Leg strength and power are important physical attributes for soccer, for competent skill execution (Cabri et al., 1988; Cometti et al., 2001) and injury prevention (Henderson et al., 2010, van Beijsterveldt et al., 2013). In particular, left-right leg strength/power asymmetries are commonly associated with increased injury risk in soccer (Croisier et al., 2002; Knapik et al., 1991). Measurements of leg strength, power and asymmetry can be an important tool to assess an athlete's physical ability, and to monitor changes that occur with training or detraining.
For practitioners to have confidence in any test, results attained from it must be considered reliable and have established ranges of meaningful change (expected natural variation around a test, needed to establish if longitudinal change can be considered 'real').
The reliability of leg strength and power assessment protocols has been reported extensively. Various studies investigating reliability of double leg jumps including unresisted and resisted countermovement jumps (Nuzzo et al., 2011,Young et al., 1997), squat jumps (Ortega et al., 2008) and broad jumps (Ortega et al., 2008, Wiklander and Lysholm, 1987) have reported test-retest coefficients of variations (CV) between 1.8%-6.0% and intraclass correlation coefficient (ICC) values between 0.88-0.93, with all studies suggesting that double leg jump tests show 'good' or 'excellent' test-retest reliability.
As the majority of movements in sport (e.g. running, cutting and kicking) involve single leg loading (Fousekis et al., 2010, Reilly 1996), to improve specificity of testing, single leg jump testing has also been utilised across the literature as a lower limb strength and power assessment protocol. As such, various single leg hop tests such as horizontal distance (Bandy et al., 1994; Paterno and Greenberger, 1996), vertical hop and triple hop for distance (Munro and Herrington, 2011) have also been assessed for reliability. These studies consistently show 'good' reliability, with ICC values between 0.76-0.96 and no significant differences between any repeated tests. Test-retest reliability of left-right leg asymmetries obtained through single leg hops has been reported to be 'good' (ICC > 0.81) in some studies (Hopper et al., 2002; Reid et al., 2007), although Risberg et al (1995) found that a 21% change in performance was needed to establish significance due the variation they found between repeated tests (7.7% coefficient of variation-CV). Differences in findings may be due to large learning effects associated with single leg jumps (Bogla and Keskula, 1997; Booher et al., 1993) suggesting that extensive familiarisation to the movement technique may be required particularly for adolescent athletes. Additionally, due to the necessity to bear load through hip, knee and ankle joints, the use of double and single leg jumps as a measure of lower limb power can be limited with load compromised individuals and alternative offloaded protocols with greater control of movement are useful for testing in the elite environment.
Isokinetic dynamometry is also commonly used to assess lower limb strength and power and has consistently been found to elicit 'moderate' to 'excellent' test-retest reliability with ICC values between 0.71-0.99 tested over a range of velocities and muscle actions (Abernethy et al., 1995; Gleeson and Mercer, 1992; Li et al., 1996; Pincivero et al., 1997). However, studies assessing the reliability of left-right leg strength/power asymmetries obtained through isokinetic dynamometry have found weaker reliability with ICC values ranging between 0.29-0.78 and standard error of mean (SEM) between 3.2% to 8.7% (Impellizzeri et al., 2007, Impellizzeri et al., 2008). Isokinetic dynamometry also lacks applicability to sporting movement (Cometti et al., 2001) and has low correlation with other sports performance measures (Mognoni et al., 1994). Therefore, isokinetic dynamometry may not be an applicable lower limb assessment tool for soccer players.
The Keiser Air 420 seated leg press (Keiser Corporation, Fresno, CA) is a pneumatic resistance-based seated leg press machine with the left and right footplates that move independently of each other. Movement from a seated position with feet elevated enables offloaded maximal strength and power testing utilising movements that may be considered more applicable to the sporting environment than isokinetic dynamometry, overcoming many of the aforementioned issues. However, to date, there has only been one study investigating the reliability of the Keiser Air 420 (LeBrasseur et al., 2008) which found a non-statistically significant increase in maximum resistance of 1.1% and 'excellent' test-rest reliability (ICC:0.990) between 2 trials. However, reliability was not established for single or double leg power values or for left-right leg asymmetries, and as participants were classified as healthy males (age range: 37-70 y), the findings cannot be generalizable to an elite soccer population.
Therefore, the current study aimed to establish the reliability of double leg maximal strength, single leg and double leg power and left-right power asymmetries in elite soccer players using a seated leg press. Additionally, the study aimed to quantify the magnitude of change between tests that can be confidently established as outside of the range of natural variability of the test.
Experimental approach to the problem
The current study used a repeated test-retest protocol in which participants undertook an incremental resistance leg press test on three separate occasions to assess for variation between tests. Each test was completed at least 72 hours following and within 192 hours of the individuals' previous test (mean interval of 132 [+ or -] 43 hours). Tests took place prior to a training session, occurred at the same time of day, following a day consisting of minimal or no physical stimulus (
Thirteen elite male professional soccer players (Age: 18.4 [+ or -] 0.8 y, height: 1.79 [+ or -] 0.09 m, weight: 72.1 [+ or -] 6.7 kg, body fat: 9.6 [+ or -] 1.4%, VO2 max: 58.1 [+ or -] 1.3 ml x [kg.sup.-1][min.sup.-1]) volunteered for this study. A small sample size was considered acceptable due to the elite level of participants and is in line with previous research conducted within the research field (Bandy et al., 1994; Li et al., 1996). All participants were playing full time academy soccer for the same premier league football club for a minimum of 9 months prior to testing. Inclusion criteria were that they were injury free (defined as a 'time loss injury' as classified in a consensus statement on injuries within soccer (Fuller et al., 2006)) for the duration of the testing period and that they had previously completed a minimum of three exposures to the testing protocol. All participants were regularly participating in soccer training sessions and lower limb gym strengthening sessions between tests and all were exposed to similar physical stimulus over this period. The study obtained ethical approval from the School of Health Research Ethics Approval Panel at University of Bath and all participants were informed of the potential benefits and risks of the research prior to providing informed consent. For participants under the age of 18 (age range: 16.8-19.5 y), parental informed consent was obtained also.
Prior to testing, participants completed a 5-minute standardised warm up at approximately 60-75 Watts on a cycle ergometer (Keiser M3+, Keiser Corporation, California) followed by ten controlled bodyweight squats and three countermovement jumps. The testing procedure, outlined in Table 1 using example resistances, involved completing an incremental leg press test from a seated position (approximately 90o knee flexion) with feet flat on each footplate beginning at low resistance and continuing until failure, with 'max resistance' pushed defined as the final load that could be moved to full knee extension with both legs whilst maintaining proper seating position (McDonagh and Davies, 1984). Resistance reached on the 10th repetition and resistance increments in-between each repetition were set dependent on participant's results from previous exposure tests and remained the same for all three trials. However, as participants were encouraged to work until max resistance was...