RPE Association with Lactate and Heart Rate during High-Intensity Interval Cycling : Medicine & Science in Sports & Exercise

Journal Logo

Applied Sciences: Psychobiology and Behavioral Strategies

RPE Association with Lactate and Heart Rate during High-Intensity Interval Cycling


Author Information
Medicine & Science in Sports & Exercise 38(1):p 167-172, January 2006. | DOI: 10.1249/01.mss.0000180359.98241.a2
  • Free


Alternating brief, higher and lower intensity bouts in a given session is known as interval training. Interval training is purported to elicit improvements in endurance performance capacity beyond the point at which improvements result from submaximal steady intensity exercise (16). A potential benefit from high-intensity interval or sprint-type training is improved buffering capacity (18) and a positively altered respiratory compensation threshold (10). RPE, a viable adjunct to oxygen consumption (V̇O2), HR, and other objective measures of intensity (1) remain very stable (13–15) at the lactate threshold regardless of training status (5,13,24). Further, RPE is sensitive to training-induced threshold changes (14). Although lactate may be a more meaningful marker of intensity (26), hygiene issues and the impracticality of taking multiple blood samples make indirectly regulating lactate via RPE an attractive option. In this regard, previous research supports the effectiveness of this paradigm (4,25).

Previous investigations demonstrate an association of RPE with blood lactate concentration [La] during graded exercise testing (3,5,13), across exercise modes (13), and steady-state exercise (4,5) suggesting [La] serves as a perceptual mediator. Others studies have produced dissonant results. Green et al. (9) evaluated [La], RPE, and HR during 60 min of constant-workload cycling showing a clear RPE–[La] divergence beginning at 30 min. Weltman et al. (27) also demonstrated an uncoupling of RPE and blood [La] with discordance observed across three repeated 30-min bouts within a single day. These studies indicate the RPE–[La] association may be disrupted under certain circumstances such as longer duration, steady workload (9), and repeated independent exercise bouts (27). It is also plausible that other factors potentially altering fatigue (hydration status, blood glucose) may alter the RPE–[La] association.

The RPE–[La] response during interval exercise is not well understood. Seiler and Sjursen (23) compared the effects of varying work duration (from 1 to 6 min) during running bouts on RPE and physiological responses. With the knowledge of the prescribed duration, interval running velocity (self-paced) declined as interval duration increased. RPE increased 2–4 units (6- to 20-point scale) from initiation to cessation of an interval session with a slight increase in [La] from midpoint to endpoint of each session. Only intervals of 1-min duration resulted in a significant increase in [La]. Authors concluded RPE increases were not associated with increased lactate. Examining the potential association between RPE and [La] and determining the strength of [La] as a mediator of exertional perceptions would be enhanced by evaluating the correspondence of these variables at multiple time points during exercise and recovery. Considering the popularity and potential performance benefits of interval exercise (16) and the convenience of using RPE to quantify intensity, research investigating the correspondence of [La] and RPE during interval exercise is warranted. This study examined RPE, [La], and HR responses during repeated, high-intensity interval cycling.



Twelve physically active male volunteers served as participants. Based on an alpha level of 0.05, an effect size of 1.5 (RPE units), an SD of 2 units for RPE, and a power of 0.80, a power analysis indicated 12 subjects would be required for this study, which suggests that the sample size for the current investigation was adequate. Before data collection, subjects completed a written informed consent outlining requirements. All procedures were approved by the local review board for protection of human subjects. Each participant arrived at the lab with previous instructions to have consumed adequate fluids in the preceding 24-h period. Subjects also reported 3 h postprandial with instructions to have abstained from caffeine and alcohol for a minimum of 24 h. Age (yr), height (cm), and mass (kg) were recorded. Height and mass were measured using a Detecto balance-type scale (Detecto, Brooklyn, NY). Body fat percentage was estimated using Lange skinfold calipers (Cambridge, MD) and a three-site method (chest, abdomen, thigh) (20).

V̇O2peak trial.

Following descriptive data collection, participants completed a maximal exertion cycling trial for V̇O2peak determination. Seat height was appropriately set on a Monark cycle Ergometer (Varberg, Sweden) with handlebars preferentially adjusted. Subjects maintained 60 rpm by using a Franz XB 700 metronome (Franz, New Haven, CT). Subjects were fitted with an appropriately sized air-cushioned face mask (Vacu-med, Ventura, CA) that was checked upon mounting for appropriate seal and to ensure there were no leaks. At this point, subjects also donned a Polar HR monitor (Stamford, CT). Participants completed a warm-up (4 min, 60 rpm, 0 W). Following the warm-up, power output was increased 60 W every 2 min. All stages following the warm-up were 2 min. Cadence was constant (60 rpm). Power output was increased in this manner until subjects achieved volitional exhaustion or could not maintain the required cadence. Metabolic data (V̇O2, V̇CO2, respiratory exchange ratio, ventilation) were collected using a Vacu-med Vista mini-cpx (silver) metabolic measurement system (Vacu-med). Turbofit software (Vacu-med), designed for use with the metabolic system, was set to report mean metabolic data over 15-s periods. The system was calibrated before each test with a gas of known composition. A 3-L syringe (Hans Rudolph, Kansas City, MO) was used to calibrate the system for ventilation. HR was recorded using a receiver interfaced with the computer. RPE relative to overall feelings were collected during the last 15 s of each stage using the Borg category (6–20) scale. Instructions for anchoring RPE were that “6” corresponds to seated rest, and “19–20” corresponds to maximal exertion. This verbal anchoring procedure was completed just before initiating exercise. Criteria for achievement of V̇O2peak were a) RPE ≥18, b) RER ≥1.1, c) plateau of V̇O2 with increased workload, and d) >85% of age-predicted maximum HR (17). Each participant met at least two of these four criteria. The highest recorded value across a 15-s period was accepted as V̇O2peak. During the V̇O2peak trial, capillary blood samples were collected at the fingertip during the final 10 s of each stage. During the warm-up, a sample was taken at 2 and 4 min. Blood samples were immediately analyzed for [La] using a YSI 1500 Sport Lactate analyzer (Yellow Springs Instruments, Yellow Springs, OH). Before each test, the YSI was calibrated using 5 mmol·L−1 standard and checked for linearity with 15 mmol·L−1 standard.

Interval cycling.

Within 14 d of the V̇O2peak trial, subjects returned to the lab with instructions to be well rested, well hydrated, and 3 h postprandial, and to have abstained from caffeine and alcohol for 24 h. Participants donned a Polar HR monitor, made appropriate adjustments to the Monark Ergometer and then completed a 10-min warm-up (60 rpm, 0 W). Interval bouts were started at 10 min. Five 2-min bouts of high-intensity cycling (INT) were completed with a 3-min recovery cycling (REC) (60 rpm, 0 W) between each INT. At the conclusion of the final 3-min recovery bout, an additional 10 min of recovery cycling was completed. Figure 1 provides a flowchart regarding the interval cycling protocol including timing of measurements. Resistance for INT was individualized. From the V̇O2peak trial, the 4-mmol·L−1 [La] threshold was identified using a graphic plot with power output on the x-axis and [La] on the y-axis. INT resistance was set at 20 W greater than power associated with the 4-mmol·L−1 [La]. This resulted in a power output for INT of 286 ± 41 W. This procedure was meant to mimic an intensity at which interval training might be completed.

FIGURE 1—Flowchart of interval cycling protocol including timeline of INT and REC periods and time points for HR, RPE, and [La] assessment.:
Shaded bars , interval (2 min); open bars , recovery (3 min); black triangles , HR, RPE, [La] measured.

During testing, capillary blood samples were collected at the end of the 10-min warm-up, at the completion of each INT, at the conclusion of each REC, and at 5- and 10-min REC. Samples were analyzed for [La] using the YSI 1500 Sport lactate analyzer (Yellow Springs Instruments), which was calibrated before each trial using 5 mmol·L−1 standard with linearity checks using 15 mmol·L−1 standard. HR and RPE were also recorded at the same time intervals as [La].


To compare values across time, a repeated-measures ANOVA was used for each variable ([La], HR, RPE) with separate analyses for INT and REC. Simple contrasts were used with INT5 and REC5 as the reference categories when analyzing INT and REC, respectively. Comparisons at each time point were also of primary importance. Therefore, RPE, [La], and HR data were standardized with each data point transformed to a z score [z = (raw score – mean)/SD] using the variable-specific grand mean and SD. Means for standardized values were then compared at each time point using repeated-measures ANOVA with simple contrasts. Results were considered significant at P ≤ 0.05. Additionally, to more thoroughly evaluate INT and REC responses, correlations were computed for [La]–RPE and HR–RPE solely at peaks (conclusion of each INT) and for values obtained at the conclusion of REC and for INT and REC values combined. For correlations, raw scores were used as opposed to standardized values.


Descriptive characteristics are presented in Table 1.

Descriptive characteristics of subjects (N = 12) and RPE responses at 4 mmol·L−1.

INT responses.

Results for INT are presented in Table 2. [La] at INT5 was significantly greater than all other INT points. HR at INT5 was significantly greater than HR at all other INT points. RPE at INT5 was significantly greater (P < 0.05) than RPE at INT1, INT2, and INT3 with the difference between INT4 approaching significance (P = 0.08).

Responses at each INT for [La], HR, and RPE.

REC responses.

Results for REC are presented in Table 3. [La] was significantly greater at REC5 than at warm-up, at each REC, and at 5- and 10-min recovery. HR at REC5 was significantly greater than at warm-up, REC1, REC2, REC3, and at 5- and 10-min recovery. RPE was significantly greater at REC5 than at warm-up, REC1, REC2, and at 10-min recovery. The difference between REC5 and the 5-min recovery approached significance (P = 0.06).

Responses at each REC for [La], HR, and RPE.

Time point comparisons.

Standardized values for HR, [La], and RPE for INT and REC are presented in Figure 2. HR was significantly greater than [La] at each INT and significantly less than [La] at each REC and at 5-min recovery. There were no significant differences between RPE and HR at any point. RPE and [La] were not significantly different at REC1, REC2, REC3, REC4, or at the 5- and 10-min recovery points. RPE was significantly less than [La] at REC5. RPE was significantly greater than [La] at INT1 with a marginal difference at INT4 (P = 0.06) and INT5 (P = 0.07). [La] was significantly less than RPE at 10-min warm-up; however, there were no significant differences for values at 10-min recovery.

FIGURE 2—Means for standardized values (z-scores) for HR, [La], and RPE at 10-min warm-up, INT1–INT5, REC1–REC5, and 5- and 10-min recovery. * HR vs [La],:
P < 0.05. § RPE < [La], P < 0.05. ‡ RPE > [La], P < 0.05. ‡‡ RPE > [La], P = 0.06. ‡‡‡ RPE > [La], P = 0.07; RPE vs HR, NSD.


Significant (P ≤ 0.05) correlations were found for HR–RPE and [La]–RPE responses for INT, for REC and INT and REC combined. These data are presented in Table 4.

Correlations for HR, [La], and RPE.


Evidence suggests the RPE–[La] association is disrupted during longer duration (9) and repeated independent exercise bouts (23,27). Research has not definitively determined responses between RPE and associated physiological variables during interval training or the potential RPE mediators during such training. This study examined HR, blood [La], and RPE during interval cycling.

[La] increased systematically upon initiating interval bouts (Fig. 2), with significantly greater sequential values for [La] within INT and REC. Visual observations also show each REC value greater than the preceding INT. Increasing concentration suggests lactate production exceeded removal. Further, with insufficient duration for [La] to decline, it can also be concluded that, through recovery, lactate continued to empty to blood faster than tissue uptake. The time lag in lactate kinetics presents at least a partial explanation for the observed RPE–[La] divergence. First, the delay between lactate production and appearance in blood creates difficulty in interpreting this measure. Higher [La] at the end of recovery (vs previous interval) discredits the concept that [La] at the immediate conclusion of an interval accurately reflects intramuscular metabolism. An alternative would be to use peak [La] following an interval as an indicator. However, this is also problematic. First, multiple measures would be required at tight time intervals to obtain a true peak. For example, if samples were taken at 1, 3, and 5 min post, it is possible a higher [La] occurred between these time points (i.e., 2 min, 2 min 30 s) (2). Additionally, an intention of interval training is to achieve an overload by repeating higher intensity bouts without permitting excessive recovery. Therefore, delaying subsequent intervals for measurements would compromise session quality. Maintaining a typical sequence of intervals consequently results in residual lactate from previous intervals contaminating serial samples, again interfering with interpretation. Finally, independent of these difficulties, lactate uptake during recovery would be unknown, also magnifying complications. Unlike [La], RPE in the current study demonstrated a quicker response to alternating workloads (Fig. 2). Consequently, it can be concluded that dissimilar response times or acute sensitivity of the different measures to changing workloads accounted for the weak correspondence between RPE and [La]. The current design used 3-min recovery periods. There was a considerable decrease in [La] at 5-min recovery following the final interval bout, and it is conceivable that extending recovery periods would have enhanced RPE–[La] correspondence. Previous work by Seiler and Sjursen (23) does not support this contention when comparing shorter and longer work–rest durations. However, in that study, [La] was only assessed at the midpoint and end of the interval session. In Weltman et al. (27) 1.5- or 3-h recovery occurred between successive 30-min steady-state (70% V̇O2max) bouts. Although this hardly classifies as interval training, there were significant main effects for time, showing a drift upward in RPE and a decline in [La]. The potential association with varying recovery durations warrants further investigation.

Unlike the [La]–RPE dissociation, perceived exertion was tightly coupled with HR throughout the session (Fig. 2). Analogous with the above rationale for the independence of [La] and RPE, HR responded more rapidly to workload alterations during intervals and therefore demonstrated a pattern similar to RPE. HR represents a primary mechanism by which acute cardiovascular demand is met. Although not a dead-end product of metabolism, lactate conceptually characterizes a sequestered source of stored energy while in the blood. Because adenosine triphosphate turnover is not dependent on lactate supply, lactate removal from blood for energy purposes is not critical owing to the relatively sluggish response to workload change for [La] compared with HR. Conversely, HR must be quicker to respond because metabolic demand changes rapidly and is partially dependent on cardiac output (and therefore indirectly reliant on HR). Whereas the correspondence between RPE and HR should be viewed as coincidental and not causal, there was clearly a closer association of RPE with HR compared with [La]. Because HR is more sensitive to acute changes in workload (as encountered during interval training), current results suggest RPE also is more sensitive to acute metabolic demand with the mediating influence of [La] being auxiliary.

Although lactate may not be a primary RPE mediator during interval training, current results do not discredit it entirely in this construct. Closer evaluation of RPE, HR, and [La] reveals significant increases across time when evaluating INT and REC bouts exclusively (Tables 2 and 3, respectively). These RPE results mirror those of Seiler and Sjursen (23) who found an upward drift for RPE during running intervals. The current study also found a drift for RPE during recovery. Conclusions from concurrent incremental changes would be consistent with previous studies identifying [La] as an RPE mediator (3,5,13,24). Whereas this is a reasonable hypothesis, an alternate explanation would be that successive intervals amplified fatigue (possibly due to a variety of factors such as acidosis, neuromuscular factors, substrate depletion), and this increase was reflected in RPE. Literature consistently indicates RPE may be attributed to a variety of factors with no single variable being exclusively dominant (19,21). It should also be emphasized that concurrent increases in [La], HR, and RPE occurred even though workloads for REC and INT were consistent throughout. This could imply that increased [La] could have influenced RPE. Alternatively, cumulative fatigue or other physiological changes resulting from multiple bouts could have played a significant role in elevating RPE. Seiler and Sjursen (23) found a 2- to 4-unit RPE drift (Borg’s 15-point scale) throughout a given interval session. Measures were taken only at a mid- and endpoint during testing. Current results are comparable, showing a 3.4-unit drift from INT1 to INT5 (Table 2). Interestingly, the present investigation also shows a similar RPE drift (4 units) from REC1 to REC5 (Table 3). Whereas concurrent [La] increases offer one explanation for RPE drift, it should also be noted that HR at REC increased significantly across time (Table 3), indicating the effect of [La] was likely not isolated. Weltman et al. (27) showed repeated exercise in a day (1.5–3 h apart) dissociates [La] and RPE. Although not mentioned by authors, it is plausible that participants incurred a level of cumulative fatigue from multiple bouts, consequently intensifying RPE. As with other variables, it is difficult to segregate independent influences of certain variables on RPE. However, within certain limitations, current results suggest [La] may play an auxiliary role in RPE during interval exercise.

Studies showing [La] as an RPE mediator have often employed incremental exercise tests (3,5,13). The tight relationships of RPE with numerous physiological measures are predictable in this paradigm. Physiological and biochemical responses to graded testing are well founded, and concomitant increases in RPE with elevated workload/power should be expected. Unfortunately, graded exercise testing does not mimic typical daily training regimens of recreational or elite athletes who practice various types of training including steady-state and interval sessions. RPE has major advantages including its convenience as a viable measure in nonlaboratory settings where objective physiological variables such as oxygen consumption (V̇O2), blood lactate, and even HR are more difficult or impossible to obtain. Research tends to support use of RPE as an adjunct to intensity regulation using other methods (e.g., HR, V̇O2) in these cases (4,6–8,11–13,22,25); however, inconsistencies exist (9,23,27; current study results). Because interval training represents a unique form of exercise, more work is needed to fully understand the utility of RPE during this type of session as well as the correspondence between perceptual and physiological variables.

Interval exercise is a common component of many training schedules. Correspondence between physiological and perceptual responses during this type of exercise is not well established, but may ultimately assist in prescribing/regulating intensity. The current study shows similar patterns for RPE and HR during interval cycling. Cyclic responses across time suggest that both RPE and HR appear to be more sensitive than [La] to acute changes in workload (and therefore immediate metabolic requirements) experienced during interval exercise. Overall, compared with HR and RPE, correspondence between [La] and RPE was weaker and attributed in part to dissimilarities in response times when workload was altered during testing. However, considering values solely for INT or REC lends some evidence that [La] may be an auxiliary mediator for RPE during interval training. Future research should further investigate the prescriptive value of RPE in regulating intensity during interval exercise.


1. American College of Sports Medicine. Guidelines for Exercise Testing and Prescription, 6th ed, Philadelphia: Lippincott Williams & Wilkins, 2000, p. 78.
2. Bishop, P., and M. Martino. Blood lactate measurement in recovery as an adjunct to training. Sports Med. 16:5–13, 1993.
3. Boutcher, S. H., R. L. Seip, R. K. Hetzler, E. F. Pierce, D. Snead, and A. Weltman. The effects of specificity of training on rating of perceived exertion at the lactate threshold. Eur. J. Appl. Physiol. 59:365–369, 1989.
4. Ceci, R., and P. Hassmen. Self-monitored exercise at three different RPE intensities in treadmill vs. field running. Med. Sci. Sports Exerc. 23:732–738, 1991.
5. Demello, J., K. J. Cureton, J. Robin, and E. Boineau. Ratings of perceived exertion at lactate threshold in trained and untrained men and women. Med. Sci. Sports Exerc. 19:354–362, 1987.
6. Dunbar, C. C., R. J. Robertson, R. Baun, et al. The validity of regulating exercise intensity by ratings of perceived exertion. Med. Sci. Sports Exerc. 24:94–99, 1992.
7. Eston, R. G., B. L. Davies, and J. G. Williams. Use of perceived effort ratings to control exercise intensity in young healthy adults. Eur. J. Appl. Physiol. 56:222–224, 1987.
8. Glass, S. C., R. G. Knowlton, and M. D. Becque. Accuracy of RPE from graded exercise to establish exercise training intensity. Med. Sci. Sports Exerc. 24:1303–1307, 1992.
9. Green, J. M., J. R. McLester, T. R. Crews, P. J. Wickwire, R. C. Pritchett, and A. Redden. RPE-lactate dissociation during extended cycling. Eur. J. Appl. Physiol. 94:145–150, 2005.
10. Green, J. M., T. R. Crews, A. M. Bosak, and W. W. Peveler. A comparison of respiratory compensation thresholds of anaerobic competitors, aerobic competitors and untrained subjects. Eur. J. Appl. Physiol. 90:608–613, 2003.
11. Green, J. M., T. Michael, and A. H. Solomon. The validity of ratings of perceived exertion for cross-modal regulation of swimming intensity. J. Sports Med. Phys. Fitness 39:207–212, 1999.
12. Green, J. M., T. R. Crews, A. M. Bosak, and W. W. Peveler. Physiological responses at 0% and 10% treadmill incline using the RPE estimation-production paradigm. J. Sports Med. Phys. Fitness 42:8–13, 2002.
13. Hetzler, R. K., R. L. Seip, S. H. Boutcher, E. Pierce, D. Snead, and A. Weltman. Effect of exercise modality on ratings of perceived exertion at various lactate concentrations. Med. Sci. Sports Exerc. 23:88–92, 1991.
14. Hill, D. W., K. J. Cureton, S. C. Grisham, and M. A. Collins. Effect of training on rating of perceived exertion at the ventilatory threshold. Eur. J. Appl. Physiol. 56:206–211, 1980.
15. Kang, J., E. C. Chaloupka, M. A. Mastronelo, M. S. Donnelly, W. P. Martz, and R. J. Robertson. Regulating exercise intensity using ratings of perceived exertion during arm and leg ergometry. Eur. J. Appl. Physiol. 78:241–246, 1998.
16. Laursen, P. B., and D. G. Jenkins. The scientific basis for high-intensity interval training. Sports Med. 32:53–73, 2002.
17. Maud, P. C. Foster. Physiological Assessment of Human Fitness. Champaign, IL: Human Kinetics, 1995, p. 14.
18. McKenna, M. J., A. R. Harmer, and S. F. Fraser. Effects of training on potassium, calcium and hydrogen ion regulation in skeletal muscle and blood during exercise. Acta Physiol. Scand. 156:335–346, 1996.
19. Mihevic. P. Sensory cues for perceived exertion: a review. Med. Sci. Sports Exerc. 13:150–163, 1981.
20. Pollock, M. L., D. H. Schmidt, and A. S. Jackson. Measurement of cardiorespiratory fitness and body composition in the clinical setting. Clin. Ther. 6:12–27, 1980.
21. Robertson, R. J., and B. J. Noble. Perception of physical exertion; methods, mediators, and applications. In: Exercise and Sport Science Reviews, J. O. Hollozy (Ed.). Baltimore: Williams & Wilkins, 1997, pp. 407–452.
22. Robertson, R. J., N. M. Moyna, K. L. Sward, N. B. Millich, F. L. Goss, and P. D. Thompson. Gender comparison of RPE at absolute and relative physiological criteria. Med. Sci. Sports Exerc. 32:2120–2129, 2000.
23. Seiler, S., and J. E. Sjursen. Effect of work duration on physiological and rating scale of perceived exertion responses during self-paced interval training. Scand. J. Med. Sci. Sports 14:318–325, 2004.
24. Seip, R. L., D. Snead, E. F. Pierce, P. Stein, and A. Weltman. Perceptual responses and blood lactate concentration: effect of training state. Med. Sci. Sports Exerc. 23:80–87, 1991.
25. Stoudemire, N. M., L. Wideman, K. A. Pass, C. L. McGinnes, G. A. Gaesser, and A. Weltman. The validity of regulating blood lactate concentration during running by ratings of perceived exertion. Med. Sci. Sports Exerc. 28:490–495, 1996.
26. Weltman, A. The Blood Lactate Response to Exercise. Champaign, IL: Human Kinetics, 1995, p. 1.
27. Weltman, A., J. Y. Weltman, J. A. Kanaley, et al. Repeated bouts of exercise alter the blood lactate-RPE relation. Med. Sci. Sports Exerc. 30:1113–1117, 1998.


©2006The American College of Sports Medicine