INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

SEVERE ANEMIA is associated with remarkable stability of pulmonary gas exchange and transport, despite a fall in the carrying capacity of blood for O₂ and CO₂. Several factors contribute to this stability, including an increased cardiac output () and increased O₂ extraction, the latter contributing also to greater CO₂ transport by the Haldane effect (11).

Previous studies in this area have suggested that the gas-exchange efficiency of the normal lung may be improved with acute hemodilution, as evidenced by higher arterial PO₂ (Pa_O2) and lower alveolar-arterial PO₂ difference [(A-a)PO₂] (although this finding is by no means consistently observed; see Table 1). In the presence of venous admixture, anemia has been postulated to result in either an increase (50) or a decrease (57) in Pa_O2, depending on the effect on mixed venous PO₂ (). Because anemia results in a fall in blood viscosity, a secondary increase in , and changes in microcirculatory blood flow, acute hemodilution may also result in reduced heterogeneity of pulmonary blood flow and an improvement in ventilation-perfusion (A/) distributions (40, 66). Because gas exchange, A/ heterogeneity, and pulmonary blood flow distribution have not been systematically studied, we explored these factors in a rabbit model of acute normovolemic hemodilution.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ten adult Flemish Giant rabbits, weighing ~5 kg each, were randomized to either control (n = 5) or hemodilution (n = 5) groups. After placement of an ear vein catheter, animals were anesthetized with intravenous ketamine and xylazine. A tracheotomy was performed, and the animals were mechanically ventilated (Siemens Servo 900B, Siemens Elema, Solna, Sweden) in the supine position using room air, a tidal volume of ~12 ml/kg, and a frequency of ~20 breaths/min. Pancuronium (0.2 mg/kg) was administered to allow control of minute ventilation (E), and infusions of ketamine and xylazine and pancuronium for maintenance of anesthesia and paralysis were begun. A carotid arterial line, a jugular venous line, and a femoral venous line were surgically placed. The jugular venous line was advanced into the right ventricle for monitoring of right ventricular pressure and sampling of mixed venous blood gases. Heparin (200 units/kg) was admininstered intravenously to prevent thrombosis of the intravascular lines. A peripheral intravenous infusion of inert gases (see below) for determination of A/ relationships, using the multiple inert-gas-elimination technique (MIGET), was begun (34, 62). Arterial blood pressure, right ventricular pressure, heart rate, and airway pressures were acquired by using a Spacelabs 511 monitor (Hillsboro, OR) and recorded by using an analog-to-digital converter, data-acquisition software (Strawberry Tree, Sunnyvale, CA), and a personal computer (Apple Macintosh, Cupertino, CA). E was adjusted to achieve an arterial PCO₂ of 30-35 Torr and fixed for the remainder of the experiment.

Twenty to thirty minutes after initiation of the inert-gas infusion, baseline measurements and samples were taken. These included arterial and mixed venous blood gases; hematocrit (Hct); arterial lactate; mixed expired PO₂, PCO₂, and nitric oxide (NO); and the hemodynamic data and airway pressures mentioned above. Fluorescent microspheres were injected intravenously immediately after withdrawal of MIGET samples for later determination of pulmonary blood flow distribution.

After baseline measurements were taken, animals in the hemodilution group underwent three serial isovolemic hemodilutions, with repeat measurements and microsphere administration ~30 min after completion of each hemodilution (1-h intervals). Hemodilution was performed by simultaneous withdrawal of blood and infusion of 6% hydroxyethyl starch at equal volumes, with exchanges ranging between 50 and 80 ml. Control animals were observed over time, with serial measurements and microsphere injections at 1-h intervals. At the conclusion of the experiment, animals were killed with a combination of an overdose of ketamine and exsanguination.

Mixed venous and arterial blood gases and mixed expired CO₂ and O₂ were measured by using an IL 1620 blood-gas machine (Instrumentation Laboratory, Lexington, MA). Venous and arterial O₂ contents were measured by using an IL 482 CO-oximeter (Instrumentation Laboratory). Mixed expired NO was measured continuously by using chemiluminescence (Sievers Instruments, Boulder, CO), with gas sampling from a 50-ml mixing chamber at a rate of 120 ml/min. Hct was measured by centrifugation of duplicate samples. Expired E was measured at the conclusion of the experiments by using a sidestream spirometer (Datex Medical Instruments, Tewksbury, MA).

Blood samples for analysis of serum lactate were spun in a microcentrifuge at 12,000 rpm for 20 s. One milliliter of plasma was drawn off and frozen at 4°C for later analysis. Lactate levels were assayed by using a YSI 1500 Sport lactate analyzer (Yellow Springs Intruments, Yellow Springs, OH).

Calculations. O₂ consumption (O₂) and CO₂ production (CO₂) were calculated from the volumes and concentrations of inspired and expired gases (by using appropriate temperature and humidity corrections) and normalized for body weight. Similarly, NO output was calculated from the corrected expired E and the concentration of NO in the expired gas. The respiratory quotient was calculated as the ratio of CO₂/O₂. was calculated with the Fick equation using O₂ and blood O₂ contents. The fraction of physiological dead space (VDS/VT Bohr) was calculated by using the Enghoff modification of the Bohr equation and reported as a percentage. O₂ extraction ratio was calculated as the ratio of arteriovenous O₂ content to arterial O₂ content and reported as a fraction. Alveolar PO₂, used to calculate the (A-a)PO₂, was calculated by using the alveolar gas equation.

Inert-gas analysis. Six inert gases (ethane, halothane, sulfurhexafluoride, cyclopropane, diethyl ether, and acetone) were dissolved in 5% dextrose solution and infused continuously at 0.7 ml/min. Duplicate arterial and venous blood samples (1 ml) and mixed expired gas samples were obtained at each measurement point; blood and gas samples were drawn simultaneously over ~30 s, with a total time for duplicate samples of ~2 min. Expired gas samples were stored at 40°C before analysis to avoid water condensation and loss of soluble gases. Concentrations of inert gases were measured on a gas chromatograph (Varian 3300, Walnut Creek, CA) equipped with a flame-ionization detector and electron-capture detector. The gas-extraction method of Wagner and associates (62, 63) was used to determine inert-gas tensions in the blood samples.

Fluorescent microsphere analysis. Fluorescent microspheres of four different colors were kept under ultrasonic agitation (Branson 1200, Branson Cleaning Equipment, Shelton, CT) until they were drawn up into a syringe immediately before injection. Each injection contained ~3.3 million microspheres of one color in 1 ml, which was calculated to give sufficient resolution based on earlier published work (1, 4, 22).

After euthanasia at the conclusion of the experiment, a median sternotomy was performed, the pulmonary artery cannulated, and the left ventricle opened. The lungs were flushed with 37°C saline containing 2% dextran for 5 min at 150 ml/min. The lungs were then dissected out of the chest, the heart and extraneous tissue were removed, and the lungs were suspended and reinflated at 20 cmH₂O for 2 days to dry. The dried lungs were encased in blocks of foam sealant and sliced perpendicular to the cranial-caudal axis every 5 mm. The lung slices were placed in a Plexiglas jig with a grid of 6-mm holes placed 1 cm on-center. A core of lung tissue was removed from each hole that fell over lung, and the piece was given x, y, and z coordinates. The core pieces (~200 per lung) were weighed, coded for any airway or residual blood content, and placed in numbered vials. The vials were filled with 1 ml of 2-ethoxyethyl acetate to dissolve the microspheres over 24 h, and the solvent was transfered to cuvettes and read for fluorescence (Perkin-Elmer LS50B, Perkin-Elmer, Norwalk, CT) by using emission and excitation wavelengths determined with standards containing all four colors. The fluorescence was converted to weight-normalized relative blood flow by dividing the result by the weight of the individual piece and the average fluorescence of the specific color for all pieces.

Data analysis. All data are presented as means ± SE at baseline and at the three subsequent measurement time points (T_1-3). Physiological data were analyzed by using Student's t-test to detect differences between groups at baseline and by repeated-measures multivariate ANOVA to detect differences between groups over time, by using the software package StatView (Abacus Concepts, Berkely, CA). A P value <0.05 was accepted as statistically significant.

MIGET analysis was performed on five control and four anemic rabbits. Inert-gas exchange was assessed by changes in the ventilation and perfusion distributions predicted by the 50-compartment model of Wagner et al. (62, 63). Inert-gas shunt (QS/QT) and inert-gas dead space (VDS/VT miget), and log SDs of the perfusion and ventilation distributions were calculated from the 50-compartment model. Global A/ heterogeneity was also assessed by the dispersion index (Disp R*-E*), derived from the retention (R) and excretion (E) data and corrected for dead space (34). Data were averaged from the duplicate samples and are presented as means ± SE for each group. If the sums of squares for an individual data set were >5.0, these data were omitted from the final analysis. This resulted in elimination of only one data set. Differences between groups at baseline were tested by using Student's t-test, and differences over time or with hemodilution were tested by using repeated-measures ANOVA.

Disp R*-E* was significantly different between groups at baseline (see RESULTS). To statistically control for this difference, the baseline value was subtracted from values at T₃ for both groups (T₃-B). A multiple-regression model was then created, which utilized baseline Disp R*-E*Group (a dicotomous indicator variable) and the interaction product of Group and baseline Disp R*-E* (Group_BDisp) as independent variables and T₃-B as the dependent variable. The significance of the difference between groups in the change in Disp R*-E* was determined by examining the coefficient and P value for the interaction term Group_BDisp.

Fluorescent microsphere analysis was performed on all rabbits. Heterogeneity of pulmonary blood flow was characterized by calculating the coefficient of variation (CV) of weight-normalized relative blood flow (CV_Q) at each measurement point for each individual rabbit. Microsphere data were also analyzed by using simple linear regression models for flow in relation to linear spatial dimensions X (transverse, right to left), Y (coronal, ventral to dorsal), Z (saggital, cranial to caudal), and D_h {distance from the hilum, expressed as [(x  x_h)² + (y  y_h)² + (z  z_h)²]^0.5, where x_h, y_h, and z_h are the coordinates of the hilum of each lung (left or right)} to characterize the distribution of pulmonary blood flow across the lung (1, 35). In a modification of previously described methods, we analyzed blood flow in the transverse, i.e., X-axis, in relation to the orientation of the individual lungs. Thus, rather than characterizing flow in terms of right to left across both lungs, we characterized flow from lateral to medial for each lung and pooled these data. To detect changes in regional blood flow between groups over time, the individual regression slopes for the above spatial dimensions were compared by using repeated-measures ANOVA. Baseline differences were determined by using an unpaired Student's t-test. Data for spatial dimensions are presented as means ± SE for each group.

Pulmonary blood flow distribution was tested for fractal structure by examining the CV of blood flow to individual pieces, using a range of piece sizes (25). By combining neighboring pieces into a conglomerate piece of a given size (based on multiples of the initial piece size), the CV of blood flow could be calculated as if the lung had been divided into larger pieces. Conglomerate piece sizes of 2, 3, 4, 6, 8, 12, 16, 24, 32, 48, and 64 were constructed as follows: a central piece was chosen at random from all available pieces and then neighboring pieces were chosen until the desired conglomerate size was reached. Individual pieces assigned to a conglomerate were then flagged as unavailable. The process was repeated, choosing another central piece at random from the remaining available pieces. If a conglomerate could not be constructed because of unavailability of neighboring pieces, the central piece was abandoned and another chosen. This was repeated until the lung was divided into as many conglomerate pieces as possible. Conglomerates were restricted to lie entirely within the left or right lung but were not restricted to be entirely within lobes. For conglomerate pieces, the CV will vary randomly, depending on which individual lung pieces were assigned to which conglomerates. Therefore, we repeated the algorithm for dividing the lung into larger conglomerate pieces. For conglomerates of size n, we repeated the process times (rounding up to the next integer value for noninteger square roots). Thus, for conglomerates of size 64, we repeated the algorithm eight times, obtaining eight CVs. We used the mean of the logarithm of the eight CVs for regression modeling [to obtain fractal dimension (D)], used the SD of the logarithm of the eight CVs for plotting error bars, and used the total number of conglomerate pieces for all eight repetitions as the regression weight at piece size 64. Log(CV) was plotted against log(piece size), and a regression line was obtained by using weighted regression analysis. The fractal dimension D was calculated as (1  slope), where slope was obtained from the regression line fitted to the log(CV) vs. log(piece-size) data. A test for linear trend in D over time was done by using weighted-regression ANOVA.

Further characterization of temporal flow patterns and their spatial location within the lung was carried out by using cluster analysis (24). The residual temporal flow for individual lung pieces was calculated as the relative flow to each piece at a given time, minus the mean of the four weight-normalized relative flows for that piece (times 0, 1, 2, and 3). Cluster dendrograms, time-series plots, and three-dimensional rotating plots of the spatial location of individual lung pieces were examined to identify distinct flow pattern groups. Distinct flow pattern clusters were chosen by using the automatic algorithm, as discussed in the APPENDIX of Glenny et al. (24). Weight-normalized relative flow values from all possible neighboring piece pairs within the lung were obtained, and a correlation coefficient was calculated by using the Pearson product-moment correlation. For each rabbit, scatter plots of microsphere counts for each piece at time i vs. time j were examined, and correlation was determined. A randomization test was then performed to compare the anemic rabbit and control rabbit correlations.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Baseline physiological variables were similar between groups, although venous PCO₂ was significantly different, possibly as a result of a slightly lower CO₂ and higher in anemic animals (Table 2). The difference in arterial PCO₂ between groups was of borderline statistical significance (P = 0.05). Serial hemodilution resulted in a linear fall in Hct from 30 ± 1.6 to 11 ± 1% (Table 2). Hct did not change significantly in the control group (P < 0.0001 for difference between groups over time). Mean arterial pressure and right ventricular pressure did not differ between groups or change with time or hemodilution, despite an ~50% increase in with hemodilution (P < 0.05 between groups over time). Arterial pH fell insignificantly in both groups, whereas arterial lactate increased slightly (Table 2).

Neither arterial nor venous PCO₂ changed with hemodilution (Table 2); however, Pa_O2 (Fig. 1) and (A-a)PO₂ (Table 2) improved with hemodilution in comparison with the control group. This improvement occurred despite a fall in the with hemodilution (Fig. 1). VD/VT Bohr, CO₂, and O₂ did not change significantly over time or with hemodilution (Table 2). As expected from the above observations, respiratorty quotient did not change with hemodilution or time (Table 2).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Arterial PO2 (PaO2) increases with hemodilution while it falls slightly over time in control animals. Mived venous PO2 () decreases with hemodilution while remaining stable in control animals. A, anemic group; C, control group. Time points represent ~1-h intervals. * P < 0.01 for differences between groups over time.

Expired NO did not differ significantly between groups at baseline, but it progressively increased in the anemic group while remaining stable in the control group (P < 0.0001 between groups over time) (Fig. 2).

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Expired nitric oxide (NO) increases during hemodilution while remaining stable over time in control animals. Time points represent ~1-h intervals. * P < 0.0001 between groups over time.

A/ distributions (MIGET). Inert-gas measurements are shown in Table 3. QS/QT was very low in both groups, and, although it was higher in anemic than in control animals throughout the experiment, the differences were not statistically significant. QS/QT remained stable and low over time and/or hemodilution. Disp R*-E* was significantly higher in anemic animals at baseline (P < 0.05). Disp R*-E* fell with each hemodilution while increasing slightly in control animals over time. The difference in groups over time was highly significant (P < 0.01) and remained significant after it was controlled for differences in groups at baseline (P < 0.05) (see METHODS).

View this table: [in this window] [in a new window]   Table 3.   A/ distributions (MIGET)

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Dispersion of pulmonary blood flow (CVQ, where CV is coefficient of variation) over time (C group) or with hemodilution (A group) is not significantly different between groups. Time points (T1-T3) represent ~1-h intervals. B, baseline.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Log-log plots of lung piece size vs. CV of pulmonary blood flow for representative rabbits in C (rabbit 9) and A (rabbit 4) groups at baseline and T3, using natural logs. Slope (Sl) of log-log plot decreases in A animal with hemodilution while remaining essentially stable in C animal. Hct, hematocrit.

View larger version (11K): [in this window] [in a new window]   Fig. 5.   Fractal dimension D (1  slope of log piece size vs. log CV) of blood flow vs. time in control (A) and anemic (B) animals. There is a significant fall in D in A animals compared with C animals (P = 0.04). Nos. are rabbit nos.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Spatial correlation of blood flow () over time or with hemodilution in C and A animals. * P = 0.04 between groups over time. Time points represent ~1-h intervals.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In agreement with previous work from our laboratory (11), the present study shows that CO₂ elimination and CO₂ remain remarkably stable in anesthetized rabbits despite severe hemodilution. Of interest, however, is the finding that arterial blood O₂ tension increases with severe anemia, particularly in comparison with control animals followed over time. This mild improvement in Pa_O2 and (A-a)PO₂ occurs in association with improvements in global indexes of A/ heterogeneity (Disp R*-E*), a reduction in the fractal dimension of blood flow D, increased spatial correlation of blood flow , and an increase in expired NO.

Anemia and Pa_O2: theoretical considerations. There are several potential reasons that anemia may affect arterial oxygenation. These include the influences of changes in , effective blood O₂ solubility, , blood viscosity, capillary transit time, and A/ heterogeneity.

Local changes which affect alveolar-arterial equilibration of O₂ can be described by the "equilibration coefficient" D/(Q · ), whereby a high D/(Q · ) value is associated with a high degree of equilibration (55). D represents the diffusive conductance of the alveolar membrane and is unlikely to be affected by anemia. Q represents pulmonary capillary blood flow, which increases during anemia, with the potential effect of reducing alveolar-arterial gas equilibration. However, the modest changes with anemia (50% with severe hemodilution) and the ameliorating effect of capillary recruitment are likely to prevent significant transit time shortening (56).  represents the capacitance of the gas in blood and is equal to the slope of the blood binding curve. The slope of the oxyhemoglobin dissociation curve is reduced in anemia (5, 52), predicting a higher degree of gas equilibration. Thus, at the alveolar-capillary level, anemia likely results in either no change or in slightly improved gas equilibration, compared with the nonanemic state.

More globally, in the presence of venous admixture, this effect is moderated by the greater saturation range Hb undergoes during anemia and the "flattening" of the oxyhemoglobin dissociation curve at high PO₂ ranges. Indeed, in the presence of venous admixture, the flattened oxyhemoglobin curve and reduced may result in an increased (A-a)PO₂ and a reduced Pa_O2 (52). In the presence of a fixed intrapulmonary shunt, if (and mixed venous O₂ saturation) were to remain constant during anemia, Pa_O2 would actually increase as anemia progressed, as illustrated by the extreme example of calculating Pa_O2 by using the shunt equation in the presence and absence of Hb (50).

Other global factors that might be expected to improve gas exchange in anemia include reduced A/ heterogeneity due to increased , reduced blood viscosity, and changes in red cell-dependent vasoactive mediators. These effects are discussed in more detail below.

Anemia and Pa_O2: studies. Over 45 years ago, Ryan and Hickam (57) found that a group of anemic patients without obvious lung disease had lower Pa_O2 and higher (A-a)PO₂ than a group of nonanemic controls. They attributed the relative hypoxemia seen with anemia to the effect of venous admixture and a likely reduction in the . However, this study was confounded by the presence of both acute and chronic anemia and the inability to control for potential cardiovascular and/or pulmonary disease in the study group.

Experimentally, the relationship between anemia and Pa_O2 is not well defined. A multitude of studies have examined the cardiopulmonary effects of hemodilution or blood transfusion in the absence of known lung disease. A thorough survey of the studies that reported blood-gas data reveals 11 that suggest an inverse relationship between Pa_O2 and Hct, five that suggest a direct relationship between Pa_O2 and Hct, and 11 that suggest either no or an inconsistent effect of Hct on Pa_O2 (see Table 1 for Ref. listing). However, only three of these studies (30, 48, 51) included concurrent controls for the effects of time and/or anesthesia, and none engaged in a systematic analysis of the effects of Hct on Pa_O2.

The effect of anemia on arterial oxygenation in the presence of lung pathology is also ill defined. In rabbits with left lung atelectasis, isovolemic hemodilution resulted in a relative increase in intrapulmonary shunt and a lower Pa_O2 than in a control group followed over time (13). Earlier, Bishop and Cheney (3) found no change in shunt after isovolemic hemodilution in dogs with oleic acid-induced lung injury; however, that study also suffered from the lack of a control group. Studies of critically ill humans suggest a deleterious effect of higher Hct on pulmonary O₂ exchange, but the potential cardiovascular and inflammatory effects of autologous banked blood transfusion, in addition to the reduced 2,3-diphosphoglycerate in banked blood, make these findings difficult to interpret (20, 38, 58).

The present study shows that isolated and compensated acute anemia is associated with an increase in Pa_O2, in comparison to nonanemic controls. We emphasize the comparison with a control group, because prolonged time under anesthesia may be associated with the development of atelectasis and a decrement in pulmonary gas exchange, thus masking any potential benefits from anemia. This may be the reason that a previous study (lacking a control group) performed in our laboratory was unable to show an improvement in Pa_O2 with hemodilution (11). Another unique facet of the present study is that E was fixed throughout. This may differ from the clinical situation where severe anemia may result in hyperventilation and respiratory alkalosis, which, in turn, may have variable and independent effects on A/ heterogeneity and O₂ exchange.

In the present study, the improvement in Pa_O2 was seen despite a fall in with anemia (Fig. 1), the result of which should be to magnify the effect of venous admixture, as suggested by Ryan and Hickam (57). In the following sections, we discuss potential mechanisms by which anemia resulted in improved O₂ exchange in our study.

A/ and pulmonary blood flow distributions and anemia. The improvement in Pa_O2 that we observed was associated with an improvement in overall A/ matching. The potential mechanisms by which A/ matching improves with anemia include an increase in pulmonary blood flow (), a reduction in blood viscosity, and an interaction between red blood cells and the vasoactive mediator NO. These factors may lead to a change in pulmonary blood flow distribution and, theoretically, a change also in ventilation distribution.