Introduction

Endurance performance, as in long distance running, imposes great demands on both the cardiovascular system and the employed locomotor organs. A model described by Pate and Kriska (1984) incorporates three major factors accounting for interindividual variance in aerobic endurance performance, namely maximal oxygen consumption (VO2max), lactate threshold (LT) and work economy. This model is supported by numerous published studies (Pollock 1977; Farell et al. 1979; Conley and Krahenbuhl 1980; Di Pampero et al. 1986; Bunc and Heller 1989; Helgerud 1994). Maximal aerobic speed (MAS), or the minimum speed needed to reach VO2max (vVO2max) has been explained by Morgan et al. (1989) as the VO2max divided by C R. Noakes et al. (1990) have shown that maximal speed reached during VO2max testing correlates better with performances for distances between 10 and 90 km than VO2max alone. In accordance with this, Lacour et al. (1990) found that C R or VO2max alone did not correlate with race velocity amongst well-trained middle and long distance runners (800–5,000 m), whereas MAS correlated well with the race velocity for distances longer than 800 m. The MAS has been found to represent 1,500–3,000-m velocity in elite and sub-elite distance runners (Lacour et al. 1990; Billat and Koralsztein 1996; Bassett et al. 2003). The time to exhaustion at MAS in these studies was approximately 3–9 min.

CR is commonly defined as the steady rate VO2 in ml kg−1 min−1 at a standard velocity (Costill et al. 1973; Conley and Krahenbuhl 1980) or as energy cost of running per metre (ml kg−1 m−1) (Di Pampero et al. 1986; Helgerud 1994; Helgerud et al. 2001). It is also expressed as mechanical efficiency, which is referred to as the ratio between work output and oxygen cost (Åstrand and Rodahl 1986).

Conley and Krahenbuhl (1980) and Helgerud (1994) have shown inter-individual variations in the oxygen cost of running (C R). The causes of this variability are not well understood, but it seems likely that anatomical trait, mechanical skill, neuromuscular skill and storage of elastic energy are important factors (Pate and Kriska 1984). Interindividual variability in C R reflects the importance of this parameter for performance in long distance running. Interindividual variability in C R expressed as standard deviations (SD) is reported by Di Pampero et al. (1986) to be ±8%, if the C R is expressed per kg of body mass. This is in line with the results from Morgan et al. (1995) and Lacour et al. (1990). These results are also expressed per kg body mass.

Energy cost for movement does not increase in the same rate as body mass (Bergh et al. 1991; Eisenmann et al. 2001; Berg 2003). According to Helgerud (1994), a lack of allometric scaling will underestimate VO2max and overestimate C R amongst the heavier runners. Consequently, part of the variation in C R in the experiments by Di Pampero et al. (1986); Morgan et al. (1995) and Lacour et al. (1990) are due to body mass differences. To use the allometric scaling when expressing C R has seemed to reveal smaller differences. Helgerud (1994) has reported SD for C R of approximately 5% within intermediate national elite marathon runners when expressing C R per kg body weight raised to the power of 0.75 and metre. In accordance with this, Helgerud et al. (2001, 2003) have shown SD for C R to be 5% in junior soccer players and 3–4% in adult players including a group of players at European Champions League level. Hoff et al. (2005) report SD for C R to be <5% amongst 36 professional soccer players.

If measurements of C R at a sub-maximal running velocity are to be used to evaluate race performance capacity amongst long distance runners, the measurements should be representative for the C R at racing velocity. Previous investigations have shown C R to be independent of running velocity up to intensities close to VO2max for long distance runners (Di Pampero et al. 1986; Helgerud 1994). However Daniels and Daniels (1992) report increasing C R at increasing relative running velocities. They also found elite middle distance runners to have better C R at velocities at or above marathon pace than elite long distance runners. The long distance runners had better C R at velocities below marathon pace than the middle distance runners.

In most studies that have investigated C R at different velocities, an incremental protocol with different sub-maximal work periods more than 3–10 min each, have been used (Di Pampero et al. 1986; Bunc and Heller 1989; Lacour et al. 1990; Helgerud 1994). A possible cause of error in these protocols is the incremental order in which the runs have been performed. In this context, it should be of interest to assess to what extent VO2 at one velocity is affected by previously completed runs.

The aim of the present study was to investigate if there are differences in running economy at different velocities for well-trained distance runners, and to what extent a commonly used incremental protocol for measuring oxygen uptake (VO2) at different velocities affects the reliability of these measurements.

The hypothesis of the present study was

  1. 1.

    There are no differences in C R at different velocities in male and female distance runners between 75 and 90% VO2max.

  2. 2.

    The commonly used incremental protocol for measuring VO2 at different intensity levels is reliable up to 90% VO2max.

Materials and methods

Subjects

Fifteen moderately to well-trained runners, 9 males and 6 females, aged 29.3 ± 7.0 years, with an average VO2max of 65.2 ± 10.4 ml kg−1 min−1 were included in the study, after giving their written consent to participate.

Test procedures

The subjects were tested on two different days, with a minimum of 1 day and a maximum of 7 days of rest or easy training in between. The first day consisted of measurements of heart rate (HR), blood lactate concentration ([La−]b) and oxygen consumption (VO2) during 5-min runs (1.5% inclination) at different velocities until exhaustion. The subjects started with a velocity assumed to be about 60% of their VO2max. This was either 8.0 or 9.5 km h−1. After completing 5 min at this velocity, the speed was increased by 1.5 km h−1 to the next 5-min work period. New steps were performed until the subject no longer could complete 5 min at the desired velocity. After 60 min of rest, a VO2max test was performed, using an incremental protocol at 5.2% inclination. The VO2max test terminated at voluntary fatigue by the subjects. HR (≥98% predicted HRmax), R (≥1.05) and [La−]b (≥8.0 mmol L−1) values, as well as a possible plateauing of the VO2 curve, was used to evaluate if VO2max was obtained. LT and MAS were calculated on the basis of these measurements. The second day of testing consisted of measurements of HR, [La−]b and VO2 at the same velocities as the first day, but performed in the opposite order (i.e. starting with the highest running velocity after a warm up). Following a 60-min rest, the subjects then performed a run to exhaustion at MAS. The protocol for one of the runners is presented in Table 1.

Table 1 Test protocol for one of the runners
Full size table

Protocol for one of the runners (Table 1)

The VO2 was measured using the metabolic test system, Sensor Medics Vmax Spectra (SensorMedics 229 CA, USA). The lactate measurements were performed using an Arkray Lactate Pro LT-1710 analyser (whole blood) (Arkray Inc. Kyoto, Japan). The LT was defined as the warm up [La−]b value (i.e. measured after the lowest velocity in day 1) + 1.5 mmol. This is in line with the protocol proposed by Helgerud et al. (1990) consisting of several 5-min steps, at 1.5% inclination, increasing the speed by 1.5 km h−1 after each step. The oxygen cost of running was measured at the same 5-min steps at 1.5% inclination. These steps were performed until the subjects no longer managed to run for 5 min (exhaustion). In day 2, the same velocities were measured, but in the opposite order. C R was calculated as gross oxygen cost per kilo bodyweight raised to the power of 0.75, per metre of running for the set running velocities. For the relative intensities, for example, 75% VO2max, running velocity was calculated in the same way as for MAS (Fig. 1).

Fig. 1
figure 1

Assessment of maximal aerobic speed. The linear regression equation in Fig. 1 is: y = 3.9393x + 2.7963, R = 0.9997, VO2max for this runner is 69.2 ml kg−1 min−1. 70% VO2max for this runner is 48.4 ml kg−1 min−1. From the equation, the corresponding velocity is 11.7 km h−1, which equals 195 m min−1. C R expressed as ml kg−1 m−1 is thus 0.284 ml kg−1 m−1. MAS is the velocity point where the horizontal line representing VO2max meets the extrapolated linear regression representing the sub-maximal VO2 measurements. Here represented by a velocity of 16.9 km h−1

Full size image

Assessment of MAS (Fig. 1)

Maximal aerobic speed was defined at the velocity at 1.5% inclination of the treadmill, where the horizontal line representing VO2max meets the extrapolated incremental line representing the sub-maximal VO2 measured in the LT assessment. By plotting VO2 data against running velocity, individual regression equations for each subject could be obtained. Based on the findings of Helgerud (1994), values in between 60 and 90% VO2max were used to calculate the incremental line. An example of the assessment of MAS is presented in Fig. 1.

Allometric scaling has been reported to decrease the SDs in C R between subjects (Helgerud 1994; Helgerud et al. 2001; Hoff et al. 2005). VO2 values are, thus, mainly expressed in ml kg−0.75 in the present study.

Statistical analysis

Descriptive statistical analyses were made to display means, standard deviations, standard error and coefficient of variance. To assess whether or not the VO2 values would be affected by the preceding step (i.e. the day to day measures described above), two-way ANOVA tests were used. Differences in C R at different velocities and possible gender differences in C R were also tested using two-way ANOVA tests. Linear regressions were made to display the equations representing the linearity between running velocity and oxygen cost of running. Correlations were calculated by the Pearson correlation test.

Results

Physical characteristics of subjects (Table 2)

Physical characteristics of the subjects are presented in Table 2. There was no significant gender differences in age, in LT velocity, in LT expressed as percentage of VO2max or in tMAS. The male subjects were taller, heavier and had a higher VO2max than the females.

Table 2 Physical characteristics of the subjects
Full size table

To compare the physiological parameters at VO2max and at MAS, only the participants that completed more than 5 min at MAS were included, independent of gender. They would, thus, have at least spent the same time at this intensity as at the set sub-maximal intensities. Further, 5 min should allow a possible steady rate to occur. These runners clearly reach their VO2max at MAS (99.4% of VO2max). Neither the results regarding HR nor [La−]b peak were different being obtained at VO2max or at MAS. However, the subjects exhibited significantly lower R values when running at MAS when compared with the R values achieved during the VO2max test. These results are presented in Table 3.

Table 3 Physiological variables at VO2max and at maximal aerobic speed (MAS)
Full size table

Physiological variables at VO2max and at MAS (Table 3)

Table 4 shows the mean values in VO2 at each of the set intensities by the ten subjects who completed at least 15.5 km h−1, independent of gender. The results are divided into the two different test days. In day 1, the subjects ran at the lowest velocity first, increasing the speed for each separate run. In day 2, they started at the highest velocity, decreasing the speed for each separate run.

Table 4 Oxygen costs of running at different test days
Full size table

Oxygen cost of running at different test days (Table 4)

There were no significant differences in oxygen costs of running (C R) expressed as VO2 ml kg−0.75 min−1 between the two test days. The coefficient of variation for the mean differences between days 1 and 2 is 0.9%. Consequently, only results from test-day 1, apart from the MAS results (day 2) are used in the following statistical analysis.

The mean data for C R expressed as VO2 ml kg−0.75 m−1 at those of the set velocities that were completed by all the 15 subjects are shown in Table 5. These results are divided by gender.

Table 5 Oxygen costs of running at set velocities
Full size table

Oxygen cost of running at set velocities (Table 5)

There were no significant differences in C R expressed as VO2 ml kg−0.75 m−1 between the different set velocities. The females had significantly lower C R than the males at the set velocities. Mean interindividual variability expressed as standard deviations (SD) for the males was ±7.1%. For the females, mean interindividual variability was ±5.2%.

By plotting VO2 data against running velocity as described in the protocol for assessing MAS, individual regression equations for each subject could be used to find the cost of running at the different relative velocities. Using running intensities representing 75–90% of VO2max, the data are within the area of velocity completed by all the subjects. Table 6 shows the mean data for the C R expressed as VO2 ml kg−0.75 m−1 at these velocities relative to VO2max.

Table 6 Oxygen costs of running at relative intensities
Full size table

Oxygen cost of running at relative intensities (Table 6; Fig. 2)

No significant differences in oxygen costs of running expressed as VO2 ml kg−0.75 m−1 between the different relative velocities was found for the male or for the female subjects. The females had significantly lower oxygen cost of running than the males at all the relative velocities.

Fig. 2
figure 2

Oxygen cost of running at relative intensities. Values are mean ± SD

Full size image

Discussion

The major findings in this study are that oxygen cost of running (C R) did not differ for velocities between 75 and 90% VO2max, and that the commonly used incremental protocol for measuring VO2 at different intensity levels is reliable up to 90% VO2max.

Cost of running at different velocities and intensities

The finding that C R did not differ at the different set velocities or relative intensities is in agreement with reports from Di Pampero et al. (1986) and Helgerud (1994). These results are, however, not in agreement with the results from Daniels and Daniels (1992) who found increasing oxygen cost of running as relative running intensity increased in the area between 70 and 100% VO2max. The results from the present study also revealed that the female runners had significantly lower CR than the male runners at all different intensities. This is in accordance with the results from Helgerud (1994), but not with results from Daniels and Daniels (1992) or Bunc and Heller (1989). There were, however, only six females participating in the present study.

Reliability of the incremental protocol for assessing oxygen cost of running

When comparing the C R from test-day 1 and test-day 2 using the two different protocols (incremental order day 1, descending order day 2), no significant differences were found. The coefficient of variation for the mean differences between days 1 and 2 is only 0.9%. Thus, it seems safe to assume that the commonly used incremental protocol for measuring VO2 at different intensity levels is reliable up to 90% VO2max.

Interindividual variability in oxygen cost of running

Interindividual variability in oxygen cost of running expressed as mean standard deviations for the three velocities completed by all 15 subjects (SD) is ±7.1% for the males and ±5.2% for the females in the present study. The variability is lowest at the lowest velocity (9.5 km h−1), and highest at the highest velocity (12.5 km h−1). Interindividual variability is reported by Di Pampero et al. (1986) to be ±8%, if the oxygen cost of running is expressed per kg of body mass. Morgan et al. (1995) and Lacour et al. (1990) have found within-group variations of about 20%. These results are also expressed per kg body mass. As energy cost for movement does not increase in the same rate as body weight, it seems likely that part of the variation in the oxygen cost of running in the experiments by Di Pampero et al. (1986), Morgan et al. (1995) and Lacour et al. (1990) is due to weight differences. If allometric scaling had been used, these results would assumingly be more in line with those reported in the present study, since the use of allometric scaling when expressing running economy has revealed smaller differences (approximately ±5%) in previous investigations (Helgerud 1994; Helgerud et al. 2001, 2003; Hoff and Helgerud 2003; Hoff et al. 2005).

Maximal aerobic speed

Eight of the fifteen subjects in our study completed more than 5 min at MAS, independent of gender. This group clearly reaches their VO2max at MAS (99.4% of VO2max at MAS). The results regarding HR and lactate confirm this. However, the subjects had significantly lower R values when running at MAS compared with the R values achieved during the VO2max test. A possible explanation for this is the lower intensity at MAS than at the end of the VO2max test.

We found a significant correlation between time at MAS and LT as percentage of VO2max for the male subjects, but not for the females. With LT representing a higher percentage of VO2max, the anaerobic contribution to the work done per metre at MAS should logically be smaller.

Participating subjects

The participating subjects in the present study are characterised as a heterogenous group with regards to running performance, VO2max and C R. However, they are all well-trained runners. Thus, it seems natural that the C R by the group as a whole is higher than reported for elite runners (Daniels and Daniels 1992), intermediate national standard marathon runners (Helgerud 1994) and recreational marathon runners (Helgerud et al. 1990). The mean C R amongst the males (0.755 ml kg−0.75 m−1) corresponds to a CR of 0.246 ml kg−1 m−1, whereas the mean C R amongst the females (0.680 ml kg−0.75 m−1) corresponds to a CR of 0.243 ml kg−1 m−1. The present results are more in agreement with the results from soccer players as reported by Helgerud et al. (2001) and by Hoff and Helgerud (2003).

Conclusion

At velocities representing intensities between 60 and 90% of VO2max, no differences in C R were found. This means that C R measured at sub-maximal velocities are representative for C R at race velocity for distances above 10,000 m for most runners. The commonly used incremental protocol for measuring oxygen uptake (VO2) at different velocities was found not to affect the reliability of these measurements.