Acute Hypoxia Decreases Maximum Fat Oxidation Rate During Step Incremental Exercise Normalized to Respiratory Compensation Point

ABSTRACT

Whether fat oxidation (FATox) is altered during exercise in hypoxia remains equivocal due to differences in experimental protocols. Furthermore, to date no investigation has reported the effect of hypoxia on maximal fat oxidation rate (MFO). Therefore, the aim of this study was to assess substrate utilization in normoxia and hypoxia and determine MFO. Seventeen active adults (12 M/5F) performed ramp and step incremental testing in normoxia (FiO₂ = 0.209; NORM) and normobaric hypoxia (FiO₂ = 0.135; HYPO). Respiratory compensation point (RCP) determined from ramp testing was used to normalize relative intensity across 6 constant workrate steps in the moderate and heavy domain. Indirect calorimetry was used to measure cardiorespiratory responses and estimate substrate utilization and MFO. Linear mixed modeling was used to compare measurements in NORM and HYPO, where intensity was expressed as a function of absolute or relative workrate. Cardiorespiratory responses to exercise were similar in NORM and HYPO when the workrate was expressed as a function of relative intensity. FATox was decreased across all stages in HYPO (p < 0.001), which was associated with a 22% decrease in MFO (HYPO: 0.26 ± 0.07 g·min⁻¹, NORM: 0.34 ± 0.07 g·min⁻¹; p < 0.001, d = 1.16). MFO occurred at a similar percentage of $\dot{\mathrm{V}}$O_2max in both NORM (38 ± 8%) and HYPO (38 ± 8%; p = 0.89, d = 0.04). MFO was decreased in HYPO regardless of whether the workrate was expressed as a function of relative or absolute intensity. This suggests that hypoxia may exert a direct effect on regulation of fuel selection during exercise, independent of the reduced absolute workrate when normalizing relative intensity to RCP.

Highlights

• FATox was lower across all workrates below RCP in HYPO when normalized for either relative or absolute intensity. Consequently, the MFO rate was lower in HYPO.

• MFO occurred at a similar relative intensity (as a %VO_2max) in both NORM and HYPO

• CHOox was similar when normalized for relative intensity in HYPO, despite a lower gross energy expenditure across each workrate

• These results suggest that oxygen availability plays an important role in regulating metabolic flexibility during exercise across all intensities in the moderate and heavy domains

1. Introduction

Maximum fat oxidation (MFO) is connected to endurance exercise performance in athletes, as well as to weight management and metabolic control in sedentary populations (Amaro‐Gahete et al. 2019). The contribution of fat oxidation (FATox) to adenosine triphosphate (ATP) resynthesis is affected by multiple factors including training status, habitual diet, relative exercise intensity and exercise duration (Amaro‐Gahete et al. 2019; Maunder et al. 2018). Consequently, conditions such as hypoxia that can modulate the relative intensity of exercise and tissue oxygenation, can also affect the rate of FATox and carbohydrate oxidation (CHOox) (Griffiths, Shannon, Matu, King, Deighton, and O’Hara 2019b). Indeed, when examining the metabolic effects of exercising in hypoxia compared to normoxia, the results can differ depending on whether similar absolute intensities or similar relative intensities are used to compare both conditions (Griffiths, Shannon, Matu, King, Deighton, and O’Hara 2019a; Young et al. 2019).

During exercise in acute hypoxia there is a decrease in convective oxygen (O₂) transport to the working muscles compared with normoxia (Lundby and Van Hall 2002). Consequently, key indicators of aerobic capacity including the gas exchange threshold (GET), respiratory compensation point (RCP), critical power (CP) and maximum O₂ consumption ($\dot{\mathrm{V}}$O_2max) have been consistently reported to decrease in a predictable, quasi‐linear manner as arterial O₂ pressure progressively declines (Shearman et al. 2016; Townsend et al. 2017; Wehrlin and Hallén 2006). Thus, when exercise is performed at similar absolute workrate, the relative intensity of exercise (i.e., as a percent of $\dot{\mathrm{V}}$O_2max) is greater in hypoxia compared to normoxia. In particular, this greater relative intensity of exercise in hypoxia may lead to increased accumulation of metabolites associated with loss of homeostasis including unsustainable changes in blood lactate concentration [La⁻]_b, inorganic phosphate, adenosine diphosphate, and H⁺ (Vanhatalo et al. 2011). When the relative intensity is matched in normoxia and hypoxia though, this has been suggested to isolate the effect of hypoxia per se (Lundby and Van Hall 2002), and thereby represents a better methodological approach to examine the effect of hypoxia independent of the synergistic impacts of exercise intensity. Also from a practical perspective, during self‐paced exercise at high altitude, people tend to select lower absolute workloads which counterbalances the elevated homeostatic disturbance and attenuates increased rating of perceived exertion (Griffiths, Shannon, Matu, King, Deighton, and O’Hara 2019a). However, even when attempting to normalize the relative exercise intensity in hypoxia, inconsistent findings in substrate utilization have been reported (Griffiths, Shannon, Matu, King, Deighton, and O’Hara 2019b; Lundby and Van Hall 2002; Matu et al. 2017; O’Hara et al. 2017; Péronnet et al. 2006). To date however, all studies that normalized for equivalent relative intensity in normoxia and hypoxia employed fixed percentages of $\dot{\mathrm{V}}$O_2max. This method of exercise intensity prescription has come under scrutiny in recent years though, and is discouraged due to between subjects variations in the relationship between the maximum metabolic steady state (MMSS) and $\dot{\mathrm{V}}$O_2max (Iannetta et al. 2020; Jamnick et al. 2020).

Matching equivalent metabolic stress between individuals during exercise requires appropriate consideration of the three exercise intensity domains that is, moderate, heavy, and severe. The physiological mechanisms which define these domains are governed by a cascade of events which originates via metabolic disturbances within the active muscles (Black et al. 2017). During incremental exercise, there are two well‐described physiological breakpoints which occur prior to attainment of $\dot{\mathrm{V}}$O_2max, which we shall refer to as the gas exchange threshold (GET) and respiratory compensation point (RCP) respectively, despite some ambiguity surrounding terminology, definitions and agreement (Binder et al. 2008; Jamnick et al. 2020). For decades it has been recognized that substrate utilization kinetics resemble a “crossover” pattern whereby the contribution of FATox declines and CHOox increases during incremental exercise and this pattern appears to be more closely related to the intensity at which the GET and RCP occur as opposed to $\dot{\mathrm{V}}$O_2max (Brooks 1997). Therefore, the choice of “anchor” to fix relative intensity could affect substrate utilization patterns and introduce unnecessary between‐subjects variability. For example, it has been shown that the GET and RCP can occur from 45% to 74% and 69%–96% of $\dot{\mathrm{V}}$O_2max respectively (Iannetta et al. 2020). Another methodological consideration is that many studies which examine FATox rate and/or MFO used 3–4 min step durations during incremental testing (Amaro‐Gahete et al. 2019). Accurate determination of substrate utilization requires a metabolic steady state (Jeukendrup and Wallis 2005), however as exercise intensity increases into the heavy domain, the duration required to achieve a steady state can extend beyond 3–4 min due to a combination of slowing $\dot{\mathrm{V}}$O₂ kinetics, emergence of the $\dot{\mathrm{V}}$O₂ slow component and stabilization of the blood bicarbonate pool (Jeukendrup and Wallis 2005; Whipp 2007).

Thus, the aim of this study was to evaluate MFO and substrate utilization in normoxia and hypoxia using an exercise intensity model that homogeneously distributes the relative intensity in each step and in each condition into similar intensities within the moderate and the heavy intensity domains, while maximizing the time allowed to reach steady (or quasi steady) state responses. We hypothesized that the MFO would decrease in hypoxia compared to normoxia but that the MFO would occur at a similar relative intensity of exercise.

2. Methods

2.1. Participants

Eighteen active adults (5 female and 13 male) were recruited for this study. A sample size of at least 12 participants was calculated based on a two‐tailed distribution and a large effect size of 0.8 using the G*Power software (version 3.1.9.6). Data from 15 active adults (5 female and 10 male) from whom successful data collection was completed are part of this study. All participants were healthy (e.g., no known cardiovascular, neuromuscular, or metabolic diseases), active (i.e., performing endurance exercise at least three times per week), non‐smokers, and did not travel to an altitude > 1500 m within the past 3 months. The participants visited the laboratory on four separate occasions to perform ramp, and step incremental testing in normoxia (NORM: FiO₂ ≈ 21%) and hypoxia (HYPO; FiO₂ ≈ 13.5% ≈ 3300 m simulated altitude). All procedures were conducted according to the Declaration of Helsinki and ethical approval was obtained via the institutional review board (HBKU‐IRB‐2024‐33). After being informed about study details and potential benefits and risks, participants signed the written consent.

2.2. Experimental Design

This study was a repeated‐measures controlled trial with counterbalanced order of condition. Participants visited the HBKU Exercise Physiology laboratory on four occasions that were separated by a minimum of 48 h. Visits 1 and three included a ramp incremental test, while visits 2 and four included a step incremental test in HYPO and NORM randomly. For HYPO, participants were connected to a gas blender system (Altitrainer 200, SMTEC, Switzerland), via a two‐way non‐rebreathing respiratory valve (part # 2730, Hans Rudolph, USA) which uses nitrogen dilution to reduce FiO₂ to a target of 13.5% that is, normobaric hypoxia equivalent to ≈3300 m simulated altitude after accounting for the effect of water vapor pressure. On the inspired side FiO2 was measured continuously using a rapid response laser diode oxygen analyzer (model 7035, Oxigraf, USA) and the Altitrainer was manually adjusted to maintain the target FiO2%. To achieve participant blinding to the condition, the Altitrainer was connected to the mask at each visit but set to FiO₂ of 21% for the NORM condition. All testing sessions for a given participant were performed at a similar time of the day (±1 h) and under similar laboratory environmental conditions (i.e., ∼20°C and ∼60% ambient relative humidity). Participants were requested to abstain from performing any high‐intensity exercise 24 h before testing, refraining from any exercise on the day of the test, and maintaining their habitual lifestyle throughout the study.

2.3. Experimental Protocols

2.3.1. Dietary Control

To control for the potential impact dietary patterns on substrate utilization participants were instructed to follow the same fasting duration of 10 ± 2 h for morning test sessions that is, overnight fast without consumption of breakfast or a minimum of 6 h for afternoon test sessions. Participants were also instructed to record their food and fluid intake for 24 h preceding the first test session using a food record diary, and thereafter consume same food amount, type, and timing prior to each additional lab visit. All participants confirmed adherence to these dietary instructions.

2.3.2. Ramp Incremental Testing

All cycling testing sessions were performed on an Excalibur Sport cycle‐ergometer (Lode BV, Groningen, The Netherlands). Participants were instructed to cycle at their preferred cadence, typically around 80–90 revolutions per minute (rpm), remain with ± 5 rpm of this cadence, and to refrain from talking during the test. Ramp tests were performed in visits one and three in NORM and HYPO to determine the GET, RCP, and maximal/peak responses in each condition. The sequence of conditions was randomized, counterbalanced, and single‐blinded. Prior to testing participants completed a validated questionnaire to predict $\dot{\mathrm{V}}$O_2max (Nes et al. 2011) which thereafter was used to estimate individualized ramp rates so each participant reached task failure in 8–12 min. The test commenced with 5 min baseline cycling at 20 W, followed by the ramp phase (range = 20–30 W.min⁻¹) until voluntary exhaustion or the inability to maintain cadence above 70 rpm despite strong verbal encouragement. The Exercise Threshold app (Keir et al. 2022) was used to plot the ventilatory and gas exchange data, and two expert researchers determined the GET and RCP. The GET was identified as the point at which $\dot{\mathrm{V}}$CO₂ started to increase disproportionally in relation to $\dot{\mathrm{V}}$O₂, there was a systematic rise the $\dot{\mathrm{V}}$E and end‐tidal oxygen tension (PetO₂) versus $\dot{\mathrm{V}}$O₂ relationship, while the ventilatory equivalent of $\dot{\mathrm{V}}$CO₂ ($\dot{\mathrm{V}}$E/$\dot{\mathrm{V}}$CO₂) and end‐tidal carbon dioxide tension (PetCO₂) remained stable. the RCP was determined as the second breakpoint in the pulmonary ventilation ($\dot{\mathrm{V}}$E) versus $\dot{\mathrm{V}}$O₂ and where the PetCO₂ began to fall after a period of isocapnic buffering. The relationship between $\dot{\mathrm{V}}$E/$\dot{\mathrm{V}}$CO₂ against $\dot{\mathrm{V}}$O₂ was also used for verification of the RCP (Keir et al. 2022). To estimate power output (PO) at the metabolic rate coinciding with GET and RCP, mean response time was fixed at 66% and 100% (respectively) of the participant individual ramp rate. $\dot{\mathrm{V}}$O_2max was considered as the highest 30 s rolling average achieved during the test.

2.3.3. Steady‐State Incremental Exercise

A step test comprising six constant workloads individualized for each participant were performed during visits two and four. A graphical representation of this procedure is presented in Figure 1. The test always commenced with a 4 min baseline cycling stage at 20 W. For the six step workrates, the PO equivalent to 80% of the RCP for each condition was first set as an “anchor point” for the sixth stage. Thereafter, 80% divided by 6 stages dictates a 13% increase in workrate per step. Note in some cases where the first workload was calculated to be lower than 20 W, the lowest workrate was used as the baseline phase. The duration of each step in order from lowest to highest was 4, 4, 4, 6, 8, and 8 min. Longer durations were chosen for the heavy domain to allow for expression of the $\dot{\mathrm{V}}$O₂ slow component.

FIGURE 1.

The step test included six constant workrate stages. An initial anchor point was set for stage 6 at 80% of the RCP corrected for estimated mean response time. The step increment was individualized to each participant (in both conditions) and equal to 13% of the power for stage 6. Step duration was 4–8 min. Nb: there was a final maximal effort stage until task failure implemented as part of another study (data from this stage are not presented in the current manuscript).

2.4. Measurements

2.4.1. Anthropometry

Stature, body mass, and body composition were taken using a Tanita MC‐780 MA (Pole) multi‐frequency segmental body composition monitor wearing minimal clothes and barefoot on the first and third visits.

2.4.2. Cardiorespiratory Variables

Participants wore a mask connected to a turbine and a gas sampling line. For all exercise testing. The FiO₂, FiCO₂ and $\dot{\mathrm{V}}$ _E were measured using a metabolic cart (Cosmed Quark PFT, Italy). Before the measurements, the metabolic cart gas analyzer was calibrated according to the manufacturer's recommendations with known atmospheric air and a calibration gas mixture containing 5.05% CO₂ and 16.01% O₂. The volume turbine was calibrated using a 3 L syringe. All gas exchange data was recorded on breath‐by‐breath basis, then filtered for outliers (> 2 standard deviations outside the nine breath rolling average). Heart rate was measured via standard two‐electrode wireless ECG sensor (Delsys Trigno) and recorded at a sample rate of 2000 Hz using an analog to digital data‐acquisition system (LabChart, ADInstruments, Australia). All cardiorespiratory variables averaged across a 15 s window centered on the time point where GET and RCP occurred, whilst all data at $\dot{\mathrm{V}}$O_2max was averaged for 30 s across the same period that $\dot{\mathrm{V}}$O_2max occurred (as mentioned).

2.4.3. Substrate Utilization

During the step incremental test, all breath‐by‐breath gas exchange data was initially filtered for outliers as described, and bin averaged across 30 s periods. Indirect calorimetry was used to estimate FATox and CHOox rates during the step tests. The average $\dot{\mathrm{V}}$O₂ and $\dot{\mathrm{V}}$CO₂ for the last two 30 s bin average periods (i.e., 60 s in total) was inserted into equations shown in Table 1 to calculate substrate utilization, assuming no contribution of protein oxidation (Jeukendrup and Wallis 2005). MFO was defined as the highest value measured at any stage during the step test (Amaro‐Gahete et al. 2019). Also, energy expenditure (EE) was used to calculate the relative contribution of CHOox (%CHOox) and FATox (%FATox) to total EE.

TABLE 1.

Equations used to estimate substrate oxidation from indirect calorimetry for CHOox (carbohydrate oxidation), FATox (fat oxidation), and EE (energy expenditure) during exercise (Jeukendrup and Wallis 2005). The expression of $\dot{\mathrm{V}}$CO₂ (volume of carbon dioxide consumption) and $\dot{\mathrm{V}}$O₂ (volume of oxygen consumption) were in L.min⁻¹ for the equations.

Substrate	Equation	Eqn #
CHOox (g·min−1)	(4.210 * $\dot{\mathrm{V}}$CO2) − (2.962 * $\dot{\mathrm{V}}$O2)	1
FATox (g·min−1)	(1.695 * $\dot{\mathrm{V}}$O2) − (1.701 * $\dot{\mathrm{V}}$CO2)	2
EE (kcal·min−1)	(4.07 * CHOox) + (9.75 * FATox)	3
% CHOox	(CHOox * 4.07)/EE * 100	4
% FATox	(FATox * 9.75)/EE * 100	5

2.4.4. Capillary Blood Sampling

Blood samples were taken during the last 30 s of each workrate in the step test for the assessment of [La⁻]_b. Measurements of [La⁻]_b were performed by wiping a finger with an alcohol swab, followed by a finger‐prick, and collection of a 20 μL blood sample with a capillary tube, which was mixed in a EKF prefilled safe lock plastic tube containing a heparinized solution for analysis using a laboratory device (Biosen C‐Line Clinic, EKF Industrie, Elektronik GmbH, Barleben, Germany). Appropriate manufacturer recommended calibration was performed on each test session.

2.5. Statistical Analysis

Initially, Q‐Q plots and Shapiro‐Wilks tests were used to check the normality of all variables. Paired t‐tests and Cohens' d were used to examine the effect of HYPO versus NORM at MFO. Two separate linear mixed model (LMM) structures were used to analyze the fixed effect of condition (NORM vs HYPO) whereby the intensity was expressed either as a fixed effect (i.e., relative intensity as %RCP: REL model) or a covariate (i.e., absolute intensity in watts; ABS model). Individual participant intercept was added as a random effect initially and then Akaike's information criterion (AIC) was assessed with or without addition of individual slope as a random factor. The model with the lowest AIC was used unless there was model convergence (or near convergence), and in these cases individual slope was removed as a random factor. The restricted maximum likelihood method was used for all LMM analyses and performed using Jamovi version 2.6.23 (GAMLj3 module). Statistically significant differences were defined as p < 0.05. All data are presented as mean ± standard deviations. Figures and data visualizations were prepared in GraphPad PRISM version 10.4.1.

3. Results

3.1. Participants

For males the mean (SD) age was 35.3 (7.8) yr, height [173.4 (10.7) cm], bodyweight [79.6 (12.2) kg], $\dot{\mathrm{V}}$O_2max [42.3 (6.3) mL·kg⁻¹·min⁻¹] and %bodyfat [20.3 (4.4) %]. For females the mean (SD) age was 35.4 (7.0) yr, height [165.2 (6.8) cm], bodyweight [64.9 (7.4) kg], $\dot{\mathrm{V}}$O_2max [37.6 (4.6) mL·kg⁻¹·min⁻¹] and %bodyfat [21.8 (5.2) %].

3.2. Ramp Incremental Test

3.2.1. GET

Although the percent of $\dot{\mathrm{V}}$O_2max at which the GET occurred was similar in NORM [59.1% (5.8)] and HYPO [57.4 (3.6); p = 0.06, d = 0.5], the absolute (L·min⁻¹) and relative (mL·kg⁻¹·min⁻¹) $\dot{\mathrm{V}}$O₂ at the GET were significantly lower in HYPO [1.46 L·min⁻¹ (0.20); 19.8 mL·kg⁻¹·min⁻¹ (2.7)] compared to NORM [1.81 L·min⁻¹ (0.39); 24.2 mL·kg⁻¹·min⁻¹ (4.7); p = 0.04, d = 1.01]. Whereas RER and HR in beats per min (bpm) were significantly higher in HYPO [0.96 (0.08); 135 bpm (13)] compared to NORM [0.86 (0.07); 130 (17) bpm; p < 0.05], $\dot{\mathrm{V}}$ _E showed no significant differences between conditions [40.0 L·min⁻¹ (7.0) at HYPO; versus 40.1 L·min⁻¹ (7.9) in NORM; p = 0.94, d = 0.01].

3.2.2. RCP

Whilst the percent of $\dot{\mathrm{V}}$O_2max at which the RCP occurred was similar in NORM [80% (5.5)] and HYPO [79.5% (5.7); p = 0.77, d = 0.11], the PO, absolute (L·min⁻¹) and relative (mL·kg⁻¹·min⁻¹) $\dot{\mathrm{V}}$O₂ at the RCP were lower in HYPO [128 (26) W; 2.03 L·min⁻¹ (0.26); 27.6 mL·kg⁻¹·min⁻¹ (4.4)] than in NORM [153 (43) W; 2.44 L·min⁻¹ (0.51); 32.7 mL·kg⁻¹·min⁻¹ (5.6); p < 0.001]. No significant differences were found for the RER, HR, and $\dot{\mathrm{V}}$ _E between conditions (1.05 (0.12); 158 bpm (14) 66.7 L· min⁻¹ (17.2) in NORM, versus 1.08 (0.08); 156 bpm (13), 64.2 L· min⁻¹ (9.9) in HYPO; p > 0.05).

3.2.3. $\dot{\mathrm{V}}$O_2max

The PPO, $\dot{\mathrm{V}}$O_2max, $\dot{\mathrm{V}}$ _E, and HR were significantly lower in HYPO [217 (49) W; 34.2 mL·kg⁻¹·min⁻¹ (6.5); 105.9 L·min⁻¹ (29.7); 171 bpm (15)] versus NORM [255 W (50); 40.5 mL·kg⁻¹·min⁻¹ (6); 114.4 L·min⁻¹ (23.1); 176 bpm (8); p < 0.05]. No significant differences were found between conditions for the rest of the measurements (p > 0.05).

3.3. Steady‐State Incremental Exercise

3.3.1. MFO

Figure 2 displays MFO data. A 22% decrease in MFO was observed in HYPO (0.26 ± 0.07 g.min⁻¹) compared to NORM (0.34 ± 0.07 g.min⁻¹; p < 0.001; d = 1.16, Figure 2, Panel A). No significant effect of hypoxia was found for CHOox at the intensity at which MFO occurred (Table 2). On average, the percent of $\dot{\mathrm{V}}$O_2max in the stage at which MFO occurred was similar during exercise in NORM (38% ± 8%) and HYPO (38% ± 8; p = 0.89; d = 0.04; Figure 2, Panel B).

FIGURE 2.

Absolute maximum fat oxidation rate (MFO; panel A) during exercising in normoxia (NORM) and hypoxia (HYPO). %VO_2max where MFO occurred (panel B).

TABLE 2.

Steady‐state physiological responses at MFO (maximal fat oxidation rate). FATox (Fat oxidation), CHOox (carbohydrate oxidation), EE (energy expenditure), %FATox (FATox as a percent of EE), %CHOox (CHOox as a percent of EE), RER (respiratory exchange ratio), PO (power output) in watts (W), [La⁻]_b (blood lactate concentration), $\dot{\mathrm{V}}$O₂ (volume of oxygen consumption). Data are presented as mean (standard deviation).

Variables at MFO	NORM	HYPO	p‐value	Cohen's d
PO (W)	41 (20)	33 (11)	0.14	0.41
% $\dot{\mathrm{V}}$O2max	38 (8)	38 (8)	0.89	0.04
%RCP	49 (12)	50 (13)	0.76	0.08
MFO (g·min−1)	0.34 (0.07)	0.26 (0.07)	< 0.001	1.16
CHOox (g·min−1)	0.60 (0.26)	0.61 (0.25)	0.88	0.04
EE (kcal·min−1)	5.76 (1.18)	5.17(1.16)	0.15	0.41
% CHOox	40.5 (12.8)	50.6 (10.4)	0.01	0.77
% FATox	58.9 (12.5)	49.4 (10.4)	0.01	0.77
RER	0.82 (0.04)	0.85 (0.03)	0.02	0.74
[La−]b (mmol·L−1)	1.42 (0.43)	1.58 (0.58)	0.34	0.25

Note: Nb: p‐values denote NORM versus HYPO assessed via paired t‐test. Bold text highlights all significant differences where p < 0.05.

3.3.2. Physiological Responses at MFO

Table 2 displays variables related to the time at which MFO occurred. Although the EE was marginally lower in HYPO compared to NORM (p = 0.08) at MFO, a significant increase in the relative contribution of CHOox existed in HYPO compared to NORM (p < 0.05). No significant differences were found between the percentage of RCP at which MFO occurred in NORM (49% ± 12; p = 0.76; d = 0.08) versus HYPO (50 ± 13).

3.3.3. Relative versus Absolute Intensity

Figure 3 displays the profiles of different components related to metabolism and intensity in HYPO and NORM against the relative and absolute PO. FATox was significantly lower for all the stages in HYPO versus NORM (p < 0.05) whether the curve was expressed against the absolute intensity (PO) or the relative intensity (as a percent of RCP; p < 0.05). On contrary, there was a significant increase in the HR and [La⁻]_b in HYPO compared to NORM (p > 0.05) only when the data were expressed against the absolute intensity but not when the data were presented in relative intensity terms (p > 0.05).

FIGURE 3.

Oxygen consumption ($\dot{\mathrm{V}}$O₂), heart rate (HR), fat oxidation rate (FATox), and blood lactate concentration ([La⁻]_b) expressed as a function of absolute workrate (panels, A, B, C, D) or matched relative intensity (panels E, F, G, H) during step incremental exercise in normoxia (NORM) and hypoxia (HYPO; FiO₂ ≈ 13.5%).

Table 3 displays the profiles of different components related to intensity, substrate utilization, gas exchange threshold, metabolic rate in HYPO and NORM during the moderate and heavy intensity stages of the step test. There were no significant differences in the percent of RCP at which each stage occurred for NORM and HYPO (p > 0.05), despite a significant increase in the $\dot{\mathrm{V}}$O₂ at each stage (Table 3, p < 0.05). Furthermore, there was a linear increase in the metabolic rate (i.e., as percent $\dot{\mathrm{V}}$O_2max) as a function of the increase in PO in NORM and HYPO.

TABLE 3.

ABS (absolute intensity mixed model analysis) and REL (relative intensity mixed model analysis) for the FATox (fat oxidation), CHOox (carbohydrate oxidation), EE (energy expenditure), RER (respiratory exchange ratio), [La⁻]_b (blood lactate concentration), HR (heart rate), $\dot{\mathrm{V}}$ _E (minute ventilation), $\dot{\mathrm{V}}$O₂ (volume of oxygen consumption), and percent $\dot{\mathrm{V}}$O_2max. Data are presented as mean (standard deviation).

	Stage						ABS model p‐values
NORM	1	2	3	4	5	6	Condition
PO (W)	20.3 (4.8)	40.7 (9.5)	61.1 (14.4)	81.6 (19.3)	101.9 (24.1)	122.3 (28.9)
VO2 (L·min−1)	0.93 (0.14)	1.15 (0.17)	1.41 (0.21)	1.64 (0.22)	1.87 (0.26)	2.12 (0.32)	0.19
%VO2max	31.9 (6.7)	39.3 (7.2)	48.0 (6.4)	56.1 (9.9)	63.9 (10.1)	72.4 (11.3)	0.15
HR (bpm)	90.5 (15.7)	98.9 (14.2)	108 (13.7)	119.2 (13.5)	132.2 (13.8)	148.9 (12)	< 0.001
V E (L·min−1)	21.5 (3.3)	26.6 (4.1)	32.3 (5)	38.5 (4.9)	46.5 (7.4)	55.2 (9.1)	0.02
EE (kcal·min−1)	4.54 (0.69)	5.72 (0.91)	6.99 (1)	8.08 (1.13)	9.22 (1.32)	10.38 (1.72)	0.46
FATox (g·min−1)	0.28 (0.08)	0.32 (0.08)	0.31 (0.09)	0.25 (0.09)	0.19 (0.08)	0.11 (0.07)	< 0.001
CHOox (g·min−1)	0.45 (0.18)	0.65 (0.25)	0.98 (0.34)	1.37 (0.33)	1.80 (0.34)	2.28 (0.41)	0.008
%FATox	60.0 (15.5)	54.6 (14.3)	44.1 (14.8)	31.2 (12)	20.7 (8.9)	10.7 (6.5)	< 0.001
%CHOox	40.0 (15.5)	45.4 (14.3)	55.8 (14.8)	68.8 (12)	79.3 (8.9)	89.3 (6.5)	< 0.001
RER	0.82 (0.05)	0.84 (0.04)	0.87 (0.04)	0.9 (0.04)	0.93 (0.03)	0.96 (0.04)	< 0.001
[La−]b (mmol·L−1)	1.23 (0.37)	1.35 (0.41)	1.51 (0.44)	1.86 (0.59)	2.72 (0.65)	3.91 (1.0)	0.002
REL model p‐values
HYPO	1	2	3	4	5	6	Condition	Interaction
PO (W)	17.0 (3.5)	34.4 (6.9)	51.3 (10.3)	68.3 (13.7)	85.5 (17.1)	102.4 (20.5)
VO2 (L·min−1)	0.84 (0.19)	1.04 (0.22)	1.24 (0.23)**	1.43 (0.25)**	1.61 (0.32)***	1.89 (0.36)***	0.004	< 0.001
%VO2max	33.3 (9.8)	40.9 (10)	48.9 (10.5)	56.3 (10.3)	63.1 (10.9)	73.7 (11.9)	0.80	0.64
HR (bpm)	93.5 (13.3)	101.5 (13.3)	111.3 (15.0)	123.5 (14.4)	136.4 (15.0)	152 (13.7)	0.10	0.77
V E (L·min−1)	20.8 (3.9)	25 (5.2)	31.6 (5.6)	37.6 (6.5)	45 (8.9)	56.5 (11.1)	0.45	0.18
EE (kcal·min−1)	4.17 (0.8)	5.18 (0.95)*	6.3 (1.1)*	7.26 (1.24)**	8.29 (1.37)**	9.57 (1.83)**	0.005	0.08
FATox (g·min−1)	0.2 (0.06)***	0.25 (0.06)***	0.21 (0.06)***	0.16 (0.06)***	0.12 (0.05)***	0.05 (0.03)**	< 0.001	0.03
CHOox (g·min−1)	0.54 (0.17)	0.68 (0.21)	1.05 (0.27)	1.4 (0.32)	1.75 (0.37)	2.23 (0.5)	0.74	0.06
%FATox	47.8 (11)***	46.8 (11.8)*	32.8 (8.7)***	21.7 (8.0)**	14.5 (6.5)*	5.9 (4.1)*	0.003	0.004
%CHOox	52.2 (11)***	53.2 (11.8)*	67.1 (8.7)***	78.3 (8.0)**	85.4 (6.5)*	94.1 (4.1)	0.003	0.005
RER	0.85 (0.03)**	0.87 (0.03)**	0.90 (0.03)**	0.93 (0.03)**	0.95 (0.03)	0.98 (0.05)	0.007	0.005
[La−]b (mmol·L−1)	1.36 (0.38)	1.46 (0.42)	1.63 (0.5)	2.09 (0.49)	2.93 (0.7)	4.1 (1.17)	0.26	0.89

*Denotes mixed model simple effects NORM versus HYPO at stage shown. *p < 0.05, **p < 0.01, ***p < 0.001. Bold text highlights all significant differences where p < 0.05.

3.3.4. Substrate Utilization

FATox was significantly lower in HYPO versus NORM in all stages (Table 3, p < 0.001), which was accompanied by a significant increase in RER (p < 0.05) and a significant decrease in EE (p = 0.005), without any condition effect on [La⁻]_b or CHOox in g·min⁻¹ at the same relative intensity in HYPO versus NORM. Moreover, CHOox and EE significantly increased following each step of exercise in both conditions (p < 0.05).

4. Discussion

This is the first study to investigate the effects of acute hypoxia on MFO using the exercise intensity domain model based on the RCP as opposed to $\dot{\mathrm{V}}$O_2max. Furthermore, this study is the first to present a statistical analysis expressing substrate utilization as a function of both absolute and relative intensity during incremental exercise. The main findings of this study were that: (i) acute hypoxia significantly reduced MFO during incremental steady state exercise; (ii) MFO occurred at the same %RCP and % $\dot{\mathrm{V}}$O_2max in both NORM and HYPO; (iii) CHOox rate was similar at all work rates when normalized for relative intensity in both conditions, whereas FATox was reduced in HYPO coincident with lower EE despite normalization of the relative intensity. When steady state $\dot{\mathrm{V}}$O_2, HR and [La⁻]_b were expressed as a percentage of either RCP or $\dot{\mathrm{V}}$O_2max in each condition, there was no condition or interaction effects, thus confirming the normalization procedure was successful and led to equivalent relative intensity in both NORM and HYPO. Thus, since relative intensity in hypoxia is a function of both the absolute exercise intensity and magnitude of the hypoxic stimulus, the observed differences in substrate utilization between conditions are likely explained by mechanisms that hypoxia impacts independently of the synergistic effect of exercise and hypoxia together. These results indicate that homeostasis of energy supply at equivalent relative intensity (and therefore reduced absolute EE) in hypoxia is largely accomplished by maintenance of CHOox at the expense of reduced FATox.

4.1. Physiological Responses

It is generally accepted that exercise under hypoxic conditions, whether terrestrial or simulated, and either acute or chronic, is associated with alterations in substrate utilization (Griffiths et al. 2019a, 2019b). However, due to a range of differences in experimental protocols, the exact nature of changes in substrate utilization remains equivocal. Previous studies typically examined substrate utilization in hypoxia at a single workrate matched for either absolute or relative intensity. In contrast, we examined the synergistic effect of exercise intensity across a range of workloads, and hypoxia using two different mixed model analyses. The advantage of this approach is that both analyses can be conducted from a one normoxic and hypoxic trial each. When expressed as a function of absolute intensity, as expected both $\dot{\mathrm{V}}$O₂ and EE across all stages during step testing were similar between conditions (Figure 3; Table 3). This observation was reported long ago by (Saltin et al. 1968) and indicates that steady‐state exercise performed at a given mechanical workrate requires an equivalent energy transfer (i.e., ATP hydrolysis to force generation) irrespective of upstream differences in oxygen transport. Thus, to maintain appropriate O₂ delivery to working musculature, both HR and V _E are elevated, which are also well‐known responses to exercise in acute hypoxia and confirmed in the present study (Figure, 3). When normalized for relative intensity in HYPO the reduction in mechanical workrate resulted in a decrease in absolute $\dot{\mathrm{V}}$O₂ and EE (Table 3), and in this case, HR and V _E showed no difference between conditions. Similarly, during the ramp incremental test we observed little difference in HR or V _E at GET or RCP, but the absolute PO and $\dot{\mathrm{V}}$O₂ were reduced. These results suggest that the whole‐body response to moderate and heavy domain exercise is regulated toward maintenance of O₂ utilization that is the aerobic energy demand of the task is prioritized, and the integrated cardiorespiratory response is adjusted accordingly.

4.2. Substrate Utilization Normalized for Absolute Intensity

In the present study, when expressed as a function of absolute workrate, we observed a decrease in FATox, and an increase in CHOox, RER and [La⁻]_b in HYPO compared to NORM. Previous studies have also reported a significant increase in CHOox and [La⁻]_b in hypoxia (equivalent to 4100–4300 m altitude) when matched for external workload (Lundby and Van Hall 2002; Péronnet et al. 2006), and consequently a greater percentage contribution from CHO to total EE with a lower percentage from FATox. Similar results were found at 3000 m simulated altitude matched for absolute intensity, and furthermore, absolute FATox in g·min⁻¹ was significantly decreased (Sumi et al. 2020). In all cases, artificial altitude was used and consequently the hypoxic “wash in” duration prior to exercise was in the range 10–20 min which is similar to the present investigation. In contrast, another study reported no differences in either CHOox or FATox at 4300 m altitude matched for absolute workrate (Young et al. 2018), however the participants were transported to natural altitude over a period of 5 h, and the substrate utilization measurements were reported over the final 40 min of an 80 min submaximal exercise bout. This is consistent with an earlier study also at 4300 m natural altitude which shows that blood free fatty acid (FFA) concentration was higher after arrival at altitude, however leg uptake of FFAs during exercise was reduced (Roberts et al. 1996). Thus, when comparing studies which examine substrate utilization, the duration of hypoxia exposure prior to the exercise bout may affect fuel partitioning, possibly due to the time course of adrenergic activation of both lipolysis and hepatic glycogenolysis which elevates blood FFA concentrations over a period of several hours as opposed to minutes (Rostrup 1998). In studies utilizing a hypoxic chamber or gas delivery system, as per the current investigation, the results appear more consistent and reveal an increase in absolute CHOox combined with a decrease in FATox during submaximal steady state exercise. In the present study, when the exercise intensity was normalized to a percentage of RCP, the decrease in FATox and increase in RER persisted, however in this case, both $\dot{\mathrm{V}}$O₂ and EE were reduced consequent to the lower mechanical workload. Whereas CHOox and [La⁻]_b remained similar in both NORM and HYPO.

4.3. Substrate Utilization Normalized for Relative Intensity

In contrast to studies that investigated the effects of hypoxia on metabolism during cycling at the same absolute intensity, numerous studies have instead matched the relative exercise intensity. Consistent with our results, a significant decrease in FATox was observed in hypoxia (4300 m simulated altitude) compared to normoxia in participants cycling for 80 min at 77% $\dot{\mathrm{V}}$O_2max (Péronnet et al. 2006). In another study a higher RER was found in hypoxia (2500 m simulated altitude) during 60 min of running at ∼82% $\dot{\mathrm{V}}$O_2max (Friedmann et al. 2004). Likewise, Katayama et al. (2010) reported higher RER in hypoxia (FiO₂ ∼ 12%) during 30 min cycling at 50% $\dot{\mathrm{V}}$O_2peak. The pre‐exercise control conditions that is duration of fasting prior to exercise, were quite similar also in this study compared to our protocol. In contrast, several studies have reported conflicting findings. For example, there were no significant changes in RER during 60 min cycling at 60% $\dot{\mathrm{V}}$O_2max at 3000 m (Bouissou et al. 1987). Similarly, during 60 min cycling at 50% VO_2max at 4100 m simulated altitude, RER, FATox and CHOox were not significantly different (Lundby and Van Hall 2002). In both of these studies the substrate utilization measurements were conducted in the latter stages of the exercise bout, which contracts our protocol in which we measured FATox and CHOox at every step, however this would not explain the discrepancies between these studies and the ones by (Péronnet et al. 2006; Friedmann et al. 2004). Lastly, O’Hara et al. (2017), in contrast to our results, reported reduced CHOox and increased FATox at 3375 m altitude. However, this study involved ascent to terrestrial altitude in Torino, Italy, which as discussed previously may have involved a period of several hours hypoxic exposure prior to initial measurements.

4.4. Maximal Fat Oxidation

A key novel aspect of the present study is that we examined substrate utilization during incremental exercise, and therefore across a range of workloads. Hence this experimental protocol allows for the determination of MFO, whereas previous studies typically only examined a single workload. The results reveal a 22% reduction in MFO in HYPO compared to NORM (Figure 2, Table 2) along with a non‐significant trend for the absolute EE at which MFO occurred to decrease (p = 0.08). Given that we observed a main effect of hypoxia on FATox when matched for relative intensity, this result is not surprising. Moreover, we found an interaction effect whereby a trend toward a smaller effect of hypoxia was observed as relative exercise intensity increases in the heavy domain versus moderate domain. This observation is relevant because MFO occurred at ≈38% $\dot{\mathrm{V}}$O_2max in both NORM and HYPO, whilst the GET occurred at ≈59% and 57% $\dot{\mathrm{V}}$O_2max in NORM and HYPO respectively. The intensity where MFO occurred in our study was thus markedly lower than the relative intensity used in all studies mentioned previously (Bouissou et al. 1987; Friedmann et al. 2004; Lundby and Van Hall 2002; Péronnet et al. 2006). It has been suggested that assessment of metabolic flexibility should involve measurement of the MFO (Chávez‐Guevara 2023), hence by examining FATox across a range of intensities this may offer a fuller picture regarding the acute effect of hypoxia. Our observation that FATox was lower regardless of whether intensity was expressed as a function of absolute or relative workrate contrasts the cardiorespiratory responses. For example, HR, V _E and [La⁻]_b all showed a characteristic “left shift” in hypoxia as a function of absolute intensity, yet this was compensated for by reducing the absolute workrate to normalize relative intensity (Figure 3). In each case these variables continuously increase as function of the external workrate, whereas the FATox curve typically displays an inverted “U” shape (Jeukendrup and Wallis 2005), which was also observed in this study. Thus, a similar left‐shift alone in the FATox curve as a consequence of matching relative intensity would be expected to make little difference to MFO. However, the entire FATox curve was “downshifted” in hypoxia at all exercise intensities (below RCP) regardless of whether the intensity was expressed as absolute or relative (Figure 3). This finding suggests that mechanisms independent of the intensity are likely to influence fuel partitioning during exercise in hypoxia.

4.5. Potential Mechanisms of Reduced Fat Oxidation in Hypoxia

It is well known that muscle contraction intensity influences metabolic regulation and substrate utilization during exercise. Specifically, during incremental exercise intracellular adenosine diphosphate [ADP]/[ATP] and adenosine monophosphate [AMP]/[ATP] ratios gradually increase leading to inhibition of beta‐oxidation and stimulates GLUT4 translocation (Steinberg and Hardie 2023). During exercise in hypoxia matched for absolute intensity, the speed of pulmonary $\dot{\mathrm{V}}$O₂ uptake kinetics decreases, which is compensated for by greater phosphocreatine breakdown (Haseler et al. 2004) thereby exacerbating the normal intensity dependent rise in [ADP]/[ATP] and [AMP]/[ATP] ratios. However, when the relative intensity is matched in hypoxia, there is little difference in pulmonary $\dot{\mathrm{V}}$O₂ uptake kinetics (Engelen et al. 1996) and only minor up‐regulation of AMPK activity (Wadley et al. 2006).

An important factor that distinguishes acute hypoxic exposure from muscle contraction though is activation of hypoxic inducible factor 1 alpha (HIF1α). Under hypoxic conditions, HIF1α is stabilized in its active form and acts somewhat like a “master switch” exerting a range of cellular responses including angiogenesis, apoptosis, regulation of erythropoiesis and glucose metabolism (Semenza 2012). In mouse fibroblasts, HIF1α was shown to strongly upregulate the activity of all glycolytic enzymes (Greijer et al. 2005), whilst later work indicated HIF1α mediated increase in AMPK phosphorylation (Sakagami et al. 2014). One study conducted in humans (D’Hulst et al. 2015) reported results consistent with the above cell/animal models showing an increase in sarcolemmal GLUT4 concentration and reduced blood glucose at rest compared to normoxia. Our data is consistent with these findings since we observed a decrease in FATox even at the lowest workload examined.

Lastly, another key mechanism which may partially explain a decrease in MFO is increased sympathetic activity. Under normoxic conditions, multiple lines of evidence indicate that elevated adrenergic activity contributes to regulation of carbohydrate metabolism such as increased hepatic glycogenolysis, increased glucose uptake in muscle, and activation of muscle glycogenolysis (Ciccarelli et al. 2013). A reduction in arterial oxygen tension stimulates the peripheral chemoreceptors, which are potent activators of the sympathetic nervous system (Dempsey et al. 2014). The time course of this response whereby the carotid chemoreceptors show elevated afferent activity in response to arterial hypoxemia is on the order of minutes (Iturriaga et al. 2021), whilst multiple studies have reported increase muscle sympathetic nerve activity at rest in hypoxia (FiO2 range = 10–14%) within 5–30 min (Rostrup 1998). In the present study, whilst we did not measure autonomic activity directly, we observed elevated HR and $\dot{\mathrm{V}}$E even at the lowest exercise intensity (i.e. baseline cycling at 20 W) which is consistent with the notion that elevated sympathetic activity was present in the current experimental protocol.

4.6. Methodological Considerations

In the present study we examined healthy recreationally active adults. This choice of inclusion criteria was deliberate for multiple reasons. It is acknowledged that MFO is generally greater in more highly trained individuals or those with greater $\dot{\mathrm{V}}$O_2max (Maunder et al. 2018), and moreover, ultra‐endurance exercise performance is positively correlated with MFO in trained male triathletes (Frandsen et al. 2017). However there remains a paucity of data on the relationship between MFO and performance in highly‐trained and elite endurance athletes, or to high‐intensity endurance performance over shorter durations for example 30–120 min or during events that are intermittent in nature such as road cycling or team sports. On the other hand, there is interest dating back decades, regarding the mechanisms of “metabolic flexibility” and substrate utilization from the perspective of weight management (Chávez‐Guevara et al. 2020). Thus, we chose a subject cohort that would be expected to lie closer to population mean values to be applicable to a wider cross section, as opposed to a specific clinical population or highly trained athletes alone. Consequently, the results have relevance to the emerging use of simulated altitude (normobaric hypoxia) for a broad range of applications including interventions related to glucose homeostasis (De Groote and Deldicque 2021), body weight management (Kayser and Verges 2013), injury rehabilitation (Yeung et al. 2025) and healthy aging (He et al. 2023).

5. Conclusions

This was the first study to examine the effect of hypoxia on MFO rate during incremental cycling exercise. The results reveal that FATox and MFO were decreased in HYPO regardless of whether the exercise intensity was expressed in absolute or relative terms. When relatively intensity was normalized between conditions, leading to lower EE in HYPO, the CHOox remained similar whereas FATox was reduced. Furthermore, cardiorespiratory responses were not significantly different in HYPO when normalizing for relative intensity. These results suggest that mechanisms related directly to arterial hypoxemia per se, may impact substrate utilization independently of changes in relative intensity during exercise in acute hypoxia.

Author Contributions

YH, DI and NT conceived the study design. YH, DI, and NT performed data collection and analysis. YH and NT wrote and edited initial drafts of the article. All authors revised and approved the final version.

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Hassanein, Youmna E. , Ibrahim Dania, Murias Juan M., and Townsend Nathan. 2026. “Acute Hypoxia Decreases Maximum Fat Oxidation Rate During Step Incremental Exercise Normalized to Respiratory Compensation Point,” European Journal of Sport Science: e70086. 10.1002/ejsc.70086.