Isokinetic Performance of Knee Flexor and Extensor Muscles in American Football Players from Brazil

The isokinetic performance of thigh muscles has been related to athletic performance and risk for non-contact injuries, such as anterior cruciate ligament ruptures and hamstring strains. Although isokinetic profile of American football players from United States (USA) is widely described, there is a lack of studies comprising players acting outside the USA. The primary objective of this study was to describe the isokinetic performance of thigh muscles in elite American football players in Brazil. Secondarily, we aimed to compare the playing positions and compare the Brazilian players with high-level athletes from USA. Knee extensor (KE) and flexor (KF) muscles of 72 Brazilian players were assessed through isokinetic tests at 60°·s-1. KE concentric peak torque was 276±56 N·m, while KF had concentric and eccentric peak torques of 151±37 N·m and 220±40 N·m, respectively. Offensive linemen players presented greater peak torque values than defensive lineman, halfbacks, and wide receivers (all comparisons are provided in the article). Brazilian players had lower scores than USA athletes for KE and KF peak torque values. In addition, a conventional torque ratio (concentric/concentric) lower than 0.6 was found in 76-83% of athletes, and a functional ratio (eccentric/eccentric) below to 1.0 in 94%. Bilateral asymmetry greater than 10% was verified in 26% and 43% of athletes for KE and KF muscles, respectively. Elite players in Brazil present high incidence of strength imbalance in thigh muscles, and they are below USA players in relation to torque production capacity of KE and KF muscles.


INTRODUCTION
American football (AF), referred to as "football" by North Americans, is a team sport evolved in United States (USA) at late nineteenth century 1 .The game is played by 11 players each team, and playing positions are grouped in defensive positions (defensive backs, DB; linebackers, LB; defensive linemen, DL), offensive positions (quarterbacks, QB; offensive linemen, OL; wide receivers, WR; halfbacks, HB) and special positions (kickers, K; punters, P) 2 .Each playing position has specific demands during the game, thus successful players present specific characteristics in relation to body size and physical conditioning 3 .For instance, strength and power are imperative for line players (OL and DL) to perform tasks like charging, blocking and tackling; while speed, agility and maneuverability are the main characteristics for WR and DB 4 .
National Football League (NFL) is the main AF league in the world.Every year, selection of new players comprises the NFL scouting combine, a weak-long event where collegiate football players perform physical and mental tests in front of NFL coaches, general managers, and scouts.Athlete's performance during the combine can affect their draft status and salary, and ultimately their career.Since lower limb strength is a highlyprized quality in AF players, isokinetic dynamometry is part of the NFL scouting combine.Zvijac et al. 5 published a retrospective study with isokinetic data collected in 1,252 college players during NFL scouting combines.These normative isokinetic data for AF players may be used to assess the conditioning status and the risk for future injuries 5 .
Due to the game characteristics, contact with players is the primary injury mechanism in AF, and 47-57% of injuries affect the lower limbs 6,7 .Non-contact injuries represent 14% and 30% of injuries in competition and regular practice, respectively 6 .Among typical non-contact injuries, anterior cruciate ligament (ACL) ruptures (1.00 injury/1000 athlete-exposures) 6 and hamstring strains (1.67 injury/1000 athlete-exposures) 8 have been associated with some athletes' intrinsic features, such as the muscle strength imbalance between hamstring and quadriceps muscles (e.g., hamstring-toquadriceps ratio) 9,10 and strength imbalance between preferred and nonpreferred limbs (e.g., bilateral asymmetry) 10 .Thus, AF teams from USA have increasingly invested in isokinetic dynamometry tests 3 .
Although AF popularity is growing and reaching fans in more than a hundred countries 2 , studies outside the USA are scarce.As far as we know, there are only two studies about the incidence of injuries in Japanese 11 and Finish 12 athletes, and two studies about physical and performance characteristics of Japanese 13 and Brazilian 14 players.However, there is a lack of studies on the isokinetic performance of thigh muscles in AF athletes who play outside the USA.Therefore, the primary aim of this study was to describe isokinetic profile of knee extensor and flexor muscles in athletes who play in the elite level of AF in Brazil.In addition, we aimed to compare the playing positions and compare the Brazilian players with high-level athletes from USA.

Study design
This is a cross-sectional study.The study protocol was prospectively approved by the institutional ethics committee.Prior to data collection, athletes were informed about the procedures and signed the written informed consent.

Participants
The study's proposal was presented and approved by the board members and coaching staff of two AF teams based in Porto Alegre (Brazil) and engaged in the national elite championships.Assessments occurred in the time interval between the state championship (first semester) and national championship (second semester).Both clubs had a similar training routine, which comprised 3 weekly training sessions within the pitch.In addition, specific strength training programs were followed by players according to their playing positions.
To be included in the study, athletes should be > 18 years old, federated, and presenting a regular participation in the club's training routine for at least four months.Athletes could not present recent history of musculoskeletal injuries that could compromise their performance on the tests or recent history of any health conditions that had drawn them from their sport activities for > 30 days within six months prior to the tests.

Procedures
All tests were performed in a single day, within a room with temperature maintained at 23±1° C and proper lighting.The subjects were instructed not to consume stimulants (e.g., caffeine) within 12 hours prior to tests, not to engage in high intensity physical activity within 24 hours prior to the tests, not to drink alcoholic beverages and not to consume any kind of analgesic and/or anti-inflammatory drugs in the 48 hours prior to tests.

Anamnesis
Identification data, information on sports practice (playing position, experience, preferred lower limb) and health information (medication usage, history of injuries) were collected.

Anthropometric evaluation
Anthropometric measurements of body mass and stature were conducted using a precision scale linked with a stadiometer (Welmy, Brazil).The body fat percentage of athletes was assessed by the bioimpedance method through a BF-900 equipment (Maltron, UK).The four-pole measurement technique was used, and all standard procedures were respected 15 .

Isokinetic evaluation
After a 10-min warm-up exercise in a cycle ergometer with self-selected load and cadence, athletes were properly positioned in the isokinetic dynamometer Biodex System 4 Pro (Biodex Medical Systems, USA).Calibration of the equipment was performed previously to data collection and gravitational correction was performed for each leg prior to testing.To prevent the influence of tests' execution order, half of players initiated the testing protocol by the preferred leg while the other half initiated by the non-preferred leg.Athletes received standardized instructions about the testing protocol and performed a specific warm-up (20 submaximal repetitions at 180°•s -1 ).A protocol usually adopted in our laboratory was used for isokinetic assessment 16 .Briefly, volunteers performed two sets of three consecutive maximum contractions of flexion-extension in concentric-concentric mode and two sets in eccentric-eccentric mode at 60°•s -1 .A two-minute rest was fulfilled between each test, and subjects were verbally encouraged during tests to produce maximal strength.
The highest peak torque value obtained in each evaluation mode was selected for statistical analysis.Some players exceeded the maximum capacity of the isokinetic dynamometer in knee extensor eccentric contractions (i.e., 450 N•m).In view of this device limitation, this study does not present results of knee extensors eccentric peak torque.We calculated the conventional ratio (knee flexors concentric peak torque divided by knee extensors concentric peak torque), the functional ratio (knee flexors eccentric peak torque divided by knee extensors concentric peak torque), and the bilateral asymmetry (weaker leg / stronger leg -1).Although there is no consensus regarding the reference values, this study considered the benchmarks of 0.6 9,17 and 1.0 9,18 for conventional and functional ratios, respectively; and a cut value of 10% for bilateral asymmetry 17 .

Statistical Analysis
Data normality was analyzed by Shapiro-Wilk test.Comparisons between playing positions (except for QB and K/P due to the small sample size) were performed using One-Way ANOVA, followed by LSD post-hoc.A significance level of 5% (α<0.05) was adopted for all comparisons.
The isokinetic peak torques found in our study for each playing position (except for QB and K/P due to the small sample size) were compared with those reported by Zvijac et al. 5 in a study involving collegiate players assessed in the NFL scouting combine.The effect size (ES) between data from Brazilian players and North-American players was calculated through the Cohen's d.ES values were considered as: "trivial" (ES≤0.2),"small" (ES>0.2),"moderate" (ES>0.6),"large" (ES>1.2),or "very large" (ES>2.0) 19.

RESULTS
This study assessed 72 AF players (see players' characteristics in Table 1).Players acting in distinct playing positions had similar ages.OL presented higher body mass than all other positions; DL were heavier than HB, WR and DB; while LB were heavier than HB and DB.HB were shorter than OL, WR, DL and LB.OL had grater fat percentage than other positions (except DL), while DL were fatter than WR.Table 2 presents the players' peak torque values.OL obtained knee extensors concentric peak torques of both legs significantly higher than WR and HB, as well as higher than DB at the non-preferred leg.The knee flexors concentric peak torque in the preferred leg did not differ between playing positions, while the non-preferred leg of the OL reached significantly higher values than HB.The OL knee flexors eccentric peak torque was greater than DB in preferred leg and greater than HB in the non-preferred leg.
There were no differences between playing positions for conventional and functional ratios (Table 3).Conventional ratio was below the reference landmark of 0.6 9,17 in 76% of the preferred legs and 83% of non-preferred legs.Functional ratio below the cut point of 1.0 9,18 was found in 94% of both preferred and nonpreferred legs.Bilateral imbalances higher than 10% 17 were found in 26% of players for knee extensor concentric peak torque, as well as in 39% and 43% of players for knee flexor concentric and eccentric peak torques, respectively.
Figure 1 presents the comparisons between Brazilian elite players (assessed in our study) and collegiate players from USA (reported by by Zvijac et al. 5 ).Brazilian players were weaker than North-Americans in all playing positions for both knee extensor and flexor muscles, as well as had poorer conventional torque ratios.International literature provided no results about the knee flexors eccentric peak torque or functional ratio of AF players for comparison with results of the present study.

DISCUSSION
This study seems to be the first one to report the isokinetic profile of the knee extensor and flexor muscles of AF players outside the USA, the country that holds the hegemony of the sport.In short, our findings suggest that: (1) thigh muscle strength levels are different among the playing positions of AF; (2) Brazilian players present high incidence of muscle strength imbalances; and (3) Brazilian players are below the North-Americans in relation to torque production capacity of thigh muscles.Physical biotype is one of the determining factors for high performance in AF 4,[20][21][22] .Therefore, call attention the dissimilar body mass and stature found in Brazilian players compared to collegiate 21,23 and professional athletes 20,22 from USA.An interesting study developed by Anzell et al. 21has documented the changes in body proportions of collegiate and professional players of AF in USA throughout the years 1942 to 2011.The average stature of Brazilian players (1.79 m) is similar to the average stature of collegiate players at 1960's, and this low stature has never been observed among NFL athletes 21 .The average body mass of Brazilian players (94 kg) is near to the values from 1980's and 1950's of collegiate and professional athletes, respectively 21 .In other words, these data suggest that the Brazilian players still need to evolve their body size to equate the world elite AF players.In addition, Brazilian players seems to have a higher fat percent than North-American players, but careful is recommended on this conclusion because North American studies used skin folds 22 , radiographic absorptiometry dual energy 23 and plethysmography displacement air 20 to assess the players' fat percent, while we used the bioimpedance technique.
The smaller body size represents a possible reason why players engaged in the elite competition level of Brazil had lower torque production capacity than collegiate players in USA 5 .However, the dissimilar level of conditioning may be the main factor behind these findings.AF is an amateur sport in Brazil, and even players from elite teams usually train only 3-4 times per week.Brazilian teams do not have specific facilities.Teams usually train in fields within public parks or fields borrowed by the town hall or universities, and depend of commercial partners to assess gyms and medical support.On the other hand, players acting at the National Collegiate Athletic Association (NCAA) has a semi-professional routine.Universities provide all facilities (training fields, gyms, etc.) and teams have specialized coaches, conditioning trainers and a full medical staff to develop their players.These specific scenarios should be taking in account when AF players from USA are compared with those ones from other countries.
Prospective studies have demonstrated that strength imbalances between knee flexor and extensor muscles significantly increases the risk of hamstring strains in Australian football 24 and soccer 10 athletes, as well as potentially contributes to knee instability and predisposes to knee sprains 9 .Due to the high incidence of injuries in knee joint and hamstring muscles in AF [6][7][8] , strength balance between the thigh muscles is an important point when it comes to injury prevention.In this sense, our findings revealed that more than 3/4 of Brazilian players had a deficient conventional ratio, and this prevalence increased to 94% when considering the functional ratio, suggesting these players have weak hamstring muscles.Interestingly, the USA collegiate players 5 have average values of conventional ratio above 0.6 in all playing positions (Figure 1-C), suggesting that their training system seems to lead to musculoskeletal adaptations that draw players from the risk zone.Unfortunately, the lack of results in the literature involving the functional ratio of North American athletes leaves open the question about their knee flexors eccentric capacity, while our results demonstrated a notably deficiency in knee flexors eccentric torque of Brazilian athletes.
The high number of athletes below the recommended standards for conventional and functional ratios suggests that efforts must to be made for balancing the torque production capacity of the knee flexor and extensor muscles and help to prevent future injuries.Petersen et al. 25 and Seagrave III et al. 26 have implemented training programs to restore muscle balance between the anterior and posterior thigh muscles on soccer and baseball players, respectively.Their training protocols involving the Nordic hamstring exercise have been proven effective in preventing primary and recurrent hamstring injuries 25 , as well as in reducing the total number of days missed due to injury 26 .Therefore, Nordic hamstring exercise could be recommended as an interesting choice for athletes with muscle imbalance assessed in our study.
Strength balance among right and left limbs is related to sports performance and injury risk 27 .The considerable number of athletes with a bilateral asymmetry higher than 10% is perhaps related to the higher demands placed on one of the lower limbs during the sport practice.An example is the initial action of the defensive and attack players, when one leg (right or left, according to individual preference) is used for the initial impulse of the race against the opponent or towards the opposing field in every move of the match.Compensation strategies should be implemented in the weakest to restore the strength symmetry on the right and left legs.
Like other studies with AF, the low number of QB and K/P may be considered a limitation of this study.This small sample size prevents a reliable statistical comparison with players from other playing positions, as well as isokinetic comparison with QB and K/P assessed by Zvijac et al. 5 .On the other hand, a positive aspect of this study was the total sample size (72 players), which is higher than most studies comprising isokinetic evaluation of AF players.Future investigations should verify how much the strength imbalances reported here affect the incidence of injuries in AF players.

CONCLUSION
AF elite athletes in Brazil have specific isokinetic performances of knee flexor and extensor muscles according to their playing position, and Brazilian players have clear disadvantage compared to elite university players from USA.The literature gap regarding AF played around the world precludes a comparison with athletes from other countries, but our findings represent a new benchmark for leagues outside the USA.Additionally, a considerable number of players presented hamstring-to-quadriceps strength imbalance and/or bilateral strength asymmetry.As a practical application, these findings denote that specific training programs should be designed and implemented to improve strength and restore the muscle strength balance in lower limbs of AF players from Brazil.

Table 1 .
Characteristics of athletes in each playing position.Results expressed as minimum and maximum values for QB and K/P, and mean±SD for other playing positions.

Table 2 .
Concentric (CON) and eccentric (ECC) peak torques of knee extensors (KE) and knee flexors (KF) in preferred and non-preferred limbs (data presented in Nm).Results expressed as minimum and maximum values for quarterbacks QB and K/P, and mean±SD for other playing positions.

Table 3 .
Conventional ratio (CR) and functional ratio (FR) in preferred and non-preferred limbs.Results expressed as minimum and maximum values for QB and K/P, and mean±SD for other playing positions.