A Quasi-Experimental Study of Chinese Top-Level Speed Skaters Training Load: Threshold Versus Polarized Model

Transcription

1 International Journal of Sports Physiology and Performance, 2012, 7, Human Kinetics, Inc. A Quasi-Experimental Study of Chinese Top-Level Speed Skaters Training Load: Threshold Versus Polarized Model Hongjun Yu, Xiaoping Chen, Weimo Zhu, and Chunmei Cao Purpose: To examine the effectiveness of threshold and polarized models in the training organization of Chinese top-level sprint speed skaters using a 2-y quasi-experimental design. Methods: Two years ( and seasons) of the Chinese national speed-skating team s daily training load (N = 9; 5 men, 23.6 ± 1.7 y, weight 76.6 ± 4.1 kg, competitive experience 5.0 ± 0.8 y, 500-m time ± 0.72 s, 1000-m time ± 2.28 s; 4 women, 25.3 ± 6.8 y, 73.0 ± 8.5 kg, 6.3 ± 3.5 y, ± 0.46 s, ± 0.81 s) were collected and analyzed. Each season s training load included overall duration (calculated in min and km), frequency (calculated by overall sessions), and training intensity (measured by ear blood lactate or estimated by heart rate), Their performances at national, World Cup, and Olympic competitions during the 2 seasons ( ), as well as lactate data measured 15 and 30 min after these competitions, were also collected and analyzed. Based on the lactate data (<2, 2 4, >4 mmol/l), training zones were classified as low, moderate, and high intensity. Results: The total durations and frequencies of the training load were similar across the seasons, but a threshold-training model distribution was used in , and a polarized-training-load organization in Under the polarized-training model, or load organization, all speed skaters performance improved and their lactate after competition decreased considerably. Conclusion: Training-intensity distribution based on a polarized-training model led to the success in top Chinese sprint speed skaters in the season. Keywords: intensity distribution, speed skating, endurance training Training load consists of training volume, intensity, and frequency and determines the nature and magnitude of the training effect. 1 Foster et al 2 reported that athletic performance was associated with an increase of training load. However, it is still not clear how training intensities (eg, low, moderate, and high) should be distributed in elite athletes. 3 6 The day-by-day and seasonal distribution of training intensity appear to be a crucial factor in training planning for athletes. Overtraining may result in injuries and decreases in performance, whereas undertraining will likely lead to a failure to achieve the desired increase of performance. There are 2 basic intensity-distribution patterns or models based on previous research literature: threshold and polarized. The threshold model emphasizes training between the first and second lactate, or ventilatory, thresholds. 3 Kindermann et al 4 suggested that training in this intensity range can significantly improve performance. They proposed using heart rate (HR) corresponding to Yu and Cao are with Tsinghua University, Beijing, China. Chen is with the Exercise Science Laboratory, Ningbo University, China. Zhu is with the University of Illinois at Urbana-Champaign, Champaign, IL. the aerobic anaerobic blood lactate threshold to determine training intensity. 4 The threshold-training model has been widely used in practice and is supported by published studies. 7,8 The polarized model emphasizes a large volume of training below the first lactate, or ventilatory, threshold combined with higher-intensity training eliciting 90% to 100% of VO 2max. 3 Seiler and Kjerland 3 suggested that the optimal training-intensity distribution for elite endurance athletes is often based on the polarized model, rather than the threshold model. In their report, they identified 3 intensity zones based on the HR threshold corresponding to ventilatory and blood lactate threshold. The polarized-training model has been demonstrated to be effective in various elite endurance-sport athletes such as international-class rowers, 9 11 marathoners, 12 distance runners, 13,14 and cyclists. 15,16 Although the polarized-training intensity distribution has been well tested in endurance athletes, less effort has been made to apply it in sprint-sport athletes. In addition, many published studies have only employed a short period of training, 9,12 15 and the impact of longterm training based on the different models is still unknown. The purpose of this study was to examine the effectiveness of the 2 training models threshold and 103

2 104 Yu et al polarized in the training of sprint speed skaters. Based on some preliminary information, we hypothesized that the training-intensity model that the Chinese top-level speed skaters employed before this study was a threshold model. A 2-stage quasi-experimental study, 1 year per stage, was designed: baseline and intervention. The specific aim of the baseline study was to verify the hypothesized threshold model, while the specific aim of the intervention study was to introduce a new training-intensity model, that is, the polarized model, and examine its effectiveness. Subjects Methods Nine Chinese top-level sprint-event (500-m and 1000-m) speed skaters, 5 men and 4 women, participated in this study. One of the male skaters was the Chinese record holder in the 500-m, who won the Speed Skating World Cup 500 m championship in the season; one of female skaters was the Chinese record holder in both the 500-m and the 1000-m, who also won the Speed Skating World Cup 500-m championship in the season; and 6 of 9 subjects took part in the 2006 Olympic Winter Games, with 3 of them performing well (1 silver, 1 bronze, and a fifth place). Descriptive statistics of the subjects are presented in Table 1. This study was approved by the National Committee for Elite Athlete Research Ethics in China. Written informed consent was obtained from all participants before data collection. Measures Training-Intensity Determination. Accurately determining training intensity is essential to this study. Since it was impractical to perform blood lactate testing Table 1 Subject Characteristics, Mean ± SD Men (n = 5) Women (n = 4) Age (y) 23.6 ± ± 6.8 Height (cm) ± ± 6.2 Mass (kg) 76.6 ± ± 8.5 VO 2max (L/min) 4.0 ± ± 0.1 VO 2max (ml min 1 kg 1 ) 52.5 ± ± 2.1 W max ± ± 5.0 HR max (beats/min) ± ± 8.5 Professional competitive experience (y) 5.0 ± ± m time (s) 35.5 ± ± m time (s) 71.2 ± ± 0.8 Abbreviations: VO 2max, maximal oxygen uptake; W max, maximal power output; HR max, maximal heart rate. 500-m and 1000-m times were the best result of the subjects in the season. Professional competitive experience was counted as number of years subjects had participated in national competitions. on athletes every day of the season, the intensity was determined using HR. The relationship between HR and blood lactate for each individual was determined in a field setting every month (see Table 2). Specifically, during the first week of every month, an investigator was present at a field setting to calibrate the HR bloodlactate relationship (HR at blood lactate concentrations of 2.0 and 4.0 mmol/l) for all the athletes. Based on target HR, measurement was made on the steady-state or interval-type sessions 30 to 60 minutes after the start of the session under field conditions using ear blood lactate with a portable device (ARKRAY Lactate Pro) and HR using radio telemetry (Polar CS600). The HR at aerobic threshold intensity (ie, <2 mmol/l blood lactate), aerobic anaerobic intensity (ie, approximately 2-4 mmol/l), and anaerobic threshold intensity (ie, >4 mmol/l) at the start, middle, and end of every training period during the first week of the season were determined. HR during the rest of the season was recorded for every training session at 15-second intervals, and the collected data were downloaded on a weekly basis. When HR data were not available due to recording problems, they were estimated using training logs and diary contents for comparable training sessions (Note: only about 0.2% of the sessions HR were estimated). Training-Load Determination. Training sessions in both seasons were categorized as the following basic types: endurance sessions, strength sessions, skating sessions, speed sessions, and dry-land skill sessions. Frequency was calculated by overall session. Endurance sessions duration was calculated in minutes, and intensity in low-, moderate-, and high-intensity zones. Strength sessions were categorized as maximum strength (<2 mmol/l blood lactate, above % of 1RM), endurance strength (>4 mmol/l blood lactate, 60 79% of 1 RM), explosive strength (<2 mmol/l blood lactate, 30 59% of 1 RM), and core strength (core stability abdominal exercises). Strength sessions were measured by sets and numbers (core strength by minutes) using a diary. Skating sessions intensities were categorized in low-, moderate-, and high-intensity zones; skating duration was measured by distance (400 m for every lap). Speed training was measured in seconds. Performance Recording and Computation. All athletes performances at national, World Cup, and Olympic competitions during the seasons were recorded, as well as lactate data measured 15 and 30 minutes after these competitions. Each athlete s season performance was determined using the best performance in those competitions within a season. All performances were expressed as a real-competition result (reported in seconds). The percentage of performance change from the previous season was also computed: Performance cross-season change % = [(best time in season best time in season)/best time in season] 100%. Other Recorded Information. All subjects performed the tests in a laboratory. Environmental conditions were

3 Threshold Versus Polarized Model 105 Table 2 Categories for Low, Moderate, and High Intensity Between Heart Rate and Blood Lactate, Mean ± SD Training-Intensity Category Low Moderate High Men heart rate (beats/min) <150 ± ± ± 10 >168 ± 10 maximal heart rate <79 ± 2 79 ± 2 89 ± 3 >89 ± 3 blood lactate (mmol/l) <2 2 4 >4 Women heart rate (beats/min) <152 ± ± ± 7 >170 ± 7 maximal heart rate <82 ± 3 82 ± 3 92 ± 4 >92 ± 4 blood lactate (mmol/l) <2 2 4 >4 Note: individual heart rates for targeted intensity ranges based on the lactate heart-rate relation in the field setting every month. similar during the study evaluations period. Maximal oxygen uptake (VO 2max ), maximal heart rate (HR max ), and maximal power output (W max ) were measured and recorded at the start of every season (in May). After ~10 minutes of warm-up, skaters performed a continuous treadmill test (Woodway, Weil am Rhein, Germany) to voluntary exhaustion. VO 2max was defined as the highest continuous 1-minute VO 2 value during the test. HR was monitored every 5 seconds by radio telemetry (Polar, Kempele, Finland) during the treadmill tests to determine HR max. W max was measured on a cycle ergometer (Lode Excalibur, Lode, Netherlands). Skaters started with an ~10-minute warm-up at 150 W, and the power output was increased by 50 W every 2.5 minutes until voluntary exhaustion. The relationship of blood lactate corresponding to HR (HR at blood lactate concentration of 2.0 and 4.0 mmol/l) for each individual also was determined at rest and the end of each exercise stage on the cycle-ergometer measurement. Blood samples for the measurement of blood lactate were taken from the ear and immediately analyzed by dry chemistry (ARKRAY Lactate Pro). HR was recorded continuously by telemetry (Polar, Kempele, Finland) during the tests to determine the HR blood-lactate relationship according to the methodology described by Kindermann et al. 4 All training sessions for both seasons were also recorded. The training dairies of both the athletes and the coaches were collected. Training activities during these 2 seasons were also recorded (eg, cycling, running, swimming, strength training, skateboard training, skating, etc). Design Stage 1: Baseline. The baseline study was conducted in the season. Using the HR monitor, all subjects HRs were recorded and used to determine total time spent in each of the 3 training zones. In addition, all athletes performances at the national and World Cup competitions during the season were recorded. Stage 2: Intervention. Because the hypothesized threshold-training model was verified by the baseline data, an intervention study in the season was designed to create a polarized-training model characterized with training plans with intensities of 85% at <2 mmol/l, 5% at 2 to 4 mmol/l, and 10% at >4 mmol/l blood lactates. Based on this intensity distribution, the intervention was designed to decrease 2 to 4 mmol/l training in the endurance session, decrease also 2 to 4 mmol/l level training in the skating session, and increase in the high-intensity skating session. The training year in stage 2 was divided into 3 periods: recovery, preparation, and competition. The recovery period was in May (lasting 3 4 wk). The main training in this period was recovery training. Athletes participated in recovery activities, for example, low-intensity running or bicycling. At the end of this period, athletes were tested for VO 2max and W max. The preparatory period went from the first week of June to the middle of September (lasting wk). During this period, the training load progressively increased from low to high. Athletes mainly performed endurance, strength, and speed training and some skateboarding-skill exercises. Endurance activities included running, cycling, swimming, and low walks (walking squats). Strength activities were general strength training to stabilize and improve hip and knee muscle power and explosive strength training (plyometrics training including multiple jumping exercises to improve pushoff muscle power). The competitive period was divided into the preseason period (September and October, lasting 6 8 wk) and the in-season period (November to the middle of March, wk). During the in-season competitive period, the overall training load and nonspecific training progressively decreased so that athletes mainly focused on specific skating training and reaching peak performance during competition. Skating activities were categorized as follows: aerobic endurance skating,

4 106 Yu et al anaerobic endurance skating, coordinated skating, sprint speed skating, sprint start skating, and recovery skating. Athletes mainly competed in the Chinese National Championships, the Speed Skating World Cup Circuit, the World Championships, and the Olympic Winter Games. During the season, the most important competition for the Chinese speed skaters was the 2006 Olympic Winter Games. Data Analysis The data were first analyzed using descriptive statistics to identify outliers. Training session by type (ie, strength, endurance, speed, skating, and skill), gender, and season were summarized and compared using a 2-way ANOVA. To determine the training model employed by a stage, the intensity-distribution zone percentage (low vs moderate vs high) was computed for endurance and skating training loads. Strength training by type (ie, maximum strength, endurance strength, explosive strength, and core strength), session, set, number, and gender was calculated and compared using a 2-way ANOVA. Performance and physiology measures in 500-m and 1000-m competitions also were summarized and compared with a 2-way ANOVA. Results obtained from the records are expressed as mean ± SD. Two-way ANOVA with Tukey post hoc tests was used to determine subgroup differences. The statistical significance level was set at.05. Results Table 3 shows training-session distribution information during the and seasons. As expected, the total sessions were substantially similar for both male and female athletes across the seasons. However, compared with the season, both male and female speed skaters performed significantly more speed sessions in the season. Tables 4 and 5 present the distribution of endurance and skating training loads between the 2 seasons in male and female speed skaters. Overall training durations and Table 3 Distribution of Training Sessions by Type and Gender in the Seasons, Mean ± SD Session Strength Endurance Speed Skating Skill Total sessions Men ± ± ± ± ± ± ± ± ± 1.8** 83.2 ± ± ± 10.8 Women ± ± ± ± ± ± ± 2.9* 97.7 ± ± 2.0** 82.0 ± ± ± 6.0 Note: Strength session and endurance session = dry-land strength and endurance training; speed session = dry-land speed training; skill session = dry-land skateboarding skill exercises; skating session = ice-skating training. *P <.05, **P <.001: Statistically significant across the seasons within the same gender. Table 4 Threshold- Versus Polarized-Model of Endurance-Training Load Distribution Between 2 Seasons by Gender in Chinese Top-Level Speed Skaters, Mean ± SD Intensity-Distribution Zone, % Duration, min Low Moderate High Men ± ± ± ± % distribution 100% 40.8 ± ± ± ± ± 420.7** ± 280.4** ± 166.5** % distribution 100% 84.4 ± ± ± 1.9 Women ± ± ± ± % distribution 100% 42.1 ± ± ± ± ± 332.1** ± 118.7** ± 120.8** % distribution 100% 87.1 ± ± ± 1.6 Note: Endurance-training load was dry-land endurance exercise. It was categorized by running, swimming, cycling training, and walking squats. **P <.001: Statistically significant across the seasons within the same gender.

5 Threshold Versus Polarized Model 107 Table 5 Threshold- Versus Polarized-Model of Skating-Training Load Distribution Between 2 Seasons by Gender in Chinese Top-Level Speed Skaters, Mean ± SD Intensity-Distribution Zone, % Duration, min Low Moderate High Male ± ± ± ± 11.9 % distribution 100% 43.2 ± ± ± ± ± 12.4** 49.5 ± 9.9** 87.5 ± 9.1** % distribution 100% 83.4 ± ± ± 1.1 Female ± ± ± ± 8.6 % distribution 100% 39.7 ± ± ± ± ± 9.5** 36.1 ± 10.2** 86.4 ± 10.2** % distribution 100% 82.0 ± ± ± 1.5 Note: Skating-training load was ice-skating exercise. It was categorized by aerobic endurance skating, anaerobic endurance skating, coordinating skating, sprint speed skating, sprint start skating, skill slowing skating, and recovery skating. ** P <.001: Statistically significant across the seasons within the same gender. frequencies were similar across the seasons, but trainingintensity distribution was very different, with a thresholdtraining model for the season (~50% at a lactate concentration of 2 4 mmol/l) but a polarized-training model (~5% at a lactate concentration of 2 4 mmol/l) for the season (see Figures 1 and 2). Table 6 indicates that the female speed skaters maximal power output increased significantly, all speed skaters performance improved significantly (average increase 2 4%), and their lactate after competition decreased significantly with the polarized-training model implemented in the season. The change trend of performance by individual skater is further illustrated in Figure 3. Discussion The key finding of this study is that a training program with a polarized-training model elicited significant performance improvement in Chinese sprint speed skaters. Hopkins et al 17 proposed that for elite athletes competing in individual sports, a very small enhancement ( %) of performance has a substantial effect on a top athlete s chances of medaling in competition. Compared with the season using the thresholdtraining model, top-level speed skaters performance had significant improvement averaging 2% to 4% in the 500-m and 1000-m events in both women (of whom 1 won a silver and 1 bronze) and men (1 placed fifth) in the 2006 Olympic Winter Games. It is true that all skaters seemed to perform better in the Olympic year. Considering the degree of improvement of 2% to 4%, however, we believe that employing the polarized model is the key contributor. Thus, it was a successful training program in the season. This study is perhaps the first training study that successfully applied the polarized model to sprint athletes. It is becoming accepted that the polarized model works for endurance athletes, as has been shown to do so in the case of world-class rowers, 11 junior male skiers, 3 and marathoners. 12 However, from a metabolic-energy perspective, 500-m and 1000-m speed-skating events are dominated by anaerobic energy, and each athlete races against the clock during the events One may wonder how the polarized-distribution model could significantly improve speed-skating performance. We believe there are 3 major reasons. First, it probably helped athletes recovery. As we know, high-intensity training load leads to stressful responses. Athletes are very sensitive to the change in training intensity. Increased training intensity is often associated with an increase in sympathetic stress. For elite athletes in this study, the polarized-training load distribution might have induced a relatively low overall stress load and therefore facilitated more rapid recovery than with the high-intensity training sessions under the lactate threshold model. 21 Seiler et al 22 found that autonomic nervous system recovery is very rapid after training bouts at 60% VO 2max for up to 120 minutes but becomes markedly delayed in highly trained subjects when the intensity increases to above 3 mm blood lactate. They suggest that polarized-training distribution could give excellent long-term results and could be effective in stimulating physiologic adaptations in athletes. 21 Second, training based on the polarized distribution may avoid training monotony, which is the index of training variability defined as the daily mean ± SD calculated over the period of a week. 23 Varying day-today training is important for athletes training program. Foster 24 examined overtraining syndrome in a training program for competitive speed skaters and introduced

6 Figure 1 Male skaters threshold- (endurance) vs polarized-model (skating) training-intensity distribution by season and training type. Figure 2 Female skaters threshold- (endurance) vs polarized-model (skating) trainingintensity distribution by season and training type. 108

7 Threshold Versus Polarized Model 109 Table 6 Performance and Physiologic Measures During Seasons, Mean ± SD Season Change season Change Men 500-m (s) ± ± ± 0.72 change (%) 0.10% ± 0.27% 2.90% ± 1.72% * 1000-m (s) ± ± ± 2.28 change (%) 0.38% ± 0.65% 3.10% ± 0.73% * 15-min postcompetition blood lactate (%) 8.52 ± 2.81 * 30-min postcompetition blood lactate (%) 4.36 ± 1.82 * maximal oxygen uptake (L/min) 3.8 ± ± 0.3 maximal oxygen uptake (ml min 1 kg 1 ) 51.0 ± ± 2.4 maximal power output ± ± 27.3 Women 500-m (s) ± ± ± 0.46 change (%) 0.11% ± 1.33% 2.15% ± 1.66% * 1000-m (s) ± ± ± 0.81 change (%) 0.38% ± 1.12% 3.78% ± 2.05% * 15-min postcompetition blood lactate (%) ± 2.21 * 30-min postcompetition blood lactate (%) 8.75 ± 2.34 * maximal oxygen uptake (L/min) 3.5 ± ± 0.1 maximal oxygen uptake (ml min 1 kg 1 ) 46.6 ± ± 2.1 maximal power output ± ± 5.0 * *P <.05: Statistically significant across the seasons within the same gender. the concept of training monotony, which is related to negative adaptations to training. The threshold mode thus may reduce the quality of higher-intensity workouts and ultimately lead to training monotony and overtraining. Finally, the polarized distribution perhaps fit biological adaptation to the top-level sprint-event athletes. Training leads to changes in gene expression and protein synthesis; unfortunately, details regarding how intensity and duration of exercise combine to modulate elite athletes cell signaling are not well reported in the literature. 21 What is known is that this regular, cellularlevel stimulation must be achieved while preserving the autonomic balance of the organism so that overtraining is avoided and the capability for maximal sympathetic mobilization is retained. 3 Compared with threshold distribution, it might be that polarized distribution can drive gene expression for mitochondrial protein proliferation through different pathways, which is optimal to adaptation for top-level sprint-event athletes. This, in fact, may be the major advantage of more polarized training: Allowing more recovery during general training enables specific training to be done harder. In other words, polarized-training distribution may make easy training easier and hard training harder. Two observations further support use of the polarized model. The first is the significant performance improvements from the season to the season. The second is that, although the polarized model is known to work in endurance athletes, this study demonstrated that it can be extended to sprint-event athletes. From the 2 seasons of data, skaters achieved the shift from the threshold model to the polarized model in their periodized annual training. The preparation period (dry-land exercises) and competitive period (ice-skating exercises) were the critical phases in the annual training for the Chinese speed skaters. During the preparation period, endurance, strength, and speed training dominated. In addition, total training types from the threshold-training model in the season shifted to the polarized-intensity model in the season. Speed skaters performed significantly more high-intensity speed sessions and endurance exercises during the season during the preparation

8 110 Yu et al Figure 3 Season-best performances over 2 seasons by skater. period. During the competition period, ice skating was the most important training. It is very clear that trainingintensity distribution, as we designed it, shifted from the threshold model to the polarized model between the 2 seasons of training. Comparing the 2 sets of periodized annual training data, we find that the intensity-distribution variety of training is a key factor for the Chinese top-level speed skaters to improve their performance. There are 2 limitations of this study that should be acknowledged. First, speed skaters significantly decreased lactates after competition may be evidence of their aerobic capability s improving under the polarized model. Since there was no control group in the study, the impact of the carryover training effect from the previous year is unknown. In addition, there are many other factors that may affect speed-skating competition results, such as ice and air-pressure conditions in different venues. Pacing strategy is also known to be a critical issue in sprint skating. 25 Another limitation of this study is that we did not apply other approaches to monitoring training load throughout the season. Several methods have been suggested to be useful to quantify training load. Banister et al 26 originally developed the training impulse (TRIMP) based on HR into a unit dose-response method to quantify training load. One of advantages of the TRIMP method is to predict an athlete s performance. 27 This method has shown great promise in better understanding training load. Foster et al 28 proposed a modification of the rating of perceived exertion (RPE) to simply quantify session training load. The session-rpe method can serve as a surrogate of HR or lactate as a measure of training intensity. The session-rpe method is a useful and advantageous method to quantify training load, especially for strength training or when HR or lactate data are not available. 29,30 Future research is necessary to apply TRIMP and session- RPE methods to quantify training load in speed skaters. The findings of this study may help coaches understand that the polarized-training model could be extended to sprint athletes. The science of training often falls behind in dealing with day-to-day training specifics of top-level athletes. While there is little evidence of why

9 Threshold Versus Polarized Model 111 the polarized-distribution model is beneficial for top-level sprint athletes, the findings of this study should encourage more research to investigate the impact of intensity distributions in training. Conclusion This study showed that training-intensity distribution based on the polarized-training model may be extended to sprint athletes. This was done for the Chinese toplevel sprint-event speed skaters, and it led the skaters to improve their performance on average 2% to 4% from the to the season. Acknowledgments We thank Guanglei Chen, the coach of Chinese national speedskating team, and all the speed skaters for their support and voluntary participation in this study. References 1. Borresen J, Lambert MI. The quantification of training load, the training response and the effect on performance. Sports Med. 2009;39: PubMed, doi: / Foster C, Danies E, Hector L, Snyder A, Welsh R. Athletic performance in relation to training load. Wis Med J. 1996;95: PubMed 3. Seiler KS, Kjerland GO. Quantifying training intensity distribution in elite endurance athletes: is there evidence for an optimal distribution? Scand J Med Sci Sports. 2006;16: PubMed, doi: /j x 4. Kindermann W, Simon G, Keul J. The significance of the aerobic-anaerobic determination of work load intensities during endurance training. Eur J Appl Physiol Occup Physiol. 1979;42: PubMed, doi: / BF Laursen PB, Jenkins DG. The scientific basis for high intensity interval training: optimising training programmers and maximising performance in highly trained endurance athletes. Sports Med. 2002;32: PubMed, doi: / Smith DJ. A framework for understanding the training process leading to elite performance. Sports Med. 2003;33: PubMed, doi: / Mader A. Evaluation of the endurance performance of marathon runners and theoretical analysis of test results. J Sports Med Phys Fitness. 1991;31:1 19. PubMed 8. Billat LV. Use of blood lactate measurements for prediction of exercise performance and for control of training. Sports Med. 1996;22: PubMed, doi: / Steinacker JM, Lormes W, Lehmann M, Altenburg D. Training of rowers before world championships. Med Sci Sports Exerc. 1998;30: PubMed, doi: / Steinacker JM, Lormes W, Kellmann M, et al. Training of junior rowers before world championships. effects on performance, mood state, and selected hormonal and metabolic responses. J Sports Med Phys Fitness. 2000;40: PubMed 11. Guellich A, Emrich E, Seiler S. Training methods and intensity distribution of young world-class rowers. Int J Sports Physiol Perform. 2009;4: PubMed 12. Billat VL, Demarle A, Slawinski J, Paiva M, Koralsztein JP. Physical and training characteristics of top-class marathon runners. Med Sci Sports Exerc. 2001;33: PubMed, doi: / Esteve-Lanao J, San Juan AF, Earnest CP, Foster C, Lucía A. How do endurance runners actually train? relationship with competition performance. Med Sci Sports Exerc. 2005;37: PubMed, doi: /01. MSS Esteve-Lanao J, Foster C, Seiler S, Lucía A. Impact of training intensity distribution on performance in endurance athletes. J Strength Cond Res. 2007;21(3): PubMed 15. Schumacher YO, Müller P. The 4000-m team pursuit cycling world record: theoretical and practical aspects. Med Sci Sports Exerc. 2002;34: PubMed, doi: / Foster C, Hoyos J, Earnest CP, Lucia A. Regulation of energy expenditure during prolonged athletic competition. Med Sci Sports Exerc. 2005;37: PubMed, doi: /01.mss bf 17. Hopkins WG, Hawley JA, Burke LM. Design and analysis of research on sport performance enhancement. Med Sci Sports Exerc. 1999;31: PubMed, doi: / Foster C, dekoning JJ, Rundell KW, Snyder AC. Physiology of speed skating. In: Garrett WE, Kirkendall DT, eds. Exercise and Sport Science. Philadelphia, PA: Lippincott Williams & Wilkins; 2000: Foster C, dekoning JJ, Hettinga F, et al. Effect of competitive distance on energy expenditure during simulated competition. Int J Sports Med. 2004;25: PubMed, doi: /s de Groot G, Hollander AP, Sargeant AJ, van Ingen Schenau GJ, de Boer RW. Applied physiology of speedskating. J Sports Sci. 1987;5: PubMed, doi: / Seiler S. What is best practice for training intensity and duration distribution in endurance athletes? Int J Sports Physiol Perform. 2010;5: PubMed 22. Seiler S, Haugen O, Kuffel E. Autonomic recovery after exercise in trained athletes: intensity and duration effects. Med Sci Sports Exerc. 2007;39: PubMed, doi: /mss.0b013e318060f17d 23. Foster C, Lehmann M. Overtraining syndrome. In: Guten GN, ed. Running Injuries. Philadelphia, PA: WB Saunders; 1997:J73 J Foster C. Monitoring training in athletes with reference to overtraining syndrome. Med Sci Sports Exerc. 1998;30: PubMed, doi: /

10 112 Yu et al 25. Hettinga FJ, De Koning JJ, Schmidt LJ, Wind NA, MacIntosh BR, Foster C. Optimal pacing strategy: from theoretical modelling to reality in 1500-m speed skating. Br J Sports Med. 2011;45(1): PubMed 26. Banister EW, Calvert TW, Savage MV, Bach T. A systems model of training for athletic performance. Aust J Sports Med. 1975;7: Banister EW, Carter JB, Zarkadas PC. Training theory and taper: validation in triathlon athletes. Eur J Appl Physiol Occup Physiol. 1999;79(2): PubMed, doi: /s Foster C, Florhaug JA, Franklin J, et al. A new approach to monitoring exercise training. J Strength Cond Res. 2001;15: PubMed 29. Impellizzeri FM, Rampinini E, Coutts AJ, Sassi A, Marcora SM. Use of RPE-based training load in soccer. Med Sci Sports Exerc. 2004;36: PubMed, doi: /01.mss f 30. Day ML, McGuigan MR, Brice G, Foster C. Monitoring exercise intensities during resistance training using a session RPE scale. J Strength Cond Res. 2004;18: