Skeletal muscle blood flow increases in response to exercise to match metabolic demand until maximal cardiac output is reached (Richardson et al., 1993). When exercise is performed using a large muscle mass and at high intensity, some degree of peripheral vasoconstriction is observed and thus vascular conductance is reduced to maintain blood pressure (Calbet et al., 2004; Volianitis and Secher, 2002). As a result, cardiac output is thought to be a significant liming factor to maximal exercise involving a large muscle mass (e.g., double-leg cycling, rowing, cross-country skiing) (Saltin and Calbet, 2006). In contrast, greater levels of peripheral vasodilation are observed when a smaller amount of muscle mass (e.g., arm cranking, knee-extension, single-leg cycling) is exercised (Volianitis and Secher, 2002) since central blood pressure is not compromised. This allows blood flow to increase linearly with exercise intensity (Richardson et al., 1993) and may potentially explain the greater exercise performance observed during small compared with large muscle mass exercise. Indeed, it has been shown that power output per leg is generally 5 to 10% greater during single-leg (SL) compared with double-leg (DL) cycling (Abbiss et al., 2011; Gordon et al., 2018; Iannetta et al., 2019). Yet, blood flow during maximal SL and DL cycling has only been assessed in a small clinical population (LeJemtel et al., 1986) or during submaximal exercise (Burns et al., 2014; Draper et al., 2022). Therefore, the improved capacity to increase blood flow and enhance oxygen delivery during a maximal cycling exercise performed using SL compared to DL remains to be demonstrated.

In addition to being affected by active muscle mass, the hyperaemic response to exercise may also be influenced by age. An attenuation of the blood flow response in active muscles is sometimes observed in older compared with younger individuals (Beere et al., 1999; Donato et al., 2006; Poole et al., 2003; Proctor et al., 2004). This may be due to a reduction in vasodilatory capacity (e.g., reduced vascular conductance), an increase in sympathetically mediated vasoconstriction (e.g., increased diaphragm metaboreceptor activity) and/or an impairment of “functional sympatholysis” (e.g., reduced ability to blunt sympathetic vasoconstriction in the locomotor muscles) with ageing (Hearon et al., 2016). Yet, blood flow values to the active muscles were similar in older (Magnusson et al., 1994) and younger (Andersen et al., 1985) individuals during small muscle mass exercise (i.e., single-leg knee extension). It is possible that the age-related decrease in local blood flow is less pronounced or prevented during such exercise modality given a greater cardiac output is available during such exercise compared with a maximal effort involving large muscle mass (Davies and Sargeant, 1974; Iannetta et al., 2019; Klausen et al., 1982). However, a lower blood flow during small muscle mass exercise was also reported in older compared with younger individuals (Donato et al., 2006; Mortensen et al., 2012). The discrepancies between these studies may partly be explained by the differences in fitness status of the participants since no age-related difference in leg blood flow were observed between younger and older individuals engaged in short- (i.e., 3 months, Beere et al. (1999)) or long-term (i.e., 30 years, Mortensen et al., 2012)) endurance training. Nonetheless, a lower blood flow was also reported in older trained individuals compared with younger trained individuals (Proctor et al., 1998). Further studies are therefore needed to clarify the effects of ageing on the hyperaemic response to exercise in older-trained individuals, and better understand the age-related changes in vascular function without the confounding effect of sedentariness.

The purpose of this study was to investigate peripheral blood flow in response to maximal exercise involving a small and a large muscle mass. A secondary aim was to assess the effect of ageing on the hyperaemic response to exercise. Therefore, we compared femoral blood flow in response to double- and single-leg cycling in younger (<40 years) and older (>60 years) endurance-trained individuals. We hypothesised that 1) femoral blood flow would be greater in response to maximal self-paced single-leg when compared with double-leg cycling but 2) attenuated in older compared to younger endurance-trained individuals.

This study was conducted in a subgroup of individuals recruited for a larger study previously conducted in the same laboratory and involving participants with varying levels of fitness. Twenty-four endurance-trained participants were recruited for this study and were classified as over 60 years old (O60, n = 12, n = 1 female and n = 11 males) or under 40 years old (U40, n = 12, n = 4 females and n = 8 males) (Table 1). Participants were participating in aerobic exercise at least 5 hours a week for at least 5 years. Smoking and presenting any musculoskeletal injury, neuromuscular disease, lower limb injury within the past 6 months were set as exclusion criteria. Prior to all visits, participants were requested to refrain from stimulants or depressants (including caffeine or alcohol) and strenuous exercise for 24 h while replicating the same dietary intake. Throughout the duration of the study, participants were asked to maintain regular training commitments. Prior to data collection, participants provided written informed consent in accordance with the Edith Cowan University Human Research Ethics Committee (REF: 2020-01267- HADDAD). Procedures used within this investigation conformed to the Declaration of Helsinki.

All participants completed three laboratory-based cycling sessions separated by at least 48 h, and no more than 10 days. During the first visit, participants performed an incremental cycling test to determine peak oxygen consumption followed by a familiarisation to maximal self-paced cycling exercise (described below). During visits two and three, participants performed experimental trials using either double-leg (DL) or single-leg (SL) cycling in a randomised and counterbalanced order (described below). All visits were completed under standard laboratory conditions (20-21°C, 40-50% relative humidity) and at a similar time of day (±1 h) to avoid any possible effect of the circadian rhythm.

Participants completed the incremental cycling test on an electromagnetically braked SRM cycle ergometer (Schoberer Rad Messtechnik, Jülich, Germany). After three stages (each 3 min) performed at 75, 100 and 125 W, intensity was further increased by 15 W every minute until volitional exhaustion or a cadence i 60 rpm could not be maintained. Verbal encouragement was given to the participants. Gas exchange was measured via a verified respiratory gas analyser (TrueOne, ParvoMedics, Sandy, UT, USA), sampled at 1 Hz and averaged into 30 sec intervals. Peak oxygen consumption (V̇O_2peak) was defined as the highest V̇O₂ value recorded over a 30 sec moving average during the incremental test. Heart rate (HR) was sampled at 1Hz (Polar, S610, Polar Electro Oy, Finland) and peak heart rate (HR_peak) was defined as the highest value recorded during the incremental test. Power output was sampled at a rate of 2 Hz and peak power output (PPO) was defined as the power output that elicited V̇O_2peak. After recovery (. 20min), participants were familiarised to maximal self-paced cycling exercise using double-leg cycling (DL) and single-leg cycling (SL) on a randomly chosen leg. The familiarisation included 4 repeated bouts of 30 sec of cycling from a stationary start where participants progressively increased power at self-selected intensity, followed by a 3 min maximal self-paced effort.

Prior to the maximal DL and SL self-paced cycling test, participants performed 5 min of passive rest on the bike in the cycling position (PRE), followed by a short duration (3min) submaximal exercise to allow participants to briefly be accustomed to the ergometer and task (i.e., SL vs DL cycling). During the SL session, the submaximal exercise was performed single-legged at 25% PPO whereas during the DL session, it was performed double-legged at 50% PPO. After some passive rest (>15 min), the participants then performed a maximal self-paced cycling test involving two intervals of 3 min interspersed by 90 s of passive recovery (i.e., two double-leg intervals during DL or four single-leg intervals during SL cycling). Such duration was chosen to spend time l 95% V̇O_2peak during DL therefore increasing the likelihood that performance during double-leg cycling would be limited by the cardiovascular system (Abbiss et al., 2011). The order of conditions (i.e., SL vs DL cycling), and the first leg during SL cycling (i.e., left vs right) was randomised and counterbalanced across participants. During SL cycling, at least 15 minutes of recovery were given after testing the first leg and prior to testing the second leg. A 10 kg counterweight was used during SL to reduce the differences in flexion and extension compared to the bilateral cycling pattern (Bini et al., 2015). Continuous feedback was provided for time only and participants were instructed and encouraged to perform each interval at the highest average power output they could sustain (Abbiss et al., 2011). The individuals were free to manipulate cadence and/or resistance at any time during the effort to maximise their power output. Perceptual responses were assessed within 30 sec after each 3 min interval using 3 different scales: perception of effort (0-10: “how much you were trying to”), rating of perceived exertion (Borg’s RPE scale, 6-20: “your level of whole-body perceived exertion”) and pain (0-10: “your own subjective definition of pain”). Leg muscle oxygen saturation (SmO₂) and total haemoglobin (THb), and femoral blood flow were assessed pre- and post-exercise using near-infrared spectroscopy (NIRS) and Doppler ultrasound respectively (described below). The first leg used during SL cycling was the leg chosen to measure NIRS and blood flow (BF) since local hemodynamic responses can be affected by the order in which each leg performs sequential SL cycling (Gordon et al., 2018). The same leg was used to measured NIRS and BF during each modality (i.e., DL and SL cycling) and was randomised and counterbalanced among participants.

Power output (SRM cycle ergometer) was averaged over each interval of the maximal self-paced cycling test and presented for each modality separately (i.e., DL vs SL cycling). Oxygen consumption, minute ventilation(metabolic analysis system), and heart rate (Polar, S610) were averaged over 60 sec prior to exercise (i.e., PRE), during each 3-min interval (i.e., INT1 and INT2), immediately post-exercise (POST INT1 and POST INT2) and throughout the passive recovery (i.e., 3-4 min, 5-6 min, 9-10 min post-exercise).

Tissue oxygenation was measured using a portable NIRS device (Moxy Monitor, Fortiori Design LLC, MN, USA). The NIRS Moxy Monitor calculates total haemoglobin (sum of oxygenated and deoxygenated haemoglobin, THb) and leg muscle oxygen saturation (oxygenated haemoglobin expressed as percentage of the total haemoglobin) based upon absorption and reflection of near-infrared light by oxygenated and deoxygenated haemoglobin (Crum et al., 2017). The sensor was placed on the vastus lateralis, one of the main contributors to power production during cycling. The leg that BF and NIRS was assessed on was equipped with a Moxy Monitor sensor at mid-point between the greater trochanter and lateral epicondyle of the femur following manufacturer guidelines, and its position was marked to ensure consistency between the trials. The sensor was secured with a light shield and athletic tape to block ambient light from interfering with the detectors. SmO₂ and THb were averaged over 60 sec prior to exercise (i.e., PRE), during each interval (i.e., INT1 and INT2), immediately post-exercise (POST INT1 and POST INT2) and throughout the passive recovery (i.e., 3-4 min, 5-6 min, 9-10 min post-exercise). SmO₂ and THb during exercise and recovery were expressed as a percentage of resting values.

Blood flow was assessed in the superficial femoral artery using a high-resolution duplex ultrasound system (T3000, Terason, Burlington, MA, USA). Ultrasound parameters were set to optimize the longitudinal B-mode images of the lumen-arterial wall interface using the lowest possible insonation angle (always <60°). Resting images and Doppler velocity were recorded prior to any exercise in a standardised seated cycling position for 1 min to quantify mean blood flow pre-exercise. After exercise, participants were instructed to immediately extend the measurement leg perpendicular to the ground, while staying seated on the cycle ergometer. It was ensured that the cycling position was standardised for each participant. The hip angle during BF assessment was not significantly different between older and younger participants (113 ± 5 vs 116 ± 7°, respectively). Leg blood flow post-exercise was assessed using the same settings and optimal images as pre-exercise. BF was averaged over 60 sec prior to exercise (i.e., PRE) and re-assessed immediately after interval 1 (POST INT1) and interval 2 (POST INT2) and later during the recovery. A qualitative signal was obtained on average 10.3 ± 6.5 and 9.7 ± 5.4 sec after interval 1 (POST INT1) and 12.2 ± 7.9 and 10.8 ± 6.2 sec after interval 2 (POST INT2) using DL and SL cycling respectively (mean ± standard deviation). A signal was deemed appropriate when a region of interest for both the vessel diameter and the velocity trace could be determined to enable offline analysis using a custom edge-detection software for offline analysis (software analysis described below). This implied the ability to clearly identify the anterior and posterior walls of the artery and its intima for a prolonged duration without interruption. To maximise the strength of blood flow analysis, the values recorded immediately post-exercise were analysed only from 15 sec post-exercise where a signal was appropriate for at least e 85% of participants (p 90% after 20 sec). Peak blood flow post-exercise (i.e., greatest flow achieved over 1 sec averaged window) was assessed within the window from 15 to 45 sec after each interval separately. Peak blood flow after intervals 1 and 2 was also averaged as a single value for DL and SL cycling respectively to investigate the relationship between femoral blood flow and power output per leg (Pearson’s correlation). Mean BF was presenter after each interval and calculated over a 30 sec window from 15 to 45 sec post-exercise. Mean BF was also calculated over a 30 sec window throughout the course of passive recovery after 3 min (3 min to 3 min 30 sec), 5 min (5 min to 5 min 30 sec) and 9 min (9 min to 9 min 30 sec). Peak blood flow post-exercise (i.e., greatest flow achieved over 1 sec averaged window) was assessed within the window from 15 to 45 sec after each interval separately.

Power output during exercise

Power output during DL cycling was the averaged power produced by the two legs simultaneously during intervals 1 and 2, whereas power output during SL cycling was the sum of the averaged power from each leg sequentially (i.e., left plus right leg) during both intervals. Power output during intervals 1 and 2 was also averaged as a single value during DL and SL cycling respectively and expressed in power output per leg (i.e., 50% power during DL and 100% power during SL) to investigate the relationship between peak femoral blood flow and power output per leg (Pearson’s correlation).

Vessel diameter and blood velocity were analysed using custom-designed, edge-detection wall-tracking software (FMD/Blood Flow 4.0, LabView 10.0, National Instruments, Austin, Texas, USA) which is validated and investigator-independent with an intra-observer coefficient of 6.7% (Woodman et al., 2001). In addition, a visual inspection of vessel diameter and blood velocity interpolated at 99Hz was performed by two operators to remove any potential outliers. Subsequent analysis was performed over data interpolated at 1Hz. Simultaneous regions of interest were identified on both the B-mode and Doppler images to analyse arterial diameter and blood velocity profile, respectively. Blood flow was calculated using the equation: blood flow = [(mean blood velocity) × (π × (vessel radius²)) x 60], where mean blood velocity is expressed in cm/s and radius is expressed in cm. Blood flow measured before and after exercise was presented as the mean blood flow (ml/min), which is calculated as a running mean of peak blood flow over three cardiac cycles that shifts by one cardiac cycle each time, and then multiplied by the constant of 0.5 and the angle correction factor based on the insonation angle (i.e., 60°). An additional analysis was performed over the time-course of recovery (i.e., after 15-45 sec, 3 to 3 min 30, 5 to 5 min 30, and 9 to 9 min 30 of recovery) where mean blood flow was expressed as a % of the peak value measured after interval 2. Blood flow data were excluded for one participant since the signal recorded in the first leg that exercised during SL cycling did not satisfy analysis requirements and using data from the second leg may have led to misinterpretation due to the influence of leg order.

Participants characteristics (i.e., age, body mass, height, peak oxygen consumption, peak power output, peak heart rate, weekly training volume, peak blood flow in response to exercise and hip angle during BF assessment were compared between older (O60) and younger (U40) individuals using independent samples t-tests. Normality of the distribution was tested and confirmed using a Shapiro-Wilk test. A linear mixed model was used to compare peak blood flow post-exercise between older (O60) and younger (U40) (age was set as a fixed effect and participants were set as random effect).

A linear mixed model was used to determine the effects of modality (i.e., DL or SL) and age (i.e., O60 or U40) on the physiological (i.e., power output, cadence, heart rate, oxygen consumption, minute ventilation, SmO₂, THb, femoral blood flow) and perceptual (i.e., effort, exertion, pain) responses during single-leg and double-leg cycling. Modality (i.e., DL or SL) and age (i.e., O60 or U40) were set as fixed effects and participants were set as random effect. Assumptions of residuals normality and homoscedasticity were confirmed by inspecting residuals plots. Bivariate associations were determined by Pearson’s correlation coefficients between femoral blood flow and power output and peak oxygen consumption. 95% confidence intervals and the equations for each linear regression were calculated. Statistical significance was set at P < 0.05. All data are presented as means ± SD. Statistical analysis was performed using Jasp 0.16 software.

A main effect for power output was observed between SL (sum of both legs) and DL cycling (295 ± 83 vs 265 ± 70 W respectively, P < 0.001). Power output was not significantly different (P = 0.053) between older vs younger endurance-trained individuals when SL and DL were considered together, although a trend was observed. Power output was 268 ± 72 vs 233 ± 52 W in older individuals and 322 ± 87 vs 296 ± 74 W in younger individuals during SL vs DL cycling, respectively.

A main effect for mean pre-exercise blood flow was observed between older compared with younger individuals (138.5 ± 62.4 vs 197.4 ± 81.5 ml/min respectively, P = 0.045, Figure 1, A and B).

V̇O₂, HR and V̇_E were significantly lower at all-time points during and after intervals 1 and 2 performed with SL compared with DL cycling (Table 2). SmO₂ was significantly lower at all-time points (except 5 min post-exercise) during DL compared with SL cycling (Table 3). THb was not different between modalities (Table 3). Mean BF was greater after intervals 1 and 2 performed using SL compared with DL cycling until 3 min post-exercise and was not significantly different thereafter (i.e., after 5 and 9 min post-exercise (Figure 1). Peak BF was greater immediately after interval 1 (1821.7 ± 632.2 vs 1322.6 ± 478.5 ml/min, P < 0.001) and after interval 2 (1667.0 ± 500.6 vs 1367.4 ± 344.8 ml/min, P < 0.001) performed using SL compared with DL cycling. When expressed as a % of peak, BF post interval 2 was not significantly different between modalities throughout the course of recovery. V̇O₂, HR and V̇_E were significantly greater in younger compared with older groups at some but not all time points during and after exercise (Table 2). SmO₂ and THb did not differ significantly between age groups throughout exercise and recovery (Table 2 and Table 3). Peak BF and mean BF at every time point were significantly lower in older compared with younger individuals (Figure 1). When expressed as a % of peak, mean BF post interval 2 was significantly lower in older compared with younger individuals throughout the course of recovery after 15-45 sec (81 ± 7 vs 88 ± 5 %, P = 0.011), 3-3.5 min (31 ± 9 vs 52 ± 12 %, P < 0.001), 5-5.5 min (24 ± 5 vs 37 ± 10 %, P = 0.005) and 9-9.5 min (19 ± 8 vs 31 ± 13 %, P < 0.001).

An interaction effect (modality x age) was observed during and after exercise for V̇O₂ and V̇_E. Younger individuals had a higher V̇O₂ (V̇O₂ during SL was 72% vs 82% DL in U40 compared with O60) and a higher V̇_E (V̇_E during SL was 76% vs 89% DL in U40 compared to O60) during SL relatively to DL (Table 2). An interaction effect (modality x age) was also observed for mean BF 3 min post-exercise. Mean BF 3 min post-exercise after SL was lower in O60 but greater in U40 compared with DL cycling (90% DL in O60 vs 130% DL in U40, P = 0.022, Figure 1).

Effort (8.7 ± 1.1 vs 9.3 ± 0.7 a.u., P < 0.001), exertion (16.7 ± 1.6 vs 18.2 ± 1.1 a.u.), and pain (6.0 ± 3.1 vs 6.9 ± 3.0 a.u., P < 0.001) were significantly lower after SL compared with DL cycling. Perceptual responses did not differ between older and younger individuals.

A significant relationship (SL: r = 0.655, P = 0.002; DL: r = 0.666, P = 0.001) between peak femoral blood flow post-exercise and power output per leg was observed, when all participants were analysed as a single group and both intervals averaged as a single measure (Figure 2). When analysing the two age groups separately, the relationship between Peak BF and power output per leg was significant during SL but not DL in the older sub-group (SL: r = 0.615, P = 0.044; DL: r = 0.514, P = 0.106) and no significant relationship was observed in the younger sub-group (SL: r = 0.618, P = 0.076; DL: r = 0.617, P = 0.077).

To better understand the effect of ageing on the hyperaemic response to maximal exercise when various amount of muscle mass is engaged, femoral blood flow in response to maximal single-leg and double-leg cycling exercise was compared in younger (<40 years) and older (>60 years) trained individuals. The major findings were that:

1. Single-leg cycling elicited a greater femoral blood flow compared with double-leg cycling in both older and younger endurance-trained individuals, which may partly explain the greater power output observed during SL compared to DL.
2. Femoral blood flow in response to maximal exercise was lower in older compared with younger endurance-trained individuals regardless of the amount of muscle mass engaged.

In the present study, single-leg cycling elicited a greater femoral blood flow compared with double-leg cycling (Figure 1). Indeed, peak blood flow measured following SL cycling corresponded to ~132 and ~130% of DL cycling in older and younger endurance-trained individuals, respectively. To our knowledge, this is the first study to show that femoral blood flow was greater in response to maximal SL compared with DL cycling in both older and younger trained populations. Yet, these results are in agreement with observations of greater femoral blood flow during SL cycling in a limited cohort of healthy (n = 4) and clinical (patients with heart failure, n = 5) individuals (LeJemtel et al., 1986) and after submaximal SL cycling exercise (Burns et al., 2014; Draper et al., 2022). As such, this study supports the finding that a transition from large to small muscle mass exercise may allow for a greater peripheral hyperaemic response (Calbet et al., 2009; Cardinale et al., 2019; Klausen et al., 1982). In turn, this greater femoral blood flow may partly explain the greater power output during SL compared with DL cycling observed here and in previous studies (Abbiss et al., 2011; Gordon et al., 2018; Iannetta et al., 2019). Indeed, during exercise involving a small muscle mass, blood pressure is not threatened to the same extent as maximal exercise involving a large muscle mass where a decrease in vascular conductance is required to avoid hypotension (Calbet et al., 2004; Davies and Sargeant, 1974; Klausen et al., 1982; Saltin and Calbet, 2006; Volianitis and Secher, 2002). As a result, a greater proportion of cardiac output was presumably redistributed to the active muscles during SL compared with DL cycling (Iannetta et al., 2019) although we were not able to measure blood pressure and vascular conductance. Overall, the present study illustrates a dissociation between exercise performance and the level of pulmonary and cardiovascular stress as previously documented (Abbiss et al., 2011; Gordon et al., 2018; Gordon et al., 2020) and indicates that the greater power output during SL cycling is accompanied by a greater blood flow but a lower heart rate, oxygen consumption, and minute ventilation compared with DL cycling (Table 2).

Noteworthy, no change in THb was observed when reducing muscle mass in the present study (Table 3) and previously (Gordon et al., 2018) despite observing a difference in BF (Figure 1). This indicates that THb may have some limitations to appropriately measure BF after a cycling task as previously suggested (Gordon et al., 2020) although others have reported a correlation between THb and blood flow after isolated muscle contractions (Alvares et al., 2020). Since leg dominance was not considered in this cohort, it is plausible that using dominant and non-dominant leg during SL without distinction may have impacted power output and NIRS, and masked some differences compared with double-leg cycling (Iannetta et al., 2019). Yet, single-leg power output differed only by ˜2.1% between left and right leg, indicating that randomising the leg used to assess NIRS and BF minimised the effect of leg dominance on tissue oxygenation or blood flow.

Muscle oxygenation was greater during SL compared with DL cycling (Table 3), indicating that oxygen extraction may be attenuated during small compared with large muscle mass exercise, in agreement with some (Cardinale et al., 2019) but in contrast to other (Gordon et al., 2018; Skattebo et al., 2020a) studies. Such discrepancies may partly be explained by the differences in the site of measurement (e.g., vastus lateralis vs rectus femoris) and in the exercise protocol (e.g., incremental exercise vs intervals). Although greater blood flow and reduced blood transit time may be partly responsible for the lower oxygen extraction (Richardson et al., 1999) observed during small compared with large muscle mass exercise in the present study, there is recent evidence to suggest that it may rather be a consequence of the attainment of maximal mitochondrial capacity (Cardinale et al., 2019). Indeed, since a greater proportion of cardiac output is available during SL compared with DL cycling (Davies and Sargeant, 1974; Iannetta et al., 2019; Klausen et al., 1982), aerobic performance during single-leg cycling seems to be primarily related to the ability to diffuse oxygen from the microvasculature to the mitochondria (Richardson et al., 1999), and the mitochondrial capacity to utilise oxygen for oxidative phosphorylation (Skattebo et al., 2020b). While some authors have previously reported that mitochondrial capacity was greater than the metabolic demand during single-leg knee extension in trained individuals (Gifford et al., 2016), the greater muscle oxygenation observed during SL, compared with DL cycling, indicates that the maximal mitochondrial activity was presumably reached during maximal self-paced single-leg but not double-leg cycling within the present study. Moreover, the strong relationship observed between peak femoral blood flow and power output during both DL and SL cycling (Figure 2) indicates that blood flow was appropriately regulated during exercise to meet the metabolic demand of the exercise task, as previously observed by others (Calbet et al., 2006), regardless of the amount of active muscle mass.

Reducing muscle mass improved oxygen delivery and exercise performance in both older and younger individuals. This is illustrated by the greater femoral blood flow and power output observed during SL compared with DL cycling in both groups (Figure 1). Moreover, no effect of age was observed on leg muscle oxygen saturation or total haemoglobin during DL and SL cycling (Table 3), indicating that the older individuals were able to accommodate an increase in blood flow in a similar way to younger individuals. It was however not accounted for the possible influence of adipose tissue thickness on the NIRS signal (Hamaoka and McCully, 2019), and these results should be interpreted with caution, although the comparison of leg muscle oxygen saturation in relative values (i.e., % THb) could minimise the risk of error during inter-individual comparison.Femoral blood flow was consistently lower at rest, immediately post-exercise, and over the course of recovery (Figure 1) in response to both large (i.e., DL) and small (i.e., SL) muscle mass exercise in the older compared to the younger group. The ~30% lower peak blood flow observed in older compared to younger individuals (Figure 1) is in agreement with previous findings showing a reduction in leg blood flow during maximal exercise with ageing (Beere et al., 1999; Donato et al., 2006; Poole et al., 2003; Proctor et al., 2004). Although not significantly different, mean power output was slightly lower during the single-leg and double-leg cycling efforts in the older, compared with the younger athletes, which may contribute to explain the lower blood flow observed. Importantly, despite evidence showing that skeletal muscle blood flow increases in response to exercise to match metabolic demand (Richardson et al., 1993), it is hard to distinguish whether the greater blood flow observed during SL cycling (in both older and younger athletes) was responsible for, or the consequence of a greater power output. Since a greater blood flow may also increase muscle glucose uptake (Malone et al., 2021), the possible contribution of nutrient delivery to the greater power output per leg observed during small vs large muscle deserves further attention in the future.

The lower femoral blood flow in the older compared with the younger group (Figure 1) observed in the present study contrasts with the findings from others who reported that blood flow was preserved with ageing during submaximal exercise involving a large muscle mass (Poole et al., 2003) and during exercise involving a small muscle mass (Donato et al., 2006; Magnusson et al., 1994; Mortensen et al., 2012). It is possible that the absence of age-related difference in blood flow was due to the intensity and/or nature of the exercise (e.g., submaximal intensity and/or small muscle mass exercise) used in previous studies. Indeed, any age-related reduction in vascular function (e.g., vasodilatory capacity) might be compensated by an increase blood flow to the locomotor muscles in response to an increase in cardiac output (Montero, 2015) during such exercise. Yet, we observed a lower hyperaemic response following exercise involving a small muscle mass in our older participants (Table 2). As a result, it is unlikely that cardiac output limited blood flow in the older group since heart rate, and presumably cardiac output, was not maximal during maximal self-paced SL cycling exercise (Davies and Sargeant, 1974; Iannetta et al., 2019; Klausen et al., 1982). Interestingly, the lower femoral blood flow in older vs younger individuals (Figure 1) was observed in the present study despite a consistent participation in aerobic activity in our sample of endurance-trained athletes. This finding extends previous results during submaximal exercise in older trained athletes (Proctor et al., 1998) and indicates that femoral blood flow may be lower in older compared with younger individuals regardless of the amount of muscle mass engaged and their training status. Further research is however needed to understand the longitudinal effects of ageing and training on vascular function.

Femoral blood flow was greater in response to maximal exercise involving a small (SL cycling) compared to a large (DL cycling) muscle mass in both younger and older endurance-trained individuals. This greater hyperaemic response likely highlights an increase in oxygen delivery to the locomotor muscles thereby explaining the observed improvement in performance. Femoral blood flow post-exercise was however lower in the older compared with the younger athletes, despite regular endurance training.