Journal Logo

SPECIAL COMMUNICATIONS: Contrasting Perspectives

The Respiratory Compensation Point and the Deoxygenation Break Point Are Valid Surrogates for Critical Power and Maximum Lactate Steady State


Author Information
Medicine & Science in Sports & Exercise: November 2018 - Volume 50 - Issue 11 - p 2375-2378
doi: 10.1249/MSS.0000000000001698
  • Free

For any type of endurance exercise, there exists a range of task-specific intensities (e.g., speed or power output) that can be sustained in steady-state conditions by oxidative metabolism without a progressive depletion of high-energy phosphates or continuous lactate accumulation. The task-specific intensity at the ceiling of this range identifies a metabolic rate above which oxidative phosphorylation must be further supported at the substrate level for exercise to continue, leading to non–steady-state conditions, accumulation of muscle metabolites, acidification of the tissue, and a hastening of exhaustion. This “critical metabolic rate” that defines sustainable from unsustainable exercise often is identified by measuring critical power (CP) or the maximal lactate steady-state (MLSS). We suggest also that the respiratory compensation point (RCP) and the near infrared spectroscopy-derived muscle deoxygenation break point (deoxy-BP) of ramp-incremental exercise reflect this important boundary.

Our group showed that the oxygen uptakes (V˙O2) associated with CP, MLSS, RCP, and deoxy-BP were not different in a single cohort of healthy volunteers (1). Although this is the only study to compare simultaneously all of these variables, other comparative studies have corroborated V˙O2 as a common denominator among these indices (2,3), supporting the idea that all represent different physiological expressions of the same underlying phenomenon. Still, others have denied the equivalence of these indexes (4,5) and have insisted that their interrelationships are coincidental on the basis that the power output associated with RCP/deoxy-BP often is reported to be greater than CP/MLSS or that these indices can exhibit low between-measure agreement. However, we argue that these findings stem from: (i) challenges with translation of the V˙O2 versus power output relationship between ramp-incremental and constant-intensity exercise paradigms and (ii) a lack of methodological validation of these indices (particularly CP).

Regarding issues with translation between ramp-incremental and constant-intensity exercise, the power output at RCP/deoxy-BP often is determined first by identifying the V˙O2 at which they occur and then by linear interpolation of the V˙O2 versus power output relationship, where V˙O2 data are back-shifted to account for the circulatory delays and V˙O2 on-kinetics at the onset of the ramp (as represented by the V˙O2 mean response time). Unfortunately, this practice incorrectly assumes that constant exercise at this power output will yield a steady-state V˙O2 corresponding to RCP/deoxy-BP. Shifting V˙O2 data by the mean response time is only effective at aligning power output at moderate intensities. On the contrary, the V˙O2 slow component that emerges with exercise above the lactate threshold increases the V˙O2 gain and slows V˙O2 kinetics to a greater extent than accounted for by the mean response time (6). The net effect is that V˙O2 falls further behind its corresponding power output as the ramp progresses (e.g., see Fig. 3 of Keir et al. [7]). For this reason, assigning to RCP/deoxy-BP a power output based on linear interpolation of the ramp V˙O2 versus power output relationship results in an overestimation of this power output. This explains why most studies report power outputs for RCP/deoxy-BP that exceed CP/MLSS despite similar V˙O2 values. Concomitantly, constant-power output performed at this level will elicit a non–steady-state V˙O2 response in excess of RCP/deoxy-BP as depicted in Figure 1.

Data from an individual whose V˙O2 associated with RCP/deoxy-BP and CP/MLSS are 3.3 L·min−1. The V˙O2 versus power output response (left) from a 30-W·min−1 ramp-incremental protocol (gray circles) and the V˙O2 versus time responses (right) from two constant power output step transitions to CP/MLSS (244 W, white circles) and the ramp-identified power output at RCP/deoxy-BP (281 W, black circles) are displayed. Note that the mean response time-corrected ramp data predict incorrectly that (i) a constant power output of 281 W will elicit a V˙O2 of 3.3 L·min−1 (overestimation) and (ii) a steady-state V˙O2 of 3.0 L·min−1 will occur at a constant power output of 244 W (underestimation).

With respect to methodological limitations, CP is said to reflect the highest intensity that can be sustained for a prolonged time predominately by oxidative energy provision. Although this is conceptually correct, seldom are physiological verifications performed to test whether the model estimated CP actually satisfies this criterion. This is a critical limitation because CP values are both test and model dependent (8), and therefore, validation of the estimated CP by corroboration of a steady-state V˙O2 during prolonged exercise (e.g., beyond 15 minutes) is the only way to ensure its accuracy. In addition, although it is relatively easy to measure with confidence the V˙O2 associated with CP/MLSS, this too rarely is evaluated. Rather, the V˙O2 often is incorrectly assigned by interpolation of the V˙O2 versus power output data from ramp exercise (2,4), which leads predictably to an underestimation of the V˙O2 associated with a given power output as observed in Figure 1.

To test definitively the equivalence and physiological connection of CP/MLSS with RCP/deoxy-BP, an interventional study designed specifically to affect these variables is required. In this context, Caen et al. (4) found that the power output associated with CP was lower than RCP and deoxy-BP and that this difference was preserved even after each was increased with 6 wk of high-intensity training. The authors concluded that these indices were not equivalent but perhaps a different interpretation is possible. Although V˙O2 data were not featured, linear interpolation of the V˙O2 versus power output relationship was used to determine the power output at RCP/deoxy-BP. As mentioned previously and depicted in Figure 1, this practice will lead to an overestimation of these variables by a difference (~25 to 40 W for a 25- to 30-W·min−1 ramp protocol) that is very close to the delta between RCP/deoxy-BP and CP/MLSS reported by Caen et al. (4). Therefore, it is more likely that these thresholds, which increased in unison with training, were equivalent both before and after training, providing evidence for the interchangeability and the causative link among these physiological indices.

To summarize, much of the literature refuting an interrelationship between RCP/deoxy-BP and CP/MLSS can be explained by (i) the methods used to translate between incremental and constant-intensity paradigms and (ii) a high probability of low accuracy in measures of CP. Although these indices are detected using different approaches, the available data support the hypothesis that each identifies the critical metabolic rate. Nevertheless, to use them interchangeably for task-specific intensity selection, we first must overcome the issue of translating with accuracy the power output associated with ramp-identified phenomena that occur above the lactate threshold. In addition, efforts need to be made to test how closely each identifies the physiological responses expected at the critical metabolic rate.


In their perspective, Broxterman and colleagues (9) argue against an equivalence between CP/MLSS and RCP/deoxy-BP citing disparate physiological determinants, inconsistent correlations among parameters, and differences across “the organism and diverse species” as supportive evidence. We disagree that any of these arguments refute the hypothesis that these indices represent different experimental manifestations of the same metabolic intensity.

Disparate Physiological Mechanisms

They argue that the RCP must, by definition, occur after the critical metabolic rate has been surpassed during ramp-incremental exercise; however, this sequential expression need not be the case. At intensities up to the critical metabolic rate, heightened ventilation offsets the increase in arterial hydrogen ion concentration ([H+]a) that is associated with metabolic and nonmetabolic carbon dioxide (CO2) production by removing CO2 at the lungs at a rate similar to its production (10). The steep rise in arterial lactate concentration associated with intensities beyond the critical metabolic rate, however, overloads this regulatory system causing a near immediate accumulation of [H+]a, a rapid compensatory reflex-increase in ventilation (i.e., the RCP), an uncoupling of ventilation relative to CO2 production, and systematic reduction in end-tidal partial pressure of CO2 (PETCO2). Importantly, RCP is identified as the oxygen uptake (V˙O2) value immediately preceding the above changes; as such, this intensity need not occur after the CP/MLSS.

Broxterman and colleagues also suggest that the RCP may be evident during constant-intensity exercise below CP/MLSS—this statement is counterintuitive and disregards the physiological changes that are required for ventilatory compensation (as it relates to RCP) to occur. As defined previously, the RCP signifies the metabolic rate at which the ability to maintain [H+]a at or near homeostatic levels fails and a ventilatory reflex response is amplified in an attempt to buffer the increase in [H+]a (11). Therefore, the RCP, by definition, cannot exist at exercise intensities below the CP/MLSS because the metabolic rates associated with these intensities are not sufficient to engender an unstable metabolic acidosis. Non–steady states in ventilation and PETCO2 that may be observed during constant-intensity exercise at or slightly below CP/MLSS likely reflect additional drives to breathe that are independent of acid–base ventilatory compensation and are not necessarily evidence of the RCP occurring below CP/MLSS.

Inconsistent Relationships among Parameters

Broxterman et al. proposed that “substantiation of this argument of equivalence, if incorrect, would greatly impede the advancement of our understanding of the mechanistic bases for exercise tolerance in health and disease.” It is difficult to understand how scientific discovery is impeded by divergent views on a topic. In our opinion, true “impediments” to the advancement of our understanding of the mechanistic bases for exercise tolerance are (i) the persistent misinterpretation and misuse of the V˙O2 response to incremental exercise (7) and (ii) the acceptance, as accurate, of the model parameter estimate of CP (12). These two factors may largely explain the high degree of “intrasubject variability” noted between CP/MLSS and RCP/deoxy-BP in previous comparative studies and question the relevance of their implications.

Regarding the first point, it must be considered that because of the muscle–lung transit delay and changes in V˙O2 kinetics that accompany exercise above the lactate threshold, the V˙O2 response to ramp-incremental exercise lags, with increasing magnitude, the change in power output (7). For this reason, the V˙O2 versus power output relationship from an incremental test cannot be used to predict the constant-intensity power output that will elicit the V˙O2 corresponding to the RCP/deoxy-BP, nor can it be used to estimate the V˙O2 associated with CP/MLSS. However, most studies comparing the metabolic rates and work rates associated with RCP/deoxy-BP and CP/MLSS have done so incorrectly by linearly interpolating the ramp-identified V˙O2 versus power output (2,4). This oversight undoubtedly contributes to the high degree of intrasubject variability reported in some studies.

The error associated with estimation of CP is of equal importance when considering studies comparing CP/MLSS to RCP/deoxy-BP. It is becoming increasingly evident that traditional CP testing methods entail a significant error in predicting the highest intensity associated with metabolic steady state. For example, estimates of CP have been shown to be on average ~20 W greater than MLSS (13) and to vary by as much as 50 W depending on the model and exhaustive trial number and duration range (8). In a recent review, Poole et al. (14) indicated that the error in CP prediction from model estimates is ±5% (190 to 210 W for a CP estimation of 200 W). For this reason, results of comparative studies that do not experimentally verify CP should be interpreted with caution.

CP/MLSS and RCP/deoxy-BP across the organism and across species

Inherent in all measurement techniques and methods are technical limitations that prohibit their application in all contexts. That the RCP (and deoxy-BP?) is not evident in small muscle mass exercise where CP (and MLSS?) may be measured does not necessarily refute their equivalence during “whole body” dynamic exercise. In small muscle exercise models, the capacity for muscle metabolism to perturb arterial acid–base homeostasis to the point at which supplementary ventilatory compensation is required would be minimal. In addition, although we appreciate the mechanistic insights that may be gleaned from animal models, we fail to see why the integrative physiological responses of an exercising snow crab, lungless salamander, or horse need reflect those of a human in this context.

Concluding Statement

Our group has repeatedly demonstrated that, when measured appropriately, the V˙O2 associated with CP/MLSS and RCP/deoxy-BP coincide with a high level of accuracy and precision (1,3). These findings support the hypothesis that CP/MLSS and RCP/deoxy-BP are surrogates that are linked together by a common metabolic stimulus. Falsification of this hypothesis, if incorrect, requires sound experimental evidence rather than a point of view. Future work on this topic will need to address the issues of translating appropriately the V˙O2 versus work rate relationship between ramp-incremental and constant-intensity exercise paradigms and the physiological validation of CP.


1. Keir DA, Fontana FY, Robertson TC, et al. Exercise intensity thresholds: identifying the boundaries of sustainable performance. Med Sci Sports Exerc. 2015;47(9):1932–40.
2. Broxterman RM, Ade CJ, Craig JC, Wilcox SL, Schlup SJ, Barstow TJ. The relationship between critical speed and the respiratory compensation point: coincidence or equivalence. Eur J Sport Sci. 2015;15(7):631–9.
3. Bellotti C, Calabria E, Capelli C, Pogliaghi S. Determination of maximal lactate steady state in healthy adults: can NIRS help? Med Sci Sports Exerc. 2013;45(6):1208–16.
4. Caen K, Vermeire K, Bourgois JG, Boone J. Exercise thresholds on trial: are they really equivalent? Med Sci Sports Exerc. 2018;50(6):1277–84.
5. Leo JA, Sabapathy S, Simmonds MJ, Cross TJ. The respiratory compensation point is not a valid surrogate for critical power. Med Sci Sports Exerc. 2017;49(7):1452–60.
6. Keir DA, Benson AP, Love LK, Robertson TC, Rossiter HB, Kowalchuk JM. Influence of muscle metabolic heterogeneity in determining the V˙O2p kinetic response to ramp-incremental exercise. J Appl Physiol (1985). 2016;120:503–13.
7. Keir DA, Paterson DH, Kowalchuk JM, Murias JM. Using ramp-incremental V˙O2 responses for constant-intensity exercise selection. Appl Physiol Nutr Metab. 2018;43(9):882–92.
8. Mattioni Maturana F, Fontana FY, Pogliaghi S, Passfield L, Murias JM. Critical power: how different protocols and models affect its determination. J Sci Med Sport. 2018;21(7):742–74.
9. Broxterman RM, Craig JC, Richardson RS. The respiratory compensation point and the deoxygenation break point are not valid surrogates for critical power and maximum lactate steady state. Med Sci Sports Exerc. 2018;50(11):2379–82.
10. Wasserman K, Beaver WL, Sun X-G, Stringer WW. Arterial H(+) regulation during exercise in humans. Respir Physiol Neurobiol. 2011;178(2):191–5.
11. Meyer T, Faude O, Scharhag J, Urhausen A, Kindermann W. Is lactic acidosis a cause of exercise induced hyperventilation at the respiratory compensation point? Br J Sports Med. 2004;38(5):622–5.
12. Keir DA, Mattioni Maturana F, Murias JM. Reply to “Discussion of ‘Can measures of critical power precisely estimate the maximal metabolic steady-state?’—Is it still necessary to compare critical power to maximal lactate steady state?”—When is it appropriate to compare critical power to maximal lactate steady-state? Appl Physiol Nutr Metab. 2018;43(1):96–7.
13. Mattioni Maturana F, Keir DA, Mclay KM, Murias JM. Can measures of critical power precisely estimate the maximal metabolic steady state? Appl Physiol Nutr Metab. 2016;41:1197–203.
14. Poole DC, Burnley M, Vanhatalo A, Rossiter HB, Jones AM. Critical power: an important fatigue threshold in exercise physiology. Med Sci Sports Exerc. 2016;48(11):2320–34.
Copyright © 2018 by the American College of Sports Medicine