MAGUIRE, MARIA A.1; WEAVER, TIMOTHY W.1,2; DAMON, BRUCE M.1,2,3,4
The microvasculature, which includes the terminal arterioles, capillaries, and postcapillary venules, is the site of oxygen delivery and substrate exchange between the blood and peripheral tissues. When oxygen supply/demand relationships are chronically challenged, such as by exercise training, the skeletal muscle microvasculature responds structurally and functionally to enhance material exchange (21). Conversely, microcirculatory structure and function impairments, such as those occurring in uncontrolled diabetes, may impair material exchange (19). Assessing microcirculatory function is, therefore, important to understanding normal, adaptational, and pathological physiology. In particular, the need exists to develop noninvasive protocols for examining microvascular function during and after challenges such as muscle contraction.
During isometric contractions, the intramuscular pressure increases linearly through contraction intensities up to and including maximum voluntary contraction (MVC) (27). This pressure compresses arterioles and venules along the muscle's fascicular lines (11), which restricts arterial flow (12), increases venous outflow (14), and decreases venous blood volume (28). During the first cardiac cycle after single isometric contractions, large-vessel blood flow increases (28), which may result from vasodilatory paracrine agents such as nitric oxide (NO), acetylcholine, prostaglandins, ATP, and potassium (3), and/or the mechanical action of muscle contraction on the arteriovenous pressure gradient (20). Postcontraction vasodilatory responses may be gender dependent, because skeletal muscle arterioles in premenopausal women have greater maximal and submaximal endothelial-dependent responses to shear stress than those of men (15).
After prolonged dynamic exercise, oxyhemoglobin saturation (%HbO2) begins to recover either immediately after exercise or nearly so (24). However, there have been several reports that %HbO2 does not begin to recover immediately after single, brief isometric contractions. In 1964, Corcondilas and colleagues (4) reported that after 0.3-s voluntary contractions of the hand and finger flexors, the deep venous %HbO2 continued to fall for about 6 s before recovering and reaching a peak value of approximately 30 s after exercise. More recently, Kime et al. (18) have shown that after brief, high-intensity, isometric handgrip contractions, %HbO2 recovery is delayed in muscles with high oxidative capacities. These authors proposed that the delayed reoxygenation resulted from an oxidative capacity so high that the oxygen demand exceeded the oxygen supply, despite the hyperemia.
Near-infrared spectroscopy (NIRS) is a noninvasive technique that is well suited to studying these processes. NIRS exploits the differential absorption of near-infrared light by oxygenated and deoxygenated heme groups, such as those in hemoglobin (Hb) and myoglobin (Mb). More than 90% of the NIRS signal measured is from Hb (23), and with frequency domain NIRS methods, the deoxy-Hb concentration ([HHb]), oxy-Hb concentration ([HbO2]), total Hb concentration ([THb]), and %HbO2 can be measured. Because most blood in skeletal muscle is in the veins and because the blood in the capillaries has a reduced hematocrit, NIRS signals from muscle arise mainly from venous Hb (23). The signal arises from small veins in particular, because light absorption in vessels greater than approximately 1 mm in diameter is so complete that the light entering these vessels is not reflected back to the detector (23).
The purpose of this study was to observe and quantify the time courses of [THb] and %HbO2 recovery after dorsiflexion MVC. We studied dorsiflexion because it is important in postural control and locomotion; for example, the inability to control foot drop may contribute to diabetic foot syndrome (1). We hypothesized that [THb] would decrease during contraction, begin to recover immediately after contraction, and, during the later stages of recovery, would transiently exceed the precontraction value. Second, we hypothesized that %HbO2 would fall during the contraction and, because of the preponderance of oxidative fibers in the tibialis anterior (TA) muscle (17), decrease further during the initial stages of recovery before beginning a delayed recovery. We also hypothesized that %HbO2 would transiently exceed the precontraction value at a later point in recovery. Finally, we hypothesized that female subjects would have a larger, more rapid increase in [THb] after contraction. We used kinetic analysis of NIRS data obtained before, during, and after brief dorsiflexion MVC to test these hypotheses, and we found evidence in support of the first two hypotheses. However, we found that in moderately active young subjects, gender influenced neither the kinetics nor the amplitude of the [THb] and %HbO2 responses to brief dorsiflexion MVC.
These procedures were approved by the Vanderbilt University institutional review board. Eighteen apparently healthy subjects (nine men, nine premenopausal women) participated in the study after providing written informed consent. The subjects were also participants in a larger study of skeletal muscle fatigue. Self-reported questionnaire data were used to exclude persons diagnosed with neurological, metabolic, cardiovascular, endocrine, pulmonary, liver, or renal disorders, and/or other medical contraindications to isometric dorsiflexion exercise (such as a prior or current history of compartment syndrome, a recent ankle sprain, etc.). The physical characteristics of the subjects are reported in Table 1. Male subjects were taller (P < 0.001), heavier (P <0.001), exerted greater dorsiflexion force (P < 0.001), and had less subcutaneous fat overlying the TA muscle (P < 0.05) than female subjects. Male and female subjects were similar with regard to age and physical activity level.
All exercise involved isometric dorsiflexion of the subject's self-reported dominant foot. The subject lay supine on an exercise table with his/her foot strapped into a home-built isometric exercise device. The foot was in the anatomic position and the knee was supported by a bolster and slighted flexed (angle of knee flexion ≍ 7.5°). The foot was firmly strapped to the exercise device using 3.8-cm nylon straps; to limit the involvement of the toe extensors, the straps were placed across the foot proximal to the base of the fifth digit. The torso was slightly inclined so that the heart was approximately 5 cm above the TA.
During the first visit to the laboratory, the subjects were informed about the study procedures, provided informed consent, performed health screening, and practiced the exercise. The skinfold thickness (THKSF) at the NIRS measurement site was measured using Harpenden skinfold calipers. The subjects reported on four more occasions for the fatigue study. The MVC data from the first day of the fatigue study were used in the present analysis. To measure the MVC force, the subjects performed two or more 3-s contractions. The subjects were instructed to pull as hard and as fast as possible and to exhale during effort to avoid a Valsalva maneuver. There were 5 min of rest between contractions. MVC were performed until the maximum force recorded during two contractions performed using good form differed by 5% or less. The greater force of the two contractions was used to represent the subject's MVC force.
The subjects were instructed not to consume caffeine or use tobacco during the six hours before a test session or to use alcohol or perform moderate or heavy physical activity for at least 24 h before a session. To ensure compliance with these instructions, the subjects completed a survey of their activity and lifestyle habits covering the 24 h before each test session.
Force Data Acquisition
The isometric dorsiflexion device included an Interface Force (Interface Force, Scottsdale, AZ) model SSM-AJ-500 load cell. The signals from the load cell were amplified (model SGA/A, Interface Force) and connected via a serial connector box (model SCB-68, National Instruments, Austin, TX) to an analog-to-digital conversion card (model 6036E, National Instruments) fit into a laptop computer. Custom data-acquisition software written in LabVIEW 7.1 (National Instruments) was used to acquire force data at 1 kHz and to provide real-time, 20-Hz visual feedback to the subjects in the form of a simulated LED panel on the computer screen. The baseline force offset was measured before each contraction to account for any changes in subject position between contractions. The force data were also sampled by the NIRS software (ISS Oximeter software, version 2.14, ISS, Inc., Champaign, IL).
NIRS Data Acquisition
Tissue-oxygenation data were collected using a frequency domain, multidistance NIRS oximeter (Model 96208, ISS, Inc.) and its accompanying software. Before each testing session, the oximeter was calibrated using a block with known absorption and scattering coefficients, and the performance was verified using a second block with different optical properties. A rigid emitter-detector head was placed over the maximum cross-sectional area of the muscle, as identified by visual inspection and palpation, and held in place using an elastic strap. The strap was adjusted to be just tight enough to prevent motion of the probe during the contraction. The oximeter head made good contact with the skin, and no motion of the probe relative to the skin was observed during or after the contractions. As additional precautions against the introduction of extraneous room light into the detector, the oximeter head was covered with an opaque material and the room lights were dimmed.
The theory of operation of the ISS oximeter has been described previously (16). Briefly, the oximeter emits light through fiber-optic cables from four lasers at a wave length of 730 nm and from four lasers at a wave length of 860 nm. The oximeter head is arranged such that the four fibers emitting light at each wave length are located at different distances from a single detector (2.0, 2.5, 3.0, and 3.5 cm). The light is intensity modulated at 110 MHz, and as it passes through the tissue, it undergoes a path length-dependent phase shift. The AC, DC, and phase components of the light at each distance are measured; in accordance with the manufacturer's recommendations, the AC and phase components were used to determine the absorption and scattering coefficients. The measured coefficients were used to calculate [HbO2] and [HHb] according to the algorithms incorporated into the software. From the [HbO2] and [HHb] data, we derived the [THb] and %HbO2, for which the software manufacturer estimates accuracies of ± 3 μM and ± 5%, respectively. Whereas NIRS data are typically averaged during acquisition, a requirement of this oximetry system is that all data channels be sampled at the same rate. To reduce errors associated with undersampling the force production, all data were sampled at 6.45 Hz, and a seven-point (1.09 s) moving average was applied during data analysis (see below).
Physical Activity Level Assessment
Physical activity level was estimated using the 7-d physical activity recall (26) and expressed as kilocalories per kilogram per day.
Force Data Analysis
Home-written LabVIEW software was used to analyze the force data. The data were low-pass filtered at 25 Hz using a fifth-order Butterworth filter. A measured calibration factor and each contraction's baseline force offset were used to convert the measured voltages to force values. The greatest force production during the MVC was used to represent the MVC strength. The point at which the force returned to the precontraction baseline value was used to represent the start of the postcontraction period.
NIRS Data Analysis
Analysis of the [THb] and %HbO2 kinetic properties was performed using Matlab version 7.01 (The MathWorks, Inc., Natick, MA). The [THb] data for the 150 s after contraction were fit to single exponential growth (SG; equation 1) and sequential exponential growth-decay (GD; equation 2) models:
where Hb represents either [THb] or %HbO2; t is the postcontraction duration; TD is the time delay; the subscripts G and D indicate decay and growth behavior, respectively; the subscript 0 denotes the baseline value; and the subscripts 1 and 2 refer to the temporal sequence of growth or decay behavior, with the designation of 2 used for this decay term because it followed the growth component. After providing reasonable initial estimates for each parameter, a Nelder-Mead simplex algorithm was used to perform an unconstrained, nonlinear minimization of the residual variances in an ordinary least-squares sense, yielding final parameter estimates. To determine whether the SG or GD model was more appropriate to describe the data, the following procedure was used:
1. Determine whether the parameter estimates for each model are scientifically appropriate.
2. Determine whether the residual plots are appropriate.
3. 3a. If only one set of parameter estimates and residual plots are acceptable, accept this model.
4. 3b. If both sets of parameter estimates and residual plots are acceptable, calculate the F statistic between the GD and SG regressions. If the F test indicates a significantly better fit of the data to the GD model at the P < 0.05 level, accept this model.
5. 3c. If neither the SG nor the GD model provides reasonable parameter estimates and residual plots, fit the data again with different initial estimates.
To determine whether the parameter estimates were scientifically appropriate, the following criteria were used: positivity (for A and τ terms); nonnegativity (for TD terms); and, for the GD model, TDD2 > TDG1. The residual plots were considered inappropriate if there was a nonrandom distribution of the residuals over time, as determined by visual inspection. The F statistic was calculated as:
Equation (Uncited)Image Tools
where SS is the sum of squared residuals and df is the number of degrees of freedom, calculated as the length of the data set minus the number of terms in the model.
The %HbO2 data were fit to the SG and GD models, as described above, and also to decay-growth (DG; equation 4) and decay-growth-decay (DGD; equation 5) models:
with the designation of 1 used for the initial decay term because it precedes the growth component. In practice, when fitting the %HbO2 data, the SG model was always rejected because of inappropriate residual plots, and either the DG or GD model was rejected because of inappropriate parameter estimates and/or inappropriate residual plots. As a result, the decision tree described above could be followed, substituting the remaining two-component model and the DGD model for the SG and GD models, respectively.
To verify these fitting procedures, simulated [THb] datasets were generated according to equations 1 and 2 (the SGT and GDT datasets, respectively), and simulated %HbO2 datasets were generated according to equations 2, 4, and 5 (the GDO, DGO, and DGDO datasets, respectively). The parameters used to generate the datasets were similar to those observed in the experimental studies and are listed inTable 2. The synthesized GDT dataset is shown in Figure 1A. Real noise from the oximeter was recorded from two periods of rest. The data were passed through a seven-point moving average, demeaned, and added together (Fig. 1B). The noise was added to the synthesized data (Fig. 1C), and the resulting dataset was fit as described above, except that the initial estimates were modified by a random-number generator to prevent investigator bias. The SGT, GDO, DGO, and DGDO datasets are shown in Figure 1E-H. A total of five noisy datasets were created for each model using noise recordings obtained from different subjects. To characterize the accuracy of the model-selection and parameter-estimation procedures, two criteria were used: the classification accuracy (i.e., whether the correct model was selected) and whether the known parameters fell within the 95% confidence intervals (95% CI) around the estimated parameter.
Statistical analysis was performed using SPSS for Windows version 13.0 (SPSS, Inc., Chicago, IL). Between-gender differences in the descriptive characteristics were evaluated using Student's t-test. The effects of THKSF on the pre-C [THb] and %HbO2 values were evaluated separately for male and female subjects using linear regression, and the slopes and intercepts were compared between genders using 95% CI around the parameter estimates. To test the hypothesis that [THb] would begin to recover immediately after contraction, a 95% CI was constructed around the mean value of TDG1 for [THb]. A 95% CI that included zero was interpreted to mean that the postcontraction increase in [THb] began immediately after contraction. To test the hypotheses that during the contraction, [THb] would decrease, and that after the contraction, [THb] would transiently exceed the precontraction value, each subject's [THb] time course was characterized with the mean value during the 10-s period before contraction (pre-C), the mean value during the last second of contraction (end-C), and the 10.1-s postcontraction (10 s-P), maximum postcontraction (max-P), and 150-s postcontraction (end-P) values in the line of best fit. These values were compared using the general linear model (GLM; gender × time, using THKSF as a covariate, with repeated measures on time). To test the hypothesis that there would be delayed reoxygenation after MVC, a 95% CI was constructed around the mean value for TDG1 for %HbO2 and interpreted in the same way as that for [THb]. To test the hypotheses that during the contraction, the %HbO2 would decrease, and that after the contraction, it would initially decrease further and later exceed the precontraction value, its time course was characterized with pre-C, end-C, minimum postcontraction (min-P), max-P, and end-P mean values. These values were compared using the GLM as described above. Multiple comparisons were made using 95% CI adjusted using the Bonferroni procedure. In addition to the gender main effects included in the models described above, gender effects were also examined by comparing the mean values of the TD, τ, and A terms from the kinetic analysis using the GLM with THKSF as a covariate. Gender effects were tested with analysis of covariance rather than by matching a subset of subjects on the basis of THKSF, because an inherent assumption of matching subjects is that the subjects differ only in the matched variable, which is not the case. In contrast, analysis of covariance considers only the shared variance between the outcome variable and the covariate. In these analyses, a subject lacking a particular component (e.g., the initial decay component of %HbO2) was considered to have an amplitude value of zero. Statistical comparisons were considered significant at P < 0.05. For subject descriptive data, the mean and standard error (SE) were calculated; for analyses performed using the GLM, marginal means and SE were calculated on the basis of the regressions of [THb] and THKSF and %HbO2 and THKSF, according to standard statistical procedures.
Analysis of Simulated NIRS Datasets
Figure 1D shows a sample curve-fitting result from the analysis of a GDT dataset. Table 2 contains the known parameters (i.e., those used to synthesize the datasets) and the mean and SE of the parameter estimates. In all cases, the correct model was selected, and the 95% CI of the mean parameter estimates included the known parameters used to synthesize the datasets.
Relationship between NIRS Variables and THKSF
Figure 2A shows the relationships between the pre-C [THb] and THKSF for male and female subjects. For male subjects, linear regression of [THb] on THKSF resulted in the following equation:
(r = 0.78). For female subjects, this analysis resulted in the following equation:
Equation (Uncited)Image Tools
(r = 0.88). Figure 2B shows the relationships between the pre-C %HbO2 and THKSF for male and female subjects. For male subjects, linear regression of %HbO2 on THKSF resulted in the following equation:
Equation (Uncited)Image Tools
(r = 0.58). For female subjects, this analysis resulted in the following equation:
Equation (Uncited)Image Tools
(r = 0.58). For both [THb] and %HbO2, the 95% CI for the slope and intercept values for males include the corresponding values for female subjects, and vice versa.
Contraction and Postcontraction [THb] Changes
Figure 3A shows typical single-subject [THb] data before, during, and after a contraction. During the contraction, there was an abrupt decrease in [THb]; this was followed by an immediate recovery and eventual overshoot of [THb] after the contraction. Also shown is the best fit of the postcontraction data to the GD model. Figure 3B shows the residual variance from the curve-fitting procedure. No structure is obvious in the residuals, and the plot is generally rectangular in shape (indicating that the variance of the residuals is consistent over all values of time). These properties of the residual plot were typical of all subjects.
Figure 4A characterizes the postcontraction [THb] time course separately for the male and female subjects at five points in time. For clarity, the group-averaged data are not shown. The end-C [THb] value was significantly lower (by 6.9 (SE 0.8) %) than the pre-C value (P < 0.001). The postcontraction [THb] returned to the pre-C [THb] at a mean postcontraction duration of 8.4 (SE 1.1) s. At a postcontraction duration of 10.1 s, which was the TD for the %HbO2 regrowth, [THb] first exceeded the pre-C value (P < 0.05). The max-P [THb] was significantly greater than the pre-C, end-C, and 10 s-P values (P < 0.001 for all contrasts). For subjects in whom there was a secondary decay component, the max-P [THb] value occurred at a time equal to TDD2. Finally, [THb] began to return to the precontraction value, and the end-P value was not significantly different from the pre-C value.
FIGURE 4-Discrete an...Image Tools
Kinetic parameter estimates describing the postcontraction [THb] responses are given in Table 3. Across all subjects, regrowth of [THb] after the contraction began immediately (95% CI for TDG1 includes zero). The amplitude term for regrowth (AG1) listed in Table 3 represents 10.9 (SE 1.0) % of the precontraction mean [THb] value of 78.5 (SE 7.3) μM. In 15 subjects, there was a secondary decay in [THb], which began at TDD2 = 59.2 (SE 5.6) s and had an amplitude representing 2.4 (SE 0.7) % of the precontraction mean [THb]. For illustrative purposes, the group-averaged [THb] postcontraction time course is shown in Figure 5A.
Contraction and Postcontraction %HbO2 Changes
Figure 6A shows typical %HbO2 data before, during, and after a contraction. The %HbO2 decreased during the contraction, and its recovery was delayed by 10.1 s in this subject. At a postcontraction duration of 34.7 s, the %HbO2 reached its maximum postcontraction value and then recovered back toward the precontraction level. The best fit of the postcontraction data to the GD model is also shown. Figure 6B shows the residual variance from the curve-fitting procedure.
These trends in %HbO2 were characteristic of those observed in all subjects (Fig. 4B). As for [THb], the male and female subjects' data are plotted separately, and for clarity, the group-averaged data are not shown. During the contraction, the %HbO2 decreased from 65.0 (SE 0.9) % to 61.7 (SE 1.3) %, a relative change of −5.0 (SE 0.9) % (P < 0.001). The relative decreases in [THb] and %HbO2 were not significantly different from each other (paired Student's t-test, P > 0.05). Immediately after contraction, the %HbO2 either remained constant (10 subjects) or decreased (eight subjects) until a postcontraction duration of 10.1 s, at which point the %HbO2 began to regrow. The max-P %HbO2 was 66.8 (SE 0.9) % and was significantly greater than the pre-C value (P < 0.01). Finally, %HbO2 began to return to the precontraction value, and the end-P value was not significantly different from the pre-C mean value.
The kinetic parameter estimates for %HbO2 are given in Table 4. Ten subjects' data were best fit with the GD model, six subjects' data were best fit with the DGD model, and two subjects' data were best fit with the DG model. The amplitude term for the initial decay represented 2.8 (SE 1.2) % of the mean precontraction %HbO2 level. The regrowth phase, beginning 10.1 (SE 1.0) s after the end of the contraction, was observed in all subjects and had an amplitude representing 12.3 (SE 1.6) % of the precontraction mean value. The secondary decay component, observed in 16 subjects, began at TDD2 = 54.4 (SE 7.5) s and had an amplitude representing 4.9 (SE 0.9) % of the precontraction mean. The mean values of TDD2 for %HbO2 and [THb] did not differ significantly from each other. The group-averaged %HbO2 postcontraction time course is shown in Figure 5B.
Gender-specific kinetic parameter estimates for [THb] and %HbO2 are provided in Tables 3 and 4, respectively. When THKSF was included as a covariate in the GLM, there were no significant differences between males and females in any of the kinetic parameter estimates. Moreover, when THKSF was included as a covariate in the GLM, there were no significant differences between genders at any of the points in either of the discrete analyses (P = 0.101 for [THb] and 0.619 for %HbO2; Fig. 4).
In this paper, we report the time course of [THb] and %HbO2 changes in the TA muscle after 3-s dorsiflexion MVC. When interpreting these responses, it is important to recognize that although we studied a recovery process, a single isometric contraction more closely resembles the first of a series of intermittent contractions than the end of a prolonged bout of exercise, in which blood flow and metabolic rates would already have been elevated. As a result, the initial hemodynamic and metabolic responses that we observed during the recovery period may be more similar to the on-kinetics for these processes than the off-kinetics. Below, we consider the influence of several factors that could potentially have influenced the NIRS measurements of THb and %HbO2 kinetics, and we then argue that there are three phases to the balance of oxygen supply and demand after MVC.
Effects of Intersubject Variation in THKSF
A complication to in vivo optical spectroscopy is that the light-penetration depth is approximately equal to the half of the light emitter-detector spacing, and the measurements are influenced by all tissues in the light path (5). For our NIRS system, the detector-emitter distances correspond to a mean light-penetration depth of 1.375 cm and a maximum depth of 1.75 cm. In these subjects, the range of adipose tissue thickness was from 0.27 to 1.23 cm (i.e., half of the THKSF measurement from the skinfold calipers). In magnetic resonance imaging studies of eight similar subjects, we observed that the TA muscle extends 2.50 (SE 0.15) cm below the subcutaneous adipose tissue along the line normal to the plane of the oximeter head (unpublished observations). Thus, the NIRS measurements correspond to subcutaneous fat and the TA muscle, but not to any other muscles.
There are two possible effects of the resulting partial-volume averaging of the NIRS data from the muscle and adipose tissues. First, it introduced additional variability into the baseline and amplitude terms of the kinetic analysis. Importantly, when the regression statistics between these variables were calculated, there were no significant differences between the slope and intercept values for male and female subjects. This finding validates an underlying assumption of the analysis of covariance, which is that the relationship between the dependent variable and the covariate is the same for all subject groups. Another potential effect of the intersubject variation in THKSF is an averaging of the kinetic properties of blood flow to the contracted muscle (in which a hyperemic response would be expected) and the subcutaneous fat (in which no response is expected). This averaging may have dulled the observed kinetic response of the [THb] growth phase, as τG1 for [THb] and THKSF were significantly correlated (r = 0.59; data not shown). However, no significant correlations were observed between THKSF and any of the other time delays or time constants.
Effects of Mb
Mb is more than 90% saturated with oxygen at rest (25) and, as discussed in the Introduction, comprises up to 10% of the NIRS signal. Assuming that the oxy-Mb saturation does not change during or after these short contractions (see below), the effects of Mb on the present observations would be to elevate the [THb] and %HbO2 A0 estimates by up to approximately 10%, with no effect on the other kinetic parameters. Because the arguments we make below do not rely on the A0 estimates, they are not confounded by the inclusion of Mb in the NIRS window.
Contraction-Induced Changes in [THb] and %HbO2
During the contraction, [THb] decreased by 6.9 (SE 0.8) %. Because the NIRS signal originates primarily from small veins (23), and considering previous observations of increased venous outflow during tetanic contractions (14), the decrease in [THb] probably resulted from a rapid ejection of blood from the postcapillary venules. Concomitantly, there was a significant decrease in the [OxyHb] (precontraction: 50.4 (SE 4.1) μM; end contraction: 43.4 (SE 3.3) μM; two-tailed paired Student's t-test, P < 0.001) but no change in the [HHb] (precontraction: 28.1 (SE 3.1) μM; end-contraction: 28.5 (SE 3.2) μM; two-tailed, paired Student's t-test, P = 0.23) (data not shown). Arithmetically, these behaviors resulted in a contraction-associated relative decrease in the %HbO2 of 5.0 (SE 0.9) %. The absence of change in [HHb] indicates that there was no increase in oxygen extraction during the contraction, consistent with the assumption above that there was no change in the oxy-Mb saturation.
Postcontraction [THb] Behavior
For 15 of 18 subjects, the [THb] recovery was best described by an initial growth phase, beginning immediately after the contraction and leading to an overshoot beyond the pre-C value and a subsequent decay back to the pre-C value; in the other three subjects, the secondary decay term was not included in the model. Temporally, the growth component can be further divided into the time required for the return of [THb] to the pre-C value (8.4 s) and the remaining time before transitioning to the secondary decay phase (an additional 50 s). The total postcontraction range of [THb] values (max-P to end-C) represents such physiological processes as vasodilation of the terminal arterioles, with a corresponding recruitment of microvascular units, and venous refilling. We would predict that the (pre-C to end-C) difference reflects venous refilling and that the (max-P to pre-C) difference represents microvascular-unit recruitment. We would also predict that these two processes are generally coincident during the first 60 s after contraction. Repeating these measurements in the presence of vasodilator inhibitors would be necessary to test these hypotheses explicitly.
Phases of Oxygen Supply-Demand Coupling after Dorsiflexion MVC
%HbO2 reflects the balance of oxygen supply and oxygen demand. Its triphasic recovery pattern (Fig. 5A) can, therefore, be used to define three phases to the matching of oxygen supply and oxygen demand after brief dorsiflexion MVC. The first phase occurred immediately after the contraction and lasted for approximately 10 s. During this phase, %HbO2 remained constant in 10 subjects and decreased further in eight subjects; overall, %HbO2 decreased (Fig. 4B). This indicates a period of oxygen demand exceeding supply and is similar to previous reports of blood-oxygenation recovery after brief, high-intensity, isometric contractions (4,18). It is also qualitatively similar to the undershoots in microvascular PO2 that have been reported during exercise on-transitions in a rat model of moderate congestive heart failure, in which oxygen delivery is impaired but citrate synthase activity is normal (7). At first inspection, the finding of delayed blood reoxygenation after a single MVC seems to differ from the immediate recovery of %HbO2 that occurs after prolonged bouts of exercise (24). However, this discrepancy probably exists because there would have been significant active hyperemia during the prolonged exercise bouts used in those studies, and thus oxygen supply already would have been elevated at the start of recovery.
Studies of oxygen consumption on-kinetics using single myocytes (13) and intact humans (6,10) have demonstrated a time delay of approximately 10 s before the rate of oxidative metabolism increases. This delay may result from a metabolic inertia resulting from a buffering of the ADP concentration by the creatine kinase and glycolytic reactions (9). In contrast, Behnke et al. (2) have calculated a temporal profile of oxygen consumption on-kinetics that indicates no time delay between the start of exercise and the onset of increase oxygen consumption. Considering that both blood flow (28) and [THb] (present study) increase immediately after isometric contractions, it is reasonable to assume that oxygen delivery also increases during this time. A constant or falling %HbO2 in the face of increasing oxygen delivery must reflect increased oxygen extraction and, presumably, oxygen consumption. This increase begins immediately after the contraction and only 3 s after the start of the contraction. Our data, therefore, agree qualitatively with those of Behnke et al. (2), but given the differences in species, preparation, and exercise models across all of these studies, and without data concerning the concentrations of ADP, inorganic phosphate, and ATP, a strict quantitative analysis is beyond the scope of the present study.
The second phase of oxygen supply-demand coupling began at a postcontraction duration of approximately 10 s. At this time, %HbO2 began to increase, indicating that the oxygen supply was increasing relative to the demand. We note that there is an exact temporal coincidence of the beginning of the relative increase in oxygen supply and the point at which [THb] began to exceed the pre-C value, but this should not be given broad mechanistic meaning, because [THb] measurements in an open system do not reflect blood flow per se. The second phase of oxygen supply-demand coupling continued for an additional 45 s. At this point, %HbO2 began to decrease back to the pre-C value. Again, this transition was almost exactly coincident with the time at which [THb] began to decrease back toward its precontraction level.
Absence of Gender Effects in the Responses
Male and female subjects differed in many anthropometric and physiological characteristics, but the potentially confounding variables of age, physical activity level, and MVC strength were either similar (age, physical activity) or not practically significant (MVC, because in all subjects the MVC force exceeded that required to produce full blood flow occlusion in the TA, as determined by Wigmore et al. (29)). When THKSF was used as a covariate, there were no gender differences observed in either the pre-, end-, and postcontraction [THb] and %HbO2 values (Fig. 4) or in the kinetic parameter estimates (Tables 3 and 4). Given the previously reported findings of greater endothelium-dependent vasodilation in premenopausal women (15), the absence of gender differences differs from our expectations. Gender differences may have not existed because approximately 6 s are required for full activation of the NO pathway (8,22,30), and the 3-s contraction may have provided an insufficiently robust response to distinguish between males and females.
Summary and Conclusions
After brief dorsiflexion MVC, there are three phases to oxygen supply-demand coupling. The first (0-10 s) is a period in which the oxygen demand increases relative to the supply. The second phase (10-55 s) is a period in which the oxygen supply increases relative to the demand. Finally (55-150 s), precontraction oxygen supply-demand matching is restored. Both the overall, qualitative pattern of the responses and the quantitative details regarding the kinetics of these processes are unaffected by gender. The measures provide information related to the kinetics of the hemodynamic and blood-oxygenation responses to brief MVC and might be useful tools for evaluating muscle microvascular and metabolic function in various states of exercise training, environmental adaptation, and peripheral vascular disease.
1. Abboud, R. J., D. I. Rowley, and R. W. Newton. Lower limb muscle dysfunction may contribute to foot ulceration in diabetic patients. Clin. Biomech. (Bristol, Avon)
2. Behnke, B. J., T. J. Barstow, C. A. Kindig, P. McDonough, T. I. Musch, and D. C. Poole. Dynamics of oxygen uptake following exercise onset in rat skeletal muscle. Respir. Physiol. Neurobiol.
3. Clifford, P. S., and Y. Hellsten. Vasodilatory mechanisms in contracting skeletal muscle. J. Appl. Physiol.
4. Corcondilas, A., G. T. Koroxenidis, and J. T. Shepherd. Effect of a brief contraction of forearm muscles on forearm blood flow. J.Appl. Physiol.
5. Cui, W., C. Kumar, and B. Chance. Experimental study of migration depth for the photons measured at sample surface. In: Time-Resolved Spectroscopy and Imaging of Tissues
, B. Chance (Ed.). Bellingham, WA: Society of Photo-optical Instrumentation Engineers, 1991, pp. 180-191.
6. DeLorey, D. S., J. M. Kowalchuk, and D. H. Paterson. Relationship between pulmonary O2 uptake kinetics and muscle deoxygenation during moderate-intensity exercise. J. Appl. Physiol.
7. Diederich, E. R., B. J. Behnke, P. McDonough, et al. Dynamics of microvascular oxygen partial pressure in contracting skeletal muscle of rats with chronic heart failure. Cardiovasc. Res.
8. Gorczynski, R. J., B. Klitzman, and B. R. Duling. Interrelations between contracting striated muscle and precapillary microvessels. Am. J. Physiol.
9. Grassi, B. Delayed metabolic activation of oxidative phosphorylation in skeletal muscle at exercise onset. Med. Sci. Sports Exerc.
10. Grassi, B., S. Pogliaghi, S. Rampichini, et al. Muscle oxygenation and pulmonary gas exchange kinetics during cycling exercise on-transitions in humans. J. Appl. Physiol.
11. Gray, S. D., E. Carlsson, and N. C. Staub. Site of increased vascular resistance during isometric muscle contraction. Am. J. Physiol.
12. Gray, S. D., and N. C. Staub. Resistance to blood flow in leg muscles of dog during tetanic isometric contraction. Am. J. Physiol.
13. Hogan, M. C. Fall in intracellular PO2 at the onset of contractions in Xenopus single skeletal muscle fibers. J. Appl. Physiol.
14. Hogan, M. C., B. Grassi, M. Samaja, C. M. Stary, and L. B. Gladden. Effect of contraction frequency on the contractile and noncontractile phases of muscle venous blood flow. J. Appl. Physiol.
15. Huang, A., D. Sun, A. Koller, and G. Kaley. Gender difference in flow-induced dilation and regulation of shear stress: role of estrogen and nitric oxide. Am. J. Physiol.
16. Hueber, D. M., S. Fantini, A. E. Cerussi, and B. Barbieri. New optical probe designs for absolute (self-calibrating) NIR tissue hemoglobin measurements. In: SPIE Conference on Optical Tomography and Spectroscopy of Tissue III
. San Jose, CA: SPIE, 1999, pp. 618-631.
17. Kent-Braun, J. A., A. V. Ng, M. Castro, et al. Strength, skeletal muscle composition, and enzyme activity in multiple sclerosis. J. Appl. Physiol.
18. Kime, R., T. Katsumura, T. Hamaoka, et al. Muscle reoxygenation rate after isometric exercise at various intensities in relation to muscle oxidative capacity. Adv. Exp. Med. Biol.
19. Kindig, C. A., W. L. Sexton, M. R. Fedde, and D. C. Poole. Skeletal muscle microcirculatory structure and hemodynamics in diabetes. Respir. Physiol.
20. Laughlin, M. H. Skeletal muscle blood flow capacity: role of muscle pump in exercise hyperemia. Am. J. Physiol.
253: H993-H1004, 1987.
21. Laughlin, M. H., R. J. Korthuis, D. J. Duncker, and R. J. Bache. Control of blood flow to cardiac and skeletal muscle during exercise. In: Handbook of Physiology. Exercise: Regulation and Integration of Multiple Systems
, L. B. Rowell and J. T. Shepherd (Eds.). Bethesda, MD: American Physiological Society, 1996, pp.705-769.
22. Marshall, J. M., and H. C. Tandon. Direct observations of muscle arterioles and venules following contraction of skeletal muscle fibres in the rat. J. Physiol.
23. McCully, K. K., and T. Hamaoka. Near-infrared spectroscopy: what can it tell us about oxygen saturation in skeletal muscle? Exerc. Sport Sci. Rev.
24. McCully, K. K., S. Iotti, K. Kendrick, et al. Simultaneous in vivo measurements of HbO2 saturation and PCr kinetics after exercise in normal humans. J. Appl. Physiol.
25. Richardson, R. S., S. Duteil, C. Wary, D. W. Wray, J. Hoff, and P. G. Carlier. Human skeletal muscle intracellular oxygenation: the impact of ambient oxygen availability. J. Physiol. (Lond)
26. Sallis, J. F., W. L. Haskell, P. D. Wood, et al. Physical activity assessment methodology in the Five-City Project. Am. J. Epidemiol.
27. Sejersted, O. M., A. R. Hargens, K. R. Kardel, P. Blom, O. Jensen, and L. Hermansen. Intramuscular fluid pressure during isometric contraction of human skeletal muscle. J. Appl. Physiol.
28. Tschakovsky, M. E., A. M. Rogers, K. E. Pyke, et al. Immediate exercise hyperemia in humans is contraction intensity dependent: evidence for rapid vasodilation. J. Appl. Physiol.
29. Wigmore, D. M., B. M. Damon, D. M. Pober, and J. A. Kent-Braun. MRI measures of perfusion-related changes in human skeletal muscle during progressive contractions. J. Appl. Physiol.
30. Wunsch, S. A., J. Muller-Delp, and M. D. Delp. Time course of vasodilatory responses in skeletal muscle arterioles: role in hyperemia at onset of exercise. Am. J. Physiol. Heart Circ. Physiol.