Determination of exercise intensity domains during upright versus supine cycling: a methodological study

Introduction

For almost a century, cardiopulmonary exercise testing (CPET) has been the gold standard of integrative physiologic assessment (Hill & Lupton, 1923), and is widely implemented as the diagnostic tool in both clinical and research settings to provide functional insights regarding general health and exercise tolerance (Kaminsky et al., 2017; Guazzi et al., 2017; Wasserman et al., 2020). However, researchers are still investigating the nuances of the moderating factors, including biological, physiological, biomechanical, and methodological aspects related to CPET (Meyler, Bottoms & Muniz-Pumares, 2021). The pertinent stages, such as the rate of oxygen uptake (V̇O₂) at the gas exchange threshold (GET) and the respiratory compensation point (RCP), provide objective outputs for clinicians and exercise physiologists to tailor personalised exercise interventions that can optimize overall cardiorespiratory health and performance while avoiding premature fatigue development. However, this is not a universal finding, especially for V̇O₂ uptake at RCP during upright exercise (Leo et al., 2017).

For upright cycling, maximal individual capabilities and exercise intensity targets are typically derived from incremental ramp CPET, performed during a single laboratory session (Whipp et al., 1981). Yet, the individual identification of optimal metabolic zones, and an accurate prescription of equivalent targeted constant work-rate exercise from CPET is still challenging. Typically, for a given cycling power output (PO), V̇O₂ assessed during a ramp CPET is often either underestimated or overestimated for constant-work rate exercise at that specific PO, due to several factors including high between-participant variability of maximal intensity capabilities (Keir et al., 2018). Thus, the heterogeneity in the response to training and potential methodological pitfalls further complicate an accurate identification of cycling metabolic zones and equivalent PO’s. Determining the PO for a given V̇O₂ associated with GET–a pivotal indicator whether the exercise intensity- domain was correctly prescribed–is possible via V̇O₂ mean response time (MRT) calculation (Keir et al., 2015; Iannetta, Murias & Keir, 2019). To date, a direct proof of the role of the V̇O₂ MRT in determining exercise intensity boundaries during supine cycling is missing. Though, the role of the V̇O₂ MRT calculated from a single-session CPET remains a controversial issue, especially when adjusting the oxygen uptake/work rate at predetermined GET or RCP (Boone & Bourgois, 2017), as these parameters represent important individual physiological limitations to performance that, crucially to note, are usually sustained over time.

There is a growing interest among the research community and clinical practitioners to investigate CPET procedures and protocols utilized in supine cycling (Dillon et al., 2021, Goulding et al., 2021a, 2021b; Steele et al., 2018). The supine cycling model has been increasingly utilized to study age-related changes in cerebral blood flow (Smirl et al., 2016), acute spleen volume changes (Zubac et al., 2021b), cardiovascular adjustments during exercise in athletes (La Gerche et al., 2013), the effects of gravity acceleration (Bonjour et al., 2010) in addition to study skeletal muscle bioenergetics (Goulding et al., 2020). Yet, concepts like GET and RCP determination received less attention during supine exercise, even though supine cycling has been gaining interest as it provides an appropriate alternative for testing certain populations, who were previously thought to be at risk during upright CPET (Mizumi et al., 2018). For example, Smirl et al. (2016) prescribed an exercise intensity domain by determining the % of HR relative to the maximum HR attained during incremental exercise. Crucially, exercise intensity domains were not categorised through gas exchange parameters for the corresponding PO, which has recently been identified as potentially misleading (Jamnick et al., 2020; Zuccarelli et al., 2018; Meyler, Bottoms & Muniz-Pumares, 2021). Although recent studies have identified that V̇O₂ peak is progressively reduced when moving from an upright position into a supine during cycling (Dillon et al., 2021), there is limited understanding of supine exercise intensity boundaries precisely determined from ramp CPET. For example, the V̇O₂ MRT during supine cycling was not performed in agreement with Iannetta, Murias & Keir (2019), who recommended that a single-visit ramp CPET to volitional exhaustion should be preceded by a moderate-intensity step-transition protocol to accurately determine the GET.

As such, the purpose of the current study was to examine the effect of posture (i.e., supine vs upright) on the most often used indicators of exercise intensity prescription: GET, RCP, V̇O₂ peak, as well as the role of the V̇O₂ MRT in determining individual exercise intensity domains. We hypothesized that supine exercise would be characterised by: (i) a prolonged V̇O₂ MRT compared to upright mainly due to sluggish O₂ transport, and (ii) a slower O₂ transport, resulting in a significantly reduced rate of oxygen uptake at GET, RCP and at VO₂ peak compared to upright exercise.

Materials and Methods

Results

Nineteen moderately-trained males (a self-reported average of 11 h of moderate-vigorous exercise per week) who had normal mean resting values for heart rate, arterial blood pressures and oxygen saturation completed all study procedures.

The oxygen uptake kinetics data during moderate intensity step-transitions in both upright and supine position are given in Table 1. Briefly, no differences were observed between the positions for either the amplitude of V̇O₂ (p = 0.114, 95% CI [−0.01 to 0.08]), time delay (p = 0.054, 95% CI [−0.05 to 6.21]) or MRT (p = 0.364, 95% CI [−9.91 to 3.80]). However, in the supine position baseline oxygen uptake was higher (p = 0.001, 95% CI [−0.24 to −0.12]), and the phase II-time constant (V̇O_2p) was 7 s slower, compared to the upright position (p = 0.001, 95% CI [−12.0 to −0.19]). The V̇O₂ averaged over the last 2 min of these moderate intensity step-transitions was significantly higher in the supine as compared to the upright position (p = 0.001, 95% CI [−0.22 to −0.10]).

The differences in the cardiorespiratory response between the body positions at GET and RCP are presented in Table 2. For the GET, no differences were observed between the two positions for V̇_E, V̇CO₂, respiratory exchange ratio (RER), and GET P_ETCO₂. In the supine position at the GET, V̇O₂ was significantly lower (p = 0.007, 95% CI [63.98–350.19]) than in the upright position, and the attained HR (p = 0.022, 95% CI [1.26–14.51]) and PO (p = 0.001, 95% CI [20.25–33.89]) were also lower, while GET P_ETO₂, p = 0.013, 95% CI [−0.67 to −2.77]) was higher. No differences were observed for V̇O₂ at GET and RCP between the body positions when V̇O₂ readings were given as a percentage of V̇O₂ peak. At RCP, no differences were observed between the two positions for V_E, V̇CO₂, RER and P_ETCO₂. Similar to GET, V̇O₂ (p = 0.012, 95% CI [0.69–5.28]), HR (p = 0.001, 95% CI [2.64–13.47]), and power output (p = 0.001, 95% CI [22.6–43.3]) attained at RCP were lower in the supine position.

Table 3 reports comparisons for the overall cardiorespiratory parameters between the upright and supine cycling protocols. No differences were observed between the positions for V_E, RER and V̇CO₂ (all p > 0.05). However, peak oxygen uptake (p = 0.002, 95% CI [111.5–406.7]) was significantly decreased in the supine position, even when corrected for body weight (by ~4 mL·kg·min⁻¹, p = 0.001, 95% CI [1.3–4.7]). Additionally, the participants attained a significantly lower HR (p = 0.001, 95% CI [4.1–12.8]) during the supine position, and irrespective of correction for body weight, reached a lower power peak output (PPO, p = 0.001, 95% CI [0.36–0.61]) compared to the upright position. The time to exhaustion during the incremental ramp test was reached almost 2 min earlier when supine was compared to upright (p = 0.001, 95% CI [1.40–2.45]). Lastly, no differences were observed between body positions for V̇O₂ peak during the constant-work rate test (p = 0.114, 95% CI [33.8–288.6]), despite substantial differences in the PPO (p = 0.001, 95% CI [26.1–44.0]).

There was no difference in the RPE data between body positions during CPET test (p = 0.368), while exercise termination was predominantly associated with pain in the leg muscles, as compared to dyspnea (9.1 ± 1.2 vs 7.6 ± 1.1, p = 0.001, in upright), and (9.5 ± 1.0 vs 7.6 ± 1.5, p = 0.001, in supine).

Figure 1 depicts Bland–Altman plots examining the agreement between V̇O₂ measured in the upright vs supine position as well as the agreements’ deviation from the mean V̇O₂. For all parameters, the Bland–Altman analysis identified a positive bias in favour of the upright position: the rate of peak oxygen uptake (Fig. 1A: bias = 258 ± 306 ml·min⁻¹); V̇O₂ peak. verification bout (Fig. 1B: bias = 127 ± 335 ml·min⁻¹); gas exchange threshold (GET, Fig. 1C: bias = 207 ± 297 ml·min⁻¹), and respiratory compensation point (RCP, Fig. 1D: bias = 251 ± 393 ml·min⁻¹), with generally wide limits of agreement. In addition, for all parameters there was a homogeneous distribution of differences between upright and supine positions across all values of V̇O₂ peak denoting homoscedasticity (R = −0.001, p = 0.999).

Discussion

The main aim of the current study was to investigate whether individual exercise intensity and the rate of O₂ uptake could be accurately determined in upright and supine cycling and subsequently used to inform task-specific training prescriptions based on corresponding PO’s. To this end, all participants completed two moderate-intensity step-transitions from 20 to 100 W, followed by an incremental CPET and a verification bout (at 90% PPO) in both supine and upright positions. The main findings of the present study can be summarized as follows: (i) no differences were observed in the V̇O₂ MRT between the two positions, although the τV̇O_2p was 7 s slower (p = 0.001) in supine compared to upright position; (ii) lower V̇O₂ in the supine position at GET and RCP compared with the upright position; (iii) a ~6% lower V̇O₂ peak (p = 0.002) was attained during supine cycling, but surprisingly comparable with the upright position (~3.6 L·min⁻¹) during the constant work rate test when both conditions reached a V̇O₂ peak corresponding to 90% PPO.

Determining exercise intensity domains for well-tolerated longer durations of exercise is challenging, both during the upright (Meyler, Bottoms & Muniz-Pumares, 2021; Keir, Pogliaghi & Murias, 2018) and supine cycling model. Compared to upright, supine exercise is characterized by slower O₂ transport (Dillon et al., 2021), greater skeletal muscle deoxygenation (Goulding et al., 2021a), and a greater type II fibre activation cycling (Goulding et al., 2021a, 2021b), altogether likely leading to earlier exercise cessation. Contrary to our hypothesis, the V̇O₂ MRT was not different between body positions likely due to an increase in the phase II-time constant (τV̇O_2p) that was counteracted by a decrease in the circulatory time delay (TD of 3 s, Table 2) hindering the difference in V̇O₂ MRT between body positions during moderate-intensity step-transitions (from 20 to 100 W). Of note, TD has no physiological equivalent (Rossiter, 2011), and subsequently confounds an accurate identification of potential metabolic zones for supine cycling, even though the phase II of the τV̇O_2p kinetics were significantly slower in supine moderate intensity step-transitions. Phase II τV̇O_2p kinetics play an essential role in exercise tolerance determination by mediating the PO at which muscle metabolite accumulation exceeds removal occurs–commonly defined as “ceiling” of tolerable endurance exercise (Goulding et al., 2021c), thereby preventing increased perception of effort and premature exercise cessation. A plausible explanation for the differences found (7 s, Table 2) would be that the supine cycling leads to kinetic disassociation between muscle O₂ uptake and V̇O₂ uptake, primarily caused by the lower baseline perfusion and sluggish O₂ delivery kinetics during supine exercise (Rossiter, 2011). Moreover, besides the well-known methodological challenges (e.g., including low reproducibility, cardiorespiratory fitness status and pedal rate) in the MRT calculation due to the left-shift, the rate of O₂ uptake and the role of the TD requires major consideration (Keir, Pogliaghi & Murias, 2019), especially when adjusting the V̇O₂/PO from a single CPET supine session (Boone & Bourgois, 2017).

While different V̇O₂ peaks between the two body-positions are plausible (Dillon et al., 2021), the scientific intrigue on whether the exercise intensity domains are different, remains. Here we sought to implement methodological procedures and to determine whether well-accepted markers of exercise intensity including the GET and RCP are transferrable from one position to the other, when corrected for V̇O₂ MRT. Significantly lower V̇O₂ uptake at GET and RCP were found for the supine compared to upright cycling (Table 3), and further discrepancy between V̇O₂ uptake at GET and RCP in different body-positions can be observed in the Bland-Altman plots in Fig. 1 (panels C&D). Specifically, the differences in V̇O₂ uptake between upright and supine postures at GET and RCP observed in our study (Table 3) are consistent with a reported ~10% higher V̇O₂ at RCP in an upright condition (Goulding et al., 2021c). In parallel to V̇O₂ uptake at RCP, Goulding et al. (2021c) measured the NIRS-derived deoxy[Hb] of the knee extensors to show that skeletal muscle deoxygenation profile and breath-by-breath O₂ uptake, do not share the same underlying physiological mechanisms in different body postures. However, these findings are inconsistent with work by Keir et al. (2015) using a classical, upright cycling model to corroborate that RCP and NIRS-derived deoxy[Hb], among other exercise intensity indicators, occur at similar V̇O₂ dependent on a common physiological phenomenon. The discrepancies between the studies can be explained by the loss of hydrostatic pressure and muscle perfusion during supine cycling, which reduces the pressure head for myoglobin-facilitated O₂ diffusion, thereby limiting the rate of the O₂ uptake increase at the onset of exercise (Poole, Musch & Colburn, 2022). An additional explanation implies that the V̇O₂ determined at GET or RCP typically lags behind the true metabolic demands of that given intensity (e.g., PO) above the GET during classical upright cycling. At times, even if the V̇O₂ MRT and the loss of mechanical efficacy between the GET and RCP are accounted for Keir et al. (2018) and Keir, Pogliaghi & Murias (2018). This typically leads to a situation where the ramp-identified PO elicits a metabolic intensity greater than the highest metabolic rate at which prolonged exercise is sustained, resulting in loss of locomotion efficiency, premature fatigue development and exercise cessation, which further exacerbate the determination of supine exercise intensities.

Recently, posture-related changes in O₂ transport and V̇O₂ peak were investigated during exhaustive CPET in both upright and supine positions (Dillon et al., 2021). Significantly lower peak cardiac outputs (computed from peak HR and peak stroke volume) were observed during supine cycling compared to upright and semi-supine postures via Doppler echocardiography. In agreement, we found a 9-beat lower HR max. on average during maximal CPET in supine compared to upright posture. Bringard et al. (2010) reported that following a 35-day bed rest, V̇O₂ peak declined by 38% and 17% in the upright and supine postures, respectively. Interestingly, the V̇O₂ peak decreases in the upright posture were predominantly explained via cardiovascular deconditioning (cardiac output ↓45%), whereas the cardiovascular responses during supine CPET remained unchanged. The authors postulated that impaired peripheral O₂ utilization and diminished skeletal muscle oxidative capacity likely underpinned the substantial decline in the supine VO₂ peak after 35-days of bed rest exposure (Bringard et al., 2010), while different mechanisms were responsible for the significant posture-related difference in V̇O₂ decline during upright exercise. These claims have recently been confirmed with skeletal muscle bioenergetics during supine CPET by Goulding et al. (2021a, 2021b), who showed that the supine V̇O₂ peak is partially limited via impaired O₂ extraction fraction of the superficial knee extensor muscles, as observed from the greater rate of muscle deoxygenation at workloads above 165 W during supine cycling. It was suggested that O₂ extraction is the limiting factor of oxygen uptake at higher supine intensities (especially associated with prolonged bed rest). Hence, there are misleading determinations of exercise intensity domains during supine cycling as a result of the lack of agreement around intensity domains determined during upright cycling.

Emerging evidence points to the notion that testing specificity is necessary during CPET procedures that inform exercise prescription and to avoid the methodological pitfalls of assuming exercise prescription is the same for different body-positions. Overall, considering the conflicting evidence observed between this study and the existing literature, further research is warranted in this area, since the commonly applied approaches may erroneously mislabel indicators of exercise intensity (including GET and RCP) that may otherwise have been correctly identified by the inclusion of the V̇O₂ MRT. Individuals are thereby prescribed inappropriate exercise intensities. Moreover, the issue pertaining TD should not be overlooked when attempting to translate a given metabolic rate (e.g., the V̇O₂ uptake) into a corresponding task specific PO, especially for the supine variant.

The current cross-sectional study is not exempt from limitations. A series of constant work-rate cycling bouts to determine critical power or maximal lactate steady state would have provided additional insight regarding position-specific markers of fatigue development. Although initially considered, such protocols were unfortunately discarded due to concerns of a high dropout rate because of Covid-19 infections or lockdowns. The practical application of a verification bout conducted at 90% PPO seems problematic and relates, potentially, to the methodological weakness of 90% PPO (see Niemeyer, Leithäuser & Beneke, 2020), but could be resolved by a verification bout performed at supramaximal PPO.

Conclusions

In conclusion, significant differences in the V̇O₂ uptake at GET, RCP and V̇O₂ peak between body positions were observed in healthy young men. Yet, the V̇O₂ MRT was not different between body positions, although the τV̇O₂ kinetics lagged behind the actual metabolic demands, which should be taken into account when attempting to prescribe individualized exercise. Overall, our work adds to the on-going debate on exercise prescription by indicating that the inclusion of moderate-intensity step-transitions prior to a maximal CPET procedure is a viable and appropriate addition to ensure accurate determination of exercise-intensity domain(s). Further research is warranted to target the exact metabolic rates at which exercise is well tolerated for longer durations and to explore different methodological pitfalls associated with the exercise intensity determination at GET and RCP in supine posture, as well as the physiological mechanisms governing central and peripheral factors that limit O₂ transport in different populations of varying cardiorespiratory fitness level.

Supplemental Information