is less than in isometric strength but the reasons are unclear. By using different methods such as the ratio of the root mean square (RMS) of the EMG recorded during MVC normalized by the muscle compound action potential (M-wave) amplitude (RMS/M), recent experiments have shown that there is a difference in the contribution of central fatigue between cycling and running. After 2 hours of prolonged cycling, the decrease in RMS (vastus lateralis and vastus medialis) was of the same extent (10%) as the decrease in M-wave amplitude,[181] showing that central fatigue did not contribute to the observed 13% decrease in strength. Contradictory central fatigue was observed after a prolonged 30 km run.[190] Two methods have been recently used to evidence deficit in muscle activation and therefore central fatigue after prolonged exercise; first, the change in the ratio between MVC and the mechanical response to an electrically evoked contraction at high frequency (80 Hz), and, second, the twitch interpolation technique where the ratio between a twitch superposed to a MCV and the twitch evoked on the muscle relaxed indicates the extent of the activation deficit. After a prolonged run, the activation deficit was shown in several studies in knee extensors[189,190] and plantar flexors,[193] whereas for the same long duration (>4 hours), this central fatigue was not observed in cycling.[181,194] However, the activation deficit was observed in cycling at higher intensity[192,195] and is probably induced at the spinal level by the presynaptic inhibition of the a-motoneuron due to metabolic causes. The activity of electromyography (EMG) has been used to examine muscle recruitment Sports Med 2009; 39 (3)

Millet et al.

200

patterns during exercise. Various authors have used EMG to establish a central fatigue component during prolonged endurance exercise.[191,196] EMG recorded during evoked contraction (M-wave) has also been used to identify peripheral fatigue as evidenced by a decrease in the action potential propagation on sarcolemna (increase in duration of M-wave) or lesser sarcolemna excitability (decrease in M-wave amplitude). There is no difference in the change in neuromuscular propagation induced by prolonged exercise between cycling and running since the decreases in amplitude (1–13%) and duration (1–24%) of M-wave are similar.[132,181,189-193,196] During prolonged exercise either in running or in cycling, the sarcolemna excitability appears to not be an important factor of peripheral fatigue. The changes in evoked twitch mechanical responses (DP; %) at low (20 Hz) and high (80 Hz) frequency tetanus and their ratio (P20/P80) indicates change in excitation-contraction coupling. High-frequency fatigue was observed after prolonged running[190] but not cycling,[194] whereas low-frequency fatigue has not been observed either after prolonged cycling or – surprisingly – after running. One may therefore speculate that the muscle damages induced by running are not important enough for inducing greater low-frequency fatigue than in cycling. However, further experiments are required. EMG measurements have not been obtained and compared during incremental cycling and running tests in triathletes or in single sport athletes. Therefore using EMG in combination with metabolic variables, it is possible that there is a difference in CNS or peripheral muscle innervation limiting muscle contraction during incremental running or cycling tasks. In one study, Bijker et al.[171] measured the relationship between mechanical power output, efficiency and EMG activity of the quadriceps muscle during incremental exercise in cycling and running. The results demonstrated that in contrast to cycling, during running EMG was not related to mechanical power output. These authors concluded that series elastic energy dictated recruitment pattern during running. This can be viewed as a considerable influence on the physiological responses during maximum running and cycling ª 2009 Adis Data Information BV. All rights reserved.

exercise. It would be of interest to replicate this study in trained cyclists, runners and triathletes. 3. Summary and Conclusions Despite treadmill running potentially utilizing more muscle mass, the majority of studies in. dicate that runners achieve a higher VO2max . on treadmill whereas cyclists can achieve a VO2max value in cycle ergometry. similar to that in treadmill running. Hence, VO2max is specific to the exercise mode (i.e. running or cycling). The data from the available studies also seems to indicate that the muscles adapt specifically to a given exercise task over a period of time resulting in an improvement in submaximal physiological variables such as the anaerobic . threshold in some cases without a change in VO2max. However, this effect is probably larger in cycling than in running. At the same time, skill influencing motor unit recruitment patterns is an important influence on the AT in cycling. Furthermore, it is likely that there is more physiological training transfer from running to cycling than vice versa. In triathletes, the majority of data demonstrate that there are generally no large differences in . VO2max measured in cycle ergometry and treadmill running. Therefore, it seems likely that triathletes .adapt in a similar way to cyclists and exhibit a VO2max in cycling that is similar to that in treadmill running. The data concerning the AT in cycling and running in triathletes are conflicting. This is likely to be due to a combination of athlete actual training load and past training history in these particular sports. The mechanisms surrounding the differences in the AT . together with VO2max in cycling and running are not largely understood but are probably due to the . relative adaptation of cardiac output influencing VO2max and also the recruitment of muscle mass in combination with the oxidative capacity of this mass influencing the AT. Several other physiological differences between cycling and running have been addressed since they are potential important factors at exhaustion: HR is different between the two activities both for maximal and submaximal intensities. The delta efficiency is higher in running. Differences in ventilatory Sports Med 2009; 39 (3)

Physiological Differences Between Cycling and Running

responses to exercise (exercise-induced arterial hypoxaemia, O2 diffusion capacity, ventilatory fatigue and pulmonary mechanics) have been reported between running and cycling, and ventilation is more impaired in cycling than running. Several mechanisms suggest that the muscle pump efficiency is greater in running than in cycling. It has also been shown that pedalling cadence affects the metabolic response during cycling but also during a subsequent running bout. However, the optimal cadence is still debated. Central fatigue and decrease in maximal force are more important after prolonged exercise in running than in cycling. All these findings might influence the training content and crosstraining effects in triathletes. However, to date very little information on volume/intensity of training in elite triathletes has been reported and there is no experiment that investigates the effects of changing one training parameter on overall triathlon performance.

201

8.

9.

10.

11.

12. 13.

14.

15.

16.

Acknowledgements No sources of funding were used to assist in the preparation of this article. The authors have no conflicts of interest that are directly relevant to the content of this article.

17.

18.

19.

References 1. Joyner MJ, Coyle EF. Endurance exercise performance: the physiology of champions. J Physiol 2008 Jan 1; 586 (1): 35-44 2. Loy SF, Hoffmann JJ, Holland GJ. Benefits and practical use of cross-training in sports. Sports Med 1995 Jan; 19 (1): 1-8 3. Tanaka H. Effects of cross-training: transfer of training effects on VO2max between cycling, running and swimming. Sports Med 1994 Nov; 18 (5): 330-9 4. Sleivert GG, Rowlands DS. Physical and physiological factors associated with success in the triathlon. Sports Med 1996 Jul; 22 (1): 8-18 5. Pechar GS, McArdle WD, Katch FI, et al. Specificity of cardiorespiratory adaptation to bicycle and treadmill training. J Appl Physiol 1974 Jun; 36 (6): 753-6 6. Withers RT, Sherman WM, Miller JM, et al. Specificity of the anaerobic threshold in endurance trained cyclists and runners. Eur J Appl Physiol Occup Physiol 1981; 47 (1): 93-104 7. Fernhall B, Kohrt W. The effect of training specificity on maximal and submaximal physiological responses to

ª 2009 Adis Data Information BV. All rights reserved.

20.

21.

22.

23.

24.

25.

treadmill and cycle ergometry. J Sports Med Phys Fitness 1990 Sep; 30 (3): 268-75 Basset FA, Boulay MR. Specificity of treadmill and cycle ergometer tests in triathletes, runners and cyclists. Eur J Appl Physiol 2000 Feb; 81 (3): 214-21 Hue O, Le Gallais D, Chollet D, et al. Ventilatory threshold and maximal oxygen uptake in present triathletes. Can J Appl Physiol 2000 Apr; 25 (2): 102-13 Schneider DA, Lacroix KA, Atkinson GR, et al. Ventilatory threshold and maximal oxygen uptake during cycling and running in triathletes. Med Sci Sports Exerc 1990 Apr; 22 (2): 257-64 Millet GP, Dreano P, Bentley DJ. Physiological characteristics of elite short- and long-distance triathletes. Eur J Appl Physiol 2003 Jan; 88 (4-5): 427-30 Kreider RB. Ventilatory threshold in swimming, cycling and running in triathletes. Int J Sports Med 1988; 9: 147-8 Millet GP, Candau RB, Barbier B, et al. Modelling the transfers of training effects on performance in elite triathletes. Int J Sports Med 2002 Jan; 23 (1): 55-63 Astrand PO, Saltin B. Maximal oxygen uptake and heart rate in various types of muscular activity. J Appl Physiol 1961 Nov; 16: 977-81 Saltin B. The interplay between peripheral and central factors in the adaptive response to exercise and training. Ann N Y Acad Sci 1977; 301: 224-31 Saltin B, Nazar K, Costill DL, et al. The nature of the training response; peripheral and central adaptations of one-legged exercise. Acta Physiol Scand 1976 Mar; 96 (3): 289-305 Gleser MA, Horstman DH, Mello RP. The effect on VO2 max of adding arm work to maximal leg work. Med Sci Sports 1974 Summer; 6 (2): 104-7 Secher NH, Ruberg-Larsen N, Binkhorst RA, et al. Maximal oxygen uptake during arm cranking and combined arm plus leg exercise. J Appl Physiol 1974 May; 36 (5): 515-8 Reybrouck T, Heigenhauser GF, Faulkner JA. Limitations to maximum oxygen uptake in arms, leg, and combined arm-leg ergometry. J Appl Physiol 1975 May; 38 (5): 774-9 Stenberg J, Astrand PO, Ekblom B, et al. Hemodynamic response to work with different muscle groups, sitting and supine. J Appl Physiol 1967 Jan; 22 (1): 61-70 Hermansen L, Saltin B. Oxygen uptake during maximal treadmill and bicycle exercise. J Appl Physiol 1969 Jan; 26 (1): 31-7 Hermansen L, Ekblom B, Saltin B. Cardiac output during submaximal and maximal treadmill and bicycle exercise. J Appl Physiol 1970 Jul; 29 (1): 82-6 McArdle WD, Magel JR. Physical work capacity and maximum oxygen uptake in treadmill and bicycle exercise. Med Sci Sports 1970 Fall; 2 (3): 118-23 Faulkner JA, Roberts DE, Elk RL, et al. Cardiovascular responses to submaximum and maximum effort cycling and running. J Appl Physiol 1971 Apr; 30 (4): 457-61 Katch FI, McArdle WD, Pechar GS. Relationship of maximal leg force and leg composition to treadmill and bicycle ergometer maximum oxygen uptake. Med Sci Sports 1974 Spring; 6 (1): 38-43

Sports Med 2009; 39 (3)

202

26. Davis JA, Vodak P, Wilmore JH, et al. Anaerobic threshold and maximal aerobic power for three modes of exercise. J Appl Physiol 1976 Oct; 41 (4): 544-50 27. Hagberg JM, Giese MD, Schneider RB. Comparison of the three procedures for measuring VO2 max in competitive cyclists. Eur J Appl Physiol Occup Physiol 1978 Jul 17; 39 (1): 47-52 28. Matsui H, Kitamura K, Miyamura M. Oxygen uptake and blood flow of the lower limb in maximal treadmill and bicycle exercise. Eur J Appl Physiol Occup Physiol 1978 Dec 15; 40 (1): 57-62 29. Miles DS, Critz JB, Knowlton RG. Cardiovascular, metabolic, and ventilatory responses of women to equivalent cycle ergometer and treadmill exercise. Med Sci Sports Exerc 1980 Spring; 12 (1): 14-9 30. Moreira-da-Costa M, Russo AK, Picarro IC, et al. Maximal oxygen uptake during exercise using trained or untrained muscles. Braz J Med Biol Res 1984; 17 (2): 197-202 31. Jacobs I, Sjodin B. Relationship of ergometer-specific VO2 max and muscle enzymes to blood lactate during submaximal exercise. Br J Sports Med 1985 Jun; 19 (2): 77-80 32. Moreira-da-Costa M, Russo AK, Picarro IC, et al. Oxygen consumption and ventilation during constant-load exercise in runners and cyclists. J Sports Med Phys Fitness 1989 Mar; 29 (1): 36-44 33. Green HJ, Sutton J, Young P, et al. Operation Everest II: muscle energetics during maximal exhaustive exercise. J Appl Physiol 1989 Jan; 66 (1): 142-50 34. Bouckaert J, Vrijens J, Pannier JL. Effect of specific test procedures on plasma lactate concentration and peak oxygen uptake in endurance athletes. J Sports Med Phys Fitness 1990 Mar; 30 (1): 13-8 35. Hill DW, Halcomb JN, Stevens EC. Oxygen uptake kinetics during severe intensity running and cycling. Eur J Appl Physiol 2003 Aug; 89 (6): 612-8 36. Scott CB, Littlefield ND, Chason JD, et al. Differences in oxygen uptake but equivalent energy expenditure between a brief bout of cycling and running. Nutr Metab (Lond) 2006; 3: 1 37. Bentley DJ, Newell J, Bishop D. Incremental exercise test design and analysis: implications for performance diagnostics in endurance athletes. Sports Med 2007; 37 (7): 575-86 38. Midgley AW, Bentley DJ, Luttikholt H, et al. Challenging a dogma of exercise physiology: does an incremental exercise test for valid determination really need to last between 8-12 minutes? Sports Med 2008; 38 (6): 441-63 39. Stromme SB, Ingjer F, Meen HD. Assessment of maximal aerobic power in specifically trained athletes. J Appl Physiol 1977 Jun; 42 (6): 833-7 40. Ricci J, Leger LA. VO2max of cyclists from treadmill, bicycle ergometer and velodrome tests. Eur J Appl Physiol Occup Physiol 1983; 50 (2): 283-9 41. Coyle EF, Coggan AR, Hopper MK, et al. Determinants of endurance in well-trained cyclists. J Appl Physiol 1988 Jun; 64 (6): 2622-30 42. Mazzeo RS, Marshall P. Influence of plasma catecholamines on the lactate threshold during graded exercise. J Appl Physiol 1989 Oct; 67 (4): 1319-22

ª 2009 Adis Data Information BV. All rights reserved.

Millet et al.

43. Kravitz L, Robergs RA, Heyward VH, et al. Exercise mode and gender comparisons of energy expenditure at selfselected intensities. Med Sci Sports Exerc 1997 Aug; 29 (8): 1028-35 44. Zeni AI, Hoffman MD, Clifford PS. Energy expenditure with indoor exercise machines. JAMA 1996 May 8; 275 (18): 1424-7 45. Sedlock DA. Post-exercise energy expenditure after cycle ergometer and treadmill exercise. J Appl Sport Sci Res 1992; 6: 19-23 46. Coyle EF. Integration of the physiological factors determining endurance performance ability. Exerc Sport Sci Rev 1995; 23: 25-63 47. Robergs RA, Ghiasvand F, Parker D. Biochemistry of exercise-induced metabolic acidosis. Am J Physiol Regul Integr Comp Physiol 2004 Sep; 287 (3): R502-16 48. Loat CE, Rhodes EC. Relationship between the lactate and ventilatory thresholds during prolonged exercise. Sports Med 1993 Feb; 15 (2): 104-15 49. Svedahl K, MacIntosh BR. Anaerobic threshold: the concept and methods of measurement. Can J Appl Physiol 2003 Apr; 28 (2): 299-323 50. Hawley JA, Stepto NK. Adaptations to training in endurance cyclists: implications for performance. Sports Med 2001; 31 (7): 511-20 51. Hassmen P. Perceptual and physiological responses to cycling and running in groups of trained and untrained subjects. Eur J Appl Physiol Occup Physiol 1990; 60 (6): 445-51 52. Albrecht TL, Foster VL, Dickinson AL. Triathletes: exercise parameters measured during bicycle, swim bench, and treadmill testing [abstract]. Med Sci Sports Exerc 1986; 18: S86 53. Kohrt WM, Morgan DW, Bates B, et al. Physiological responses of triathletes to maximal swimming, cycling, and running. Med Sci Sports Exerc 1987 Feb; 19 (1): 51-5 54. O’Toole ML, Hiller DB, Crosby LO, et al. The ultraendurance triathlete: a physiological profile. Med Sci Sports Exerc 1987 Feb; 19 (1): 45-50 55. O’Toole M, Hiller WDB, Douglas PS. Cardiovascular responses to prolonged cycling and running. Ann Sports Med 1987; 3: 124-30 56. Roalstad MS. Physiologic testing of the ultraendurance triathlete. Med Sci Sports Exerc 1989 Oct; 21 (5 Suppl.): S200-4 57. Flynn MG, Costill DL, Kirwan JP, et al. Muscle fiber composition and respiratory capacity in triathletes. Int J Sports Med 1987 Dec; 8 (6): 383-6 58. Kreider RB, Boone T, Thompson WR, et al. Cardiovascular and thermal responses of triathlon performance. Med Sci Sports Exerc 1988 Aug; 20 (4): 385-90 59. Loftin M, Warren BL, Zingraf S, et al. Peak physiological function and performance of recreational triathletes. J Sports Med Phys Fitness 1988 Dec; 28 (4): 330-5 60. Dengel DR, Flynn MG, Costill DL, et al. Determinants of success during triathlon competition. Res Q Exerc Sport 1989 Sep; 60 (3): 234-8 61. Stein TP, Hoyt RW, Toole MO, et al. Protein and energy metabolism during prolonged exercise in trained athletes. Int J Sports Med 1989 Oct; 10 (5): 311-6

Sports Med 2009; 39 (3)

Physiological Differences Between Cycling and Running

62. Kohrt WM, O’Connor JS, Skinner JS. Longitudinal assessment of responses by triathletes to swimming, cycling, and running. Med Sci Sports Exerc 1989 Oct; 21 (5): 569-75 63. Millard-Stafford M, Sparling PB, Rosskopf LB, et al. Carbohydrate-electrolyte replacement during a simulated triathlon in the heat. Med Sci Sports Exerc 1990 Oct; 22 (5): 621-8 64. Rehrer NJ, Brouns F, Beckers EJ, et al. Gastric emptying with repeated drinking during running and bicycling. Int J Sports Med 1990 Jun; 11 (3): 238-43 65. Butts NK, Henry BA, McLean D. Correlations between VO2max and performance times of recreational triathletes. J Sports Med Phys Fitness 1991 Sep; 31 (3): 339-44 66. Deitrick RW. Physiological responses of typical versus heavy weight triathletes to treadmill and bicycle exercise. J Sports Med Phys Fitness 1991 Sep; 31 (3): 367-75 67. Medelli J, Maingourd Y, Bouferrache B, et al. Maximal oxygen uptake and aerobic-anaerobic transition on treadmill and bicycle in triathletes. Jpn J Physiol 1993; 43 (3): 347-60 68. Sleivert GG, Wenger HA. Physiological predictors of short-course triathlon performance. Med Sci Sports Exerc 1993 Jul; 25 (7): 871-6 69. Miura H, Ishiko T. Cardiorespiratory responses during a simulated triathlon. International council for health, physical education and recreation (ICHPER) 36th World Congress; 1993; Yokohama: 157-61 70. Murdoch SD, Bazzarre TL, Snider IP, et al. Differences in the effects of carbohydrate food form on endurance performance to exhaustion. Int J Sport Nutr 1993 Mar; 3 (1): 41-54 71. Miura H, Kitagawa K, Ishiko T, et al. Characteristics of VO2max and ventilatory threshold in triathletes. Jpn J Exerc Sports Physiol 1994; 1 (1): 99-106 72. Zhou S, Robson SJ, King MJ, et al. Correlations between short-course triathlon performance and physiological variables determined in laboratory cycle and treadmill tests. J Sports Med Phys Fitness 1997 Jun; 37 (2): 122-30 73. Roberts A, McElligott M. The relationship between strength and endurance in female triathletes. NSRC Scientific Report. Canberra (ACT): University of Canberra, 1995 74. Ruby B, Robergs R, Leadbetter G, et al. Cross-training between cycling and running in untrained females. J Sports Med Phys Fitness 1996 Dec; 36 (4): 246-54 75. Kerr CG, Trappe TA, Starling RD, et al. Hyperthermia during Olympic triathlon: influence of body heat storage during the swimming stage. Med Sci Sports Exerc 1998 Jan; 30 (1): 99-104 76. Derman KD, Hawley JA, Noakes TD, et al. Fuel kinetics during intense running and cycling when fed carbohydrate. Eur J Appl Physiol Occup Physiol 1996; 74 (1-2): 36-43 77. Miura H, Kitagawa K, Ishiko T. Economy during a simulated laboratory test triathlon is highly related to Olympic distance triathlon. Int J Sports Med 1997 May; 18 (4): 276-80 78. Hue O, Le Gallais D, Chollet D, et al. The influence of prior cycling on biomechanical and cardiorespiratory response profiles during running in triathletes. Eur J Appl Physiol Occup Physiol 1998; 77 (1-2): 98-105

ª 2009 Adis Data Information BV. All rights reserved.

203

79. Hue O, Le Gallais D, Boussana A, et al. Ventilatory responses during experimental cycle-run transition in triathletes. Med Sci Sports Exerc 1999 Oct; 31 (10): 1422-8 80. Miura H, Kitagawa K, Ishiko T. Characteristic feature of oxygen cost at simulated laboratory triathlon test in trained triathletes. J Sports Med Phys Fitness 1999 Jun; 39 (2): 101-6 81. Schabort EJ, Killian SC, St Clair Gibson A, et al. Prediction of triathlon race time from laboratory testing in national triathletes. Med Sci Sports Exerc 2000 Apr; 32 (4): 844-9 82. Hue O, Le Gallais D, Boussana A, et al. Performance level and cardiopulmonary responses during a cycle-run trial. Int J Sports Med 2000 May; 21 (4): 250-5 83. Toraa M, Friemel F. Fatigue of the respiratory muscles due to maximal exercise on 2 different ergometers. Can J Appl Physiol 2000 Apr; 25 (2): 87-101 84. Hue O, Le Gallais D, Boussana A, et al. Catecholamine, blood lactate and ventilatory responses to multi-cycle-run blocks. Med Sci Sports Exerc 2000 Sep; 32 (9): 1582-6 85. Hue O, Le Gallais D, Prefaut C. Specific pulmonary responses during the cycle-run succession in triathletes. Scand J Med Sci Sports 2001 Dec; 11 (6): 355-61 86. Hue O, Galy O, Le Gallais D, et al. Pulmonary responses during the cycle-run succession in elite and competitive triathletes. Can J Appl Physiol 2001 Dec; 26 (6): 559-73 87. Vercruyssen F, Brisswalter J, Hausswirth C, et al. Influence of cycling cadence on subsequent running performance in triathletes. Med Sci Sports Exerc 2002 Mar; 34 (3): 530-6 88. Basset F, Boulay MR. Treadmill and cycle ergometer tests are interchangeable to monitor triathletes annual training. J Sports Sci Med 2003; 2 (3): 110-6 89. Vercruyssen F, Suriano R, Bishop D, et al. Cadence selection affects metabolic responses during cycling and subsequent running time to fatigue. Br J Sports Med 2005 May; 39 (5): 267-72 90. Schneider DA, Pollack J. Ventilatory threshold and maximal oxygen uptake during cycling and running in female triathletes. Int J Sports Med 1991 Aug; 12 (4): 379-83 91. O’Toole ML, Douglas PS. Applied physiology of triathlon. Sports Med 1995 Apr; 19 (4): 251-67 92. Miura H, Kitagawa K, Ishiko T. Characteristics of cardiorespiratory responses to the latter stage of a simulated triathlon. Jpn J Phys Fitness Sports Med 1994; 43: 381-8 93. De Vito G, Bernardi M, Sproviero E, et al. Decrease of endurance performance during Olympic triathlon. Int J Sports Med 1995 Jan; 16 (1): 24-8 94. 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 95. Bernard T, Vercruyssen F, Grego F, et al. Effect of cycling cadence on subsequent 3 km running performance in well trained triathletes. Br J Sports Med 2003 Apr; 37 (2): 154-18; discussion 9 96. Galy O, Hue O, Boussana A, et al. Effects of the order of running and cycling of similar intensity and duration on pulmonary diffusing capacity in triathletes. Eur J Appl Physiol 2003 Nov; 90 (5-6): 489-95

Sports Med 2009; 39 (3)

204

97. Millet GP, Bentley DJ. The physiological responses to running after cycling in elite junior and senior triathletes. Int J Sports Med 2004 Apr; 25 (3): 191-7 98. Galy O, Manetta J, Coste O, et al. Maximal oxygen uptake and power of lower limbs during a competitive season in triathletes. Scand J Med Sci Sports 2003 Jun; 13 (3): 185-93 99. Bassett Jr DR, Howley ET. Limiting factors for maximum oxygen uptake and determinants of endurance performance. Med Sci Sports Exerc 2000 Jan; 32 (1): 70-84 100. Sloniger MA, Cureton KJ, Prior BM, et al. Lower extremity muscle activation during horizontal and uphill running. J Appl Physiol 1997 Dec; 83 (6): 2073-9 101. Bolognesi M. Ventilatory threshold and maximal oxygenuptake during cycling and running in duathletes. Med Sport 1997; 50: 209-16 102. Vleck VE, Bentley DJ, Millet GP, et al. Pacing during an elite Olympic distance triathlon: comparison between male and female competitors. J Sci Med Sport 2008; 11 (4): 424-32 103. Roecker K, Striegel H, Dickhuth HH. Heart-rate recommendations: transfer between running and cycling exercise? Int J Sports Med 2003 Apr; 24 (3): 173-8 104. DiCarlo LJ, Sparling PB, Millard-Stafford ML, et al. Peak heart rates during maximal running and swimming: implications for exercise prescription. Int J Sports Med 1991 Jun; 12 (3): 309-12 105. O’Toole ML, Douglas PS, Hiller WD. Use of heart rate monitors by endurance athletes: lessons from triathletes. J Sports Med Phys Fitness 1998 Sep; 38 (3): 181-7 106. Ray CA, Cureton KJ, Ouzts HG. Postural specificity of cardiovascular adaptations to exercise training. J Appl Physiol 1990 Dec; 69 (6): 2202-8 107. Kenny GP, Reardon FD, Marion A, et al. A comparative analysis of physiological responses at submaximal workloads during different laboratory simulations of field cycling. Eur J Appl Physiol Occup Physiol 1995; 71 (5): 409-15 108. Gilman MB. The use of heart rate to monitor the intensity of endurance training. Sports Med 1996 Feb; 21 (2): 73-9 109. Foster C, Lucia A. Running economy: the forgotten factor in elite performance. Sports Med 2007; 37 (4-5): 316-9 110. Saunders PU, Pyne DB, Telford RD, et al. Factors affecting running economy in trained distance runners. Sports Med 2004; 34 (7): 465-85 111. Hausswirth C, Bigard AX, Berthelot M, et al. Variability in energy cost of running at the end of a triathlon and a marathon. Int J Sports Med 1996 Nov; 17 (8): 572-9 112. Hausswirth C, Bigard AX, Guezennec CY. Relationships between running mechanics and energy cost of running at the end of a triathlon and a marathon. Int J Sports Med 1997 Jul; 18 (5): 330-9 113. Hausswirth C, Brisswalter J, Vallier JM, et al. Evolution of electromyographic signal, running economy, and perceived exertion during different prolonged exercises. Int J Sports Med 2000 Aug; 21 (6): 429-36 114. Hausswirth C, Lehenaff D. Physiological demands of running during long distance runs and triathlons. Sports Med 2001; 31 (9): 679-89 115. Guezennec CY, Vallier JM, Bigard AX, et al. Increase in energy cost of running at the end of a triathlon. Eur J Appl Physiol Occup Physiol 1996; 73 (5): 440-5

ª 2009 Adis Data Information BV. All rights reserved.

Millet et al.

116. Millet GP, Millet GY, Hofmann MD, et al. Alterations in running economy and mechanics after maximal cycling in triathletes: influence of performance level. Int J Sports Med 2000 Feb; 21 (2): 127-32 117. Boone T, Kreider RB. Bicycle exercise before running: effect on performance. Ann Sports Med 1986; 3: 25-9 118. Millet GP, Vleck VE. Physiological and biomechanical adaptations to the cycle to run transition in Olympic triathlon: review and practical recommendations for training. Br J Sports Med 2000 Oct; 34 (5): 384-90 119. Millet GP, Millet GY, Candau RB. Duration and seriousness of running mechanics alterations after maximal cycling in triathletes: influence of the performance level. J Sports Med Phys Fitness 2001 Jun; 41 (2): 147-53 120. Jones AM. The physiology of the world record holder for the women’s marathon. Int J Sports Sci Coaching 2006; 1 (2): 101-15 121. Lucia A, Esteve-Lanao J, Olivan J, et al. Physiological characteristics of the best Eritrean runners-exceptional running economy. Appl Physiol Nutr Metab 2006 Oct; 31 (5): 530-40 122. Lucia A, Olivan J, Bravo J, et al. The key to top-level endurance running performance: a unique example. Br J Sports Med 2008; 42 (3): 172-4 123. Billat V, Lepretre PM, Heugas AM, et al. Training and bioenergetic characteristics in elite male and female Kenyan runners. Med Sci Sports Exerc 2003 Feb; 35 (2): 297-304; discussion 5-6 124. Billat VL, Demarle A, Slawinski J, et al. Physical and training characteristics of top-class marathon runners. Med Sci Sports Exerc 2001 Dec; 33 (12): 2089-97 125. Conley DL, Krahenbuhl GS, Burkett LN, et al. Following Steve Scott: physiological changes accompanying training. Phys Sportsmed 1984; 12: 103-6 126. Saltin B, Larsen H, Terrados N, et al. Aerobic exercise capacity at sea level and at altitude in Kenyan boys, junior and senior runners compared with Scandinavian runners. Scand J Med Sci Sports 1995 Aug; 5 (4): 209-21 127. Saltin B, Kim CK, Terrados N, et al. Morphology, enzyme activities and buffer capacity in leg muscles of Kenyan and Scandinavian runners. Scand J Med Sci Sports 1995 Aug; 5 (4): 222-30 128. Gaesser GA, Poole DC. The slow component of oxygen uptake kinetics in humans. Exerc Sport Sci Rev 1996; 24: 35-71 129. Bijker KE, De Groot G, Hollander AP. Delta efficiencies of running and cycling. Med Sci Sports Exerc 2001 Sep; 33 (9): 1546-51 130. Asmussen E, Bonde-Petersen F. Apparent efficiency and storage of elastic energy in human muscles during exercise. Acta Physiol Scand 1974 Dec; 92 (4): 537-45 131. Zacks RM. The mechanical efficiencies of running and bicycling against a horizontal impeding force. Int Z Angew Physiol 1973 Jul 20; 31 (4): 249-58 132. Avela J, Kyrolainen H, Komi PV, et al. Reduced reflex sensitivity persists several days after long-lasting stretchshortening cycle exercise. J Appl Physiol 1999 Apr; 86 (4): 1292-300 133. Farley CT, Gonzalez O. Leg stiffness and stride frequency in human running. J Biomech 1996 Feb; 29 (2): 181-6

Sports Med 2009; 39 (3)

Physiological Differences Between Cycling and Running

134. Kram R. Muscular force or work: what determines the metabolic energy cost of running? Exerc Sport Sci Rev 2000 Jul; 28 (3): 138-43 135. Kram R, Taylor CR. Energetics of running: a new perspective. Nature 1990 Jul 19; 346 (6281): 265-7 136. Richardson RS, Harms CA, Grassi B, et al. Skeletal muscle: master or slave of the cardiovascular system? Med Sci Sports Exerc 2000 Jan; 32 (1): 89-93 137. di Prampero PE. Factors limiting maximal performance in humans. Eur J Appl Physiol 2003 Oct; 90 (3-4): 420-9 138. Noakes TD. Maximal oxygen uptake: ‘‘classical’’ versus ‘‘contemporary’’ viewpoints: a rebuttal. Med Sci Sports Exerc 1998 Sep; 30 (9): 1381-98 139. Levine BD. VO2max: what do we know, and what do we still need to know? J Physiol 2008; 586: 25-34 140. Prefaut C, Durand F, Mucci P, et al. Exercise-induced arterial hypoxaemia in athletes: a review. Sports Med 2000 Jul; 30 (1): 47-61 141. Galy O, Le Gallais D, Hue O, et al. Is exercise-induced arterial hypoxemia in triathletes dependent on exercise modality? Int J Sports Med 2005 Nov; 26 (9): 719-26 142. Powers SK, Lawler J, Dempsey JA, et al. Effects of incomplete pulmonary gas exchange on VO2 max. J Appl Physiol 1989 Jun; 66 (6): 2491-5 143. Green HJ, Carter S, Grant S, et al. Vascular volumes and hematology in male and female runners and cyclists. Eur J Appl Physiol Occup Physiol 1999 Feb; 79 (3): 244-50 144. Galy O, Hue O, Boussana A, et al. Blood rheological responses to running and cycling: a potential effect on the arterial hypoxemia of highly trained athletes? Int J Sports Med 2005 Jan-Feb; 26 (1): 9-15 145. Boussana A, Galy O, Hue O, et al. The effects of prior cycling and a successive run on respiratory muscle performance in triathletes. Int J Sports Med 2003 Jan; 24 (1): 63-70 146. Boussana A, Hue O, Matecki S, et al. The effect of cycling followed by running on respiratory muscle performance in elite and competition triathletes. Eur J Appl Physiol 2002 Aug; 87 (4-5): 441-7 147. Boussana A, Matecki S, Galy O, et al. The effect of exercise modality on respiratory muscle performance in triathletes. Med Sci Sports Exerc 2001 Dec; 33 (12): 2036-43 148. Hue O, Boussana A, Le Gallais D, et al. Pulmonary function during cycling and running in triathletes. J Sports Med Phys Fitness 2003 Mar; 43 (1): 44-50 149. Smith TB, Hopkins WG, Taylor NA. Respiratory responses of elite oarsmen, former oarsmen, and highly trained non-rowers during rowing, cycling and running. Eur J Appl Physiol Occup Physiol 1994; 69 (1): 44-9 150. Gavin TP, Stager JM. The effect of exercise modality on exercise-induced hypoxemia. Respir Physiol 1999 May 3; 115 (3): 317-23 151. Hopkins SR, Barker RC, Brutsaert TD, et al. Pulmonary gas exchange during exercise in women: effects of exercise type and work increment. J Appl Physiol 2000 Aug; 89 (2): 721-30 152. Hill NS, Jacoby C, Farber HW. Effect of an endurance triathlon on pulmonary function. Med Sci Sports Exerc 1991 Nov; 23 (11): 1260-4

ª 2009 Adis Data Information BV. All rights reserved.

205

153. Bonsignore MR, Morici G, Abate P, et al. Ventilation and entrainment of breathing during cycling and running in triathletes. Med Sci Sports Exerc 1998 Feb; 30 (2): 239-45 154. Ekblom B. Effect of physical training on oxygen transport system in man. Acta Physiol Scand Suppl 1968; 328:1-45 155. Saltin B, Blomqvist G, Mitchell JH, et al. Response to exercise after bed rest and after training. Circulation 1968 Nov; 38 (5 Suppl.): VII1-78 156. Delp MD, Laughlin MH. Regulation of skeletal muscle perfusion during exercise. Acta Physiol Scand 1998 Mar; 162 (3): 411-9 157. Laaksonen MS, Kivela R, Kyrolainen H, et al. Effects of exhaustive stretch-shortening cycle exercise on muscle blood flow during exercise. Acta Physiol (Oxf) 2006 Apr; 186 (4): 261-70 158. Rowland TW. The circulatory response to exercise: role of the peripheral pump. Int J Sports Med 2001 Nov; 22 (8): 558-65 159. Sheriff DD. Muscle pump function during locomotion: mechanical coupling of stride frequency and muscle blood flow. Am J Physiol Heart Circ Physiol 2003 Jun; 284 (6): H2185-91 160. Noakes TD, St Clair Gibson A. Logical limitations to the ‘‘catastrophe’’ models of fatigue during exercise in humans. Br J Sports Med 2004 Oct; 38 (5): 648-9 161. St Clair Gibson A, Noakes TD. Evidence for complex system integration and dynamic neural regulation of skeletal muscle recruitment during exercise in humans. Br J Sports Med 2004 Dec; 38 (6): 797-806 162. Holloszy JO, Coyle EF. Adaptations of skeletal muscle to endurance exercise and their metabolic consequences. J Appl Physiol 1984 Apr; 56 (4): 831-8 163. Coyle EF, Feltner ME, Kautz SA, et al. Physiological and biomechanical factors associated with elite endurance cycling performance. Med Sci Sports Exerc 1991 Jan; 23 (1): 93-107 164. Weston AR, Myburgh KH, Lindsay FH, et al. Skeletal muscle buffering capacity and endurance performance after high-intensity interval training by well-trained cyclists. Eur J Appl Physiol Occup Physiol 1997; 75 (1): 7-13 165. Green HJ, Patla AE. Maximal aerobic power: neuromuscular and metabolic considerations. Med Sci Sports Exerc 1992 Jan; 24 (1): 38-46 166. Aunola S, Marniemi J, Alanen E, et al. Muscle metabolic profile and oxygen transport capacity as determinants of aerobic and anaerobic thresholds. Eur J Appl Physiol Occup Physiol 1988; 57 (6): 726-34 167. Ivy JL, Costill DL, Maxwell BD. Skeletal muscle determinants of maximum aerobic power in man. Eur J Appl Physiol Occup Physiol 1980; 44 (1): 1-8 168. Coyle EF, Sidossis LS, Horowitz JF, et al. Cycling efficiency is related to the percentage of type I muscle fibers. Med Sci Sports Exerc 1992 Jul; 24 (7): 782-8 169. Horowitz JF, Sidossis LS, Coyle EF. High efficiency of type I muscle fibers improves performance. Int J Sports Med 1994 Apr; 15 (3): 152-7 170. Parkhouse WS, McKenzie DC, Hochachka PW, et al. Buffering capacity of deproteinized human vastus lateralis muscle. J Appl Physiol 1985 Jan; 58 (1): 14-7

Sports Med 2009; 39 (3)

Millet et al.

206

171. Bijker KE, de Groot G, Hollander AP. Differences in leg muscle activity during running and cycling in humans. Eur J Appl Physiol 2002 Oct; 87 (6): 556-61 172. Marcinik EJ, Potts J, Schlabach G, et al. Effects of strength training on lactate threshold and endurance performance. Med Sci Sports Exerc 1991 Jun; 23 (6): 739-43 173. Chapman AR, Vicenzino B, Blanch P, et al. Does cycling effect motor coordination of the leg during running in elite triathletes? J Sci Med Sport 2008; 11 (4): 371-80 174. Borg G, Van den Burg M, Hassmen P. Relationships between perceived exertion, HR and HLa in cycling, running and walking. Scand J Sports Sci 1987; 9: 69-77 175. Marsh AP, Martin PE. Effect of cycling experience, aerobic power, and power output on preferred and most economical cycling cadences. Med Sci Sports Exerc 1997 Sep; 29 (9): 1225-32 176. Patterson RP, Moreno MI. Bicycle pedalling forces as a function of pedalling rate and power output. Med Sci Sports Exerc 1990 Aug; 22 (4): 512-6 177. Takaishi T, Yasuda Y, Ono T, et al. Optimal pedaling rate estimated from neuromuscular fatigue for cyclists. Med Sci Sports Exerc 1996 Dec; 28 (12): 1492-7 178. Lucia A, Hoyos J, Chicharro JL. Preferred pedalling cadence in professional cycling. Med Sci Sports Exerc 2001 Aug; 33 (8): 1361-6 179. Marsh AP, Martin PE, Foley KO. Effect of cadence, cycling experience, and aerobic power on delta efficiency during cycling. Med Sci Sports Exerc 2000 Sep; 32 (9): 1630-4 180. Marsh AP, Martin PE. The relationship between cadence and lower extremity EMG in cyclists and noncyclists. Med Sci Sports Exerc 1995 Feb; 27 (2): 217-25 181. Lepers R, Hausswirth C, Maffiuletti N, et al. Evidence of neuromuscular fatigue after prolonged cycling exercise. Med Sci Sports Exerc 2000 Nov; 32 (11): 1880-6 182. Vercruyssen F, Hausswirth C, Smith D, et al. Effect of exercise duration on optimal pedaling rate choice in triathletes. Can J Appl Physiol 2001 Feb; 26 (1): 44-54 183. Brisswalter J, Hausswirth C, Smith D, et al. Energetically optimal cadence vs. freely-chosen cadence during cycling: effect of exercise duration. Int J Sports Med 2000 Jan; 21 (1): 60-4 184. Gottschall JS, Palmer BM. The acute effects of prior cycling cadence on running performance and kinematics. Med Sci Sports Exerc 2002 Sep; 34 (9): 1518-22 185. Bentley DJ, Millet GP, Vleck VE, et al. Specific aspects of contemporary triathlon: implications for physiological

ª 2009 Adis Data Information BV. All rights reserved.

186.

187.

188.

189.

190.

191.

192.

193.

194.

195.

196.

analysis and performance. Sports Med 2002; 32 (6): 345-59 Bernard T, Vercruyssen F, Mazure C, et al. Constant versus variable-intensity during cycling: effects on subsequent running performance. Eur J Appl Physiol 2007 Jan; 99 (2): 103-11 Vleck VE, Burgi A, Bentley DJ. The consequences of swim, cycle, and run performance on overall result in elite olympic distance triathlon. Int J Sports Med 2006 Jan; 27 (1): 43-8 Millet GY, Lepers R. Alterations of neuromuscular function after prolonged running, cycling and skiing exercises. Sports Med 2004; 34 (2): 105-16 Millet GY, Lepers R, Maffiuletti NA, et al. Alterations of neuromuscular function after an ultramarathon. J Appl Physiol 2002 Feb; 92 (2): 486-92 Millet GY, Martin V, Lattier G, et al. Mechanisms contributing to knee extensor strength loss after prolonged running exercise. J Appl Physiol 2003 Jan; 94 (1): 193-8 Lepers R, Maffiuletti NA, Rochette L, et al. Neuromuscular fatigue during a long-duration cycling exercise. J Appl Physiol 2002 Apr; 92 (4): 1487-93 Lepers R, Millet GY, Maffiuletti NA. Effect of cycling cadence on contractile and neural properties of knee extensors. Med Sci Sports Exerc 2001 Nov; 33 (11): 1882-8 Racinais S, Girard O, Micallef JP, et al. Failed excitability of spinal motoneurons induced by prolonged running exercise. J Neurophysiol 2007 Jan; 97 (1): 596-603 Millet GY, Millet GP, Lattier G, et al. Alteration of neuromuscular function after a prolonged road cycling race. Int J Sports Med 2003 Apr; 24 (3): 190-4 Bentley DJ, Smith PA, Davie AJ, et al. Muscle activation of the knee extensors following high intensity endurance exercise in cyclists. Eur J Appl Physiol 2000 Mar; 81 (4): 297-302 Takaishi T, Yasuda Y, Moritani T. Neuromuscular fatigue during prolonged pedalling exercise at different pedalling rates. Eur J Appl Physiol Occup Physiol 1994; 69 (2): 154-8

Correspondence: Dr Gregoire P. Millet, ISSEP, University of Lausanne, CH-1015, Lausanne, Switzerland. E-mail: [email protected]

Sports Med 2009; 39 (3)

Sports Med 2009; 39 (3): 207-224 0112-1642/09/0003-0207/$49.95/0

REVIEW ARTICLE

ª 2009 Adis Data Information BV. All rights reserved.

Treatment of Common Deficits Associated with Chronic Ankle Instability Alison Holmes and Eamonn Delahunt School of Physiotherapy and Performance Science, University College Dublin, Health Sciences Centre, Belfield, Dublin, Ireland

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Proprioceptive Deficits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Interventions for Proprioceptive Deficits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Neuromuscular Deficits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Interventions for Neuromuscular Deficits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Postural Control Deficits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Interventions for Postural Control Deficits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. Strength Deficits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. Interventions for Strength Deficits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

207 208 209 212 212 214 215 218 219 221

Lateral ankle sprains are amongst the most common injuries incurred by athletes, with the high rate of reoccurrence after initial injury becoming of great concern. Chronic ankle instability (CAI) refers to the development of repetitive ankle sprains and persistent residual symptoms post-injury. Some of the initial symptoms that occur in acute sprains may persist for at least 6 months post-injury in the absence of recurrent sprains, despite the athlete having returned to full functional activity. CAI is generally thought to be caused by mechanical instability (MI) or functional instability (FI), or both. Although previously discussed as separate entities, recent research has demonstrated that deficits associated with both MI and FI may co-exist to result in CAI. For clinicians, the main deficits associated with CAI include deficits in proprioception, neuromuscular control, strength and postural control. Based on the literature reviewed, it does seem that subjects with CAI have a deficit in frontal plane ankle joint positional sense. Subjects with CAI do not appear to exhibit any increased latency in the peroneal muscles in response to an external perturbation. Preliminary data suggest that feed-forward neuromuscular control may be more important than feed-back neuromuscular control and interventions are now required to address deficits in feed-forward neuromuscular control. Balance training protocols have consistently been shown to improve postural stability in subjects with CAI. Subjects with CAI do not experience decreased peroneus longus strength, but instead may experience strength deficits in the ankle joint invertor muscles. These findings are of great

Holmes & Delahunt

208

clinical significance in terms of understanding the mechanisms and deficits associated with CAI. An appreciation of these is vital to allow clinicians to develop effective prevention and treatment programmes in relation to CAI.

Recent research has found the ankle to be the second most commonly injured body site in sport, with ankle sprain being the most common type of ankle injury.[1] Garrick[2] was one of the first to conclude that the lateral ligament complex of the ankle was one of the most commonly injured structures in athletes, with further research finding lateral ankle sprains to account for 85% of all ankle sprains.[3] Re-injury rates reported following lateral ankle sprains vary between studies, but have been found to be as high as 70%[4] and 80%[5] in certain sporting populations. A study by Gerber et al.[6] found that 55% of subjects at 6 weeks and 40% of subjects at 6 months continued to experience residual symptoms following initial ankle joint injury. Braun[7] found that 72% of subjects experienced residual symptoms up to 18 months post-injury. The development of residual symptoms, such as a subjective feeling of ‘giving way’, and recurrent inversion injury of the ankle has been termed chronic ankle instability (CAI).[8] CAI is generally attributed to mechanical instability (MI) or functional instability (FI) of the ankle joint.[8,9] FI is a subjectively reported phenomenon initially defined as a tendency for the ankle to ‘give way’ during normal activity.[10,11] More recently this definition has expanded to indicate that subjects with FI could experience one or more of the following: neuromuscular deficits, proprioceptive deficits, strength deficits and impaired postural control.[8,12] MI is defined as ankle movement beyond the physiological limit of the ankle’s range of motion and is often used in conjunction with the term ‘laxity’.[13] More recently, Hertel[8] has described MI as pathological laxity of the ankle secondary to damage of the ligamentous tissues that support the ankle joint, also concluding that arthrokinematic restrictions, degenerative changes and synovial changes are contributing factors associated with MI. Some authors have attempted to define MI and FI as entirely separate entities,[14] with various studies ª 2009 Adis Data Information BV. All rights reserved.

having provided evidence that FI can exist as a separate entity to MI.[15,16] However, other studies have proven that the two may also coexist.[17] It is generally accepted that although FI and MI may exist as separate entities, CAI usually occurs due to a combination of both.[8,12,18,19] This article examines the neuromuscular contributing factors and specific deficits associated with CAI, in particular with FI, as defined by Hertel,[8] and more specifically reviews the current different approaches used in treating these common deficits in subjects with CAI. This article addresses proprioceptive deficits, neuromuscular deficits, postural control deficits and strength deficits as these are readily amenable to physiotherapeutic intervention.[8] Degenerative and synovial changes are not readily treatable from a conservative perspective, while the exact contribution of arthrokinematic restrictions to the development of CAI is only an emerging concept.[8] 1. Proprioceptive Deficits Proprioception is the result of the cumulative neural input to the CNS, from mechanoreceptors in the joint capsules and ligaments, as well as the surrounding muscles and overlying skin.[20] The assessment of proprioception is divided into two distinct components, kinesthesia and joint position sense (JPS).[21] Ankle joint kinesthesia is usually assessed using the threshold-to-detection of passive motion (TTDPM), while JPS is assessed using both active and passive joint reproduction.[21] The ability to detect motion in the foot and to respond by making postural adjustments is widely considered crucial in the prevention of ankle injury.[22] Konradsen and Voigt[23] have proposed a biomechanical model to connect a deficit in ankle joint proprioception with the increased frequency of inversion injuries and ‘giving way’ episodes experienced in subjects with FI. This model suggests that during the terminal Sports Med 2009; 39 (3)

Treatment of Common Ankle Instability Deficits

swing phase of the gait cycle, a deficit in ankle joint proprioception could result in the lateral border of the foot colliding with the support surface and consequently resulting in a forced inversion injury or ‘giving way’ episode. Direct in vivo evidence to support this hypothesis has recently been provided by Delahunt et al.,[24] who have shown that during treadmill walking, subjects with FI have a decreased foot-floor clearance during the terminal swing phase of the gait cycle when compared with non-injured subjects. A number of studies have examined ankle joint kinesthesia in subjects with FI.[25-29] Of these studies, three examined sagittal plane movement.[25,27,28] The studies by Garn and Newton[25] and Mulloy Forkin et al.[27] indicated that subjects with FI had a significantly increased difficulty in detecting passive motion in the injured compared with the non-injured ankle joint. However, the study by Refshauge et al.[28] failed to show a difference in kinesthetic plantar flexion or dorsiflexion performance in subjects with FI compared with a non-injured control group. Two other studies have examined frontal plane movement,[26,29] the results of which are conflicting, with the study by Lentell et al.[26] showing a deficit in ankle joint inversion kinesthetic awareness, whereas the study by Hubbard and Kaminski[29] showed no deficits in either inversion or eversion kinesthesia. A number of studies have assessed ankle joint inversion JPS,[22,30-32] with all studies indicating that subjects with FI have deficits in either active or passive joint position replication. Based upon the hypothesis by Freeman,[10] it was widely believed that damage to the ankle joint lateral ligament complex produced deafferentation, consequently resulting in repeated inversion injury and episodes of ‘giving way’. However, a study by Konradsen et al.[33] refutes this hypothesis. These authors have shown that active JPS is not affected by the induction of ankle joint anaesthesia, with the authors concluding that muscle and tendon mechanoreceptors subserve active proprioceptive function. This observation concurs with current physiological understanding about the mechanisms underlying proprioception,[28] which indicate that ª 2009 Adis Data Information BV. All rights reserved.

209

muscle afferents play the most important proprioceptive role.[34] Possible reasons for the differences noted between studies for ankle joint sagittal plane kinesthetic performance include differences used for the classification of FI. The differences in inclusion criteria are highlighted in table I, which outlines the inclusion criteria used to define FI by all the studies included in the current review examining proprioceptive function in subjects with FI. This problem has been previously emphasized by Hubbard and Kaminski,[29] who state ‘‘We recommend that a standard set of criteria be used in the determination of FAI status and made available for future research. This precaution should enable a more accurate assessment of FAI, more consistency across studies, and easier comparison of data pools, while lessening the variability in study designs.’’ Another common limitation of the studies is that the majority of them do not conform to the principles of psychophysical testing procedures, which state that when examining thresholds, movement stimuli should be presented a large number of times, due to the variation in threshold with time.[28] The only study adhering to these principles was that of Refshauge et al.[28] Based on the literature presented above, there is conflicting evidence regarding the presence or absence of kinesthetic deficits in subjects with FI, while it does appear that subjects with FI do exhibit a deficit in frontal plane movement JPS. 2. Interventions for Proprioceptive Deficits A number of studies have been performed to examine the effectiveness of various intervention programmes aimed at addressing proprioceptive deficits in subjects with FI.[35-38] The results of these studies are discussed below and outlined in table II. Bernier and Perrin[35] examined the effects of a 6-week coordination training intervention on proprioception as measured by JPS, and postural sway in subjects with FI. They found post-test scores for JPS improved significantly from pretest scores in both the intervention and control groups, suggesting the presence of a learning Sports Med 2009; 39 (3)

Study

No. of sprains

Activity level

Unilateral or bilateral instability

Willems et al.[22]

>3 inversion sprains

Garn and Newton[25]

‡2 lateral ankle sprains, range 2–20

Involved in sports

Unilateral

Lentell et al.[26]

History of an inversion injury requiring protected weight-bearing and/or immobilization

Recreational athletic classes

Unilateral

Mulloy Forkin et al.[27]

>1 inversion injury

Gymnasts

Unilateral or bilateral

Refshauge et al.[28]

>3 inversion sprains

Hubbard and Kaminski[29]

Recruitment based on FAIQ. History of an inversion injury requiring protected weight-bearing and/or immobilization

Unilateral

Presence of weakness

Presence of pain

Presence of MI

Pain during heavy, intense loading

Did not measure MI

Weaker than stable ankle

More painful than stable ankle

Weaker than stable ankle

More painful than stable ankle

No MI (presence of MI was considered an exclusion criteria)

Function of the ankle

210

ª 2009 Adis Data Information BV. All rights reserved.

Table I. Inclusion criteria used to define functional instability (FI) by all the studies included in the current review examining proprioceptive function in subjects with FI (a blank cell indicates that this information was not specified) Presence of ‘giving way’ ‘Giving way’ present

Less functional than the noninjured ankle

‘Giving way’ present

Athletic activities; average 8.13 h/wk

Jerosch and Bischof[30]

>2 inversion injuries

Konradsen and Magnusson[32]

Repeated inversion injuries >7/wk

2 sessions per wk

FAIQ = Functional Ankle Instability Questionnaire; MI = mechanical instability.

MI present

Holmes & Delahunt

Sports Med 2009; 39 (3)

Boyle and Negus[31]

Treatment of Common Ankle Instability Deficits

effect. Overall, their results suggested that although coordination training may improve some aspects of postural sway, it had no significant effect on JPS. The authors concluded that their study did have limitations, such as lack of specificity between training and assessment measures and perhaps insufficient duration of training to produce an effect. Another limitation of this study was that it did not conform to the principles of psychometric testing procedures, with the authors only using two trials per subject for the calculation of mean joint angles. Eils and Rosenbaum[36] examined the effects of a 6-week multistation proprioceptive exercise programme in subjects with self-reported ankle instability. They found the intervention group showed a significant improvement in JPS (plantar flexion and dorsiflexion), postural sway and muscle reaction times. Furthermore, the results of a 1-year follow-up questionnaire (90% response rate) showed a significantly reduced frequency of ankle inversion episodes after the exercise programme of almost 60%, with most

211

subjects reporting a better feeling of stability and safety. These results suggest that an increased ability to detect the angle of the ankle joint as well as enhanced postural sway may help to prevent recurrent ankle joint injuries. The authors concluded that the multistation proprioceptive exercise programme could be recommended for prevention and rehabilitation of recurrent ankle inversion injuries and that the main advantage compared with other programmes is the relatively low training frequency of once per week and the possibility to perform this training in class or team situations. Two studies have examined the effect of ankle joint strengthening programmes on JPS,[37,38] both of which have shown positive findings. The study by Docherty et al.[37] indicated that subjects who took part in the exercise protocol exhibited a statistically significant improvement in inversion (pre-test 6.8 – 5.0, post-test 2.8 – 2.8) and plantar flexion (pre-test 7.9 – 6.0, post-test 1.4 – 0.9) JPS. The authors do provide an interesting hypothesis as to why strength training produced a change in proprioception.

Table II. Interventions for proprioceptive deficits Study

Design

Definition of FI and intervention applied

Outcome measures

Results

Bernier and Perrin[35]

Controlled group pre-/ post-intervention n = 45 (control = 14; Sham treatment = 14; exercise = 17)

FI = at least one significant ankle inversion sprain followed by repeated injury or repetitive episodes of ‘giving way’ 6-wk coordination and balance programme 3 ·/wk

JPS Postural sway

No significant differences found for postural sway or JPS post-intervention

Eils and Rosenbaum[36]

Controlled group pre-/ post-intervention n = 48 (exercise = 31; control = 17)

FI = repeated ankle inversion sprains and a self-reported subjective feeling of instability or giving way 6-wk progressive physiotherapeutic exercise programme consisting of 12 neuromuscular exercises 1 ·/wk

JPS 12 mo follow-up questionnaire

In the exercise group, the results showed a significant › in JPS. 90% of the exercise group returned the questionnaire 1 year after training. Evaluation showed an almost 60% ﬂ in frequency of ankle inversion sprains

Docherty et al.[37]

Controlled group pre-/ post-intervention n = 20 (exercise = 10; control = 10)

FI = >3 sprains in the last 5 y. One episode of ‘giving way’ in the last 12 mo 6-wk progressive resistance ankle joint strengthening 3 ·/wk

JPS

› in plantar flexion and inversion JPS

Sekir et al.[38]

Pre-/post-intervention n = 24

FI = 2 moderate sprains, repeated episodes of ‘giving way’ 6-wk isokinetic concentric inversion and eversion protocol 3 ·/wk

JPS

Improved JPS

FI = functional instability; JPS = joint position sense; post = post-test; pre = pre-test; ﬂ indicates decrease; › indicates increase.

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (3)

Holmes & Delahunt

212

They hypothesize that the strengthening protocol may have increased g-efferent activity, which in turn could render the muscle spindles more sensitive to stretch, resulting in greater acuity in sensing joint position. However, this hypothesis requires further validation. In the study by Sekir et al.,[38] there was a significant improvement in proprioception (pre-test at 10 inversion 2.35 – 1.16, post-test at 10 inversion 1.33 – 0.62; pre-test at 20 inversion 3.10 – 2.16, post-test at 20 inversion 2.19 – 0.98) following the strengthening programme. The link between strength training and proprioception requires further investigation. The reviewed literature does indicate that various exercise protocols can enhance the proprioceptive capabilities of subjects with FI. However, the exact consequences of this enhanced proprioceptive capability requires further investigation, as only the study by Eils and Rosenbaum[36] examined the long-term effects of their intervention; it must be noted though that the authors could not determine whether the effect was due to enhanced proprioception or postural control. 3. Neuromuscular Deficits Neuromuscular control has been defined as the unconscious activation of dynamic restraints occurring in preparation for and in response to joint motion and loading for the purpose of maintaining and restoring functional joint stability.[39] The basic premise underlying the articular deafferentation theory of FI originally developed by Freeman[10] was that damage to the ankle joint capsule and ligaments produced delayed and diminished reflex responses in the ankle joint evertor muscles. Following injury, the ankle joint evertors would not be able to respond quickly enough to any unexpected perturbation, rendering the ankle joint vulnerable to repeated inversion injury. Thus, many investigators have examined the response times of the peroneal muscles to an unexpected inversion perturbation in subjects with FI.[40-46] A full and comprehensive review of these studies can be found in a recently published article by Delahunt.[47] This article also outlines some of ª 2009 Adis Data Information BV. All rights reserved.

the potential reasons for the discrepancy in observed results, including different methodologies and different subject inclusion criteria, which are points also highlighted in an article by Riemann and Lephart.[39] It has been suggested that reflexive activity of the peroneus longus is unlikely to provide adequate joint protection during dynamic activity, and thus the functional relevance of investigating muscle response times to inversion perturbations requires consideration.[47] It must be noted that two studies examining peroneal muscle activity during jump landing in subjects with FI have failed to identify any decrease in post-landing reactive muscle activity.[48,49] Thus, both authors concluded that subjects with FI do not exhibit any deficits in peroneus longus reflexive muscle activity. However, both of these studies did find a decrease in peroneus longus muscle activity prior to contact with the ground, indicating that subjects with FI may experience a deficit in feed-forward neuromuscular control. Furthermore, the study by Delahunt et al.[49] also found a significant increase in ankle joint frontal plane movement (increased inverted position compared with noninjured control subjects) prior to contact with the ground, which coincided with the decrease in preinitial contact peroneus longus muscle activity. The combination of these two observed deficits could leave the ankle joint vulnerable to re-injury as an unexpected contact with the ground, such as alighting on a rutted playing field or an opponents foot, could force the ankle joint further into inversion resulting in a hyperinversion injury and subsequent sprain of the ankle joint lateral ligaments. The inclusion criteria used by the two studies investigating feed-forward neuromuscular control in subjects with FI are outlined in table III. 4. Interventions for Neuromuscular Deficits Two studies were identified in the literature that examined the effects of neuromuscular training on the response times of the lower limb musculature to sudden inversion perturbations. The results of these studies are discussed below and outlined in table IV. Sports Med 2009; 39 (3)

‘Giving way’ present during sporting activity Less functional than opposite ankle

‘Giving way’ present during sporting activity Less functional than opposite ankle

Presence of ‘giving way’

More painful than opposite ankle Chronically weaker than opposite ankle

More painful than opposite ankle Chronically weaker than opposite ankle

Presence of weakness

Presence of pain

Presence of MI

Function of the ankle

213

Minimum of two inversion sprains to one ankle Delahunt et al.[49]

ª 2009 Adis Data Information BV. All rights reserved.

MI = mechanical instability.

Minimum of two inversion sprains to one ankle Caulfield et al.[48]

Involved in sporting activity (type not specified)

Unilateral or bilateral instability Activity level No. of sprains Study

Table III. Inclusion criteria used to define functional instability (FI) by all the studies included in the current review examining feed-forward neuromuscular function in subjects with FI (a blank cell indicates that this information was not specified)

Treatment of Common Ankle Instability Deficits

Clark and Burden[50] examined the effects of a 4-week wobble-board training programme on the onset of tibialis anterior and peroneus longus muscle activity and the perception of stability in subjects with FI. Nineteen male subjects with FI were randomly assigned to a training or control group. The training group exercised three times per week by performing a range of exercises on a wobble board. Outcome measures were the onset latency of the tibialis anterior and peroneus longus muscles to a sudden inversion perturbation of 20 as well as perception of ankle joint stability as measured by the Ankle Joint Functional Assessment Tool (AJFAT). Results of the study indicated that subjects in the training group had a significant decrease in both tibialis anterior (29.9% decrease in the exercise group when compared with pre-intervention scores) and peroneus longus (31.2% decrease in the exercise group when compared with pre-intervention scores) onset latencies following completion of the exercise programme, which was not observed in the control group. Furthermore, subjects in the training group also reported a significant improvement in their pre-exercise AJFAT score (the mean percentage change over the programme for the exercise group was 28.4%). The authors do concede that the reduced latency observed following the training programme is unlikely to significantly enhance the torque-producing capabilities of the tibialis anterior and peroneus longus muscles. However, the observed decrease in latency could signify that the muscles are intrinsically stiffer. It has been shown that stiffer muscles can enhance muscle spindle activation via feed-forward neuromuscular control, resulting in a decreased time to initiation of a reflex response.[52] The authors did find that subjects who took part in the training programme had a significant improvement on the AJFAT. However, one limitation of this measure is that the subjects were not followed up for a period of time after the intervention, and thus the true significance of this improvement cannot be quantified as we do not know if the programme produced a decrease in ankle joint inversion injury incidence. A more recent study has shown that a specific balance training programme can reduce the Sports Med 2009; 39 (3)

Holmes & Delahunt

214

Table IV. Interventions for neuromuscular deficits (examination of response times to sudden inversion perturbation) Study

Design

Definition of FI

Outcome measures

Results

Clark and Burden[50]

Controlled group pre-/ post-intervention n = 19 (exercise = 10; control = 9)

FI = subjective complaint of weakness often in the absence of mechanical instability. Wobble board exercise programme 3 ·/wk for 4 wk

AJFAT. Tibialis anterior and peroneus longus onset latency in response to sudden 20 inversion

Subject’s perception of ankle stability › during and after the exercise programme in the intervention group. Onset latency of tibialis anterior and peroneus longus in response to 20 sudden inversion was significantly ﬂ

Akhbari et al.[51]

Pre-/post-intervention n = 15

FI = repetitive sprains or repetitive episodes of ‘giving way’ 4-wk Biodex Stability System, 12 min per session 3 ·/wk

Musculature onset and peak latency of peroneals, tibialis anterior and soleus

Post-training there was a significant ﬂ in muscle onset and peak latency for peroneals and tibialis anterior

AJFAT = Ankle Joint Functional Assessment Tool; FI = functional instability; post = post-test; pre = pre-test; ﬂ indicates decrease(d); › indicates increase(d).

peroneal and tibialis anterior onset times and peak latencies.[51] In this study, 15 subjects with FI underwent a 4-week balance training programme on the Biodex Stability System. Each training session lasted 12 minutes and was performed three times per week. No measures of perceived ankle joint stability or long-term follow-up of patients were measured. Thus, the clinical significance of the observed decrease in lower-limb onset timing and peak latencies requires consideration. It seems that the reflex response of the peroneal muscles is not fast enough to prevent a sudden unexpected inversion perturbation.[47] Thus, the theory of articular deafferentation and the subsequent reflex stabilization role of the peroneal muscles seems to be redundant. However, the feed-forward mechanism of ankle joint stability requires further investigation. Recent research has shown that subjects with FI exhibit a deficit in feed-forward neuromuscular control, whereby the peroneal muscles do not prepare the ankle joint complex for expected contact with the ground during jump landing.[48,49] Further intervention studies should examine the effect of specific training protocols on pre-initial contact ankle joint muscle activity and joint movement. 5. Postural Control Deficits Impaired postural control is frequently evident in subjects with both acute and repetitive ª 2009 Adis Data Information BV. All rights reserved.

ankle sprains.[10,25,27] Functional deficits in postural control are consistently recognized in subjects with CAI.[53,54] These postural deficits are most likely secondary to a combination of impaired neuromuscular control and proprioception.[8] The pronation and supination of the foot in single-leg stance to maintain the body’s centre of gravity above its base of support is termed ‘ankle strategy’.[13] Subjects with FI have been shown to lack ankle strategy and instead adopt a less efficient hip strategy to maintain single-leg stance.[55] Hip strategy creates large shear forces with the ground, thus potentially increasing ankle inversion and resulting in the ankle ‘giving way’.[56] Changes in central neural control in the presence of joint dysfunction are thought to lead to these postural alterations.[8] Originally deficits in balance and postural control were shown using a one-legged standing balance test.[10,14] This static test has been supplemented by the use of more dynamic tests such as the star excursion balance test (SEBT),[57-59] timeto-stabilization (TTS)[60,61] and the recently developed dynamic postural stability index,[62] all of which have shown postural control deficits in subjects with CAI. Using the SEBT, Olmsted et al.[58] found that subjects with FI had significantly decreased reach when compared with uninjured controls or their unaffected side. Gribble et al.[59] furthered these findings by demonstrating decreased reach as well as Sports Med 2009; 39 (3)

Treatment of Common Ankle Instability Deficits

decreased knee and hip flexion angles in subjects with FI during the SEBT. Both Ross et al.[60] and Brown and Mynark[61] have shown that subjects with CAI take longer to stabilize following dynamic activity than non-injured controls. Ross et al.[60] hypothesized that this increased time-to-stabilization could leave the ankle joint vulnerable to recurrent injury due to the inability to correctly execute proper landing techniques. Similar findings were reported by Wikstrom et al.[62] using another measure of dynamic postural control. Thus, review of the literature indicates that subjects with CAI do exhibit deficits in dynamic postural control, which may not be without potential functional consequence from an injury perspective. The inclusion criteria used by those studies investigating postural control in subjects with FI are outlined in table V. 6. Interventions for Postural Control Deficits The use of dynamic testing to highlight postural deficits in subjects with CAI has in turn led to the development of more dynamic rehabilitation methods. The results of these studies are discussed below and outlined in table VI. Two studies have investigated the effect of a balance training programme performed on a commercially available piece of equipment, the Biodex Stability System (Biodex Inc., Shirley, NY, USA).[51,63] Both studies indicated that the specific training programmes used resulted in increased balance performance as noted by a reduction in post-intervention stability index scores as compared with pre-intervention scores. However, both studies are limited because the outcome measures used were the same as techniques used for training, which could have led to a degree of learning technique. The use of another objective outcome measure in addition to a mode-specific one may have helped to clarify results. A number of studies have examined the effect of balance training on the excursion of the centre of pressure, or centre of pressure velocity.[36,64,65] All of these studies reported an improvement in balance performance as evidenced by a reducª 2009 Adis Data Information BV. All rights reserved.

215

tion in centre of pressure excursion or velocity. However, it must be noted that each of these studies used very different protocols. Eils and Rosenbaum[36] used a multistation proprioceptive training protocol, while the study by Kidgell et al.[65] examined the effects of a 6-week dura disc or mini-trampoline training programme. An initial study by Ross and Guskiewicz[66] examined the effects of a 6-week coordination training programme with and without stochastic resonance stimulation (SR) on dynamic postural stability of subjects with FI and with stable ankles. Subjects were assigned to one of three groups: traditional coordination training, coordination training with SR or a control group. Jump landing tests were performed to monitor progress, using anterior/posterior and medial/ lateral TTS figures. The intervention groups had decreased TTS results post-treatment, while no improvements were seen in the control group. Although statistically the improvements in both intervention groups were the same, effect sizes indicated that coordination training with SR improved TTS to a greater degree than coordination training alone. The authors concluded that although coordination training does improve postural stability in subjects with FI, SR stimulation may be considered as an alternative therapy as it may have more beneficial effects over a shorter duration. The effects of these two interventions were only compared over a 2- and 4-week period, however, with no longer term follow-up to monitor if there was any significant difference between recurrence of sprain figures between groups. A further study by Ross[68] examined noiseenhanced postural stability in subjects with FI. Subjects performed single-leg balance tests under SR and under control conditions. Subsensory stimulation was delivered to the ankle muscles and ligaments. Subjects were blinded to test conditions. The study concluded that postural instabilities associated with FI can be improved with an optimal noise intensity level of SR stimulation. The basic premise underlying SR is that it can cause sub-threshold sensorimotor signals to exceed threshold, thus allowing weak sensorimotor signals to become detectable.[69] Sports Med 2009; 39 (3)

Unilateral or bilateral instability

Presence of weakness

Presence of pain

Presence of MI

Function of the ankle

Presence of ‘giving way’

Study

No. of sprains

Tropp et al.[14]

Recurrent sprain, number not specified

Olmsted et al.[58]

At least one episode of lateral ankle sprain

Gribble et al.[59]

At least one episode of lateral ankle sprain

Ross et al.[60]

Three ankle sprains

Physically active for at least 3 h/wk

At least two ‘giving way’ episodes

Brown and Mynark[61]

Minimum of two ankle sprains. Average of 7–10 sprains

Recreationally active. At least 20 min 3 ·/wk

Minimum of two episodes of ‘giving way’

Unilateral General athletic CAI population from an NCAA Division III university

Pathological anterior drawer sign

‘Giving way’ present

Could be present, but not measured

Recurrent episodes of ‘giving way’ present

Multiple episodes of ankle ‘giving way’

Unilateral

Weaker than stable ankle

More painful than stable ankle

No MI (presence of MI was considered an exclusion criteria)

Less functional than the noninjured ankle

CAI = chronic ankle instability; FAIQ = Functional Ankle Instability Questionnaire; MI = mechanical instability; NCAA = National Collegiate Athletic Association.

‘Giving way’ present

Holmes & Delahunt

Sports Med 2009; 39 (3)

Wikstrom et al.[62] FAIQ. History of an inversion injury requiring protected weightbearing and/or immobilization

Activity level

216

ª 2009 Adis Data Information BV. All rights reserved.

Table V. Inclusion criteria used to define functional instability (FI) by all the studies included in the current review examining postural control deficits in subjects with FI (a blank cell indicates that this information was not specified)

Treatment of Common Ankle Instability Deficits

217

Table VI. Interventions for postural control deficits Study

Design

Definition of FI and intervention applied

Outcome measures

Results

Akhbari et al.[51]

Pre-/post-intervention n = 15

FI = repetitive sprains or repetitive episodes of ‘giving way’ 4-wk BSS, 12 min per session 3 ·/wk

Balance

Significant › in balance measures

Rozzi et al.[63]

Controlled group pre-/ post-intervention n = 26 (exercise = 13; control = 13)

FI = repeated episodes of ‘ankle rolling’ and/or the ankle ‘giving way’ Static and dynamic balance training using BSS 3 ·/wk for 4 wk

AJFAT Single leg static balance assessment at level 2 and 6 on the BSS using SI

› in AJFAT scores. Post-SI scores of both groups were significantly ﬂ vs their respective pre-scores reflecting an improvement in balance ability

Eils and Rosenbaum[36]

Controlled group pre-/ post-intervention n = 48 (exercise = 31; control = 17)

FI = repeated ankle inversion sprains and a self-reported subjective feeling of instability or ‘giving way’. 6-wk progressive physiotherapeutic exercise programme consisting of 12 neuromuscular exercises 1 ·/wk

Postural sway

In the exercise group, results showed a significant ﬂ in postural sway. 90% of the exercise group returned the questionnaire 1 y after training. Evaluation showed an almost 60% ﬂ frequency of ankle inversion sprains

Michell et al.[64]

Pre-/post-intervention n = 32 Group 1: exercise sandal group (8 stable ankle subjects, 8 unstable ankle subjects) Group 2: shoe group (8 stable ankle subjects, 8 unstable ankle subjects)

FI = at least two sprains, and sensations of ‘giving way’ Group 1: functional balance training programme with exercise sandals 3 ·/wk for 8 wk Group 2: functional balance training programme with exercise shoes 3 ·/wk for 8 wk

Postural sway

Postural stability improved in both groups

Kidgell et al.[65]

Controlled group pre-/ post-intervention n = 20 (group 1 = 7; group 2 = 6; control = 7)

FI = lateral ankle sprain within past 2 y Group 1: progressive balance training on mini-trampoline 3 ·/wk for 6 wk Group 2: progressive balance training on dura disc 3 ·/wk for 6 wk

Postural sway

After 6 wk there was a significant improvement in postural sway in subjects in group 1 and group 2 with no significant differences in results between groups

Ross and Guskiewicz[66]

Controlled group pre-/ post-intervention n = 60 (group 1 = 20: 10 FI and 10 stable; group 2 = 20: 10 FI and 10 stable; group 3 = 20: 10 FI and 10 stable)

FI = at least two sprains, and sensations of ‘giving way’ Group 1: coordination training 5 ·/wk for 6 wk Group 2: coordination training and stochastic resonance 5 ·/wk for 6 wk Group 3: no training

TTS

Subjects with FI improved their A/P and M/L TTS scores. Stochastic resonance did not influence the results

Hale et al.[67]

Controlled group pre-/ post-intervention n = 48 (FI exercise = 16; FI control = 13; healthy control = 19)

FI = unilateral sprain with chronic residual symptoms – subjective feeling of ‘giving way’ in past 6 mo Exercise group underwent 4-wk rehabilitation programme including ROM, strength, neuromuscular control and functional tasks

Postural sway SEBT FADI

Subjects with FI had ﬂ postural control and SEBT ability pre-training. Post-training the FI exercise group showed › SEBT and FADI scores

AJFAT = ankle joint functional assessment tool; A/P = anterior/posterior; BSS = Biodex Stability System; FADI = Foot and ankle disability index; M/L = medial/lateral; FI = functional instability; post = post-test; pre = pre-test; ROM = range of motion; SEBT = star excursion balance test; SI = stability index; TTS = time-to-stabilization; ﬂ indicates decrease(d); › indicates increase(d).

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (3)

Holmes & Delahunt

218

The exact neurological mechanism underlying the improvement of postural stability following the application of SR is still not understood, but one proposed mechanism is that it can bring muscle spindles and Golgi tendon organs closer to the threshold of stimulation.[69] Thus Ross[68] suggested that SR may enhance signal detection of weak sensorimotor signals associated with postural control in subjects with FI. However, this hypothesis requires further validation. Hale et al.[67] examined the effects of a 4-week comprehensive rehabilitation programme that consisted of range of motion, strength, neuromuscular control and functional tasks on postural control and function in subjects with CAI. Although the programme had positive effects on postural control and functional limitations, the exact mechanism behind these improvements could not be identified. However, the study does lend support to the need for a comprehensive rehabilitation approach in subjects with CAI. Based on the literature presented above, there is evidence to support the use of balance training programmes in clinical practice. However, one limitation of the majority of the studies reviewed was the lack of subject follow-up over an extended period of time, as perhaps the most important clinical variable is whether these programmes actually produce a reduction in future ankle sprain incidence. 7. Strength Deficits Strength training has typically been a major element in rehabilitation programmes following ankle sprains and is most often initiated once pain-free range of motion is regained.[70] The ankle joint musculature plays an integral role in dynamic stabilization of the ankle joint, which is achieved by co-contraction of the musculature surrounding the joint.[71] During high-load dynamic activities such as running, cutting and jump landing, muscular co-contraction is required to minimize resulting ground reaction forces.[72] Thus, an imbalance in this muscular co-contraction as a result of strength deficits may leave the ankle joint susceptible to injury due ª 2009 Adis Data Information BV. All rights reserved.

to less efficient dissipation of ground reaction forces.[71] Consequently abnormal forces could be transmitted to the ankle joint ligamentous and articular structures causing injury, a point that has recently been emphasized by Caulfield and Garrett[73] and Delahunt et al.[49] A potential cause of FI historically cited in the literature involves weakness of the peroneal musculature.[74] The main mechanism of incurring a lateral ankle ligament complex sprain involves a forced plantar flexion and inversion injury. Thus it has been suggested that deficits in evertor strength would reduce the ability of the peroneal muscles to resist the inversion moment and return the foot to a neutral position, thereby preventing an inversion injury.[75] It has also been suggested that deficits in inversion strength could play a role in the development of FI and repeated episodes of ankle joint inversion injury.[19] During closed kinetic chain activities, Wilkerson and Nitz[19] have suggested that the ankle joint invertors play an integral role in controlling the rate of calcaneal eversion as the body’s centre of mass is displaced laterally beyond the base of support. If the calcaneus reaches its limit of eversion the medial border of the forefoot rises from the supporting surface creating an inversion torque about the forefoot and rearfoot, thus potentially producing a hyper-inversion injury and subsequent damage to the lateral ligament complex. Because of the biomechanical importance of the ankle joint musculature, many studies have examined the relationship between strength deficits and FI. Previous authors have suggested that subjects with FI have eversion strength deficits;[76] however, more recent and methodologically rigorous research using isokinetic dynamometry has failed to support the theory that evertor weakness is a common deficit in subjects with FI.[15,17,77-79] The consensus in the literature is that evertor strength deficits are not a common finding in subjects with FI. However, despite this consensus, Caulfield[70] concluded that strength training of the peronei forms a central component of rehabilitation programmes for FI, illustrating that despite the lack of evidence, evertor weakness is considered a significant factor in FI amongst clinicians. Sports Med 2009; 39 (3)

Treatment of Common Ankle Instability Deficits

The exact reason for this is unclear, but perhaps there is a lack of knowledge among some clinicians regarding the exact role of neuromuscular deficits in the aetiology of FI. Another interesting observation is that Caulfield[70] does not present any data regarding the composition of rehabilitation programmes for FI and perhaps the observation alluded to is pure conjecture. Other authors have suggested that perhaps it is a lack of endurance in the evertors that contributes to FI[48,80] and that the timing of the evertor muscle activity may be of greater importance in FI.[24,81] A number of studies have shown that subjects with FI exhibit strength deficits in their invertor musculature.[17,75,81] The presence of invertor strength deficits could pose a problem for subjects with FI, especially if one considers the biomechanical and neuromuscular potential mechanism of injury proposed by Wilkerson and Nitz.[19] Ryan[17] has suggested that the presence of invertor strength deficits occurs as a result of selective inhibition. The development of selective inhibition may be the result of arthrogenic muscle inhibition, which has been described by Hopkins and Ingersoll[82] as an ongoing reflex reaction of musculature surrounding a joint after distension or damage to the structures of that joint, and is thought to impede rehabilitation following joint injury by preventing appropriate activation of the surrounding muscles. The process of selective inhibition has been postulated to occur in order to protect injured structures by inhibiting muscles that are capable of increasing tensile stress on damaged ligamentous and articular structures.[83] Thus, in accordance with the process of selective inhibition, the ankle joint invertors may be reflexively inhibited due to their ability to produce ankle joint inversion and subsequent stress on the lateral ankle ligament complex. The inclusion criteria used by those studies investigating strength deficits in subjects with FI are outlined in table VII. 8. Interventions for Strength Deficits The peronei are integral to the control of supination of the rearfoot and protection against ª 2009 Adis Data Information BV. All rights reserved.

219

lateral ankle sprains.[70] Despite the volume of literature refuting the role of eversion strength deficits as a contributing factor to the development of FI, a number of studies have been performed to examine the effectiveness of strengthening programmes in subjects with FI.[37,38,84] The results of these studies are discussed below and outlined in table VIII. Docherty et al.[37] examined the effects of strength training on ankle joint strength and JPS in subjects with FI. Ankle joint strength was measured using a hand-held dynamometer while subjects performed maximal isometric muscle contractions against manual resistance. The exercise protocol consisted of progressive resistive exercises using resistive tubing. The study found that ankle strengthening exercises significantly improve ankle joint evertor (pre-test 30.9 – 6.5 N, post-test 45.0 – 4.9 N) and dorsiflexor (pre-test 33.3 – 4.8 N, post-test 50.6 – 6.3 N) strength in subjects with FI. The control group, which consisted of subjects with FI who did not carry out the exercise protocol, showed no statistically significant change in either evertor (pre-test 30.8 – 6.0 N, post-test 27.7 – 11.6 N) or dorsiflexor (pre-test 33.8 – 7.2 N, post-test 33.9 – 5.0 N) strength. This study does have a number of limitations that are worth considering. First, strength testing was examined using a hand-held dynamometer utilizing maximal isometric muscle contractions. The authors do not present any data pertaining to the reliability of this method of strength testing. Furthermore, there was no longterm follow-up of subjects to determine whether the changes in strength noted produced any meaningful functional consequence, in terms of a reduction in the incidence of episodes of ankle sprain or sensations of ‘giving way’. Kaminski et al.[84] examined the effects of strength and proprioception training on eversion (E) to inversion (I) ratios in subjects with unilateral FI. Thirty-eight subjects were randomly assigned to one of four treatment groups: strength training, proprioception training, strength and proprioception training, or control/no training. The results of the study found no significant differences in average torque and peak torque E/I ratios post-training, with the authors concluding Sports Med 2009; 39 (3)

Presence of weakness

Presence of pain

An inversion injury

Unilateral

Chronically weaker than opposite ankle

More painful than opposite ankle

Ryan[17]

At least three inversion sprains

Unilateral

At least six episodes of ‘giving way’ into inversion

Tropp[76]

Recurrent sprains. No. not specified

Unilateral

Episodes of ‘giving way’

Munn et al.[75]

>1 inversion sprain

Unilateral

Kaminski et al.[77]

At least one inversion injury

Unilateral

Bernier et al.[78]

Ankle inversion injuries. No. not specified

Unilateral

McKnight and Armstrong[79]

Minimum of two ankle sprains. Average of 7.5 sprains.

Wilkerson et al.[81]

Multiple ankle sprains

No. of sprains

Lentell et al.[15]

MI = mechanical instability.

Activity level

Weight-bearing activity 3 ·/wk

Presence of weakness

Presence of MI

Function of the Presence ankle of ‘giving way’ Less functional than opposite ankle

Presence of pain

Reduced function

No MI (presence of MI was considered an exclusion criteria)

Episodes of ‘giving way’

Repeated episodes of ‘giving way’

Repeated episodes of ‘giving way’

Unilateral

Frequent episodes of ‘giving way’

Holmes & Delahunt

Sports Med 2009; 39 (3)

Unilateral or bilateral instability

Study

220

ª 2009 Adis Data Information BV. All rights reserved.

Table VII. Inclusion criteria used to define functional instability (FI) by all the studies included in the current review examining strength deficits in subjects with FI (a blank cell indicates that this information was not specified)

Treatment of Common Ankle Instability Deficits

221

Table VIII. Interventions for strength deficits Study

Design

Definition of FI and intervention applied

Outcome measures

Results

Docherty et al.[37]

Controlled group pre-/postintervention n = 20 (exercise = 10; control = 10)

FI = >3 sprains in the last 5 y. One episode of ‘giving way’ in the last 12 mo

Strength testing

› in dorsiflexor and evertor strength

Kaminski et al.[84]

Controlled group pre-/postintervention n = 38 Randomly assigned to one of four groups

FI defined according to FAIQ Group 1: progressive Theraband strengthening programme ·6 wk Group 2: progressive proprioception training using T-band kicks ·6 wk Group 3: combination of both programmes ·6 wk Group 4: control

Eversion to inversion strength ratios using isokinetic testing

No significant results were found post

Sekir et al.[38]

Pre-/post-intervention n = 24

FI = two moderate sprains, repeated episodes of ‘giving way’ 6-wk isokinetic concentric inversion and eversion protocol 3 ·/wk

Strength testing

› in concentric invertor strength

FAIQ = functional ankle instability questionnaire; FI = functional instability; post = post-test; pre = pre-test; › indicates increase.

that 6 weeks of strength and proprioceptive training (either alone or combined) had no effect on isokinetic measures of strength in subjects with self-reported unilateral FI. However, weaknesses were identified within the study such as a lack of consistency in mode specificity between training techniques and outcome measures used. The authors also concluded that the exercise regimen used may not have been rigorous enough to produce changes in stronger subjects. Further research is needed to offer guidelines on how strenuous the training regimens must be to bring about changes in strength. The results of this study cannot be directly compared with the study by Docherty et al.[37] because of the different methods of strength testing used in each study. Again a common limitation of this study similar to that of Docherty et al.[37] was the lack of subject followup to determine whether the training protocol produced any change in functional status as measured by incidence of recurrent ankle sprains. Sekir et al.[38] studied the effect of isokinetic training on strength, functionality and proprioception in athletes with FI. Twenty-four subjects with unilateral FI performed a concentric inversion and eversion strengthening protocol on an isokinetic dynamometer three times per week for 6 weeks. Outcome measures were concentric and eccentric invertor and evertor peak torques, passive JPS and a range of more functional ª 2009 Adis Data Information BV. All rights reserved.

measures including a single-leg hopping course and various hopping drills. Following the exercise protocol, subjects exhibited significant improvements in concentric invertor torque, which had been shown to be decreased compared with the non-injured ankle joint prior to the strengthening protocol and functional performance. However, correlation analyses showed no relationship between the improvements of these parameters, indicating that there is not necessarily a cause-effect relationship between these variables, a factor not investigated by Docherty et al.[37] 9. Conclusion Subjects with CAI appear to exhibit deficits in ankle joint frontal plane positional sense. The exact consequence of this has yet to be fully elucidated. The present literature suggests that the feed-back neuromuscular response time of the peroneal muscles is not affected in the presence of CAI. Of more importance may be feedforward neuromuscular control mechanisms. Studies are now required to examine the effects of dynamic neuromuscular training, incorporating the principles of feed-forward neuromuscular control on ankle joint injury risk factors. Balance training protocols improve postural control in subjects with CAI. However, the most Sports Med 2009; 39 (3)

Holmes & Delahunt

222

efficient protocol has yet to be developed. Evertor strength deficits do not appear to contribute to the development of CAI, so the use of evertor strengthening protocols in clinical practice requires careful consideration. Strengthening protocols addressing invertor muscle deficits may be more efficacious as subjects with CAI exhibit invertor strength deficits. It is the responsibility of the clinician to complete a thorough evaluation and apply an exercise programme that addresses those specific deficits that are present. This will ensure better functional outcomes following ankle sprains and help prevent CAI as well as improving physiotherapeutic and athletic training methods in general.

Acknowledgements There are no conflicts of interest and no funding was received.

References 1. Fong DT, Hong Y, Chan LK, et al. A systematic review on ankle injury and ankle sprain in sports. Sports Med 2007; 37: 73-94 2. Garrick JG. The frequency of injury, mechanism of injury, and epidemiology of ankle sprains. Am J Sports Med 1977; 5: 241-2 3. Diamond JE. Rehabilitation of ankle sprains. Clin Sports Med 1989; 8: 877-91 4. Yeung M, Chan K, So C, et al. An epidemiological survey on ankle sprain. Br J Sports Med 1994; 28: 112-6 5. Smith RW, Reischl SF. Treatment of ankle sprains in young athletes. Am J Sports Med 1986; 14: 465-71 6. Gerber JP, Williams GN, Scoville CR, et al. Persistent disability associated with ankle sprains: a prospective examination of the athletic population. Foot Ankle Int 1998; 19: 654-60 7. Braun BL. Effects of an ankle sprain in a general clinical population 6 to 18 months after medical evaluation. Arch Fam Med 1999; 8: 143-8 8. Hertel J. Functional anatomy, pathomechanics and pathophysiology of lateral ankle instability. J Athl Train 2002; 37: 364-75 9. Monaghan K, Delahunt E, Caulfield C. Ankle function during gait in patients with chronic ankle instability compared to controls. Clin Biomech 2005; 21: 168-74 10. Freeman M. Instability of the foot after injuries to the lateral ligament of the ankle. J Bone Joint Surg 1965; 47B: 669-77 11. Demeritt KM, Schultz SJ, Docherty CL, et al. Chronic ankle instability does not affect lower extremity functional performance. J Athl Train 2002; 37: 507-11

ª 2009 Adis Data Information BV. All rights reserved.

12. Hertel J. Functional instability following lateral ankle sprain. Sports Med 2000; 29: 361-71 13. Tropp H. Functional ankle instability revisited. J Athl Train 2002; 37: 512-5 14. Tropp H, Odenrick P, Gillquist J. Stabilometry recordings in functional and mechanical instability of the ankle joint. Int J Sports Med 1985; 6: 180-2 15. Lentell G, Katzman L, Walters M. The relationship between muscle function and ankle stability. J Orthop Sports Phys Ther 1990; 11: 605-11 16. Birmingham TB, Chesworth BM, Hartsell HD, et al. Peak passive resistive torque at maximum inversion range of motion in subjects with recurrent ankle inversion sprains. J Orthop Sports Phys Ther 1997; 25: 342-8 17. Ryan L. Mechanical stability, muscle strength, and prorioception in the functionally unstable ankle. Aust J Physiother 1994; 40: 41-7 18. Denegar CR, Miller SJ. Can chronic ankle instability be prevented? Rethinking management of lateral ankle sprains. J Athl Train 2002; 37: 430-5 19. Wilkerson GB, Nitz AJ. Dynamic ankle stability: mechanical and neuromuscular interrelationships. J Sports Rehabil 1994; 3: 43-57 20. Rowinski MJ. Orthopaedic and sports physical therapy. St Louis (MO): CV Mosby, 1990 21. Lephart SM, Pincivero DM, Rozzi SL. Proprioception of the ankle and knee. Sports Med 1998; 25: 149-55 22. Willems T, Witvrouw E, Verstuyft J, et al. Proprioception and muscle strength in subjects with a history of ankle sprains and chronic instability. J Athl Train 2002; 37: 487-93 23. Konradsen L, Voigt M. Inversion injury biomechanics in functional instability: a cadaver study of simulated gait. Scand J Med Sci Sports 2002; 12: 329-36 24. Delahunt E, Monaghan K, Caulfield B. Altered neuromuscular control and ankle joint kinematics during walking in subjects with functional instability of the ankle joint. Am J Sports Med 2006; 34: 1970-6 25. Garn SN, Newton RA. Kinesthetic awareness in subjects with multiple ankle sprains. Phys Ther 1988; 11: 1667-71 26. Lentell G, Baas B, Lopez D, et al. The contributions of proprioceptive deficits, muscle function, and anatomic laxity to functional instability of the ankle. J Orthop Sports Phys Ther 1995; 21: 206-15 27. Mulloy Forkin D, Koczur D, Battle R, et al. Evaluation of kinaesthetic deficits indicative of balance control in gymnasts with unilateral chronic ankle sprains. J Orthop Sports Phys Ther 1996; 23: 245-50 28. Refshauge KM, Kilbreath SL, Raymond J. The effect of recurrent ankle inversion sprain and taping on proprioception at the ankle. Med Sci Sports Exerc 2000; 32: 10-15 29. Hubbard TJ, Kaminski TW. Kinesthesia is not affected by functional ankle instability status. J Athl Train 2002; 37: 481-6 30. Jerosch J, Bischof M. Proprioceptive capabilities of the ankle in stable and unstable joints. Sports Exerc Injury 1996; 2: 167-71 31. Boyle J, Negus V. Joint position sense in the recurrently sprained ankle. Aust J Physiother 1998; 44: 159-63

Sports Med 2009; 39 (3)

Treatment of Common Ankle Instability Deficits

32. Konradsen L, Magnusson P. Increased inversion angle replication error in functional ankle instability. Knee Surg Sports Traumatol Arthrosc 2000; 8: 246-51 33. Konradsen L, Ravn JB, Sørensen AI. Proprioception at the ankle: the effect of anaesthetic blockade of ligament receptors. J Bone Joint Surg 1993; 75B: 433-6 34. Refshauge KM, Fitzpatrick RC. Perception of movement at the human ankle: effects of leg position. J Physiol 1995; 488: 243-8 35. Bernier JN, Perrin DH. Effect of coordination training on proprioception of the functionally unstable ankle. J Orthop Sports Phys Ther 1998; 27: 264-75 36. Eils E, Rosenbaum D. A multi-station proprioceptive exercise program in patients with ankle instability. Med Sci Sports Exerc 2001; 33: 1991-8 37. Docherty CL, Moore JH, Arnold BL. Effects of strength training on strength development and joint position sense in functionally unstable ankles. J Athl Train 1998; 33: 310-14 38. Sekir U, Yildiz Y, Hazneci B, et al. Effect of isokinetic training on strength, functionality and proprioception in athletes with functional ankle instability. Knee Surg Sports Traumatol Arthrosc 2007; 15: 654-64 39. Riemann BL, Lephart SM. The sensorimotor system, part I: the physiological basis of functional joint stability. J Athl Train 2002; 37: 71-9 40. Konradsen L, Ravn JB. Ankle instability caused by prolonged peroneal reaction time. Acta Orthop Scand 1990; 61: 199-204 41. Karlsson J, Andreasson GO. The effect of external ankle support in chronic lateral ankle joint instability. Am J Sports Med 1992; 20: 257-61 42. Lo¨fvenberg R, Ka¨rrholm J, Sundelin G, et al. Prolonged reaction time in patients with chronic lateral instability of the ankle. Am J Sports Med 1995; 23: 414-7 43. Ebig M, Lephart SM, Burdett RG, et al. The effect of sudden inversion stress on EMG activity of the peroneal and tibialis anterior muscles in chronically unstable ankles. J Orthop Sports Phys Ther 1997; 26: 73-7 44. Fernandes N, Allison GT, Hopper D. Peroneal latency in normal and injured ankles at varying angles of perturbation. Clin Orthop Relat Res 2000; 375: 193-201 45. Vaes P, van Gheluwe B, Duquet W. Control of acceleration during sudden ankle supination in people with unstable ankles. J Orthop Sports Phys Ther 2001; 31: 741-52 46. Vaes P, Duquet W, van Gheluwe B. Peroneal reaction times and eversion motor response in healthy and unstable ankles. J Athl Train 2002; 37: 475-80 47. Delahunt E. Peroneal reflex contribution to the development of functional instability of the ankle joint. Phys Ther Sport 2007; 8: 98-104 48. Caulfield B, Crammond T, O’Sullivan A, et al. Altered ankle-muscle activation during jump landings in participants with functional instability of the ankle joint. J Sport Rehabil 2004; 13: 189-200 49. Delahunt E, Monaghan K, Caulfield B. Changes in lower limb kinematics, kinetics, and muscle activity in subjects with functional instability of the ankle joint during a single leg drop jump. J Orthop Res 2006; 24: 1991-2000

ª 2009 Adis Data Information BV. All rights reserved.

223

50. Clark VM, Burden AM. A 4-week wobble board exercise program improved muscle onset latency and perceived stability in individuals with a functionally instable ankle. Phys Ther Sport 2005; 6: 181-7 51. Akhbari B, Takamjani IE, Salavati M, et al. A 4-week biodex stability exercise program improved ankle musculature onset, peak latency and balance measures in functionally unstable ankles. Phys Ther Sport 2007; 8: 117-29 52. Grabiner MD. Bioelectric characteristics of the electrochemical delay proceeding concentric contraction. Med Sci Sports Exerc 1986; 18: 37-43 53. Mattacola CG, Lloyd JW. Effects of a 6 week strength and proprioception training program on measures of dynamic balance: a single case design. J Athl Train 1997; 32: 127-35 54. Holme E, Magnusson SP, Becher K, et al. The effect of supervised rehabilitation on strength, postural sway, position sense and reinjury risk after acute ligament sprain. Scand J Med Sci Sports 1999; 9: 104-9 55. Pintsaar A, Brynhildsen J, Tropp H. Postural corrections after standardised perturbations of single leg stance: effect of training and orthotic devices in patients with ankle instability. Br J Sports Med 1996; 30: 151-5 56. Gauffin H. Knee and ankle kinesiology and joint instability [dissertation]. Linkoping: Linkoping University, 1991 57. Hertel J, Miller S, Denegar C. Intratest an intertest reliability during the star excursion balance test. J Sports Rehab 2000; 9: 104-16 58. Olmsted LC, Carcia CR, Hertel J, et al. Efficacy of the star excursion balance test in detecting reach deficits in subjects with chronic ankle instability. J Athl Train 2002; 37: 501-6 59. Gribble PA, Hertel J, Denegar CE, et al. The effects of fatigue and chronic instability on dynamic postural control. J Athl Train 2004; 39: 321-9 60. Ross SE, Guskiewicz KM, Yu B. Single-leg jump-landing stabilization times in subjects with functionally unstable ankles. J Athl Train 2005; 40: 298-304 61. Brown CN, Mynark R. Balance deficits in recreational athletes with chronic ankle instability. J Athl Train 2007; 42: 367-73 62. Wikstrom EA, Tillman MD, Chmielewski TL, et al. Dynamic postural stability deficits in subjects with selfreported ankle instability. Med Sci Sports Exerc 2007; 39: 397-402 63. Rozzi SL, Lephart SM, Sterner R, et al. Balance training for persons with functionally unstable ankles. J Ortho Sports Phys Ther 1999; 29: 478-86 64. Michell TB, Ross SE, Blackburn JT, et al. Functional balance training, with or without exercise sandals, for subjects with stable or unstable ankles. J Athl Train 2006; 41: 393-8 65. Kidgell DJ, Horvath DM, Jackson BM. Effect of six weeks of dura disc and mini trampoline balance training on postural sway in athletes with functional ankle instability. J Strength Cond Res 2007; 21: 466-9 66. Ross SE, Guskiewicz KM. Effect of co-ordination training with and without stochastic resonance stimulation on dynamic postural stability of subjects with functional ankle instability and subjects with stable ankles. Clin J Sports Med 2006; 16: 323-8 67. Hale SA, Hertel J, Olmsted-Kramer LC. The effect of a four week comprehensive rehabilitation program on postural

Sports Med 2009; 39 (3)

Holmes & Delahunt

224

68.

69. 70. 71.

72. 73.

74.

75.

76. 77.

control and lower extremity function in individuals with chronic ankle instability. J Ortho Sports Phys Ther 2007; 37: 303-11 Ross SE. Noise enhanced postural stability in subjects with functional ankle instability. Br J Sports Med 2007; 41: 656-9 Cordo P, Inglis J, Verschueren S, et al. Noise in human spindles. Nature 1996; 383: 769-70 Caulfield B. Functional instability of the ankle joint: features and underlying causes. Physiotherapy 2000; 86: 401-11 Kaminski TW, Hartsell HD. Factors contributing to chronic ankle instability: a strength perspective. J Athl Train 2002; 37: 394-405 Dvir Z. Isokinetics: muscle testing interpretation and clinical application. London: Churchill Livingstone, 1995: 1-22 Caulfield BM, Garret M. Changes in ground reaction force during jump landing in subjects with functional instability of the ankle joint. Clin Biomech 2002; 19: 617-21 Boisen WR, Staples OS, Russell SW. Residual disability following acute ankle sprains. J Bone Joint Surg Am 1955; 37: 1237-43 Munn J, Beard D, Refshauge KM, et al. Eccentric muscle strength in functional ankle instability. Med Sci Sports Exerc 2003; 35: 245-50 Tropp H. Pronator muscle weakness in functional instability of the ankle joint. Int J Sports Med 1986; 7: 291-4 Kaminski TW, Perrin DH, Gansneder BM. Eversion strength analysis of uninjured and functionally unstable ankles. J Athl Train 1999; 34: 239-45

ª 2009 Adis Data Information BV. All rights reserved.

78. Bernier JN, Perrin DH, Rijke AM. Effect of unilateral functional instability of the ankle on postural sway and inversion and eversion strength. J Athl Train 1997; 32: 226-32 79. McKnight CM, Armstrong CW. The role of ankle strength in functional ankle instability. J Sports Rehab 1997; 6: 21-9 80. Delahunt E. Neuromuscular contributions to functional instability of the ankle joint. J Bodywork Move Ther 2007; 11: 204-13 81. Wilkerson G, Pinerola J, Caturano R. Invertor versus evertor torque and power deficiencies associated with lateral ankle ligament injury. J Orthop Sports Phys Ther 1997; 26: 78-86 82. Hopkins J, Ingersoll C. Arthrogenic muscle inhibition: a limiting factor in joint rehabilitation. J Sport Rehabil 2000; 9: 135-59 83. Swearingen RL, Dehne E. A study of pathological muscle function following injury to a joint. J Bone Joint Surg 1964: 46A: 1364 84. Kaminski TW, Buckley BD, Powers ME, et al. Effect of strength and proprioception training on eversion to inversion strength ratios in subjects with unilateral functional ankle instability. Br J Sports Med 2003; 37: 410-5

Correspondence: Dr Eamonn Delahunt, School of Physiotherapy and Performance Science, University College Dublin, Health Sciences Centre, Belfield, Dublin 04, Ireland. E-mail: [email protected]

Sports Med 2009; 39 (3)

Sports Med 2009; 39 (3): 225-234 0112-1642/09/0003-0225/$49.95/0

REVIEW ARTICLE

ª 2009 Adis Data Information BV. All rights reserved.

The Anatomy of the Pubic Region Revisited Implications for the Pathogenesis and Clinical Management of Chronic Groin Pain in Athletes Brett A. Robertson,1 Priscilla J. Barker,2 Marius Fahrer2 and Anthony G. Schache3 1 School of Physiotherapy, The University of Melbourne, Melbourne, Victoria, Australia 2 Department of Anatomy, The University of Melbourne, Melbourne, Victoria, Australia 3 Department of Mechanical Engineering, Melbourne School of Engineering, The University of Melbourne, Melbourne, Victoria, Australia

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Aetiology of Chronic Groin Pain. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Epidemiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Musculoskeletal Risk Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Scope of the Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Composition of Adductor Longus at its Pubic Attachment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Arrangement of the Lower Fibres of Internal Oblique and Transversus Abdominis . . . . . . . . . . . . 4.4 Confluence of Soft Tissue Structures Anterior to the Pubic Symphysis . . . . . . . . . . . . . . . . . . . . . . . 4.5 Review Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Implications for Pathogenesis and Clinical Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

225 226 227 227 227 227 228 229 229 230 231 232

Chronic groin pain is a common complaint for athletes participating in sports that involve repetitive sprinting, kicking or twisting movements, such as Australian Rules football, soccer and ice hockey. It is frequently a multifactorial condition that presents a considerable challenge for the treating sports medicine practitioner. To better understand the pathogenesis of chronic groin pain in athletes, a precise anatomical knowledge of the pubic symphysis and surrounding soft tissues is required. Several alternative descriptions of pubic region structures have been proposed. Traditionally, chronic groin pain in athletes has been described in terms of discrete pathology requiring specific intervention. While this clinical reasoning may apply in some cases, a review of anatomical findings indicates the possibility of multiple pathologies coexisting in athletes with chronic groin pain. An appreciation of these alternative descriptions may assist sports medicine practitioners with diagnostic and clinical decision-making processes. The purpose of this literature review is to reappraise the anatomy of the pubic

Robertson et al.

226

region, considering findings from cadaveric dissection and histology studies, as well as those from diagnostic imaging studies in athletes.

Chronic groin pain is a common complaint for athletes participating in sports that involve repetitive sprinting, kicking or twisting movements, such as Australian Rules football, soccer and ice hockey.[1-3] It is typically a multifactorial condition that presents a considerable challenge for the treating sports medicine practitioner.[4] Symptoms are often vague and diffuse, extend between the lower abdomen and medial thigh, and may be attributed to a variety of diagnostic entities.[5-9] Additionally, it is possible that many of these diagnostic entities share a common origin. Recent research[10,11] has challenged several details in the anatomy of the pubic region as described in modern texts. This new information may have important clinical implications. If the textbook descriptions are inaccurate, then models of chronic groin pain pathogenesis based on these descriptions may be inherently flawed. The purpose of this literature review is to reappraise the anatomy of the pubic region. The outcomes will provide a background for further investigation, contribute to the understanding of chronic groin

pain pathogenesis, and provide a context for evaluating current management strategies. 1. Aetiology of Chronic Groin Pain Numerous conditions are reported in the literature as possible causes of acute or chronic groin pain in athletes.[4,12-20] A composite table of these conditions reveals a differential diagnosis algorithm that is impractical for the average clinician (table I). The potential for a medical cause of groin pain in athletes should be recognized and appreciated by clinicians.[5,9,14,15,18,21,22] However, the majority of conditions resulting in chronic groin pain in athletes are indicated as being of musculoskeletal origin.[4,12,14,23,24] The specific source/diagnosis premise may be obscured by vague symptoms or an insidious onset indicating the possibility of sinister pathology. That said, if the presenting symptoms are aggravated by activity and relieved by rest, then symptom behaviour is suggestive of a musculoskeletal disorder. The balance of this review

Table I. Possible causes of groin pain in athletes reported in the literature[4,12-20] Abdominal aortic aneurysm

Hydrocoele/varicocoele

Acetabular disorders

Inflammatory bowel disease

Postpartum symphysis separation Prostatitis

Adductor strain

Inguinal or femoral hernia

Pubic instability

Adductor tendinitis

Intra-abdominal abscess

Sacroiliac joint problems

Apophysitis

Legg-Calve´-Perthes disease

Seronegative spondyloarthropathy

Appendicitis

Lumbar spine pathology

Slipped capital femoral epiphysis

Avascular necrosis of femoral head

Lymphadenopathy

Snapping hip syndrome

Avulsion fracture

Muscle strain

Sports hernia

Bursitis

Myositis ossificans

Stress fractures

Conjoined tendon dehiscence

Nerve entrapment

Synovitis

Crohn’s disease

Obturator nerve entrapment

Testicular neoplasm

Diverticular disease

Osteitis pubis

Testicular torsion

Epididymitis

Osteoarthritis

Urethritis

Femoroacetabular impingement

Ovarian cyst

Urinary tract infection

Herniated nucleus pulposus

Pelvic inflammatory disease

Hockey player’s syndrome

Pelvic stress fracture

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (3)

The Anatomy of the Pubic Region

therefore focuses on chronic sports-related or athletic groin pain that originates from musculoskeletal structures in the pubic region. Specifically, this review addresses the anatomy and relations of the adductor and lower abdominal musculature, the pubic bone, and the pubic symphysis and its capsular tissues. 2. Epidemiology The incidence of injury to the groin region represents 5–18% of reported athletic injuries.[18,25-28] However, caution is required when interpreting these data, as there are currently no universal standards for the definition or classification of these conditions.[3,4,29] The use of generic classifications such as ‘groin injury’ or ‘groin strain’ may identify which athletes have groin symptoms, but the true value of the data is questionable. For example, acute traumatic injuries would not be differentiated from insidious overuse injuries, even though they may have a distinctly different pathogenesis. Furthermore, if an injury is recorded only when it results in a missed training session or match, then players who continue to train and play with groin symptoms would not be included in the data.[20] Additionally, epidemiological studies do not appear to adequately identify which musculoskeletal conditions are most commonly involved in groin injuries. This deficiency could possibly be addressed by adopting the clinical entities approach[7] to classify athletic groin pain in future studies. Within this approach, an injury may be classified as comprising one or more of adductorrelated, abdominal wall-related, pubic bonerelated or psoas-related clinical entities.[7,30] 3. Musculoskeletal Risk Factors The musculoskeletal risk factors for groin injury provide a context for evaluating the relevance of disparities identified in the anatomical literature. There are relatively few published studies with prospectively collected data that investigate the risk factors for groin injury in athletes. Similar to the epidemiological studies, there is no consistency in the definition of the groin ª 2009 Adis Data Information BV. All rights reserved.

227

injuries investigated or in the reporting of their degree of chronicity. Furthermore, these studies involve a limited range of sports, namely soccer,[31-33] ice hockey[34,35] and Australian Rules football.[36] When interpreting the findings of these studies, it is important to consider the task demands of the sport being investigated. For instance, the excursion of movement or the magnitude and velocity of loading may influence the structures at risk. Reduced hip abduction[31] and total hip rotation[33] range of movement (ROM) have been reported as risk factors for groin strain in soccer players. Reduced total hip rotation ROM was also found to be a risk factor for pubic bone stress injury in Australian Rules footballers.[36] Conversely, reduced hip abduction ROM was reported as having no association with adductor muscle injury in both soccer[32] and ice hockey players.[34,35] Reduced hip adductor strength was reported as a risk factor for adductor strain,[35] but also as having no association with groin strain[34] in ice hockey players. The former study also found that a reduced ratio of hip adductor to hip abductor strength was a risk factor for adductor strain.[35] Although conflicting findings have been reported, the proposed risk factors suggest that musculoskeletal structures in the pubic region are of primary interest in athletic groin pain.[12] 4. Literature Review 4.1 Scope of the Review

The review was limited to descriptive studies that investigated the anatomy of the pubic region in humans. These include studies that performed gross anatomy or histological examination in human cadavers, or diagnostic imaging in athletes. The human cadaver studies provide greater anatomical detail than the imaging studies, yet may be limited in application due to specimen age and athletic development. Conversely, the imaging studies include subjects within an appropriate demographic, but without the strength of evidence provided by dissection or histology. Surgical studies were not included, as their Sports Med 2009; 39 (3)

Robertson et al.

228

purpose is to describe surgical technique or response to intervention. Subsequently, the review was restricted to primary research studies that applied a standardized procedure in previously uninjured specimens or healthy subjects. The specific focus of the review was the musculoskeletal anatomy of the pubic region, as these are the structures commonly implicated in the pathogenesis of chronic groin pain in athletes. A number of disparities in the descriptions of pubic region anatomy were identified when comparing relevant research studies with modern anatomical texts.[37-42] Interestingly, some of the recently published data are consistent with anatomical descriptions contained in older texts.[43,44] It is important to note that editors of modern texts are challenged to describe human anatomy both accurately and concisely. However, this review concludes that there are three key disparities in musculoskeletal anatomy descriptions of the pubic region that have important implications for the pathogenesis and clinical management of athletes with chronic groin pain. These disparities relate to: (i) the composition and arrangement of the pubic attachments of the adductor longus (AL) muscle; (ii) the arrangement of the lower fibres of the internal oblique (IO) muscle and the lower part of the transversus abdominis (TrA) aponeurosis; and (iii) the relations of the soft tissue structures anterior to the pubic symphysis. The modern textbook descriptions of these structures will be presented and compared with the relevant research evidence that challenges their accuracy. 4.2 Composition of Adductor Longus at its Pubic Attachment

The AL muscle is described in modern texts as arising by a narrow tendon from the body of the pubis in the angle between the crest and the symphysis[37-41] (figure 1). A dissection study by Tuite et al.[45] (n = 37 elderly cadavers) reported that AL attaches to the pubis by a thin tendon anteriorly, consistent with the textbook description. However, they also found that in 92% of specimens, the deep surface of AL was characterized by muscular fibres attaching to the pubis. ª 2009 Adis Data Information BV. All rights reserved.

b

a

Fig. 1. Pubic attachments of (a) adductor longus and (b) rectus abdominis, according to textbook descriptions.

Furthermore, in 24% of specimens, the lateral 5–11 mm of the anterior attachment of AL was composed of muscular fibres. More recently, Strauss et al.[10] undertook a dissection and histological study in 28 elderly cadavers to describe the cross-sectional anatomy of AL at its pubic attachment. The proximal 10 cm of 42 AL muscles were harvested and cut into three cross-sectional samples at 0, 1 and 2 cm from their bony attachment. The cross-sectional area was measured using microcalipers and the fibre composition was determined by microscopic analysis. Consistent with Tuite et al.,[45] the authors found that AL was composed of a thin tendon anteriorly, and muscular fibres on the deep surface of its pubic attachment. The relative contribution of tendon fibres was found to be only 38% (–13.0), and this proportion decreased with further distance from the origin (figure 2). This implies that 62% of the pubic attachment of Sports Med 2009; 39 (3)

The Anatomy of the Pubic Region

229

AL is composed of muscle fibres, which is in striking contrast with textbook descriptions of an entirely tendinous origin.[37-41] 4.3 Arrangement of the Lower Fibres of Internal Oblique and Transversus Abdominis

Fibre type (% of cross-sectional area)

The composition and pubic attachments of the IO and TrA muscles are variably described in modern texts.[37,39,40,42] The lower fibres of IO and the lower part of the TrA aponeurosis are described as fusing to form a conjoint tendon, which then turns downward to attach to the pubic crest and pectineal line.[37,39,40,42] Additionally, Hollinshead and Rosse[40] and Sandring et al.[39] describe medial fibres from the conjoint tendon that extend medially to decussate at, and fuse with, the linea alba.[39,40] The reported functions of IO and TrA provide context for determining the clinical relevance of any variation in their description. Previous research has identified that TrA is controlled independently from IO, external oblique, RA and multifidus.[46] Furthermore, its transverse orientation and attachments suggest its potential involvement in lumbo-pelvic stability during postures and movement.[46,47] The onset of TrA activation has been found to occur in a feed-forward manner during lower limb movements in healthy subjects,[48,49] consistent with an anticipatory stabilization function. Interestingly, although it still occurred in a feed-forward manner, this onset was found to 80 Muscle

70 60 50 40 30

Tendon

20 10 0 1 2 Distance from pubic attachment (cm)

Fig. 2. Fibre composition of adductor longus proximally. Muscle and tendon fibre composition of adductor longus at 0, 1 and 2 cm from its pubic attachment, as reported by Strauss et al.[10]

ª 2009 Adis Data Information BV. All rights reserved.

be significantly delayed in a cohort of Australian Rules footballers with chronic groin pain.[49] A precise anatomical description of the attachments of IO and TrA is required to interpret the proposed functions of these muscles. Several authors have questioned the existence of a true conjoint tendon, albeit mainly from surgical observations rather than descriptive anatomical studies.[50-53] Condon[50] performed detailed dissections of the groin unilaterally in 135 male cadavers that were without hernia or previous regional injury. The author found that a true conjoint tendon occurred in only 4 of 135 (3%) specimens. Of the remaining specimens, there were three patterns of attachment noted. A direct pubic attachment of the TrA aponeurosis was identified in 11/135 (8%) specimens. The remaining 120/135 (89%) specimens were found to attach into the rectus sheath, the majority of which (101/120) were attached more than 0.5 cm above the pubic tubercule, as was the (separate) insertion of IO. 4.4 Confluence of Soft Tissue Structures Anterior to the Pubic Symphysis

The relationship of the soft tissues anterior to the pubic symphysis has been briefly described in modern texts.[39,40,42] However, growing clinical interest in the functional stability of the pelvic ring necessitates reassessment of these relations in detail. The stability of the posterior pelvic ring during the transfer of the weight of the trunk from the sacrum to the hips is derived from the arch-like morphology of the pelvic bones.[54-56] The ‘pelvic arch’ acts to resist shearing forces at the nearly vertical surfaces of the sacroiliac joint. Its strength is provided by fixation of its lateral ends, which requires activation of transversely oriented muscles such as IO, TrA and piriformis.[55,56] Anteriorly, the two pelvic arches are joined at the pubic symphysis.[54] Subsequently, the anatomical arrangement of structures adjacent to the symphysis may be important in developing a comprehensive model of pelvic ring stability. The rectus sheath has been described to fuse with the periosteum adjacent to its pubic insertion.[39] Additionally, confluence of the rectus abdominis (RA), gracilis and fascia lata attachments has Sports Med 2009; 39 (3)

Robertson et al.

230

a b

c

d

e

Fig. 3. Muscle attachments in the pubic region: (a) rectus abdominis; (b) decussating fibres of rectus abdominis; (c) inguinal ligament; (d) attachments of rectus abdominis and adductor longus in continuity; and (e) adductor longus (adapted from Eisler[43]).

also been detailed in the pubic region of males.[39] Schilders[28] reported that the anterior rectus sheath joins the fascia of the common adductors to form an aponeurosis covering the anterior surface of the pubic bone. However, limited details of their methods were provided and findings appear to be based on visual observation only. Similarly, Shortt et al.[57] described the pubic attachments of AL and RA as being continuous both on magnetic resonance imaging (MRI) and in cadaveric specimens. Both AL and RA were reported to fuse in the midline with the capsule of the pubic symphysis,[57] although the study methodology that generated these findings was not reported. A more detailed study was performed by Robinson et al.,[11] who dissected the pubic region of 17 elderly cadavers (male and female) and then compared the findings with MRIs of the anterior pelvis in a group of ten healthy male athletes (mean age 17 years). Dissection revealed that the ª 2009 Adis Data Information BV. All rights reserved.

pubic attachment of AL consisted of tendon anteriorly and muscle fibre more deeply, in agreement with Strauss et al.[10] and Tuite et al.[45] Additionally, AL was attached to the pubic symphysis capsular tissues in all specimens. The composition of this capsular attachment was mixed tendon and muscle fibre in 53% of specimens, and entirely muscle fibre in 47%. RA was also found to attach to the pubic symphysis capsular tissues in all specimens, becoming continuous with AL. More deeply, the pubic symphysis capsular tissues were found to merge with the anterior surface of the interpubic disc and the articular cartilage. The attachment of the AL and RA muscles to the pubic symphysis capsular tissues was also found to be evident on MRI for all subjects. These findings suggest that an intimate relationship exists between AL, RA and the anterior capsular soft tissues of the pubic symphysis (figures 3 and 4). 4.5 Review Summary

This review has identified three key disparities in descriptions of the musculoskeletal anatomy of the pubic region that have implications for the pathogenesis and clinical management of athletes with chronic groin pain. The proximal attachment of AL may be predominantly muscular, rather than entirely tendinous as previously described. The lower fibres of IO and TrA appear to exist more commonly as separate entities attaching into the rectus sheath than as a ‘conjoint tendon’ into the pubic bone. The AL and RA are b a

Fig. 4. Pubic symphysis, coronal section: (a) adductor longus and (b) decussation of adductor longus anterior to the pubic symphysis (adapted from Testut and Latarjet[44]).

Sports Med 2009; 39 (3)

The Anatomy of the Pubic Region

reported to attach in continuity via the capsular tissues of the pubic symphysis. These tissues were also reported to merge with the interpubic disc and adjacent articular cartilage. Finally, the rectus sheath is reported to be continuous with AL via the pubic symphysis capsular tissues. This confluence of soft tissue structures anterior to the pubic symphysis may provide the anatomical substrate for a composite stabilizing or force transmission mechanism. 5. Implications for Pathogenesis and Clinical Management While limited by availability of suitable studies, the findings of this review have several possible implications for the pathogenesis and clinical management of chronic groin pain. The predominantly muscular attachment of AL into the pubic bone implies that the pathogenesis of an adductor-related component of groin pain is likely to be best explained as an enthesopathy, rather than tendinopathy, such as that described for the Achilles or patellar tendons.[58,59] A treatment strategy based exclusively on a tendinopathy paradigm may therefore yield indifferent results. The proximal attachment of AL was also reported to extend beyond the pubic bone to attach in continuity with the RA via the anterior capsular tissues of the pubic symphysis. These descriptions provide a context for evaluating the mechanism of the limited adductor tenotomy procedure, in which the anterior aspect of AL is surgically divided. Orchard et al.[60] hypothesized that this procedure encourages more normal tendon loading, thereby assisting recovery in accordance with a ‘stress-shielding’ aetiology. However, the predominantly muscular composition of AL at its pubic attachment suggests that this hypothesis requires further consideration. An alternative explanation for the apparent clinical benefit of the limited adductor tenotomy procedure is that it may disrupt connections between AL and the capsular tissues, and thus reduce stress loading on the anterior pubic symphysis. The anterior relations of the pubic symphysis identify that the AL, RA, IO and TrA ª 2009 Adis Data Information BV. All rights reserved.

231

muscles have the potential to contribute to a composite mechanism that could provide pelvic ring stability and force transmission (figure 5). These functions are particularly relevant when considering that repeated loading of the pubic region is commonly described in the aetiology of chronic groin pain.[5,14,61] While the functional interaction between these structures has been reported previously,[62,63] the anatomical connections described in this review have further implications. A direct anatomical connection provides a mechanism for overlapping pathologies, thereby strengthening the hypothesis that chronic groin pain is likely to involve multiple structures. If a single musculoskeletal structure was involved, it would be reasonable to expect that symptoms, assessment findings and outcomes would be relatively predictable. However, clinical evidence suggests that chronic groin pain

a

b

c

Fig. 5. Potential load transfer pathways. The arrangement of the pubic fibres of adductor longus and rectus abdominis may allow direct load transfer between the muscles ipsilaterally (a) and contralaterally (b). Decussating fibres of adductor longus may allow load transfer between the muscle pair (c).

Sports Med 2009; 39 (3)

Robertson et al.

232

is often associated with vague, diffuse symptoms, inconsistent clinical findings, and a varied response to interventions.[3,5,14,18,64] In this context, the pathogenesis of chronic groin pain is more consistent with a multiple entities paradigm, perhaps reflecting a common reaction pattern to repetitive pubic region loading.[5,14,61] The confluence of structures anterior to the pubic symphysis may question the validity of some clinical examination tests used routinely in the assessment of chronic groin pain. The direct anatomical connections between AL, RA, IO and TrA muscles suggest that it is unlikely that pain provocation or stress tests load single anatomical structures in isolation. Active hip adduction against resistance is a test that is commonly performed in the clinical examination of athletes with chronic groin pain. In the authors’ experience, this test can in some instances reproduce suprapubic rather than infrapubic pain. This clinical finding may be attributable to synergies between AL and the lower abdominal musculature. Alternatively, the finding may be due to a provocative force initiated by AL, and then delivered to the superior aspect of the pubis via connections with the pubic symphysis capsular tissues and lower abdominals. The reliability and validity of clinical examination tests commonly used in the assessment of athletes with chronic groin pain have been investigated in a variety of populations.[7,36,63,65-68] However, the diagnostic validity of these tests is yet to be established. The complex, interdependent anatomy revealed in this review suggests that it is challenging for the sports medicine practitioner to precisely diagnose structures associated with chronic groin pain. With the current available evidence, the role of clinical examination in the assessment of chronic groin pain may be limited to identifying the musculoskeletal deficits associated with the condition. 6. Conclusion This review of the literature has identified several descriptions of pubic region anatomy that differ from those provided in modern texts. These variations have possible implications for underª 2009 Adis Data Information BV. All rights reserved.

standing the pathogenesis of chronic groin pain and its clinical management. The proximal attachment of AL was reported to be predominantly muscular, rather than fibro-tendinous. Subsequently, the pathogenesis of insertional adductor pain may be best explained as an enthesopathy, rather than tendinopathy. The IO and TrA muscles were found to attach medially into the distal rectus sheath, rather than directly into the pubis. In addition to an extension from AL, the rectus sheath then forms a confluence of aponeurotic structures anterior to the pubic symphysis. While a functional connection between AL, IO, TrA and RA has been previously suggested, a direct anatomical connection has broader implications. A composite structure anterior to the pubic symphysis provides the anatomical substrate for the multiple clinical entities model of chronic groin pain. It provides an alternative mechanism for the surgical procedure of limited adductor tenotomy, in which disruption of the connections between AL and the capsular tissues may reduce the stress loading on the pubic symphysis. It also suggests that it is unlikely that pain provocation or stress tests load single anatomical structures in isolation. This feature may provide an explanation for the authors’ clinical observation that resisted hip adduction can in some instances reproduce suprapubic rather than infrapubic pain. Additionally, the role of clinical examination in the assessment of chronic groin pain may be limited to identifying the musculoskeletal deficits associated with the condition. Subsequently, a functional approach to conservative management may be appropriate in the absence of definitive clinical findings. The findings of this review have several possible implications for understanding the pathogenesis of chronic groin pain and its clinical management. It is important to note that this review was limited to the small number of descriptive studies that specifically investigated the anatomy of pubic region in humans. Subsequently, the findings of this review should be hypothesis generating, with further anatomical and clinical studies warranted to establish their veracity. Sports Med 2009; 39 (3)

The Anatomy of the Pubic Region

Acknowledgements No funding was provided for the preparation of this article and the authors have no conflicts of interest directly relevant to its contents.

References 1. Ekstrand J, Hilding J. The incidence and differential diagnosis of acute groin injuries in male soccer players. Scand J Med Sci Sports 1999 Apr; 9 (2): 98-103 2. Emery CA, Meeuwisse WH, Powell JW. Groin and abdominal strain injuries in the National Hockey League. Clin J Sport Med 1999 Jul; 9 (3): 151-6 3. Orchard J, Verrall GM. Groin injuries in the Australian football league [online]. Int SportMed J 2000; 1 Suppl. 1 [Accessed 2008 Apr 23] 4. Orchard J, Read JW, Verrall GM, et al. Pathophysiology of chronic groin pain in the athlete [online]. Int SportMed J 2000; 1 Suppl. 1 [Accessed 2008 Apr 23] 5. Besjakov J, Von Scheele C, Ekberg O, et al. Grading scale of radiographic findings in the pubic bone and symphysis in athletes. Acta Radiol 2003 Jan; 44 (1): 79-83 6. Ekberg O, Persson NH, Abrahamsson PA, et al. Longstanding groin pain in athletes. A multidisciplinary approach. Sports Med 1988 Jul; 6 (1): 56-61 7. Holmich P. Long-standing groin pain in sportspeople falls into three primary patterns, a ‘‘clinical entity’’ approach: a prospective study of 207 patients. Br J Sports Med 2007 Apr; 41 (4): 247-52; discussion 52 8. Lovell G. The diagnosis of chronic groin pain in athletes: a review of 189 cases. Aust J Sci Med Sport 1995 Sep; 27 (3): 76-9 9. Orchard J, Wood T, Seward H, et al. Comparison of injuries in elite senior and junior Australian football. J Sci Med Sport 1998 Jun; 1 (2): 83-8 10. Strauss EJ, Campbell K, Bosco JA. Analysis of the crosssectional area of the adductor longus tendon: a descriptive anatomic study. Am J Sports Med 2007 Jun; 35 (6): 996-9 11. Robinson P, Salehi F, Grainger A, et al. Cadaveric and MRI study of the musculotendinous contributions to the capsule of the symphysis pubis. AJR Am J Roentgenol 2007 May; 188 (5): W440-5 12. Anderson K, Strickland SM, Warren R. Hip and groin injuries in athletes. Am J Sports Med 2001 Jul-Aug; 29 (4): 521-33 13. Fon LJ, Spence RA. Sportsman’s hernia. Br J Surg 2000 May; 87 (5): 545-52 14. Lynch SA, Renstrom PA. Groin injuries in sport: treatment strategies. Sports Med 1999 Aug; 28 (2): 137-44 15. Macintyre J, Johson C, Schroeder EL. Groin pain in athletes. Curr Sports Med Rep 2006 Dec; 5 (6): 293-9 16. Moeller JL. Sportsman’s hernia. Curr Sports Med Rep 2007 Apr; 6 (2): 111-4 17. Morelli V, Espinoza L. Groin injuries and groin pain in athletes: part 2. Prim Care 2005 Mar; 32 (1): 185-200 18. Morelli V, Smith V. Groin injuries in athletes. Am Fam Physician 2001 Oct 15; 64 (8): 1405-14

ª 2009 Adis Data Information BV. All rights reserved.

233

19. Morelli V, Weaver V. Groin injuries and groin pain in athletes: part 1. Prim Care 2005 Mar; 32 (1): 163-83 20. Nicholas SJ, Tyler TF. Adductor muscle strains in sport. Sports Med 2002; 32 (5): 339-44 21. Albers SL, Spritzer CE, Garrett Jr WE, et al. MR findings in athletes with pubalgia. Skeletal Radiol 2001 May; 30 (5): 270-7 22. Swan Jr KG, Wolcott M. The athletic hernia: a systematic review. Clin Orthop Relat Res 2007 Feb; 455: 78-87 23. Lovell G, Galloway H, Hopkins W, et al. Osteitis pubis and assessment of bone marrow edema at the pubic symphysis with MRI in an elite junior male soccer squad. Clin J Sport Med 2006 Mar; 16 (2): 117-22 24. Orchard JW, Read JW, Neophyton J, et al. Groin pain associated with ultrasound finding of inguinal canal posterior wall deficiency in Australian Rules footballers. Br J Sports Med 1998 Jun; 32 (2): 134-9 25. Brennan D, O’Connell MJ, Ryan M, et al. Secondary cleft sign as a marker of injury in athletes with groin pain: MR image appearance and interpretation. Radiology 2005 Apr; 235 (1): 162-7 26. Ekstrand J, Gillquist J. Soccer injuries and their mechanisms: a prospective study. Med Sci Sports Exerc 1983; 15 (3): 267-70 27. Holmich P, Uhrskou P, Ulnits L, et al. Effectiveness of active physical training as treatment for long-standing adductor-related groin pain in athletes: randomised trial. Lancet 1999 Feb 6; 353 (9151): 439-43 28. Schilders E. Groin injuries in athletes. Current Orthopaedics 2000 Nov; 14 (6): 418-23 29. Schilders E, Bismil Q, Robinson P, et al. Adductor-related groin pain in competitive athletes: role of adductor enthesis, magnetic resonance imaging, and entheseal pubic cleft injections. J Bone Joint Surg Am 2007 Oct; 89 (10): 2173-8 30. Bradshaw C, Holmich P. Longstanding groin pain. In: Brukner P, Khan K, editors. Clinical sports medicine. 3rd ed. Sydney (NSW): McGraw-Hill, 2007 31. Arnason A, Sigurdsson SB, Gudmundsson A, et al. Risk factors for injuries in football. Am J Sports Med 2004 Jan-Feb; 32 (1 Suppl.): 5S-16S 32. Witvrouw E, Danneels L, Asselman P, et al. Muscle flexibility as a risk factor for developing muscle injuries in male professional soccer players: a prospective study. Am J Sports Med 2003 Jan-Feb; 31 (1): 41-6 33. Ibrahim A, Murrell GA, Knapman P. Adductor strain and hip range of movement in male professional soccer players. J Orthop Surg (Hong Kong) 2007 Apr; 15 (1): 46-9 34. Emery CA, Meeuwisse WH. Risk factors for groin injuries in hockey. Med Sci Sports Exerc 2001 Sep; 33 (9): 1423-33 35. Tyler TF, Nicholas SJ, Campbell RJ, et al. The association of hip strength and flexibility with the incidence of adductor muscle strains in professional ice hockey players. Am J Sports Med 2001 Mar-Apr; 29 (2): 124-8 36. Verrall GM, Hamilton IA, Slavotinek JP, et al. Hip joint range of motion reduction in sports-related chronic groin injury diagnosed as pubic bone stress injury. J Sci Med Sport 2005 Mar; 8 (1): 77-84 37. McMinn RMH, editor. Last’s anatomy: regional and applied. Edinburgh: Churchill Livingstone, 1990

Sports Med 2009; 39 (3)

234

38. Moore K, Dalley 2nd AF. Clinically oriented anatomy. 5th ed. Philadelphia (PA): Lippincott Williams & Wilkins, 2005 39. Sandring S, Ellis H, Healy JC, et al., editors. Gray’s anatomy: the anatomical basis of clinical practice. 39th ed. Edinburgh: Churchill Livingstone, 2005 40. Hollinshead WH, Rosse C. Textbook of anatomy. 4th ed. Philadelphia (PA): Harper & Row, 1985 41. Romanes GJ. Cunningham’s manual of practical anatomy. Volume I: upper and lower limbs. 15th ed. Oxford: Oxford University Press, 1986 42. Romanes GJ. Cunningham’s manual of practical anatomy. Volume II: thorax and abdomen. 15th ed. Oxford: Oxford University Press, 1986 43. Eisler P. Die muskeln des stammes. Jena: Gustav Fisher, 1912 44. Testut L, Latarjet A. Traite d’anatomie humaine. Paris: G. Doin & Compagnie, 1948 45. Tuite DJ, Finegan PJ, Saliaris AP, et al. Anatomy of the proximal musculotendinous junction of the adductor longus muscle. Knee Surg Sports Traumatol Arthrosc 1998; 6 (2): 134-7 46. Hodges PW, Richardson CA. Contraction of the abdominal muscles associated with movement of the lower limb. Phys Ther 1997 Feb; 77 (2): 132-42; discussion 42-4 47. Richardson CA, Snijders CJ, Hides JA, et al. The relation between the transversus abdominis muscles, sacroiliac joint mechanics, and low back pain. Spine 2002 Feb 15; 27 (4): 399-405 48. Hodges PW, Richardson CA. Feedforward contraction of transversus abdominis is not influenced by the direction of arm movement. Exp Brain Res 1997 Apr; 114 (2): 362-70 49. Cowan SM, Schache AG, Brukner P, et al. Delayed onset of transversus abdominus in long-standing groin pain. Med Sci Sports Exerc 2004 Dec; 36 (12): 2040-5 50. Condon RE. Reassessment of groin anatomy during the evolution of preperitoneal hernia repair. Am J Surg 1996 Jul; 172 (1): 5-8 51. Fagan SP, Awad SS. Abdominal wall anatomy: the key to a successful inguinal hernia repair. Am J Surg 2004 Dec; 188 (6A Suppl.): 3S-8S 52. Skandalakis JE, Gray SW, Skandalakis LJ, et al. Surgical anatomy of the inguinal area. World J Surg 1989 Sep-Oct; 13 (5): 490-8 53. Skandalakis JE, editor. Surgical anatomy: the embryologic and anatomic basis of modern surgery. 1st ed. Athens: PMP, 2004 54. Gamble JG, Simmons SC, Freedman M. The symphysis pubis: anatomic and pathologic considerations. Clin Orthop Relat Res 1986 Feb; (203): 261-72 55. Snijders CJ, Ribbers MT, de Bakker HV, et al. EMG recordings of abdominal and back muscles in various standing postures: validation of a biomechanical model on sacroiliac joint stability. J Electromyogr Kinesiol 1998 Aug; 8 (4): 205-14

ª 2009 Adis Data Information BV. All rights reserved.

Robertson et al.

56. Snijders CJ, Vleeming A, Stoeckart R. Transfer of lumbosacral load to iliac bones and legs. Part 1: biomechanics of self-bracing of the sacroiliac joints and its significance for treatment and exercise. Clin Biomech 1993; 8 (6): 285-94 57. Shortt CP, Zoga AC, Kavanagh EC, et al. Anatomy, pathology, and MRI findings in the sports hernia. Semin Musculoskelet Radiol 2008 Mar; 12 (1): 54-61 58. Benjamin M, McGonagle D. The anatomical basis for disease localisation in seronegative spondyloarthropathy at entheses and related sites. J Anat 2001 Nov; 199 (Pt 5): 503-26 59. Benjamin M, Moriggl B, Brenner E, et al. The ‘‘enthesis organ’’ concept: why enthesopathies may not present as focal insertional disorders. Arthritis Rheum 2004 Oct; 50 (10): 3306-13 60. Orchard JW, Cook JL, Halpin N. Stress-shielding as a cause of insertional tendinopathy: the operative technique of limited adductor tenotomy supports this theory. J Sci Med Sport 2004 Dec; 7 (4): 424-8 61. LeBlanc KE, LeBlanc KA. Groin pain in athletes. Hernia 2003 Jun; 7 (2): 68-71 62. Major NM, Helms CA. Pelvic stress injuries: the relationship between osteitis pubis (symphysis pubis stress injury) and sacroiliac abnormalities in athletes. Skeletal Radiol 1997 Dec; 26 (12): 711-7 63. Mens J, Inklaar H, Koes BW, et al. A new view on adductionrelated groin pain. Clin J Sport Med 2006 Jan; 16 (1): 15-9 64. Kavanagh EC, Koulouris G, Ford S, et al. MR imaging of groin pain in the athlete. Semin Musculoskelet Radiol 2006 Sep; 10 (3): 197-207 65. Holmich P, Holmich LR, Bjerg AM. Clinical examination of athletes with groin pain: an intraobserver and interobserver reliability study. Br J Sports Med 2004 Aug; 38 (4): 446-51 66. Slavotinek JP, Verrall GM, Fon GT, et al. Groin pain in footballers: the association between preseason clinical and pubic bone magnetic resonance imaging findings and athlete outcome. Am J Sports Med 2005 Jun; 33 (6): 894-9 67. Verrall GM, Slavotinek JP, Barnes PG, et al. Description of pain provocation tests used for the diagnosis of sportsrelated chronic groin pain: relationship of tests to defined clinical (pain and tenderness) and MRI (pubic bone marrow oedema) criteria. Scand J Med Sci Sports 2005 Feb; 15 (1): 36-42 68. Verrall GM, Slavotinek JP, Fon GT. Incidence of pubic bone marrow oedema in Australian Rules football players: relation to groin pain. Br J Sports Med 2001 Feb; 35 (1): 28-33

Correspondence: Dr Anthony G. Schache, Department of Mechanical Engineering, Melbourne School of Engineering, The University of Melbourne, VIC 3010, Australia. E-mail: [email protected]

Sports Med 2009; 39 (3)

Sports Med 2009; 39 (3): 235-256 0112-1642/09/0003-0235/$49.95/0

REVIEW ARTICLE

ª 2009 Adis Data Information BV. All rights reserved.

Annual Age-Grouping and Athlete Development A Meta-Analytical Review of Relative Age Effects in Sport Stephen Cobley,1 Joseph Baker,2 Nick Wattie1 and Jim McKenna1 1 Carnegie Research Institute, Leeds Metropolitan University, Leeds, West Yorkshire, UK 2 School of Kinesiology and Health Science, York University, Toronto, Ontario, Canada

Contents Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Explanations for Relative Age Effects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Rationale for a Meta-Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Study Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Sample of Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Study Review Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Data Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Overall Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Subgroup Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1 Sex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2 Age Category . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3 Skill Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.4 Sport Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 General Findings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Context-Specific Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Eliminating Relative Age Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4 Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

235 236 236 237 238 238 238 238 239 239 239 240 240 240 240 240 240 240 251 251 253 254

Annual age-grouping is a common organizational strategy in sport. However, such a strategy appears to promote relative age effects (RAEs). RAEs refer both to the immediate participation and long-term attainment constraints in sport, occurring as a result of chronological age and associated physical (e.g. height) differences as well as selection practices in annual agegrouped cohorts. This article represents the first meta-analytical review of RAEs, aimed to collectively determine (i) the overall prevalence and strength of RAEs across and within sports, and (ii) identify moderator variables. A total of 38 studies, spanning 1984–2007, containing 253 independent samples across 14 sports and 16 countries were re-examined and included in a single analysis using odds ratios and random effects procedures for combining

Cobley et al.

236

study estimates. Overall results identified consistent prevalence of RAEs, but with small effect sizes. Effect size increased linearly with relative age differences. Follow-up analyses identified age category, skill level and sport context as moderators of RAE magnitude. Sports context involving adolescent (aged 15–18 years) males, at the representative (i.e. regional and national) level in highly popular sports appear most at risk to RAE inequalities. Researchers need to understand the mechanisms by which RAEs magnify and subside, as well as confirm whether RAEs exist in female and more culturally diverse contexts. To reduce and eliminate this social inequality from influencing athletes’ experiences, especially within developmental periods, direct policy, organizational and practitioner intervention is required.

1. Background Within many sport contexts, the youth stages of participation are often organized into annual age-groups using specific cut-off dates (e.g. 1 September in the UK). Whilst with honourable intention and for the purposes of competition organization and values of fair play, such a policy remains insensitive to the subtle chronological age differences (referred to as ‘relative age’ differences) between members within an annual cohort.[1] These differences are associated with immediate and long-term consequences, commonly known as ‘relative age effects’ (RAEs).[1-3] Grondin et al.[2] were the first to assess the consequences of annual age-grouping in sports, following consistent reports of attainment differentials according to relative age in education.[4-6] They examined the birth-date distributions of Canadian ice-hockey and volleyball players, participating at recreational, competitive and senior professional levels for the 1981–2 season. Their results identified significant and repeated over-representations of ice-hockey players born in the first quartile (i.e. the 3 months after age-group cut-off dates) for each age-group category and level of competition, including professionals, while in volleyball over-representations were observed for the elite representative levels. Barnsley et al.[1] also identified birth-date differentials amongst ice-hockey players in the Canadian elite developmental leagues and National Hockey League (NHL) for 1983–4, and later found similar inequalities in the junior representative leagues (at ages ‡11 years).[3] ª 2009 Adis Data Information BV. All rights reserved.

Together, these studies suggested that being relatively older within an annual sporting cohort provided significant attainment advantages when compared with those who were relatively younger. Many studies have identified similar differentials in birth-date patterns across youth agegroups and levels of competition for the sports of baseball,[7-8] ice hockey,[9] soccer[10-12] and tennis.[13,14] Studies have also identified RAEs in other sports, but essentially in high performing samples, including Australian Rules football,[15] cricket,[16,17] netball[17] and both codes of rugby.[15] It is important to note that RAEs are not universal. In fact, in several contexts (e.g. golf[18]) RAEs have not been identified or predicted to occur. These contexts are typically free of annual age-grouping and other requisite precursor conditions (e.g. selection processes in tiers of youth competition). 1.1 Explanations for Relative Age Effects

Although previous studies (until recently[19]) did not include physical or maturational indices, most suggested physical differences (i.e. greater chronological age and likelihood of more advanced physical characteristics) as being primarily responsible for RAEs.[3,20,21] Attributes of greater height, mass (to a degree), aerobic power, muscular strength, endurance and speed do provide performance advantages in most sports.[22-23] Furthermore, during adolescence, a time when annual age-groupings are employed and where sport competition can be intensive, a 1-year age difference, especially during the stages of puberty Sports Med 2009; 39 (3)

Relative Age Effects in Sport

(i.e. 13–15 years of age in boys; 12–14 years in girls) can heighten physical[24,25] and performance[23,26] differences. Thus, relatively older athletes may have an increased likelihood of exhibiting advanced physical characteristics and entering puberty earlier, compared with their relatively younger peers. In sports where body size, strength and power convey advantages, elite junior athletes have been identified as above average for height and weight when compared with age-matched normative data (e.g. soccer[27,28]). Likewise in gymnastics, where height and mass gain impedes flexibility, rotational speed and the strength to mass ratio, maturational delay in more highly skilled gymnasts has been observed.[14] In fact, a greater frequency of relatively younger gymnasts has been reported in high performance contexts.[29] A complementary and interacting mechanism, relating to selection and experience, has also been proposed to account for the long-term propagation of RAEs. Being relatively older is more likely to provide a performance and selection advantage when assessed or evaluated (by coaches) against annual age-group peers. This selection advantage increases the likelihood of access to higher levels of competition, training and coaching.[30] It is likely that such access will be accompanied by increases in volumes of practice, training load and competition frequency, thereby generating an experience advantage over nonselected and likely relatively younger peers. In contrast, those not selected are considered less able to access practice and coaching expertise facilities, or higher levels of competition, constraining their sporting involvement and development. Events associated with selection, trials or talent identification are thus postulated to differentiate an individual’s ability to invest in practice and accumulate sport-specific skill and experience, factors deemed critical for attainment.[31,32] Selection and exposure to practice and match-play may provide significant technical and game intelligence advantages[33,34] to selected relatively older players, accounting for their overrepresentation in senior professional sports. Other interacting psychological and broader sociocultural mechanisms have also been preª 2009 Adis Data Information BV. All rights reserved.

237

sented to account for RAEs. Linked with selection and experience differences according to relative age, psychological disparities have also been suggested.[21] Relatively older players may be more likely to develop higher perceptions of competence[35,36] and self-efficacy.[37] In comparison, relatively younger athletes, faced with consistent sport selection disadvantages, may be more likely to have negative sport experiences, develop low competence perceptions, and thus terminate sport involvement.[38-40] Related to sociocultural influences, two studies have associated population and sport participation growth with heightened competition in youth sport contexts, and thereby an inflated likelihood of RAEs.[41,42] Likewise, sport policies that have attempted to address performance concerns on the international stage by adopting earlier competition, talent identification and streaming have also been associated with the first appearances of RAEs in sport.[41] Such sociocultural forces should be kept in mind with reference to the rationale and purpose for the present study. 1.2 Rationale for a Meta-Analysis

RAEs appear to be complex phenomena, with sociocultural antecedents combining with interindividual age and physical differences to affect sport attainment. To date, a variety of sports contexts differing in age categories, levels of competition and cultures have been assessed for RAEs. Since the narrative review of Musch and Grondin,[21] many more samples within studies have been collected and examined, yet several questions remain unresolved. For example, how prevalent and robust are RAEs across and within sports contexts? What factors modify the risk of RAEs in a sports context? By employing meta-analytical methods, these questions can be addressed as most previous studies provided consistent sample and sport context information, often presenting birth-date frequency data (i.e. the proxy measure to identify relative age in an annual age-grouping) in quarterly or halfyearly distributions. Such information is valuable if support toward direct intervention is to be generated. Furthermore, by identifying potential Sports Med 2009; 39 (3)

Cobley et al.

238

factors that moderate risk size, sports contexts can consider strategies that will help address and remove the unnecessary RAE inequality. 1.3 Study Purpose

The purpose of this meta-analytical review was to generate a broad picture of RAE prevalence in sport by systematically re-examining the numerous ‘snap-shots’ taken of sport contexts in previous studies. For the first time, metaanalytical methods were used to ascertain the risk size and moderating factors of RAEs across and within sports. We hypothesized that RAEs were apparent across sexes in highly popular team sport contexts where (i) annual age-grouping policies were employed in youth participation stages, and (ii) youth stages would include intensive competition and a skill level hierarchy, which involved selection mechanisms regulating access to higher levels of competition. In an attempt to directly explain and account for the consistent birth-date discrepancies in senior professional sport, we further hypothesized that RAE risk size increased with skill level and chronological age. 2. Methods

polated along with the type of sport, level of competition, country and competition year(s) of data collection. Those studies reporting birthdate distribution in quartiles (i.e. per 3 months), halves (i.e. 6 months) or both were included. After obtaining all listed studies, reference sections were further examined to locate other relevant studies. All articles written in French and English were identified and interpreted. At the time of writing, we were not aware of any RAE-related publications in other languages. Finally, where sample and sport context information was not presented, authors were contacted for respective information. Five authors were contacted for further information, with three able to return required information, which often related to sample characteristics (e.g. sport context, skill level). These authors were also asked if they were aware of any additional studies not included in our list. No study additions were made through this procedure.1 The overall process yielded 38 studies, spanning 1984 (i.e. the first published study of RAEs in sport[2]) to January 2007,2 with 253 independent samples in 14 sports, across 16 countries. Participants involved in identified studies were either current or former sports players who competed at a range of levels, including recreational, junior and national senior representatives.

2.1 Sample of Studies

Published research papers, including those published in peer-reviewed conference proceedings, were tracked, collected and analysed over a 3-month period, specifically November 2006 to January 2007. This included searches of PubMed, PsychINFO and PsychARTICLES databases using the keywords ‘relative age effect’, ‘birthdate effect’, ‘season of birth’, ‘age position’ and their derivations. Additional criteria for inclusion in the meta-analysis were that papers reported both sample characteristics and information regarding the sport context. Sample characteristics pertaining to birth date, sex and chronological age at the time of data collection were extra-

2.2 Study Review Procedure

All articles were read and examined in full by the first author. Information extracted from each of the studies was categorized and cross-validated by an independent reviewer. No categorization or reporting accuracy errors were evident in the extraction of variable information. Of the 38 studies, only three (reflecting ten samples) failed to report complete sample information for either total sample size or the distribution of birth dates after contacting the authors. These samples therefore had to be excluded from particular or whole aspects of the data analysis.

1 The search and data collection procedure missed one study, which came to light following data analysis.[43] 2 Since the time of data collection and analysis several more studies have been published.

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (3)

Relative Age Effects in Sport

2.3 Data Analysis

Across sport contexts, studies generally standardized relative age differences by categorizing birth date into quartiles (i.e. 3-month periods) following cut-off dates. From these original data, odds ratios (ORs) and 95% confidence intervals (CIs) were calculated for both quartile and halfyear distributions.3 Specifically, for each sample reported, birth-date distributions (e.g. number of people in quartile 1) were compared against an expected frequency, assuming an equal distribution (e.g. N = 100, expected quartile count = 100/4 = 25). This comparator (or control group) value was utilized similar to previous studies (except for Grondin et al.[2]). To clarify this assumption, however, national population statistics were checked and findings suggest that variations in birth-date distribution occur across a calendar year. For example, for the period of 1970–2000 in Canada and the UK (countries where many relative age studies have been conducted), a consistently higher number of births occurred in the spring–summer (i.e. April–August) months.[45,46] In comparison, the months/quartiles coinciding with dates used for age-grouping in sport for these countries reported a consistently lower number of births. So, while the assumption of equal distribution was not completely accurate, birth distributions were not correlated with age-grouping and participation trends in sport, thus national census data were deemed unlikely to bias or influence sample OR calculations. When comparing quartiles and half-year distributions in all OR analyses, quartile 4 (i.e. the relatively youngest members) and the second 6 months of annual age-groupings were assigned as referent groups. Overall summary effect sizes were calculated using DerSimonian and Laird[47] methods for combining samples (see Sutton et al.[48]). Since heterogeneity between studies was expected because of variety in sport contexts and samples characteristics, a random effects model

239

was used. The outcomes were weighted by the inverse variance. Heterogeneity was assessed using the Cochran Q value.[49] When heterogeneity was detected, sources of heterogeneity were explored using sub-stratification analysis. All analyses were conducted using either Microsoft Excel or RevMan 4.2.[50] 3. Results 3.1 Overall Results

For quartile analyses, the birth dates of 124 524 sport participants (former or present) in 246 samples were compared. Descriptive analyses identified an uneven distribution of birth dates in the overall sample (i.e. quartile 1 [Q1] = 31.2%; quartile 2 [Q2] = 26.1%; quartile 3 [Q3] = 22.3%; quartile 4 [Q4] = 20.6%). For half-year comparisons, data from seven additional samples were included,[16,51] raising total participants to 130 108 across 253 independent samples. This sample equated to 57.26% (born in the first 6 months of an age-grouping year) and 42.74% (born in the second 6 months of an age-grouping year). Based on established criteria for interpreting effect sizes,[52] DerSimonian and Laird[47] procedures revealed a significant overall, but small, OR of 1.65 (95% CI 1.54, 1.77; Z = 14.46, p < 0.001) across all samples for the likelihood of sports participants to be born in Q1 versus Q4 of an age-grouping year. Heterogeneity was also evident between samples (Q value = 1731.1, degrees of freedom [df] = 245, p < 0.0001). A decreasing linear trend of RAE risk was identified following comparisons between Q2 and Q4, with an overall OR of 1.37 (95% CI 1.30, 1.44; Z = 11.59, p < 0.001) and Q3 and Q4, OR 1.13 (95% CI 1.10, 1.16; Z = 7.88,p < 0.001). Heterogeneity was also apparent for these comparisons (Q2 and Q4: Q value = 957.9, df = 245, p < 0.0001; Q3 and Q4: Q value = 306.2, df = 245, p < 0.005). Relative age effect sizes were also small when comparing

3 An odds ratio is considered as a comparison between the odds of exposure (i.e. to a sport context) compared to the odds of exposure (i.e. general population). Confidence intervals quantify the uncertainty in measurement. It is usually reported as 95% CI, which is the range of values within which we can be 95% sure that the true value for the whole population lies. See Rudas[44] for an introduction.

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (3)

Cobley et al.

240

between the first 6 months and the second 6 months, with an OR of 1.39 (95% CI 1.32, 1.47, Z = 12.73, p < 0.0001) found. Heterogeneity between studies was again apparent (Q value = 1416.9, df = 252, p < 0.001). For all analyses, funnel plot assessments did not suggest publication bias was evident.[53,54] However, as evidence for heterogeneity was consistent, follow-up subgroup stratification analyses were conducted (as recommended by Gelber and Goldhirsch[55] as well as Yusuf et al.[56]) to identify possible sources of influence. Similar procedures were used; however, only comparisons between Q1 versus Q4 and the first 6 months versus the second 6 months were made. 3.2 Subgroup Results 3.2.1 Sex

Tables I and II, respectively, show the results of ORs for individual male and female samples as well as overall summary analyses. Of these samples, only 24 directly examined relative age effects in female athletes, comprising 3321 (or 2%) of all participants. Considering samples available, sex made little difference to the overall ORs, whether based on quartile or half-yearly distributions (males Q1 vs Q4 = OR: 1.65, 95% CI 1.54, 1.77; first 6 months vs second 6 months = OR: 1.39, 95% CI 1.32, 1.47; females Q1 vs Q4 = OR: 1.21, 95% CI 1.10, 1.33; first half vs second half = OR: 1.39, 95% CI 1.26, 1.54). 3.2.2 Age Category

To consider age as a moderator of risk, ages within samples were categorized into child (<11 years), junior (11–14 years), adolescent (15–18 years) and senior (>18 years) for subcategory analyses. We excluded samples from the analysis where ages spanned across these categories (e.g. Baxter-Jones et al.[25]). Table III summarizes results of age-category analyses. Overall summary calculations identified small significant effects across age categories, regardless of whether relative age was considered in quarter- or half-yearly distributions. Risk progressively increased with age from the child category to the adolescent (15–18 years) age ª 2009 Adis Data Information BV. All rights reserved.

range. For the comparison between Q1 and Q4, small-moderate effects (OR: 2.36, 95% CI 2.00, 2.79) were evident at the adolescent stage, before declining at the senior (19 years plus) age category (OR: 1.44, 95% CI 1.35, 1.53). 3.2.3 Skill Level

Prior to analysis, all samples were categorized into one of four skill levels: recreational (e.g. leisure and house leagues), competitive (often associated with juniors and amateurs), representative (often associated with regional and national representation) and elite (regarded as professional or senior national representative). Overall, summary results identified small significant ORs regardless of skill category (see table IV); however, risk increased with skill level, with the highest risk evident at the representative (preelite) stage (i.e. Q1 vs Q4 = OR: 2.77, 95% CI 2.36, 3.24). Interestingly, summary ORs suggest that the risks of RAEs are lower at the elite stage than in the representative stage (OR: 1.42, 95% CI 1.34, 1.51). 3.2.4 Sport Context

While 14 sports have been assessed for relative age effects, most studies focused upon ice hockey (32.8%), soccer (30%) and baseball (13%). Regardless of whether quartile or half-yearly summary ORs were considered, small effects were apparent in these sports, as well as in basketball and volleyball (i.e. the next two mostly examined contexts; see table V). Only in American Football were ORs non-significant. Other sports contexts were not examined due to low sample numbers. 4. Discussion 4.1 General Findings

This article represents the first meta-analytical study of RAEs, synthesizing data from samples in previous research (spanning 1984 to January 2007) into a single analysis, whilst partially controlling for wider population trends in this period. Its primary purpose was to determine the overall prevalence and strength of RAEs in sport. A secondary purpose was to identify risk change according to moderator variables, with such Sports Med 2009; 39 (3)

Study

Subject age (y)

Sport

Level of competition

No. of subjects

Grondin et al.[2]

8–9

Ice hockey

Junior AA

8–9

Ice hockey

Junior BB

171

2.59 (1.35, 4.96)

2.27 (1.17, 4.38)

1.90 (0.97, 3.72)

1.67 (0.59, 2.82)

8–9

Ice hockey

Junior CC

256

2.37 (1.43, 3.94)

1.67 (0.99, 2.82)

1.35 (0.78, 2.3)

1.72 (0.65, 2.64)

8–9

Ice hockey

Novice recreation

110

1.60 (0.74, 3.45)

2.05 (0.96, 4.34)

0.85 (0.36, 1.95)

1.97 (0.51, 3.81)

10–11

Ice hockey

Junior AA

124

5.27 (2.33, 11.9)

2.90 (1.24, 6.78)

2.09 (0.87, 5.01)

2.64 (0.52, 4.99)

10–11

Ice hockey

Junior BB

206

2.31 (1.31, 4.07)

1.93 (1.08, 3.44)

1.18 (0.64, 2.18)

1.94 (0.61, 3.14)

10–11

Ice hockey

Junior CC

273

1.58 (0.96, 2.61)

2.04 (1.25, 3.32)

1.30 (0.78, 2.17)

1.57 (0.66, 2.38)

10–11

Ice hockey

Novice recreation

138

1.24 (0.62, 2.44)

1.03 (0.51, 2.07)

1.48 (0.76, 2.88)

0.91 (0.56, 1.63)

12–13

Ice hockey

Junior AA

120

3.05 (1.45, 6.44)

1.94 (0.89, 4.2)

1.05 (0.45, 2.43)

2.42 (0.52, 4.61)

12–13

Ice hockey

Junior BB

202

3.09 (1.66, 5.74)

3.04 (1.63, 5.65)

2.04 (1.07, 3.88)

2.01 (0.61, 3.28)

12–13

Ice hockey

Junior CC

298

1.21 (0.76, 1.93)

1.46 (0.93, 2.3)

0.96 (0.6, 1.55)

1.36 (0.67, 2.02)

12–13

Ice hockey

Novice recreation

90

1.09 (0.47, 2.48)

1.00 (0.43, 2.29)

1.00 (0.43, 2.29)

1.04 (0.48, 2.13)

14–15

Ice hockey

Youth AA

131

2.30 (1.15, 4.58)

1.56 (0.76, 3.19)

0.82 (0.37, 1.79)

2.11 (0.54, 3.89)

14–15

Ice hockey

Youth BB

194

2.00 (1.12, 3.56)

1.39 (0.76, 2.53)

1.48 (0.81, 2.69)

1.36 (0.61, 2.22)

14–15

Ice hockey

Youth CC

301

1.29 (0.81, 2.05)

1.65 (1.05, 2.59)

0.98 (0.6, 1.58)

1.48 (0.67, 2.2)

14–15

Ice hockey

Novice recreation

67

0.87 (0.32, 2.34)

1.31 (0.51, 3.35)

1.00 (0.37, 2.63)

1.09 (0.43, 2.5)

Senior

Ice hockey

NHL professional

386

1.71 (1.14, 2.58)

1.42 (0.93, 2.15)

1.29 (0.85, 1.96)

1.36 (0.7, 1.93)

Senior

Ice hockey

Varsity

177

1.30 (0.72, 2.33)

1.17 (0.64, 2.12)

0.95 (0.51, 1.74)

1.26 (0.59, 2.11)

94

OR comparisons [Q1–4/1st and 2nd 6 mo] (95% CI) Q1 vs Q4 Q2 vs Q4 Q3 vs Q4

1st vs 2nd

3.09 (1.27, 7.51)

1.93 (0.48, 3.95)

2.54 (1.03, 6.27)

1.90 (0.75, 4.82)

Ice hockey

College AAA

150

2.20 (1.13, 4.27)

1.91 (0.98, 3.74)

1.12 (0.55, 2.29)

1.94 (0.56, 3.41)

16–19

Ice hockey

Junior elite developmental

171

3.47 (1.82, 6.62)

2.00 (1.01, 3.92)

1.66 (0.83, 3.31)

2.05 (0.58, 3.49)

14–15

Ice hockey

Youth AAA elite developmental

167

3.05 (1.57, 5.91)

2.60 (1.32, 5.08)

1.70 (0.84, 3.42)

2.09 (0.58, 3.58)

12–13

Volleyball

Junior

46

1.66 (0.45, 6.12)

3.00 (0.87, 10.3)

2.00 (0.55, 7.16)

1.55 (0.36, 4.26)

14–15

Volleyball

Youth cadet

31

0.77 (0.19, 3.16)

0.88 (0.22, 3.52)

0.77 (0.19, 3.16)

0.93 (0.29, 3.17)

16–17

Volleyball

Youth juvenile

24

0.83 (0.16, 4.29)

1.00 (0.2, 4.95)

1.16 (0.24, 5.61)

0.84 (0.24, 3.38)

14–15

Volleyball

Provincial youth cadet

211

1.21 (0.7, 2.08)

1.17 (0.67, 2.01)

1.10 (0.63, 1.91)

1.13 (0.62, 1.8)

16–17

Volleyball

Provincial youth juvenile

210

1.17 (0.67, 2.03)

0.93 (0.53, 1.64)

1.45 (0.85, 2.48)

0.85 (0.62, 1.37)

17–19

Volleyball

Provincial youth junior

64

1.14 (0.42, 3.09)

0.92 (0.33, 2.58)

1.50 (0.56, 3.95)

0.82 (0.42, 1.93)

Senior

Volleyball

Provincial senior

50

2.80 (0.77, 10.1)

3.20 (0.89, 11.4)

3.00 (0.83, 10.7)

1.50 (0.38, 3.94)

Continued next page

241

Sports Med 2009; 39 (3)

Senior

Relative Age Effects in Sport

ª 2009 Adis Data Information BV. All rights reserved.

Table I. Unadjusted odds ratios (ORs) for independent male subjects examining relative age effect in sport

Study

Barnsley et al.

Daniel and Janssen[41]

Barnsley and Thompson[3]

[1]

242

ª 2009 Adis Data Information BV. All rights reserved.

Table I. Contd Subject age (y)

Sport

Level of competition

16–20

Ice hockey

WHL amateur developmental

16–20

Ice hockey

OHL amateur developmental

Senior

Ice hockey

NHL professional

Senior

Basketball

NBA professional

297

1.11 (0.7, 1.75)

0.94 (0.59, 1.5)

1.18 (0.75, 1.86)

0.94 (0.67, 1.39)

Senior

Baseball

MLB professional

682

1.24 (0.92, 1.67)

1.11 (0.82, 1.5)

0.90 (0.66, 1.23)

1.23 (0.77, 1.60)

Senior

Football

CFL professional

342

1.00 (0.64, 1.54)

1.17 (0.76, 1.79)

1.32 (0.87, 2.02)

0.93 (0.69, 1.34)

Senior

Football

CFL professional

436

0.89 (0.61, 1.3)

0.96 (0.66, 1.4)

0.93 (0.64, 1.35)

0.96 (0.72, 1.33)

Senior

Football

AFC professional

777

1.25 (0.94, 1.66)

1.21 (0.91, 1.61)

1.18 (0.89, 1.57)

1.12 (0.78, 1.44)

Senior

Football

NFC professional

749

1.15 (0.86, 1.53)

0.98 (0.73, 1.31)

1.09 (0.82, 1.46)

1.01 (0.78, 1.3)

Senior

Ice hockey

NHL professional

103

1.10 (0.52, 2.33)

0.60 (0.26, 1.36)

0.96 (0.45, 2.06)

0.87 (0.51, 1.7)

Senior

Ice hockey

NHL professional

318

0.92 (0.59, 1.41)

0.87 (0.56, 1.35)

0.81 (0.52, 1.27)

0.98 (0.68, 1.44)

Senior

Ice hockey

NHL professional

320

1.05 (0.67, 1.63)

0.97 (0.62, 1.51)

1.12 (0.73, 1.74)

0.95 (0.68, 1.39)

Senior

Ice hockey

NHL professional

355

0.94 (0.62, 1.43)

1.03 (0.68, 1.56)

0.96 (0.63, 1.46)

1.00 (0.69, 1.44)

Senior

Ice hockey

NHL professional

775

2.14 (1.59, 2.88)

2.05 (1.52, 2.75)

1.37 (1.00, 1.87)

1.76 (0.78, 2.26)

Senior

Ice hockey

NHL professional

217

2.21 (1.29, 3.78)

1.31 (0.74, 2.31)

1.18 (0.66, 2.09)

1.61 (0.62, 2.57)

7–8

Ice hockey

Junior-minor-mite

1 676

1.12 (0.93, 1.36)

1.08 (0.89, 1.31)

0.98 (0.81, 1.19)

1.11 (0.84, 1.31)

9–10

Ice hockey

Junior-minor-mite

1 839

1.10 (0.91, 1.32)

1.17 (0.97, 1.4)

1.07 (0.89, 1.29)

1.09 (0.85, 1.28)

11–12

Ice hockey

Junior-minor-pee wee

1 536

1.13 (0.92, 1.38)

1.23 (1.01, 1.51)

1.05 (0.86, 1.29)

1.15 (0.84, 1.37)

13–14

Ice hockey

Junior-minor-bantam

1 112

1.37 (1.08, 1.75)

1.24 (0.97, 1.59)

0.89 (0.68, 1.15)

1.39 (0.8, 1.71)

15–16

Ice hockey

Youth-minor midget

815

1.18 (0.89, 1.56)

1.39 (1.05, 1.83)

1.19 (0.9, 1.58)

1.17 (0.78, 1.48)

17–18

Ice hockey

Youth-minor-juvenile

220

1.45 (0.85, 2.48)

1.52 (0.89, 2.59)

1.02 (0.58, 1.78)

1.47 (0.63, 2.33) 0.82 (0.53, 1.55)

Ice hockey

Youth representative

7–8

Ice hockey

Mite-lowest tier

9–10

Ice hockey

9–10 9–10 11–12

OR comparisons [Q1–4/1st and 2nd 6 mo] (95% CI) Q1 vs Q4 Q2 vs Q4 Q3 vs Q4

1st vs 2nd

698

4.56 (3.23, 6.42)

3.23 (2.27, 4.59)

2.10 (1.46, 3.03)

2.50 (0.76, 3.27)

350

3.76 (2.36, 5.98)

2.84 (1.76, 4.56)

1.60 (0.97, 2.65)

2.53 (0.68, 3.69)

715

1.97 (1.45, 2.67)

1.83 (1.35, 2.49)

1.35 (0.98, 1.85)

1.61 (0.77, 2.09)

115

0.90 (0.44, 1.85)

0.66 (0.31, 1.4)

0.90 (0.44, 1.85)

1 676

1.12 (0.93, 1.36)

1.08 (0.89, 1.31)

0.98 (0.81, 1.19)

1.11 (0.84, 1.31)

Mite-low tier

764

0.80 (0.6, 1.06)

0.88 (0.67, 1.17)

0.96 (0.72, 1.27)

0.86 (0.78, 1.1)

Ice hockey

Mite-mid tier

789

1.08 (0.81, 1.43)

1.18 (0.89, 1.56)

1.06 (0.8, 1.41)

1.09 (0.78, 1.39)

Ice hockey

Mite-upper tier

286

3.15 (1.88, 5.28)

2.90 (1.73, 4.88)

1.87 (1.09, 3.21)

2.10 (0.66, 3.18)

Ice hockey

Pee wee low tier

461

0.73 (0.5, 1.07)

1.06 (0.74, 1.52)

1.00 (0.7, 1.44)

0.89 (0.72, 1.23)

11–12

Ice hockey

Pee wee mid tier

746

1.02 (0.76–1.36)

1.21 (0.91, 1.61)

1.02 (0.76, 1.37)

1.10 (0.77, 1.41)

11–12

Ice hockey

Pee wee upper tier

329

2.40 (1.54, 3.75)

1.69 (1.06, 2.67)

1.23 (0.76, 1.98)

1.83 (0.68, 2.68)

Continued next page

Cobley et al.

Sports Med 2009; 39 (3)

19–20

No. of subjects

Study

Subject age (y)

Sport

Level of competition

No. of subjects

OR comparisons [Q1–4/1st and 2nd 6 mo] (95% CI) Q1 vs Q4 Q2 vs Q4 Q3 vs Q4

Barnsley and Thompson[3]

13–14

Ice hockey

Bantam low tier

586

0.81 (0.58, 1.13)

0.86 (0.62, 1.2)

1.14 (0.83, 1.56)

0.78 (0.75, 1.04)

13–14

Ice hockey

Bantam mid tier

206

1.47 (0.85, 2.55)

1.28 (0.73, 2.24)

1.14 (0.64, 2.01)

1.28 (0.62, 2.07)

13

Ice hockey

Junior AA minor

183

4.23 (2.16, 8.26)

3.23 (1.63, 6.39)

2.29 (1.13, 4.62)

2.26 (0.59, 3.8)

14

Ice hockey

Junior AA major

137

3.35 (1.63, 6.88)

2.58 (1.24, 5.38)

1.11 (0.49, 2.5)

2.80 (0.54, 5.15)

15–16

Ice hockey

Youth midget low

463

0.91 (0.63, 1.33)

1.18 (0.82, 1.69)

1.10 (0.77, 1.59)

0.99 (0.72, 1.36)

15–16

Ice hockey

Youth midget mid

227

1.24 (0.72, 2.13)

1.60 (0.94, 2.7)

1.20 (0.69, 2.05)

1.29 (0.63, 2.03)

15

Ice hockey

Youth AA major

125

2.81 (1.32, 5.98)

2.25 (1.04, 4.85)

1.75 (0.79, 3.85)

1.84 (0.53, 3.41)

17–18

Ice hockey

Youth mid tier

220

1.88 (1.06, 3.31)

1.08 (0.59, 1.98)

2.20 (1.26, 3.85)

0.92 (0.62, 1.48)

19–20

Ice hockey

Youth mid tier

115

1.21 (0.61, 2.38)

0.66 (0.32, 1.38)

0.90 (0.45, 1.83)

0.98 (0.54, 1.8) 1.75 (0.81, 2.15)

1st vs 2nd

Boucher and Halliwell[57]

Senior

Ice hockey

NHL professional

1 116

2.15 (1.68, 2.74)

1.86 (1.45, 2.38)

1.28 (0.99, 1.66)

8–17

Ice hockey

Junior regional representative

1 085

2.53 (1.96, 3.26)

2.18 (1.68, 2.83)

1.76 (1.35, 2.29)

1.70 (0.81, 2.1)

Grondin and Trudeau[58]

Senior

Ice hockey

NHL professional

388

1.55 (1.02, 2.34)

1.82 (1.21, 2.74)

1.24 (0.81, 1.9)

1.50 (0.7, 2.12)

Senior

Ice hockey

NHL professional

79

1.62 (0.67, 3.92)

1.00 (0.39, 2.54)

1.31 (0.53, 3.23)

1.13 (0.46, 2.43)

Senior

Ice hockey

NHL professional

54

1.45 (0.49, 4.26)

1.54 (0.53, 4.5)

0.90 (0.29, 2.84)

1.57 (0.39, 3.99)

Senior

Baseball

MLB professional

682

1.36 (1.00, 1.84)

1.30 (0.95, 1.76)

1.10 (0.81, 1.5)

1.26 (0.77, 1.64)

Senior

Baseball

MLB professional

837

1.29 (0.99, 1.7)

1.12 (0.85, 1.47)

1.03 (0.78, 1.36)

1.19 (0.79, 1.5)

Thompson et al.[7] Barnsley et al.[11]

Senior

Soccer

Professional national

528

1.50 (1.05, 2.12)

1.38 (0.96, 1.96)

1.40 (0.98, 1.99)

1.20 (0.74, 1.61)

16–17

Soccer

Junior elite national

287

5.86 (3.35, 10.2)

4.40 (2.5, 7.77)

1.77 (0.95, 3.28)

3.70 (0.64, 5.7)

288

19–20

Soccer

Developmental elite national

Brewer et al.[27]

16–17

Soccer

National junior

59

Glamser and Marciani[59]

18–24

Football

College

59

18–24

Baseball

College

26

18–24

Football

College

49

Thompson et al.[8]

6.13 (3.51, 10.7)

4.27 (2.42, 7.53)

1.68 (0.90, 3.12)

3.88 (0.64, 5.99)

12.0 (1.37, 104)

12.0 (1.37, 104)

3.53 (0.38, 9.13)

4.75 (1.29, 17.3)

5.00 (1.37, 18.2)

4.00 (1.07, 14.8)

1.95 (0.4, 4.8)

1.75 (0.33, 9.02)

2.75 (0.56, 13.3)

1.00 (0.17, 5.82)

2.25 (0.25, 8.85)

5.00 (0.9, 27.7)

6.50 (1.20, 35.0)

2.26 (0.36, 6.15) 1.22 (0.21, 5.59)

34.0 (4.09, 281)

12.0 (2.31, 62.2)

Baseball

College

20

1.50 (0.25, 8.81)

1.25 (0.20, 7.61)

1.25 (0.20, 7.61)

4–6

Baseball

T-ball beginners

335

1.47 (0.96, 2.27)

1.14 (0.73, 1.78)

1.23 (0.79, 1.91)

1.17 (0.68, 1.70)

7–9

Baseball

Junior

894

0.99 (0.76, 1.29)

0.93 (0.72, 1.22)

0.90 (0.69, 1.17)

1.01 (0.79, 1.27)

10–12

Baseball

Junior

1 235

1.13 (0.90, 1.41)

0.97 (0.78, 1.22)

0.95 (0.76, 1.19)

1.07 (0.82, 1.30)

13–15

Baseball

Junior league

823

0.97 (0.74, 1.27)

0.81 (0.62, 1.07)

0.96 (0.73, 1.26)

0.91 (0.78, 1.15)

16–18

Baseball

Youth league

127

1.40 (0.70, 2.82)

1.37 (0.68, 2.75)

0.92 (0.44, 1.92)

1.44 (0.54, 2.64)

10

Baseball

Minor

321

1.00 (0.64, 1.53)

0.78 (0.50, 1.23)

0.98 (0.64, 1.52)

0.89 (0.68, 1.31)

10

Baseball

Major

35

1.50 (0.41, 5.47)

1.00 (0.25, 3.88)

0.87 (0.21, 3.48)

1.33 (0.31, 4.21)

Continued next page

243

Sports Med 2009; 39 (3)

18–24

Relative Age Effects in Sport

ª 2009 Adis Data Information BV. All rights reserved.

Table I. Contd

244

ª 2009 Adis Data Information BV. All rights reserved.

Table I. Contd Study

Subject age (y)

Sport

Level of competition

Thompson et al.[8]

11

Baseball

Minor

11

Baseball

12

Baseball

12

Verhulst

[10]

Boucher and Mutimer[9]

Dudink[60]

OR comparisons [Q1–4/1st and 2nd 6 mo] (95% CI) Q1 vs Q4 Q2 vs Q4 Q3 vs Q4

1st vs 2nd

105

0.93 (0.44, 1.96)

0.61 (0.27, 1.34)

0.83 (0.39, 1.77)

0.84 (0.51, 1.63)

Major

342

1.62 (1.05, 2.48)

1.40 (0.91, 2.17)

1.15 (0.73, 1.79)

1.40 (0.69, 2.03)

Minor

49

0.40 (0.12, 1.25)

0.30 (0.08, 1.00)

0.75 (0.26, 2.11)

0.40 (0.36, 1.09)

Baseball

Major

383

1.09 (0.73, 1.63)

1.11 (0.75, 1.66)

0.86 (0.57, 1.29)

1.18 (0.70, 1.68)

13

Baseball

Senior

145

0.88 (0.45, 1.72)

0.88 (0.45, 1.72)

1.37 (0.72, 2.58)

0.74 (0.56, 1.31)

13

Baseball

Junior

239

1.26 (0.77, 2.06)

0.75 (0.44, 1.27)

0.90 (0.54, 1.50)

1.06 (0.64, 1.64)

14

Baseball

Senior

207

0.92 (0.54, 1.59)

0.89 (0.52, 1.53)

0.80 (0.46, 1.39)

1.00 (0.62, 1.61)

14

Baseball

Junior

10–14

Baseball

Lower level junior

10–14

Baseball

High level junior

10–14

Baseball

Tournament juniors

10–14

Baseball

Senior Senior

23

0.44 (0.08, 2.31)

0.33 (0.05, 1.90)

0.77 (0.17, 3.55)

0.43 (0.23, 1.87)

827

0.90 (0.69, 1.18)

0.76 (0.57, 1.00)

0.96 (0.73, 1.25)

0.85 (0.78, 1.07)

1 022

1.27 (0.99, 1.62)

1.07 (0.83, 1.37)

0.94 (0.73, 1.22)

1.20 (0.80, 1.48)

410

1.83 (1.22, 2.74)

1.82 (1.21, 2.72)

1.36 (0.90, 2.07)

1.54 (0.71, 2.16)

Recreation juniors

951

0.99 (0.77, 1.27)

0.81 (0.63, 1.05)

0.81 (0.63, 1.05)

0.99 (0.80, 1.24)

Soccer

Professional div 1

369

1.11 (0.74, 1.68)

1.28 (0.85, 1.92)

0.98 (0.65, 1.50)

1.20 (0.70, 1.72)

Soccer

Professional div 2

342

1.47 (0.96, 2.25)

1.22 (0.79, 1.89)

1.18 (0.76, 1.83)

1.23 (0.69, 1.78)

Senior

Soccer

Professional div 1

411

2.19 (1.47, 3.26)

1.52 (1.00, 2.29)

1.43 (0.94, 2.16)

1.52 (0.71, 2.13)

Senior

Soccer

Professional div 2

768

1.90 (1.43, 2.53)

1.22 (0.90, 1.65)

1.39 (1.03, 1.87)

1.30 (0.78, 1.67)

Senior

Soccer

Professional div 2

399

1.77 (1.19, 2.65)

1.55 (1.03, 2.33)

1.22 (0.8, 1.85)

1.50 (0.71, 2.11)

401

1.76 (1.18, 2.61)

1.45 (0.97, 2.17)

1.12 (0.73, 1.69)

1.51 (0.71, 2.13)

68

1.83 (0.69, 4.85)

1.50 (0.55, 4.04)

1.33 (0.48, 3.64)

1.42 (0.43, 3.26)

Senior

Soccer

Professional div 2

8–9

Ice hockey

Junior novice

10–11

Ice hockey

Atom

213

2.75 (1.56, 4.87)

1.96 (1.09, 3.53)

1.62 (0.89, 2.94)

1.80 (0.62, 2.89)

12–13

Ice hockey

Pee wee

224

3.34 (1.84, 6.07)

3.13 (1.72, 5.69)

2.26 (1.22, 4.18)

1.98 (0.62, 3.15)

14–15

Ice hockey

Bantam

302

3.10 (1.9, 5.06)

2.18 (1.32, 3.62)

1.86 (1.11, 3.10)

1.84 (0.67, 2.75)

16–17

Ice hockey

Midget

144

3.60 (1.72, 7.51)

2.66 (1.25, 5.65)

2.33 (1.09, 4.99)

1.88 (0.56, 3.34)

Senior

Ice hockey

NHL professional

884

2.28 (1.72, 3.01)

2.09 (1.58, 2.77)

1.36 (1.01, 1.83)

1.85 (0.79, 2.33)

Senior

Soccer

Professional premier

761

2.11 (1.59, 2.81)

1.39 (1.03, 1.88)

1.08 (0.79, 1.47)

1.68 (0.77, 2.16)

Senior

Soccer

Professional div 1

734

1.79 (1.34, 2.39)

1.14 (0.85, 1.55)

1.04 (0.77, 1.42)

1.43 (0.77, 1.85)

Senior

Soccer

Professional div 2

673

1.91 (1.41, 2.58)

1.28 (0.93, 1.75)

0.93 (0.67, 1.30)

1.64 (0.76, 2.14)

Senior

Soccer

Professional div 3

609

2.12 (1.53, 2.94)

1.65 (1.18, 2.31)

1.18 (0.83, 1.67)

1.73 (0.75, 2.28)

Senior

Cricket

Professional county

292

NR

NR

NR

1.16 (0.67, 1.73) Continued next page

Cobley et al.

Sports Med 2009; 39 (3)

Edwards[16]

No. of subjects

Study

Subject age (y)

Sport

Level of competition

No. of subjects

OR comparisons [Q1–4/1st and 2nd 6 mo] (95% CI) Q1 vs Q4 Q2 vs Q4 Q3 vs Q4

Baxter-Jones[14]

11–17

Soccer

Elite junior

65

7.40 (2.32, 23.5)

3.20 (0.94, 10.8)

1.40 (0.36, 5.33)

4.41 (0.39, 11.1)

11–17

Swimming

Elite junior

54

2.85 (0.90, 8.97)

2.71 (0.86, 8.56)

1.14 (0.32, 4.04)

2.60 (0.38, 6.79) 2.52 (0.44, 5.72)

1st vs 2nd

9–17

Tennis

Elite junior

74

3.54 (1.40, 8.97)

1.27 (0.45, 3.52)

0.90 (0.31, 2.65)

Baxter-Jones et al.[25]

9–18

Gymnastics

Elite junior

38

1.25 (0.34, 4.55)

1.12 (0.30, 4.16)

1.37 (0.38, 4.94)

1.00 (0.33, 3.00)

Brewer et al.[61]

16–17

Soccer

Youth elite development

59

34 (4.09, 281)

12 (1.37, 104)

12 (1.37, 104)

3.53 (0.38, 9.13)

Senior

Soccer

Professional national

16

17 (0.71, 405)a

11 (0.44, 272)a

7 (0.27, 184)a

Stanaway and Hines[62]

Senior

Baseball

MLB professional

600

1.22 (0.88, 1.68)

1.25 (0.90, 1.72)

0.97 (0.69, 1.34)

Senior

Football

Hall of fame

167

1.20 (0.64, 2.25)

1.58 (0.86, 2.91)

1.11 (0.59, 2.1)

1.31 (0.59, 2.23)

Helsen et al.[30]

Senior

Soccer

Professional

408

1.78 (1.2, 2.64)

1.44 (0.96, 2.16)

1.13 (0.74, 1.71)

1.51 (0.71, 2.13)

10–16

Soccer

Junior elite national

369

4.62 (2.92, 7.30)

2.37 (1.47, 3.84)

1.97 (1.20, 3.21)

2.35 (0.69, 3.39)

6–16

Soccer

Youth elite club

485

2.62 (1.79, 3.82)

1.95 (1.32, 2.88)

1.77 (1.19, 2.62)

1.65 (0.73, 2.25)

6–10

Soccer

Youth leagues

270

1.84 (1.13, 2.97)

1.46 (0.89, 2.39)

1.10 (0.66, 1.83)

1.57 (0.65, 2.38)

12–16

Soccer

Junior leagues

226

Senior

Ice hockey

NHL professional

16–18

Ice hockey

Junior national

19–26

Ice hockey

College representative

Senior

Ice hockey

8–16

Montelpare et al.[51]

Musch and Hay[63]

Helsen et al.[64]

Relative Age Effects in Sport

ª 2009 Adis Data Information BV. All rights reserved.

Table I. Contd

4.33 (0.15, 28.1) 1.25 (0.75, 1.65)

1.16 (0.68, 1.96)

1.08 (0.63, 1.84)

1.28 (0.75, 2.15)

0.98 (0.63, 1.54)

1 090

NA

NA

NA

1.78 (0.81, 2.19) 1.85 (0.64, 2.92)

NA

NA

NA

NA

NA

NA

1.44 (0.86, 1.67)

Amateur representative

476

NA

NA

NA

1.78 (0.73, 2.45)

Ice hockey

Junior and youth representative

474

NA

NA

NA

1.49 (0.73, 2.05)

8–16

Ice hockey

Junior minor

974

NA

NA

NA

1.17 (0.80, 1.46)

Senior

Soccer

Professional

207

1.58 (0.91, 2.77)

1.38 (0.78, 2.43)

1.33 (0.75, 2.34)

1.27 (0.62, 2.04)

Senior

Soccer

Professional

61

1.76 (0.66, 4.72)

1.00 (0.35, 2.84)

0.92 (0.32, 2.65)

1.44 (0.41, 3.45)

Senior

Soccer

Professional

486

1.70 (1.18, 2.45)

1.49 (1.03, 2.16)

1.39 (0.95, 2.01)

1.33 (0.73, 1.82)

Senior

Soccer

Professional

355

1.31 (0.87, 1.98)

1.06 (0.69, 1.61)

0.95 (0.62, 1.45)

1.21 (0.69, 1.74)

Senior

Soccer

Professional

360

2.20 (1.44, 3.35)

1.78 (1.15, 2.74)

1.01 (0.64, 1.61)

1.97 (0.69, 2.84)

10–12

Soccer

Junior representative

410

2.15 (1.45, 3.19)

1.57 (1.04, 2.35)

1.12 (0.73, 1.72)

1.75 (0.71, 2.46)

10–12

Soccer

Junior representative

507

1.78 (1.25, 2.54)

1.13 (0.78, 1.65)

1.52 (1.06, 2.18)

1.15 (0.73, 1.56)

12–14

Soccer

Junior representative

452

2.15 (1.45, 3.19)

1.95 (1.31, 2.90)

1.84 (1.23, 2.75)

1.44 (0.72, 1.98)

12–14

Soccer

Junior representative

520

1.87 (1.33, 2.63)

1.02 (0.71, 1.48)

1.14 (0.79, 1.64)

1.35 (0.74, 1.82)

14–16

Soccer

Youth representative

449

3.58 (2.41, 5.32)

1.98 (1.3, 2.99)

1.44 (0.94, 2.22)

2.27 (0.71, 3.16)

Continued next page

245

Sports Med 2009; 39 (3)

231 2 047

246

ª 2009 Adis Data Information BV. All rights reserved.

Table I. Contd Study

Subject age (y)

Sport

Level of competition

Helsen et al.[64]

14–16

Soccer

Youth representative

16–18

Soccer

16–18 15–16

Hoare[65]

Grondin and Koren[66]

Simmons and Paull[67]

No. of subjects

OR comparisons [Q1–4/1st and 2nd 6 mo] (95% CI) Q1 vs Q4 Q2 vs Q4 Q3 vs Q4

1st vs 2nd

385

1.35 (0.95, 1.92)

1.03 (0.72, 1.49)

1.17 (0.82, 1.68)

1.09 (0.73, 1.49)

Youth representative

458

2.24 (1.55, 3.23)

1.26 (0.86, 1.87)

1.07 (0.72, 1.59)

1.69 (0.72, 2.33)

Soccer

Youth representative

501

0.68 (0.47, 0.97)

0.66 (0.46, 0.94)

0.97 (0.69, 1.36)

0.68 (0.73, 0.92)

Basketball

Junior regional representative

130

9.14 (3.64, 22.9)

4.85 (1.88, 12.5)

3.57 (1.35, 9.41)

3.06 (0.53, 5.74)

15–16

Basketball

Junior regional representative

113

5.22 (2.15, 12.6)

3.44 (1.39, 8.53)

2.88 (1.15, 7.24)

2.22 (0.51, 4.29)

17–18

Basketball

Junior regional representative

118

3.25 (1.52, 6.93)

1.75 (0.78, 3.88)

1.37 (0.60, 3.12)

2.10 (0.52, 3.99)

Senior

Basketball

Professional

89

2.42 (1.03, 5.71)

1.78 (0.74, 4.30)

1.14 (0.45, 2.88)

1.96 (0.48, 4.09)

Senior

Baseball

MLB professional

5 033

1.12 (1.00, 1.25)

1.02 (0.91, 1.14)

0.97 (0.86, 1.08)

1.09 (0.90, 1.20)

Senior

Baseball

MLB professional

1 123

1.25 (0.99, 1.58)

1.16 (0.92, 1.47)

0.99 (0.78, 1.26)

1.21 (0.81, 1.48)

Senior

Baseball

MLB professional

1 260

1.17 (0.94, 1.46)

1.08 (0.87, 1.35)

0.95 (0.76, 1.19)

1.15 (0.82, 1.40)

Senior

Baseball

MLB professional

1 032

1.04 (0.82, 1.33)

1.01 (0.79, 1.28)

0.89 (0.70, 1.14)

1.08 (0.80, 1.34)

Senior

Baseball

MLB professional

1 000

0.98 (0.76, 1.25)

0.98 (0.77, 1.26)

0.88 (0.69, 1.13)

1.04 (0.80, 1.29)

Senior

Baseball

MLB professional

1 303

1.21 (0.97, 1.5)

1.08 (0.87, 1.35)

1.01 (0.81, 1.26)

1.13 (0.82, 1.37)

Senior

Baseball

MLB professional

1 514

1.31 (1.07, 1.61)

1.19 (0.97, 1.46)

1.02 (0.83, 1.26)

1.23 (0.83, 1.47)

Senior

Baseball

MLB professional

1 405

1.48 (1.19, 1.83)

1.37 (1.10, 1.69)

1.25 (1.01, 1.56)

1.26 (0.83, 1.51)

Senior

Baseball

Professional

744

2.39 (1.77, 3.23)

1.82 (1.34, 2.48)

1.47 (1.07, 2.02)

1.70 (0.77, 2.19)

15–16

Soccer

Youth national developmental

9–16

Soccer

Junior/youth representative

14–15

Soccer

15–16

Soccer

7–8 9–10

12.2 (3.70, 40.4)

4.50 (1.28, 15.7)

2.00 (0.51, 7.73)

5.58 (0.41, 13.4)

5.04 (4.59, 5.54)

2.84 (2.58, 3.13)

1.09 (0.98, 1.22)

3.76 (0.92, 4.06)

Junior national

78

16.6 (4.43, 62.6)

5.33 (1.33, 21.2)

3.00 (0.70, 12.7)

5.50 (0.41, 13.2)

Junior national

63

2.23 (0.85, 5.80)

0.61 (0.19, 1.89)

1.00 (0.35, 2.82)

1.42 (0.42, 3.36)

Soccer

Junior leagues

4 795

0.97 (0.87, 1.09)

0.90 (0.81, 1.01)

0.86 (0.77, 0.97)

1.00 (0.90, 1.11)

Soccer

Junior leagues

5 332

1.12 (1.01, 1.25)

0.96 (0.86, 1.07)

1.04 (0.93, 1.16)

1.02 (0.91, 1.12)

11–12

Soccer

Junior leagues

5 417

1.14 (1.03, 1.27)

1.01 (0.91, 1.13)

1.00 (0.89, 1.11)

1.08 (0.91, 1.18)

13–14

Soccer

Junior leagues

4 478

1.23 (1.09, 1.38)

1.13 (1.00, 1.27)

0.96 (0.85, 1.08)

1.20 (0.90, 1.33)

15–16

Soccer

Junior leagues

3 266

1.24 (1.08, 1.42)

1.17 (1.02, 1.34)

1.02 (0.88, 1.17)

1.19 (0.88, 1.34)

17–18

Soccer

Junior leagues

2 033

1.24 (1.04, 1.48)

1.12 (0.94, 1.34)

1.05 (0.88, 1.25)

1.15 (0.86, 1.34)

Glamser and Vincent[69]

17–18

Soccer

Youth regional representative

147

3.00 (1.48, 6.05)

2.66 (1.31, 5.41)

1.50 (0.70, 3.18)

2.26 (0.56, 4.03)

O’Donoghue et al.[17]

Senior

Cricket

Professional

120

1.37 (0.67, 2.78)

1.33 (0.65, 2.71)

0.74 (0.34, 1.59)

1.55 (0.53, 2.90)

Senior

Cricket

Professional

75

1.00 (0.41, 2.43)

0.90 (0.36, 2.22)

0.85 (0.34, 2.11)

1.02 (0.45, 2.24)

Senior

Netball

Professional

128

1.58 (0.80, 3.11)

0.86 (0.41, 1.78)

0.96 (0.47, 1.97)

1.24 (0.54, 2.27)

Senior

Netball

Professional

119

0.90 (0.43, 1.85)

0.80 (0.38, 1.67)

1.12 (0.55, 2.27)

0.80 (0.53, 1.49)

Musch[68]

Sports Med 2009; 39 (3)

Continued next page

Cobley et al.

79 8 857

Study

Subject age (y)

Sport

Level of competition

Edgar and O’Donoghue[70]

Senior

Soccer

Professional

Senior

Soccer

Senior Senior

Abernethy and Farrow[15]b

No. of subjects

OR comparisons [Q1–4/1st and 2nd 6 mo] (95% CI) Q1 vs Q4 Q2 vs Q4 Q3 vs Q4

1st vs 2nd

345

1.40 (0.92, 2.14)

1.01 (0.65, 1.57)

1.24 (0.81, 1.90)

1.07 (0.69, 1.55)

Professional

92

1.13 (0.50, 2.56)

0.90 (0.39, 2.09)

1.13 (0.50, 2.56)

0.95 (0.49, 1.94)

Soccer

Professional

69

1.21 (0.49, 2.98)

0.73 (0.28, 1.92)

0.68 (0.25, 1.80)

1.15 (0.44, 2.62)

Soccer

Professional

92

0.85 (0.36, 2.01)

1.23 (0.54, 2.79)

1.28 (0.57, 2.89)

0.91 (0.49, 1.86)

Senior

Soccer

Professional

598

1.25 (0.90, 1.72)

0.99 (0.71, 1.37)

1.15 (0.83, 1.59)

1.04 (0.75, 1.37)

Senior

Soccer

Professional

115

1.03 (0.50, 2.11)

0.45 (0.19, 1.02)

1.22 (0.60, 2.47)

0.66 (0.52, 1.26)

Senior

Soccer

Professional

23

0.12 (0.01, 1.34)

0.87 (0.18, 4.07)

0.87 (0.18, 4.07)

0.53 (0.23, 2.25)

Senior

Soccer

Professional

138

0.84 (0.43, 1.63)

0.53 (0.26, 1.09)

1.15 (0.60, 2.18)

0.64 (0.55, 1.15)

Senior

Soccer

Professional

736

1.16 (0.87, 1.54)

0.89 (0.66, 1.19)

1.15 (0.86, 1.53)

0.95 (0.77, 1.22)

Senior

Australian Football

Professional

627

1.44 (1.05, 1.96)

1.26 (0.92, 1.74)

0.97 (0.69, 1.34)

1.37 (0.76, 1.80)

Senior

Rugby union

Professional

74

2.16 (0.84, 5.54)

1.66 (0.63, 4.36)

1.33 (0.49, 3.58)

1.64 (0.45, 3.64)

Senior

Rugby union

National professional

37

1.42 (0.37, 5.39)

1.71 (0.46, 6.31)

1.14 (0.29, 4.46)

1.46 (0.32, 4.50)

Senior

Rugby league

Professional

418

2.39 (1.60, 3.56)

1.90 (1.26, 2.86)

1.23 (0.80, 1.89)

1.92 (0.71, 2.69)

Senior

Rugby league

National professional

92

2.46 (1.07, 5.67)

1.53 (0.64, 3.66)

1.13 (0.45, 2.79)

1.87 (0.48, 3.85)

Senior

Cricket

Professional

151

1.13 (0.60, 2.11)

1.00 (0.52, 1.89)

0.84 (0.43, 1.61)

1.15 (0.57, 2.01)

Senior

Cricket

National professional

385

2.33 (0.63, 8.60)

1.83 (0.48, 6.95)

1.33 (0.33, 5.30)

1.78 (0.33, 5.37)

Senior

Basketball

Professional

94

1.94 (0.86, 4.35)

0.94 (0.39, 2.26)

1.33 (0.57, 3.07)

1.23 (0.49, 2.49)

Senior

Basketball

National professional

18

2.66 (0.41, 17.1)

1.33 (0.18, 9.72)

1.00 (0.12, 7.89)

2.00 (0.19, 10.2)

Senior

Tennis

ATP professional

237

1.65 (0.97, 2.81)

1.68 (0.99, 2.85)

1.46 (0.85, 2.50)

1.35 (0.64, 2.10)

14–18

Tennis

ITF national juniors

237

2.19 (1.28, 3.74)

1.97 (1.15, 3.38)

1.41 (0.81, 2.47)

1.72 (0.63, 2.69)

Helsen et al.[12]

14–18

Soccer

Youth national

99

1.68 (0.79, 3.54)

1.45 (0.68, 3.09)

0.90 (0.40, 2.02)

1.64 (0.52, 3.15)

14–18

Soccer

Youth national

90

4.12 (1.56, 10.8)

3.62 (1.36, 9.62)

2.50 (0.91, 6.84)

2.21 (0.47, 4.61)

14–18

Soccer

Youth national

94

2.93 (1.31, 6.57)

0.81 (0.32, 2.05)

1.12 (0.46, 2.72)

1.76 (0.49, 3.58)

14–18

Soccer

Youth national

41

3.00 (0.84, 10.6)

2.16 (0.59, 7.93)

0.66 (0.14, 3.08)

3.10 (0.32, 9.51)

14–18

Soccer

Youth national

77

12.0 (3.15, 45.6)

6.0 (1.51, 23.7)

6.66 (1.69, 26.1)

2.34 (0.45, 5.21)

14–18

Soccer

Youth national

50

3.60 (1.01, 12.7)

4.4 (1.26, 15.3)

1.00 (0.23, 4.33)

4.00 (0.35, 11.3)

14–18

Soccer

Youth national

36

17.0 (1.84, 156)

11.0 (1.16, 103)

7.00 (0.70, 69.1)

3.50 (0.29, 11.7)

14–18

Soccer

Youth national

101

3.07 (1.33, 7.08)

1.61 (0.66, 3.91)

2.07 (0.87, 4.91)

1.52 (0.50, 3.01)

14–18

Soccer

Youth national

72

6.6 (2.09, 20.7)

5.00 (1.56, 15.9)

1.80 (0.50, 6.43)

4.14 (0.41, 9.94)

15–16

Soccer

Youth national

288

6.4 (3.67, 11.1)

3.22 (1.80, 5.75)

2.45 (1.35, 4.44)

2.78 (0.65, 4.24)

Continued next page

247

Sports Med 2009; 39 (3)

Edgar and O’Donoghue[13]

Relative Age Effects in Sport

ª 2009 Adis Data Information BV. All rights reserved.

Table I. Contd

248

ª 2009 Adis Data Information BV. All rights reserved.

Table I. Contd Study

Subject age (y)

Sport

Level of competition

Helsen et al.[12]

17–18

Soccer

Youth national

19–21

Soccer

11–14

OR comparisons [Q1–4/1st and 2nd 6 mo] (95% CI) Q1 vs Q4 Q2 vs Q4 Q3 vs Q4

1st vs 2nd

144

1.65 (0.84, 3.23)

1.69 (0.86, 3.30)

1.19 (0.59, 2.39)

1.52 (0.56, 2.69)

Youth national

159

1.07 (0.58, 1.97)

0.85 (0.45, 1.60)

0.95 (0.51, 1.76)

0.98 (0.58, 1.69)

Soccer

Junior club tournament

677

2.01 (1.47, 2.76)

1.71 (1.24, 2.35)

1.47 (1.06, 2.04)

1.50 (0.76, 1.96)

Senior

Soccer

Professional

1 930

1.44 (1.20, 1.73)

1.40 (1.17, 1.68)

1.20 (0.99, 1.44)

1.29 (0.85, 1.51)

Senior

Soccer

Professional

827

1.45 (1.10, 1.91)

1.16 (0.87, 1.54)

1.21 (0.91, 1.60)

1.18 (0.78, 1.49)

Vincent and Glamser[72]

16–17

Soccer

Developmental national

24

3.25 (0.66, 15.9)

0.50 (0.06, 3.84)

1.25 (0.22, 7.08)

1.66 (0.24, 6.76)

Esteva and Drobnic[73]

Youth

Basketball

Youth representative

157

9.75 (4.16, 22.8)

6.62 (2.78, 15.7)

2.25 (0.87, 5.77)

5.03 (0.54, 9.27)

Senior

Basketball

Professional

404

1.42 (0.95, 2.12)

1.62 (1.09, 2.41)

1.12 (0.74, 1.70)

1.43 (0.71, 2.01)

Senior

Basketball

NBA Professional

382

0.98 (0.66, 1.47)

0.93 (0.62, 1.4)

0.92 (0.62, 1.38)

1.00 (0.70, 1.41)

Senior

Ice hockey

NHL professional

151

1.79 (0.94, 3.40)

1.13 (0.58, 2.22)

1.27 (0.65, 2.47)

1.28 (0.57, 2.24)

Senior

Basketball

NBA professional

436

1.21 (0.83, 1.77)

1.05 (0.71, 1.54)

1.22 (0.84, 1.78)

1.01 (0.72, 1.40)

Senior

Baseball

MLB professional

907

1.38 (1.06, 1.78)

1.15 (0.88, 1.50)

1.00 (0.76, 1.30)

1.26 (0.79, 1.58)

Senior

Golf

PGA professional

197

1.12 (0.64, 1.94)

0.91 (0.52, 1.61)

0.97 (0.55, 1.71)

1.03 (0.61, 1.67)

Senior

Ice hockey

NHL professional

549

1.54 (1.09, 2.16)

1.56 (1.11, 2.19)

1.12 (0.78, 1.60)

1.46 (0.74, 1.95)

Senior

Ice hockey

NHL professional

146

1.07 (0.57, 2.03)

0.97 (0.51, 1.85)

0.78 (0.4, 1.53)

1.14 (0.56, 2.01)

Senior

Ice hockey

NHL professional

206

1.32 (0.76, 2.28)

1.13 (0.64, 1.96)

1.02 (0.58, 1.79)

1.21 (0.62, 1.95)

Senior

Ice hockey

NHL professional

252

1.14 (0.69, 1.90)

1.20 (0.72, 1.98)

1.31 (0.79, 2.16)

1.01 (0.65, 1.55)

Senior

Ice hockey

NHL professional

282

1.23 (0.77, 1.97)

1.17 (0.73, 1.88)

1.06 (0.66, 1.71)

1.16 (0.66, 1.75)

Senior

Ice hockey

NHL professional

284

0.81 (0.51, 1.28)

0.77 (0.49, 1.23)

0.70 (0.44, 1.12)

0.93 (0.66, 1.39)

Senior

Ice hockey

NHL professional

423

0.76 (0.51, 1.11)

0.87 (0.59, 1.27)

0.98 (0.67, 1.42)

0.82 (0.71, 1.14)

Senior

Ice hockey

NHL professional

698

1.50 (1.11, 2.03)

1.34 (0.99, 1.82)

1.20 (0.88, 1.63)

1.29 (0.77, 1.67)

Senior

Ice hockey

NHL professional

798

1.90 (1.43, 2.54)

1.79 (1.34, 2.40)

1.20 (0.88, 1.62)

1.68 (0.78, 2.14)

Senior

Ice hockey

NHL professional

600

1.55 (1.12, 2.15)

1.47 (1.06, 2.04)

1.05 (0.74, 1.47)

1.47 (0.75, 1.95)

Senior

Ice hockey

NHL professional

76

0.76 (0.30, 1.89)

1.33 (0.56, 3.12)

0.52 (0.19, 1.37)

1.37 (0.45, 3.00)

121 159d

1.65 (1.54, 1.77)

NA

NA

1.39 (1.32, 1.47)

Vaeyens et al.[71]

Coˆte´ et al.[18]

[42]c

Wattie et al.

Summary effect size

No. of subjects

0.5 added to raw data as Q4 = 0, preventing OR calculation. Procedure recommended by Sutton et al.[48]

b

Authors also reported data for soccer at junior national (16–17 years old), developmental national (21–23 years old) and national professional (senior) levels. However, sample totals were not available for OR calculation.

c Figure excludes total sample numbers from Edwards[16] and Montelpare et al.[51] d

At the time of data collection this paper was in press and accepted for publication, but not published until June 2007.

AA, AAA, BB, CC = levels of ice hockey competition, where ‘As’ are more competitive or a higher level than subsequent letters of the alphabet; AFC = American Football Conference; ATP = Association of Tennis Professionals; CFL = Canadian Football League; div = division; MLB = Major League Baseball; NBA = National Basketball Association; NFC = National Football Conference; NR = sample information not reported; NA = quartile data not available following contact with lead author; original data presented in bi-monthly distributions; NHL = National Hockey League; OHL = Ontario Hockey League; PGA = Professional Golfers’ Association; Q = quartile; WHL = Western Hockey League.

Cobley et al.

Sports Med 2009; 39 (3)

a

Relative Age Effects in Sport

information hopefully beneficial to strategies motivated toward eradicating RAEs. Findings suggest RAEs are robust and generally prevalent across the sports contexts examined to date. Across all samples, summary ORs indicate that for every two participants born in the last quartile of an annual age-group, over three are participating from the first quartile of the same age-group. Risk likelihoods increased when the number of months away from the referent group (i.e. quartile 4) was amplified, suggesting a linear profile to RAEs. Findings identified that the relatively youngest sport participants within annual age-groups were (i) less likely to participate in recreational and competitive sport from under 14 years of age; (ii) certainly less likely to participate on representative teams during the 15- to 18-year-old bracket; and (iii) less likely to become an elite athlete in the sport contexts examined. In combining previous literature (e.g. Helsen et al.[30]) with findings from the present study, it seems that sport is less likely to be an activity or career pathway for relatively younger individuals, whose birth dates coincide with the last 3 months of an annual age-grouping strategy. Several factors influenced the magnitude of RAEs, notably age category, skill level and sport context. Analyses identified sport contexts with distinctive RAE risks; higher risks were associated with basketball, soccer and ice hockey. In these sports, mid to late adolescence (15–18 years) and the representative level of competition (i.e. regional and national representation) were most vulnerable to RAEs. In contrast, while small significant effects remained, childhood (under 11 years) and recreational sports contexts reported the lowest risk of RAEs. In the seven available American Football samples, no evidence of RAEs was found. Related to these contexts, low numbers of samples were available, so findings should be evaluated with caution. Taken together, findings partially reinforced our hypothesis that RAEs are probably most likely to occur in highly popular sports, prevailing due to a combination of mechanisms primarily associated with maturation and selection of athletes within the developmental tiers and ª 2009 Adis Data Information BV. All rights reserved.

249

structures of a sport.[3,19,30] Contrary to our hypotheses, RAE risk did not increase linearly with skill level or age category. Rather, at the elite level (professional or senior national representative) risks decreased to below that of the youth representative. At senior ages (i.e. >18 years) RAE risk also decreased to below that of the adolescent ages. Nevertheless, RAEs persisted into older cohorts. The reduction of RAEs at the senior and elite stages is difficult to explain, with several mechanisms possible. For example, whilst acknowledging that annual age-groupings within sport generally terminate in senior sport (i.e. often 19–21 years old), it could be that differences according to physical maturity become redundant at the senior years,[26] allowing the relatively younger athlete to perform on a more equal footing. Nevertheless, such an explanation is reliant upon relatively younger athletes remaining actively engaged in sport through years of unfavourable selection and attainment. It is worth reminding that Helsen et al.[30] and Barnsley et al.[11] reported higher drop-out rates in relatively younger players across junior and adolescent ages. Another possibility is that senior athletes transfer from one sport to another (even from lower levels of involvement and in contexts where RAEs are more or less likely), thereby avoiding the disadvantaged developmental environment. In elite team sports, Baker et al.[74] noted that elite athlete development profiles were highly variable, suggesting that this type of late-stage transfer is possible, depending on the compatibility of performance requirements. An alternative explanation is that relatively older athletes, originally selected for additional training and higher levels of skill representation during their junior and adolescent years, withdraw from competitive levels of participation preceding or during their senior years due to injury, overtraining, burnout or boredom. In popular highly competitive sports (e.g. soccer and ice hockey), many talented athletes in the adolescent years (15–18) do not fulfil their early potential by attaining a professional contract. Some limited evidence suggests that highly specialized training environments, such as those conducive to RAE Sports Med 2009; 39 (3)

250

ª 2009 Adis Data Information BV. All rights reserved.

Table II. Unadjusted odds ratios (ORs) for female independent samples examining relative age effect in sport Study

Subject age (y)

Sport

Level of competition

Grondin et al.[2]

12–13

Volleyball

Junior

14–15

Volleyball

16–17

OR comparisons [Q1–4/1st and 2nd 6 mo)] (95% CI) Q1 vs Q4 Q2 vs Q4 Q3 vs Q4

1st vs 2nd

96

1.56 (0.67, 3.63)

2.37 (1.05, 5.35)

1.06 (0.43, 2.57)

1.9 (0.49, 3.86)

Youth cadet

97

1.00 (0.46, 2.15)

1.1 (0.51, 2.36)

0.35 (0.14, 0.89)

1.55 (0.49, 3.11)

Volleyball

Youth juvenile

56

1.14 (0.4, 3.2)

1.00 (0.35, 2.85)

0.85 (0.29, 2.49)

1.15 (0.4, 2.86)

14–15

Volleyball

Provincial youth cadet

219

2.28 (1.3, 3.99)

2.12 (1.21, 3.73)

1.43 (0.79, 2.58)

1.8 (0.62, 2.87)

16–17

Volleyball

Provincial youth juvenile

188

1.25 (0.7, 2.25)

1.43 (0.8, 2.55)

1.12 (0.62, 2.03)

1.26 (0.6, 2.07)

17–19

Volleyball

Provincial youth

59

1.06 (0.39, 2.86)

0.81 (0.29, 2.27)

0.81 (0.29, 2.27)

1.03 (0.41, 2.5)

Senior

Volleyball

Provincial senior

40

0.87 (0.22, 3.34)

1.62 (0.46, 5.62)

1.5 (0.42, 5.24)

1.00 (0.34, 2.92)

11–18

Swimming

Elite junior

60

2.11 (0.72, 6.14)

2.11(0.72, 6.14)

1.44 (0.47, 4.38)

1.72 (0.41, 4.19)

9–18

Tennis

Elite junior

81

2.23 CI (0.9, 5.47)

1.84 CI (0.74, 4.6)

1.15 CI (0.43, 3.02)

1.89 CI (0.46, 4.07)

Baxter-Jones et al.[25]

9–18

Gymnastics

Elite junior

81

1.64 (0.69, 3.89)

1.23 (0.5, 3)

0.88 (0.34, 2.23)

1.53 (0.46, 3.27)

Hoare[65]

15–16

Basketball

Junior regional representative

130

5.72 (2.56, 12.7)

3.81 (1.67, 8.69)

1.27 (0.5, 3.21)

4.2 (0.52, 8.07)

15–16

Basketball

Junior regional representative

100

6.71 (2.54, 17.6)

3.42 (1.25, 9.39)

3.14 (1.13, 8.67)

2.44 (0.49, 4.94)

17–18

Basketball

Junior regional representative

98

1.77 (0.79, 3.97)

1.27 (0.55, 2.93)

1.38 (0.6, 3.16)

1.27 (0.5, 2.54)

Senior

Basketball

Professional

78

3.00 (1.15, 7.77)

1.8 (0.66, 4.87)

2.00 (0.74, 5.35)

1.6 (0.46, 3.47)

Senior

Netball

National professional

128

1.58 (0.8, 3.11)

0.86 (0.41, 1.78)

0.96 (0.47, 1.97)

1.24 (0.54, 2.27)

Senior

Netball

National professional

119

0.9 (0.43, 1.85)

0.8 (0.38, 1.67)

1.12 (0.55, 2.27)

0.8 CI (0.53, 1.49)

Senior

Tennis

ATP professional

211

1.94 (1.11, 3.38)

1.61 (0.91, 2.83)

1.3 (0.73, 2.32)

1.54 (0.62, 2.47)

239

1.85 (1.09, 3.13)

1.47 (0.86, 2.52)

1.65 (0.96, 2.8)

1.25 (0.64, 1.94)

72

1.61 (0.62, 4.18)

2.00 (0.78, 5.08)

0.92 (0.33, 2.56)

1.88 (0.44, 4.24)

804

1.11 (0.84, 1.47)

1.14 (0.86, 1.51)

1.10 (0.83, 1.45)

1.07 (0.78, 1.36)

Baxter-Jones[14]

O’Donoghue et al.[17]

Edgar and O’Donoghue [13]

No. of subjects

Tennis

ITF national juniors

17–18

Soccer

Youth national

Vincent and Glamser[72]

17–18

Soccer

State representative

17–18

Soccer

Regional representative

71

1.33 (0.52, 3.4)

1.53 (0.6, 3.86)

0.86 (0.32, 2.33)

1.53 (0.44, 3.45)

17–19

Soccer

Developmental national

39

3.00 (0.78, 11.5)

1.40 (0.32, 5.97)

2.40 (0.6, 9.44)

1.29 (0.33, 3.84)

Senior

Ice hockey

Senior women

299

1.04 (0.65, 1.65)

1.29 (0.82, 2.03)

1.05 (0.66, 1.67)

3321

1.21 (1.10, 1.33)

Wattie et al. [42] Summary effect size

ATP = Association of Tennis Professionals; ITF = International Tennis Federation; Q = quartile.

1.13 (0.67, 1.68) 1.39 (1.26, 1.54)

Cobley et al.

Sports Med 2009; 39 (3)

14–18 Helsen et al.[12]

Relative Age Effects in Sport

251

Table III. Summary odds ratios (ORs) for relative age effects in sport according to age categorya Age category

Q1 vs Q4 no. of samples (% of total)

1st vs 2nd 6 mo summary OR (95% CI)

no. of samples (% of total)

summary OR (95% CI)

£10 y

17 (6.91)

1.22 (1.08, 1.39)

17 (6.71)

1.12 (1.03, 1.22)

Junior (11–14 y)

42 (17.07)

1.29 (1.29, 1.96)

44 (17.39)

1.36 (1.15, 1.60)

Adolescent (15–18 y)

69 (28.04)

2.36 (2.00, 2.79)

70 (27.66)

1.72 (1.54, 1.92)

107 (43.49)

1.44 (1.35, 1.53)

110 (43.47)

1.29 (1.24, 1.35)

Senior (‡19 y) a

11 samples from the Q1 vs Q4 comparison and 12 samples from the 1st vs 2nd 6-mo comparisons were excluded from the analysis due to participant samples crossing age-category boundaries applied.

Q = quartile.

occurrence (i.e. through selection and identification processes), are related to shorter playing careers and increased rates of dropout at the senior level.[75] On the whole, we can only speculate as to why RAEs decline at the elite stage. Nonetheless, these results demonstrate that a slightly ‘more even playing-field’ exists for those relatively younger individuals within senior and elite echelons of sport. 4.2 Context-Specific Findings

Meta- and substratification analyses were less able to accurately account for the potential role of context specificity, whereby unique sociocultural variables could amplify or reduce RAEs. For example, soccer and ice hockey show consistent RAEs, regardless of age and skill level. However, identifying basketball as a context for heightened risk of RAEs contradicts the equivocal findings of some individual studies (e.g. Daniel and Janssen[41]). Upon further examination of basketball samples, over half of the samples (i.e. eight) included in our OR calculations were derived from the data of Hoare[65] examining Australian Basketball. In this context, RAEs were exceptionally high, lending credence to the suggestion that factors distinctive to the developmental structure of Australian Basketball may escalate RAEs. Likewise, factors distinct to the developmental structure of American football (e.g. drafting and selection at later ages) may also reduce the likelihood of RAEs (as argued by Daniel and Janssen[41]). These context-specific findings suggest RAE risk is variable and that ª 2009 Adis Data Information BV. All rights reserved.

those responsible for sport structures can modify and potentially eradicate RAE inequalities. 4.3 Eliminating Relative Age Effects

Several recommendations have been proposed to resolve RAEs. Initially these addressed annual age-groupings, by advocating a change in the age-group cut-off date (e.g. from January to June), rotating cut-off dates from year to year (Barnsley et al.[1]), or altering age-grouping bandwidths. However, changing cut-off dates only leads to a transfer of RAEs,[71] as exemplified in Australian,[63] Belgian[64] and English[67] youth soccer. To prevent a ‘fixed-bias’ across sport development, Grondin et al.[2] proposed an expansion of age-group bandwidths to 15 and 21 months, as opposed to the typical 12-month groupings, to rotate cut-off dates across particular ages and constantly change group composition. Similarly, Boucher and Halliwell[57] proposed a 9-month bandwidth (referred to as the Novem system) to reduce potential age inequalities in a given group, whilst also ensuring that the same participants (i.e. relatively older or younger) were not disadvantaged year after year during youth stages of competition (i.e. present Under 10s to Under 16s). To address the RAE inequality in Canadian ice hockey specifically, Hurley et al.[76] presented the relative age fair (RAF) cycle, whereby cut-off dates altered for each and every consecutive year of participation. In their plan, cut-off dates changed by 3 months between seasons of competition, to ensure players experienced being in each quartile position Sports Med 2009; 39 (3)

Cobley et al.

252

Table IV. Summary odd ratios (ORs) for relative age effects according to skill levela Skill level

Q1 vs Q4

1st vs 2nd 6 mo

no. of samples (% of total)

summary OR (95% CI)

no. of samples (% of total)

summary OR (95% CI)

Recreational

28 (11.38)

1.12 (1.05, 1.20)

28 (11.06)

1.09 (1.03, 1.15)

Competitive

53 (21.54)

1.63 (1.35, 1.97)

55 (21.73)

1.40 (1.21, 1.62)

Representative

70 (28.45)

2.77 (2.36, 3.24)

73 (28.85)

1.87 (1.68, 2.07)

Elite

95 (38.61)

1.42 (1.34, 1.51)

97 (38.33)

1.28 (1.22, 1.33)

a

Samples that could not be clearly categorized into one of the above were excluded from comparison analyses.

Q = quartile.

(i.e. Q1, Q2, Q3 and Q4) across the competitive junior structure of ice hockey. Both the Novem and RAF strategy may help address the RAE problem; however, there is foreseeable complexity in re-structuring and implementing these options within youth sports. On a separate but related point, Musch and Grondin[21] noted that across sports contexts (and education) many cut-off dates used for agegrouping are actually similar (e.g. 1 September to 31 August in the UK). So, to prevent repeated and consistent (dis)advantages from occurring, they recommended that deliberate variation of cut-off dates be used for across sports contexts. This step would prevent generic RAEs across sports contexts and may reduce the likelihood of persistent negative experiences of sporting involvement relative to age-matched peers, and may help maintain sporting involvement in contexts with favourable conditions (i.e. in which you are relatively older). Nevertheless, it may not prevent RAEs within a given sport, thereby not preventing their occurrence. Other possible solutions have targeted maturational differences and the process by which athletes are selected. Barnsley and Thompson[3] advocated implementing player quotas, where selection must meet specified birth-date distributions to prevent favouring of relatively older players. More substantially, the average age of a whole team,[30,64] the number of selections and the distribution of playing time could be regulated. Another popular solution has been to suggest grouping participants according to physical (i.e. height and weight) classification,[14,21] similar to that routinely adopted in boxing and wrestling. ª 2009 Adis Data Information BV. All rights reserved.

More sensitive to individual variability in physical characteristics, this may be sensible particularly during developmental stages. Again, these strategies may prove difficult to integrate into sport systems and are as yet unproven in their value for resolving RAEs. A less challenging solution is to delay the processes of selection, identification and representation beyond stages of puberty and maturation (i.e. 15–16 years of age). Governing bodies and coaches should reconsider the necessity for early selection, intensive training and levels of representation at junior and child ages. Admittedly, the path to success in sport does require intensive long-term training and commitment, often referred to as the ‘10-year rule of attainment’.[77] Yet, peak performance in many sports (e.g. soccer, ice hockey) is often not attained until the late twenties and thirties, providing a sufficient window for training and development subsequent to adolescence. This position is further substantiated by an expanding literature illustrating concerns for the physiological and psychosocial welfare of athletes involved in intensive training from early ages.[78-80] Delaying selection might reduce RAEs and indirectly help reduce the risk of compromising health during an athlete’s development. Another possibly beneficial approach would be to raise awareness of RAEs among those responsible for the infrastructure and coordination of youth sport. Sports contexts with higher risks of RAEs (e.g. Canadian ice hockey and European soccer) should be targeted. During adolescence, coaches need to be attentive to the possibility that physical attributes, such as height and weight Sports Med 2009; 39 (3)

Relative Age Effects in Sport

253

(which underpin speed, power and strength) are being overlooked during early stages of athlete development (i.e. 13–16 years old), conveying selection advantages to the relatively older at a time coinciding with intense identification and selection to competition at representative levels. Whilst monitoring for RAEs in selection and participation, coaches should integrate more movement and skill-based (e.g. movement accuracy, consistency and adaptability) criteria in selection, reducing the association and dependence upon physical attributes. Up to the mid-teenage years (i.e. 15–16), governing bodies and coaches should re-assess whether tiers of competition and levels of representation could be removed to support selection and representation in later years. To assist, it should be considered that coaches within sport development systems are pressured to obtain immediate performance success. As a result, coaches may face a constant battle of selecting individuals/teams that help guarantee immediate success in youth ages (i.e. likely to be relatively older with advanced physical characteristics at the moment), as opposed to individuals/teams that may be more successful in the longer term. Strategies focusing on raising awareness with recommendations pertaining to the importance of delayed selection may help reduce the emphasis on striving for immediate performance success in youth. Considerate of the findings from the present study, it is certainly feasible that if a relatively younger athlete maintains sport involvement, despite the constant and disadvantaging annual age-grouping policy in

youth, this may prove beneficial in the senior years. 4.4 Future Directions

This meta-analytical review has identified several areas where further research is needed. Data consistently and recurrently support the presence of RAEs in specific sports, yet in other contexts the data are less conclusive. For example, data from basketball are somewhat equivocal, while data from women’s sports are particularly sparse. Initial data in women’s sports (e.g. Wattie et al.[42]) suggests these contexts may not be as susceptible to RAEs. One explanation for the discrepancy associates differences in participation rates and lower competition for selection into representative skill levels as reasons for a reduced likelihood of RAE risk. Researchers should also more broadly consider a range of sociocultural contexts. Largely limited by studies in North America, Europe and Australia to date, examinations of RAEs in African, South American and Asian countries might provide valuable information about the role of sport infrastructure in perpetuating RAEs. Similarly, a comparison of RAEs between sport development systems utilizing early talent identification systems (e.g. Australia) and those without explicit programmes may be useful. This would determine whether RAEs are the result of greater exposure to high-quality resources or due to the effects of early success on the development of self-efficacy and other feelings of competence. Moreover, this research might prove useful for

Table V. Summary odds ratios (ORs) for relative age effects according to sport contexta Sport context

Q1 vs Q4

1st vs 2nd 6 mo

no. of samples (% of total)

summary OR (95% CI)

no. of samples (% of total)

Ice hockey

77 (31.30)

1.62 (1.45, 1.79)

83 (32.80)

1.40 (1.31, 149)

Soccer

76 (30.89)

2.01 (1.73, 2.32)

76 (30.03)

1.55 (1.37, 1.74)

Baseball

33 (13.41)

1.20 (1.12, 1.30)

33 (13.04)

1.14 (1.08, 1.20)

Basketball

15 (6.09)

2.66 (1.80, 3.93)

15 (5.92)

1.77 (1.34, 2.33)

Volleyball

14 (5.69)

1.33 (1.07, 1.65)

14 (5.53)

1.24 (1.03, 1.49)

7 (2.84)

1.24 (0.93, 1.65)

7 (2.76)

1.08 (0.94, 1.23)

American Football a

summary OR (95% CI)

Samples examining other sport contexts (e.g. tennis) were not included in any Q1 vs Q4 or 1st vs 2nd 6 mo comparison analyses.

Q = quartile.

ª 2009 Adis Data Information BV. All rights reserved.

Sports Med 2009; 39 (3)

Cobley et al.

254

determining the effectiveness of talent identification systems, since one measure of utility is certainly the degree to which ‘talented’ or ‘gifted’ performers are missed by the system (a type II error). For example, countries and/or sportgoverning bodies employing intensive early talent identification and development systems in sport may, ironically, be achieving the opposite effect by constraining and reducing their talent pool through early selection processes and generating RAEs. In this scenario, young athletes may depart sport prior to full maturity, without opportunity to nurture their skills and inherent interest. Qualitative idiographic investigations examining developmental sport structures, coaching practice and the child/athlete experience within them, will certainly strengthen our understanding of how RAEs manifest and operate. Atheoretical work has dominated the study of RAEs to date and future studies should be grounded in more theoretically sound foundations. In addition to providing pieces to the puzzle that are currently missing, such studies would be valuable for creating a sound theoretical understanding of (i) the origins of RAEs, (ii) their implications in human development, and (iii) ways in which development systems can be modified to reduce or remove RAEs in the future. 5. Conclusions This meta-analysis suggests consistent small risks of RAEs are apparent across sport, with the relatively younger members of annual age-group cohorts persistently disadvantaged. Risk size is moderated by several factors, including chronological age differences (i.e. number of months) between cohort members, age category, skill level and sport context. Practices that produce RAEs need to be revised, whilst interventions that reduce or eradicate this sporting inequality need to be implemented and evaluated. These steps are necessary as annual age-grouping and associated processes appear to constrain the likelihood of immediate and long-term participation as well as attainment in sport. Whether you are motivated toward realizing the positive effects of sport on youth development (e.g. promoting fun, ª 2009 Adis Data Information BV. All rights reserved.

enjoyment and inclusive participation), or are interested in elite athlete development, the presence of RAEs (we argue) appears contradictive to both these outcomes. Certainly, notions of competition, selection and talent identification, which seem to create RAEs across developmental stages, should be addressed by sport organizations. This is the main challenge facing researchers, sport governing bodies, coaches, parents and athletes alike. Acknowledgement No funding was received for this review, and the authors have no conflicts of interest that are directly relevant to the content of this review.

References 1. Barnsley RH, Thompson AH, Barnsley PE. Hockey success and birthdate: the RAE. Can Assoc Health Phys Educ Recreation J 1985; 51: 23-8 2. Grondin S, Deschaies P, Nault LP. Trimesters of birth and school output [in French]. Apprent Social 1984; 16: 169-74 3. Barnsley RH, Thompson AH. Birthdate and success in minor hockey: the key to the NHL. Can J Behav Sci 1988; 20: 167-76 4. Armstrong HG. A comparison of the performance of summer and autumn-born children at eleven and sixteen. Br J Educ Psychol 1966; 36: 72-6 5. Freyman R. Further evidence on the effect of date of birth on subsequent school performance. Educ Res 1965; 8: 58-64 6. Alton A, Massey A. Date of birth and achievement in GCSE and GCE A-level. Educ Res 1998; 40: 105-9 7. Thompson A, Barnsley R, Stebelsky G. ‘Born to play ball’: the relative age effect and major league baseball. Sociol Sport J 1991; 8: 146-51 8. Thompson A, Barnsley R, Stebelsky G. Baseball performance and the relative age effect: does little league neutralize birthdate selection bias? Nine 1992; 1: 19-30 9. Boucher J, Mutimer B. The relative age phenomenon in sport: a replication and extensions with ice-hockey players. Res Q Exerc Sport 1994; 65: 377-81 10. Verhulst J. Seasonal birth distribution of West European soccer players: a possible explanation. Med Hypotheses 1992; 38: 346-8 11. Barnsley RH, Thompson AH, Legault P. Family planning: football style – the relative age effect in football. Int Rev Sociol Sport 1992; 27: 77-88 12. Helsen WF, Van Winckel J, Williams MA. The relative age effect in youth soccer across Europe. J Sports Sci 2005; 23: 629-36 13. Edgar S, O’Donoghue P. Season of birth distribution of elite tennis players. J Sports Sci 2005; 23: 1013-20 14. Baxter-Jones A. Growth and development of young athletes: should competition levels be age related? Sports Med 1995; 20: 59-64

Sports Med 2009; 39 (3)

Relative Age Effects in Sport

15. Abernethy B, Farrow D. Contextual factors influencing the development of expertise in Australian athletes. Proceedings of the 11th World Congress of Sport Psychology; 2005 Aug 15-19; Sydney (NSW) 16. Edwards S. Born too late to win? [letter] Nature 1994; 370: 186 17. O’Donoghue P, Edgar S, McLaughlin E. Season of birth bias in elite cricket and netball. J Sports Sci 2004; 22: 256-7 18. 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: 1065-73 19. Sherar LB, Baxter-Jones ADG, Faulkner RA, et al. Does physical maturity and birth date predict talent in male youth ice hockey players? J Sports Sci 2007; 25(8): 879-86 20. Malina RM. Physical growth and biological maturation of young athletes. Exerc Sport Sci Rev 1994; 22: 389-434 21. Musch J, Grondin S. Unequal competition as an impediment to personal development: a review of the relative age effect in sport. Dev Rev 2001; 21: 147-67 22. Malina RM, Bouchard C, Bar-Or O. Growth, maturation and physical activity. 2nd ed. Champaign (IL): Human Kinetics, 2004 23. Malina RM, Ribeiro B, Aroso J, et al. Characteristics of youth soccer players aged 13-15 years classified by skill level. Br J Sports Med 2007; 41: 290-5 24. Baxter-Jones A, Helms P, Baines-Preece J, et al. Growth and development of male athletes: implications for identification of talent [abstract]. J Sports Sci 1994; 12: 156 25. Baxter-Jones A, Helms P, Maffuli N, et al. Growth and development of male gymnasts, swimmers, soccer and tennis players: a longitudinal study. Ann Hum Biol 1995; 22: 381-94 26. Levfvre J, Beunen G, Steens G, et al. Motor performance during adolescence and age thirty as related to age at peak height velocity. Ann Hum Biol 1990; 17: 423-35 27. Brewer J, Balsom PD, Davis JA, et al. The influence of birth date and physical development on the selection of a male junior international soccer squad. J Sports Sci 1992; 10: 561-2 28. Garganta J, Maia J, Silva R, et al. A comparative study of explosive leg strength in elite and non-elite young soccer players. In: Reilly T, Clarys J, Stibbe A, editors. Science and football. London: Spon, 1993: 304-6 29. Ste-Marie DM, Starkes J, Cronin A, et al. The relative age effect: a ‘flip-flop’ phenomenon in gymnasts. Proceedings of the Canadian Society for Psychomotor Learning and Sport Psychology Conference; 2000 Oct; Waterloo (ON) 30. Helsen WF, Starkes JL, Van Winckel J. The influence of relative age on success and dropout in male soccer players. Am J Hum Biol 1998; 10: 791-8 31. Ericsson KA, Krampe RT, Tesch-Ro¨mer C. The role of deliberate practice in the acquisition of expert performance. Psychol Rev 1993; 100: 363-406 32. Baker J, Horton S. A review of primary and secondary influences on sport expertise. High Ability Studies 2004; 15: 211-28 33. Ward P, Williams AM. Perceptual and cognitive skill development in soccer: the multidimensional nature of expert performance. J Sport Exerc Psychol 2003; 25: 93-111

ª 2009 Adis Data Information BV. All rights reserved.

255

34. Williams AM. Perceptual skill in soccer: implications for talent identification and development. J Sports Sci 2000; 18: 737-50 35. Harter S. Effectance motivation reconsidered: toward a developmental model. Hum Dev 1978; 21: 34-64 36. Harter S. Causes and consequences of low self-esteem in children and adolescents. In: Baumeister RF, editor. Selfesteem: the puzzle of low self-regard. New York: Plenum Press, 1993: 87-116 37. Bandura A. Social foundations of thought and action: a social cognitive theory. Englewood Cliffs (NJ): Prentice Hall; 1986 38. Smith RE. Toward a cognitive-affective model of athletic burnout. J Sport Psychol 1986; 8: 36-50 39. Gould D, Pelichkoff L. Participation motivation and attrition in young athletes. In: Smoll FL, Magill RA, Ash MJ, editors. Children in sport. 3rd ed. Champaign (IL): Human Kinetics, 1988: 161-78 40. Weiss R, Chaumeton N. Motivational orientations in sport. In: Horn TS, editor. Advances in sport psychology. Champaign (IL): Human Kinetics, 1992: 61-99 41. Daniel TE, Janssen CTL. More on the relative age effect. Can Assoc Health Phys Educ Recreation J 1987; 53: 21-4 42. Wattie N, Baker J, Cobley S, et al. A historical examination of relative age effects in Canadian hockey players. Int J Sport Psychol 2007; 38: 178-86 43. Pyne DB, Gardner AS, Sheehan K, et al. Positional differences in fitness and anthropometric characteristics in Australian football. J Sci Med Sport 2006; 9: 143-50 44. Rudas T. Odds ratios in the analysis of contingency tables: qualitative application in the social sciences series. Thousand Oaks (CA): Sage Publications, 1997 45. Statistics Canada. Adapted from CANSIM database. Table 053-0001 vital statistics, births, deaths and marriages, quarterly (1970-2000) [online]. Available from URL: http://cansim2.statcan.ca. [Accessed 2007 Feb 25] 46. Office for National Statistics (UK). Birth Statistics FM1, volume numbers 11-32 [online]. Available from URL: http://www.statistics.gov.uk. [Accessed 2007 Jan 26] 47. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled Clin Trials 1986; 7: 177-88 48. Sutton AJ, Abrams KR, Jones DR, et al. Methods for meta-analysis in medical research. Chichester: Wiley and Sons, 2002 49. Cochran WG. The combination of estimates from different experiments. Biometrics 1954; 10: 101-29 50. Review manager (Rev Man). Version 4.2 for Windows. Copenhagen: the Nordic Cochrane Centre, the Cochrane Collaboration 2003 [online]. Available from URL: http:// www.cc-ims.net/RevMan [Accessed 2007 Jan 28] 51. Montelpare WJ, Scott D, Pelino M. Tracking the relative age effect across minor, amateur and professional icehockey leagues. In: Ashare AB, editor. Safety in ice hockey. vol. 3. Philadelphia (PA): American Society for Testing and Materials, 2000: 250-60. ASTM STP 1341 52. Haddock CK, Rindskopf D, Shadish WR. Using odds ratios as effect sizes for meta-analysis of dichotomous data: a primer on methods and issues. Psychol Methods 1998; 3: 339-53

Sports Med 2009; 39 (3)

256

53. Begg CB. Publication bias. In: Cooper H, Hedges LV, editors. The handbook of research synthesis. New York: Russell Sage Foundation, 1994: 399-409 54. Egger M, Smith GD, Schneider M, et al. Bias in metaanalysis detected by a simple graphical test. BMJ 1997; 315: 629-34 55. Gelber RD, Goldhirsch A. The interpretation of results from sub-set analyses within overviews of randomized clinical trials. Stat Med 1987; 6: 371-88 56. Yusuf S, Wittes J, Probstfield J, et al. Analysis and interpretation of treatment effects in subgroups of patients in randomized clinical trials. JAMA 1991; 266: 93-8 57. Boucher J, Halliwell W. The Novem system: a practical solution to age grouping. Can Assoc Health Phys Educ Recreation J 1991; 57: 16-20 58. Grondin S, Trudeau F. Date of birth and the National Hockey League: analyses of different parameters [in French]. Rev Sci Tech Act Phys Sport. 1991; 26: 37-45 59. Glamser F, Marciani L. The birthdate effect and college athletic participation: some comparisons. J Sport Behav 1992; 15: 227-38 60. Dudink A. Birthdate and sporting success [letter]. Nature 1994; 368: 592 61. Brewer J, Balsom PD, Davis J. Seasonal birth distribution amongst European soccer players. Sports Exerc Inj 1995; 1: 154-7 62. Stanaway KB, Hines T. Lack of a season of birth effect among American athletes. Percept Mot Skills 1995; 81: 952-4 63. Musch J, Hay R. The relative age effect in soccer: crosscultural evidence for a systematic discrimination against children born late in the competition year. Sociol Sport 1999; 16: 54-64 64. Helsen WF, Starkes JL, Van Winckel J. Effect of a change in selection year on success in male soccer players. Am J Hum Biol 2000; 12: 729-35 65. Hoare D. Birthdate and basketball success: is there a relative age effect? Proceedings of the Pre-Olympic Congress: Sports Medicine and Physical Education International Congress on Sports Science; 2000 Sep 7-13; Brisbane (QLD) 66. Grondin S, Koren S. The relative age effect in professional baseball: a look at the history of major league baseball and at current status in Japan. Avante 2000; 6: 64-74 67. Simmons C, Paull GC. Season-of-birth bias in association football. J Sports Sci 2001; 19: 677-86

ª 2009 Adis Data Information BV. All rights reserved.

Cobley et al.

68. Musch J. Birthdate and success in youth soccer: investigating the development of the relative age effect. Sportonomics 2002; 8: 22-8 69. Glamser FD, Vincent J. The relative age effect among elite American youth soccer players. J Sport Behav 2004; 27: 31-8 70. Edgar S, O’Donoghue P. The distribution of season-of-birth among players of the 2002 FIFA World Cup. J Sports Sci 2004; 22: 578-9 71. Vaeyens R, Philippaerts RA, Malina RM. The relative age effect in soccer: a match-related perspective. J Sports Sci 2005; 23: 747-56 72. Vincent J, Glamser FD. Gender differences in the relative age effect among US Olympic development program youth soccer players. J Sports Sci 2006; 24: 405-13 73. Esteva S, Drobnic F. Birthdate and basketball success. FIBA Assist Magazine 2006; 18: 64-6 74. Baker J, Coˆte´ J, Abernethy B. Sport specific training, deliberate practice and the development of expertise in team ball sports. J Appl Sport Psychol 2003; 15: 12-25 75. Barynina II, Vaitsekhovskii SM. The aftermath of early sports specialization for highly qualified swimmers. Fitness Sports Rev Int 1992; 27: 132-3 76. Hurley W, Lior D, Tracze S. A proposal to reduce the age discrimination in Canadian minor hockey. Can Public Policy 2001; 27 (1): 65-75 77. Simon HA, Chase WG. Skill in chess. Am Scientist 1973; 61: 394-403 78. American Academy of Pediatrics. Intensive training and sports specialization in young athletes. Pediatrics 2000; 106: 154-7 79. Wiersma LD. Risks and benefits of youth sport specialization: perspectives and recommendations. Pediatric Exerc Sci 2000; 12: 13-22 80. Baker J. Early specialization in youth sport: a requirement for adult expertise? High Ability Stud 2003; 14: 85-94

Correspondence: Dr Stephen Cobley, Room 124, Fairfax Hall, Carnegie Faculty of Sport and Education, Headingley Campus, Leeds Metropolitan University, West Yorkshire, LS6 3QS, UK. E-mail: [email protected]

Sports Med 2009; 39 (3)