The V̇o_2peak vs. V̇o_2max Expedient

For young, healthy subjects that are used to driving themselves to exhaustion while cycling or running, the incremental/ramp cardiopulmonary exercise test yields a highly reproducible V̇o_2max irrespective of exercise test protocol, work rate forcing function, or pacing strategy (10). However, this may not be assured with exercise-naïve, unmotivated, and/or clinical populations. A serious concern here is that, when a training or therapeutic paradigm is being tested, the efficacy of the intervention may be overestimated in repeated testing as the subject gains experience and possibly enhanced confidence. Thus, if an accurate V̇o_2max was not measured initially, whether or not the limit of the integrated O₂ transport/utilization system has actually been increased cannot be determined. In an attempt to mitigate this substantial source of experimental error, investigators have resorted to the term V̇o_2peak (~5,200 PubMed references to date) as simply the highest V̇o₂ reached on a given test. Unfortunately, this procedure cannot discriminate among subjects who cease exercise because of lack of motivation, perceived discomfort, or a plethora of other reasons, none of which are related necessarily to their maximal rate of O₂ transport/utilization (38).

So-Called Secondary Criteria Used to Validate V̇o_2max

In those instances where a V̇o₂ plateau is not attained as definitive evidence of V̇o_2max, investigators commonly elect to substantiate that V̇o_2max was actually achieved by utilization of so-called secondary criteria assumed to validate V̇o_2max. These criteria most commonly include one or more of the following: heart rate (HR) ≤10 beats/min or ≤5% of the age-predicted (220-age) maximum, blood lactate concentration ≥8 mM, or respiratory exchange ratio (RER) >1.00, 1.10, or 1.15 (31, 38). The obvious problem with this “one size fits all” approach is that there is a broad range of maximal values for each of these variables in healthy and especially patient populations. Thus perusal of the literature reveals that HRmax has a 95% confidence interval of ± 35 beats/min (5, 11, 26), maximal RER varies from 1.0 to 1.44 (46, 50), and maximal blood lactate concentration varies from 4 to 17 mM (5, 16, 50). Moreover, in the same subjects, changing the slope of the ramp (i.e., speed of the work rate increment) can alter these criteria in opposite directions. Specifically, faster ramps yield higher RERs but lower HRs at the same V̇o_2max (7, 8, 50, 52, 54). Also, equally pernicious with regard to so-called V̇o_2max criteria is that distinct populations may have very different HRmax values. For instance, extremely sedentary adults and also some populations of children may have lower HRmax values than predicted (11, 19). For North American children/young adults aged 8–18 yr, HRmax averages 187 beats/min (11) compared with 205 beats/min in Scandinavian children (5). An additional consideration here is that patient populations taking β-blockers will experience bradycardia during exercise and thus evince very low HRmax values.

Utilization of these criteria can allow for a 30–40% underestimation of the true V̇o_2max and/or an errant rejection of tests in which subjects had actually achieved their V̇o_2max (31, 38). At its least destructive, employment of secondary criteria increases variability and weakens experimental design, thereby increasing the likelihood of finding a false negative (type II error, incorrectly retaining a false null hypothesis); at its worst, it may create a false positive (type I error, incorrectly rejecting a true null hypothesis).

A Viable Strategy for V̇o_2max Validation

In contrast to the notion that a discrete work rate yields a predictable V̇o₂, recognition of the existence of the V̇o₂ slow component (17, 35), especially during constant work rate exercise in the severe-intensity domain (i.e., above critical power, CP) (39), means that, provided the exercise is of sufficient duration, V̇o_2max will be achieved for any supra-CP work rate (Fig. 2) (35–37, 42). However, there is a smaller range of constant work rates, but also see above the maximum obtained on the ramp test, which are sustainable for several minutes, each of which will yield V̇o_2max before exhaustion (Fig. 3). Accordingly, and consistent with the earliest recognition that “an obligatory plateau of V̇o₂ provides unequivocal identification of V̇o_2max” (21, 50), a plateau of V̇o₂ vs. work rate (or speed) can be simply constructed from just two exercise bouts (i.e., a maximal ramp/incremental and a subsequent exhausting constant work rate test at a higher work rate than that achieved at the end of the ramp test, Fig. 4). The key is that, although all work rates above CP but below that at which the highest V̇o₂ was measured on the ramp test can bring V̇o₂ to V̇o_2max, the subsequent validation test must be at a higher work rate than attained in the ramp test. There is not a straightforward answer to the question: How much higher? The investigators must select a work rate that is sufficiently higher than that attained on the ramp test to give the V̇o₂ signal for the higher work rate the opportunity to emerge from the extant noise. In the event that the subsequent test produces a plateau signifying V̇o_2max, this signal would be lower than expected for the work rate based on the previous V̇o₂-work rate slope, thereby allowing construction of the relationship shown in Fig. 4. On the other hand, the work rate must be sustainable for sufficient duration that the kinetics allow achievement of the same or a greater V̇o₂ if such is possible. To date, a work rate ~10% higher than that completed on the previous ramp test has proven effective (3, 29, 31, 40) but may not be ideal for all populations or circumstances. Resolution of this specific issue would be valuable. Imposition of work rates below the maximum achieved on the ramp test may allow construction of a plateau of V̇o₂ against work rate but will not extend the relationship to the higher work rate critical to define V̇o_2max (45; see also the RISE-95 test validated in chronic heart failure patients, Ref. 7).

For exercise tests run on the motor-driven treadmill, the same principles apply. However, the potential flexibility of increasing treadmill speed and/or incline is available. For experienced runners, Midgley and colleagues (29, 31) have advocated an increased running speed of 0.5 km/h for the verification bout and, in addition to the levelling-off criteria for V̇o₂, a difference of less than 4 beats/min in HR between the maximum reached on the ramp/incremental and the verification bout (30). As was true for the cycle ergometry paradigm, identifying the correct work rate (i.e., speed/incline) for the treadmill verification bout within the specific pertinent populations tested is a valuable goal. Even when, or if, this goal is achieved, it is important that the investigator build into the experimental design the latitude to select sufficient verification bouts to provide definitive evidence for V̇o_2max.

The specific rest period intervening between the ramp and subsequent validation test does not seem to be critical. Specifically, 20 min has been recommended (33), but, whereas longer may be advisable for patient populations, healthy subjects may want or require far less (i.e., 5 or 10 min). Procedurally, asking subjects or patients to perform two exhausting bouts of exercise within the same laboratory or clinical testing session is feasible, and typically subjects prefer the constant work rate test to the ramp test. The limit of tolerance during this constant work rate test is usually 3–6 min. Depending on the subject’s V̇o₂ kinetics, work rates that induce exhaustion in much less than 3 min might be in the so-called “extreme” work rate domain and therefore fail to achieve V̇o_2max (22, 35, 56).

To date, this validation protocol has been performed successfully with children (6), healthy sedentary (2), athletic (51), and obese (43, 57) adults, and patient populations (e.g., cystic fibrosis, 44). It is also notable that, if, in a particular investigation, CP is being measured using repeated exhaustive severe-intensity constant work rate tests, one or more of these can be used to validate V̇o_2max provided that they are above the maximum power output achieved on the ramp test. Selecting a particular work rate that is, for example, 10% above that which achieved the highest V̇o₂ on the ramp test means that the absolute level to which V̇o₂ rises is essentially independent of the contribution of the V̇o₂ slow component, V̇o₂ kinetics, or the tolerable duration of the exercise (22, 23, 34, 47, 56).

One alternative protocol to the exhausting constant work rate test recommended above is the 3-min all-out effort (49). This entails the subject cycling as fast as possible against a fixed resistance causing the power output to peak within a few seconds before falling precipitously to asymptote at CP after ~2 min. On this test, V̇o_2max is typically achieved within ~1 min and sustained for the remainder of the test despite developed power falling far below that achieved at V̇o_2max during ramp incremental exercise. Although the full potential of this test across populations remains to be determined, its extremely strenuous nature may be contraindicated in at least some patient populations and even in healthy individuals who are unused to maximal exhausting exercise. Murgatroyd et al. (32) have reported that an exhausting ramp incremental test followed immediately by a maximal-effort 3-min variable-power sprint test effectively validates V̇o_2max, suggesting that key parameters of aerobic fitness can be appraised within a single exercise test. Unfortunately, the variable power condition for both of these options confounds construction of the relationship demonstrated in Fig. 4 that defines V̇o_2max. Recently, there has been interest in the use of self-paced incremental exercise tests (using 5 × 2 min stages at fixed increments of rating of perceived exertion) for the determination of V̇o_2max (27). Such a procedure, which has been shown to produce similar V̇o_2max values to traditional ramp incremental tests (9, 20, 24), might be considered to have greater ecological validity than conventional protocols in which work rate is externally controlled and incremented in a strictly linear fashion. Moreover, the “closed-loop” nature of the self-paced protocol obviates the necessity for the researcher or clinician to estimate starting work rates and ramp rates. The utility of self-paced maximal exercise tests in the laboratory or clinical setting requires further evaluation but does not negate the value of the verification phase advocated herein, which has the singular advantage of allowing construction of the graphical solution to V̇o_2max shown in the top panel of Fig. 4.

What Constitutes a “Plateau” for V̇o_2max Identification

During incremental exercise, a plateau in V̇o₂ is considered to represent the attainment of the definitive V̇o_2max. However, exactly how the existence of this plateau is defined is rarely considered. The criterion proposed by Taylor et al. (48) of a change in V̇o₂ of <2.1 ml·kg^-1·min^-1 between consecutive stages as the subject approaches exhaustion is often applied indiscriminately and without consideration of the expected increment in V̇o₂ for the ramp or incremental rate being utilized. Moreover, the window over which V̇o₂ is averaged (i.e., 15, 30, or 60 s) can significantly influence both the absolute V̇o₂ and the proportion of plateaus identified (1, 23). The inclusion of a verification phase in the test battery helps to circumvent these issues. For example, consider a subject who completes a ramp incremental test and achieves a V̇o_2peak of 2,500 ml/min at a peak work rate of 200 W (i.e., baseline of 500 ml/min during “unloaded” cycling and a V̇o₂ gain of 10 ml·min^-1·W^-1 throughout the test). If the verification phase is completed at 10% above the peak work rate (i.e., 220 W), then the expected V̇o₂ would be 2,700 ml/min. Given a possible error in V̇o₂ measurement of ~3% (~80 ml/min), a plateau could be accepted if V̇o₂ was ≤ 2,620 ml/min.

Midgley and colleagues have advocated prediction of the “expected” V̇o₂ for the verification test using a least-squares linear regression line fitted to the ramp test V̇o₂ data for minute −6 to −2 before exhaustion (30, 31). The slope of this line is then extended to the absolute work rate (e.g., 100% of maximum work rate on the ramp test). A V̇o₂ plateau and therefore V̇o_2max verification has been defined as a difference between the modeled and actual V̇o₂ of >50% of the regression slope as determined above. This approach is advocated in preference to choosing some arbitrary V̇o₂ plateau “threshold” of 100, 200, or even 280 ml/min (31).

Conclusions

The V̇o_2max is an essential paradigm in the evaluation of both the integrated capability of an individual’s pulmonary-cardiovascular-metabolic systems to transport and utilize O₂ to support muscle contraction and the efficacy of interventions designed to enhance one or more components of the O₂ transport/utilization pathway. Measurement of V̇o_2max will therefore remain a vital technique in the arsenal of researchers in exercise physiology, sport scientists, and clinical exercise specialists. In this article, we have highlighted some of the methodological issues that can thwart accurate assessment of V̇o_2max and cautioned against the acceptance of V̇o_2peak measured during ramp incremental exercise as a maximum value in all but those who are familiar with maximal exercise testing and are highly motivated. We advocate the inclusion of a short constant work rate verification phase, completed at a higher work rate than that achieved during the ramp test, in the exercise test battery to enable verification of the V̇o_2max. Fundamental tenets of this approach are that it is objective, specific to the particular subject and exercise testing format or modality, and sufficiently robust that it is not impacted substantially by day-to-day differences in physiological responses (29, 30, 31).