Near-infrared spectroscopy (NIRS) was introduced by Jöbsis (15) in 1977, and it is a technique that has been extensively used to monitor muscle and brain oxygenation status noninvasively at the microvascular/tissue level. In particular, NIRS has been widely used in the areas of skeletal muscle physiology (6), sport, health, exercise, and medical sciences (12). The application of NIRS is based on the fact that near-infrared light, with wavelengths in the window of 650 to 900 nm, can penetrate tissues to a greater depth than visible light, where it is then absorbed or scattered in the tissue (26). The overall light absorbance is affected by the oxygenation status of chromophore compounds such as hemoglobin (Hb), myoglobin (Mb), and cytochrome c oxidase. However, the contribution of cytochrome c oxidase to the NIRS signals is estimated to be negligible, only approximately 2%–5% (1). Therefore, when light is emitted and received at two wavelengths, the differences in light absorbance at these two wavelengths can be used to determine the relative changes of oxygenated Hb and Mb (HbMbO2) and deoxygenated Hb and Mb (HHbMb) (e.g., [2,3,8,11]), or with some devices, absolute changes (e.g., ).
Given that venous O2 concentration represents the balance between O2 uptake and O2 supply and that NIRS should reveal tissue oxygenation status, one would expect a relationship between the two. The rationale for this can be ascertained from application of the Fick principle to tissue (muscle) metabolic rate as follows:
in which V˙O2 is the oxygen uptake per min, Q˙ is the blood flow per min, CaO2 is the arterial O2 concentration, and CvO2 is the venous O2 concentration. Further, we assume that venous O2Hb% is proportional to CvO2, where venous O2Hb% is the percent saturation of Hb with O2 in the venous blood. The question then is to what extent changes in tissue HHbMb or HbMbO2 reflect changes in CvO2 and venous O2Hb% under various conditions (13). For example, the correlation between NIRS-derived signals and venous O2Hb% has been investigated by comparing 1) jugular vein and brain measures (17), 2) blood samples drawn from a forearm vein to NIRS spectra during an exercise of lifting weight by depressing a lever in humans (23), and 3) electrically stimulated twitch contractions in canine gracilis muscle (29); high correlation coefficients were reported for these experiments. However, this high correlation has not been observed under all conditions (2,3,22), and the overall number of validation studies is small.
Therefore, the purpose of the present study in isolated, perfused canine skeletal muscle was to examine the relationship between NIRS signals and both CvO2 and venous O2Hb% under a wider variety of O2 delivery and V˙O2 conditions than has been previously evaluated. The model used allows greater confidence in aligning the venous drainage with the tissue being interrogated by NIRS and offers the potential in future experiments for distinguishing the Hb versus Mb contribution to the NIRS signal. Other advantages of this model include the absence of overlying skin and adipose tissue and the interrogation of a greater percentage of the total muscle volume. For these experiments, a continuous-wave NIRS instrument (Oxymon Mk III; Artinis Medical Systems BV, Zetten, Netherlands) was used to assess muscle oxygenation; this device provides relative changes in oxyhemoglobin/oxymyoglobin (HbMbO2), deoxyhemoglobin/deoxymyoglobin (HHbMb), and total hemoglobin/myoglobin (HbMbtot) in comparison with a biased baseline.
METHODS AND PROCEDURES
All experimental procedures performed in this study were approved by the Auburn University Institutional Care and Use Committee. Six female adult mongrel hounds were used. The dogs were housed at the Division of Laboratory Animal Health facility at the Auburn University Veterinary School, with access to food and water ad libitum. They were fasted for 24 h before experiments. For unknown reasons, the HbMbO2 for one of the dogs (dog 1) varied widely and unrealistically during trials 1 and 2. All other data for this animal, including HHbMb, were consistent with that of the other five dogs. However, none of this dog’s data have been included, and we report here only the results of the other five animals.
On each experimental day, one dog was anesthetized to a deep surgical plane of anesthesia. An initial bolus of sodium pentobarbital was given intravenously at 30 mg·kg−1 body weight via a prominent cephalic vein of a forelimb. This established a deep surgical plane, which was maintained throughout the experiments with additional doses (65–100 mg) of sodium pentobarbital given as necessary into an isolated jugular vein to maintain absence of pedal, palpebral, and corneal reflexes. An endotracheal tube was inserted, and the dogs were ventilated with a respirator (Respirator model 613; Harvard Apparatus, Holliston, MA). Rectal temperatures were maintained at 37°C with a heat lamp and heating pad. After surgical preparation, all dogs were treated with heparin in divided doses totaling 3000 U·kg−1. Ventilation was adjusted to a level that established and maintained normal arterial PO2, PCO2, and pH; initial settings were 20 mL·kg−1 body weight for tidal volume and a frequency of 15–20 breaths per minute.
The left gastrocnemius plus superficial digital flexor (GS) was surgically isolated as previously described in detail (10,14,27), resulting in a single arterial supply (popliteal artery) and single venous drainage (popliteal vein). The popliteal vein was cannulated, and a flow-through–type transit-time ultrasonic flow probe (6NRB440; Transonic Systems, Ithaca, NY) was placed to measure blood flow rate (Q˙). A reservoir was attached to a cannula in the left jugular vein, allowing blood draining from the popliteal vein to be returned to the animal. Either the right carotid artery or the right femoral artery was cannulated to route blood through a peristaltic pump (Gilson Minipuls 3) to the GS, thus allowing experimental control of muscle perfusion. A transducer (Model RP-1500; Narco Biosystems, Austin, TX) was inserted to measure perfusion pressure to the muscle. The distal GS tendons were connected to an isometric myograph via a load cell (Interface SM-250, Scottsdale, AZ). Two bone nails were used to fix the femur and tibia to the base of the myograph. The sciatic nerve was isolated and ligated, and the distal nerve stump was pulled into a small tubular electrode for stimulation. A saline-soaked gauze and a plastic sheet were used to cover exposed tissues to minimize drying and cooling. The GS was set to optimal length (Lo) before each contraction protocol. At the end of each experiment, the animal was euthanized by administering an overdose of sodium pentobarbital along with saturated potassium chloride.
The experiment consisted of five trials (trials 1, 2, 3, 4, and 5) and one pretrial. The protocols for trials 1, 2, 3, and 4 are outlined in the Supplemental Digital Content (see Figure, Supplemental Digital Content 1, Illustration of protocols for trials 1–4, https://links.lww.com/MSS/A706). Trials 1 and 2 were performed on resting muscles. For trials 1 and 2, Q˙ via the peristaltic pump was adjusted to maintain average control perfusion pressure in the range of 120–140 mm Hg. Trial 1 examined the influence of convective O2 delivery on the NIRS signals; the animal was respired on normoxic air throughout this trial. When the NIRS signals had stabilized, Q˙ was raised by increasing the pump speed by 50%. Again, when the NIRS signals were steady, the muscle Q˙ was brought back to the control level. Next, the flow was decreased by 50% until the NIRS signals stabilized again. Trial 2 examined the effects of varying inspired gas concentrations on the NIRS signals during constant blood flow. This trial began with the animal ventilated on normoxic air, and the control Q˙ was maintained throughout the trial. The inspired gas was then switched to 100% O2, back to normoxic air (≈21% O2), and then to hypoxic gas (12% O2); the NIRS signals were allowed to stabilize during each step.
After trials 1 and 2, a pretrial for trials 3 and 4 was performed. During this pretrial, the GS was first stimulated to contract tetanically (8 V, 50 Hz, 0.2-ms pulse, and 200-ms duration) at the rate of one contraction per 2 s (1/2 s). Q˙ was adjusted with the pump to maintain average perfusion pressure around 200 mm Hg; this pump setting was subsequently used in trial 3. Without stopping contractions, the stimulation rate was then increased to 2/3 s. The pump setting was again adjusted to provide a Q˙ that caused perfusion pressure to stabilize around 200 mm Hg; this setting was subsequently used in trial 4. Setting the pump to establish a high perfusion pressure ensured a sufficient flow rate and minimized fatigue during each trial.
Trials 3 and 4 were performed on contracting muscles. After the pretrial, the muscle was allowed to recover for 35 min, after which trial 3 began with the GS stimulated to contract tetanically at 1/2 s for 2 min at the predetermined flow rate. The flow rate was then increased by 20%, taken back to the control level, and finally decreased by 20%. The NIRS signals were allowed to stabilize at each flow rate. After another 35-min recovery, trial 4 was performed with a similar protocol to that of trial 3, except that the GS was stimulated at the rate of 2/3 s.
Trial 5 was conducted to determine the maximal range of NIRS changes. First, the muscle remained at rest while the animal was ventilated on 100% O2 with muscle Q˙ increased by 50% above the control level; the levels of HHbMb and HbMbO2 that were achieved were considered to be minimal and maximal, respectively. Next, the muscle was stimulated to contract tetanically at a rate of 1/2 s with the popliteal artery clamped to stop the blood flow; the final plateau levels of HHbMb and HbMbO2 that were achieved were considered to be maximal and minimal, respectively.
Popliteal arterial samples were collected before and at the end of each trial for trials 1, 3, and 4, as well as at the end of each perturbation of trial 2. Popliteal venous samples were drawn before and after each perturbation within each trial for the determination of PO2, PCO2, and pH with a blood gas, pH analyzer (GEM Premier 3000 Instrumentation Laboratory, Lexington, MA), and total Hb concentration and O2Hb% via CO-Oximeter (IL-682, Instrumentation Laboratory). CvO2 was calculated as follows (e.g., ):
where PvO2 is the partial pressure of oxygen in the venous blood.
A continuous-wave NIRS system (Oxymon Mk III, Artinis Medical Systems BV) was used to assess relative changes in muscle oxygenation. Two fiber-optic bundles were used to emit and receive light at two different wavelengths (760 and 860 nm). The optodes were placed over the belly of the medial head of the left GS and held in place with an elastic band. The muscle was covered with a dark plastic sheet throughout the experiment to block extraneous light. The two optodes were held 25 mm apart, resulting in a penetration depth of approximately 12.5 mm (7). Relative changes in oxyhemoglobin/oxymyoglobin (HbMbO2), deoxyhemoglobin/deoxymyoglobin (HHbMb), and total hemoglobin/myoglobin (HbMbtot) were output from the system to represent the oxygenation/deoxygenation state of Hb and Mb. The NIRS signal was biased to zero before the start of each trial. For analysis of the NIRS results, HHbMb and HbMbO2 values were 10-s averages of HHbMb and HbMbO2 signals before drawing the venous sample in trials 1 and 2, the average of the signals during the 12 s before the venous blood draw for trial 3 (six contractions) and trial 4 (eight contractions). The 12-s period was chosen so that the average HHbMb and HbMbO2 values were calculated based on NIRS changes within complete contraction cycles. HHbMb% and HbMbO2% were calculated by dividing HHbMb and HbMbO2 in arbitrary units, by the maximal range of HHbMb and HbMbO2, respectively, as determined in trial 5.
Data are presented as means ± SD. One-way repeated-measures ANOVA was used to compare the different perturbations within each trial, as well as to compare the parameters of different trials. The level of significance was set at P ≤ 0.05. If a difference was found, a Student–Newman–Keuls post hoc test was performed to determine where the differences were located in the data set. The correlations between HHbMb, HbMbO2, and venous O2Hb% were assessed with linear regression analysis. All statistical computations were performed using KaleidaGraph 4.5 (Synergy Software, Reading, PA).
Arterial PaO2 was not significantly different between the beginning (91 ± 8 mm Hg) and the end (93 ± 10 mm Hg) of this trial (P = 0.283). Control and experimental values of Q˙, blood pressure, CvO2, venous O2Hb%, HHbMb%, and HbMbO2% are displayed in Table 1, and significant differences are indicated. Q˙ varied with the experimental manipulations as planned, and blood pressure (perfusion pressure) responded accordingly. Although the statistical differences were not completely consistent, the numerical changes across the different blood flow conditions were uniform for CvO2, venous O2Hb%, HHbMb%, and HbMbO2%; i.e., when blood flow and, therefore, O2 delivery were increased, the oxygenation status was higher and deoxygenation lower, and vice versa.
The control and treatment values of PaO2, CvO2, venous O2Hb%, HHbMb%, and HbMbO2% in trial 2 are shown in Table 2. The PaO2 changes were as planned on the basis of the inspired O2 percentage provided to the animals. With only one exception, all of the statistical differences across the four PaO2 conditions were identical for CvO2, venous O2Hb%, HHbMb%, and HbMbO2%. With increased PaO2 and therefore increased O2 delivery, venous and muscle oxygenation indicators increased whereas deoxygenation indicators decreased; lower PaO2 resulted in the opposite effects.
Arterial PaO2 was not different between the beginning (110 ± 20 mm Hg) and the end (107 ± 23 mm Hg) of this trial (P = 0.336). Q˙, blood pressure, CvO2, venous O2Hb%, HHbMb%, and HbMbO2% in trial 3 are presented in Table 3. Q˙ changes were as planned in all cases. Blood pressure differences were consistent with the blood flow and duration of contraction times across the conditions. As was the case for trial 2, with only one exception, all of the statistical differences across the four blood flow conditions were identical for CvO2, venous O2Hb%, HHbMb%, and HbMbO2%. With increased blood flow and therefore increased O2 delivery, venous and muscle oxygenation indicators increased whereas deoxygenation indicators decreased; lower blood flow resulted in the opposite effects.
Arterial PaO2 was not different between the beginning (97 ± 10 mm Hg) and the end (92 ± 8 mm Hg) of this trial (P = 0.197). Values for Q˙, blood pressure, CvO2, venous O2Hb%, HHbMb%, and HbMbO2% are displayed in Table 4. Q˙ varied as experimentally manipulated. Blood pressure responded as expected on the basis of the blood flows and contraction durations. There was greater variability in the statistical differences in this trial than that in trial 3. Nevertheless, all of the numerical changes and most of the statistical differences across the four blood flow conditions were consistent for CvO2, venous O2Hb%, HHbMb%, and HbMbO2%. With increased blood flow and therefore increased O2 delivery, venous and muscle oxygenation indicators increased whereas deoxygenation indicators decreased; lower blood flow resulted in the opposite effects.
The maximal changes of HHbMb and HbMbO2 spectra in this trial were considered to reflect the maximal range of NIRS changes under blood perfusion. The average maximal HHbMb change was 81.2 ± 14.8 arbitrary units (au), and 150.5 ± 20.8 au for the maximal HbMbO2 change. The blood Hb concentration varied from 14.9 to 20.2 g·dL−1 among the different animals in trials 1 to 4.
Correlation between HHbMb% and HbMbO2% versus O2Hb% and CvO2
The correlations between normalized HHbMb (HHbMb%) and HbMbO2 (HbMbO2%) versus O2Hb% and CvO2 were assessed for the data collected in trials 1–4. The correlational relationships between HHbMb% and HbMbO2% versus venous O2Hb% and CvO2 for each animal are depicted in the Supplemental Digital Content (see Figure, Supplemental Digital Content 2, Individual plots of HHbMb% vs. venous O2Hb% for Dogs 2–6, https://links.lww.com/MSS/A707; see Figure, Supplemental Digital Content 3, Individual plots of HbMbO2% vs. venous O2Hb% for Dogs 2–6, https://links.lww.com/MSS/A708; see Figure, Supplemental Digital Content 4, Individual plots of HHbMb% vs. CvO2 for Dogs 2–6, https://links.lww.com/MSS/A709; see Figure, Supplemental Digital Content 5, Individual plots of HbMbO2% vs. CvO2 for Dogs 2–6, https://links.lww.com/MSS/A710). A high linear correlation was seen between HHbMb% and venous O2Hb% (R2 = 0.92 ± 0.05), between HbMbO2% and venous O2Hb% (R2 = 0.92 ± 0.03), between HHbMb% and CvO2 (R2 = 0.89 ± 0.06), and between HbMbO2% and CvO2 (R2 = 0.90 ± 0.05).
The overall relationships between HHbMb% and venous O2Hb% (Fig. 1A) and between HHbMb% and CvO2 (Fig. 1B) for the combined data are illustrated in Figure 1. The overall relationships between HbMbO2% and venous O2Hb% (Fig. 2A) and between HbMbO2% and CvO2 (Fig. 2B) for the combined data are illustrated in Figure 2. These overall relationships were linear and highly correlated with R2 values ranging from 0.81 to 0.90.
This study investigated the relationship between deoxy- and oxy-NIRS signals and venous O2Hb% and venous oxygen concentration (CvO2). The major finding is that high correlations existed between HHbMb% and HbMbO2% versus venous O2Hb% and CvO2 under a variety of inspired gas conditions, and with different blood flows both at rest and at different frequencies of muscle contractions.
The relationship between HHbMb/HbMbO2 and each of venous O2Hb% and CvO2 was evaluated under four conditions in each of trials 1–4. To our knowledge, this is the first study to evaluate the correlation between NIRS signals and venous O2Hb% and CvO2 over such a wide variety of physiological conditions (normoxia, hyperoxia, and hypoxia, as well as various blood flow rates, both at rest and during muscle contractions of differing metabolic intensities). It is expected that HbMbtot would affect the correlation between venous O2Hb% and CvO2 and the relative NIRS changes with devices that do not measure absolute concentrations (16,18,31). For example, jugular venous bulb O2 saturation (jugular venous O2Hb%) was measured and compared with NIRS-measured cerebral oxygenation during surgical operations in humans 10 min after the start of the operation, after 400 mL blood loss, after 800 mL blood loss, and right before blood transfusion (31). In that study (31), NIRS values did not change in parallel with O2 saturation in the jugular venous sample. Instead, with blood loss, jugular venous O2Hb% remained unchanged whereas NIRS values decreased significantly. Therefore, in the present study, we sought to minimize the influence of various HbMbtot concentrations in different animals ([Hb] in the range of 14.9 to 20.2 g·dL−1) by normalizing NIRS measurements to the maximal range of NIRS changes. However, despite the normalization of the NIRS signal, we still observed interindividual variability in that the overall correlation of HHbMb% with venous O2Hb% and with CvO2 in the combined data generally displayed a weaker relationship (R2 from 0.81 to 0.90) in comparison with the relationship within each individual animal (R2 from 0.85 to 0.98). In a study conducted by Wilson et al. (29), which used a similar experimental design as the present one, canine gracilis muscle was isolated and electrically stimulated to elicit twitch contractions at 0.25, 0.5, 1, 2, 3, 4, and 5 Hz. A high linear correlation was observed between HHbMb and venous O2Hb% (R = −0.97 ± 0.01). Similar to the present study, an interindividual variability in the slope and intercept of the correlation curve was observed. Therefore, our present study confirmed and extended the results of Wilson et al. (29). Overall, these results suggest that for a particular animal/muscle, the deoxy-/oxy-NIRS signals reflect changes in the O2 level of the tissue’s venous drainage under a wide variety of physiological conditions. Our previous work with this same model has shown good (R2 = 0.69 for spontaneous blood flow) to very good correlation (R2 = 0.93 for pump-controlled blood flow) between the change in HHbMb and venous O2Hb% during the transition from rest to steady-state contractions, particularly after the first 20 s (30).
The relationship between venous O2Hb% and HHbMb/HbMbO2 has also been investigated in humans in vivo. HbO2 saturation as indicated by a NIRS-based cerebral oximeter (5) was reported to be linearly correlated (R2 = 0.90 ± 0.09) with jugular venous O2Hb% as determined from blood samples under hypercapnic/hypoxic and normocapnic/hypoxic conditions (17). In another study by Mancini et al. (23), NIRS spectra (measured with absorption at 760 and 800 nm) and venous O2Hb% in deep arm muscles were compared during an exercise of depressing a lever that lifted a weight every 4 s for 2 min. In that study (23), a high correlation (R2 = 0.92 ± 0.02) was observed between NIRS and venous O2Hb%. More recently, Vogiatzis et al. (28) used three NIRS instruments with six optodes to assess oxygenation indices and local blood flow via indocyanine green dye in the vastus lateralis of cyclists exercising at several different intensities during both normoxia and hypoxia (FIO2 = 0.12). They (28) found linear correlations (median R2 = ≈0.7) between a flow-weighted average NIRS O2 signal and a femoral venous O2Hb% for each of their six subjects.
However, other evidence has suggested that the oxy-NIRS signal (HbMbO2) and venous O2Hb% are not correlated under all conditions in humans in vivo. MacDonald et al. (22) determined the relationship between a calculated NIRS relative change in muscle tissue oxygenation and measures of femoral venous O2Hb% during the transition from rest to leg-kicking exercise at a constant intensity for 5 min. Although NIRS oxygenation generally mimicked the femoral venous O2Hb% changes during the first minute, the NIRS oxygenation measure increased significantly during the succeeding minutes whereas the femoral venous O2Hb% remained at low levels (22). In another human study by Boushel et al. (2), a cuff was first placed around the upper arm and inflated to 280 mm Hg for 10 min while deep antecubital venous blood samples were drawn every minute and subsequently in recovery. In two other trials, handgrip exercises were performed at 15% and 30% of maximal voluntary contraction; in both cases, the handgrip exercise was followed by a postexercise muscle ischemia period for 3 min and then a recovery period. Low but significant correlations were found for most conditions, e.g., ischemia at rest (R2 = 0.36), 15% MVC handgrip (R2 = 0.41), and 30% MVC handgrip (R2 = 0.31). However, HbMbO2 and O2Hb% were not correlated during the postexercise muscle ischemia, with venous O2Hb% increasing throughout the ischemic period and NIRS-HbO2 decreasing. Finally, Costes et al. (3) reported that NIRS-measured O2 muscle saturation (NIRS-O2Hb%) was weakly correlated (R2 = 0.30) with femoral HbO2 saturation (femoral venous O2Hb%) during 30 min of constant workload cycling under hypoxic (FIO2 = 0.14), but not normoxic, conditions. The changes in NIRS-O2Hb% and femoral venous O2Hb% followed a parallel pattern during hypoxia. However, during exercise in normoxia, venous O2Hb% decreased up to 15 min and then remained stable, whereas NIRS O2Hb% only decreased slightly at minute 5 and then returned back to the resting level or even higher.
Grassi et al. (11) postulated that investigations using NIRS oxygenation indices, e.g., MacDonald et al. (22), were less likely to find high correlations with venous O2Hb% because the oxygenation signals are more heavily influenced by changes in perfusion/blood volume in the field of NIRS interrogation. Accordingly, they (11) proposed that the HHbMb signal is a more robust indicator of the balance between blood flow and O2 extraction. Indeed, Koga et al. (19) recently reported data showing that the oxygenation signal (HbMbO2) from a continuous-wave NIRS instrument (which provides relative changes in NIRS signals) did not accurately track local muscle O2 extraction during changes in skin blood flow, as determined by a time-resolved NIRS instrument (which determines absolute changes in NIRS signals). However, relative changes in the deoxygenation signal (HHbMb) as assessed by the continuous-wave NIRS device displayed the same temporal pattern as the absolute measure provided by the time-resolved NIRS instrument. We can also note here that, as often seen in human studies, there was a decided asymmetry between the HbMbO2 and the HHbMb signals when we performed our physiological calibration that involved increased blood flow and occlusion. The change in the HbMbO2 signal was much greater than that in the HHbMb signal, likely because of the effect of blood volume on the oxygenation signal. The absence of skin and adipose in our model indicates that those factors are not the sole cause of this asymmetry. In light of the preceding discussion, our overall results suggest that the HHbMb signal should accurately track the venous O2Hb% of the interrogated tissue.
For examining the correlation between O2Hb% and NIRS measurements, the isolated perfused canine GS model offers several advantages. First, the depth to which near-infrared light can penetrate tissue is approximately 1.5–2.5 cm (7). Therefore, excessive subcutaneous fat makes NIRS interrogation less effective or even impossible for detecting tissue oxygenation status. However, in the GS model, all overlying fat and other tissue have been removed. Second, the amount of tissue interrogated is a greater percentage of the total muscle volume in our experiments in comparison with typical human experiments. If, for simplicity, we assume a) that the volume of interrogation is one-half of an ellipsoid with axes equal to one-half the distance between the optodes (≈12.5 mm) in all three dimensions and b) the muscle volume is equal to its weight divided by a density of 1.06 g·mL−1 (25), then in our model, we are interrogating approximately 6% of the muscle volume (≈4 mL/≈68.5 mL). For a typical human muscle (e.g., vastus lateralis; ≈286 mL ), the percentage would be ≈1.4%, disregarding the effect of skin and fat. Third, venous O2Hb% measured from blood samples in human studies in vivo reflect the oxygenation status of the mixture of venous blood draining local tissues. For example, in the study by Mancini et al. (23), although the catheter was inserted deep into the vein to collect blood that was primarily from exercising muscles, the authors were still only able to conclude that the collected blood originated partly from the target muscle. However, in the GS model, all venous blood is derived from the isolated GS. Therefore, our results indicate that for an individual muscle or perhaps muscle group, NIRS HHbMb and HbMbO2 signals reliably reflect the mean venous O2 saturation of blood derived from the tissue being interrogated by NIRS.
Note that our results absolutely do not exclude an unknown contribution of Mb to the NIRS signal. Extensive review, calculations, and modeling have produced values of 25%–80% for the contribution of Mb to the total NIRS signal (4,20). Our current study does not offer any insight into this question but simply indicates that regardless of the exact Mb contribution, there is a strong correlation of the NIRS signals with the venous O2Hb% under the conditions of our experiments.
Despite the strengths outlined previously, there are several limitations to the present study. First, the canine GS muscle is highly oxidative, almost exclusively made up of type S (oxidative) and type FR (oxidative/glycolytic) muscle fibers (data and nomenclature as in Maxwell et al. ). Therefore, the findings of the present study must be interpreted with caution, especially when applied to other species or more glycolytic muscle groups. Second, the blood flow in this study was pump controlled. This allowed us to investigate the correlation between NIRS signals and venous O2Hb% at various controlled flow rates. However, this also means that the results of the present study might not always be applicable to circumstances with spontaneously regulated flow, for example, during exercise. In addition, the canine GS model induced isometric muscle contractions via electrical stimulation of the sciatic nerve. In this type of contraction, all the motor units are recruited simultaneously, which is different from the recruitment pattern of spontaneous exercise. Third, our correlations were possibly enhanced by the fact that the oxygenation/deoxygenation levels, because of the protocols used, tended to cluster in high and low values. Nevertheless, the generality of these correlations is supported by our previous work with this same model during non–steady-state contractions, which also showed good correlations across the entire range of oxygenation levels from rest to contractions at 1/3 s and 2/3 s (see Fig. 3 of Wüst et al. ). Finally, the Oxymon Mk III was the only NIRS instrument used, so it is possible that instruments from other manufacturers could give different results. In particular, the Oxymon is a continuous-wave device that provides relative rather than absolute changes in the heme signals.
In conclusion, a linear correlation between normalized NIRS-derived HHbMb, HbMbO2, and venous O2Hb% and CvO2 for individual animals/muscles was confirmed under various blood flows, various inspired O2 percentages, and at two different intensities of muscle contractions. These relationships were weaker but still strong when the data from different animals/muscles were combined.
Some of the funding for this study was provided by an Auburn University Graduate Student Research Grant to Yi Sun. There are no conflicts of interest. The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Boushel R, Langberg H, Olesen J, Gonzales-Alonzo J, Bülow J, Kjaer M. Monitoring tissue oxygen availability with near infrared spectroscopy (NIRS) in health and disease. Scand J Med Sci Sports
2. Boushel R, Pott F, Madsen P, et al. Muscle metabolism from near infrared spectroscopy during rhythmic handgrip in humans. Eur J Appl Physiol Occup Physiol
3. Costes F, Barthélémy JC, Féasson L, Busso T, Geyssant A, Denis C. Comparison of muscle near-infrared spectroscopy and femoral blood gases during steady-state exercise in humans. J Appl Physiol (1985)
4. Davis ML, Barstow TJ. Estimated contribution of hemoglobin and myoglobin
to near infrared spectroscopy. Respir Physiol Neurobiol
5. Delpy DT, Cope M, van der Zee P, Arridge S, Wray S, Wyatt J. Estimation of optical pathlength through tissue from direct time of flight measurement. Phys Med Biol
6. Ferrari M, Muthalib M, Quaresima V. The use of near-infrared spectroscopy in understanding skeletal muscle physiology: recent developments. Philos Trans A Math Phys Eng Sci
7. Ferrari M, Quaresima V. Near infrared brain and muscle oximetry: from the discovery to current applications. J near Infrared Spec
8. Ferrari M, Wei Q, Carraresi L, De Blasi RA, Zaccanti G. Time-resolved spectroscopy of the human forearm. J Photochem Photobiol B
9. Gladden LB, Crawford RE, Webster MJ. Effect of lactate concentration and metabolic rate on net lactate uptake by canine skeletal muscle. Am J Physiol
. 1994;266(4 Pt 2):R1095–101.
10. Goodwin ML, Hernández A, Lai N, Cabrera ME, Gladden LB. V˙O2
on-kinetics in isolated canine muscle in situ during slowed convective O2
delivery. J Appl Physiol (1985)
11. Grassi B, Pogliaghi S, Rampichini S, et al. Muscle oxygenation and pulmonary gas exchange kinetics during cycling exercise on-transitions in humans. J Appl Physiol
12. Hamaoka T, McCully KK, Niwayama M, Chance B. The use of muscle near-infrared spectroscopy in sport, health and medical sciences: recent developments. Philos Trans A Math Phys Eng Sci
13. Henson LC, Calalang C, Temp JA, Ward DS. Accuracy of a cerebral oximeter in healthy volunteers under conditions of isocapnic hypoxia. Anesthesiology
14. Hernández A, Goodwin ML, Lai N, Cabrera ME, McDonald JR, Gladden LB. Contraction-by-contraction V˙O2
and computer-controlled pump perfusion as novel techniques to study skeletal muscle metabolism in situ. J Appl Physiol (1985)
15. Jöbsis FF. Noninvasive, infrared monitoring of cerebral and myocardial oxygen sufficiency and circulatory parameters. Science
16. Jones AM, Davies RC, Ferreira LF, Barstow TJ, Koga S, Poole DC. Evaluation of the dynamics of muscle oxygenation by near-infrared-based tissue oximeters reply. J Appl Physiol
17. Kim MB, Ward DS, Cartwright CR, Kolano J, Chlebowski S, Henson LC. Estimation of jugular venous O2
saturation from cerebral oximetry or arterial O2
saturation during isocapnic hypoxia. J Clin Monit Comput
18. Kishi K, Kawaguchi M, Yoshitani K, Nagahata T, Furuya H. Influence of patient variables and sensor location on regional cerebral oxygen saturation measured by INVOS 4100 near-infrared spectrophotometers. J Neurosurg Anesthesiol
19. Koga S, Poole DC, Kondo N, Oue A, Ohmae E, Barstow TJ. Effects of increased skin blood flow on muscle oxygenation/deoxygenation: comparison of time-resolved and continuous-wave near-infrared spectroscopy signals. Eur J Appl Physiol
20. Lai N, Zhou H, Saidel GM, et al. Modeling oxygenation in venous blood and skeletal muscle in response to exercise using near-infrared spectroscopy. J Appl Physiol (1985)
21. Le Troter A, Fouré A, Guye M, et al. Volume measurements of individual muscles in human quadriceps femoris using atlas-based segmentation approaches. MAGMA
22. MacDonald MJ, Tarnopolsky MA, Green HJ, Hughson RL. Comparison of femoral blood gases and muscle near-infrared spectroscopy at exercise onset in humans. J Appl Physiol (1985)
23. Mancini DM, Bolinger L, Li H, Kendrick K, Chance B, Wilson JR. Validation of near-infrared spectroscopy in humans. J Appl Physiol (1985)
24. Maxwell LC, Barclay JK, Mohrman DE, Faulkner JA. Physiological characteristics of skeletal muscles of dogs and cats. Am J Physiol
25. Mendez J, Keys A. Density and composition of mammalian muscle. Metabolism
26. Owen-Reece H, Smith M, Elwell CE, Goldstone JC. Near infrared spectroscopy. Br J Anaesth
27. Stainsby WN, Welch HG. Lactate metabolism of contracting dog skeletal muscle in situ. Am J Physiol
28. Vogiatzis I, Habazettl H, Louvaris Z, et al. A method for assessing heterogeneity of blood flow and metabolism in exercising normal human muscle by near-infrared spectroscopy. J Appl Physiol (1985)
29. Wilson JR, Mancini DM, McCully K, Ferraro N, Lanoce V, Chance B. Noninvasive detection of skeletal muscle underperfusion with near-infrared spectroscopy in patients with heart failure. Circulation
30. Wüst RC, McDonald JR, Sun Y, et al. Slowed muscle oxygen uptake kinetics with raised metabolism are not dependent on blood flow or recruitment dynamics. J Physiol
. 2014;592(Pt 8):1857–71.
31. Yoshitani K, Kawaguchi M, Iwata M, et al. Comparison of changes in jugular venous bulb oxygen saturation and cerebral oxygen saturation during variations of haemoglobin concentration under propofol and sevoflurane anaesthesia. Br J Anaesth