Introduction

To perform exercise for more than a few minutes skeletal muscle blood flow \( {\left( {\ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} } \right)} \) must meet the metabolic demand. Importantly, muscle contraction has direct effects on muscle blood flow that depend on exercise mode, intensity and frequency of contractions (Bellemare et al. 1983; Buchler et al. 1985; Walloe et al. 1988; Lutjemeier et al. 2005). Even with low-to-moderate force development the intramuscular pressure generated during contraction exceeds arterial pressure, such that blood flows through the muscle in the relaxation phase (Walloe et al. 1988). Because of the shorter period of time spent in the relaxation phase at high contraction frequencies and duty cycles blood flow may become limited during such contractions, as is indeed observed in the dog diaphragm (Bellemare et al. 1983; Buchler et al. 1985). Specifically, there is an interaction between duty cycle and frequency such that for constant duty cycle \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) becomes limited at frequencies above 80 contractions/min (cpm) (Buchler et al. 1985).

Hoelting et al. (2001) extended the findings from the diaphragm to human muscles to show that shortening the relaxation phase by increasing both the contraction frequency and duty cycle results in a smaller increase in \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) during incremental knee-extension exercise than expected from the increased power output. However, others find that when knee-extension exercise is performed at a constant duty cycle, elevations in contraction frequency do not impair \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) (Sjogaard et al. 2002; Osada et al. 2002). In fact, the greater muscle oxygen consumption \( {\left( {\ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} } \right)} \) during heavy exercise performed at 100 than at 60 cpm is matched by an increase in \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) such that leg O2 extraction is similar at 60 and 100 cpm (Ferguson et al. 2001).

The primary concern regarding the adequacy of \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) during muscle contractions is capillary O2 exchange. Near-infrared spectroscopy (NIRS) uses the optical properties of light-absorbing chromophores (predominantly hemoglobin and myoglobin) to noninvasively estimate tissue O2 saturation \( {\left( {{\text{St}}_{{{\text{O}}_{{\text{2}}} }} } \right)} \) or O2 extraction in the skeletal muscle microcirculation. During cycling (Takaishi et al. 2002) and plantar flexion exercise (Quaresima et al. 2001) \( {\text{St}}_{{{\text{O}}_{{\text{2}}} }} \) is similar at contraction rates ranging from 40 to 80–85 cpm suggesting that \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) in the microcirculation is not limited during exercise performed ≤85 cpm. However, a shortcoming of \( {\text{St}}_{{{\text{O}}_{{\text{2}}} }} \) measurements is that they are affected by changes in blood volume under the NIRS probe (Ferrari et al. 1997). Moreover, these studies (Quaresima et al. 2001; Takaishi et al. 2002) did not evaluate the effects of increasing contraction frequency on the \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2M}}}} }} {\text{/}}\ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) balance across a wide range of exercise intensities. The deoxyhemoglobin concentration ([HHb]) signal from NIRS is less sensitive to blood volume changes, and thus is considered a better estimate of capillary O2 extraction (Kowalchuk et al. 2002; Grassi et al. 2003; Ferreira et al. 2005), reflecting the \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2M}}}} }} {\text{/}}\ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) in the microcirculation.

The current investigation examined the effects of pedal cadence (60 cpm vs. 100 cpm) on [HHb] at low, moderate and peak exercise. We anticipated one of the following outcomes: (a) If the increase in \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} \) at elevated contraction frequencies is matched by an increase in \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \), [HHb] will be similar at 60 and 100 rpm; conversely (b) if \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) during moderate (~lactate threshold) and peak exercise is limited at elevated contraction frequencies when \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} \) is expected to be similar (Jones et al. 2004), [HHb] will be greater at 100 rpm compared to 60 rpm. Since increasing pedal cadence during cycling exercise is similar to knee-extension exercise where increases in contraction frequency are accomplished with relatively constant duty cycles (Ferguson et al. 2001; Osada et al. 2002), we hypothesized that [HHb] would be similar at 60 and 100 cpm.

Methods

This study involved healthy, physically active subjects (seven males, two females; 24.7±6.3 years; 67.9±12.2 kg, mean±SD) who volunteered to participate after receiving detailed explanation of the experimental protocol, which was approved by the Institutional Review Board and followed the principles outlined in the Declaration of Helsinki. Each subject provided written informed consent prior to performance of any procedure.

The protocol involved two visits to the laboratory, separated by at least 48 h, but no longer than 1 week, for incremental cycling exercise on an electronically braked cycle ergometer (Corival 400, Lode, The Netherlands). Incremental exercise involved 4 min of baseline cycling (20 W) at the desired cadence (60 or 100 cpm) followed by a progressive (ramp) increase in exercise intensity (15–30 W/min) until volitional fatigue or until the cadence was lower than 55 or 95 cpm, respectively, for 5–10 s despite strong verbal encouragement. The incremental work rate, seat height and handlebar position on the cycle ergometer were reproduced on the second test. Pulmonary gas exchange and skeletal muscle O2 extraction, estimated by the deoxy-hemoglobin concentration ([HHb]) determined with NIRS, were measured during both exercise tests. Peak oxygen uptake \( {\left( {\ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} } \right)} \) was defined as the highest \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} \) achieved during the test averaged over a 15-s interval. The LT was estimated from gas exchange measurements using the V-slope method (i.e., the change in the slope of the pulmonary \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{CO}}_{2} }} {\text {--}} \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{2} }} \) relationship), ventilatory equivalents and end-tidal gas tensions (Wasserman et al. 1973; Beaver et al. 1986).

Pulmonary gas exchange \( {\left( {\ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} {\text{ and }}\ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{CO}}_{{\text{2}}} }} } \right)} \) and minute expired ventilation \( {\left( {\ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{\text{E}}} } \right)} \) were measured breath-by-breath (CardiO2, Medical Graphics Corporation, St. Paul, MN). Before each test, the volume signal was calibrated by pumping a 3-l syringe at flow rates spanning the range expected during the exercise studies, while the O2 and CO2 analyzers were calibrated with gases of known concentration. Heart rate was recorded from the electrocardiogram using a modified lead I configuration and stored in the breath-by-breath file.

Skeletal muscle [HHb] was evaluated by a frequency-domain multidistance (FDMD) NIRS system (OxiplexTS, ISS, Champaign, IL). The NIRS probe consists of eight light-emitting diodes (LED), operating at wavelengths of 690 (four LEDs) and 830 nm (four LEDs), and one detector fiber bundle. Source–detector separations were 2.0, 2.5, 3.0 and 3.5 cm. The modulation frequency of the light-source intensity was 110 MHz and the cross-correlation frequency for heterodyne detection was 5 KHz. The NIRS data were stored at a frequency of 31.25 Hz. The probe was positioned longitudinally on the belly of the muscle vastus lateralis, bonded to the skin (Skin-Bond, Smith & Nephew, Largo, FL) and secured with Velcro® straps around the thigh. The probe position was pen-marked and in the second visit the probe was placed at the same site. The NIRS was calibrated on each test day following the manufacturer’s recommendations.

The FDMD NIRS provides measurement of absolute concentration (expressed in μM) of oxy-([HbO2]) and deoxy-hemoglobin ([HHb]). [HHb] and [HbO2] were calculated incorporating the dynamic measurement of reduced scattering coefficients made throughout the exercise test. Total hemoglobin concentration ([THb]) was the sum of [HHb] and [HbO2], while \( {\text{St}}_{{{\text{O}}_{{\text{2}}} }} \) = [HbO2]/[THb] (%). The NIRS data were averaged in 15-s bins and the [HHb], [HbO2], [THb] and \( {\text{St}}_{{{\text{O}}_{{\text{2}}} }} \) corresponding to the baseline cycling (60 s average), LT and \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} \) (15 s average for both) were determined throughout exercise.

The [HHb] signal is less sensitive to changes in blood volume under the area of the probe than \( {\text{St}}_{{{\text{O}}_{{\text{2}}} }} \) (Ferrari et al. 1997) and is considered an estimate of O2 extraction in the microcirculation (Kowalchuk et al. 2002; Grassi et al. 2003; Ferreira et al. 2005). The NIRS–FDMD provides absolute measurements of [HHb]; however, the day-to-day reproducibility is unclear. To examine the reproducibility of [HHb] one subject visited the laboratory on four different days and [HHb] was measured during 4 min of baseline pedaling (20 W) at 60 cpm. Sixty seconds of data were averaged to determine [HHb] each day.

Data analysis

To determine differences between two means, a two-tailed Student’s paired t test was performed. A repeated-measures analysis of variance was performed to compare more than two means and the Bonferroni post hoc test was used for pairwise comparisons. Statistical significance was accepted when P<0.05 (NCSS 2000—Statistical Software, Kaysville, UT). Values were reported as mean±SD, unless otherwise specified.

Results

The coefficient of variation for [HHb] on one subject during baseline cycling at different days was 7.0% ([HHb]; Day 1=16.8 μM, Day 2=18.5 μM, Day 3=18.3 μM and Day 4=20.0 μM).

The responses of pulmonary \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} \) and [HHb] for a representative subject are shown in Fig. 1, while the results are summarized in Fig. 2. Baseline and peak \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} \) were greater at 100 cpm compared with 60 cpm. However, the lactate threshold at 60 and 100 cpm was not significantly different, either as absolute value or as percent of \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} \) (LT60rpm=56.9±7.8% and LT100rpm=57.2±7.0%; P=0.89). The [HHb] increased from baseline pedaling to peak exercise. However, there was no significant difference for the 60- and 100-cpm trials.

Fig. 1
figure 1

Pulmonary \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} \) (Top) and [HHb] (Bottom) from a representative subject during incremental exercise performed at 60 and 100 cpm. Dotted line depicts the onset of incremental exercise

Full size image
Fig. 2
figure 2

Pulmonary \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} \) (Top) and Deoxy-hemoglobin concentration ([HHb], Bottom) at Baseline (BSL), estimated lactate threshold (LT) and peak exercise (Peak) during cycling at 60 and 100 cpm. Values are mean±SD (error bars). Significant difference between pedal cadences (*) and exercise intensities (δ)

Full size image

The \( {\text{St}}_{{{\text{O}}_{{\text{2}}} }} \) decreased from BSL to Peak and \( {\text{St}}_{{{\text{O}}_{{\text{2}}} }} \) at 100 cpm was slightly, but significantly greater than for the 60-cpm trial (Table 1). Similarly, [HbO2] was greater at 100 than at 60 cpm; however, [HbO2] did not change significantly from baseline to peak exercise. The [THb] tended to be higher at 100 cpm compared to 60 cpm. In addition, [THb] increased from baseline to the lactate threshold but did not change further at peak exercise. Peak work rate was not significantly different between 60 and 100 cpm.

Table 1 Muscle oxygenation and power during incremental cycling exercise at 60 and 100 cpm
Full size table

Discussion

The principal new finding was the similar [HHb], assumed to reflect capillary O2 extraction, at baseline (20 W), estimated lactate threshold and \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} \) during incremental cycling exercise performed at 60 and 100 cpm. This was so even though \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} \) was higher at 100 cpm compared to 60 cpm. These data suggest that during cycling exercise at 100 cpm \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) was not limited any further than during cycling at 60 cpm. Thus, our data indicate that the likely greater (internal) work performed at 100 cpm was ‘matched’ by a higher \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) compared to 60 cpm at a given external power.

The NIRS is a noninvasive estimate of muscle oxygenation that has limitations which cannot be circumvented (Ferrari et al. 2004). The main concerns regarding NIRS are an inability to separate light absorption by hemoglobin from that by myoglobin, and the relative contribution of arterioles, capillaries and venules to the signal at rest and during exercise. We assumed that changes in [HHb] provide an estimate of O2 extraction, and that the relative contribution of vascular compartments to the NIRS signal was similar and unchanging across conditions and during exercise. Myoglobin deoxygenation is relatively unchanged from 50–60% WRpeak to peak work rate (Richardson et al. 1995) while [HHb] continued to increase above 50% WRpeak, which is qualitatively similar to leg O2 extraction (Knight et al. 1992; Richardson et al. 1993), suggesting that changes in the NIRS signal originated primarily from changes in hemoglobin. The validity of the assumption regarding the relative contribution of arterioles, capillaries and venules to the NIRS signal is not clear; nor is it clear if these relative proportions remain constant across exercise intensities. However, as capillaries comprise the majority of the muscle microvascular volume (Poole et al. 1995) it appears reasonable to consider that the changes in [HHb] will primarily reflect capillary O2 extraction.

We observed excellent reproducibility (CV=7.0%) for measurements during cycling at 20 W. We did not specifically test the reproducibility of measurements at the lactate threshold and \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} \). However, previous data from our laboratory suggest a CV of ~16% for changes in [HHb] during cycling (60 cpm) from 20 W to work rates representing ~50% and 80% \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} \) (Ferreira et al. 2005).

There is an interaction between pedal frequency and power that can affect muscle recruitment during cycling (MacIntosh et al. 2000). We studied the oxygenation profile of the vastus lateralis, as this muscle is a predominant contributor to the power, and thus pulmonary \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} , \) during cycling (Li et al. 1998). However, a possible increase in recruitment of other muscles, which would contribute to the elevated pulmonary \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} \) at 100 cpm, could not be assessed with the current protocol. Nevertheless, muscle-shortening velocity is the main determinant of metabolic cost of cycling at increased pedal cadence (McDaniel et al. 2002). Therefore, at least part of the increase in pulmonary \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} \) and presumably muscle \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} \) at 100 cpm was due to elevated vastus lateralis metabolism.

We observed that [HHb] during the baseline pedaling, estimated LT and \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} \) were not different at the 100 and 60 cpm trials, indicating a similar \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2M}}}} }} {\text{/}}\ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) in the microcirculation. These results are in agreement with investigations showing that O2 extraction (Ferguson et al. 2001) or muscle oxygen saturation by NIRS (Quaresima et al. 2001; Takaishi et al. 2002) are not affected by alterations in the contraction frequency. Thus, our data suggest that during cycling exercise performed at a duty cycle of ~1/3, muscle blood flow in the microcirculation was not mechanically impeded by increasing pedal cadence from 60 to 100 cpm. In fact, two aspects of exercise at higher contraction frequencies that might optimize muscle perfusion are the lower intramuscular pressures developed during the contraction period and a possibly greater contribution of the muscle pump (Gotshall et al. 1996; Sheriff et al. 2001). However, the effectiveness of the muscle pump as an important mechanism to increase \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) during exercise in humans may be confined to light work rates (Lutjemeier et al. 2005) while in animals its effect has been disputed (Hamann et al. 2004; Valic et al. 2005). In addition, higher mean arterial pressure during exercise at 100 cpm would aid to maintain (or increase) muscle blood flow at levels similar to those observed at 60 cpm (Gotshall et al. 1996), although this is only a partial explanation since systemic vascular resistance also decreases when pedal frequency is increased from 70 to 110 cpm (Gotshall et al. 1996).

The [HHb] determined with NIRS is less sensitive to changes in blood volume during exercise than \( {\text{St}}_{{{\text{O}}_{{\text{2}}} }} \) (Ferrari et al. 1997; Grassi et al. 2003) and is considered a reliable estimate of fractional capillary O2 extraction (Kowalchuk et al. 2002; Grassi et al. 2003; Ferreira et al. 2005). Based on the Fick principle, O2 extraction [C(a–v)O2] is equal to \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2M}}}} }} {\text{/}}\ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) (assuming constant CaO2). \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) is linearly related to \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2M}}}} }} \) and has a positive intercept on the \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) axis (Andersen et al. 1985; Richardson et al. 1993). This predicts that [C(a–v)O2] increases hyperbolically as a function of work rate, or more precisely \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2M}}}} }}, \) considering that CaO2 remains relatively constant in normal healthy subjects (Richardson et al. 1993; Whipp 1994). However, in subjects with low exercise capacity the plateau region of C(a–v)O2 seen at a higher \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} \) may not be achieved. In seven subjects the [HHb] response followed a hyperbolic profile (e.g., Fig. 1). Therefore, these observations suggest that the [HHb] response approximates predictions of the O2 extraction profile in the skeletal muscle microcirculation, assuming that the linear relationship between \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) and \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2M}}}} }} \) measured in larger vessels holds true for the microcirculation.

There is debate surrounding the limiting factors of maximal \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{\text{2}}} }} \) (di Prampero 2003). In healthy subjects, performing large muscle mass exercise under normoxia, \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} \) is determined primarily by the capacity to deliver O2 to the contracting muscle (Andersen et al. 1985; Knight et al. 1993; Richardson et al. 1993; Cardus et al. 1998; di Prampero 2003). We found an approximately 4% greater \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} \) at 100 cpm than at 60 cpm. However, an increase in \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} \) when exercise is performed at higher pedal frequencies is not a universal finding. In other studies (Zoladz et al. 1995; Jones et al. 2004) similar \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} \) was found for pedal cadences ranging from 40 to 120 cpm in subjects similar to those presently studied. It appears that in our study the greater \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} \) at 100 cpm was caused by an increased muscle O2 delivery, and not by an increased extraction as reflected by the similar [HHb] at \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} \) at 60 and 100 cpm.

Conclusion

We demonstrate that vastus lateralis oxygenation is similar during cycling exercise at 60 and 100 cpm, suggesting no impairment of muscle blood flow with the increase in contraction frequency. The similar muscle oxygenation and greater muscle metabolism at 100 cpm than at 60 cpm indicates a higher blood flow at 100 cpm, so that the ratio between O2 uptake and blood flow was similar during exercise at 60 and 100 cpm. In addition, a greater \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2peak}}}} }} \) was achieved during the 100 cpm trials, further suggesting increased muscle O2 delivery compared to the 60 cpm trial. Therefore, we conclude that vastus lateralis O2 extraction was similar at 60 and 100 cpm, suggesting that the \( \ifmmode\expandafter\dot\else\expandafter\.\fi{V}_{{{\text{O}}_{{{\text{2M}}}} }} {\text{/}}\ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) in the microcirculation was not altered and, presumably, no impairment of \( \ifmmode\expandafter\dot\else\expandafter\.\fi{Q}_{{\text{M}}} \) occurred with the increase in pedal frequency.