A total of 32 research papers involving 49 experimental studies were considered for this review. Each study provided data on playing level, number of subjects, measurement process and VO_2Max values. Of these, 17 studies provided full details of age, body mass and stature and 15 studies provided full or partial details of these parameters. Only 7 of these studies provided full details of VO_2Max values for the 5 positional categories and a further 7 studies provided details for only some of the positional categories. Tables 1 and 2 illustrate the full profile of studies according to the inclusion criteria and the values for U-17, U-20 and First Division players.

The anthropometric profiles for body mass and stature of soccer players in Brazil provide some clear comparisons with other values reported for a similar era. The U-17 players in Brazil were of similar stature to professional youth soccer players in the UK but appeared to have less body mass (Franks et al., 1999, McMillan et al., 2005). The Brazilian based studies reported a range of mean values of 173-177cm and 60-71kg for the U-17 group. The U-20 soccer players ranged from mean values between 174-181cm and 66-75.5kg which were similar values to those players of equivalent age reported for Tunisian players (Chamari et al., 2004) and Greek players Metaxas et al., 2005) as well as similar body mass, but inferior stature to Norwegian players (Helgerud et al., 2001).

The profiles of the First Division players in Brazil were similar in stature and body mass to players in Greece (Kalapotharakos et al., 2006) and Saudi Arabia (Al-Hazzaa et al., 2001). The ranges of mean values for these players were 173-181cm and 61-73kg. However, the players in Brazil were similar only in body mass with players of similar positions from Spain (Bloomfield et al., 2005, Casajús, 2001), Germany (Bloomfield et al., 2005) Italy (Bloomfield et al., 2005, Mohr et al., 2003), Denmark (Krustrup et al., 2006, Krustrup et al. 2003, Mohr et al., 2003), Norway (Wisløff et al., 1998), England (Bloomfield et al., 2005) and were generally smaller in stature with equivalent players from these European countries. Table 3 illustrates biographical, anthropometric and performance characteristics of professional and elite soccer players for other Nations since 1995.

The majority of tests for the assessment of VO_2Max for U-17 and U-20 categories were performed through indirect evaluation. For the U-17 players, the majority of protocols used were either the 20m Multi-Stage Fitness Test (MSFT), originally proposed by Leger et al., 1988, or forms of incremental running to maximum on a treadmill. There were only 3 studies that were conducted with online gas analysis to exhaustion on a treadmill (Azevedo et al., 2006; Ley et al., 2002; Silva et al., 1997b). The highest values reported for this age group were by Lopes and Marins, 1996 who used a maximal effort on a treadmill to predict a VO_2Max of 62.9 ± 3.37 ml·kg^-1·min^-1 for forwards and 61.39 ± 0.65 ml·kg^-1·min^-1 for external defenders. Silva et al., 1997b also reported a value of 65.97 ± 4.81 ml·kg^-1·min^-1 through gas analysis from a treadmill test in a group of 19 players aged approximately 17 years old but unfortunately did not indicate which positions these players played. Similarly, for the U-20 players, the most used protocol was also the MSFT, however, there has been a wider range of test protocols reported than at U-17 level. This includes maximum indirect protocols in treadmill, Yo-Yo Endurance Test (Bangsbo, 1994), Yo-Yo Intermittent Test (Krustrup, 2003), a 3600m Test, Margaria (Margaria, 1963) and a 5min running test known as V_amax(5) (Berthon et al., 1997). Direct measurements occurred on only two occasions in this age category although one study contained 42 subjects but provided no coverage of positional differences (Silva et al., 1997b). The highest values reported were by Pereira (2003) who measured 26 First National Level players aged 18-19 using the MSFT and predicted an overall value of 63.23 ± 3.79 ml·kg^-1·min^-1 with only small differences between the outfield positions with the midfielders reaching a predicted peak of 65.60 ± 3.67 ml·kg^-1·min^-1. In contrast, Campeiz et al., 2005 provided values below 60 ml·kg^-1·min^-1 through use of the Yo-Yo Endurance Test in a group of 13 First State players aged approximately 18 years old. In this trial, external defenders reached the highest values of all the positions, although these players were of a lower playing standard to the other studies performed at this age level. It is also important to note that the Yo-Yo test is an assessment for intense intermittent exercise and requires high rates of aerobic and anaerobic energy turnover which may act as a limiting factor to determine accuracy in aerobic capacity evaluation.

At First Division level, a greater number of VO_2Max assessments that have been reported in the scientific press were performed through direct measurement protocols compared to U-17 and U-20 groups. Of all VO_2Max evaluation studies performed with this level, 47.83% used gas analysis techniques with the remaining studies used a wide range of indirect protocols similar to the U-20 group, including the MSFT and Yo-Yo Tests. The reported overall values were slightly lower than those reported in the U-20 group with gas analysis mean values higher than 60 ml·kg^-1·min^-1 provided by Silva et al., 1999 and Silva et al., 2000 for 18 and 26 First National players respectively and Balikian et al., 2002 for 25 Second State players. The majority of remaining studies reported mean values between 50-57 ml·kg^-1·min^-1. The highest reported value for a positional group was 62.31 ± 0.86 ml·kg^-1·min^-1 for external defenders within a group of 45 First National players, although the details of the methodology were not provided (Pavanelli et al., 2004).

In Brazil, a country with more than 180 million inhabitants, soccer is regarded as the national sport with a reported more than 30 million recreational players and 12,000 players with professional registrations in clubs all over the world (CBF, 2008). At an elite level, there however appears to be distinct physical and physiological differences between Brazilian players and their European counterparts. In this respect, it appears that although players in Brazil are shorter in stature, they are similar in body mass. In addition, the talent selection process in Brazil is regarded as different to Europe due the Brazilian style of play which prioritizes player selection through the technical aspects of the game rather than recruiting through physical size (Drubsky, 2003, p. 187). In contrast, the stature and body mass of players has been suggested to be an important policy for player recruitment in major European Leagues (England, Italy, Spain, Germany) in order to suit the style of the play and the specific physical demands of each playing position (Bloomfield et al., 2005). In this respect, the top leagues in Spain and Italy had a much higher number of players from South America and on average, shorter and lighter players compared to those in England and Germany, particularly in midfield and forward positions (Bloomfield et al., 2005). These authors also postulated that this phenomenon could additionally be attributed to cultural and climatic reasons for adaptation to lifestyle in a European country when migrating from South America, rather than simply for the style of play adapted by teams in these respective leagues. Of course, other criteria for success include tactical awareness, teamwork abilities, fitness and psychological variables.

Interestingly, although clear trends exist in terms of age, stature and body mass for different positions in different countries, physical size and shape is not necessarily a priority predictor for success (Reilly et al., 2000). There are also a wide range of physiological values reported for players of different levels in different countries. One possible reason for these differences is a lack of consistency or uniformity in testing protocols to assess VO_2Max. This not only relates to the wide choice of possible tests to implement but also the variety in fitness status of the players, the stage of the season when they are tested, the preparation of the players prior to testing, their motivation, the testing arena itself as well as the environmental conditions. Furthermore, due to the high volume of games per season in most countries, it can be very difficult to schedule appropriate testing occasions. All these factors may influence the outcome of the tests and some or all of this information is often not reported in the articles reviewed in this study, particularly in assessments of indirect measurements. This can therefore make it difficult to make true comparisons between studies and only assumptions should be made when assessing any differences. Furthermore, it is important to recognize that a high VO_2Max value is not the sole predictor of high aerobic performance in soccer, as this is dependent of two other important elements, namely anaerobic threshold and running economy (Bunc et al., 1987; Chamari et al., 2005; Edwards et al., 2003; Helgerud et al., 2001; Kemi et al., 2003). The VO_2Max assessment also possesses some limitations, particularly in the interpretation of results based upon ml.kg^-1.min^-1 as the valid oxygen consumption necessary for an intermittent running pattern as in soccer does not rise in direct proportion to body mass (Wisløff et al., 1998). This results in players with lower body mass having an overestimated VO_2Max value and the heaviest players to having an underestimation in real work capacity terms (McMillan et al., 2005). It is therefore important to factor a correction for body mass (ml·kg^-0.75·min^-1) to make direct comparisons between players of different body mass (Wisløff et al., 1998; Al-Hazzaa et al., 2001; McMillan et al., 2005). Furthermore, there should be some caution in using VO_2Max testing for direct soccer performance assessment, which usually involves an incremental protocol of increasing speeds on a treadmill at set time intervals. As soccer involves an exercise pattern taxing both the aerobic and anaerobic energy pathways in random and intermittent intervals, using dynamic movement patterns (Bloomfield et al., 2007) players are conditioned specifically for this activity pattern and are not ideally prepared for the VO_2Max assessment protocol. Although it is suggested that players should have VO_2Max values superior to 60 ml·kg^-1·min^-1 in order to be competitive at the highest levels in soccer (Reilly et al., 2000), some of the top players would be expected to have superior values in anaerobic components, often to the detriment of their aerobic capacity, therefore VO_2Max values must be viewed alongside speed, agility and power values for a more appropriate assessment (Stølen et al., 2005; Wisløff et al., 1998).

In addition, one of the main limitations for the majority (70.7%) of Brazilian soccer studies is that they even lack detail on the breakdown for different positions and only 15.7% of the studies provide complete data for different playing positions. The absence of this information makes it difficult to characterize and create a more detailed profile of the Brazilian soccer players' aerobic fitness in their different positions as well as different training periods. It has been suggested that at an elite level, an improvement between 3% and 10% should be expected from a physical conditioning period of 4 to 15 weeks during the season (Bangsbo, 1994; Helgerud et al., 2001; Silva et al., 1997a). This highlights the importance of testing through the season, so that a thorough evaluation can occur and training programmes can be adjusted at critical moments of the season.

Overall, the results of VO_2Max in studies relating to Brazilian soccer players suggest that they possess lower levels of aerobic power than players from other countries, particularly in Europe. This may be due to a series of factors, including the total training time dedicated for development of aerobic capacity, the different training regimens; or even a lower innate aerobic capacity, but mainly it may be suggested it is due a different style of play and subsequently a different physical demand needed for Brazilian soccer. This difference was highlighted by Rienzi et al., 2000 who reported a mean total distance of 8,638 ± 1,158m covered by players in the Copa América, was significantly lower (p < 0.05) than the mean total distance covered by players from the English Premier League 10,104 ± 703m. There were also a higher number of passes made in Brazilian soccer and a less direct style of play identified which could relate to the lower levels of aerobic power also reported between these groups of players and was also a similar finding to that reported by Barros et al., 1998 of 7,490 ± 775m for a Brazilian First Division team. However, other studies have found similar distances in Brazilian soccer to European studies including Ananias et al., 1998 who found superior total distance covered by a small number of Brazilian players (n = 6) from First Team soccer (2 central defenders, 2 midfielders and 2 forwards) where the first half mean distance of 5,446 ± 550m was significantly higher (p < 0.05) than the second half mean distance of 4,945 ± 366m and a mean total distance covered of 10,392 ± 849m, a value similar to those found by studies with First Division players from other countries (Bangsbo et al., 2006; Di Salvo et al., 2007; Mohr et al., 2003; Santos et al., 2001). Recently, Barros et al., 2007 also found a similar mean total distance covered in comparison with European soccer of 10,012 ± 1,024m for a larger group of Brazilian First Division players (n = 55) evaluated in 55 games over the course of three seasons between 2001-2004. This may suggest either a change in playing style, pitch size, playing environment, or in fitness capacity in Brazilian players. One possible reason for this potential change could be for Brazilian players to be spotted by European talent scouts on the basis on a player's suitability to compete in the respective leagues and therefore migrate to the more financially lucrative European clubs. However, it should also be noted that methodological differences in assessing work-rate may also account for some of the observed differences between various studies.

Within Brazilian soccer, it appears that from the ages of 17-20 there is an important period of physical development emphasized for preparation for performance at First Division Level. This physical development phase is attributed to a higher volume of intense training performed by players at U-20 level in comparison to U-17 and even First Division (Drubsky, 2003, p. 177). The training includes formal (coached) speed, strength and aerobic power development as well as superior tactical knowledge of the Brazilian style of play, although this development is built on a solid foundation of sound execution of the technical aspects of the game which have been developed up to and beyond U-17 level (Drubsky, 2003, p. 174). However, as the data reported in the literature are often mean values, it would be useful for more of these studies to also have reported the peak values to understand the optimal profile of physical fitness for soccer players in Brazil. Of all the studies reviewed, only 3 provided any peak or range values. These were all in studies of First Division players, all from direct measurements, with individual peak VO_2Max values reported at 73.21 ml·kg^-1·min^-1 (Silva, 1999), 58.6ml.kg^-1.min^-1 (Teixeira, 1999) and 69.8 ml·kg^-1·min^-1 (Silva, 2000) although positional specificity was not supplied. These values would provide conditioning coaches with a much better framework to establish ideal profiles for players of different ages in different positions, and perhaps also the profile of soccer-specific fitness variables at different stages of the season.

It is also important to recognize early in the specific physical development phase which the optimal playing position is for the young player (Gil et al., 2007). It is well understood that different playing positions require superior fitness variables for soccer than other playing positions (Balikian et al., 2002; Stølen et al., 2005; Wisløff et al., 1998). Therefore, identifying each player's specialized position is especially important in order to optimize their physical, physiological, psychological, technical and tactical development so that to prepare them for higher playing levels later in their career. It has been suggested that this should occur as young as 12 to 14 years old (Gil et al., 2007; Strøyer et al., 2004). In Brazil, there exists no governmental policy to provide soccer development in schools and physical education provision is generic to all sports. Furthermore, it appears that 'street soccer' is becoming less practiced and young players are increasingly being taught soccer in specialized soccer schools and clubs. From this system, talented players from 10 years old begin to play more systematically, integrating into the playing style of the club, yet with an emphasis placed on skill development (Drubsky, 2003, p. 147).

In terms of aerobic power, it is common to see the midfield players possessing the highest VO_2Max values in the team, however it is evident that in studies of Brazilian soccer players, the highest values are held by the external defenders although they are only slightly higher than the midfielders. A possible reason for this is in the tactical system incorporated by Brazilian teams. A 4-4-2 playing system is frequently employed and the external defenders are expected to also contribute to attacking play on top of their defensive duties. This may be a characteristic of Brazilian soccer as no differences were found between total distance covered by the First Division external defenders and midfielders in Brazil (Barros et al., 1998; Barros et al., 2007). On the contrary, some studies on European soccer found significantly higher distances covered by midfielders (p < 0.05) than the other positions (Di Salvo et al., 2007; Bangsbo, 1994; Bangsbo et al., 1991), as well as distance covered at high intensity speeds (p < 0.05) (Di Salvo et al., 2007). Furthermore, the high VO_2Max values in this position may also reflect the training effects from the match-play in competition as well as in training, in addition to the organised conditioning performed by these players.

• Physical and physiological differences exist between Brazilian soccer and European soccer.
• Players in Brazil appear to be shorter in stature, similar in body mass and have a lower overall aerobic capacity to their European equivalents
• In Brazil, there seems to be a physical development phase for players at U-20 level which prepares them for the demands at First Division level.