In response to moderate-intensity step-transition exercise, oxygen uptake ( V ˙ O₂) increases in a mono-exponential manner (Margaria et al., 1965; Margaria et al., 1963; Poole and Jones, 2012). The increase in V ˙ O₂ is closely related to the metabolic demand of exercise and is achieved through an integrated response of the pulmonary, cardiovascular, and muscle metabolic systems (Grassi, 2003; Poole et al., 2007; Grassi et al., 2003; Grassi, 2005; Grassi et al., 2021; Rossiter, 2011). The rate at which V ˙ O₂ adjusts at the onset of exercise, commonly referred to as V ˙ O₂ kinetics (τ V ˙ O₂), varies across individuals/populations and serves as an important index of aerobic metabolic function, with implications for metabolic stability (Grassi et al., 2015; Grassi et al., 2011; Grassi et al., 1996) and exercise tolerance (Poole and Jones, 2012; Grassi et al., 2011; Goulding et al., 2021).

Convective and diffusive O₂ delivery to active muscles involves a complex interaction between motor unit recruitment, cardiac output, sympathetic vasoconstriction, and local vasodilation that matches muscle blood flow (BF) to metabolism (Pittman, 2016; Pittman, 2011; Poole et al., 2011; Poole and Musch, 2023; Poole et al., 2022; Marinari et al., 2025; DeLorey and Clifford, 2022; Joyner and Casey, 2015; Zoladz et al., 2016). Previous studies in healthy active individuals have reported that bulk leg BF increases at a rate similar to or faster than τ V ˙ O₂ (Macdonald et al., 1998; Fukuba et al., 2004; Nyberg et al., 2017; duManoir et al., 2010; DeLorey et al., 2007; MacPhee et al., 2005; Koga et al., 2005; Endo et al., 2005; Harper et al., 2006; Jones et al., 2012; Schlup et al., 2015; Paterson et al., 2005; Love et al., 2023), suggesting that bulk O₂ delivery may not be a limiting factor for τ V ˙ O₂. In contrast, in disease states where pulmonary and cardiovascular function may be impaired and muscle O₂ delivery is significantly slowed, τ V ˙ O₂ is also slowed (Poole and Jones, 2012; Poole et al., 2007), suggesting that there may be a critical rate of O₂ delivery required to support an increase in V ˙ O₂ at the onset of exercise. However, measures of bulk leg BF do not provide any information related to the distribution of leg BF and may not reflect BF/O₂ delivery to active muscle(s). Near-infrared spectroscopy (NIRS)-derived [deoxyhemoglobin] ([HHb]) provides an index of changes in muscle O₂ extraction in response to exercise in a discrete region of the active muscle microcirculation (Grassi et al., 2003; DeLorey et al., 2003; Grassi and Quaresima, 2016). Consistent with the Fick equation, changes in [HHb] reflect the balance between changes in microvascular O₂ delivery and those in muscle V ˙ O₂. A faster τ[HHb] (i.e., O₂ extraction) relative to τ V ˙ O₂ may result from either slowed muscle O₂ delivery relative to muscle V ˙ O₂ or faster muscle V ˙ O₂ relative to the rate of muscle O₂ delivery. Previous studies have reported that [HHb] adapts at a faster rate than V ˙ O₂ in healthy individuals (Grassi et al., 2003; DeLorey et al., 2003; Adami et al., 2011), demonstrating a mismatch between microvascular O₂ delivery and muscle V ˙ O₂ at the onset of exercise. The faster τ[HHb] relative to τ V ˙ O₂ observed in these studies could reflect the following: (i) a slowed O₂ delivery relative to the metabolic demand, with O₂ delivery potentially limiting τ V ˙ O₂; (ii) a faster O₂ intracellular utilization relative to O₂ delivery, with the slower O₂ delivery not being rate-limiting for τ V ˙ O₂. Unfortunately, these studies did not include measures of leg BF/O₂ delivery to facilitate resolution of the underlying cause of the faster τ[HHb].

Individuals with high aerobic fitness appear to have faster τ V ˙ O₂ than those with low aerobic fitness (Caputo et al., 2003; Caputo and Denadai, 2004; Cerretelli et al., 1979; George et al., 2018; Grey et al., 2015; Inglis et al., 2021; Koppo et al., 2004a), and an inverse relationship between maximal oxygen uptake ( V ˙ O_2max) and τ V ˙ O₂ has been reported (Inglis et al., 2021; Murias et al., 2011a; Fawkner et al., 2002; Chilibeck et al., 1996; Murgatroyd et al., 2011; Norris and Petersen, 1998; Powers et al., 1985). Although it seems unlikely that V ˙ O_2max and τ V ˙ O₂ have a direct mechanistic basis, the relationship between V ˙ O_2max and τ V ˙ O₂ suggests that aerobic fitness may influence the physiological determinants of τ V ˙ O₂.

Therefore, the purpose of this study was to investigate the following hypotheses: 1) V ˙ O_2max and τ V ˙ O₂ would be inversely correlated; and 2) the limiting factor to τ V ˙ O₂ would be a function of V ˙ O_2max, with O₂ delivery limiting τ V ˙ O₂ in participants with “poor” aerobic fitness, but not in participants with “good” and “superior” aerobic fitness.

All testing was completed in the Integrative Human Exercise Physiology Laboratory at the University of Alberta. Participants reported to the laboratory on three separate occasions. Participants were instructed to abstain from exercise, caffeine, alcohol, and ibuprofen for 24 h prior to testing and to eat a light meal ∼2 h before exercise testing. Laboratory temperature was maintained between 20 °C and 22 °C.

Day 1. Participants completed an incremental exercise test to volitional exhaustion on a cycle ergometer (Ergoselect 200 K, Ergoline, Bitz, Germany) to determine V ˙ O_2max. Following 2 min of baseline data collection, participants began pedaling, and the work rate was progressively increased in a ramp-like fashion at 30 watts (W)·min⁻¹ to volitional exhaustion. Criteria used to establish a maximal test included the observation of a plateau in peak V ˙ O₂, despite an increased work rate, a respiratory exchange ratio (RER) > 1.1, achievement of >90% age-predicted maximal heart rate (HR), and volitional exhaustion. Participants were then assigned to pre-determined aerobic fitness groups described above.

Day 2. Participants completed an incremental alternate-leg knee-extension (KE) exercise test to volitional exhaustion on a custom-built KE ergometer, as previously described (DeLorey et al., 2007). This test was conducted to determine individual work rates for moderate-intensity KE exercise. After 2 min of resting baseline data collection, participants completed 1-min of passive (unloaded) KE exercise, followed by alternate-leg KE exercise at a cadence of 30 contractions per leg per minute (cpm) from an initial work rate of 18 W. The work rate was then increased 3 W·min⁻¹ until volitional exhaustion or until participants were unable to maintain a cadence of 30 cpm. Criteria used to establish a maximal test were a plateau in V ˙ O₂ despite an increase in the work rate, an RER >1.10, and volitional exhaustion. The gas exchange threshold (GET) was defined as the V ˙ O₂ at which CO₂ production ( V ˙ CO₂) began to increase disproportionately with respect to V ˙ O₂, concurrent with an increase in the ratio of minute ventilation ( V ˙ E) to V ˙ O₂ and end-tidal PO₂, while V ˙ E/ V ˙ CO₂ and end-tidal PCO₂ remained stable. Moderate-intensity KE exercise was defined as a work rate corresponding to 80% of the work rate at the GET.

Day 3. Participants completed two step transitions from passive KE exercise to moderate-intensity KE exercise to determine the on-transient τ V ˙ O₂, τ[HHb], and τLVC. Testing began with 2 min of resting baseline data collection, followed by 1-min of passive KE exercise. Thereafter, the work rate was increased in a stepwise manner, and participants performed 5 min of constant-load moderate-intensity KE exercise. Following a 15-min recovery period, the protocol was repeated, and the data from the two exercise transitions were ensemble-averaged (see data analysis). All KE exercise was performed at a cadence of 30 cpm and was preceded by 1-min of passive, unloaded exercise to minimize the effects of mechanical inertia and muscle mechanical factors and to increase the amplitude of the on-transient responses, thereby improving the signal-to-noise ratio and the accuracy of kinetics analysis. Participants’ legs were secured to the lever arms of the KE ergometer to facilitate passive exercise.

For all exercise testing, participants breathed through a mouthpiece, with their nose occluded. A low-resistance mass-flow meter was used to measure pulmonary gas exchange ( V ˙ O₂, V ˙ CO₂, and RER) and V ˙ E breath-by-breath via open-circuit indirect calorimetry (Vmax® 229d; Viasys™ Healthcare, Palm Springs, CA). Prior to each test, the flow meter was calibrated with a 3-L calibration syringe, and the O₂ and CO₂ analyzers were calibrated with gases of known concentrations.

A three-lead electrocardiogram (ECG) was measured continuously (Power Laboratory 16/30, AD Instruments, Colorado Springs, CO), and HR was derived from the ECG. Beat-by-beat arterial blood pressure (BP) was measured using photoplethysmography on the middle finger of the right hand (Finometer™, Finapres Medical Systems, Amsterdam, Netherlands). BP was also measured using a sphygmomanometer, and Finometer BP was corrected to manually measured pressures when pressure differences were observed. Mean arterial pressure (MAP) was calculated on a beat-by-beat basis.

Mean blood velocity (MBV) of the right femoral artery was measured using pulsed-Doppler ultrasonography (Vivid I, General Electric, Waukesha, WI). Data were acquired continuously using a 7.5 MHz probe positioned 2–3 cm distal to the inguinal ligament and proximal to the femoral artery bifurcation, while the probe was maintained at a 45-degree angle of insonation. Prior to exercise testing on days 2 and 3, the resting diameter of the femoral artery was measured in triplicate during diastole. The three measures were then averaged to determine the baseline femoral artery diameter. Previous studies have demonstrated that the common femoral artery diameter does not change from resting values during exercise. Thus the resting diameter was used for blood flow calculations during exercise (Macdonald et al., 1998; MacPhee et al., 2005; Paterson et al., 2005; Rådegran and Saltin, 2000). Mean blood velocity in cm·s⁻¹ was measured on a beat-by-beat basis. Limb BF was calculated as BF (mL·min⁻¹) = MBV·π· r ²·60, where r is the measured radius of the femoral artery. LVC was then calculated as follows: LVC (L·min⁻¹·mmHg⁻¹) = BF·MAP⁻¹. Data were recorded using a PowerLab 16/30 system and Chart 7 data acquisition software (AD Instruments) at a sampling frequency of 100 Hz.

Relative [HHb] was measured in the vastus lateralis (VL) muscle using NIRS (NIRO 300, Hamamatsu Photonics, Hamamatsu, Japan), as described previously (DeLorey et al., 2007). In short, optodes were placed on the belly of the VL at the midpoint between the lateral epicondyle and the greater trochanter of the femur. These optodes were contained within an optically dense plastic holder to minimize extraneous light and the loss of NIR light from the field of interrogation and ensure the position of optodes relative to each other. This optode assembly was affixed to the skin using tape and was wrapped in an elastic bandage to further prevent movement of the optodes and interference of extraneous light.

The intensities of incident and transmitted light were continuously recorded, along with relevant extinction coefficients and estimated optical path length, assuming a differential path length factor of 3.83 (DeLorey et al., 2007). These values were used for online estimation and display of changes in concentrations of oxyhemoglobin (O₂Hb), HHb, and total hemoglobin (Hb_tot). The raw attenuation signal in optical density units was sampled at 1Hz and transferred to a computer and stored for future analysis. Prior to testing, the NIRS unit was “zeroed” to a stable, resting baseline.

The purpose of the present study was to investigate the relationship between τ V ˙ O₂ and V ˙ O_2max, and determine whether the capacity to deliver or to utilize O₂ limits τ V ˙ O₂ in an aerobic fitness-dependent manner. V ˙ O_2max was inversely correlated with τ V ˙ O₂, and τ V ˙ O₂ was faster in the S and G fitness groups than in the P fitness group. τLVC was not different between groups, and it was not positively correlated with V ˙ O_2max or τ V ˙ O₂ across and within groups, indicating that O₂ delivery kinetics were similar between groups despite large differences in aerobic fitness. τ[HHb] was faster in the S than in the P fitness group and was inversely correlated with V ˙ O_2max and positively correlated with τ V ˙ O₂ across groups. Collectively, these data indicate that O₂ delivery was not limiting for τ V ˙ O₂ and that intracellular oxidative metabolism may limit τ V ˙ O₂ in healthy individuals, regardless of aerobic fitness.

In the present study, τ V ˙ O₂ was slower in the P than in the G and S fitness groups, whereas τ V ˙ O₂ was not different between the S and G fitness groups. Consistent with the present data, several cross-sectional studies have also reported faster τ V ˙ O₂ in trained or active individuals than in untrained or sedentary individuals (Caputo et al., 2003; Caputo and Denadai, 2004; Cerretelli et al., 1979; George et al., 2018; Grey et al., 2015; Inglis et al., 2021; Koppo et al., 2004a), and although the data are not conclusive, an inverse relationship between V ˙ O_2max and τ V ˙ O₂ has often been reported (Inglis et al., 2021; Murias et al., 2011a; Fawkner et al., 2002; Chilibeck et al., 1996; Murgatroyd et al., 2011; Norris and Petersen, 1998; Powers et al., 1985). Inglis et al. (2021) investigated the relationship between V ˙ O_2max and τ V ˙ O₂ in untrained ( V ˙ O_2max ∼ 40 mL·kg⁻¹·min⁻¹) and trained ( V ˙ O_2max ∼ 58 mL·kg⁻¹·min⁻¹) participants and reported an inverse relationship between V ˙ O_2max and τ V ˙ O₂ within the untrained group, but not within the trained group. The kinetics of cardiac output (Q; measured through impedance cardiography) were not different between untrained and trained groups and were similar to or even faster than τ V ˙ O₂ in both groups, suggesting that O₂ delivery was not limiting in either group. Although Q kinetics do not provide information related to the distribution of Q or muscle blood flow, Inglis et al. (2021) reported a higher [HHb]/ V ˙ O₂ ratio in the vastus lateralis muscle of the untrained group than that of the trained group, suggesting that microvascular O₂ delivery may be slower within active muscle and contribute to a slower V ˙ O₂ response in untrained participants. Whether τ V ˙ O₂ is limited by O₂ delivery (DeLorey et al., 2003; Hughson et al., 1996; Murias et al., 2011b; Murias et al., 2011c), intracellular oxidative metabolism (Zoladz et al., 2016; Grassi, 2000; Christensen et al., 2011; Korzeniewski and Rossiter, 2015), or a combination of both (Murias et al., 2014) within the active limbs across fitness levels remains controversial.

This study, in line with others (Macdonald et al., 1998; Fukuba et al., 2004; Nyberg et al., 2017; duManoir et al., 2010; DeLorey et al., 2007; MacPhee et al., 2005; Koga et al., 2005; Endo et al., 2005; Harper et al., 2006; Jones et al., 2012; Schlup et al., 2015; Paterson et al., 2005; Love et al., 2023; Inglis et al., 2021), indicated that O₂ delivery (i.e., τLVC) was faster or as fast as τ V ˙ O₂ within each fitness group. Furthermore, τLVC and LVC_GAIN were not different between groups despite large differences in aerobic fitness, and τLVC was not correlated with either τ V ˙ O₂ or V ˙ O_2max, suggesting that O₂ delivery is not a rate-limiting factor for τ V ˙ O₂ across fitness levels. Consistent with this notion, pump perfusion of canine muscle to eliminate temporal delays in O₂ delivery and enhancement of muscle O₂ diffusive capacity did not accelerate τ V ˙ O₂ during moderate-intensity exercise (Grassi et al., 1998a; Grassi et al., 1998b).

Although the present study suggests that at the onset of exercise, O₂ delivery to the active muscles is not limiting for τ V ˙ O₂ across fitness levels (evidenced by similar τLVC and LVC_GAIN in all groups), it is important to acknowledge that LVC reflects bulk O₂ delivery to the whole limb and does not reflect microvascular O₂ delivery to active muscle fibers. Therefore, it could be argued that either better microvascular O₂ distribution or intracellular oxidative mechanisms underlie differences in τ V ˙ O₂ between fitness levels. If microvascular O₂ delivery limits τ V ˙ O₂ across fitness levels, an inverse relationship between τ[HHb] and τ V ˙ O₂ should be observed, and faster τ[HHb] would be expected in the P fitness group (i.e., the group with the slowest τ V ˙ O₂) than in the G and S fitness groups. However, τ[HHb] was significantly faster in the S fitness group than in the P fitness group and positively correlated with τ V ˙ O₂ and negatively correlated with V ˙ O_2max across fitness groups. Potential contributions of microvascular O₂ delivery to τ V ˙ O₂ in our P fitness group, where τ[HHb] was faster than τ V ˙ O₂, cannot be dismissed. The direction of the relationships between τ[HHb] with τ V ˙ O₂ and V ˙ O_2max, as well as the faster τ[HHb] in the S than in the P fitness group, and, more importantly, the similar τLVC and LVC_GAIN between groups support the notion that intracellular oxidative metabolism is the rate-limiting factor for τ V ˙ O₂ across fitness levels.

Supporting enhanced intracellular oxidative metabolism in the S fitness group is the observed initial [HHb] overshoot. With τLVC not being limiting for τ V ˙ O₂, the [HHb] overshoot may be attributed to a transient enhanced oxidative metabolism response and/or to specific motor unit recruitment strategies in highly trained individuals at the onset of exercise (Grassi et al., 2011; Marinari et al., 2025; Korzeniewski and Zoladz, 2003; Bowen et al., 2013; do Nascimento Salvador et al., 2023). Accordingly, it has recently been demonstrated that muscle excitation increases “disproportionally” at the onset of a step-transition exercise without prior warm-up in recreationally active individuals, which was connected to a greater [HHb] response (Marinari et al., 2025). A similar interpretation was proposed to explain the overshoot in V ˙ O₂ observed in trained cyclists at intensities below the GET (Kilding and Jones, 2008; Koppo et al., 2004b).

Interestingly, end-exercise gains in V ˙ O₂, LVC, and [HHb] were not different between fitness groups (Table 3), indicating similar adaptations at different levels of the O₂ cascade system per unit work rate.

In this study, V ˙ O_2max and τ V ˙ O₂ were inversely correlated across three fitness levels. Although τ V ˙ O₂ was faster in the “good” and “superior” fitness groups than in the “poor” fitness group, τLVC was similar between fitness groups and not correlated to either τ V ˙ O₂ or V ˙ O_2max. Conversely, τ[HHb] was inversely correlated to V ˙ O_2max and positively correlated to τ V ˙ O₂ across fitness groups. Furthermore, τ[HHb] was faster in the “superior” fitness group than in the “poor” fitness group. Collectively, the present study suggests that O₂ delivery is not a rate-limiting factor for τ V ˙ O₂ and that the intracellular capacity to utilize O₂ may be the primary limiting factor for τ V ˙ O₂ in healthy individuals, regardless of aerobic fitness.