Journal Logo

APPLIED SCIENCES: Physical Fitness and Performance

Relationship between %HRmax, %HR Reserve, %V˙O2max, and %V˙O2 Reserve in Elite Cyclists


Author Information
Medicine & Science in Sports & Exercise: February 2007 - Volume 39 - Issue 2 - p 350-357
doi: 10.1249/01.mss.0000246996.63976.5f
  • Free


Physical exercise performed at an appropriate level of intensity has beneficial effects on health in the general population and improves maximal oxygen consumption (V˙O2max) and performance of athletes (3).

The relationship between V˙O2 and heart rate (HR) during exercise is considered linear (4,12) and, therefore, justifies the popular use of HR for estimation, prescription, and monitoring of exercise intensity in athletes and sedentary subjects (3). However, the precision of the HR-V˙O2 relationship can vary from subject to subject as a function of the metabolic stress or physical training level, which means that it must be individually determined (26), especially in highly trained endurance athletes such as amateur and professional cyclists in which these factors are obviously amplified.

For estimation of exercise intensity, Karvonen et al. (15) were the first to recommend the use of a target HR corresponding to a defined percentage of the difference between HRmax and resting HR (heart rate reserve or HRR). Since then, %HRR has been considered equivalent to %V˙O2max without any validation of this substitution and despite reservations expressed by several authors (5,22,30). Recently, Swain and Leutholtz (28) demonstrated in healthy adults that %HRR was not equivalent to %V˙O2max. This finding was confirmed by other studies conducted in healthy adults (29) or in patients with diabetes (8), cardiovascular disease (6), or obesity (7). Swain and Leutholtz (28) also proposed the concept of V˙O2 reserve or V˙O2R (difference between V˙O2max and resting V˙O2) and showed that %HRR was equivalent to %V˙O2R in healthy adults. The equivalency of these two parameters was also confirmed in obese subjects (7). Although other studies (6,8,29) failed to demonstrate the equivalency of %HRR and %V˙O2R, they clearly established that %V˙O2R was a better predictor of %HRR than was %V˙O2max. Consequently, when revising its position stand in 1998, ACSM proposed new guidelines, in which the use of %V˙O2R instead of %V˙O2max is recommended for exercise-intensity estimation based on %HRR (3).

However, the results of a recent study (11) are in total contradiction with previous conclusions. Using cycle ergometers, this study has investigated the relationship between %V˙O2max, %V˙O2R, and %HRR before and after a 20-wk training period in 630 initially sedentary subjects aged 17-65 yr. According to their preliminary results, the authors concluded that %V˙O2max is the best predictor of %HRR compared with %V˙O2R. Thus, this study challenges the relative intensity scale indicated by ACSM.

For exercise prescription, in 1990, ACSM (2) recommended values of 55, 62, 70, 85, and 90% HRmax, corresponding to values of 40, 50, 60, 80, and 85% V˙O2max, respectively. Similarly, in 1998, ACSM (3) proposed values of 35, 55, 70, and 90% HRmax, corresponding to values of 20, 40, 60, and 85% V˙O2R or HRR. Several authors have expressed some reservations about these recommendations. One of the major criticisms is that the ACSM recommendations were established with incorrect transposition of linear regression equations and were based on a collective method of data analysis, leading to a unique equation applicable to all subjects (27). According to Swain et al. (27,28), a method based on individual regression analysis would be more appropriate, because it would more accurately take into account individual variability in groups of subjects. By using this method in 81 healthy females and 81 healthy males aged 18 to 34 yr, Swain et al. (27) observed %HRmax values significantly higher than the recommended values at each level of %V˙O2max. Similar inconsistencies were reported in adults with chronic obstructive pulmonary disease (COPD), in whom %HRmax values at 50, 60, and 80% (but not 85%) of V˙O2max were significantly higher than the ACSM recommendations (25). In view of the difficulty of continually adjusting correspondence scales between the various markers of intensity, the equivalency between %HRR and %V˙O2R seems to be a more practical alternative, because no conversion factor is required.

Amateur and professional cyclists rely heavily on HR as a marker of exercise intensity for training prescription and for exercise-intensity monitoring during competitive events (18,23). Achievement of optimal performance in highly trained cyclists requires an individually tuned training program in which the workload is precisely defined and monitored (17,20). To the best of our knowledge, the HR-V˙O2 relative intensity scales included in the ACSM recommendations for the general population (2,3) have never been validated in, and may not be suitable for, highly trained endurance athletes, such as elite amateur or professional cyclists. In highly fit athletes, the HR-V˙O2 relationships may have specific characteristics as suggested by data from a Swain and Leutholtz (28) study, in which the slope and intercept of the %HRR versus %V˙O2max regression were smaller in high-fit (V˙O2max > 50 mL·min−1·kg−1: y = 1.05x − 4.1) compared with low-fit subjects (V˙O2max < 30 mL·min−1·kg−1: y = 1.13x − 15.1). Therefore, a difference of approximately 20 mL·min−1·kg−1 in V˙O2max seems to significantly affect the HR-V˙O2 relationship. On the basis of the published data, it is estimated that in Swain and Leutholtz (28), the V˙O2max of the fittest subjects ranged from 50 to 53.32 mL·min−1·kg−1. Considering that the difference in V˙O2max between elite endurance athletes and the highly fit subjects of Swain et al. (28) is also approximately 20 mL·min−1·kg−1, could the results obtained by Swain and Leutholtz (28) be readily extended to elite endurance athletes?

In consideration of the apparent effect of fitness, the lack of validation in elite endurance athletes, and the current controversy, the determination of the relationship between %HRR and %V˙O2max or %V˙O2R in elite cyclists may provide valuable information. Our hypothesis was that in a population of elite endurance athletes, %HRR is equivalent to %V˙O2R and not to %V˙O2max. Our study was, therefore, designed to examine the HR-V˙O2 relationship and to assess the relative exercise-intensity scale currently recommended by ACSM in a group of elite road cyclists during the racing season.



Twenty-six (26) highly trained, healthy male road cyclists, 11 elite amateurs (AMAT) and 15 professionals (PRO), tested in the context of the mandatory medical monitoring required by the FFC (Fédération Française de Cyclisme) and the UCI (Union Cycliste Internationale), were selected for this study. Written informed consent was obtained after each subject was fully informed of the purpose and aim of this observational study. All subjects were familiar with the protocols and materials used for the current study. All amateur cyclists had at least 2 yr of race experience at the elite amateur level (Elite-2 UCI classification). Professional cyclists were actively racing and had participated in at least one major national or international competition of the UCI calendar. The subjects' physical and physiological characteristics are presented in Table 1. The professional cyclists in this study were elite riders according to the classification proposed by Jeukendrup et al. (14).

Characteristics of elite amateur (AMAT) and professional (PRO) road cyclists.


All subjects performed an exercise test between 9:00 a.m. and 12:00 a.m. from May to September, during the racing period of the cycling season. A clinical interview and examination eliminated any contraindications to this test. Subjects were asked to refrain from any hard physical activities for 24 h before testing and from the consumption of any stimulants (coffee, tobacco, alcohol) or drugs that could influence heart rate. They ingested their last meal at least 3 h before the start of the test.

After measuring the subjects' height and weight and adjusting the bike as necessary, each subject remained seated calmly for 20 min. Minimal HR and V˙O2 values recorded during the last 5 min of this period were considered to be resting values. Environmental conditions were maintained constant during each test (temperature: 19.6 ± 0.01°C, relative humidity: 62.0 ± 0.05%, atmospheric pressure: 1026 ± 0.1 bars; mean ± SEM).

All incremental exercise tests were performed on an electromagnetically braked cycle ergometer (Lode Excalibur Sport, Groningen, The Netherlands) equipped with a narrow racing saddle, clipless pedals, and drop handlebars. The subjects used their own pedals and racing shoes, and adjusted the seat height and handlebars position to the individual morphological characteristics and habitual racing position. All tests were performed in the conventional seated posture. The test protocol was a graded procedure of 3-min stages starting at 0 W, with workload increasing by 50 W·3 min−1 from rest until exhaustion. Every subject chose his own cadence and was requested to keep it constant for the duration of the test with the aid of a visual revolution counter. Average cadence for the group was 90 rpm, with individual values ranging from 80 to 110 rpm. Elite cyclists have been shown to prefer a pedaling cadence within this range.

The test was terminated either voluntarily by the subject, or when the cadence could not be maintained at a minimum of 80 rpm. Maximal power output (Wmax) was calculated as Wmax = Wlast + (t × Δ/T in which Wlast is the last completed workload, t is the number of seconds in the final stage, Δ is the power increment in watts, and T is the duration in seconds of a complete stage.

During the tests, subjects breathed room air through a Hans-Rudolph (Kansas City, MO) two-way nonrebreathing valve. Expired gases were continuously collected into the mixing chamber of an Oxycon V open-circuit calorimetric system (Mijnhardt, Groningen, The Netherlands). The dry-gas flowmeter was calibrated before each test with a 3-L syringe. The paramagnetic O2 and infrared CO2 analyzers were calibrated before each test with known gases analyzed by a Perkin-Elmer MGA-100 mass spectrometer. The mixing chamber was automatically sampled every second. Oxygen consumption (V˙O2, L·min−1 STPD), carbon dioxide production (V˙CO2, L·min−1 STPD), ventilatory flow rate (VE, L·min−1 BTPS), and respiratory exchange ratio (RER) data were calculated, averaged, and recorded every 30 s. Heart rate was measured and recorded with a heart rate monitor (Polar, Kempele, Finland) and with continuous 12-lead ECG monitoring on an electrocardiograph (Marquette Case 15, Milwaukee, WI).

Data collection and analysis.

Only data from tests considered to be exhaustive according to criteria reviewed in Howley et al. (13) were analyzed. Selected exercise tests fulfilled at least two of the following three criteria: 1) RER greater than 1.08; 2) HRmax ± 10 bpm of the age-predicted HRmax; and 3) V˙O2max ≈ 6 mL·kg−1·min−1 of predicted V˙O2max according to the Storer equation recommended for endurance-trained athletes (19).

Maximum HR and V˙O2 values during exercise were determined as the highest values observed during the exercise test. HRR and V˙O2R values were calculated by subtracting the value at rest from the respective maximum value of each parameter. HR and V˙O2 values recorded during the last minute of each 3-min complete stage were averaged and expressed as percentages of their respective reserve or maximum values, that is, %HRmax, %HRR, %V˙O2max, or %V˙O2R. For each subject, data obtained at rest, at the last minute of each completed stage, and at maximum workload were used to perform four types of linear regression (%HRmax vs %V˙O2max, %HRmax vs %V˙O2R, %HRR vs %V˙O2max, and %HRR vs %V˙O2R) and to calculate the respective slopes, intercepts, and squared correlation coefficients. The mean intercepts and slopes for each type of regression were then used to calculate the regression equations related to each group of subjects.

To evaluate the HR-V˙O2 relative intensity scale recommended by ACSM, %HRmax versus %V˙O2max regression slopes and intercepts for each subject were used to predict the %HRmax values corresponding to 40, 50, 60, 80, and 85% V˙O2max, respectively. Predicted values were then compared with the values recommended by ACSM (2), that is, 55, 62, 70, 85, and 90% HRmax, respectively. According to the same principle, %HRmax values predicted according to the %HRmax versus %V˙O2R regression and corresponding to 20, 40, 60, and 85% V˙O2R, respectively, were compared with the values recommended by ACSM (3), that is, 35, 55, 70, and 90% HRmax, respectively.

For evaluating the level of equivalency between %HRR, %V˙O2max, and %V˙O2R, the slopes and intercepts of the %HRR versus %V˙O2max regression and the %HRR versus %V˙O2R regression were calculated for each subject and were used to predict values of %V˙O2max and %V˙O2R corresponding to selected values of 35, 45, 55, 65, 75, 85, and 95% HRR, respectively.


Results are reported as mean ± standard error of the mean (SEM), similar to previous studies (6-8,25,28,29), because SEM reflects the theoretical dispersion of the sample means and, therefore, characterizes uncertainty about the true value of the population mean, providing an estimate of the accuracy of the mean. The data were analyzed using computer software (Statview 5, SAS Institute Inc.). Mean values of subject characteristics were compared between AMAT and PRO with a nonparametric Mann-Whitney test. Student t-tests were used to determine whether the mean slopes and intercepts of each type of regression in each group of subjects were significantly different from the line of identity (i.e., slope = 1; intercept = 0). Nonparametric tests based on 95% confidence intervals were used to determine whether mean predicted values of %HRmax, %V˙O2max, and %V˙O2R were significantly different from the respective recommended or indicated values. Spearman correlation coefficients were calculated to investigate possible relationships between subject characteristics and intercepts of the %HRR versus %V˙O2max regression. For all statistical analyses, the null hypothesis was rejected at a probability of P < 0.05.


Subject characteristics.

Subjects' anthropometrical characteristics are presented in Table 1. VE, V˙O2, and power-output maximal values are representative of highly trained cyclists (14,17). No significant difference was observed between elite amateur (AMAT) and professional (PRO) cyclists for any physical or physiological parameters. The seven subjects with the highest V˙O2max values (77.53 ± 1.51 kg·mL−1·min−1) and the seven subjects with the lowest V˙O2max values (63.03 ± 2.54 kg·mL−1·min−1) do not differ significantly concerning slopes (1.071 ± 0.045 and 1.066 ± 0.029, respectively) and intercepts (−6.836 ± 3.995 and −5.720 ± 2.129 respectively) of the %HRR versus %V˙O2max regression, as well as slopes (1.005 ± 0.042 and 0.999 ± 0.021 respectively) and intercepts (−0.251 ± 3.888 and 0.894 ± 1.508 respectively) of the %HRR versus %V˙O2R regression. Therefore, the population of this study was considered homogeneous relative to the HR-V˙O2 relationship.

Comparison with the ACSM recommendations.

The results of prediction analysis are presented in Table 2. The mean values were calculated by using the results of the 26 individual linear regressions studied. Values of %HRmax at 40, 50, 60, and 80% (but not 85%) of V˙O2max were significantly higher (P < 0.001) than the ACSM guidelines. Similarly, values of %HRmax at 20, 40, and 60% (but not 85%) of V˙O2R were significantly higher (P < 0.001) than the ACSM recommendations. The two mean regression equations concerning the relationship of %HRmax to %V˙O2max and to %V˙O2R are (means ± SEM):

Predicted values of %HRmax corresponding with indicated values of %V˙O2max or %V˙O2R; comparison with the ACSM guidelines.

These two equations are not redundant, because the terms of V˙O2 vary, and, above all, the respective slopes are significantly different as well as the intercepts (P < 0.0001).

Linear regression analysis.

This analysis was performed to study the characteristics of the relationship between %HRR and %V˙O2max and between %HRR and %V˙O2R (Table 3). In the AMAT group, PRO group, or TOTAL sample, the %HRR versus %V˙O2max regressions did not coincide with the line of identity, that is, an intercept equal to 0 and a slope equal to 1. These regressions had a mean intercept of −5.274 ± 0.97 %HRR units for AMAT, −6.093 ± 1.22 %HRR units for PRO, and −5.747 ± 0.80 %HRR units for the TOTAL sample. The mean slope was 1.070 ± 0.01 for AMAT, 1.069 ± 0.01 for PRO, and 1.069 ± 0.01 for the TOTAL sample. In each group, the mean intercept was significantly different from 0 (P < 0.0001), and the mean slope was significantly different from 1 (P < 0.0001). The corresponding mean coefficients of determination (R2) varied according to the groups between 0.992 and 0.994 (Table 2 and Fig. 1).

Intercepts, slopes, and correlation coefficients of linear regression analyses for %HRR vs %V˙O2max and %HRR vs %V˙O2R regressions.
Regression line of percent heart rate reserve versus %V˙O2max for the TOTAL sample of elite cyclists (N = 26). The regression equation is the average of 26 individual regressions. The corresponding mean slope and intercept were significantly different from 1 and 0, respectively (P < 0.0001).

The %HRR versus %V˙O2R regressions were not distinguishable (P > 0.2) from the line of identity in AMAT, PRO, and the TOTAL sample. It had a mean intercept of 1.298 ± 0.85 %HRR units for AMAT, 0.359 ± 1.03 % HRRunits for PRO, and 0.756 ± 0.69 %HRR units for the TOTAL sample. In each group, the mean slope was 1.003 ± 0.01. The mean intercepts were not significantly different from 0, and the mean slopes were not significantly different from 1. The corresponding mean coefficients of determination (R2) varied from 0.992 to 0.994 according to the groups (Table 2 and Fig. 2).

Regression line of percent heart rate reserve vs %V˙O2 Reserve for the TOTAL sample of elite cyclists (N = 26). The regression equation is the average of 26 individual regressions. The corresponding mean slope and intercept were not significantly different from 1 and 0, respectively (P > 0.2).

Because there was no significant difference between AMAT (N = 11) and PRO (N = 15) in terms of physical and physiological characteristics or regression parameters, the two subgroups were merged (N = 26) for correlation and prediction analyses.


The intercept of the %HRR versus %V˙O2max regression was significantly correlated with resting V˙O2 (r = −0.46; P < 0.05) but not with V˙O2max (r = −0.08).

Prediction of %V˙O2max and %V˙O2R from %HRR.

Table 4 shows that the predicted values of %V˙O2R were not significantly different from the indicated values of %HRR for intensities ranging from 35 to 95% HRR. On the other hand, predicted values of %V˙O2max were significantly higher than indicated values of %HRR or %V˙O2R, from 35 to 75% HRR. The disparity between %V˙O2max and %HRR gradually decreased with increasing intensity of exercise, with the error decreasing from 9% at 35% HRR to 2% at 65% HRR. This disparity was not significant at 75, 85, and 95% HRR (Fig. 3).

Values of %V˙O2max and %V˙O2R corresponding with indicated values of %HRR.
Regression lines for predicted values of %V˙O2max and %V˙O2R at a given value of %HRR. Significant differences between %V˙O2max and the corresponding %HRR value: * P < 0.05; ** P < 0.001.

Verification with a theoretical V˙O2 at rest.

The mean value of V˙O2 at rest measured in our subjects (4.3 ± 0.9 kg·mL−1·min−1) was significantly different fromthe theoretical value of 1 MET, that is, V˙O2 of 3.5 kg·mL−1·min−1 (P < 0.001). For the purposes of verification, the measured value of V˙O2 at rest was replaced by the theoretical value in each subject. Replacement of this value did not significantly affect the final result of regression analysis: the modified %HRR versus %V˙O2R regression still coincides with the line of identity. The modified mean intercept (0.026 ± 0.734) was not significantly different from 0, and the modified mean slope (1.010 ± 0.007) was not significantly different from 1. Predicted values of %V˙O2R at 35, 45, 55, 65, 75, 85, and 95% HRR on the basis of the modified regression were (means ± SEM): 34.6 (± 0.6), 44.5 (± 0.5), 54.4 (± 0.5), 64.3 (± 0.5), 74.2 (± 0.5), 84.1 (± 0.5), and 94.1 (± 0.5), respectively. These %V˙O2R values were not significantly different from %HRR values or from the %V˙O2R values predicted with the nonmodified regression analysis.


To the best of our knowledge, this is the first study to specifically test the HR-V˙O2 relationship in elite cyclists. In consideration of the lack of validation in elite endurance athletes, the apparent effect of fitness (28), and the recent publication of divergent results (11) concerning the HR-V˙O2 relationship, the current study was designed to examine the relationships between %HRR, %V˙O2max, and %V˙O2R in 26 elite amateur and professional male cyclists. No significant difference in the HR-V˙O2 relationship was observed between highly trained elite amateur and professional cyclists. The main findings of the present study are that in elite cyclists, the ACSM recommendations, in terms of correspondence between %HRmax and %V˙O2max or %V˙O2R, underestimate exercise intensity; that %HRR is equivalent to %V˙O2R but not to %V˙O2max; and that %HRR, %V˙O2max, and %V˙O2R become equivalent above 75% exercise intensity.

Below 85% V˙O2max, predicted values of %HRmax in our subjects were significantly higher than the 1990 ACSM-recommended values (Table 2). These results are in agreement with several other studies (7,22,25,27). Our results are especially similar to those reported by Byrne and Hills (7), who observed mean values of 59, 66, 72, and 86% HRmax corresponding to 40, 50, 60, and 80% V˙O2max. This concordance of results observed between elite cyclists and sedentary subjects should be noted. However, at 50 and 60% V˙O2max, %HRmax values were 3-6% lower in our highly trained cyclists than those observed in healthy 18- to 34-yr-old sedentary subjects (27) or in 55- to 80-yr-old COPD patients (25). This finding is in contrast to the conclusion that "high-fit men averaged 2% higher in %HRmax than low-fit men at any given value of %V˙O2max" (27). Several authors have reported that the method based on HRmax underestimates V˙O2R by about 15% (16,22,28,29). Consequently, new recommendations have been made by ACSM (3). Our study shows that the recent recommendations based on HRmax underestimate V˙O2R in elite cyclists. However, the mean value of 90% for predicted %HRmax observed at 85% V˙O2max or V˙O2R is in agreement with the ACSM recommendations. For values above 80% V˙O2max or V˙O2R, conformity with the recommendations remains controversial, because some authors (7,22), in contrast with others (9,27), have reported higher values at 85% V˙O2max or V˙O2R (i.e., 92-93% HRmax).

The 3% disparity in predicted HRmax values in relation to %V˙O2max between the present study and the 1990 ACSM recommendations (Table 2) would lead to errors of estimation of 6, 5, 4, and 3.5% at 40, 50, 60, and 80% of V˙O2max, respectively. The degree of disparity between predicted and recommended HRmax values is greater when related to %V˙O2R: 13, 6, and 4%, leading to errors of estimation of 37, 11, and 6% at 20, 40, and 60% of V˙O2R, respectively. Could such errors of estimation be problematic for elite cyclists? With reference to an exercise-intensity scale (21), these errors would have little impact on high-intensity exercise (> 90% HRmax or > 85% V˙O2max), whereas they would mainly affect the intensities of recovery (< 70% HRmax or < 60% V˙O2max), moderate aerobic intensities (70-80% HRmax or 60-75% V˙O2max), and, to a lesser degree, sustained aerobic intensities (80-90% HRmax or 75-85% V˙O2max). It should be noted that the lower exercise intensities are the most usual during training and competition. It has been shown that riders spend as much as 60% of the duration of a stage race at moderate (50-70% V˙O2max) and at low or recovery exercise intensities (< 50% V˙O2max) during a major cycling tour (10). For exercise prescription and monitoring in elite cyclists, we suggest two alternatives: either apply the equations of %HRmax in relation to %V˙O2max or V˙O2R proposed in the current study (see equations 1 and 2) or use %HRR in relation to %V˙O2R.

Our study clearly demonstrated an equivalency between %HRR and %V˙O2R in elite cyclists. These subjects presented a high fitness level (peak work rate = 20 to 21 METs), which is in agreement with the literature (17). The equivalency between %HRR and %V˙O2R was established by individual regression analysis, with the result that the %HRR versus %V˙O2R regression was not distinguishable with the line of identity in terms of slope and intercept. This equivalency was also established by analyses of prediction, which showed that predicted %V˙O2R values were not significantly different from values of %HRR indicated on an intensity scale ranging from 35% to 95% of HRR. These results are in agreement with other studies (7,9,28). Davis and Convertino (9) suggested an equivalency between %HRR and percentage of net V˙O2max at four 5-min steady-state workloads on the treadmill, ranging from 25 to 85% of V˙O2R, net V˙O2max being the same as V˙O2R. Testing 63 healthy adults of both sexes aged 18-40 yr with low to moderate fitness levels (7.4-15.9 METs) on a bicycle ergometer, Swain and Leutholtz (28) were the first to report an equivalency relationship between %HRR and %V˙O2R on the basis of individual regression and prediction analysis. In another treadmill study, Byrne and Hills (7) also showed that the %HRR versus %V˙O2R regression coincided with the line of identity in 32 male and female obese adults with low to moderate fitness levels (8.9-13.8 METs).

In healthy, moderately fit (10.9-14.8 METs) sedentary subjects (29), and particularly in very-low-fit (4.3-6 METs) patients with heart disease (6) or diabetes (8), the equivalency between %HRR and %V˙O2R has not been formally established. However, these studies reported that %V˙O2R was a better predictor of %HRR than %V˙O2max, by showing that the %HRR versus %V˙O2R regression was both closer to the line of identity and significantly different from the %HRR versus %V˙O2max regression in terms of slopes and intercepts. The wide range of the subjects' fitness level and its obvious impact on the completion of an exhaustive test and, consequently, on the quality of regressions, could therefore explain the variations in results. Paradoxically, a recent study performed on cycle ergometers (11), using individual regression and prediction analyses in 630 sedentary subjects aged 17-65 yr, before and after a 20-wk training program, reported a contradictory result indicating that %V˙O2max, rather than %V˙O2R, was a better predictor of %HRR, challenging the established guidelines.

One special, unexpected but coherent, finding in our study concerns a particular relationship observed between %HRR and %V˙O2max. At intensities equal to or below 65% HRR, %V˙O2max was significantly higher than %HRR, whereas at intensities equal to or above 75% HRR, an equivalency was observed between the two parameters (Fig. 3). Two factors related to the disparity between %HRR and %V˙O2max could explain this biphasic behavior. The first factor described in our study and confirmed in the literature (5,28) is the pattern of the disparity between %HRR and %V˙O2max. Initially marked, the disparity gradually decreases with increasing intensity of exercise. This is well illustrated by the various regression lines that gradually converge towards the line of identity and merge with this line at 100% exercise intensity (Fig. 3). The equivalency between %V˙O2max and %V˙O2R above 75% HRR exercise intensity has never been reported and could be linked to the high fitness level of our population. The second factor would be the initial level of disparity, which, in our subjects, was very low-only +3% at 35% HRR. At this intensity, this is the lowest value of disparity reported, with values in the literature ranging from 7 to 18% (5,28). This very low value of initial disparity could be explained by the very high fitness level of our subjects, because in healthy adults, Swain and Leutholtz (28) showed that the level of initial disparity between %V˙O2max and %HRR was negatively correlated with V˙O2max. This relationship was also observed in obese adults when V˙O2max was expressed relative to fat-free mass and not to total body weight (7). In the present study, this relationship was not observed. In contrast, the initial disparity was inversely correlated with V˙O2 at rest, which itself was positively correlated with V˙O2max.

Our subjects presented a higher mean V˙O2 resting value (4.3 ± 0.9 mL·kg−1·min−1) than the theoretical equivalent of 1 MET (equal to 3.5 mL·kg−1·min−1 of V˙O2). The elevated value of V˙O2 at rest may be explained by the fact that it was measured when subjects were sitting on the bike in a position that might require additional muscles' solicitation for postural control. However, this mean value was similar to that reported under similar conditions by Swain and Leutholtz (28) in adult males, and it was not significantly different from the value of 4.2 mL·kg−1·min−1, equivalent to 1.2 METs, reported by Ainsworth et al. (1) for subjects resting in a quiet environment. It is not surprising to find a relatively high V˙O2 at rest in cyclists characterized by a high level of training, a high neuroendocrine activation, a low fat mass, and a high fat-free mass-factors thought to promote elevation of basal metabolism (RMR) and, therefore, V˙O2 at rest (24). Moreover, the use of an identical at-rest V˙O2 value of 3.5 mL·kg−1·min−1 for all subjects did not modify the conclusions of our study related to the HR-V˙O2 relationship. This is an important result, especially for clinical practice and exercise in the field or in the laboratory, when the required conditions of absolute rest before exercise are difficult to obtain. Elite cyclists could use either a carefully measured at-rest V˙O2 value or a theoretical value equivalent to 1 MET or 1.2 METs for exercise prescription.


The present study is the first to examine and validate the relationships between HR and V˙O2 in elite cyclists, thus complementing the body of work in this area. A constant equivalency relationship between %HRR and %V˙O2R can be observed in these highly trained endurance athletes, in line with the ACSM recommendations for the general population. It also appears that in these subjects, %V˙O2max is equivalent to %HRR for intensities equal to or above 75% of HRR, which was unexpected, considering the results obtained in the general population. However, because %V˙O2max is inaccurate for lower levels of intensity, estimation based on this parameter are of very little value because these levels of intensity are extensively used during training and competition in cycling. Additionally, compared with the general population, the relative exercise-intensity scale indicated by ACSM underestimates exercise intensity in elite cyclists, particularly when based on the %HRmax-%V˙O2R relationship. When exercise intensity is based on the %HRmax-%V˙O2max relationship, the underestimation is much reduced and is probably of lower practical impact.

Consequently, the better predictor of %HRR would be %V˙O2R, not %V˙O2max, in elite cyclists. These athletes should, therefore, use %HRR in relation to %V˙O2R for estimation of exercise intensity. Our results support the conclusions of Swain and Leutholtz (28) and the ACSM (3) recommendations on this particular point.

We sincerely thank the participants and their coaches for their interest and cooperation. We appreciate the assistance of Pascale Biorel, state registered nurse, for technical support and data collection.


1. Ainsworth, B. E., W. L. Haskell, M. C. Whitt, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med. Sci. Sports Exerc. 32:498-504, 2000.
2. American College of Sports Medicine. Position Stand: The recommended quantity and quality of exercise for developing and maintaining cardiorespiratory and muscular fitness, and flexibility in adults. Med. Sci. Sports Exerc. 22:265-274, 1990.
3. American College of Sports Medicine. Position Stand: The recommended quantity and quality of exercise for developing and maintaining cardiorespiratory and muscular fitness, and flexibility in adults. Med. Sci. Sports Exerc. 30:975-991, 1998.
4. Arts, F. J., and H. Kuipers. The relation between power output, oxygen uptake and heart rate in male athletes. Int. J. Sports Med. 15:228-231, 1994.
5. Belman, M. J., and G. A. Gaesser. Exercise training below and above the lactate threshold in the elderly. Med. Sci. Sports Exerc. 23:562-568, 1991.
6. Brawner, C. A., S. J. Keteyian, and J. K. Ehrman. The relationship of heart rate reserve to V˙O2 reserve in patients with heart disease. Med. Sci. Sports Exerc. 34:418-422, 2002.
7. Byrne, N. M., and A. P. Hills. Relationships between HR and V˙O2 in the obese. Med. Sci. Sports Exerc. 34:1419-1427, 2002.
8. Colberg, S. R., D. P. Swain, and A. I. Vinik. Use of heart rate reserve and rating of perceived exertion to prescribe exercise intensity in diabetic autonomic neuropathy. Diabetes Care 26:986-990, 2003.
9. Davis, J. A., and V. A. Convertino. A comparison of heart rate methods for predicting endurance training intensity. Med. Sci. Sports 7:295-298, 1975.
10. Fernandez-Garcia, B., J. Perez-Landaluce, M. Rodriguez-Alonso, and N. Terrados. Intensity of exercise during road race pro-cycling competition. Med. Sci. Sports Exerc. 32:1002-1006, 2000.
11. Gaskill, S. E., C. Bouchard, T. Rankinen, et al. Percent heart rate reserve is better related to %V˙O2max than to %V˙O2 reserve: the Heritage Family Study. Med. Sci. Sports Exerc. 36(Suppl.):S3, 2004.
12. Hawley, J. A., and T. D. Noakes. Peak power output predicts maximal oxygen uptake and performance time in trained cyclists. Eur. J. Appl. Physiol. Occup. Physiol. 65:79-83, 1992.
13. Howley, E. T., D. R. Bassett Jr., and H. G. Welch. Criteria for maximal oxygen uptake: review and commentary. Med. Sci. Sports Exerc. 27:1292-1301, 1995.
14. Jeukendrup, A. E., N. P. Craig, and J. A. Hawley. The bioenergetics of world class cycling. J. Sci. Med. Sport. 3:414-433, 2000.
15. Karvonen, M. J., E. Kentala, and O. Mustala. The effects of training on heart rate: a longitudinal study. Ann. Med. Exp. Biol. Fenn. 35:307-315, 1957.
16. Londeree, B. R., and S. A. Ames. Trend analysis of the % V˙O2max-HR regression. Med. Sci. Sports 8:123-125, 1976.
17. Lucia, A., J. Pardo, A. Durantez, J. Hoyos, and J. L. Chicharro. Physiological differences between professional and elite road cyclists. Int. J. Sports Med. 19:342-348, 1998.
18. Lucia, A., J. Hoyos, A. Carvajal, and J. L. Chicharro. Heart rate response to professional road cycling: the Tour de France. Int. J. Sports Med. 20:167-172, 1999.
19. Malek, M. H., D. E. Berger, T. J. Housh, J. W. Coburn, and T. W. Beck. Validity of V˙O2max equations for aerobically trained males and females. Med. Sci. Sports Exerc. 36:1427-1432, 2004.
20. Mujika, I., and S. Padilla. Physiological and performance characteristics of male professional road cyclists. Sports Med. 31:479-487, 2001.
21. Neumayr, G., R. Pfister, G. Mitterbauer, A. Maurer, and H. Hoertnagl. Effect of ultramarathon cycling on the heart rate in elite cyclists. Br. J. Sports Med. 38:55-59, 2004.
22. Panton, L. B., J. E. Graves, M. L. Pollock, et al. Relative heart rate, heart rate reserve, and V˙O2 during submaximal exercise in the elderly. J. Gerontol. A. Biol. Sci. Med. Sci. 51:165-171, 1996.
23. Palmer, G. S., J. A. Hawley, S. C. Dennis, and T. D. Noakes. Heart rate responses during a 4-d cycle stage race. Med. Sci. Sports Exerc. 26:1278-1283, 1994.
24. Poehlman, E. T. A review: exercise and its influence on resting energy metabolism in man. Med. Sci. Sports Exerc. 21:515-525, 1989.
25. Simmons, D. N., M. J. Berry, S. I. Hayes, and S. A. Walschlager. The relationship between %HRpeak and %V˙O2peak in patients with chronic obstructive pulmonary disease. Med. Sci. Sports Exerc. 32:881-886, 2000.
26. Skinner, J. S., S. E. Gaskill, T. Rankinen, et al. Heart rate versus %V˙O2max: age, sex, race, initial fitness, and training response- HERITAGE. Med. Sci. Sports Exerc. 35:1908-1913, 2003.
27. Swain, D. P., K. S. Abernathy, C. S. Smith, S. J. Lee, and S. A. Bunn. Target heart rates for the development of cardiorespiratory fitness. Med. Sci. Sports Exerc. 26:112-116, 1994.
28. Swain, D. P., and B. C. Leutholtz. Heart rate reserve is equivalent to %V˙O2 reserve, not to %V˙O2max. Med. Sci. Sports Exerc. 29:410-414, 1997.
29. Swain, D. P., B. C. Leutholtz, M. E. King, L. A. Haas, and J. D. Branch. Relationship between % heart rate reserve and % V˙O2 reserve in treadmill exercise. Med. Sci. Sports Exerc. 30:318-321, 1998.
30. Weltman, A., J. Weltman, R. Rutt, et al. Percentages of maximal heart rate, heart rate reserve, and V˙O2peak for determining endurance training intensity in sedentary women. Int. J. Sports Med. 10:212-216, 1989.


©2007The American College of Sports Medicine