The effect of resistive breathing on leg muscle oxygenation using near-infrared spectroscopy during exercise in men

Transcription

1 The effect of resistive breathing on leg muscle oxygenation using near-infrared spectroscopy during exercise in men John M. Kowalchuk *, Harry B. Rossiter *, Susan A. Ward and Brian J. Whipp * * Department of Physiology, St George s Hospital Medical School, Cranmer Terrance, Tooting, London SW17 0RE, UK, Canadian Centre for Activity and Aging, School of Kinesiology and Department of Physiology, The University of Western Ontario, London, Ontario, Canada N6A 3K7 and Sport and Exercise Science Research Centre, South Bank University, London SE1 0AA, UK (Manuscript resubmitted 23 July 2002; accepted 24 July 2002) The effect of added respiratory work on leg muscle oxygenation during constant-load cycle ergometry was examined in six healthy adults. Exercise was initiated from a baseline of 20 W and increased to a power output corresponding to 90 % of the estimated lactate threshold (moderate exercise) and to a power output yielding a tolerance limit of 11.8 min (± 1.4, S.D.) (heavy exercise). Ventilation and pulmonary gas exchange were measured breath-by-breath. Profiles of leg muscle oxygenation were determined throughout the protocol using near-infrared (NIR) spectroscopy (Hamamatsu NIRO 500) with optodes aligned midway along the vastus lateralis of the dominant leg. Four conditions were tested: (i) control (Con) where the subjects breathed spontaneously throughout, (ii) controlled breathing (Con Br) where breathing frequency and tidal volume were matched to the Con profile, (iii) increased work of breathing (Resist Br) in which a resistance of 7 cmh 2 Ol _1 s _1 was inserted into the mouthpiece assembly, and (iv) partial leg blood flow occlusion (Leg Occl), where muscle perfusion was reduced by inflating a pressure cuff (~90 mmhg) around the upper right thigh. During Resist Br and Leg Occl, subjects controlled their breathing pattern to reproduce the ventilatory profile of Con. An ~3 min period with respiratory resistance or pressure cuff was introduced ~4 min after exercise onset. NIR spectroscopy data for reduced haemoglobin myoglobin (D[Hb]) were extracted from the continuous display at specific times prior to, during and after removal of the resistance or pressure cuff. While the D[Hb] increased during moderate- and heavy-intensity exercise, there was no additional increase in D[Hb] with Resist Br. In contrast, D[Hb] increased further with Leg Occl, reflecting increased muscle O 2 extraction during the period of reduced muscle blood flow. In conclusion, increasing the work of breathing did not increase leg muscle deoxygenation during heavy exercise. Assuming that leg muscle O 2 consumption did not decrease, this implies that leg blood flow was not reduced consequent to a redistribution of flow away from the working leg muscle. Experimental Physiology (2002) 87.5, The provision of adequate blood flow to working muscle is dependent on increases in both cardiac output and local vascular conductance. However, vascular resistance increases in other tissues including skin, viscera and inactive muscle to regulate blood pressure within narrow tolerance limits. It has now been established that, during high-intensity exercise that can be sustained for only 2 3 min, deliberate activation of other muscle groups to increase the active muscle mass results in reduction in muscle perfusion in the previously-active working muscle as a result of local vasoconstriction (Harms et al. 1997, 1998). Harms et al. demonstrated that during leg cycling exercise at a power output requiring > 95 % maximum O 2 uptake ( J,max ), a period of resistive breathing to increase inspiratory muscle work was associated with redistribution of blood flow away from the working leg muscles, presumably to 2456 *Corresponding author s address: School of Kinesiology, 3M Centre, The University of Western Ontario, London, Ontario, Canada N6A 3K7. jkowalch@uwo.ca Present address: Centre for Exercise Science and Medicine, Institute of Biomedical and Life Sciences, West Medical Building, University of Glasgow, Glasgow G12 8QQ, UK. Publication of The Physiological Society

2 602 J. M. Kowalchuk, H. B. Rossiter, S. A. Ward and B. J. Whipp Exp. Physiol support the newly recruited respiratory muscles. In these studies, the decrease in leg muscle blood flow was associated with a decrease in leg J, as leg muscle O 2 extraction could not increase to compensate for the reduction in blood flow. The authors argued that when exercise was truly maximal, both cardiac output and the arterio-venous O 2 content difference would be at maximal levels, and thus the blood flow requirements for supporting an increase in the supplementary active muscle mass could only be met through redistribution of blood flow away from other working muscle groups. Whether this steal phenomenon functions during submaximal exercise is controversial. Secher et al. (1977) demonstrated that when arm ergometry was performed in combination with leg ergometry, leg blood flow was reduced. In this study, the reduction in leg blood flow was associated with a compensatory increase in its arterio-venous O 2 content difference with leg J being maintained. Others, using combinations of leg and arm exercise to increase the mass of active muscle, have been unable to confirm that recruitment of additional muscle groups leads to reductions in blood flow to active leg muscle during submaximal exercise (Savard et al. 1989; Richter et al. 1992; Richardson et al. 1995; Strange, 1999). Recently, Wetter et al. (1999) demonstrated that when exercising at a submaximal power output requiring 75 % J,max, leg J and leg muscle blood flow were lower (with no change in arterio-venous O 2 content difference) during a period of inspiratory resistive breathing compared to a period of respiratory unloaded breathing, although no such differences were found relative to normal breathing or when exercising at only 50 % J,max. Muscle oxygenation, as reflected by muscle end-capillary and muscle tissue O 2 content, is dependent on the balance between O 2 delivery (i.e. muscle blood flow w arterial O 2 content) and muscle O 2 consumption. Changes in local muscle oxy- or deoxygenation status, as determined by changes in the combined tissue oxyhaemoglobin (HbO 2 ) and oxymyoglobin (MbO 2 ) concentrations and the combined reduced haemoglobin (Hb) and reduced myoglobin (Mb) concentrations, respectively, can be monitored noninvasively using transcutaneous near-infrared (NIR) spectroscopy (Chance et al. 1992; Mancini et al. 1994; Elwell, 1995). As myoglobin levels are small relative to those of haemoglobin, it is conventional to take changes in the oxygenation and deoxygenation signals to reflect changes in [HbO 2 ] and [Hb], respectively, although a recent study comparing NIR spectroscopy with 1 H- magnetic resonance spectroscopy suggests that NIR signals may correlate more closely to changes in Mb (Tran et al. 1999). During submaximal exercise for which cardiac output and muscle O 2 extraction are not yet maximal, a reduction in muscle perfusion would be expected to be associated with an increase in muscle O 2 extraction, assuming muscle O 2 consumption not to be altered. This, in turn, would be expected to be seen as an increase in the NIR muscle deoxygenation signal (D[Hb]). That is, while interpretation of the D[HbO 2 ] signal is complicated by its dependence on changes in perfusion to the field of NIR interrogation (which therefore influences delivery of HbO 2 to the field (Ferrari et al. 1997)), the D[Hb] signal is dependent on changes in O 2 extraction and normally is essentially unaffected by perfusion (or increases in arterial blood volume in the field of NIR interrogation) as arterial levels of [Hb] are trivial at normal arterial PJ levels (De Blasi et al. 1994). The aim of this study, therefore, was to examine leg muscle oxygenation (and thus the balance between muscle perfusion and muscle oxygen consumption) during cycle ergometry at both moderate and heavy exercise intensities during periods of increased respiratory muscle loading (by application of an external resistive load) using NIR spectroscopy of the quadriceps muscle group. We hypothesised that, during both the moderate- and heavy-intensity exercise, leg muscle deoxygenation (D[Hb]) would follow its normal contour (i.e. no additional deoxygenation) during the resistive breathing because the cardiac output, not being limiting, would increase to provide the compensatory flow for the required increases in respiratory muscle work. METHODS Subjects Six healthy male subjects (mean age (± S.D.) 29 ± 16 years; range years) participated in this study; all were non-smokers. The protocol was approved by the Institutional Ethics Committee for Human Experimentation and informed written consent was provided by each of the subjects after the experimental protocol and possible risks associated with participation in the study had been explained, in accordance with the Declaration of Helsinki. Exercise protocol Subjects reported to the laboratory at approximately the same time on each of four testing days. Involvement in other studies in our laboratory within the previous year had established each subject s peak O 2 uptake ( J,peak, 4.4 ± 0.2l min _1 ) and estimated lactate threshold (U L, 2.6 ± 0.4 l min _1 ) using a conventional cluster of pulmonary gas exchange indices (Whipp et al. 1986). On arriving at the laboratory, the subjects were seated comfortably while the NIR optodes were positioned and secured on the leg (see Near-infrared spectroscopy ). The laser diodes were then activated and approximately 15 min was allowed to ensure that the diodes had established an optimal working temperature before beginning the exercise protocol. During this time, the subject remained seated. The subject then moved to the cycle ergometer (Excalibur Sport, Lode, Groningen, The Netherlands) and was instrumented for pulmonary gas exchange and heart rate measurements. The NIR system was initialised while the subject rested quietly on the ergometer with the legs relaxed and still. When a steady baseline signal was achieved, the system was again zeroed and all subsequent changes in NIR signal intensity were made relative to this resting value. The NIR signal was monitored continuously for the remainder of the study, with no further adjustments being made to the NIR system zero value. The exercise protocol consisted of 10 min of constantload moderate- (~90 % U L ; 140 ± 21 W) and heavy-intensity exercise (244 ± 20 W), each preceded and followed by a baseline of 20 W cycling. The work rate assigned for the heavy-intensity exercise was estimated to lead to fatigue in min (actual time to fatigue, 11.9 ± 1.4 min). At the completion of the exercise

3 Exp. Physiol Resistive breathing and leg oxygenation 603 protocol, the subject moved from the cycle ergometer and sat comfortably in a chair while the NIR signals continued to be monitored. Two maximal voluntary contractions (MVCs) of the quadriceps-hamstring muscle groups were then performed, each separated by approximately 5 10 min resting recovery, to provide a physiological calibration for the D[Hb] signal (Chance et al. 1992; Belardinelli et al. 1995) (see Near-infrared spectroscopy ). Subjects were studied during four conditions at both the moderate- and heavy-intensity work rates. (i) In the control (Con) condition, there were no experimental interventions and breathing was spontaneous throughout the exercise protocol. (ii) To increase the work of breathing, a 3 4 min period of resistive breathing (Resist Br) was initiated 4 min after exercise onset. The resistor was positioned in the breathing apparatus immediately distal to the volume turbine and consisted of a Perspex block with a narrowed opening (i.d. 10 mm) designed to produce a resistance of 6.5 cmh 2 O l _1 s _1 at a flow of 3 l s _1. The resistance was both introduced and subsequently removed abruptly at end-expiration and without warning to the subject; the change in resistance occurred within one breath. (iii) To ensure that the D[Hb] signal was not maximised during the heavy-intensity exercise trials and that any changes in muscle oxygenation due to blood flow changes could therefore be detected by NIR spectroscopy, a period of partial blood flow restriction was included. To reduce muscle perfusion to the leg musculature, blood flow to the right leg was partially occluded for a 3 4 min period (Leg Occl) 4 min after exercise onset. A pressure cuff was wrapped around the right thigh, proximal to the NIR optodes, and inflated to a pressure of ~90 mmhg. In this instance, the NIR signal should be interpreted with caution as an increase in D[Hb] signal could reflect an increase in either O 2 extraction (due to a reduction in muscle perfusion) or venous (i.e. deoxygenated) volume, or a combination of both. However, an increase in D[Hb] signal combined with a decrease in D[HbO 2 ] and no change in total [Hb] (D[HbT]) most likely reflects greater muscle O 2 extraction (Ferrari et al. 1997). (iv) To ensure that the ventilatory profile for the moderate and heavy exercise tests would be the same in each of the four conditions, even during the period of resistive breathing, subjects volitionally controlled both their tidal volume (V T ) and breathing frequency (f R ) (Con Br) at the levels achieved during spontaneous breathing in the control condition. Controlled breathing was initiated 3 min after exercise onset and continued until the end of each exercise period. The values for V T and f R assigned during the period of controlled breathing represented the average values for corresponding 1 min periods of spontaneous breathing in the control condition. Inspired and expired volumes were visually displayed on an oscilloscope (located, in clear view, directly in front of the subject), and f R was presented audibly by a metronome. The periods of resistive breathing and partial leg occlusion always occurred within a period of controlled breathing. As a consequence, the conditions were assigned in order of Con and Con Br, with Resist Br and Leg Occl randomly assigned as the final two conditions. Near-infrared spectroscopy Local muscle oxygenation profiles were monitored continuously using NIR spectroscopy (Hamamatsu NIRO 500, Hamamatsu Photonics KK, Japan). The transmitting and receiving optodes were positioned on the vastus lateralis muscle of the dominant leg, at mid-thigh level and parallel with the long axis of the muscle. The optodes were housed in an optically dense plastic holder, thus ensuring that the position of the optodes, relative to each other, was fixed and invariant. The optode assembly was secured on the skin surface with tape, and then covered by an optically dense, black nylon sleeve, thus minimising the intrusion of extraneous light and loss of transmitted NIR light from the field of interrogation. The thigh, with attached optodes and covering, was wrapped with an elasticised cloth, to minimise movement of the optodes while still permitting freedom of movement for cycling. This preparation essentially prevented any optode movement relative to the skin surface. The theory of NIR spectroscopy and operating characteristics of the NIRO 500 spectrometer are described in detail by Elwell (1995). Briefly, one fibre optic bundle carried the NIR light produced by laser diodes to the tissue of interest while a second fibre optic bundle returned the transmitted light from the tissue to a photon detector (photomultiplier tube, PMT) in the spectrometer. Four different wavelength laser diodes (776, 826, 845 and 905 nm) provided the light source. The diodes were pulsed in rapid succession and the light detected by the PMT. The use of four laser diodes enables more chromophores to be detected and also increases the sensitivity of the instrument, thus providing an advantage of the NIRO 500 over other simpler NIR detection systems (Chance et al. 1992; Mancini et al. 1994; Belardinelli et al. 1995). The intensity of incident and transmitted light was recorded continuously and, along with the relevant specific extinction coefficients and optical pathlength, used for on-line estimation and display of the changes from the resting baseline of [HbO 2 ] (D[HbO 2 ]), [Hb] (D[Hb]) and total [Hb] (D[HbT] = D[HbO 2 ]+D[Hb]) (see Elwell, 1995, for details). The raw attenuation signal (in OD units) was transferred to a computer and stored for further analysis. In the present study, the interoptode spacing was 4 cm. While values for differential pathlength factors (DPF) in muscle have been published for calf and forearm (see Table 1.1 in Elwell, 1995, and van der Zee et al. 1992; Duncan et al. 1995), there are presently no published values for the quadriceps muscle. In the present study we used a value for DPF of 3.83 (personal communication, Professor D. T. Delpy, Dept Medical Physics and Bioengineering, University College London, London, UK). The D[Hb] signal was used to estimate changes in intramuscular oxygenation status in the field of interrogation, independent of blood volume changes (De Blasi et al. 1994; Ferrari et al. 1997). To provide equivalence for D[Hb] signal sensitivity across experiments (i.e. allowing for minor differences arising from differences in optode placement and/or day-to-day variability in signal intensity), the D[Hb] profiles were normalised between the minimum value (min) measured during the protocol and the maximum value (max) achieved at the end of heavy-intensity exercise or at the peak of the MVC, whichever was higher, according to the following equation: S norm =(S t _ S min )(S max _ S min ) _1 w 100, where S norm is the normalised signal value (%), S t is the signal value at time t, S min is the signal minimum value and S max is the signal value at the end of heavy-intensity exercise or during the MVC manoeuvres, whichever was higher. Arterial blood oxyhaemoglobin saturation was determined noninvasively using pulse oximetry of the finger (Biox 3740, Ohmeda, CO, USA). Gas exchange Ventilation ( E ), breathing pattern (V T, f R ), gas exchange (O 2 uptake, J; CO 2 output, CO2 ) and end-tidal partial pressures for O 2 and CO 2 (P ET,O2 and P ET,CO2, respectively) were measured breath-by-breath throughout each of the exercise protocols. The subjects breathed through a mouthpiece connected to a lowdeadspace (90 ml) low-resistance (< 1.5 cmh 2 O at 3 l s _1 ) turbine volume transducer (Interface Associates, Irvine, CA, USA) for

4 604 J. M. Kowalchuk, H. B. Rossiter, S. A. Ward and B. J. Whipp Exp. Physiol Table 1. Summary of power output (PO) and J values for the submaximal constant-load and incremental exercise tests PO (W) J (l min _1 ) Subject Peak Moderate Heavy U L Peak Moderate Heavy (< U L ) (> U L ) Mean (± S.D.) (42) (21) (34) (0.5) (0.65) (0.29) (0.64) measurement of inspiratory and expiratory volumes and flow; the volume turbine was calibrated prior to each test using a syringe of known volume (3.0 l). Respired gas was sampled continuously from the mouthpiece (1 ml s _1 ) and analysed for the fractional concentrations of O 2, CO 2 and N 2 by mass spectrometry (QP9000, Morgan Medical, Gillingham, UK); the mass spectrometer was calibrated prior to each test using two precision-analysed gas mixtures. The time delay between the volume and gas concentration signals was measured by passing a bolus of gas through the system (Beaver et al. 1973). Following analog-to-digital conversion, the electrical signals from these devices were sampled every 20 ms and processed online by digitial computer for computation and display, breathby-breath, of pulmonary gas exchange variables. The calibration and validation procedures have been described previously (Beaver et al. 1981). Statistical analysis The effect of the experimental interventions on the D[Hb] profile was summarised by comparing the data averaged over a 30 s period at times corresponding to the period immediately before the start of the intervention (Pre); the period at the end of the intervention (Mid); and a period beginning 15 s after removal of the intervention (Post). A two-way analysis of variance for repeated measures was performed with main effects of condition and time. A significant F ratio was further analysed using Student- Neuman-Keuls post hoc analysis. Statistical significance was accepted at P < All values are reported as the mean ± S.D. Figure 1 Profile of expired volume (V E ; top panel), respired P CO2 and PJ (middle panels), and mouth pressure (P M ; bottom panel) during the transition to a period of resistive breathing (at arrow) in heavy-intensity exercise. exp, expiration; insp; inspiration.

5 Exp. Physiol Resistive breathing and leg oxygenation 605 RESULTS An on-line trace showing expired volume (V E ), respired partial pressures for CO 2 (P CO2 ) and O 2 (PJ), and mouth pressure (P M ) for a representative subject prior to and during the period of resistive breathing in heavy-intensity exercise is presented in Fig. 1. Note the constancy of V E (V T ; f R ) and gas exchange variables (end-tidal P CO2 and PJ) despite an ~4-fold increase in P M. Ventilation and gas exchange The breath-by-breath responses of E, f R and V T during a single trial of the exercise protocol in the Con and Resist Br conditions are presented in Fig. 2 for a representative subject. E, f R and V T were similar between the two conditions for both moderate- and heavy-intensity exercise, despite the progressive rise in E that characterised heavy-intensity exercise. In general, the control breathing pattern responses (V T ; f R ) for both moderate- and heavyintensity exercise were successfully reproduced across all other conditions (e.g. Fig. 2). As a result, the end-tidal gas tension profiles were similarly unaffected (Fig. 1), as was, presumably, alveolar ventilation. Patterns for breath-by-breath J and heart rate were also similar across conditions (Fig. 2). End-exercise J averaged 2.23 ± 0.29 l min _1 (i.e. 50 ± 5% J,peak ) and 3.90 ± 0.64 l min _1 (i.e. 94 ± 7% J,peak ) for moderate-and heavyintensity exercise, respectively, while end-exercise heart rate averaged 127 ± 7 and 186 ± 6 beats min _1. During heavy-intensity exercise, J and heart rate increased throughout the exercise without any differences between conditions. These increases were reflected in the average values measured during the periods immediately before (Pre) and following (Post) the experimental interventions. For J the Pre and Post values were, respectively (l min _1 ): Con: 3.53 ± 0.26 (Pre) and 3.90 ± 0.33 (Post); Con Br: 3.48 ± 0.32 and 3.79 ± 0.33; Resist Br: 3.58 ± 0.34 and 4.12± 0.49; Leg Occl: 3.68 ± 0.36 and 4.22± The corresponding values for heart rate were, respectively (beats min _1 ): Con: 168 ± 12 (Pre) and 180 ± 10 (Post); Con Br: 164 ± 10 and 175 ± 8; Resist Br: 168 ± 9 and 182± 8; Leg Occl: 170 ± 11 and 183 ± 8. Importantly, there were no further increases in J or heart rate during the periods in which resistive breathing or leg occlusion were imposed. Figure 2 Profiles for pulmonary O 2 uptake ( J; top panel), heart rate (HR, beats min _1 ; second panel), expired ventilation ( E ; middle panel), breathing frequency (f R, breaths min _1 ; fourth panel) and tidal volume (V T ; bottom panel) during moderate- (Ex MOD ) and heavy-intensity exercise (Ex HVY ) in a representative subject in control (A) and resistive breathing (B) conditions. Dotted vertical lines represent the period of constantload exercise. Horizontal bars represent the period of controlled breathing (CB) and resistive breathing (RB).

6 606 J. M. Kowalchuk, H. B. Rossiter, S. A. Ward and B. J. Whipp Exp. Physiol Near-infrared spectroscopy Leg oxygenation profiles for a representative subject during a single transition to constant-load moderate- and heavy-intensity exercise during the Con and Resist Br conditions are presented in Fig. 3. In general, for all subjects, the deoxygenated Hb signal increased rapidly and without discernible delay at the onset of both moderateand heavy-intensity exercise, and approached a plateau during moderate exercise but continued to rise throughout heavy exercise (Fig. 3; Table 2). The increase in D[Hb] from baseline was greater during heavy- than for moderate-intensity exercise (Figs 3 5; Table 2); the endexercise value represented 90 ± 6 % and 57±17 % of maximal muscle deoxygenation (i.e. 100 % D[Hb]) during heavy- and moderate-intensity exercise, respectively. Generally, D[HbO 2 ] demonstrated a rapid but transient, intensity-dependent decrease at exercise onset which was followed by an increase which continued throughout moderate exercise but showed a further sustained reduction during heavy-intensity exercise (Fig. 3). Similarly, D[HbT] decreased transiently at the onset of moderate and heavy exercise. This was followed by an increase in D[HbT] which, in moderate exercise, tended to continue throughout exercise, while in heavy exercise, the increase in D[HbT] was only transient and was followed by a gradual decrease for the remainder of the exercise (Fig. 3). In general, the profile of D[HbO 2 ] appeared qualitatively similar to that of the D[HbT] profile, reflecting a dependency of the D[HbO 2 ] signal on the total Hb within the field of NIR interrogation. In contrast, the D[Hb] did not demonstrate this dependency (Fig. 3) and thus we chose to report only the changes in the D[Hb] signal. Also shown in Fig. 3 are the changes in NIR profiles associated with the two MVCs performed at the end of each of the exercise periods, demonstrating the good reproducibility that was typical also of the other subjects. While it had been expected that peak muscle deoxygenation would be seen during an MVC (Chance et al. 1992), this occurred in only 48 % of the trials; in the remaining trials, peak values were seen either at the end of heavy exercise (24 %) or during the period of partial occlusion of leg blood flow (28 %). As a consequence, at the end of the heavy exercise bout, the D[Hb] represented 86 ± 5 %, 97 ± 1 %, and 91 ± 5 % of observed peak muscle deoxygenation for those trials where deoxygenation was greatest during the MVC, at end-exercise, and during partial leg occlusion, respectively. The normalised profiles for the D[Hb] signal for a representative subject during single transitions to constantload moderate- and heavy-intensity exercise during each of the four experimental conditions are shown in Fig. 4. In general, the profiles for the single transitions were similar between conditions, demonstrating the good reproducibility of the response even on single transitions. As seen in this Figure 3 Profiles for changes in deoxyhaemoglobin concentration (D[Hb]), oxyhaemoglobin concentration (D[HbO 2 ]), and total haemoglobin concentration (D[HbT]) during moderate- (Ex MOD ) and heavyintensity exercise (Ex HVY ) in a representative subject in control (A) and resistive breathing (B) conditions. Dotted vertical lines represent the period of constant-load exercise. Horizontal bars represent the period of controlled breathing (CB) and resistive breathing (RB). The near-infrared (NIR) profile changes for each of the two maximal voluntary contractions (MVC) are shown at the end of each of the traces.

7 Exp. Physiol Resistive breathing and leg oxygenation 607 Table 2. Change in concentration of deoxy-hb (D[Hb]) during heavy-intensity exercise in control (Con), controlled breathing (Con Br), resistive breathing (Resist Br) and leg occlusion (Leg Occl) D[Hb] (mm) D[Hb] (%) Pre* Mid* Post* Pre* Mid* Post* End Ex* Con 28.3 a, b 30.9 a 31.7 b 80.1 a, b 82.3 c, d 86.1 a, c 88.1 b, d (8.6) (8.4) (8.5) (9.8) (7.8) (8.2) (8.2) Con Br a, b, c 89.7 a 90.2 b 91.7 c (16.5) (16.9) (17.4) (9.3) (7.5) (6.8) (5.7) Resist Br (11.9) (12.7) (12.8) (11.9) (11.8) (11.8) (6.8) Leg Occl 34.3 a 39.6 a 36.6 a 84.6 a 95.3 a, b, c 88.9 b 89.4 c (13.0) (13.6) (13.9) (6.9) (4.3) (7.0) (5.2) Values are mean (± S.D.). * NIR values were determined as an average for a period corresponding to the experimental intervention: a 30 s period immediately prior to the onset of the intervention (Pre); a 30 s period at the end of the intervention (Mid); and a 30 s period starting 15 s after removal of the intervention (Post); a 15 s period immediately preceding end-exercise (End Ex). a, b, c, d Values with the same symbols are significantly different (P < 0.05) within a condition. subject and summarised for all subjects in Table 2, there was no additional increase in D[Hb] signal during the period of resistive breathing, the relative intensity of which corresponded to a J of 93 ± 3 % of that attained at endexercise (or 90 ± 8 % peak J as determined during the initial ramp exercise test to fatigue). During heavy exercise, the normalised D[Hb] signal was 87 ± 12% during, and 87 ± 12% immediately after, the period of resistive breathing. Partial leg occlusion (Fig. 4; Table 2) was associated with a further increase in D[Hb] and decrease in D[HbO 2 ], consistent with increased muscle O 2 extraction. Importantly, during partial leg occlusion D[HbT] increased (P < 0.05) only during moderate exercise (Fig. 5). For heavy exercise, the normalised D[Hb] was greater (P < 0.05) during (95 ± 4 %) compared to immediately before (85 ± 7 %) or after (89 ± 7 %) the period of partial leg occlusion. For these same periods, the normalised D[HbO 2 ] was lower (P < 0.05) Figure 4 Profiles for changes in deoxyhaemoglobin concentration (D[Hb]) during moderateand heavy-intensity exercise in a representative subject in control (A), controlled breathing (B), resistive breathing (C), and partial leg occlusion (D) conditions. Dotted vertical lines represent the period of constant-load exercise. Note that this is the same subject as in Fig. 3. Horizontal bars represent the periods of resistive breathing (C) and partial leg occlusion (D). The increase in D[Hb] during each of the two maximal voluntary contractions is shown at the end of each of the traces.

8 608 J. M. Kowalchuk, H. B. Rossiter, S. A. Ward and B. J. Whipp Exp. Physiol during (30 ± 18 %) compared to before (39 ± 15 %) and after (35 ± 13 %), while the normalised D[HbT] was similar during each time period (Mid, 71 ± 12%; Pre, 71 ± 21 %; Post, 72± 12%). DISCUSSION This study demonstrated that during constant-load moderate- (~90 % U L or ~50 % J,peak ) and heavy-intensity (~94 % J,peak ) cycling, leg muscle deoxygenation (as reflected by an increase in the NIR D[Hb] signal) increased in an intensity-dependent manner. Also, this study is unique in that it used a period of combined inspiratory and expiratory resistive breathing to increase the work of breathing during exercise and showed that increases in active muscle mass had no discernible effect on leg muscle Figure 5 Profiles for changes in deoxyhaemoglobin concentration (D[Hb]), oxyhaemoglobin concentration (D[HbO 2 ]), and total haemoglobin concentration (D[HbT]) during moderate- (Ex MOD ) and heavy-intensity exercise (Ex HVY ) in a representative subject in the partial leg occlusion condition. Note that this is the same subject as in Figs 3 and 4. Dotted vertical lines represent the period of constant-load exercise. Horizontal bars represent the period of controlled breathing (CB) and partial leg occlusion (PC). The NIR profile changes for each of the two maximal voluntary contractions (MVC) are shown at the end of each of the traces. oxygenation as assessed by NIR spectroscopy. Furthermore, we demonstrated that this lack of change in oxygenation status could not simply be ascribed to the muscle O 2 extraction already being maximal, as partial occlusion of thigh vasculature was consistently associated with an increase in the NIR D[Hb] signal (and decrease in D[HbO 2 ]), implying further muscle O 2 extraction (Ferrari et al. 1997). Our observations in moderate-intensity exercise are similar to those of Nielsen et al. (2001), who observed greater deoxygenation of the leg muscle on going from rest to leg cycling exercise (at 150 W) but no further change in deoxygenation with increases in work of breathing until the diameter of their resistance device had decreased to 4.5 mm (compared with 10 mm in the present study). In spite of the need to recruit both inspiratory and expiratory muscles to support the increased work of breathing during heavy-intensity exercise, our data differ from those of Harms et al. (1997, 1998) during maximal intensity exercise that led to fatigue in 2 3 min, and suggest that when the exercise intensity does not demand a maximal J during the exercise, resistive breathing is not associated with a compensatory increase in leg muscle O 2 extraction as a consequence of a reduction in leg blood flow (i.e. a respiratory steal phenomenon). Also, in contrast to the findings of Secher et al. (1977), but consistent with the results of Richardson et al. (1995), Richter et al. (1992), Savard et al. (1989), Strange (1999), and most recently Wetter et al. (1999), our data suggest that when exercise intensity, and thus cardiac output and muscle O 2 extraction, is not maximal, blood flow to active muscle is not reduced as a result of increases in active muscle mass that occur during a period of resistive breathing. Near-infrared spectroscopy Near-infrared spectroscopy provides a non-invasive means of monitoring muscle oxygenation, and has been used to study muscle oxygenation during exercise in healthy subjects and in patients with disease (Chance et al. 1992; McCully et al. 1994; Mancini et al. 1994; Belardinelli et al. 1995; Hamaoka et al. 1996). Specifically, the NIR signals provide information primarily on changes in oxy- and deoxyhaemoglobin of arteriolar, capillary and venular blood and, to a lesser extent, myoglobin in muscle (Wang et al. 1990; Chance et al. 1992). As light is almost completely absorbed by large blood vessels, changes in NIR signals have been argued to reflect mainly changes in oxygenation in capillaries and small blood vessels (Chance et al. 1992). Changes in skin blood flow have been shown to contribute minimally to the NIR signals measured in muscle (Mancini et al. 1994), possibly as a consequence of the small volume of skin relative to muscle volume sampled by the NIR probes (Mancini et al. 1994). In the present study, the similarity in the NIR response profiles of D[HbO 2 ] and D[HbT] (e.g. Fig. 3) suggests that the D[HbO 2 ] signal is influenced by the total amount of Hb (predominantly oxygenated Hb) in the field of NIR interrogation, and thus can be influenced by changes in blood flow, haemoconcentration, or both. Specifically,

9 Exp. Physiol Resistive breathing and leg oxygenation 609 during moderate exercise, the similar profile of increase in both D[HbO 2 ] and D[HbT] may be related to an expansion of microvascular blood volume due to capillary recruitment at a relatively constant O 2 extraction, as reflected by a relatively constant D[Hb]. During heavy exercise, D[HbT] increased only transiently before decreasing which most likely represents tonic restraint (i.e. by sympathetic vasoconstriction or mechanical restriction) to further increases in local blood volume (despite higher flow rates). However, of this existing blood volume, there is a greater D[Hb] and lower D[HbO 2 ], presumably due to higher O 2 extraction. While it is possible that the D[Hb] signal might be influenced by a blood volume-dependent increase in [Hb] resulting from venular vasodilation or venous volume expansion, the influence of this effect is likely to be minimised by the intermittent mechanical compression of the muscle vasculature that occurs during the contraction phase of the exercise duty cycle. Also, during the period of cuff inflation, when an increase in venular volume is most likely, the D[HbT] was unchanged suggesting that blood volume did not change appreciably during that period. As a consequence, we chose to report only the D[Hb] response profiles. The profiles for D[Hb] for each of the four conditions were found to be highly reproducible, even though they derived from four separate exercise protocols performed on different days (Figs 3 and 4). When placing the NIR transmitting and receiving optodes over the leg muscle prior to exercise, care was taken to secure the optode assembly to the skin surface to minimise movement of the optodes relative the skin surface. However, movement of the underlying tissues relative to the optode probes is unavoidable during dynamic exercise; the extent to which this might influence pathlength or scattering of the photons is unknown. While we recognise these as possible limitations to this technique, we believe that their influence on the interpretation of the data is small for the following reasons: (i) optode placement and cycling cadence were similar between conditions; (ii) comparisons were made relative to events occurring immediately before and after the experimental intervention; (iii) each subject performed all conditions and thus acted as his own control; and (iv) the metabolic response to the exercise and conditions within the muscle are assumed to be similar for each of the conditions, as reflected by the similarity of the J response between conditions. Resistive breathing and muscle oxygenation During maximal leg cycling exercise, Harms et al. (1997) showed that leg blood flow was related to the work of breathing in an inverse and curvilinear manner, with the relative reduction in leg blood flow during inspiratoryonly resistive breathing (i.e. 3 7 cmh 2 Ol _1 s _1 ) being greater than the increase associated with assisted breathing using a proportional-assist ventilator. These authors considered the intensity of exercise to be maximal, as it was sustained for only min. Furthermore, cardiac output and both whole-body and leg arterio-venous O 2 differences were regarded as being maximal, as no additional increase in these variables was seen during the period of resistive breathing (Harms et al. 1997, 1998). As a consequence of the reduction in muscle blood flow during resistive breathing, leg muscle O 2 consumption was reduced but with no effect on pulmonary J (Harms et al. 1997, 1998). In contrast, during cycling exercise at 50 % and 75 % J,max, Wetter et al. (1999) reported that increasing the work of breathing by resistive loading was associated with an increase in pulmonary J, while assisted breathing was associated with a decrease in pulmonary J without any significant changes in leg vascular resistance or leg muscle blood flow. Secher et al. (1977) demonstrated that recruiting additional muscle mass during submaximal leg cycling exercise ( J legs only = 2.7 l min _1 or 67 % of J,max ) by simultaneous arm cranking ( J arms = 1.4 l min _1 or 44 % of J,max ) also reduced blood flow to already active leg muscle. Thus, for the combined arm leg exercise ( J= 3.4 l min _1 or 78 % of J,max ), leg blood flow was reduced by ~15 % (from 12.4 to 10.5 l min _1 ) and the arterio-venous O 2 difference across the exercising leg muscle increased by ~13 % (from 156 to 176 ml l _1 ) to compensate for the reduced flow, with no change in pulmonary or leg J (Secher et al. 1977). However, several other studies using similar exercise protocols with exercise intensities ranging between 24 and 82 % J,max have been unable to confirm these findings (Savard et al. 1989; Richter et al. 1992; Richardson et al. 1995). These differences may reflect the increased noradrenaline spillover in these latter studies, the associated lower leg vascular conductance suggesting that vasoconstriction had constrained further increases in leg blood flow (Savard et al. 1989; Richter et al. 1992). In the present study the use of a combined inspiratory expiratory resistance (7 cmh 2 Ol _1 s _1 ), thereby requiring recruitment of both inspiratory and expiratory muscles, presumably induced a greater active (respiratory) muscle mass than in the studies of Harms, Dempsey and colleagues (Harms et al. 1997, 1998; Wetter et al. 1999). The power output and thus J for the heavy-intensity exercise used in the present study was lower than that used by Harms et al. (1997; 1998), but higher than that used by Secher et al. (1977) and others (Savard et al. 1989; Richter et al. 1992; Richardson et al. 1995), and could be tolerated for ~12± 1 min before fatigue. Consistent with the supra-u L intensity, the associated J response in the present study increased slowly and progressively throughout the exercise period (i.e. J slow component: Whipp, 1994; Gaesser & Poole, 1996) with no differences between conditions or as a consequence of the experimental interventions. Although the J achieved near-maximal levels by end-exercise (94 ± 7 % of peak J), at the time corresponding to the period of resistive breathing J represented only 90 ± 8 % of peak J (or 93 ± 3 % end-exercise J). Thus, in contrast to the studies of Harms et al. (1997, 1998) where resistive breathing was introduced during exercise which achieved maximum J, in the present study resistive breathing was added during heavy, but not yet a maximal exercise. The important feature of our results was that, during the period of resistive breathing, there was no additional increase in

10 610 J. M. Kowalchuk, H. B. Rossiter, S. A. Ward and B. J. Whipp Exp. Physiol the NIR D[Hb] signal (or decrease in D[HbO 2 ]) implying that muscle O 2 extraction did not increase further during this period. Also, in the present study the increase in heart rate during heavy- (and moderate-) intensity exercise was similar in each of the conditions, with no further increase in heart rate observed during the period of resistive breathing. If leg muscle blood flow had been reduced during the period of resistive breathing and assuming that leg O 2 consumption was not changed, an increase in the NIR D[Hb] signal would have been expected; i.e. reflecting a compensatory increase in muscle O 2 extraction following a reduction in muscle perfusion. This was clearly shown in the present study: the NIR D[Hb] signal increased (and D[HbO 2 ] decreased with no change in D[HbT]) when leg blood flow was intentionally reduced by partial leg occlusion with the inflation of a pressure cuff (Fig. 5). Limb muscle sympathetic nerve activity (MSNA) or noradrenaline spillover were not measured in this study but would be expected to increase if limb blood flow were reduced by a respiratory steal phenomenon (Harms et al. 1997). Increases in MSNA in an inactive limb have been demonstrated during fatiguing diaphragmatic contractions induced by addition of either inspiratory (St Croix et al. 2000; Sheel et al. 2001) or expiratory (Derchak et al. 2002) resistive loads. In these studies (St Croix et al. 2000; Sheel et al. 2001; Derchak et al. 2002), limb MSNA increased in a time-dependent manner following the application of the resistive load, and was maximal within 2 3 min, and thus, the duration for resistive breathing used in the present study (i.e. 3 4 min) should have provided sufficient time for limb MSNA to increase. That a decrease in limb blood flow and thus increase in NIR D[Hb] signal (and decrease in D[HbO 2 ]) and/or increase in heart rate (and presumably cardiac output) did not occur with resistive breathing in the present study suggests one of the following: (i) that the combined inspiratory expiratory resistance and increased work of breathing were not severe enough to activate an intercostal diaphragmatic metaboreflex (St Croix et al. 2000; Sheel et al. 2001; Derchak et al. 2002) and/or elicit intercostal diaphragmatic fatigue (St Croix et al. 2000; Sheel et al. 2001; Derchak et al. 2002), (ii) that any limitation on cardiac output is presumably not severe enough at all but the highest work rates, or (iii) that redistribution of blood flow occurred from regions other than active muscle or from regions within active leg muscle that were not interrogated by the NIR signal. Therefore, the findings of the present study demonstrate that recruitment of additional working muscle mass by means of resistive breathing performed during heavy, but submaximal, exercise is not associated with a reduction in blood flow to the working leg muscles, at least up to work rates corresponding to ~90 95 % J,peak. This suggests that with recruitment of additional muscle mass during submaximal exercise, blood flow requirements for the newly recruited muscles are met by either increases in cardiac output, redistribution of blood flow from non-working tissues, or both, without the need to steal blood flow from other active muscles. As a further increase in heart rate, which would normally accompany an increase in cardiac output, was not observed during the period of resistive breathing, we speculate that blood flow redistribution rather than increases in cardiac output were important in providing blood flow to the newly recruited respiratory muscles during resistive breathing in this study. BEAVER, W. L., LAMARRA, N. & WASSERMAN, K. (1981). Breath-bybreath measurement of true alveolar gas exchange. Journal of Applied Physiology: Respiration, Environmental and Exercise Physiology 51, BEAVER, W. L., WASSERMAN, K. & WHIPP, B. J. (1973). On-line computer analysis and breath-by-breath graphical display of exercise function tests. Journal of Applied Physiology 34, BELARDINELLI, R., BARSTOW, T. J., PORSZASZ, J. & WASSERMAN, K. (1995). Skeletal muscle oxygenation during constant work rate exercise. Medicine and Science in Sports and Exercise 27, CHANCE, B., DAIT, M., ZHANG, C., HAMAOKA, T. & HAGERMAN, F. (1992). Recovery from exercise-induced desaturation in the quadriceps muscles of elite competitive rowers. American Journal of Physiology 262, C DE BLASI, R. A., FERRARI, M., NATALI, A., CONTI, G., MEGA, A. & GASPARETTO, A. (1994). Noninvasive measurement of forearm blood flow and oxygen consumption by near-infrared spectroscopy. Journal of Applied Physiology 76, DERCHAK, P. A., SHEEL, A. W., MORGAN, B. J. & DEMPSEY, J. A. (2002). Effects of expiratory muscle work on muscle sympathetic nerve activity. Journal of Applied Physiology 92, DUNCAN, A., MEEK, J. H., CLEMENCE, M., ELWELL, C. E., TYSZCZUK, L., COPE, M. & DELPY, D. T. (1995). Optical pathlength measurements on adult head, calf and forearm and the head of the newborn infant using phase resolved optical spectroscopy. Physics in Medicine and Biology 40, ELWELL, C. E. (1995). A Practical Users Guide to Near Infrared Spectroscopy, pp Hamamatsu Photonics KK, London. FERRARI, M., BINZONI, T. & QUARESIMA, V. (1997). Oxidative metabolism in muscle. Philosophical Transactions of the Royal Society of London B 352, GAESSER, G. A. & POOLE, D. C. (1996). The slow component of oxygen uptake kinetics in humans. Exercise and Sport Sciences Reviews 24, HAMAOKA, T., IWANE, H., SHIMOMITSU, T., KATSUMURA, T., MURASE, N., NISHIO, S., OSADA, T., KUROSAWA, Y. & CHANCE, B. (1996). Noninvasive measures of oxidative metabolism on working human muscle by near-infrared spectroscopy. Journal of Applied Physiology 81, HARMS, C. A., BABCOCK, M. A., MCCLARAN, S. R., PEGELOW, D. F., NICKELE, G. A., NELSON, W. B. & DEMPSEY, J. A. (1997). Respiratory muscle work compromises leg blood flow during maximal exercise. Journal of Applied Physiology 82, HARMS, C. A., WETTER, T. J., MCCLARAN, S. R., PEGELOW, D. F., NICKELE, G. A., NELSON, W. B., HANSON, P. & DEMPSEY, J. A. (1998). Effects of respiratory muscle work on cardiac output and its distribution during maximal exercise. Journal of Applied Physiology 85, MCCULLY, K. K., IOTTI, S., KENDRICK, K., WANG, Z., POSNER, J. D., LEIGH, J. & CHANCE, B. (1994). Simultaneous in vivo measurements of HbO 2 saturation and PCr kinetics after exercise in normal humans. Journal of Applied Physiology 77, 5 10.

11 Exp. Physiol Resistive breathing and leg oxygenation 611 MANCINI, D. M., BOLINGER, L., LI, H., KENDRICK, K., CHANCE, B. & WILSON, J. R. (1994). Validation of near-infrared spectroscopy in humans. Journal of Applied Physiology 77, NIELSEN, H. B., BOESEN, M. & SECHER, N. H. (2001). Near-infrared spectroscopy determined brain and muscle oxygenation during exercise with normal and resistive breathing. Acta Physiologica Scandinavica 171, RICHARDSON, R. S., KENNEDY, B., KNIGHT, D. R. & WAGNER, P. D. (1995). High muscle blood flows are not attenuated by recruitment of additional muscle mass. American Journal of Physiology 269, H RICHTER, E. A., KIENS, B., HARGREAVES, M. & KJAER, M. (1992). Effect of arm-cranking on leg blood flow and noradrenaline spillover during leg exercise in man. Acta Physiologica Scandinavica 144, ST CROIX, C. M., MORGAN, B. J., WETTER, T. J. & DEMPSEY, J. A. (2000). Fatiguing inspiratory muscle work causes reflex sympathetic activation in humans. Journal of Physiology 529, SAVARD, G. K., RICHTER, E. A., STRANGE, S., KIENS, B., CHRISTENSEN, N. J. & SALTIN, B. (1989). Norepinephrine spillover from skeletal muscle during exercise in humans: role of muscle mass. American Journal of Physiology 257, H SECHER, N. H., CLAUSEN, J. P., KLAUSEN, K., NOER, I. & TRAP-JENSEN, J. (1977). Central and regional circulatory effects of adding arm exercise to leg exercise. Acta Physiologica Scandinavica 100, SHEEL, A. W., DERCHAK, P. A., MORGAN, B. J., PEGELOW, D. F., JACQUES, A. J. & DEMPSEY, J. A. (2001). Fatiguing inspiratory muscle work causes reflex reduction in resting leg blood flow in humans. Journal of Physiology 537, STRANGE, S. (1999). Cardiovascular control during concomitant dynamic leg exercise and static arm exercise in humans. Journal of Physiology 514, TRAN, T.-K., SAILASUTA, N., KREUTZER, U., HURD, R., CHUNG, Y., MOLÉ, P. A., KUNO, S. & JUE, T. (1999). Comparative analysis of NMR and NIRS measurements of intracellular PJ in human skeletal muscle. American Journal of Physiology 276, R1682 R1690. VAN DER ZEE, P., COPE, M., ARRIDGE, S. R., ESSENPREIS, M., POTTER, L. A., EDWARDS, A. D., WYATT, J. S., MCCORMICK, D. C., ROTH, S. C., REYNOLDS, E. O. R. & DELPY, D. T. (1992). Experimentally measured optical pathlengths for the adult head, calf and forearm and the head of the newborn infant as a function of inter optode spacing. Advances in Experimental Medicine and Biology 316, WANG, D.-J., WANG, Z., NOYSZEWSKI, E., NIOKA, S., HIRAO, K., CHENG-DU, T. & CHANCE, B. (1990). Correlation of optical and 1 H NMR of Hb and Mb deoxygenation in canine gastrocnemius. Magnetic Resonance in Medicine 1, 175 (Abstract). WETTER, T. J., HARMS, C. A., NELSON, W. B., PEGELOW, D. F. & DEMPSEY, J. A. (1999). Influence of respiratory muscle work on VO 2 and leg blood flow during submaximal exercise. Journal of Applied Physiology 87, WHIPP, B. J. (1994). The slow component of O 2 uptake kinetics during heavy exercise. Medicine and Science in Sports and Exercise 26, WHIPP, B. J., WARD, S. A. & WASSERMAN, K. (1986). Respiratory markers of the anaerobic threshold. Advances in Cardiology 33, Acknowledgements Financial support for this research was provided by the MRC (UK) grant no. G j and NSERC (Canada). The authors wish to thank the subjects for their time and commitment given to this study. We would also like to acknowledge the technical support provided by Ms Mandy Skasick.