European Journal of Applied Physiology

, Volume 91, Issue 1, pp 94–99

Prediction of sprint triathlon performance from laboratory tests

Authors

  • R. Van Schuylenbergh
    • Exercise Physiology and Biomechanics Laboratory, Department of Kinesiology, Faculty of Physical Education and Physiotherapy K.U. Leuven
  • B. Vanden Eynde
    • Exercise Physiology and Biomechanics Laboratory, Department of Kinesiology, Faculty of Physical Education and Physiotherapy K.U. Leuven
    • Exercise Physiology and Biomechanics Laboratory, Department of Kinesiology, Faculty of Physical Education and Physiotherapy K.U. Leuven
Original Article

DOI: 10.1007/s00421-003-0911-6

Cite this article as:
Van Schuylenbergh, R., Eynde, B.V. & Hespel, P. Eur J Appl Physiol (2004) 91: 94. doi:10.1007/s00421-003-0911-6

Abstract

This study investigated whether sprint triathlon performance can be adequately predicted from laboratory tests. Ten triathletes [mean (SEM), age 21.8 (0.3) years, height 179 (2) cm, body mass 67.5 (2.5) kg] performed two graded maximal exercise test in random order, either on their own bicycle which was mounted on an ergometer or on a treadmill, to determine their peak oxygen consumption (O2peak). Furthermore, they participated in two to three 30-min constant-load tests in both swimming, cycling and running to establish their maximal lactate steady state (MLSS) in each exercise mode. Swim tests were performed in a 25-m swimming pool (water temperature 27°C). During each test heart rate (HR), power output (PO) or running/swimming speed and blood lactate concentration (BLC) were recorded at regular intervals. Oxygen uptake (O2) was continuously measured during the graded tests. Two weeks after the laboratory tests all subjects competed in a triathlon race (500 m swim, 20-km bike, 5-km run) [1 h 4 min 45 s (1 min 38 s)]. Peak HR was 7 beats·min−1 lower in the graded cycle test than in the treadmill test (p<0.05) at similar peak BLC (~10 mmol·l−1) andO2peak (~5 L·min−1). High correlations were found betweenO2peak during cycling (r=−0.71, p<0.05) or running (r=−0.69, p<0.05) and triathlon performance. Stepwise multiple regression analysis showed that running speed and swimming speed at MLSS, together with BLC in running at MLSS, yielded the best prediction of performance [1 h 5 min 18 s (1 min 49 s)]. Thus, our data indicate that exercise tests aimed to determine MLSS in running and swimming allow for a precise estimation of sprint triathlon performance.

Keywords

CyclingMaximal lactate steady stateRace resultsRunningSwimming

Introduction

The popularity of triathlon is rapidly increasing, in particular the shorter events, such as the Olympic distance (1500 m, 40 km, 10 km) or the sprint triathlon (500 m, 20 km, 5 km). Triathlon competition time ranges from 50–70 min for the sprint events to several hours for the Olympic and long distance races. Accordingly, top level triathletes are characterized by a very high aerobic power (maximal rate of oxygen consumption). Peak oxygen uptake in running (O2peak) in elite triathletes as a rule is in the range of 70–90 ml O2 · min−1 · kg−1 (O'Toole et al. 1987, 1989; Schneider and Pollack 1991; Basset and Boulay 2000; Hue et al. 2000). Corresponding values in cycling are slightly lower (O'Toole et al. 1989; Schneider and Pollack 1991; Basset and Boulay 2000; Hue et al. 2000). Still,O2peak has not appeared to be a good predictor of triathlon performance in elite triathletes (O'Toole et al. 1987; Le Gallais et al. 1999). Submaximal exercise intensity markers such as the so-called anaerobic threshold (Schabort et al. 2000) or the economy of effort (Dengel et al. 1989; Toussaint 1991; Miura et al. 1997) conceivably yield a higher validity to predict triathlon performance. In this respect, the maximal lactate steady state (MLSS) is generally acknowledged to be a good marker of functional aerobic power during prolonged exercise. Indeed, the MLSS represents the highest exercise intensity which can be performed in the absence of progressively increasing blood lactate concentration, which means that oxidative energy metabolism, including oxidation of the lactate produced, accounts for the bulk of energy provision in active muscles (Heck et al. 1985; Beneke 1995).

Therefore, the primary purpose of this study was to evaluate whether sprint triathlon performance can be adequately predicted from laboratory exercise tests aimed to assess MLSS exercise intensity. The MLSS is presumably specific to the exercise mode (Beneke and von Duvillard 1996; Beneke et al. 2001). Thus, separate tests were performed to determine the exercise intensities and heart rates corresponding with the MLSS in swimming, cycling and running, respectively.

Methods

Subjects

Ten male physical education students volunteered for this study after being informed about the nature and risks involved in participating in the experiments. All subjects had been consistently involved in triathlon training for at least 1 year and they had competed in at least three races prior to participation in the study. All subjects were tested in April, which means approximately 1 month before the start of triathlon competitions in Belgium. They were 21.8 (0.3) (range: 20.3–22.9) years old and their body weight and height were 179 (2) cm, and 67.5 (2.5) kg, respectively.

Protocol

Study protocol

All subjects performed 8 to 11 exercise tests over a 3-week period. A 2-day rest period, during which the subjects were instructed to only perform light recovery training for 90 min at most, interspersed the tests. Furthermore, to facilitate the replenishment of carbohydrate stores in between the tests, the subjects received specific instructions to increase their dietary carbohydrate intake during the period of the study. In addition, 2–3 h before each test they received a standardized carbohydrate-rich meal (1500 kcal: 85% carbohydrate, 10% protein, 5% fat). The first two tests were graded exercise tests on a bicycle ergometer and on a treadmill to determine peak oxygen in cycling and running, respectively. Thereafter the subjects performed six to nine 30-min constant-load tests (two to three tests per sports discipline), to precisely determine the exercise intensities and heart rates corresponding with their MLSS in swimming, cycling and running, respectively. The latter tests were performed in random order. During the tests in the air-conditioned laboratory (19°C) the subjects were continuously cooled by two frontal fans [air velocity 2.9 (0.2) m·s−1]. The swim tests were performed in the local 25-m indoor swimming pool with a constant water temperature at 27°C.

Graded tests

The subjects performed a graded cycling test to volitional exhaustion on their own race bicycle, which was mounted on an electromagnetically braked ergometer (Avantronic Recordtrainer, Leipzig, Germany).

After a 20-min warm up at 100 W, power increments were made every 6 min by two-thirds of the subject's body weight expressed in watts. To facilitate optimal power output, the subjects were allowed to use their routine cadence [95 (3) revolutions·min−1] (Swain and Wilcox 1992; Chavarren and Calbet 1999). Power output was continuously measured by the ergometer. The graded running test until exhaustion was performed on a motor-driven treadmill (Woodway ELG2, Weil am Rhein, Germany) which was calibrated prior to the experiments. The slope of the ergometer was set at 1% to simulate outdoor running conditions (Heck et al. 1985). The speed at warming up was set at 2.5 m·s−1 for 20 min. Thereafter the speed was increased by 0.5 m·s−1 every 6 min. During both the bicycle and the treadmill tests heart rate (HR) was continuously monitored (Polar, Kempele, Finland), and oxygen uptake, carbon dioxide output and ventilation were measured using a breath-by-breath open circuit system (Jaeger Oxycon Alpha, Hoechberg, Germany). Furthermore, capillary blood samples were taken from a hyperaemic earlobe at 3-min intervals for determination of lactate concentration (Analox LM07, London, UK).

Constant-load tests

Each MLSS trial started with a 20-min standardized warm up eliciting a HR of 120–130 (swimming and cycling) or 135–145 (running) beats·min−1. In swimming, the subjects were instructed to sustain the highest possible constant speed for 30 min. The mean swimming speed between 100 m and 200 m of the first trial was used to pace the swimmer during the test. The subjects were paced every length by visual feedback to ensure constant swimming speed. If fluctuations in swimming speed were above 0.01 m·s−1 the trial was cancelled. For the cycling and running MLSS tests, the workload in the first trial was set at the power (cycling) or speed (running) corresponding to the 4-mmol·l−1 lactate threshold as determined during the corresponding graded exercise tests. This intensity was previously shown to approximate MLSS well (Heck et al. 1985). Heart rate (Polar), workload (Avantronic Recordtrainer) or speed (Woodway ELG2) were continuously recorded. Lactate samples were taken from a hyperaemic earlobe (Forapin) at 10-min intervals, which required a 30-s break in the swimming and running tests but not in the cycling trials. In the next constant-load test the workload (cycling) or speed (swimming, running) was adjusted to meet the MLSS criterion. Thus, if blood lactate concentration was found to decrease or was stable, exercise intensity was increased by 10–20 watt (cycling), 0.1–0.25 m · s−1 (running) or 0.01–0.06 m · s−1 (swimming). Conversely, if blood lactate showed a progressive increase throughout the test, intensity was decreased by 10–20 watt (cycling), 0.1–0.25 m · s−1 (running) or 0.01–0.06 m · s−1 (swimming). The MLSS was defined as the highest blood lactate concentration that increased no more than by 1.0 mmol·l−1 during the final 20 min of the 30-min constant workload test (Heck et al. 1985). If this criterion for a given discipline was not met after two tests, an additional test was performed.

Calibration of the bicycle ergometer

Two ergometers were used in this study. Prior to the experiments the ergometers were calibrated using an isokinetic dynamometer, which was instrumented with a torque transducer (Lebow 1605, 0.05% accuracy level) and connected to the front chain ring of the bike. The axis of the dynamometer was aligned with the axis of the crank. In the MLSS power range (250–350 watt) and at the cadences (90–110 revolutions·min−1) and gear ratio (53×16) used in this study, ergometer A power measurements showed a 3.6% to 6.5% error, versus 3.1% to 10.3% for ergometer B. At the end of the experiments the entire calibration procedure was repeated. Power curves were found to be identical to those before the experiments for both ergometers. Thus, the error of measurement was constant throughout the experiments, and within each ergometer. Still, to control for this error of measurement, each subject performed all cycling tests on the same ergometer. Therefore, the bicycle power measurements made throughout this study are valid and reproducible.

Triathlon race performance

Within 2 weeks after the last laboratory test all subjects competed in the national university triathlon championship. Weather conditions were mild (no rain, 19°C). The race consisted of a 500-m swim, a 20-km bike race and a 5-km run. Drafting was allowed. The time to complete the race was recorded by the race organizers to the nearest second.

Data analysis

Determination of lactate thresholds

For both graded tests the HR, oxygen uptake and power output or speed corresponding to the 4-mmol·l−1 blood lactate (TH-La4) were determined by linear interpolation of the power/speed:lactate curve. The HR, blood lactate concentration (BLC), oxygen uptake and power output or speed corresponding the individual lactate-threshold (TH-Dm) was determined using the Dmax-method adapted from Cheng and co-workers (Cheng et al. 1992). Briefly, for each individual, the lactate curve was first fitted using a third-degree polynomial equation. Thereafter the intersection of the best-fit curve and the tangent parallel to the straight line linking the lowest and the highest lactate concentration measured was taken as TH-Dm. Furthermore, HR, BLC, oxygen uptake and power output or speed corresponding to the lactate threshold (TH-Lt), wherein the lactate threshold was defined as the first breakpoint in the lactate curve (Cabrera and Chizeck 1996), were calculated. In addition, after the power/speed corresponding to MLSS intensity had been precisely established by the constant-load tests, HR, BLC and oxygen uptake corresponding to MLSS intensity during the graded tests were calculated by linear interpolation of the power/speed:lactate curve (MLSSe). Gross mechanical efficiency was calculated using the equations of Lusk at MLSSe for the cycling test and the treadmill test (Lusk 1928).

Statistical analyses

The results are expressed as means (SEM). Pearson correlation coefficients between the physiological measurements during the incremental and constant-load tests on the one hand, and triathlon race time on the other hand, were calculated. Furthermore, stepwise multiple regression was performed with triathlon performance (expressed in minutes) as the dependent variable and the physiological and anthropometric measurements as independent variables.

Comparisons between cycling and running for the different submaximal and peak measurements were performed by means of a 2 [sport (cycling, running)] × 4 [measurements (LT, LT4, MLSSe, Max)] way analysis of variance. Comparison of HR and BLC corresponding to MLSS in swimming, cycling and running was made using three-way analysis of variance with repeated measures (T10, T20, T30). Where appropriate, post hoc analysis was performed using a Tukey HSD test. Statistical analyses were carried out using a statistical software package (Statsoft, Statistica 5.5, Tulsa, USA). For all statistics, the significance level was set at p<0.05.

Results

Graded exercise tests

Measurements at submaximal exercise

Heart rate, BLC and rate of oxygen uptake corresponding to TH-Lt, TH-Dm and TH-La4 were determined during a graded exercise test on both the bicycle ergometer and the treadmill. At TH-Lt, TH-Dm and TH-La4, heart rate on average was 12–18 beats·min−1 lower during the cycle test than during the treadmill test (p<0.05), whilst blood lactate levels were similar (~2 mmol·l−1 at TH-Lt versus ~4 mmol·l−1 at TH-Dm). Compared with cycling, oxygen uptake at TH-Lt intensity was about 20% higher in treadmill running (p<0.05). However at TH-Dm and TH-La4 oxygen uptake was similar between tests. At the workload corresponding to the MLSS, which was determined by means of the 30-min constant-load tests, heart rate and oxygen uptake were similar in cycling and running. However, blood lactate was markedly higher during cycling than during running (p<0.05) (Table 1).
Table 1.

Heart rate, blood lactate concentration and oxygen uptake (V̇O2), at the lactate threshold (TH-Lt), the Dmax method threshold (TH-Dm), the 4 mmol·l–1 lactate threshold (TH-La4), MLSS workload (MLSSe) and at exhaustion during the graded bicycle and treadmill tests. The exercise intensity corresponding to MLSS was determined during a 30-min constant-load test. Thereafter heart rate, blood lactate and oxygen uptake (V̇O2) corresponding to this intensity during the graded exercise test were calculated (MLSSe)

Parameter

Bicycle

Treadmill

Mean (SEM) (n=10)

Pearson correlation coefficienta

Mean (SEM) (n=10)

Pearson correlation coefficienta

TH–Lt

HR (beats · min–1)

147 (3)

0.01

165 (3)$†

0.02

BLC (mmol · l–1)

1.9 (0.2)

0.54

2.1 (0.1)

0.01

O2 (L · min–1)

3.2 (0.2)

–0.76*

3.8 (0.2)$†

–0.51

TH–Dm

HR (beats · min–1)

170 (3)

–0.02

182 (3)$

0.02

BLC (mmol · l–1)

4.3 (0.3)

–0.02

4.1 (0.3)

–0.52

O2 (l · min–1)

4.0 (0.2)

–0.70*

4.3 (0.2)

–0.59

TH–La4

HR (beats · min–1)

165 (3)

–0.10

183 (3)$

0.26

O2 (l · min–1)

4.3 (0.2)

–0.79*

4.4 (0.3)

–0.53

MLSSe

HR (beats · min–1)

176 (3)

0.09

179 (3)

–0.45

BLC (mmol · l–1)

5.7 (0.5)

–0.11

3.9 (0.4)$†

–0.73*

O2 (l · min–1)

4.3 (0.2)

–0.77*

4.4 (0.3)

–0.72*

Exhaustion

HR (beats · min–1)

193 (3)

0.37

200 (2)$†

0.33

BLC (mmol · l–1)

10.7 (0.7)

0.24

10.2 (0.6)

–0.37

O2peak (l · min–1)

5.0 (0.2)

–0.71*

5.1 (0.2)

–0.69*

aPearson correlation coefficients with triathlon race time

$p<0.05 compared with cycling; p<0.05 compared with MLSS; *a significant correlation with triathlon race time (p<0.05)

Measurements at peak exercise

Time to exhaustion was 34 min 54 s (1 min 6 s) for the cycle test and was slightly shorter for the treadmill test 31 min 36 s (1 min 30 s) (p<0.05). Peak heart rate on average was 7 beats·min−1 higher in the run test than in the cycle test (p<0.05) at similar peak lactate concentration (~10 mmol·l−1) and peak oxygen uptake (~5 l·min−1).

Constant-load MLSS tests

During the 30-min constant-load tests BLC and heart rate were measured at 10-min intervals. In each exercise mode (swimming, cycling, running) BLC were constant between min 10 and min 30. However, lactate levels were highest during the swim test (~8.0–8.5 mmol·l−1), slightly lower in the cycling test (~6.5–7.0 mmol·l−1, p<0.05), and lowest in the treadmill test (~5.0–5.5 mmol·l−1, p<0.05 compared with swimming and cycling). In each exercise mode heart rate increased by about 5–6 beats·min−1 from min 10 to min 30. Heart rates were similar between swimming and cycling (~165–170 beats·min−1), yet were higher during treadmill running (~175–180 beats·min−1) (Table 2).
Table 2.

Heart rate and blood lactate concentration during the MLSS tests in swimming, cycling and running. Values are mean (SEM) of 10 observations. A 30-min constant-load test was performed. Heart rate and blood lactate concentrations were measured at 10-min intervals (T10, T20, T30)

T0

T10

T20

T30

Blood lactate (mmol · l–1)

  Swimming

5.0$†

(0.8)

8.1$

(0.6)

8.5$

(0.6)

8.0$

(0.8)

  Cycling

3.7*

(0.3)

7.1*

(0.6)

7.2*

(0.6)

6.6*

(0.6)

  Running

2.7$†*

(0.3)

5.3*$

(0.4)

5.6*$

(0.4)

5.2*$

(0.3)

Heart rate (beats · min–1)

  Swimming

125†$

(3)

165

(3)

169

(3)

171

(3)

  Cycling

145*

(2)

165

(3)

170

(3)

172

(3)

  Running

155*$†

(4)

173*$

(4)

177*$

(4)

179*$

(3)

*p<0.05 compared with swimming, $p<0.05 compared with cycling, p<0.05 compared with T10

Relationship between triathlon performance and laboratory exercise test measurements

Pearson correlations between HR, BLC and oxygen uptake values measured at submaximal and maximal work intensities during the graded exercise tests were correlated with triathlon performance (~ race time). High correlations were found between oxygen uptake and triathlon performance at TH-Lt, TH-Dm, TH-La4, MLSSe and at exhaustion during the bicycle test (r=−0.70 to −0.79, p<0.05). Correlations were lower for the run test (r=−0.51, n.s. to −0.73, p<0.05). Blood lactate concentration corresponding to MLSS exercise intensity during the treadmill test, but not during the cycle test, yielded a high correlation with triathlon performance (r=−0.73, p<0.05). There were no significant correlations between the heart rate at TH-Lt, TH-Dm, TH-La4, MLSSe and at exhaustion during either the cycle or the run test, and triathlon performance (r<−0.50, n.s.). Stepwise multiple regression analysis, including all physiological measures from the graded and constant load tests (data from Tables 1 and 2), indicated running speed at MLSS as the primary predictor of triathlon performance (r2=0.68). Swimming speed at MLSS and BLC during running at MLSS intensity together accounted for the variance in triathlon performance (r2=0.98) (Table 3). When only the results from the graded exercise tests were included in the regression analysis, oxygen uptake at TH-La4 during cycling emerged as the primary determinant of performance. Peak BLC during the treadmill test was an additional factor that contributed significantly to explain the variance in triathlon race time (r2=0.77) (Table 4).
Table 3.

Stepwise multiple regression analysis with triathlon performance as the dependent variable and the measured physiological parameters as independent variables. (Vmlss-run running velocity at MLSS, Vmlss-swim swimming velocity at MLSS, BLCmlss-run blood lactate concentration at MLSS, SEE standard error of estimate.) See Methods for further details

Step

Stepwise multiple regression analysis

R2

SEE

1

Triathlon performance (min)=110.8–11.4·Vmlss–run (m · s–1)

0.68

3.23

2

Triathlon performance (min)=137.3–9.2·Vmlss–run (m · s–1)–35.6·Vmlss–swim (m · s–1)

0.93

1.69

3

Triathlon performance (min)=130–9.2·Vmlss–run (m · s–1)–25.9·Vmlss–swim (m · s–1)+1.4·BLCmlss–run (mmol · l–1)

0.98

0.95

Table 4.

Stepwise multiple regression analysis with triathlon performance as the dependent variable and the measured physiological variables during graded bicycle and treadmill testing as independent variables. (BLCpeak-run Peak blood lactate concentration during a graded treadmill test, V̇O2 TH-LA4cyc oxygen uptake during a graded bicycle test at the 4 mmol lactate threshold, SEE standard error of estimate.) See Methods for further details

Step

Stepwise multiple regression analysis

R2

SEE

1

Triathlon time (min) =95.54–7.86·O2 TH–La4cyc (l·min–1)

0.63

3.71

2

Triathlon time (min) =104.95–7.51·O2 TH–La4cyc (l·min–1)–1.05·BLCpeak–run (mmol·l–1)

0.77

2.77

Discussion

The main result of present study is that sprint triathlon performance can be adequately predicted from laboratory exercise tests aimed to assess the exercise intensities corresponding to the MLSS in swimming, cycling and running. Our data demonstrate that MLSS tests can indeed yield a very accurate prediction of sprint triathlon performance. Running and swimming speed at MLSS in conjunction with BLC during running at MLSS pace were selected as the strongest set of predictors of triathlon performance and explained nearly all the variance in triathlon finishing time (~98%).

Running speed at MLSS was selected as the strongest predictor of global triathlon performance, which is compatible with earlier observations by Landers et al. (2000). In world class triathletes during an Olympic distance triathlon they found running performance to be much more variable than swimming or cycling performances. This finding is compatible with running performance being the primary determinant of success in high-level short distance triathlon races. In keeping with most (Miura et al. 1997; Zhou et al. 1997; Schabort et al. 2000) but not all (Sleivert and Wenger 1993) previous observations, we also found rather high correlations between peak oxygen uptake rate (l·min−1) and triathlon time (−0.69 to −0.71, p<0.05). Still, MLSS indices rather than peak oxygen uptake were selected by the multiple regression analysis to explain the variance in triathlon time. Indeed, it is well documented that in a population of well trained athletes, submaximal exercise intensity markers are more closely related to triathlon performance than peak oxygen uptake measurements (Lucia et al. 1998; Demarle et al. 2001). Furthermore, the subjects enrolled in the present study were not elite triathletes, but physical education students who were involved in other high-intensity sport activities as well as their daily specific triathlon training. This unspecific training pattern may help to explain the high level of aerobic power in our subject population (peak oxygen uptake), yet at the same time the limited value of their peak oxygen uptake values for predicting their triathlon performance.

Similar to the observations of Beneke at al. (2000), we found no significant correlation between BLC during the MLSS treadmill test on the one hand, and the corresponding MLSS running speed or triathlon performance on the other hand. Despite this, multiple regression analysis revealed that the blood lactate level during MLSS treadmill testing could predict a fraction of the variance in triathlon time. According to our data, the higher BLC during running at MLSS corresponds with slower triathlon performance. Such correlation obviously does not indicate that the steady-state blood lactate level during running at MLSS is causally implicated in triathlon performance. However, it is tempting to speculate that the higher MLSS lactate level may reflect a greater proportion of fast-twitch muscle fibers relative to slow fibers. Along the same line of reasoning, a higher proportion of slow-twitch muscle fibers may conceivably predispose to better endurance exercise performance.

It is interesting to note that none of the measurements during bicycle ergometry were selected by the regression analysis to predict triathlon time. This is probably because competitors are tactically allowed to take advantage of drafting during the bike race in a triathlon. By drafting, the energy cost of cycling is remarkably reduced and the subsequent running performance and triathlon outcome are improved (Hausswirth et al. 1999, 2001). Thus, the weight of physiological determinants of cycling ability to explain actual cycling performance during racing is markedly reduced.

Our current observations, together with earlier ones (Beneke and von Duvillard 1996; Beneke et al. 2001) clearly show that the BLC occurring during exercise at MLSS intensity is very different between exercise modes. Lactate concentrations during swimming, cycling, and running were largely different within individuals, with the highest levels occurring during swimming. Steady-state BLC in swimming, cycling, and running ranged over 5.8–11.0 mmol·l−1, 3.6–9.5 mmol·l−1 and 3.7–6.9 mmol·l−1, respectively. By analogy, steady-state lactate levels have been reported to show large intra-individual variation between rowing, skating, and cycling (Beneke and von Duvillard 1996; Harnish et al. 2001). Another important observation is that in most of our subjects the steady-state lactate concentrations measured during MLSS exercise exceeded the 4 mmol·l−1 lactate threshold, previously proposed to be a valid index of MLSS intensity (Heck et al. 1985). Thus, clearly "fixed" lactate thresholds (e.g. 4 mmol·l−1 lactate threshold) are irrelevant for estimating MLSS in individual athletes.

In conclusion, our data show that sprint triathlon performance in moderately trained triathletes can be precisely predicted from constant-load tests aimed to determine the exercise intensities corresponding to MLSS in running and swimming. Furthermore, heart rates and blood lactate levels during swimming, cycling, and running at MLSS intensity are very different within a given individual. Thus, if heart rates are to be used to monitor training intensity, adequate exercise testing should involve separate tests for swimming, cycling, and running.

Acknowledgements

The authors thank Veronique Colman and Ulrich Persyn for their skillful technical assistance during the experiments.

Copyright information

© Springer-Verlag 2003