Mathematical simulations of oxygen delivery to tissue from capillaries that take into account the particulate nature of blood flow predict the existence of oxygen tension (Po2) gradients between erythrocytes (RBCs). As RBCs and plasma alternately pass an observation point, these gradients are manifested as rapid fluctuations in Po2, also known as erythrocyte-associated transients (EATs). The impact of hemodilution on EATs and oxygen delivery at the capillary level of the microcirculation has yet to be elucidated. Therefore, in the present study, phosphorescence quenching microscopy was used to measure EATs and Po2 in capillaries of the rat spinotrapezius muscle at the following systemic hematocrits (Hctsys): normal (39%) and after moderate (HES1; 27%) or severe (HES2; 15%) isovolemic hemodilution using a 6% hetastarch solution. A 532-nm laser, generating 10-μs pulses concentrated onto a 0.9-μm spot, was used to obtain plasma Po2 values 100 times/s at points along surface capillaries of the muscle. Mean capillary Po2 (Pc; means ± SE) significantly decreased between conditions (normal: 56 ± 2 mmHg, n = 45; HES1: 47 ± 2 mmHg, n = 62; HES2: 27 ± 2 mmHg, n = 52, where n = capillary number). In addition, the magnitude of Po2 transients (ΔPo2) significantly decreased with hemodilution (normal: 19 ± 1 mmHg, n = 45; HES1: 11 ± 1 mmHg, n = 62; HES2: 6 ± 1 mmHg, n = 52). Results suggest that the decrease in Pc and ΔPo2 with hemodilution is primarily dependent on Hctsys and subsequent microvascular compensations.
- oxygen transport
- oxygen tension gradients
- phosphorescence quenching microscopy
quantitative measurements of oxygen tension (Po2) gradients in capillaries as erythrocytes pass through them have been made only recently (7). In 1977, the existence of intracapillary Po2 gradients, which take into account the discrete nature of blood flow in capillaries, was predicted theoretically by Hellums (11). Recently, an in vivo experiment substantiated that these Po2 gradients in the plasma between red blood cells (RBCs) do indeed exist (7). The exact nature of these gradients and their effect on the diffusion of oxygen from the blood to tissue has yet to be fully elucidated.
Hellums coined the term “erythrocyte-associated transients” (EATs) (11), which refers to the variation in Po2 caused by the alternate passage of RBCs and plasma gaps as observed at a stationary observation point along a capillary. These transients, which can also be described as spatial gradients, are a manifestation of the Po2 heterogeneity in the plasma between RBCs (13). Therefore, rapid variations of Po2 can be observed as a function of microcirculatory factors, such as hematocrit (Hct), blood flow, or oxygen content. The amplitude of this inter-RBC gradient has been predicted by theoretical models to vary according to cell spacing, shape, and orientation, as well as oxygen flux at the capillary wall, RBC-to-wall spatial clearance, and oxygen consumption rate of the tissue (2, 5, 6, 10, 12, 14, 21, 29).
Theoretical considerations of the discrete nature of capillary blood flow have suggested that inter-RBC Po2 heterogeneity increases the intracapillary fraction of resistance to oxygen transport to twice that compared with previous models assuming a homogenous distribution of hemoglobin (11). This stems from the low solubility of oxygen in the plasma, which hinders diffusive oxygen movement from the RBC through the surrounding plasma sheath and plasma gaps into the tissue. Consequently, RBC separation distance plays an important role in the rate of oxygen transport from RBC to tissue (13). Under conditions of rest or maximal tissue oxygen consumption, low plasma Po2 can become limiting for oxygen exchange, dependent on a critical RBC separation distance defined by a threshold for maintaining uniform oxygen flux at a point along a capillary (6).
Previous difficulties preventing the measurement of capillary Po2 transients have been overcome by substantially increasing the temporal and spatial resolution of the phosphorescence quenching method with pulsed laser excitation of the phosphorescent probe. Golub and Pittman (7) previously reported measurements of EATs in rat mesenteric capillaries using rapid laser excitation to detect fast Po2 transients. The purpose of the present study was to extend these observations to a three-dimensional tissue and to observe EATs in the capillaries of the rat spinotrapezius muscle.
Furthermore, hemodilution has been used therapeutically to increase peripheral perfusion to ischemic tissues. Rheological alterations, such as a decrease in blood viscosity (a function of hematocrit), combined with peripheral vasodilation have been credited with increased flow through vascular networks. In addition, hemodilution has been found to influence the distribution of RBC flow and Hct in microvascular networks (22). During hemodilution, the systemic hematocrit (Hctsys) is reduced, and therefore the oxygen-carrying capacity of the blood is diminished. The role of capillaries in oxygen delivery under such conditions has yet to be investigated.
The present study attempts to understand the localized changes in capillary oxygen tension as a function of reduced Hctsys. Isovolemic hemodilution, resulting in a dramatic reduction of Hctsys, could affect capillary Hct and cause an increase in RBC separation distance. Therefore, defining the nature of EATs where Hctsys is reduced to a moderate or severe level could provide further information regarding capillary oxygen exchange and microvascular compensations due to hemodilution.
MATERIALS AND METHODS
A total of 12 female Sprague-Dawley rats (244 ± 3.7 g; Harlan, Indianapolis, IN) were used in this investigation. One group (normal, n = 4; Hct = 38.8%) was used to assess nonhemodiluted conditions, and two other groups (HES1, n = 4; HES2, n = 4) were used to assess isovolemic hemodilution conditions at two different systemic hematocrits (26.6 and 14.8%, respectively) using Hespan (6% hetastarch, a high-molecular-weight hydroxyethyl starch; DuPont Pharma, Wilmington, DE) as the hemodiluent. Animals were initially anesthetized with a combination of ketamine-acepromazine (72 and 3 mg/kg ip, respectively). Following cannulation of the right jugular vein for administration of supplemental anesthetic, surgical preparation and measurements were conducted under a continuous intravenous infusion of alfaxalone/alfadolone (0.1 mg·kg−1·min−1 Saffan; Schering-Plough Animal Health, Welwyn Garden City, UK). Upon conclusion of experimentation, the animals were euthanized with administration of 0.4 ml/kg iv Euthasol (390 mg/ml pentobarbital sodium and 50 mg/ml phenytoin; Delmarva Laboratories, Midlothian, VA). All procedures and protocols were approved by the Virginia Commonwealth University Institutional Animal Care and Use Committee.
Initially, the trachea was cannulated (PE-240; Clay Adams, Parsippany, NJ) to ensure a patent airway and spontaneous breathing of room air. Subsequent cannulation of the right jugular vein with PE-90 tubing and the left carotid artery with PE-50 tubing allowed for continuous intravenous administration of anesthetic and monitoring of mean arterial pressure (CyQ 103/301; Cyber Sense, Nicholasville, KY), respectively. The femoral artery was cannulated with PE-10 tubing and used for the withdrawal of blood during the isovolemic hemodilution procedure and blood sampling for arterial blood gas measurements, including Po2, Pco2, and pH (ABL 705; Radiometer America, Westlake, OH). In addition, total hemoglobin (Hb) and percent arterial hemoglobin oxygen saturation (SaO2) were measured using a hemoximeter (Radiometer OSM3; Radiometer, Crawley, UK). Hct was determined by capillary tube centrifugation of arterial blood. The femoral vein was cannulated with PE-20 tubing and utilized for infusion of Hespan and the albumin-bound Pd-porphyrin phosphorescence probe. The spinotrapezius muscle was partially exteriorized using a modification of the procedure originally described by Gray (9) and subsequently modified by Bailey et al. (1). Care was taken to keep the muscle moist with continuous superfusion of Ringer solution, and sutures were placed at 1-cm intervals along the lateral margins to facilitate surgical and experimental manipulation, ensuring minimal trauma to the spinotrapezius muscle. After exteriorization of the spinotrapezius, the animal was placed on its side on a thermostatic animal platform described by Golub and Pittman (8). The muscle was secured dorsal side up and spread to its physiological length by using the sutures along the lateral margins. Finally, Ringer solution was applied topically, and the muscle was covered with Saran film (Dow Chemical, Indianapolis, IN) to prevent tissue desiccation and atmospheric gas exchange.
Phosphorescence quenching microscopy.
The experimental setup previously described by Golub and Pittman (7) was used to measure plasma Po2 and maximum phosphorescence amplitude (PA) in capillaries and to measure light transmission (LT) through a capillary. The animal platform was placed on the stage of an Ortholux microscope (Leitz, Stuttgart, Germany) configured for epi-illumination through a ×100/1.30 objective (Leitz). Immersion oil was placed between the objective and the Saran film covering the spinotrapezius muscle.
Phosphorescence excitation was achieved with a 532-nm laser (GMS1-038-15T; Lasermate, Pomona, CA), which delivered 60 nJ for a duration of 10 μs to a 0.9-μm-diameter region at 100 Hz. A photomultiplier tube (R3896 with high-voltage socket HC123-01; Hamamatsu, Bridgewater, NJ) detected the emitted phosphorescence collected by the ×100 objective. The signal from the photomultiplier tube was directed through a gated amplifier (OP37EP; Analog Devices, Norwood, MA), designed to be a current-to-voltage converter, and then to a 12-bit analog-to-digital converter (PC-MIO-16E-4; National Instruments, Austin, TX). The digital data were stored on a Dell personal computer (Round Rock, TX).
The laser light passing through the capillary was detected with a T-5 photodiode (Intor, Socorro, NM) integrated with a 532-nm laser line filter. The LT signal collected from the diode was passed through an amplifier and directed to the second channel of the analog-to-digital converter. Both the phosphorescence and LT signals were monitored with an oscilloscope (72-3060; Tenma, Springboro, OH) for amplitude adjustments.
Capillary selection for measurement and positioning the laser spot on the vessel were carried out under transillumination using an OG-570 filter (Edmund Optics, Barrington, NJ) to prevent excitation of the phosphor. Measurements of phosphorescence excitation and LT were carried out in a darkened room, and other background light sources were minimized to reduce noise caused by stray light.
Data acquisition was performed using a program written in LabVIEW (National Instruments) at a sampling rate of 200 kHz with 400 data points per curve and 100 curves per measurement sequence. The analysis program was also written in LabVIEW and was used for the off-line determination of Po2 for the 100 sequential curves collected. Nonlinear fitting of the phosphorescence decay curves was based on the rectangular Po2 distribution model (30) where t (μs) is the time from the beginning of phosphorescence decay, I(t) (V) is the magnitude of the phosphorescence signal, I(0) (V) is the amplitude of the phosphorescence signal at t = 0, M (mmHg) is the mean Po2, δ (mmHg) is the half-width of the Po2 distribution, T (μs) is the lifetime of the fast postexcitation transient, A (V) is the amplitude of the fast postexcitation transient and B (V) is the baseline offset. Stern-Volmer constants k0 = 18.3 × 10−4 μs−1 and kq = 3.06 × 10−4 μs−1·mmHg−1 were determined in a separate study (30).
Capillary inclusion parameters and probe administration.
Only capillaries supporting continuous RBC flow were selected for study. Before data collection, each capillary chosen for study had to meet the following additional specified selection criteria: 1) sufficient LT with evident RBC fluctuations as observed on the oscilloscope and 2) spatial separation of the capillary from other microvessels. Measurements were made in the centerline of the capillary in a region of high visual contrast.
The phosphor probe (30 mg/kg) was administered via the right femoral vein before measurements were taken, and ∼1 min of equilibration time was allowed before initiation of data collection to ensure a homogeneous probe distribution in the plasma. Based on a cardiac output of 100 ml/min and a circulating volume of 15 ml, the circulatory time for a 244-g rat would be 9 s (25). During a period of 1 min, probe was distributed throughout the circulation and a stable amplitude of the phosphorescence signal detected. Experimental measurements were then made within a 30-min period to exclude nonintracapillary sources from the phosphorescence signals that would arise following eventual extravasation of the albumin-bound probe.
LT, PA, and Po2.
As independent corroborative measures associated with the detection of Po2 fluctuations, LT and maximum PA were measured. LT was measured via photodiode registration of the intensity of the laser light passing through a capillary. Maximum PA was determined by analyzing the beginning of the decay curve after the end of the excitation pulse. Upon analysis, the maximum amplitude of the phosphorescence decay curve and the light transmittance pulses were averaged using the three highest points at the peak of each signal.
The maximum amplitude of the phosphorescent signal is proportional to the amount of probe present in the detection volume. A homogeneous distribution of the dissolved probe in plasma was demonstrated by Golub and Pittman (7) using densitometric scanning and phosphorescence imaging of the Pd-porphyrin probe location, confirming that probe does not diffuse into RBCs. Therefore, PA increases when there is more probe in the detection region, which occurs when there is more plasma present (i.e., a plasma gap between RBCs). Observations of the relative minima and maxima of LT and PA are expected to be correlated with observed maxima and minima of Po2, respectively, as RBCs and plasma gaps alternately pass through the detection region.
Isovolemic hemodilution procedure.
A stepwise isovolemic blood exchange protocol was followed, using Hespan as the hemodiluent. The volume of each withdrawal/infusion step was calculated as a percentage of blood volume estimated at 6% of the body weight. The target Hcts were 25% for moderate hemodilution and 15% for severe hemodilution. Each step took ∼90 min to complete. An additional 15 min of equilibration time for the animal was allowed posthemodilution before systemic measurements and probe infusion. For measurements in nonhemodiluted animals, the animals were allowed a 15-min equilibration period after surgery before systemic measurements and probe infusion.
The “peak detection” method in Origin version 6.1 (OriginLab, Northampton, MA) was used to determine peak values on the unfiltered data, corresponding to the near-RBC Po2, and troughs, corresponding to the lowest Po2 in the plasma gap. Po2 peak values, as well as the corresponding times these peaks occurred during the 1-s measurement period, were determined using this time series analysis method. In addition, the same method was applied to detect the times when peak minima and maxima occurred in the LT and PA time tracings corresponding to the Po2 time tracings for a given capillary. This information was used to help substantiate the presence or absence of RBCs in the detection volume.
The Origin analysis tool “pick peaks” uses a search rectangle in which width and height (the percentage of the total number of points in the data range, i.e., time) and the total amplitude of the data in the range (i.e., LT, PA, or Po2, respectively) can be specified. A peak is rejected if it does not meet the minimum height criterion, the percentage of the total amplitude of the data in the range relative to the minimum determined within the specified width. The test point is at the top center of the rectangle, so any peak found may not correspond to an absolute local maximum. This situation can occur when data are asymmetrically distributed around a relative maximum (Origin). A 5 × 5 search rectangle with a minimum height of 5 was used to detect both positive and negative peaks simultaneously.
For the 100 consecutive Po2 measurements collected over the 1-s measurement period, all positive peaks detected were averaged, giving a mean peak Po2 (Po2Max), and all negative peaks detected were averaged, resulting in the mean trough Po2 (Po2Min). This was done for each capillary Po2 time course for all three conditions (total of 159 capillaries). The individual mean capillary Po2 (Pc) is an average Po2 of the 100 sequential curves collected (unfiltered) over the acquisition period of 1 s. The difference between Po2Maxand Po2Min is termed the ΔPo2. In addition, RBC flux was estimated, based on the number of positive peaks detected by the time series analysis (i.e., peak detection method) using the unfiltered Po2 data from each capillary.
A mass balance approach was used in the determination of oxygen loss between the systemic arterial blood and the site of capillary Po2 measurements for each capillary (see appendix). SaO2 was measured from arterial blood obtained from the carotid artery. Capillary oxygen saturation (Sc) was estimated from the Hill equation, using capillary Po2 values and P50, recalculated due to differences in pH and Pco2 under the different conditions (i.e., normal, HES1, HES2). To account for the shift in the oxygen dissociation curve, the Severinghaus equation (24) was used as follows: where P is the standard P50 of Sprague-Dawley rats (36 mmHg; Ref. 3), pH is the experimental value determined by the blood gas analyzer, T is the stabilized temperature of the rat spinotrapezius, and BE refers to the base excess, also determined by the blood gas analyzer. Changes in blood pH and BE between the arterial and capillary measurement sites were neglected. Sc was then estimated from the Hill equation, where Po2Max is the maximum capillary Po2 associated with the presence of a RBC, P50 is the shifted value determined as described above, and nH is 2.7, the Hill coefficient (15). Equation A10 (see appendix) was then used to calculate the oxygen loss per RBC (LR) as follows: where ΔSo2 is SaO2 − Sc, [Hb]RBC is the average hemoglobin concentration per RBC (see Table 1), and CHb is 1.34 ml O2/g Hb, the oxygen binding capacity of hemoglobin.
ANOVA and the Tukey-Kramer post hoc test were conducted with the software JMP version 4 (SAS Institute, Cary, NC) to assess statistical significance of reported values between conditions. Correlation coefficients were determined between paired data sets derived from the Po2, LT, and PA time tracings. A paired t-test was used to determine significance between the time points where both Po2Max and Po2Min occurred with the Po2 values of the preceding time points to establish signal quality. All data are presented as means ± SE. A statistical significance level of P < 0.05 was accepted.
Systemic data, capillary Po2, and Hct.
A total of 159 capillaries were evaluated in the 12 female Sprague-Dawley rats. There were no significant differences in mean arterial pressure (MAP) among normal (99.8 ± 7.4 mmHg), HES1 (114.7 ± 9.9 mmHg), and HES2 conditions (96.2 ± 8.9 mmHg). SaO2 also was not significantly different among the three conditions (93.6 ± 2.1, 92.8 ± 0.7, and 94.0 ± 1.5% for normal, HES1, and HES2, respectively).
Arterial Po2 (PaO2) varied significantly between the normal (97.0 ± 6.1 mmHg) and HES1 conditions (76.1 ± 1.7 mmHg), whereas PaO2 for the HES2 condition (82.6 ± 6.0 mmHg) was not significantly different from either the normal or HES1 value. Following hemodilution, the Hctsys was reduced significantly from 38.8 ± 0.9% (normal condition) to 26.6 ± 1.3% (HES1, moderate hemodilution) and then to 14.8 ± 1.3% (HES2, severe hemodilution). As the Hctsys was reduced, Pc also decreased. Pc was 56.5 ± 2.2 mmHg (n = 45) for the normal condition, 47.2 ± 1.7 mmHg (n = 62) after moderate hemodilution, and 26.8 ± 1.5 mmHg (n = 52) after severe hemodilution, with a significant difference evident among all conditions. Table 1 summarizes the systemic variables for the three conditions, and comparisons of PaO2, Pc, and Hctsys among the three conditions are shown in Fig. 1.
Evidence of EATs in the spinotrapezius muscle.
Significant fluctuations in Po2 were measured at selected observation points along spinotrapezius muscle capillaries during the 1-s time course at 100-Hz flash excitation under the three Hct conditions. An example of Po2 fluctuations in a HES2 capillary is shown in Fig. 2, where Po2Min is 24 mmHg, Po2Max is 29 mmHg, and Pc is 27 mmHg. In addition, Fig. 2 shows the time points at which maximum and minimum peaks were determined. Since Po2 maxima and minima should correspond with the minima and maxima, respectively, of LT and PA, the specific time points at which peaks occurred in the measurement period were correlated. An example of these correlations from the same HES2 capillary are depicted in Fig. 3, where the time points of peak occurrences for both Po2 maxima and minima (t and t) are plotted against the corresponding time points of peak occurrences of minima and maxima, respectively, for LT (t and t; A) and PA (t and t; B). Results demonstrate that a high correlation (R = 0.99, P < 0.0001) exists between the time points of LT or PA minima and maxima with the time points at which Po2Max and Po2Min, respectively, occur. The correlation between LT and PA time points (Fig. 3C) is also good (R = 0.99, P < 0.0001).
Po2Max and Po2Min were determined in all capillaries in each of the three conditions and are summarized in Table 2 using the peak detection method. Briefly, the Po2 fluctuated between averages of 47.4 and 66.0 mmHg, 41.9 and 52.7 mmHg, and 23.6 and 30.0 mmHg for the normal, HES1, and HES2 conditions, respectively. The Po2 values of the preceding time points for the occurrences of Po2Max and Po2Min [Po2(ti − 1)] were compared with the Po2 values of the specific time points where Po2Max and Po2Min occurred (Po2ti). This was done in five capillaries chosen at random to determine signal quality and to further substantiate the use of the peak detection method for the determination of relative minima and maxima. For the HES2 example capillary, the mean difference between Po2ti and Po2(ti − 1) (0.8 mmHg) was not significant, as was the case in all five capillaries. However, all ΔPo2 values (12, 17, 6, 9, and 15 mmHg) for these five capillaries were significant compared with the local Po2 differences (0.07, 0.02, 0.9, 0.2, and 0.3 mmHg, respectively) around the chosen peak values.
Also, from the peak detection method of analysis using individual 1-s time tracings of Po2, capillary cell flux was determined to be 10.9 ± 0.23, 11.3 ± 0.19, and 11.4 ± 0.23 cells/s for normal, HES1, and HES2 conditions, respectively, that passed through the detection region with no significant differences among conditions. As an approximation, each peak can be assumed to represent one RBC, yet under some conditions where cells might pass the detection region in tandem, this cell flux value could be an underestimate.
Po2Max, corresponding to the Po2 value associated with the RBCs, decreased as the Hctsys was reduced by isovolemic hemodilution with Hespan. The difference between Po2Max and Po2Min (ΔPo2) was found to decrease from 18.5 ± 1.1 mmHg under the normal condition to 10.9 ± 0.6 mmHg after moderate hemodilution and then to 6.5 ± 0.5 mmHg following severe hemodilution (Table 2). Thus Po2 fluctuations were reduced by 41 and 65% for moderate and severe hemodilution, respectively, compared with the normal condition. Figure 4 shows ΔPo2 plotted against Po2Max. This graph indicates that during isovolemic hemodilution with Hespan, ΔPo2 is related to Hct and RBC Po2 (Fig. 5).
Oxygen saturation and oxygen loss.
Table 3 compares SaO2 and Sc, in addition to the difference between these two values (ΔSo2) and the calculated oxygen loss per RBC (LR) for the normal, HES1, and HES2 groups. There were significant differences in Sc among all conditions, whereas only the HES2 group was significantly different for both ΔSo2 and LR. The oxygen saturation differences (ΔSo2 = SaO2 − Sc) were 17.4 ± 2.2, 24.1 ± 2.3, and 54.8 ± 2.9% for the normal, HES1, and HES2 conditions, respectively. As previously stated, LR provides an estimate of oxygen loss per RBC as blood flows from large arteries and arterioles predominately characterized by bulk convective oxygen delivery to the capillaries, characterized by diffusive delivery of oxygen to the tissue. LR was not significantly different between the normal and HES1 conditions (8.2 ± 1.1 and 10.9 ± 1.0 ml O2/100 ml, respectively); however, HES2 (24.7 ± 1.3 ml O2/100 ml) was significantly different from the HES1 and normal conditions (data also summarized in Table 3). Values shown in Table 3 are also plotted against Hct in Fig. 6.
With the use of an isovolemic hemodilution procedure with Hespan, Hctsys in rats was significantly reduced from the normal value (38.8 ± 0.9%) to a moderate (26.6 ± 1.3%) or severe (14.8 ± 1.3%) level. At each of the three levels, fluctuations in capillary Po2 were measured at randomly selected points in rat spinotrapezius muscle capillaries using phosphorescence quenching microscopy. The magnitude of Po2 fluctuations, ΔPo2, as well as the mean capillary Po2 and near-RBC Po2 (Po2Max) decreased with hemodilution. In addition, by using the peak detection method, maximum and minimum peaks were identified in the Po2, LT, and PA time tracings. The high correlations of the corresponding time points, where specific maxima and minima occurred in the collected signals, demonstrate that these three independent signals corroborate the presence or absence of RBCs in the detection region. Despite hemorheological and hemodynamic changes that might have occurred with hemodilution, results indicate that RBC flux did not vary significantly among experimental groups. Furthermore, Sc decreased with hemodilution, whereas oxygen loss per RBC increased significantly following severe hemodilution, relative to the normal and HES1 conditions.
Laser excitation of the Pd-porphyrin phosphorescent probe confined to the plasma allowed the measurement of plasma Po2 transients and PA, proportional to the amount of plasma in the detection region. The particulate nature of blood flow was also revealed by the simultaneous measurement of LT through the capillary. As a result of the intrinsic nature of RBCs to absorb light, the signal would decrease in the presence of a RBC and increase once the RBC moved out of the detection region. The signal for PA would correspondingly increase, reflecting the presence of plasma measured by the amount of light emitted by the excited phosphor probe in the detection region, and decrease when an RBC was present (i.e., less plasma).
Although the time points of maxima and minima were highly correlated among the three independent variables (P < 0.0001 in all cases), some noise is still present in the signals. Golub and Pittman (7) found that, due to the complex scattering of light (i.e., reflection and refraction) by a RBC with random size, shape, and orientation, PA was a better indicator of the presence of plasma or RBC. In the present study we could not conclude whether LT or PA was a better independent corroborative measure, perhaps because of other factors that might have contributed to variability between signals in skeletal muscle, including muscle thickness, refraction of transmitted light by muscle tissue, and the presence of multiple microvessels along the incident light beam.
Mathematical predictions of oxygen tension associated with RBCs and plasma in capillaries suggest that significant Po2 gradients can exist in the plasma between RBCs (11, 13). In early modeling studies, ΔPo2 was predicted to be no greater than 6 mmHg for conditions of rest, anemia, or exercise (13, 16). Using RBC shape as a determinant of oxygen transport, Wang and Popel (29) calculated Po2 amplitudes as great as 20 mmHg under simulated conditions of maximal tissue oxygen demand. Assuming normal Hct under resting conditions, a maximum Po2 oscillation of 10 mmHg has also been predicted (14). Homer et al. (13) calculated that the maximum Po2 difference between RBCs and plasma would be more pronounced under conditions of severe acute anemia, reasoning that the distance between red cells would increase and therefore cause the Po2 to be depressed in the plasma gap between cells. In the current study this expectation was not observed; rather, the largest difference in oxygen tension was found to occur at normal Hctsys. A conclusion reached in several mathematical models analyzing oxygen tension between RBCs and plasma is that cell spacing is a critical determinant of the difference observed between the near-red cell and plasma gap Po2 (5, 6, 13, 16, 17). Therefore, many models predict that with hemodilution, the spatial distribution of RBCs changes so that they become spaced further apart as they flow through the capillaries. This prediction assumes that in the process of hemodilution, where the Hctsys is reduced, there is a concomitant decrease in capillary hematocrit whereby RBC spacing increases. This assumption implies that the smaller number of RBCs is distributed among the same capillaries, thereby increasing the average RBC spacing. Results in the current study show that with isovolemic hemodilution, the difference between the near-RBC and plasma gap Po2 instead decreases, suggesting that the distribution of RBCs among capillary networks is significantly altered.
Tsai et al. (28) studied hemodynamic parameters in resting rabbit skeletal muscle as the Hctsys was gradually reduced to 26% of normal by isovolemic hemodilution with a 6% dextran solution. Their results suggest that the spacing between RBCs during capillary transit is independent of the reduction in Hctsys. They also reported that RBC flux was maintained by an increase in RBC velocity. Similar results have been reported by Sarelius (23), who measured capillary blood flow variables in the hamster cremaster muscle after isovolemic hemodilution. In this case, no change was found in capillary RBC flux or velocity at Hctsys values 32 and 49% of normal. This study also concluded that capillary RBC flow and capillary RBC content (cells/100 μm) were not directly proportional to the changes in Hctsys. In the present study, capillary RBC flux was estimated by counting the number of peaks in the Po2 signal over the 1-s measurement period, and the results are consistent with previous studies indicating no significant change in capillary RBC flux with hemodilution in skeletal muscle (18, 23, 28).
Because EATs were the focus of the current investigation, only capillaries perfused by RBCs were selected for study. This raises the issue of how systemic hemodilution affects the distribution of RBCs within the microcirculation, especially in capillary networks. From previous studies (22), it appears that hemodilution results in higher RBC velocity and increased axial migration of RBCs in arterioles. Side branches from the proximal parts of the arteriolar network are more likely to contain blood that is much lower in Hct due to plasma skimming. The capillary networks derived from those branches will hence be perfused mostly with plasma and a paucity of RBCs. Accordingly, the capillary networks derived from the more distal parts of the arteriolar network will be the recipients of most of the RBCs, consistent with the finding that in those capillaries perfused by RBCs, capillary flux remains relatively unchanged following hemodilution. This overall increase in the heterogeneity of RBC flow among capillary networks results in a maldistribution of oxygen delivery, where some capillary networks receive an inadequate supply of RBCs and therefore oxygen, whereas others are adequately supplied with RBCs with reduced So2 under hemodiluted conditions. Also, as a result of redistribution of RBCs, cell spacing in the smaller number of RBC-perfused capillaries will not vary much from that at normal Hct. This could be part of the explanation as to why the magnitude of the EATs (i.e., ΔPo2) does not increase with reduced Hctsys, and it appears that the distribution of RBCs within capillary networks with hemodilution may result in maintained RBC spacing in fewer networks (i.e., functional capillary density decreases roughly in proportion to the degree of hemodilution).
Also, in contrast to the predictions of mathematical models, as the animals were hemodiluted with a 6% hetastarch solution, the Po2 in the capillaries became more homogeneous as reflected by the significant decrease in the ratio ΔPo2/Pc(proportional to the coefficient of variation) from normal (see Table 2). The exchange of blood with a non-oxygen-carrying fluid caused a decrease in the overall systemic oxygen-carrying capacity of the blood via a reduction in Hctsys, which also contributed to the decrease in plasma Po2. From the data collected, it can be concluded that there was a greater change in near-RBC Po2 than there was in plasma gap Po2 as the animals were hemodiluted, indicating that homogeneity on a local level is more a function of the RBC Po2 decrease than the plasma gap Po2 decrease (see Fig. 5). This can be explained by considering oxygen exchange between discrete regions of relatively high Po2 (i.e., RBCs) or relatively low Po2 (i.e., plasma gaps) and the even lower Po2 in the tissue. As a result of the oxygen diffusion gradient between these respective regions and the tissue, the higher oxygen flux from the RBC region would lead to a larger Po2 drop for the RBCs than for the plasma gaps.
In acutely anemic animals with significantly reduced Hct and total Hb concentration, normal oxygen consumption is maintained by parallel increases of cardiac output and oxygen extraction (19). Although cardiac output was not measured in this study, we report that ΔSo2 increased significantly at 15% Hct, indicating that oxygen extraction increased significantly between the large arterioles and the capillary measurement sites. Swain and Pittman (27) reported that about two-thirds of the oxygen loss occurring between arterioles and venules actually takes place in precapillary arterioles. The current study also demonstrates that a substantial amount of oxygen loss must occur before the blood reaches the center of capillaries (where, on average, measurements were made) and that the degree of loss is augmented by hemodilution. Although the systemic arterial Po2 stayed relatively steady, the mean capillary Po2 showed a significant decreasing trend (Tables 1 and 2), reflecting greater oxygen loss from the arteries to the middle of capillaries under such conditions. The finding of the oxygen saturation difference above shows that as the oxygen-carrying capacity is compromised, the burden of tissue supply of oxygen from each erythrocyte increases. This oxygen loss per RBC is also represented in Table 3 as LR, demonstrating that the largest oxygen loss occurs under severe hemodilution conditions, three times the loss per RBC compared with normal. Therefore, isovolemic hemodilution results in the depression of RBC Po2 with a lesser depression of plasma Po2 contributing to dampened EATs (shown in graphic form in Figs. 5 and 6).
In addition, on the basis of theoretical modeling, we hypothesized that ΔPo2 would increase with isovolemic hemodilution. Results demonstrate that this does not occur. Rather, the analysis of Po2 data revealed that as the Hctsys was reduced, the ΔPo2 decreased from roughly 18 mmHg (normal) to 11 mmHg after moderate hemodilution (HES1) and then to 6 mmHg after severe hemodilution (HES2). Figure 7 shows the relationship between ΔPo2 from the different conditions and LR. As the hemodilution becomes more severe, the LR increases and ΔPo2 decreases. This is consistent with the explanation above concerning the greater burden, with decreasing Hctsys, of each erythrocyte to supply oxygen to the tissue under unchanged conditions of demand, despite an increase in cardiac output.
Stein and Ellsworth (26) evaluated So2 and Po2 at both ends of capillary networks in the hamster retractor muscle. They found that So2 and Po2 at the venular end of capillary networks were significantly less in hemodiluted animals where the Hctsys was reduced by 40%. Our results showed significant decreases in So2 and Po2 after Hctsys was reduced by 31 and 62% in HES1 and HES2 groups, respectively. In the extensor digitorum longus muscle of rats, Ellis et al. (4) measured So2 at the arteriolar and venular ends of capillaries and found that in control animals, the So2 was 67.7 ± 7.1% at the arteriolar end and 56.7 ± 21.0% at the venular end. These data are consistent with our nonhemodiluted animal group and validate our method of So2 determination.
Although this study used the most current method of phosphorescence quenching microscopy, difficulties still exist in making such measurements in muscle. Initially, the rat mesentery was used to study EATs, where the two-dimensional microvascular network of the mesentery assured a limited contribution of extravasated albumin-bound probe to the phosphorescence signal from the interstitial space or from any microvessels in the laser light path. The three-dimensional anatomy and high vascular spatial density of the spinotrapezius muscle increased the probability of signal artifact from other compartments, including microvessels and the interstitial space, along the incident light path. To minimize the effects on LT and PA from extraneous microvessels and the tissue, we employed a 30-min time limit for capillary measurements based on published observations that extravasation of the phosphorescent probe over 35 min of perfusion was insufficient to yield a detectable phosphorescence signal in skeletal muscle (20). Each capillary chosen for the present study also met strict inclusion criteria to ensure that no other microvessels were in the incident laser light path.
In conclusion, the present study shows that EATs can be identified in the rat spinotrapezius muscle as independently corroborated by the Po2, LT, and PA signals under normal conditions and after isovolemic hemodilution. Furthermore, the largest ΔPo2 occurs under normal conditions, and the Po2 gradients in plasma, as well as mean capillary Po2, decrease as the Hctsys is reduced via isovolemic hemodilution. These results differ from some theoretical predictions, in part because of assumptions in the models that were required to simplify the complex conditions of the microcirculation. Several theoretical predictions have emphasized the importance of RBC spacing on plasma Po2 gradients. The current study found no evidence of significant changes in RBC spacing (near constant RBC flux for different Hct values) so that capillary EATs under the conditions of study appear to be mostly dependent on the Po2 and So2 of RBCs entering the capillary. The decreased oxygen-carrying capacity caused by reduced hematocrit increases the burden of each erythrocyte to supply oxygen to the tissue, especially under conditions of severe hemodilution. In addition, the effect of systemic hematocrit on the distribution of RBC flow in microcirculatory networks due to rheological alterations (i.e., reduced viscosity and increased axial migration) has a profound influence on the magnitude of EATs. These conclusions strongly suggest that during isovolemic hemodilution, passive compensatory mechanisms of oxygen delivery exist in the microvasculature that alter EATs in contrast to what might be hypothesized on the basis of theoretical works.
To interpret capillary Po2 measurements under conditions of hemodilution where some variables could not be measured, we employed the following scenario. One of two cases is assumed to exist in a given capillary: 1) it is perfused only by plasma, or 2) it is perfused by RBCs and plasma. In the case of the plasma-perfused capillaries, the number of total plasma-perfused capillaries is NP, flow is Qc,P, and capillary hematocrit is by definition Hctc,P = 0. For RBC-perfused capillaries, the total number of capillaries is NR, flow is Qc,R, and capillary discharge hematocrit is Hctc,R.
Note that functional capillary density can be defined as FCD = NR/(NP + NR). The input artery, in this case the spinotrapezius feed arteriole, is assumed to have the same values for oxygen tension (PaO2), oxygen saturation (SaO2), and hematocrit (Hcta) as systemic arterial values (i.e., negligible loss of oxygen and flow separation, respectively, between the systemic artery and feed arteriole). All blood flow entering through this artery (Qa) is assumed to go through either the plasma- or RBC-perfused capillaries. Therefore, conservation of blood flow gives (A1) Assuming that all RBCs entering the artery go through the RBC-perfused capillaries, then conservation of RBC flow gives the following relationship, since the plasma-perfused capillaries do not contain any RBCs: (A2) The convective oxygen flow for the upstream inflow vessel (neglecting oxygen carried by the plasma) is (A3) where [O2]a is arterial oxygen content, SaO2 is arterial oxygen saturation, [Hb]RBC is the average Hb concentration in a single RBC, and CHb is the oxygen-binding capacity of Hb. For the capillaries, Qc, is (A4) where the term pertaining to plasma flow can be neglected because [O2]c,P is very small compared with [O2]c,R. Therefore, the equation can be simplified as (A5) where ScR is calculated from the Hill equation using the Po2 of the RBCs.
The rate at which oxygen diffuses from the arteriolar network between the arterial inflow and the middle of the RBC-perfused capillaries (“middle” reflects that the capillary sites are chosen at random positions along a capillary) is given by the following: (A6) The rate of oxygen loss from the network between the inflow artery and middle of the capillaries relative to the oxygen inflow to the network is (A7) From conservation of RBC flow (Eq. A2), we have (A8) where ΔSo2 = SaO2 − ScR. The average oxygen loss per RBC is then (A9) Again, using Eq. A2, we have (A10)
This work was supported in part by National Heart, Lung, and Blood Institute Grant HL-18292, American Heart Association (Mid-Atlantic Affiliate) Grant 065449U, and the A. D. Williams Foundation.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 by the American Physiological Society