Cross-country (XC) skiing represents one of the most

Transcription

1 Evaluation of an Upper-Body Strength Test for the Cross-Country Skiing Sprint THOMAS STÖGGL 1,2, STEFAN LINDINGER 1,2, and ERICH MÜLLER 1,2 1 Department of Sport Science and Kinesiology, University of Salzburg, Salzburg, AUSTRIA; and 2 Christian Doppler-Laboratory, Biomechanics in Skiing, Salzburg, AUSTRIA ABSTRACT STÖGGL, T., S. LINDINGER, and E. MÜLLER. Evaluation of an Upper-Body Strength Test for the Cross-Country Skiing Sprint. Med. Sci. Sports Exerc., Vol. 39, No. 7, pp , Purpose: The scope of the study was (a) to develop a test concept for specific upper-body power and strength diagnostics of cross-country (XC) skiing sprint athletes, (b) to check test reliability and validity, and (c) to test the hypothesis that maximal power, explosive strength, and power endurance predict double-poling (DP) sprint performance over race distance. Methods: Nineteen elite XC skiers performed test retest of the two-phase test (2PT) on a rollerboard, with the four-repetition maximal test (4R max T) as phase 1 and the 40-repetition test (40RT) as phase 2, both for determination of specific upper-body power and explosive strength. To check validity, 31 subjects performed the 2PT and a DP sprint test for 50 m, and a subgroup (N = 19) also performed a DP maximal-speed test and a 1000-m DP sprint test, both on a treadmill. Results: The 4R max T was highly reliable (r = , P G 0.001), except for explosive force and time to peak force. The 40RT was highly reliable for all variables concerning velocity and power (r = , P G 0.001). Peak lactate showed only low reliability (r = 0.69, P G 0.01). The peak values (maximal power, peak velocity, etc.) measured in the 4R max T contributed to up to 84% of the variation in 50-m DP sprint time and up to 61% of the variation in 1000-m performance. Moderate to high correlations in 1000-m DP sprint performance were found between the mean values and the fatigue indices of the 40RT. Conclusions: The 2PT is a reliable, valid, short-lasting test. The relationship between maximal power output (measured in the 4R max T) to 50- and 1000-m sprint performance suggests increasing the proportion of training aimed at the improvement of specific explosive strength and maximal power. Key Words: DOUBLE POLING, MAXIMAL SPEED, POWER, PREDICTORS, ROLLERBOARD Cross-country (XC) skiing represents one of the most demanding endurance sports: compared with athletes of other sports, XC skiers are in the lead concerning maximal aerobic power (23) (see Table 1 for all article abbreviations). However, in recent decades, several studies have shown potential for improving XC skiing performance by increasing strength and power of the upper body. Additionally, a close relationship between aerobic and anaerobic upper-body capacity and XC skiing performance has been found (9,11,12,25). Owing to the recent introduction of the sprint disciplines, the importance of strength, power, and, therefore, the development of appropriate testing concepts, have increased. To date published investigations regarding the XC skiing sprint are rare. Stöggl et al. (26,28) developed test concepts for the XC skiing sprint and analyzed a simulated sprint competition in the classical style, using roller skiing on a Address for correspondence: Thomas Stöggl, Department of Sport Science and Kinesiology, Rifer Schlossallee 49, 5400 Taxach/Rif; thomas. stoeggl@sbg.ac.at. Submitted for publication September Accepted for publication February /07/ /0 MEDICINE & SCIENCE IN SPORTS & EXERCISE Ò Copyright Ó 2007 by the American College of Sports Medicine DOI: /mss.0b013e treadmill. Sprint performance in the classical style was strongly dependent on the maximal speed in the single techniques, fatigue resistance, technical skills, and anaerobic capacity, whereas maximal oxygen uptake (V O 2max ) showed no correlation. The fastest sprint skiers were able to produce longer cycle lengths at an optimal poling frequency and, therefore, needed fewer pole plants during a sprint heat. Hence, faster athletes could produce more propulsion at equal poling frequency. Holmberg et al. (7) have illustrated that the fastest athletes used a more sprinter-like double-poling (DP) strategy, characterized by higher peak pole forces and higher impulses of pole force. All of these aspects suggest that the difference between faster and slower skiers is based on different strength capacities, especially in the upper body, and on technical aspects. Nilsson et al. (13) found a strong correlation between peak and mean power output during a 30-s test and also between the mean power output during the 30-s test and a 6-min test, all performed on a DP ergometer. The transfer into DP performance on snow or roller skis was not confirmed. Nevertheless, these findings coincide with the strong relationship of sprint performance with maximal speed in the diagonal stride and DP, as shown by Stöggl et al. (26,28). Comparable results have been found in several studies on running ( m), stressing the crucial role of so-called muscle power factors, tested during a maximal anaerobic running test (MART) and short-term 1160

2 TABLE 1. List of abbreviations, in order of first use in text. 2PT 4R max T 40RT ACC CV DP DPI DP 50m test DP Vmax test DP 1000m test F exp FI FI 1000 FI Vact,FI Vpeak,W P act PF t 50 t PF t Vpeak V max V 1000 V act V act40,v peak40,w V con V peak XC skiing sprint tests for m, for performance (17,20 22). The question remains whether the strength level measured (either generally or specific) is a parameter for discriminating better-performing XC skiing sprinters from slower ones, performing on snow or roller skis. Several concepts for specific standardized testing and training for XC skiing have been developed (1 4,11 13, 18,25 28). For the most part, incremental tests to exhaustion were performed, to measure upper-body aerobic and anaerobic performance. Only in a few studies was specific maximal upper-body strength measured (3 5,12,18). Either by use of a cable pulley or Vasa Trainer, one-repetition maximum (3 5,18) or 10-repetition maximum (12) was determined. Stöggl et al. (27) introduced and biomechanically validated a simple, low-cost rollerboard concept for the simulation of DP, illustrating the possibility of measurement and training of specific upper-body explosive strength, maximal power, and power endurance. The specific aims of the present study were 1) to develop a rollerboard test concept for determination of specific upper-body explosive strength, maximal power, and power endurance for the XC skiing sprint, 2) to examine the reliability of the test concept, and 3) to check the validity by testing the hypothesis that XC skiing specific power and strength determine maximal DP speed and sprint performance over race distance. MATERIALS AND METHODS Subjects Two-phase test on the rollerboard Four-repetition maximal test on the rollerboard 40-repetition test on the rollerboard Peak acceleration during active phase on the rollerboard Coefficient of variation Double poling Double-poling imitation on the rollerboard 50-m double-poling maximal-speed test on the track Double-poling maximal-speed test on the treadmill 1000-m double-poling sprint test on the treadmill Explosive force, as the rate of force development Fatigue index Fatigue index in the 1000-m double-poling test Fatigue index in the 40RT, calculated for the single variables Mean power output during active phase on the rollerboard Peak strap force on the rollerboard Time for the 50-m double-poling sprint Time to peak force Time to peak velocity on the rollerboard Maximal speed during the DP Vmax test Mean velocity in the 1000-m test Mean velocity during the active phase on the rollerboard Mean value for the 40 repetitions in the 40RT, calculated for the single values Mean velocity calculated for total upward movement of rollerboard (active and passive rollout phase) Peak velocity during the active phase on the rollerboard Cross-country skiing A total of 31 elite XC skiers of the Austrian, Slovakian, and Swiss national and student national teams (25 men and six women, age 26 T 5 yr, height 180 T 10 cm, weight 72 T 10 kg) volunteered as subjects in the study, which was approved by the ethics committee of the University of Salzburg. All skiers were fully acquainted with the nature of the study before they gave their written informed consent to participate. Rollerboard Testing Concepts Specification of the DP imitation on the rollerboard. The testing device was a rollerboard (weighing 7.5 kg) with an attachment allowing a convenient and stable kneeling position of the athletes (Fig. 1). The rollerboard can be moved up and down on a steel rail inclined to wall bars (at 16- for the used tests). The athletes pulled themselves towards the elevated end of the rail with the use of two straps secured at the upper part of the wall bars, 1.5 m above the end of the rail (strap angle to horizontal), simulating the DP technique. The straps were connected to two pole grips. The DP imitation (DPI) was investigated for its biomechanical validity; high similarities to DP, except for longer cycle time, a more extended angle of the elbow at the start of the cycle, and higher activation of the biceps brachii in DPI, were found (27) (Fig. 1). Two-phase test. The two-phase test (2PT) was developed in accordance with the 2PT of Martin et al. (10) on a bench pulling device; it also follows the principles of the Wingate anaerobic test (8). Before testing, a standardized warm-up was performed. It included 15 min of running, 10 DPI repetitions at moderate speed, and five repetitions at maximal speed. The first phase of the 2PT was the four-repetition maximum test (4R max T), consisting of four repetitions at maximal possible movement velocity during the active phase of DPI. For every measured variable, the highest value of the four repetitions was recorded as maximal value during DPI. After a break of 3 min, the subject performed the second phase, the 40-repetition test (40RT), where the athlete had to perform 40 repetitions with maximal exercise speed at each repetition, simulating the sprint heat duration (2 3 min). No break was allowed between the repetitions. Calculated variables in the 2PT. The DPI cycle was divided into three phases. In the first phase, the athlete generates propulsion and then moves upward on the rail (active phase). The active phase started when velocity became positive (upward) and ended when the force in the straps fell below zero. The second phase of the upward movement was called the passive phase, according to the roll out of the board, and ended when the velocity became negative (start of roll downward) (Fig. 2). In the third phase (roll-down phase), the whole system was lowered and actively decelerated back to the starting position. The following variables were calculated: mass (m), represented by the downhill-slope force calculated as mass of the whole system (subject + rollerboard) times sin (16-), whereas 16- is the inclination of the rail in the test; mean velocity during STRENGTH TEST FOR CROSS-COUNTRY SPRINT SKIING Medicine & Science in Sports & Exercise d 1161

3 FIGURE 1 Picture series of the double-poling imitation on the rollerboard compared with double poling on roller skis. active phase (V act ); peak velocity (V peak ); time to peak velocity (t Vpeak ); peak acceleration (V exp ), calculated as V peak /t Vpeak ; mean velocity for the whole concentric phase (active phase and passive phase) (V con ); mean mechanical power during the active phase (P act ), calculated as m ( V act /t act )V act, where t act is the duration of the active phase; peak strap force (PF); time to PF (t PF ); rate of force development (F exp ) calculated as PF/t PF ; impulse of strap force (IF). For the 40RT, mean values for the 40 repetitions in the single variables (V act40, V peak40, V con40, P act40, and PF 40 ); a fatigue index (FI), calculated as the percent difference between the mean values in the 40RT and the maximal values in the 4R max T(e.g.,FI Vact =(V act j V act40 ) 100/V act [%]); peak heart rate during test (HR); peak lactate (LA) as the maximal value in the first, third, fifth, and seventh minutes after the end of the test; and test time (t). Data collection and proceeding. Force and velocity parameters during DPI were collected by a complete measurement system consisting of an input box connected to A/D converter cards (DAQ 6024E A/D card-12 bit, National Instruments), strain gauge force transducers (Ergotest Technology, Langesund, Norway) integrated at the top of the straps, a linear encoder (Ergotest Technology, Langesund, Norway, Typ ET-Enc-01, resolution G mm, 1000 Hz) for velocity measurements, a laptop (IBM T42, New York, NY), and specially developed measurement software. Heart rate was measured by the use of a heart rate monitor (Polar S610, Polar Electro OY, Helsinki, Finland). Blood samples (20 KL) were taken from the earlobe before and in the first, third, fifth, and seventh minutes after the ends of both tests, for later determination of peak lactate (Biosen 5140, EKF-Diagnostic GmbH, Magdeburg, Germany). Treadmill and Field Tests on Roller Skis The tests on roller skis in the field and on the treadmill were used to test validity of the roller board concept. They have already been described in detail and evaluated on reliability and validity by Stöggl et al. (28). Hence, just a FIGURE 2 Definition of active and passive phase when using the double-poling imitation on the rollerboard. Dashed line, strap forces; solid line, velocity during active phase; dotted line, velocity during passive phase. PF, peak strap force; V peak, peak velocity Official Journal of the American College of Sports Medicine

4 brief overview is given in the following section. Roller skis in all the tests were Pro-Ski C2 (Sterners, Nyhammar, Sweden), with every subject using the same pair of roller skis, which were warmed up before beginning the testing session by roller skiing for 20 min on the treadmill or track, respectively, to prevent a warm-up effect of the wheels and bearings during testing. Subjects used their own poles for the classical technique (pole length: 82 T 1% of body height). All treadmill tests were performed on a large treadmill (Pomer, Wiege Data, Leipzig, Germany, belt dimensions of m) on which roller skis could easily be used. The belt of the treadmill consisted of a nonslip rubber surface, allowing subjects to use their own poles with special carbide tips. Each subject was familiar with roller skiing on the treadmill at high speeds from numerous training and testing sessions for a period of at least 2 months before this study. Before the start of each treadmill test, the athletes were secured with a safety harness, which was connected to an emergency brake suspended from a metal bracket above the treadmill. Heart rate was measured, and blood samples from the earlobe were taken in the first, third, fifth, and seventh minutes after the ends of both tests, for later determination of lactate. DP maximal-speed test on the treadmill. The DP maximal-speed test on the treadmill (DP Vmax ) served for determination of maximal DP speed on the treadmill. All subjects performed a standardized warm-up program of 15 min using the test roller ski. Test started with a 30-s stage of medium intensity (4 mis j1 ), and treadmill speed was then increased to 7 mis j1 with linear acceleration of 0.2 mis j2. Speed was then increased every 5 s, using increments of 0.3 mis j1. The test was stopped by the security button at a velocity where the subject could no longer maintain the roller ski_s front wheels ahead of a marker placed 1.5 m from the front of the treadmill. Maximal speed at the stop of the test (V max ) was recorded m DP test on the treadmill. Ten minutes after the end of the DP Vmax test, subjects continued with the 1000-m DP test on the treadmill (DP 1000m ). Initial speed of the treadmill was set at 7 mis j1, with an acceleration of 2.5 mis j2 at the start. Using a speed-control device (Wiege-Data, Leipzig, Germany), treadmill speed could be regulated by the athlete. At the end of the treadmill, a linear encoder was positioned, whereas a line fixed at the hip measured the position of the athlete with respect to the treadmill. If the athlete moved forward, treadmill speed increased, whereas it decreased when the athlete moved towards the rear end of the treadmill. Acceleration of the speed-control device was set at 2.5 mis j1 for each meter. Similar to all-out test concepts (e.g., Wingate test), athletes had to DP as fast as possible from the beginning and had to maintain maximal possible DP speed throughout the 1000 m. Time for the 1000 m (t 1000 ), mean velocity for the 1000 m (V 1000 ), maximal velocity for the 1000 m (V max1000 ), FI for the 1000 m (FI 1000 ; calculated as FI 1000 =((V max1000 j V 1000 ) 100/V max1000 ) (%)), lactate, and heart rate were recorded. 50-m DP maximal-speed test. The 50-m DP maximal-speed test (DP 50m ) was performed on a flat, straight, indoor tartan track of 80 m, and only the use of the DP technique was allowed. Fifty-meter time (t 50 ) was measured by fixed light sensors (ALGE-TIMING, Lustenau, Austria). Starting position was standardized by positioning the roller ski with the front wheel at the starting line and placing the poles on the ground. At the start, no push-off with legs was allowed. Time measurement started when the athlete passed through the first light sensor installed at the starting line (height of the athlete_s lower leg) and ended when he passed the 50-m mark. After a 20-min warm-up on roller skis on the track, each subject had two trials for the 50 m. Reliability- and Validity-Testing Procedures Nineteen subjects (14 men and five women) performed test retest of the 2PT to check reliability. Time between test and retest was approximately 1 wk, and tests were performed at the same time of day. Thirty-one subjects (25 men and six women) performed the 2PT and the DP 50m to test the validity of the 2PT as an indicator for shortduration maximal DP speed. Nineteen subjects (15 men and two women) performed the 2PT, the DP Vmax, and the DP 1000m, to check for validity of the 2PT as a predictor of DP sprint performance over race distance. Statistical Analysis All data were checked for normality, calculated with conventional procedures and presented as means and standard deviations (T SD). In addition, coefficient of variation (CV), using the equation CV = SD/mean 100 [%], was used for reliability purposes. Individual CV were calculated for each subject and were averaged within the tests to obtain an overall CV. Differences between test and retest were assessed with a paired-sample t-test, which was done to determine whether learning effects between tests occurred. Reliability for each variable in the test retest and relationships between the variables of the different tests were examined by pairwise comparisons using Pearson product moment correlation coefficient tests. The r values were categorized as follows: excellent, 0.9 1; high, ; moderate, ; and low, G 0.7. The statistical level of significance was set at P G 0.05 for all analyses. All statistical tests were processed using SPSS 12.0 software (SPSS Inc, Chicago, IL) and Office Excel 2003 (Microsoft Corporation, Redmond, WA). RESULTS The descriptive data of the subjects are shown in Table 2. t 50 in the DP 50m was 8.38 T 0.58 s. V max in the DP Vmax was 8.14 T 0.93 mis j1. t 1000 was T 22.9 s, with a mean velocity of 6.1 T 0.7 mis j1. Test duration for the 40RT was STRENGTH TEST FOR CROSS-COUNTRY SPRINT SKIING Medicine & Science in Sports & Exercise d 1163

5 TABLE 2. Subject characteristics (N = 31). Age (yr) 26 T 5 Height (cm) 180 T 10 Mass (kg) 72 T 10 Peak LA (mm) 13.0 T 2.1 Peak HR (bpm) 181 T 5 V maxdp (mis j1 ) 8.2 T 0.9 t 50 (s) 8.38 T 0.58 Values are means T SD for 31 subjects. Peak LA, peak blood lactate after DP 1000m test; peak HR, maximal heart rate during DP 1000m test; V maxdp, double-poling maximal speed measured in the DP Vmax test; t 50, time for the DP 50m test. 124 T 3 s, with a DPI frequency of around 0.32 Hz. Peak lactate in the DP 1000m ranged from 9.2 to 16.4 mm (13.0 T 2.1) and in the 40RT from 5.4 to 9.3 mm (7.1 T 1.2); it was significantly higher (P G 0.001) in the DP 1000m compared with the 40RT. Peak heart rate during the DP 1000m was higher (P G 0.01) compared with the 40RT (181 T 5vs172T 12 bpm). FI 1000 was 24.0 T 5.9%. FI in the 40RT were 10.4 T 3.4, 12.7 T 4.5, 10.3 T 3.5, 13.6 T 4.1, and 13.8 T 5.1% for FI Vact,FI Vpeak,FI Vcon,FI Pact, and FI PF, respectively. All FI values measured in the 40RT were lower (P G 0.001) compared with the FI for the DP 1000m (FI 1000 ). Reliability of 2PT. Reliability statistics of the 2PT are shown in Table 3. No statistical differences for any of the measured variables were found between test and retest. Test retest correlations for the 4R max T ranged from r = 0.77 to 0.99 (P G 0.001). Lowest correlations were found for the variables concerning time to peak force and explosive force, whereas all variables related to velocity and power showed excellent correlations. Overall CV ranged from 1.35 to 6.82% for all variables of the tests, again showing the highest values for time to peak force, explosive force, but also impulse of force. The 40RT showed excellent reliability for all measured variables related to velocity, power, and force (0.92 G r G 0.99; P G 0.001). Peak lactate showed low reliability (r = 0.69, P G 0.01). Overall CV ranged from 1.10 to 6.96% for all measured variables. Test retest correlation for the FI variables related to velocity and power ranged from r = 0.77 to 0.92 (P G 0.001), showing the highest value for FI Vpeak. FI PF showed only low reliability (r = 0.54, P G 0.05). Overall CV ranged from 8.06% (FI Vpeak ) to 23.18% (FI PF ). Validity of the 2PT. Detailed information about all the correlations between the single variables of the 2PT to the tests on roller skis is presented in Table 4. In the following section, only the highest values are presented. Validity of 4R max T. An excellent correlation to t 50 was found for V peak (r = j0.92, P G 0.001), and there were high correlations for V act and P act (both r = j0.81, P G 0.001), accounting for up to 84% of the total variation in the DP 50m (Fig. 3). A low correlation between the length of the passive roll-out phase on the rollerboard and DP 50m sprint performance was found (r = j0.55, P G 0.001). P act highly correlated (r = 0.81, P G V 0.001), and V act, V peak, and V con moderately (r = , all P G 0.001) correlated to V max, as measured in the DP Vmax on the treadmill. V act, V peak, and P act showed moderate correlations to DP 1000m sprint performance (r = , all P G 0.001). P act accounted for 61% of the variation in the DP 1000m sprint test (Fig. 4). Moderate correlations to the FI in the DP 1000m were found for V act (r = j0.71, P G 0.001) and P act (r = j0.73, P G 0.001). No correlations to any of the variables measured in the tests on roller skis were found for the variables of time to peak velocity, peak acceleration, or time to peak force. Low to moderate correlations were found between the measured values of the active phase (V act, V peak, P act ) and the distance of the passive roll-out phase (r = , P G 0.001). Validity of 40RT mean values. High correlations to the DP 50m maximal speed were found for V peak40, V con40, and t (r = j0.80 to j0.84, all P G 0.001). High correlations to V max were found for V act40, V peak40, V con40, and test duration of the 40RT (r = , all P G 0.001). V peak40 showed the highest correlation to V 1000 (r = 0.83, P G 0.001), whereas just moderate correlations were TABLE 3. Reliability values (N = 19) of the two-phase test (2PT) on the rollerboard. 4R max T V act V peak t Vpeak V exp V con P act PF t PF F exp IF r CV RT Mean Values V act40 V peak40 V con40 P act40 PF 40 HR LA t r CV RT Strength Endurance Values FI Vact FI Vpeak FI Vcon FI Pact FI PF r * CV * P G 0.05; P G 0.01; P G N, number of subjects in the test retest; r, Pearson correlation coefficient; CV, coefficient of variation; 4R max T, four-repetition maximal test on the rollerboard; 40RT, 40-repetition test on the rollerboard; V act, mean velocity during active phase; V peak, peak velocity; t Vpeak, time to peak velocity; V exp, rate of velocity development calculated as V peak /tv peak ; V con, mean velocity calculated over the whole forward movement of the rollerboard; P act, mechanical power during propulsion phase; PF, peak strap force during propulsion phase; t PF, time to PF; F exp, rate of force development; IF, impulse of force (NIs); V act40, mean values for the 40 repetitions in the 40RT calculated for the variable V act ; HR, peak heart rate during test; LA, peak lactate after exercise; FI Vact, fatigue index calculated for the variable V act Official Journal of the American College of Sports Medicine

6 TABLE 4. Relationships between variables of the two-phase test to 50-m double-poling (DP) maximal-speed test (DP 50m ), DP maximal-speed test (DP Vmax ), and 1000-m DP maximal-speed test (DP 1000m ) performance. 4R max T V act V peak t Vpeak V exp V con P act PF t PF F exp IF t 50 (n = 31) j0.83 j j0.52 j0.78 j0.83 j0.69 j0.23 j0.35 j0.57 V max (n = 19) j V 1000 (n = 19) j j * 0.60 FI 1000 (n = 19) j0.71 j0.59* 0.28 j0.38 j0.48 j0.73 j0.57* 0.31 j0.58* j0.49* 40RT Mean Values V act40 V peak40 V con40 P act40 PF 40 HR LA t t 50 (n = 19) j0.74 j0.84 j0.80 j0.70 j0.42 j0.06 j0.38 j0.80 V max (n = 19) * V 1000 (n = 19) * j0.17 j FI 1000 (n = 19) j0.68 j0.67 j0.64 j0.68 j j RT Strength Endurance Values FI Vact FI Vpeak FI Vcon FI Pact FI PF t 50 (n = 19) j0.01 V max (n = 19) j0.70 j0.56* j0.22 j0.75 j0.04 V 1000 (n = 19) j0.79 j0.69 j0.40 j0.81 j0.25 FI 1000 (n = 19) * * P G 0.05; P G 0.01; P G N, number of subjects in the test retest; t 50, 50-m running time during DP 50m test; V max, maximal velocity during DP Vmax test; t 1000, 1000-m running time in DP 1000m test; FI 1000, fatigue index during DP 1000m test; V act, mean velocity during active phase; V peak, peak velocity; t Vpeak, time to peak velocity; V exp, rate of velocity development calculated as V peak /tv peak ; V kon, average velocity calculated over the whole forward movement of the rollerboard; P act, mechanical power during active phase; PF, peak strap force during active phase; t PF, time to PF; F exp, rate of force development; IF, impulse of force (NIs); V act40, mean values for the 40 repetitions in the 40RT calculated for the variable V act ; HR, peak heart rate during test; LA, peak lactate after exercise; FI Vact, fatigue index calculated for the variable V act. found for the other variables related to velocity and power. All variables except for PF 40 correlated to the FI in the DP 1000m (r = j0.64 to j0.68, P G 0.01). No correlations to any of the variables measured in the tests on roller skis were found for the physiological variables of heart rate and lactate. Additionally, no correlation between peak lactate and heart rate measured in the 40RT and in the DP 1000m was found. Among relationships between the single variables of the 4R max T to the respective variables of the 40RT, excellent positive correlations were found for V act, V peak, V con, P act, and PF (r = , all P G 0.001). Validity of 40RT FI. Low correlations to t 50 were found for FI Vact and FI Pact (r = 0.65 and 0.68, both P G 0.01). Moderate correlations to V max were found for FI Vact and FI Pact (r = j0.70 and j0.75, both P G 0.001). FI Pact FIGURE 3 Relationship between peak velocity measured in 4R max T on the rollerboard and 50-m time measured in the DP 50m on the track (t 50 ). *** P G showed a high correlation to V 1000 (r = j0.81, P G 0.001). Low correlations to the FI measured in the DP 1000m were found for FI Vact,FI Vpeak, and FI Pact (r = , all P G 0.01). Moderate to low correlations between the mean values of the 40 repetitions for V peak40, V act40, P act40, and their respective FI values were found (r = j0.58 to j0.75, all P G 0.01). Regarding relationships between the single variables of the 4R max T to the respective FI variables of the 40RT, low correlations were found for V act, V peak, and PF (r = j0.39 to 0.49, all P G 0.01). DISCUSSION The designed rollerboard and the performed DP imitation have already been investigated for biomechanical validity by Stöggl et al. (27). Especially with modifications to the rail inclination, adding extra weights and changing movement speed introduced the possibility to train and test upper-body maximal strength, power, and power endurance in a specific way. Together with the very reliable, valid 2PT, it is now possible to test specific upper-body maximal power, explosive strength, and power endurance of XC skiers with short test durations of only about 5 min. Reliability of the 2PT. According to Schabort et al. (24), a reliable performance test is one that has small changes in mean values, a small within-individual variation represented by CV, and a high test retest correlation in the measured variables. The strong correlation coefficients and the small, nonsignificant changes in the mean values between test and retest, as well as the small CV values, provide strong evidence for the reliability of the 2PT. With the exception of lactate, time to peak force, explosive force, STRENGTH TEST FOR CROSS-COUNTRY SPRINT SKIING Medicine & Science in Sports & Exercise d 1165

7 FIGURE 4 Relationship between maximal power output during the active phase measured in the 4R max T on the rollerboard and mean velocity measured in the DP 1000m roller skiing test on the treadmill (V 1000 ). *** P G impulse of force, and the FI calculated for the variable peak force, all measured variables of the 2PT showed excellent reliability. The low reliability of lactate (r = 0.69; CV = 6.96%) is in accordance with an investigation of Nummela et al. (16), arguing for low reproducibility of peak lactate in a MART (r = 0.6; CV = 19.2). Additionally, no real pattern between peak lactate and sprint performance was found. The moderate reliability of time to peak force and explosive force might be attributed to the calculation process of those variables by the used software. The variable of time to peak force was rounded up to two decimal places (hundredths of a second), and accuracy might, therefore, be too low. Consequently, the single values of the athletes may have had insufficient variety to discriminate. To improve the significance of these values, the measurement accuracy has to be changed in the future software version. The variables concerning the FI on the rollerboard were less stable compared with the calculated mean values in the 40RT. Only FI Vpeak showed excellent reliability (r = 0.92), followed by FI Vact and FI Pact, which showed high reliability (r = 0.86 to 0.89). Although correlation coefficients were moderate to excellent, these variables showed the highest CV values, ranging from 8.06 to 23.18%. The method to calculate the FI variables might be one explanation for the high variation of these variables. Two variables measured in separate test modes (4R max T and 40RT) were used in the calculation procedure. Thus, small variability in the single variables might cause higher variability in the calculation of FI. Another explanation might be that tactical behaviors of some athletes in the 40RT might have been different between test and retest. If an athlete does not exert maximally at the beginning of the test, the rate of power or velocity decrease should be lowered; consequently, the subject is probably able to produce more mean work in the 40RT (29). A solution might be to skip the 4R max T and take the maximal values at the beginning of the 40RT as peak values. Nevertheless, as proposed by Inbar et al. (8), the duration of the pending test might affect the effort at the beginning, and thus the peak values might be lower in the 40RT compared with the 4R max T. Validity of the 2PT. Validity of the 2PT was verified by comparing the test results on the rollerboard with the results of the tests on roller skis (DP 50m,DP Vmax, and DP 1000m ), which were found to be reliable, valid tests for sprint roller skiing (28). Excellent to high validity of the 4R max T was found by correlating t 50 of the DP 50m to the variables V peak, P act, and V act. V peak did account for up to 84% of the total variation in the performance of the DP 50m. Interestingly, the validity statistics of the 4R max T were lower when compared with the DP Vmax test on the treadmill, a test also diagnosing short-duration sprint abilities. The higher correlations of the 4R max T to the DP 50m, compared with the correlations to the DP Vmax, might be caused by 1) the higher number of subjects compared with the DP 50m, and 2) the fact that during the DP Vmax, only the maximal speed counts, whereas in the DP 50m, the starting technique, the acceleration phase, and the speed development for 50-m influence the test results. Hence, although there was a high correlation between t 50 and V max, as shown by Stöggl et al. (28), the prerequisites for performance in the DP 50m are more complex. The acceleration phase at the start should be closely connected to the ability to produce high force and power for propulsion, as measured in the 4R max T. Interestingly, the maximal values (variables V act, V peak, and P act ) of just four repetitions during the 4R max T accounted for up to 61% of the total variation of DP 1000m performance (V 1000 ; r = ), a variable standing for sprint performance over race distance. In comparison, the highest correlation between the 40RT and V 1000 was found for V peak40, accounting for 69% of the variation of the V 1000, a percentage that is not much higher than that for the 4R max T. This might be attributed to the excellent correlation (r = ) between the peak values in the 4R max T and the mean values recorded in the 40RT. Further determinants accounting for DP sprint performance might be certain technical aspects in DP, and the role of the lower body for DP performance, which was excluded from the measurement concept of the 2PT on the rollerboard. Summarizing the data, it can be shown that P act and V peak were closely connected to the maximal DP speed and DP sprint performance for 1000 m on roller skis. Hence, those variables seem to be the most stable and highest sprint predicting variables measured in the 2PT. The variable V peak stands for the ability to produce maximal movement velocity on the rollerboard during the active phase. In the calculation of P act, the parameters of mass, mean velocity during the active phase (V act ), and the time of the active phase (t act ) were used. Thus, to achieve a high P act, the athlete has to produce a high mean velocity during an optimally short active phase. Additionally, if an athlete of higher body weight produces the same V act in the same time as a lighter one, he or she consequently achieves higher P act values. These results highlight the hypothesized connection between DP sprint performance and the ability to produce 1166 Official Journal of the American College of Sports Medicine

8 high maximal movement speed in specific movement modes, strongly connected to maximal strength, explosive strength, and power. In summary, the high relationship of 4R max T variables such as peak velocity and maximal power output to 40RT, DP 50m, and DP 1000m sprint performance shows, on the one hand, the high specificity of the DPI (27), and on the other hand, that the 4R max T alone serves as a simple, reliable, and valid test concept for diagnostics of specific upper-body performance and also for DP sprint performance in XC skiing. Fatigue and the role of the lower body in DP. The measured FI values of the 40RT showed moderate correlations to their respective mean values in the 40RT, and moderate to high (FI Pact ) correlations to 1000-m sprint performance (V 1000 ). This aspect stresses the importance of a well-developed power endurance level of the upper body at high power output for sprint performance, thus coinciding with the prerequisites for performance in the DP 1000m suggested by Stöggl et al. (28). Interestingly, only moderate correlations were found for the FI values of the 40RT to the FI value of the DP 1000m. As already mentioned, one explanation might be the just moderate to high reliability of FI in the 2PT. Yet, when analyzing the measured physiological variables of peak lactate and heart rate, both were higher in the DP 1000m compared with the 40RT, indicating a higher metabolic loading when performing the 1000-m test using the DP technique compared with DPI on the rollerboard. It was observed that peak lactate did not correlate to performance in the 40RT or to performance in the DP 1000m. Interestingly, peak lactate and heart rate during the 2PT did not correlate with the lactate and heart rate values in the DP 1000m. Additionally, the FI values in DPI were all lower compared with FI 1000 (24 vs 10 16%). All those aspects might be explained by the fact that DPI on the rollerboard is restricted to upper-body work only, whereas the lower body is totally excluded. In XC skiing, the legs are the main energy consumers. Surprisingly, this is also true in DP (30), which is often looked on as primarily an upper-body exercise (6). Holmberg et al. (7) have shown that the legs play a critical role in DP performance. There was a marked electromyography activity in several of the lower-body muscles and substantial flexion extension movements in the hip, knee, and ankle joints. Another factor might be that although the arm, shoulder, and trunk muscles show similar activation patterns and activation height when comparing DP with DPI, cycle duration and, consequently, frequency is much longer or lower, respectively, in the DPI situation (27). The 40 repetitions were performed at a frequency of around 0.3 Hz, compared with a higher DP frequency of around 0.9 Hz at submaximal DP speed (7) up to 1.5 Hz at maximal DP speed (26). Thus, the athlete had a longer time of recovery between two active phases in DPI compared with DP. Interestingly, studies using the Wingate test (8) have revealed that athletes with higher peak power output achieved higher power output in the test but had a steeper fatigue curve slope, representing a weaker FI. It was found that athletes who specialized in events requiring high mechanical power output had a higher ratio of fast-twitch to slow-twitch muscle fibers compared with endurance athletes. The findings of previous studies (26,28) regarding the positive relationships between maximal speed in DP and DIAG and sprint performance partly coincide with those findings. In contrast, athletes who showed the highest maximal speed in DP and the highest power output in the 4R max T tended toward a better FI in the 40RT and the DP 1000m (Table 4). It might be speculated that DP at high skiing velocities is even more economic when using the sprinter-like DP technique, as shown by Holmberg et al. (7). This more explosive DP technique led to shorter relative poling times and longer relative recovery phases. Consequently, they made a shift toward shorter time of high activation and longer time for recovery, a technical aspect that might positively affect work economy in DP at high velocities. However, future studies are required to verify that speculation. Upper-body power and strength versus DP speed and performance. In a previous study (28), we have shown that V max in DP on roller skis was highly related to 1000-m DP sprint performance. Furthermore, in the simulated XC skiing sprint (26), V O 2max was rather low (r = 0.5; P ), whereas maximal speed in the DP and DIAG techniques again showed high to excellent correlations with classical sprint performance. These results are also supported by several studies in running, highlighting that maximal running velocity in the MART (V MART ), and/or short duration and maximal velocity (20 m), largely determine running performance for distances from 400 to 5000 m (17,20 22). Additionally, several studies (14,15,19) have concluded that endurance performance and peak treadmill running performance on level terrain were influenced not only by central factors related to V O 2 but also by so-called muscle power factors, which are also related to anaerobic and, especially, neuromuscular characteristics (e.g., voluntary and reflex neural activation, rate and force of myofibrillar cross-bridge cycle activity, muscle force, running mechanics). All of these findings stress the influence of muscle power factors on performance in XC sprint skiing and running. The findings of the current study also support the strong connection between DP sprint performance and specific explosive strength and maximal power levels measured in the 2PT. One explanation for the role of specific maximal power and explosive strength on DP sprint performance might be seen in the demands on sprint DP. As shown by Holmberg et al. (7), the duration of the propulsion phase at submaximal speeds in DP was around 300 ms, and it might even decrease toward maximal speeds. Additionally, athletes using a sprinter-like DP technique even had shorter relative poling phases, but they were able to produce higher STRENGTH TEST FOR CROSS-COUNTRY SPRINT SKIING Medicine & Science in Sports & Exercise d 1167

9 peak pole forces and impulse of forces within shorter ground-contact times, compared with the conventional DP technique. In a simulated sprint competition (26), we have observed that the fastest skiers were able to achieve a higher cycle length at equal poling frequency. Hence, because of their higher skiing speed, they had to produce more propulsion at each pole thrust. The difference in cycle length between faster and slower skiers might be explained, on the one hand, by different strength capacities, and, on the other hand, by technical aspects. In total, the prerequisites for using the more explosive sprinter-like DP technique and, consequently, for achieving higher maximal speeds and performance over sprint race distance, could be found in the level of explosive strength and maximal power. Practical applications and measurement concept modifications. Another point is the question about the necessity and efficiency of the specially developed measurement concept for the DP imitation on the rollerboard, using both force and velocity parameters. The main purpose was to get a trigger for the active phase and to cut out the passive roll-out phase in the calculation of variables in the 2PT. To answer that question, the height of the correlation coefficient of the mean velocity over the whole upward movement (V con ) (including the active phase and passive roll-out phase) was compared with the correlation coefficients of the variables attributed to the active phase (e.g., V act, V peak, P act ). V con showed the poorest (but still moderate to high) correlation coefficients towards the tests on roller skis. This might be explained by the possibility of incorrect movements, especially in the passive roll-out phase, which was just considered in the calculation of the variable V con. Thus, by including that special measurement concept, it was possible to increase the quality of the 2PT. However, when just calculating V con, the force measurements can be excluded, and moderate to high relationships to DP sprint performance were still achieved. Hence, when using conventional velocity-measurement systems, sufficient test quality still can be guaranteed. In settings where no velocity- and force-measurement apparatus is available, such as small clubs or regional teams, an easily measurable variable as an indicator of performance might be the distance of the passive roll-out phase. From a mechanical point of view, the passive roll-out phase is independent of the mass of the athlete and only depends on the velocity at the end of the active phase. Hence, theoretically, the higher the end speed an athlete can reach, the longer the distance will be in the passive-roll-out. This connection can be found by the low to moderate correlation between power and velocity output during the active phase and the length of the passive roll-out phase. CONCLUSION AND PERSPECTIVE The developed 2PT for specific upper-body strength diagnostics in XC skiing showed high reliability. Both parts of the 2PT, the 4R max T and the 40RT, were valid tests for predicting short-duration, maximal-speed and for sprint performance over race distance. The 4R max T accounted for up to 84% of the variation of DP 50m maximal speed and up to 61% of DP 1000m sprint performance. V peak and P act were the highest predicting variables measured in the 2PT. The fastest sprinters for the 50- and 1000-m tests produced the highest maximal values in the 4R max T and showed a lower FI during the 40RT. Altogether, the ability to produce a high power output with explosive speed, together with a welldeveloped power endurance level, are the prerequisites for performance in sprint DP. Physiological parameters such as lactate and heart rate showed low reliability and were not related to performance in any of the sprint tests on roller skis. Hence, that aspect calls into question the importance of lactate and heart rate measurements in these test concepts. Further aspects that contribute to sprint performance, such as skiing technique and the role of the lower body during DP, cannot be measured by the 2PT. The positive influence of the ability to produce maximal power output in DPI on XC skiing sprint performance suggests that it might be beneficial to increase the proportion of training aimed specifically at improving explosive strength, maximal power, and power endurance. In the future, 2PT data could be used to determine standards and norms to reveal deficits, enervations, and athlete development, thus serving as an instrument for guiding or controlling the training. We thank H-C. Holmberg for his valuable comments on the manuscript. The authors would also like to express appreciation to participating athletes for their enthusiasm and cooperation during the investigation. Special thanks for the kind support of both the coaches Peter Reithofer and Roland Diethart. REFERENCES 1. BILODEAU, B., B. ROY, and R. M. BOULAY. Upper body testing of cross-country skiers. Med. Sci. Sports Exerc. 27: , GASKILL, S., R. SERFASS, and K. W. RUNDELL. Upper body power comparison between groups of cross-country skiers and runners. Int. J. Sports Med. 20: , HELGERUD, J., J. T. VIK, and J. HOFF. The effect of maximal strength training on endurance performance in upper body for highly trained male cross-country skiers. Corpus Psyche Societas 8:90 103, HOFF, J., A. GRAN, and J. HELGERUD. Maximal strength training improves aerobic endurance performance. Scand. J. Med. Sci. Sports 12: , HOFF, J., J. HELGERUD, and U. WISLOFF. Maximal strength training improves work economy in trained female cross-country skiers. Med. Sci. Sports Exerc. 31: , HOFFMAN, M. D., and P. S. CLIFFORD. Physiological aspects of competitive cross-country skiing. J. Sports Sci. 10:3 27, Official Journal of the American College of Sports Medicine

10 7. HOLMBERG, H-C., S. LINDINGER, T. STÖGGL, E. EITZLMAIR, and E. MÜLLER. Biomechanical analysis of double poling in elite cross-country skiers. Med. Sci. Sports Exerc. 37: , INBAR, O., O. BAR-OR, and J. SKINNER. The Wingate Anaerobic Test. Leeds, UK: Human Kinetics, MAHOOD, N. V., R. W. KENEFICK, R.KERTZER, and T. J. QUINN. Physiological determinants of cross-country ski racing performance. Med. Sci. Sports Exerc. 33: , MARTIN, D., K. CARL, and K. LEHNERTZ. Handbuch Trainingslehre. Schorndorf, Germany: Hofmann, MYGIND, E., B. LARSSON, and T. KLAUSEN. Evaluation of a specific test in cross country skiing. J. Sports Sci. 9: , NESSER, T. W., S. CHEN, R. C. SERFASS, and S. E. GASKILL. Development of upper body power in junior cross-country skiers. J. Strength Cond. Res. 18:63 71, NILSSON, J. E., H-C. HOLMBERG, P. TVEIT, and J. HALLÉN. Effects of 20-s and 180-s double poling interval training in cross-country skiers. Eur. J. Appl. Physiol. 92: , NOAKES, T. D. Implications of exercise testing for prediction of athletic performance: a contemporary perspective. Med. Sci. Sports Exerc. 20: , NOAKES, T. D., K. H. MYBURGH, and R. SCHALL. Peak treadmill running velocity during the V O2max test predicts running performance. J. Sports Sci. 8:35 45, NUMMELA, A., M. ALBERTS, R. P. RIJNTJES, P. LUHTANEN, and H. K. RUSKO. Reliability and validity of the maximal anaerobic running test. Int. J. Sports Med. 2:97 102, NUMMELA, A. T., L. M. PAAVOLAINEN, K. A. SHARWOOD, M. I. LAMBERT, T. D. NOAKES, and H. K. RUSKO. Neuromuscular factors determining 5 km running performance and running economy in well-trained athletes. Eur. J. Appl. Physiol. 97:1 8, OSTERAS, H., J. HELGERUD, and J. HOFF. Maximal strength-training effects on force-velocity and force-power relationships explain increases in aerobic performance in humans. Eur. J. Appl. Physiol. 88: , PAAVOLAINEN, L. M., A. NUMMELA, and H. RUSKO. Muscle power factors and V O2max as determinants of horizontal and uphill running performance. Scand. J. Med. Sci. Sports 10: , PAAVOLAINEN, L. M., A. T. NUMMELA, and H. K. RUSKO. Neuromuscular characteristics and muscle power as determinants of 5-km running performance. Med. Sci. Sports Exerc. 31: , RUSKO, H., A. NUMMELA, and A. MERO. A new method for the evaluation of anaerobic running power in athletes. Eur. J. Appl. Physiol. 66:97 101, RUSKO, H. K., and A. T. NUMMELA. Measurement of maximal and submaximal anaerobic power. Int. J. Sports Med. 2:89 130, SALTIN, B. The physiology of competitive c.c. skiing across a four decade perspective; with a note on training induced adaptations and role of training at medium altitude. In: Science and Skiing, Erich Müller, Hermann Schwameder, E. Kornexl, et al. (Eds). St. Christoph am Arlberg, London, UK, Chapman & Hall, pp , SCHABORT, E. J., J. A. HAWLEY, W. G. HOPKINS, and H. BLUM. High reliability of performance of well-trained rowers on a rowing ergometer. J. Sports Sci. 17: , STAIB, J. L., J. IM, Z. CALDWELL, and K. W. RUNDELL. Crosscountry ski racing performance predicted by aerobic and anaerobic double poling power. J. Strength Cond. Res. 14: , STÖGGL, T., S. LINDINGER, and E. MÜLLER. Analysis of a simulated sprint competition in classical cross country skiing. Scand. J. Med. Sci. Sports (OnlineEarly Articles). doi: /j x. 27. STÖGGL, T., S. LINDINGER, and E. MÜLLER. Biomechanical validation of a specific upper body training and testing drill in cross-country skiing. Sports Biomech. 5:23 46, STÖGGL, T., S. LINDINGER, and E. MÜLLER. Reliability and validity of test concepts for the cross-country skiing sprint. Med. Sci. Sports Exerc. 38: , VANDEWALLE, H., G. PÉRÈS, and H. MONOD. Standard anaerobic exercise tests. Sports Med. 4: , VAN HALL, G., M. JENSEN-URSTAD, H. ROSDAHL, H-C. HOLMBERG, B. SALTIN, and J. A. CALBET. Leg and arm lactate and substrate kinetics during exercise. Am. J. Physiol. Endocrinol. Metab. 284:E193 E205, STRENGTH TEST FOR CROSS-COUNTRY SPRINT SKIING Medicine & Science in Sports & Exercise d 1169