Nordic skiing is an endurance sport that requires the use of both the lower and upper body. When skiing, the grade of the terrain alters how physical work is distributed between the upper- and lower-body musculature. In fact, changes in inclination have been suggested to affect mechanical efficiency, cardiovascular function, and substrate utilization (2,8,13,16,19-21). Differences in muscle mass, intrinsic muscle characteristics, blood flow, and oxygen consumption between the upper and lower body create a disparity between these segments in terms of maximal work capacity and metabolic demand to exercise (4,10,15,22,24).
At an equivalent absolute oxygen consumption, aerobic arm cranking has been shown to result in higher blood lactate (BLa) concentration, a higher respiratory exchange ratio, and increased plasma epinephrine levels than leg cycling exercise (10). When cross-country skiing at intensities of 70-95% of age-predicted maximum heart rate (HR), the use of double poling (a technique that shifts much of the workload to the upper body) elicits higher BLa values than during skating (a technique where the load is more equally shared between the arms and legs) (9). An increase in grade during skate skiing similarly places a greater load on the upper-body musculature (2,4,8,13,15,19-21). During cross-country skiing, the upper-body musculature is less capable of extracting and using oxygen during exercise due to a faster transit time of blood through the capillaries and a decreased capillary diffusion area (4). A greater use of the upper body during uphill skating should therefore increase the dependence on anaerobic metabolism, and BLa values would be expected to be higher for a given level of exercise.
The velocity at which lactate threshold (LT) occurs has long been known to correlate highly to endurance performance (5,7). For runners or cyclists, velocity may be a good measure of intensity, but for skiers, frequently varying environmental conditions (e.g., snow consistency, course elevation profile) do not allow accurate measurement of exercise intensity by pace (6). Thus, HR-based training zones have become a very important method of intensity control in Nordic skiing. These zones are routinely determined using the relationship between BLa, HR, and oxygen consumption (o2) from laboratory-based running or roller ski tests (18,24). It is important for skiers, coaches, and exercise specialists to consider the extent that grade might alter the relationship between these variables to determine and maintain accurate training intensities. It was therefore the purpose of this study to examine the effects of level vs. graded skate skiing on capillary BLa kinetics. The second purpose was to evaluate the effects of grade-induced alterations in the blood lactate vs. HR relationship on training intensity prescriptions. It was hypothesized that a greater use of the upper body during uphill skiing will increase the dependence on anaerobic metabolism and result in greater BLa accumulation for a given exercise intensity.
Experimental Approach to the Problem
This study used a within-subject repeated measures design to determine if the grade of the terrain during cross-country skiing alters the BLa vs. intensity relationship. To do this, differences in prethreshold BLa, the HR at LT, and o2 at LT were compared between level and uphill skiing for each subject. With the order of conditions randomly assigned, subjects were able to complete both conditions without eliciting an order effect.
Eleven, male, Nordic ski racers, who competed for 6 different National Collegiate Athletic Association-Division I University Ski Teams, served as subjects. All skiers were proficient in all skate and classic skiing techniques and had national- or international-level race experience. They were 20.8 ± 1.2 years, 71 ± 5.4 kg, and, over the month before testing, completed 10.7 ± 2.9 hours of aerobic training per week, 7.3 ± 2.5 hours of which were spent skiing or roller skiing. Additional subject descriptive characteristics can be found in Table 1. The research protocol was approved by the University's Human Subjects Committee, and all subjects gave their informed written consent.
In total, 4 roller ski treadmill protocols were used, 2 submaximal and 2 maximal. To assess the effects of grade and speed on capillary BLa during submaximal Nordic skate skiing, 2 incremental roller ski protocols were used, one increased intensity by grade (Ginc), while the other protocol increased intensity using speed (Sinc). The order of the speed- and grade-based tests was randomized between the participants, and the tests were completed 24 hours apart. Each of the submaximal protocols were discontinuous using 3-minute exercise stages followed by 1 minute of rest for blood sampling. Comparisons were made for prethreshold BLa concentrations, HR and o2 at LT, and the HR/o2 relationship between the grade- and speed-based protocols. To calculate the shift in target HR when going from level to graded skiing (or vice versa), all the pairs of data points (HR and BLa) for each person, at each stage, were plotted in a single XY scatter plot. Then, second-order polynomials were fit to the group HR vs. BLa curves for both Ginc and Sinc. The 2 polynomials were subtracted algebraically with the resultant equation allowing the determination of the target HR shift for any given BLa greater than 2.5 mmol·L−1. Additionally, peak oxygen consumption (o2peak), peak lactate (BLapeak), and peak heart rate (HRpeak) were assessed for both skate skiing and double poling using 2 additional maximal treadmill roller ski tests. This was done to compare aerobic capacity between primarily upper-body work (double poling) and combined upper- and lower-body work (skating). All protocols used roller ski ergometry on a 2.4- × 3.1-m motorized treadmill (Figure 1), and all participants used the same model roller skis (Figure 2) (Model 610; Marwe, Hyvinkää, Finland) fit with either New Nordic Norm (Rottefella, Klokkarstua, Norway) or Salomon Nordic System (Salomon, Annecy, France) skating bindings to accommodate the use of each subject's own ski boots. Competition skating ski boots and carbon fiber skating ski poles at each individual's competition length were used for all trials. The model and manufacturer of boots and poles varied from subject to subject based on personal preference, but within a subject, the same roller skiing equipment was used for all trials. With the exception of the double-pole protocol, participants were instructed to use the skating technique only but were free to choose the specific stride within that discipline.
Treadmill Roller Ski Protocols
An effort was made to design the submaximal speed and grade protocols such that the stage workloads and steps were of similar magnitude for each. The 2 maximal protocols were used to measure the peak aerobic capacity and HR for both the skate skiing and double-poling techniques.
Submaximal Speed Protocol for Lactate Threshold Determination
This protocol incrementally increased intensity by speed and used 3-minute stages. Subjects performed a 5-minute warm-up period at 2 m·s−1, at 5% grade. Each stage then increased by 0.5 m·s−1 until 3.5 m·s−1 was reached when speed then increased by 0.2 m·s−1 per stage. Heart rate was recorded in the final 30 seconds of each stage. Expired gas was continuously sampled during each 3-minute stage, and stage o2 was calculated as the mean of the last minute of each stage. A blood sample for BLa determination was taken immediately after each 3-minute stage during a 1-minute rest period. The test was terminated when BLa surpassed 5 mmol·L−1.
Submaximal Grade Protocol for Lactate Threshold Determination
The protocol started at a constant speed of 2.7 m·s−1 and 2% grade and used 3-minute stages. Grade was increased by 2% at each stage until BLa surpassed 5 mmol·L−1 at which point the test was terminated. Expired gas, HR, and blood samples were collected as described above.
Maximal Skate Skiing Aerobic Capacity Protocol
To measure o2peak, HRpeak, and BLapeak during skate skiing, a maximal continuous incremental treadmill protocol was used. The protocol employed 1-minute stages with a constant treadmill velocity of 3.8 m·s−1. The first stage began at a 1% grade, and subsequently, each stage increased by 1% grade each minute until volitional exhaustion. Expired gas was collected continuously and the highest 30-second average was recorded as o2peak. Heart rate was recorded at the end of each stage, and HRpeak was recorded as the highest HR observed in the final minutes of the test. Two minutes post test, a blood sample was taken for BLapeak determination.
Maximal Double-Pole Aerobic Capacity Protocol
A maximal, continuous, incremental treadmill protocol using the double-pole technique (mostly arm and torso work) was used to measure o2peak, HRpeak, and BLapeak for this ski technique. The test began with a 2-minute warm-up at 2.5 m·s−1 at a constant 7% grade. This grade was chosen to prevent a long coasting time between poling that occurred at lower grades and to elicit exhaustion without excessive over ground speed. The first stage began at 2.6 m·s−1 at a 7% grade for 2 minutes; then, the speed increased 0.2 m·s−1 every 2 minutes until volitional exhaustion. Expired gas, HR, and blood samples were collected as for the o2peak protocol.
Measurement of Blood Lactate
Following a fingerstick, a 25 μL blood sample was collected using heparinized microcentrifuge capillary tubes and was immediately diluted to 1/3 of its original concentration in a 50 μL solution containing Triton X-100 (cell lysing agent) and sodium fluoride (antiglycolytic agent). Although only a single BLa measure was recorded at each exercise intensity, the dilution and storage of the samples allowed for repeated measurement in the event of a measurement error or aberrant result. BLa concentration was determined using a Yellow Springs Instrument 2300 blood lactate analyzer (YSI 2300 Stat Plus; Yellow Springs Instrument, Yellow Springs, OH, USA). The reliability of this analyzer to repeatedly measure blood lactate has been reported to be r = 0.99 with an SEM of 0.06 mmol·L−1 (17). Prethreshold BLa values were calculated by averaging the first 3 BLa measures (which were below LT for all subjects) obtained during each of the 2 submaximal protocols. To determine the HR and o2 at LT, BLa was plotted vs. HR and o2, and values were converted to the log scale to produce a more definitive breakpoint according to the methods of Beaver et al. (1). After this conversion, the statistical analysis software (Statistica, StatSoft, Inc., Tulsa, OK, USA) was used to mathematically determine the LT. In the model used, the LT represents the intersection of upper and lower linear trends fit to the BLa curve. Using similar graded exercise protocols in runners, Weltman et al. (25) demonstrated reliability coefficients of r = 0.82 for the o2 at LT and r = 0.95 for the HR at LT.
Measurement of Oxygen Consumption and Heart Rate
Oxygen consumption was assessed by open-circuit spirometry using a Sensormedics 2900 indirect calorimeter (Sensormedics 2900, Sensormedics Corp., Conshohocken, PA, USA). Oxygen consumption was recorded continuously and reported using 30-second average periods. Heart rate was recorded via telemetry (Vantage XL Heart Rate Monitor, Polar, Kempele, Finland).
All comparisons were made using paired t-tests, and the level of significance was set at p ≤ 0.05 a priori. To control the inflated type I error rate due to the 8 tests performed, a Bonferroni correction was used adjusting the critical p value to p ≤ 0.006. Comparisons of mean for prethreshold BLa, HR at LT, and o2 at LT between Sinc and Ginc protocols were made. Differences in mean o2peak, HRpeak, and BLapeak between maximal double poling and skating were also examined. To compare the HR/o2 relationship between protocols, the slopes and intercepts of linear regression lines fit to the graph of HR vs. o2 were compared. Values are given as mean ± SD. With 11 subjects, the study was sufficiently powered to detect differences in the HR at LT and o2 at LT with power = 0.65 and 0.89, respectively.
Based on the HR (p = 0.63) and o2 (p = 0.53) at LT, there was no significant order effect between the group of subjects that completed Ginc on day 1 (Sinc on day 2) and those that completed Ginc on day 2 (Sinc on day 1). Curvilinear regression demonstrated a leftward shift of the BLa/HR relationship during graded skiing as compared with level skiing (Figure 3). The average prethreshold BLa value for Ginc and Sinc were not statistically different (2.3 ± 0.6 mmol·L−1 vs. 2.1 ± 0.6 mmol·L−1, respectively, p = 0.29) (Figure 4A). o2 at LT during Ginc was 6% lower than that seen during Sinc (46.3 ± 2.8 ml·kg−1·min−1 vs. 49.1 ± 1.6 ml·kg−1·min−1, respectively, p = 0.004) (Figure 4B). The mean HR at LT during Ginc was 4% lower than that during Sinc (155 ± 7 b·min−1 vs. 162 ± 9 b·min−1, respectively, p = 0.001) (Figure 4C). When comparing the work capacity during double poling and skating, the mean double-pole o2peak was 7% lower than the skating o2peak (60.3 ± 2.8 ml·kg−1·min−1 vs. 64.6 ± 1.8 ml·kg−1·min−1, p = 0.0005) (Table 2). Similarly, during the double-pole protocol, HRpeak was significantly lower (188 ± 6 b·min−1) than skating HRpeak (193 ± 6 b·min−1, p = 0.0015). There were however no differences in BLapeak between double poling (12.5 ± 1.8 mmol·L−1) and skating techniques (13.1 ± 1.90 mmol·L−1, p = 0.37). Examination of the linear HR/o2 relationship indicated that there were no significant differences for the slopes (p = 0.38) and intercepts (p = 0.53) between protocols.
An important finding of this study is that BLa was greater for a given exercise HR or oxygen consumption during uphill skate skiing, suggesting a greater dependency on anaerobic metabolism during climbing (Figure 4B, C). This leftward shift of the BLa vs. intensity relationship is only evident at intensities at or above the LT. This indicates that at lower intensities, lactate production and clearance are in equilibrium (Figure 4A). Because the relationship between HR and o2 is not altered by grade during submaximal skate skiing, HR can be considered a consistent indicator of aerobic energy expenditure regardless of terrain. Finally, the reduced HRpeak and o2peak seen during double poling compared with skating illustrates the difference in work capacity between the upper and lower body and provides the basis for comparing the metabolic effects of varying the workload between these segments.
For a given amount of external work, the upper body relies more heavily on anaerobic metabolism, and the present study suggests that grade facilitates this shift during cross-country skiing. Although the contribution of the upper body to forward propulsion during skiing was not measured in the present study, previous biomechanical research shows that the use of the arms increases with grade, and poling during uphill skiing may account for as much as 50% of forward propulsion (2,3,8,19-21). Others have shown that an increase in BLa occurs during double poling and double poling uphill when compared with skate skiing (9,11,14). The greater percentage of glycolytic fibers, lower blood flow, and reduced oxygen-diffusing capacity characterizing the muscle of the upper body are thought to account for the higher BLa production for a given workload in comparison to the lower body (4,15). Therefore, as the percentage of the total workload is shifted toward the upper body musculature, glycolytic rate and lactate production should and did increase. Equally as important, the reduced density of oxidative fibers in the upper body would impair the oxidation of BLa, and clearance of lactate could also be reduced (15). Combined, these factors increase net lactate accumulation when a greater portion of the workload is partitioned to the upper-body musculature, as is the case during uphill skiing. Our observation of a reduced HR and o2 at LT and the leftward movement of the lactate vs. intensity curve complement this theory.
The lack of difference in prethreshold BLa between level and graded skiing indicates that at lower intensities, lactate production and clearance are similar in both conditions. It is likely that the metabolic capacity to produce adenosine triphosphate (ATP) aerobically is sufficient at these intensities, and mechanisms of lactate clearance keep pace with production. Alternatively, the differences in the magnitude of the grade during the early stages of the protocols might have been too little to have an effect on muscle activation patterns and metabolism as it did at the higher grades/intensities. This was demonstrated by the athletes being more likely to use the V2 and V2 alternate techniques at the early stages of each protocol that shifted toward the V1 technique only during the higher grades of the graded protocol. The lack of difference in BLa at low grades indicates that the prescribed HR- and BLa-based training intensities do not have to be adjusted for lower inclines or at intensities below LT when transitioning from level to graded terrain.
As expected, and in agreement with Mahood et al. (12), the present study demonstrated that both the o2peak and HRpeak during double poling were significantly lower (7 and 3%, respectively) than that seen during skating. Thus, at a given absolute workload, the upper body is working at a higher percentage of its peak aerobic capacity and HR, leading to a greater reliance on anaerobic metabolism, which leads to higher exercise BLa levels. This could partially explain the leftward shift in the BLa/HR relationship seen in this study during graded skiing. Interestingly, there were no differences in BLapeak between double poling and skate skiing, indicating that pH tolerance and possibly systemic buffering capacity are not different during upper-body work only and combined upper- and lower-body work.
Terrain variations and climatic conditions do not allow the use of velocity as an accurate measure of intensity during cross-country skiing. Therefore, the use of the BLa/HR relationship is essential to the control of exercise intensity in the field. The leftward shift of the BLa/intensity relationship seen during uphill skiing may cause the athlete to train at a higher than expected BLa when using HR as a guide. This potentially places a greater than desired emphasis on anaerobic energy production during low-intensity or recovery training sessions. For coaches and athletes, it is important to recognize the additional metabolic load during graded skate skiing and adjust the target HR downward or intentionally choose flat or rolling terrain for long slow distance and recovery workouts. This will help assure that skiers recover adequately during easy days and do not overreach or overtrain due an unexpectedly high training volume. The discrepancy between prescribed and actual intensity is more likely to occur if the laboratory or field-based testing protocols were performed on level terrain and training takes place on graded terrain. Verges et al. (23) tested the comparability of LT derived from laboratory running tests to LT derived from field roller skiing and showed a leftward shift of the BLa/HR curve during roller skiing. Similar to the present study, they concluded that the greater contribution of the upper-body musculature may be the cause of this shift and that laboratory testing procedures should more closely mimic training and racing conditions. Therefore, the establishment of BLa- and HR-based training intensity zones should be performed in the field or on a treadmill with protocols that replicate ski course grade profiles.
In the absence of laboratory protocols that mimic field conditions, it may be of interest for coaches and athletes to have the ability to alter training intensity HRs to account for variations in grade. The HR at LT during skiing uphill (8-12% grade) was approximately 7% lower than that during level skiing. To shift training zone HRs from graded skiing to level skiing (or vice versa), the following equation can be used:
where x = any BLa.
For a given lactate level, the value obtained is the number of heart beats that should be added to the target HR when going from graded skiing to level skiing, or the number of heart beats that should be subtracted from the target HR when going from level to graded skiing. This equation is most useful for blood lactate values greater than 2.5 mmol·L−1 as there were no differences in prethreshold BLa between protocols, but its validity should be assessed across a wider range of skiing abilities. For simple estimation purposes or when the data describing the BLa/HR relationship are absent, coaches and athletes may add 5 b·min−1 to the target HR when going from graded to level skiing or subtract 5 b·min−1 when going from level to graded skiing. Alternatively, coaches and athletes may “spot-check” blood lactate levels during ski training over varied terrain using a portable lactate meter to assure that skiers are achieving the desired exercise intensity. These portable meters are becoming increasingly economical, small, portable, and easy to use, making them probable tools for controlling exercise intensity even at the high school or club level.
An assumption of this article, which is supported by the lower o2peak seen during double poling, is that the upper body is less capable of regenerating ATP aerobically than the lower body. The lower extremity musculature is used during almost all types of aerobic training routinely performed by cross-country skiers (skiing, running, hiking, cycling, hill bounding, or ski imitation) whereby the upper-body musculature is only targeted during ski-specific exercises. This is likely to lead to a lower level of aerobic conditioning of the upper-body musculature and may further explain the greater reliance on anaerobic metabolism when work is shifted to this segment. Therefore, a practical recommendation from this research is that cross-country skiers focus a larger portion of their training on ski-specific exercises that target the aerobic conditioning of the upper body by itself or in conjunction with lower-body training. To support this notion, the importance of upper-body conditioning in ski racing can be demonstrated by the work of Mahood et al. (12) who showed that a 1-km uphill double pole time trial strongly predicted 10 km time trial performance (r = 0.92) and overall ranking during a competitive ski season (r = 0.95).
In summary, a greater reliance on the upper-body musculature during uphill cross-country skiing, and the associated lower aerobic capacity of this body segment, decreases both the HR and oxygen consumption at LT and shifts the BLa vs. intensity curve to the left. Ideally, laboratory testing protocols should mimic the predominant terrain experienced by the skier, and in an effort to accurately prescribe training intensity, lactate testing may more appropriately be performed in the field on undulating terrain. Regardless, athletes, coaches, and researchers should be aware of the change toward anaerobic metabolism experienced during uphill skiing to properly prescribe and monitor training intensity as well as interpret laboratory results. Additionally, greater emphasis on upper-body training in cross-country skiers may increase the aerobic capacity of this body segment and mitigate the effects of grade seen in this article. The alteration in the BLa/HR relationship described here is specific to Nordic skate skiing, but the findings may be relevant to other activities that distribute workload between the upper and lower body.
The authors would like to thank Joohee Im, PhD, for her help with data collection. This study was supported by a Sport Science and Technology Grant from the United States Olympic Committee.
1. Beaver, WL, Wasserman, K, and Whipp, BJ. Improved detection of lactate threshold
during exercise using a log-log transformation. J Appl Physiol
59: 1936-1940, 1985.
2. Bilodeau, B, Boulay, MR, and Roy, B. Propulsive and gliding phases in four cross-country skiing techniques. Med Sci Sports Exerc
24: 917-925, 1992.
3. Boulay, MR, Rundell, KW, and King, DL. Effect of slope variation and skating technique on velocity in cross-country skiing. Med Sci Sports Exerc
27: 281-287, 1994.
4. Calbet, JAL, Holmberg, H-C, Rosdahl, H, van Hall, G, Jensen-Urstad, M, and Saltin, B. Why do the arms extract less oxygen than legs during exercise? Am J Physiol Regul Integr Comp Physiol
289: R1448-R1458, 2005.
5. Coyle, EF. Integration of the physiological factors determining endurance performance ability. Exerc Sport Sci Rev
23: 25-63, 1995.
6. Eisenman, PA, Johnson, SC, Bainbridge, CN, and Zupan, MF. Applied physiology of cross-country skiing. Sports Med
8: 67-79, 1989.
7. Farrell, PA, Wilmore, JH, Coyle, EF, Billing, JE, and Costill, DL. Plasma lactate accumulation and distance running performance. Med Sci Sports Exerc
11: 338-344, 1979.
8. Gregory, RW, Humphreys, SE, and Street, GM. Kinematic analysis of skating techniques of Olympic skiers in the women's 30-k race. J Appl Biomech
10: 382-292, 1994.
9. Hoffman, MD, Clifford, PS, Snyder, AC, O'Hagan, KP, Mittelstadt, SW, Roberts, MM, Drummond, HA, and Gaskill, SE. Physiological effects of technique and rolling resistance in uphill roller skiing
. Med Sci Sports Exerc
2: 311-317, 1998.
10. Hooker, SP, Wells, CL, Manore, MM, Philip, SA, and Martin, N. Differences in epinephrine and substrate response between arm and leg exercise. Med Sci Sports Exerc
22: 779-784, 1990.
11. Larson, AJ. Variations in heart rate at blood lactate threshold
due to exercise mode in elite cross-country skiers. J Strength Cond Res
20: 855-860, 2006.
12. Mahood, NV, Kenefick, RW, Kertzer, R, and Quinn, TJ. Physiological determinants of cross-country ski racing performance. Med Sci Sports Exerc
33: 1379-1384, 2001.
13. Millet, GY, Hoffman, MD, Candau, RB, and Clifford, PS. Poling
forces during roller skiing
: Effects of grade. Med Sci Sports Exerc
30: 1637-1644, 1998.
14. Mittelstadt, SW, Hoffman, MD, Watts, PB, O'Hagan, KP, Sulentic, JE, Drobish, KM, Gibbons, TP, Newbury, VS, and Clifford, PS. Lactate response to uphill roller-skiing: Diagonal stride versus double pole techniques. Med Sci Sports Exerc
27: 1563-1568, 1995.
15. Mygind, E. Fibre characteristics and enzyme levels of arm and leg muscles in elite cross-country skiers. Scand J Med Sci Sports
5: 76-80, 1995.
16. Mygind, E, Andersen, LB, and Rasmussen, B. Blood lactate and respiratory variables in elite cross-country skiing at racing speeds. Scand J Med Sci Sports
4: 243-251, 1994.
17. Rodriquez, FA, Banquells, M, Pons, V, Drobnic, F, and Galilea, PA. A comparative study of blood lactate analytic methods. Int J Sports Med
13: 462-466, 1992.
18. Rundell, KW and Bacharach, DW. Physiological characteristics and performance of top U.S. biathletes. Med Sci Sports Exerc
27: 1302-1310, 1995.
19. Smith, GA and Heagy, BS. Kinematic analysis of skating technique of Olympic skiers. J Appl Biomech
10: 79-88, 1994.
20. Smith, GA, McNitt-Gray, J, and Nelson, RC. Kinematic analysis of alternate stride skating in cross-country skiing. Int J Sports Biomech
4: 49-58, 1988.
21. Smith, GA, Nelson, RL, Feldman, A, and Rankinen, JL. Analysis of V1 skating technique of Olympic cross-country skiers. Int J Sports Biomech
5: 185-207, 1989.
22. Toner, MM, Glickman, EL, and McArdle, WD. Cardiovascular adjustments to exercise distributed between the upper and lower body. Med Sci Sports Exerc
22: 773-778, 1990.
23. Verges, S, Flore, P, and Favre-Juvin, A. Blood lactate concentration/heart rate relationship: Laboratory running test vs. field roller skiing
test. Int J Sports Med
24: 446-451, 2003.
24. Verges, S, Flore, P, Laplaud, D, Guinot, M, and Favre-Juvin, A. Laboratory running test vs. field roller skiing
test in cross-country skiers: A longitudinal study. Int J Sports Med
27: 307-313, 2006.
25. Weltman, A, Snead, D, Stein, P, Seip, R, Schurrer, R, Rutt, R, and Weltman, J. Reliability and validity of a continuous incremental treadmill protocol for the determination of lactate threshold
, fixed blood lactate concentrations, and VO2
max. Int J Sports Med
11: 26-32, 1990.