Skip Navigation

Cardiovascular Research 1998 37(1):225-235; doi:10.1016/S0008-6363(97)00226-5
© 1998 by European Society of Cardiology
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow E-letters: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when E-letters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Pries, A. R
Right arrow Articles by Gaehtgens, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Pries, A. R
Right arrow Articles by Gaehtgens, P.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Copyright © 1998, European Society of Cardiology

Blood flow resistance during hemodilution: effect of plasma composition

Axel R Priesa,*, Timothy W Secombb, Markus Sperandioc and Peter Gaehtgensc

aDeutsches Herzzentrum Berlin, Augustenburger Platz 1, D-13353 Berlin, Germany
bDept. of Physiology, University of Arizona, Tucson, AZ 85724, USA
cDept. of Physiology, Freie Universität Berlin, Arnimallee 22, D-14195 Berlin, Germany

* Corresponding author. Tel. (+49-30) 838 2087; Fax (+49-30) 838 4916; E-mail: pries@zedat.fu-berlin.de

Received 3 July 1997; accepted 15 August 1997


   Abstract
Top
 Abstract
1 Introduction
 2 Materials and methods
 3 Results
 4 Discussion
 References
 
Objectives: To investigate the causes of wide variations in reported effects of hemodilution on flow resistance of vascular beds. Methods: (a) In a meta-analysis of 28 prior studies, resistance values at hematocrits of zero (R0) and 0.45 (R0.45) were derived. Study design characteristics (presence of vasodilatory reserve or leukocytes, species, tissue, hemodiluent) were tested by ANOVA for their relation to the ratio R0/R0.45. (b) Experiments were performed to determine flow resistance during hemodilution in the rat mesentery with (n = 8) and without (n = 11) pretreatment with heparinase, which modifies the endothelial glycocalyx. (c) A mathematical flow simulation for mesenteric microvascular networks was used to predict resistance effects of hemodilution and of a hypothetical layer on the endothelial surface. Results: (a) In prior studies using native plasma for hemodilution R0 averaged 59±8% of R0.45, while in studies using artificial solutions R0 averaged 32±12% of R0.45. The larger reduction of flow resistance upon dilution with artificial media is independent of viscosity and oncotic pressure. Other design characteristics did not show strong significant effects. (b) Present experiments showed large reductions of flow resistance with saline hemodilution which were nearly halved after heparinase pretreatment. (c) Resistance effects of hemodilution with plasma or after heparinase treatment agree with model predictions based on tube flow rheology of blood. The larger resistance effects of dilution with artificial media can be explained by the removal of an endothelial surface layer of ~1.5 µm thickness. Conclusions: The results imply that changes of plasma composition, due to use of artificial infusion media, influence peripheral resistance and tissue perfusion. They are consistent with the hypothesis that interactions between endothelial glycocalyx structures and plasma components lead to formation of a thick layer at the endothelial surface which increases flow resistance.

KEYWORDS Hemodilution; Optimal hematocrit; Organ perfusion; Glycocalyx; Plasma proteins; Rat mesentery microvascular networks


   1 Introduction
Top
 Abstract
 1 Introduction
2 Materials and methods
 3 Results
 4 Discussion
 References
 
Since the classical study of Whittaker and Winton [1], the reduction of resistance to blood flow in vascular beds during progressive hemodilution has been the subject of many experimental investigations [2–28]. The relationship between flow resistance and hematocrit has important implications for physiological phenomena related to blood rheology and for the pathophysiological or therapeutic effects of hematocrit changes in patients. It also determines the so-called "optimal hematocrit" (Hopt) at which overall oxygen delivery to the tissue via hemoglobin transport is maximized. However, experimental attempts to quantify the reduction in flow resistance resulting from hemodilution have led to highly variable results and the existence of a single value of the "optimal hematocrit" has been the subject of controversy. It is therefore not even possible to predict unequivocally from experimental studies whether oxygen supply to a tissue will increase or decrease for a given hemodilution procedure.

Several attempts have been made to explain this variability. Benis et al. [5, 9]suggested that inertial effects at high blood flow seen at low hematocrit add to the flow resistance measured if constant pressure perfusion (rather than constant volume flow) is imposed. This would tend to exaggerate the dependence of flow resistance on hematocrit. Gaehtgens et al. [16]reported stronger flow dependence on hematocrit if the vascular system was not intentionally dilated, suggesting that hematocrit reduction was accompanied by vasodilation due to metabolic or shear stress dependent mechanisms. Recruitment of a ‘vasodilatory reserve’ would then lead to overestimation of the purely rheological effects of hemodilution.

Despite these explanations, published information on the relationship between flow resistance and hematocrit gathered over the last 65 years has remained controversial and difficult to interpret. The present study uses a meta-analysis of available hemodilution studies to examine possible causes and mechanisms of the observed differences. To this end, the quantitative effects of hemodilution on blood flow resistance were determined for all studies and related to the characteristics of experimental design including the animal species and tissue studied, the suppression of vascular tone, the presence or absence of white blood cells in the perfusate, and the use of plasma or artificial media for hemodilution. This analysis shows that the hematocrit/resistance relationship critically depends on the use of plasma or saline as dilution medium, suggesting the existence of a layer of plasma components adsorbed on the luminal vessel surface which can be washed out by artificial infusion media. This hypothesis was tested in additional experiments involving direct estimation of blood flow resistance in rat mesenteric microcirculatory beds during intentional hemodilution with and without enzymatic modification of the endothelial glycocalyx.


   2 Materials and methods
Top
 Abstract
 1 Introduction
 2 Materials and methods
3 Results
 4 Discussion
 References
 
2.1 Meta-analysis of prior studies
By computer search (Medline) and by analysis of related published literature, 28 reports were identified describing experimental studies in which the hematocrit (H) of blood flowing through vascular beds was acutely manipulated over a substantial range and the flow resistance (R) was measured. The studies were published between 1933 and 1995. The experimental procedures employed in these studies were classified according to the following characteristics: species; tissue; perfusate at reference hematocrit; composition of dilution medium; hematocrit range tested; presence or absence of white blood cells (WBC) in the perfusate; mode of perfusion (constant flow or constant pressure); level of perfusion pressure; suppression of vascular tone.

The experimental designs of the studies show considerable variations. For example, approaches used to suppress changes of vascular tone by inducing vasodilation may differ in effectiveness (and thus remaining vasodilatory reserve) and possible side effects. However, the present analysis was aimed at investigating the impact of major systematic differences in the experimental procedures on flow resistance changes during hemodilution. Possible effects of the numerous secondary differences between studies were neglected as their analysis would require much larger numbers of studies than are actually available. Therefore, the number of classes or categories used to describe each of the characteristics listed above was restricted to the primary factors so as to ensure a statistically useful number of studies in each class. With respect to vasodilation, two classes (yes/no) were used. A similar strategy was applied in defining the classes for presence of white cells (yes/no) and for the medium used for hemodilution. For the latter characteristic, three classes were defined: ‘plasma’ when only plasma or serum was used throughout the experiment; ‘plasma/artificial’ when plasma was present at the reference hematocrit, but artificial saline-based media with or without added colloids were used for dilution; and ‘artificial’ when artificial media were used both at reference hematocrit and for dilution. For the characteristics ‘species’ and ‘tissue’ the numbers of classes were unavoidably larger, and the small number of studies in some classes limits the expected discrimination of the results with respect to these characteristics.

Data on flow resistance and hematocrit were derived from tables or figures in each publication. If possible (e.g. if pressure-flow curves for different hematocrit levels are given), values were determined for conditions of fixed flow rate and fixed pressure, and from the slope of the pressure-flow curve (I/P curve). The data for fixed pressure conditions were extracted from the given pressure-flow curves by selecting a level of driving pressure close to the physiological range (80 mmHg for most tissues) and reading the corresponding values of volume flow from the graph. A similar procedure was used to obtain data for fixed flow. For each data set the flow rate reported at hematocrit 0.45 and at the driving pressure used to obtain the fixed pressure data was selected as the fixed flow level.

For most individual data sets conductance (C = 1/R) increased in an approximately linear fashion with decreasing hematocrit. Therefore, a linear regression of C on H was used to derive values of R0/0.45 (which is equal to C0.45/0), the ratio of flow resistance at hematocrits of zero (R0) and 0.45 (R0.45), and Hopt, the optimal hematocrit at which the product of hematocrit and conductance and thus calculated oxygen delivery at a given driving pressure across the vascular bed is maximal. Fig. 1 shows four examples. Steep regression lines correspond to strong dependence of flow resistance on hematocrit (i.e. low values of R0/0.45). In 26 out of 31 cases the data were adequately fitted by this procedure, as judged from the correlation coefficient (r2 ≥0.85) and from the distribution of residuals. In these cases values of R0/0.45 and Hopt were obtained from the regression equation. For the remaining 5 cases the resistance at zero hematocrit (R0) was extrapolated using the graph of conductance versus hematocrit, and Hopt was chosen to correspond to the maximal value of H·C obtained from the experimental data. For 8 studies, including those in which flow resistance is given only for a standard or control hematocrit and for red cell-free perfusion (hematocrit=0), the shape of the relation between resistance and hematocrit could not be analyzed. Therefore a regression was not calculated and values for Hopt are not given.


Figure 1
View larger version (27K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1 Examples of the dependence of conductance of vascular beds (C = 1/Resistance) on hematocrit as derived from four literature studies. Conductance was normalized with respect to the value at zero hematocrit. From these graphs R0/0.45, which is identical to C0.45/0, is given by the normalized conductance at hematocrit 0.45. In addition to original data points, the results of a linear regression analysis (continuous line, r2) are given.

 
In some studies the suspending medium at the control hematocrit (i.e. plasma) and the medium used for hemodilution had different viscosities. In order to correct the obtained R0/0.45 values for changes in medium viscosity during hemodilution, the uncorrected value was multiplied by the average ratio of suspending medium viscosity to plasma viscosity. This ratio was determined from viscosities given in the respective study or from appropriate data in the literature. The ratio values used were 1.4 for dilution with 6% dextran [7, 8], 0.58 for saline, and 0.89 for serum.

2.2 Experimental data
Experiments were conducted in the microvasculature of the rat mesentery using intravital microscopy following approval of the procedures used by university and governmental committees on animal care. The flow resistance of microvascular networks in the rat mesentery was determined at various levels of systemic hematocrit during intravital microscopy by measuring the pressure drop between feeding arterioles and draining venules as well as volume flow rate. The animal preparation and the setup used for intravital microscopy have been described in detail elsewhere [29]. Male Wistar rats (body weight 300–450 g) were prepared for intravital microscopy of the mesenteric microcirculation following premedication (atropine 0.1 mg/kg i.m. and pentobarbital sodium 20 mg/kg i.m.), anaesthesia (ketamine 100 mg/kg i.m.), cannulation of trachea, jugular vein, and carotid artery, and abdominal midline incision. The animals were then transferred to a special stage mounted on an intravital microscope. The small bowel was exteriorized and fat-free portions of the mesentery selected for investigation with a 25x/N.A.0.6 salt-water immersion objective (Leitz). During the experiments the level of anaesthesia and fluid balance were maintained by i.v. infusion of physiological saline (24 ml/kg per h) containing 0.3 mg/ml pentobarbital sodium. In this preparation of the exposed mesentery, vessels generally exhibit no spontaneous smooth muscle tone. However, as a precaution to prevent the development of tone and thus variation of flow resistance during the measurement period papaverine (10–4 M) was continuously superfused. Heart rate and arterial blood pressure (range 105 to 140 mmHg) were continuously monitored via the catheter in the carotid artery.

The networks selected for this study were supplied by feeding arterioles with inner diameters of about 30 µm and drained by venules of about 45 µm. As in previous studies in the same preparation the networks consist of about 450 individual vessel segments [27]. Intravascular pressures were measured by the servo-nulling technique using a micropipette with a tip diameter of about 1 µm connected to a micropressure measuring system (Model 5, IPM). Venular pressures were found to change less than 2 mmHg during the experimental procedures, and were therefore typically measured only at the beginning and the end of an experiment. After the initial measurement of venular pressure the micropipette was introduced into the main feeding arteriole where it remained during the course of the experiment. Systemic hematocrit (H) was varied systematically by infusion of homologous red cells and hydroxyethyl starch solution (100 g/l, MW 200 000/0.5 in physiological saline; Fresenius). After elevation of H up to 0.65 by slow infusion of red cells (in 3 control experiments), H was successively lowered by hemodilution. In each hemodilution step between 1 and 2 ml blood was withdrawn from the carotid catheter and replaced by an equal amount of the dilution fluid via the venous catheter. At levels of H below about 0.2, up to 2 ml of dilution fluid were given in excess to the amount of blood withdrawn in order to increase the dilution effect and to reduce the fall in systemic blood pressure. At the lowest hematocrit level reached (0.15±0.06), the measurement of venular pressure was repeated. Venular pressure levels for the intermediate hematocrit steps were interpolated from the values taken at the beginning and the end of the experimental procedure.

At each hemodilution step the microscopic image of the feeding arteriole was recorded via a CCD camera (MX, AIS) on videotape for about 30 s using asynchronous flash illumination (11360-1, Chadwick-Helmuth). These recordings were analyzed off-line to determine arteriolar centerline flow velocity and vessel diameter using a digital image analysis system [30]. The spatial correlation analysis of recordings obtained with asynchronous illumination allowed measurements of blood flow velocities up to about 40 mm/s [31]. Since vessel diameters remained constant during the experiments, changes in network flow resistance were calculated from changes in arterio-venous pressure drop and centerline flow velocity in the feeding arteriole. In 8 out of 19 experiments, hemodilution was preceded by perfusion of the network with heparinase (100 U/ml, 10 min, Sigma) via a micropipette introduced into the feeding arteriole. The pipette was pressurized and the pressure adjusted so that during microperfusion with the enzyme solution a small amount of blood (about 10%) was still flowing in the arteriole in order to limit possible effects of high transmural pressure. The flow resistances of networks of different sizes and geometrical hindrances varied considerably. Therefore, resistance values for each hematocrit level were normalized with respect to the resistance of the same network at hematocrit 0.45. This allowed comparison of data from different experiments. From the normalized values, R0/0.45 was calculated for each experiment according to the procedures used for the data sets from literature studies.

2.3 Flow simulation
To allow comparison of the experimental data with predictions on the basis of blood viscosity measurements in blood-perfused glass tubes, flow simulations were performed using a previously described mathematical model [29]. Using an iterative algorithm, this model calculates the distribution of flow and pressure in a microvascular network with experimentally determined vessel diameters, lengths, and interconnections between the vessel segments. 6 Microvascular network structures were used to obtain the simulation results reported here. These networks have been recorded by intravital microscopy in the rat mesentery in a previous study [31], and were similar in size and structure to those networks used for the present experimental measurements described above.

In calculating flow and pressure distributions in the microvascular networks, the model takes into account the partition of red cells and plasma at diverging bifurcations (phase separation effect), and the dependence of effective blood viscosity on hematocrit and vessel diameter (Fahraeus—Lindqvist effect). For these phenomena, parametric descriptions of previously published experimental results were used [29, 32]. While the description of the phase separation effect is based on an in vivo investigation in the same tissue used also in the present study, the description of the Fahraeus—Lindqvist effect is based on measurements of the rheological properties of blood in small-diameter glass tubes (in vitro viscosity law [33]). Simulations were performed for each network increasing the hematocrit in the feeding arterioles from values of 0 to 0.86 in steps of 0.01. For each hematocrit step flow resistance was calculated. To estimate the effect on flow resistance elicited by the removal of a completely immobile layer on the endothelial surface with a given thickness (d), simulations were performed in which all vessel diameters were increased by 2d above the experimentally determined values.


   3 Results
Top
 Abstract
 1 Introduction
 2 Materials and methods
 3 Results
4 Discussion
 References
 
3.1 Meta-analysis of prior studies
The results of the literature compilation are shown in Table 1, including the derived values for Hopt, and R0/0.45. The studies are listed in order of publication date. For those studies (n = 12) providing data for both fixed flow and fixed pressure perfusion, R0/0.45 is consistently higher for fixed flow conditions than for fixed pressure conditions with a difference of 0.081±0.052 (mean ±SD, p<0.005, two tailed t-test). A similar result was obtained for Hopt where values obtained under fixed flow conditions are higher by 0.076±0.054 (n = 11; p<0.005). These results indicate that the impact of hemodilution on flow resistance was consistently weaker under fixed-flow conditions, but the difference was small compared to the observed variability between studies.


View this table:
[in this window]
[in a new window]

  		Table 1 Compilation of literature data (ordered according to publication data) on the relationship between hematocrit and blood flow resistance in vascular beds

 
According to Table 1 the R0/0.45 values range between 0.106 and 0.714. Within this range the frequency distribution of R0/0.45 (Fig. 2) exhibits two peaks suggesting the existence of two distinctive clusters with strong and weak dependence of flow resistance on hematocrit, respectively. To investigate possible mechanisms responsible for the bimodal shape of the frequency distribution, the influence of experimental study design characteristics on the values of R0/0.45 is shown in Fig. 3. Differences between R0/0.45 values for the different classes of the experimental parameters ‘white cells’, ‘dilation’, and ‘medium’ were tested using simple factorial ANOVA. A significant effect is seen for the parameter ‘medium’ (p<0.001), while the influence of ‘dilation’ was small (p = 0.067) and that of ‘white cells’ negligible (p = 0.772). An analysis including the parameters ‘species’ and ‘tissue’ yielded similar results and also showed significance for the difference between studies in the rat and in other species (p<0.05, one-way ANOVA, Bonferroni correction for post hoc multiple comparisons). However, the latter result must be interpreted with caution due to the low number of available studies compared to the number of classes for the parameter ‘species’. Furthermore, all published experiments on rats used artificial saline-based solutions as dilution medium, and it is therefore not possible to test whether the difference in R0/0.45 between rats and other species is in fact due to the choice of dilution medium.


Figure 2
View larger version (14K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2 Frequency distribution of R0/0.45 for 29 individual values from 28 publications. Literature values were obtained from measurements at constant perfusion pressure and corrected for changes in the viscosity of the suspending medium during hemodilution.

 

Figure 3
View larger version (19K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 3 The effects of different experimental parameters on the mean value (±SD) of R0/0.45. Significant effects are seen for the parameters ‘species’ and ‘medium’ (simple factorial ANOVA, p<0.003). However, the result for ‘species’ may not be valid, as discussed in the text. With respect to the dilution medium used, three groups have been formed: PLASMA: plasma (or serum) used as suspending medium during control perfusion with a hematocrit of 0.45 and for dilution, PLASMA/ARTIFICIAL: plasma used as suspending medium during control conditions and saline based media used for dilution, ARTIFICIAL: artificial saline-based solutions used as suspending medium throughout the experiment.

 
Fig. 4 shows frequency distributions for R0/0.45 in the three classes of the parameter ‘medium’. Studies in which plasma was the suspending medium at reference hematocrit, and was also used for dilution (‘plasma’), showed significantly higher values (p<0.05, one-way ANOVA, Bonferroni correction) for R0/0.45 than those in which artificial media were used for dilution only (‘plasma/artificial’) or for both control conditions and dilution (‘artificial’). No significant difference was found between R0/0.45 values for the ‘plasma/artificial’ and the ‘artificial’ group.


Figure 4
View larger version (23K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 4 Frequency distributions of R0/0.45 and mean values (±SD) for the three classes of dilution medium (see Fig. 4). The results for the PLASMA group (black bars) is significantly (ANOVA) different from those of the PLASMA/ARTIFICIAL and the ARTIFICIAL group (hatched bars).

 
3.2 Experimental data and flow simulation
For the experimental data obtained in the present study, Fig. 5 gives the flow resistance of mesenteric microvessel networks at different systemic hematocrits. Flow resistance is expressed relative to that obtained at a hematocrit of 0.45. Combined results are presented both for control experiments (n = 11) and for measurements following heparinase microinfusion into the arteriole feeding the network (n = 8). The value of R0/0.45 determined for the control experiments is 0.34±0.11 (mean±SD). Heparinase microinfusion leads to a substantially flatter relation between hematocrit and network flow resistance as quantified by the higher value of R0/0.45 (0.65±0.14). For comparison, a prediction is shown which is obtained by the mathematical simulation of flow through mesenteric microvascular networks using the in vitro viscosity law [32]. Results after heparinase treatment are very similar to those of the model simulations which yield a R0/0.45 value of 0.62±0.015. As reported earlier [34], heparinase microinfusion leads to an average reduction in absolute flow resistance of about 21% in addition to its effects on the relation between hematocrit and flow resistance. This effect is not shown in Fig. 5 which gives normalized resistance values.


Figure 5
View larger version (19K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 5 Experimental data obtained in the present study for the relation between mesenteric microvascular flow resistance (normalized with respect to the resistance at a hematocrit of 0.45) and hematocrit for untreated networks (control, n = 11) and after heparinase micro-injection (heparinase, n = 8). The symbols indicate averages (±SE) of the two experimental groups for consecutive hematocrit classes. Also shown are predictions (mean: dashed line, ±SD: thin lines) obtained with mathematical flow simulations using the in vitro viscosity law for 6 experimentally determined network structures.

 

   4 Discussion
Top
 Abstract
 1 Introduction
 2 Materials and methods
 3 Results
 4 Discussion
References
 
The results of studies performed over the last 65 years exhibit a remarkable variability in the quantitative effect of hemodilution on blood flow resistance of vascular beds, and consequently also in the estimated value for the optimal hematocrit. Because of the limited number and the heterogeneity of the available studies, a meta-analysis can only address the effects of major differences in experimental design. Nonetheless, such an analysis can reveal information about factors influencing flow resistance, even when investigating these factors was not among the aims of the original studies.

As blood viscosity decreases during hemodilution, blood flow increases in experimental protocols in which the driving pressure is held constant. This leads to an increase in the inertial effect on energy dissipation with decreasing hematocrit and might therefore reduce the effect of hemodilution on flow resistance [5, 9]. Alteration of inertial effects during hemodilution is prevented in experiments in which the volume flow rate is held constant at all hematocrit levels. The dependence of flow resistance on hematocrit is consistently weaker in this type of experiments compared to experiments in which perfusion pressure is held constant and flow can vary with hematocrit. However, these differences can mostly be explained by the fact that the pressure-flow curves (I/P curves) exhibit a positive intercept with the pressure axis (‘flow cessation pressure’). This effect has been tentatively attributed to blood rheological properties (‘yield shear stress’) or the passive compression of vessels (‘critical closing pressure’). In contrast, a nonlinear shape of I/P curves which could reflect inertial effects is only seen in about half of the studies analyzed and is of minor importance for the observed differences between the fixed-flow and the fixed-pressure data. In addition, the variability of R0/0.45 within both the fixed-flow and the fixed-pressure group is much larger than the average difference between these groups, precluding the use of this mechanism to explain a significant part of the observed variability of the relation between hematocrit and flow resistance.

On the basis of the hypothesis that reduction of O2-carrying capacity of the blood might cause metabolically mediated vasodilation, the effect of hemodilution on flow resistance might be increased, if a dilatory reserve was present in the perfused tissue [16]. The elevation of flow rate associated with hematocrit reduction might also cause vasodilation due to a flow-dependent release of vasoactive autacoids, e.g. EDRF or prostacyclin. The data in Fig. 3 show that vasodilatory responses may indeed have some effect on R0/0.45. However, in the set of studies analyzed, this effect is quantitatively small and not significant at the p<0.05 level in the ANOVA. Therefore, the presence of a vasodilatory reserve present in many studies analyzed does not appear to be the main cause for the variable dependence of flow resistance on hematocrit reported.

White cells have been shown to potentially occlude capillary vessels especially under conditions of low perfusion pressure [35]. If activated, these cells exhibit even lower deformability and thus might act as microemboli in cell suspensions that have been manipulated during preparation of different hematocrits in the perfused blood. Also, variation of hematocrit is achieved in many studies by dilution of whole blood, and white cell concentration might therefore vary in parallel to hematocrit in the perfusates used thus causing reduced plugging at low compared to high hematocrit. However, the difference in R0/0.45 between studies with and without white blood cells in the perfusate was very small and not significant in the present database, suggesting the presence of white cells has little effect on the hematocrit dependence of flow resistance.

4.1 Composition of dilution medium
Unexpectedly, the experimental parameter most strongly correlated with the resistance effect of hemodilution was the choice of the dilution medium (Figs. 3 and 4Go). R0/0.45 for the ‘plasma’ group equals 0.59, indicating that dilution of the blood with plasma to zero hematocrit led to an (extrapolated) average reduction of flow resistance to about 59% of the value seen at a hematocrit of 0.45. Since the suspending medium was not altered during hemodilution in these experiments, this value reflects the resistance effect of an isolated reduction in red cell concentration (Fig. 6). The magnitude of this resistance reduction is in agreement with extrapolations of blood rheological behavior in glass tubes to the perfusion of microvascular networks [31, 32]. In contrast, the use of artificial saline-based media for hemodilution resulted in a stronger reduction of flow resistance and a R0/0.45 value of 0.32. This indicates an additional effect of artificial media on flow resistance, which does not result from differences in suspending medium viscosity since the data were corrected for such differences. This additional effect is quantitatively comparable to the effect of hematocrit reduction alone.


Figure 6
View larger version (42K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 6 Schematic diagram indicating hypothesized relations between the thickness of the glycocalyx and changes in flow resistance with hematocrit. See text for explanation.

 
The additional effect of artificial media on resistance is independent of the presence and nature of the colloids added to the saline and is therefore not related to differences in colloid osmotic pressures. Evidently, some property specific to plasma, which is not present in the artificial media used in the dilution studies analyzed here, is responsible for maintaining high flow resistance during hemodilution. This observation can be accounted for by the following hypothesis: The reversible adsorption of plasma components, most likely proteins, to the endothelial glycocalyx [36, 37]leads to the formation of a zone of impeded flow immediately adjacent to the luminal vessel surface [38, 39]. Accordingly, the resistance changes seen in studies in which artificial solutions are used for hemodilution reflect not only the effect of lowered hematocrit, but also the effect of removal of adsorbed proteins. This removal increases the effective cross-section of vessels available for free blood flow. In contrast, experiments in which dilution was performed with plasma alone represent only the effects of hematocrit reduction. These effects are shown schematically in Fig. 6.

To exert a significant hemodynamic effect, the adsorbed layer creating a zone of impeded flow must be of considerable thickness. The values of R0/0.45 obtained for the ‘plasma’ and the ‘plasma/artificial’ group in the present analysis (0.59 and 0.32, resp.) suggest that washout of plasma proteins leads to a reduction of flow resistance which is similar to that elicited by the removal of red blood cells. Theoretical simulations show that the removal of a completely immobile layer of about 1.5 µm in all microvessels is necessary to reduce flow resistance by the required amount in the networks studied. (Due to the fourth-power relationship between vessel radius and flow resistance, such a layer would affect flow resistance predominantly in microvessels and its presence in larger vessels cannot be determined from the present analysis.) Direct intravital microscopic observations have demonstrated an exclusion zone for macromolecules and red cells with a thickness of about 0.6 µm on the endothelial surface [40]. Indirect estimations using different techniques [31, 38, 39, 41–43]indicated a layer thickness of about 0.5 to 1 µm. These estimates are much higher than values between 0.05 and 0.1 µm reported for the thickness of glycocalyx components bound directly to the endothelial cell membrane [44, 45]. It is therefore conceivable that the major portion of the flow-retarding layer consists of adsorbed plasma components which are washed out during hemodilution with protein-free solutions.

This concept is supported by the results of the present experiments performed to test the effect of enzymatic treatment on the hematocrit/resistance relationship. Heparinase treatment modifies the endothelial glycocalyx by cleaving oligosaccharide side chains of glycoproteins residing on the endothelial surface. This has been shown to increase capillary hematocrit [46]and to decrease flow resistance of microvascular networks [34]. Both phenomena have been attributed to reductions by about 0.5 to 1 µm of the thickness of a macromolecular layer on the endothelial surface which lead to increases of the vascular cross-section available for free flow of red cells and plasma. As pointed out above, this effect cannot primarily result from changes in the thickness of membrane-bound glycoproteins due to the enzymatic treatment. However, it could represent a reduction in effective layer thickness due to enzymatic cleavage of anchoring points required for the adsorption of plasma proteins.

For untreated microvascular beds, the R0/0.45 value obtained (0.34±0.11) is similar to the mean value (0.32±0.12) of the ‘plasma/saline’ class in the literature compilation. However, if the hemodilution is preceded by heparinase treatment of the microvessels (Fig. 5), the value of R0/0.45 (0.65±0.14) is close to the mean value (0.59±0.08) for the ‘plasma’ class and to predictions based on glass tube rheology (0.62±0.015), suggesting that only the red cell dependent effect of hemodilution is present. Therefore, the present results are consistent with the concept that adsorption of plasma proteins to the glycocalyx increases flow resistance and that this effect is inhibited by heparinase treatment. According to this concept, the results of studies in which saline solutions are used for hemodilution reflect not only the effect of lowered hematocrit, but also the effect of increased effective vessel cross-section due to removal of adsorbed proteins. In contrast, experiments with plasma dilution only represent the effects of hematocrit reduction as such (Fig. 6). In such experiments the dependence of flow resistance on hematocrit is remarkably close to predictions based on glass tube rheology [32].

It is currently not known which plasma proteins may constitute the adsorbed layer. Albumin and fibrinogen are obvious candidates since they are known to bind to the endothelial surface [47–50]. However, the available results suggest that neither of these proteins alone is sufficient to generate or maintain the layer. Some studies in the ‘artificial’ and ‘plasma/artificial’ groups used saline with added albumin as dilution medium and yielded values for R0/0.45 which are in the same low range as those obtained for studies using other colloidal substances or saline alone. In an in vitro study Reinhart et al. [49]found that endothelium-coated beads exhibit reduced sedimentation velocity if suspended in plasma rather than in saline. Reduced sedimentation velocity was not found in saline with added albumin or fibrinogen alone. Interaction of two or more plasma proteins may thus be required for the creation of an adsorbed protein layer, and other plasma components could interfere with layer formation as in the competitive inhibition of fibrinogen binding to endothelial surfaces by albumin [48, 49].

The adsorbed layer may also be affected by changes in concentration or composition of the plasma proteins in the absence of hematocrit changes. Such changes could substantially influence vascular permeability [51, 52]in addition to peripheral resistance and perfusion. This may explain the observation of the strong predictive value of (rather small) increases in plasma viscosity and fibrinogen content for circulatory diseases including stroke, hypertension, peripheral arterial occlusive disease and myocardial infarction [53–56].

Time for primary review 27 days.


   Acknowledgements
 
This study was supported by the Deutsche Forschungsgemeinschaft (Pr 271/1-1, 1-2, 5-1 and 5-2) and by NIH grant HL34555.


   References
Top
 Abstract
 1 Introduction
 2 Materials and methods
 3 Results
 4 Discussion
 References
 

  1. Whittaker S.R.F., Winton F.R. The apparent viscosity of blood flowing in the isolated hindlimb of the dog, and its variation with corpuscular concentration. J Physiol (London) (1933) 78:339–369.[Free Full Text]
  2. Levy M.N., Share L. The influence of erythrocyte concentration upon the pressure-flow relationships in the dog's hind limb. Circ Res (1953) 1:247–255.[Abstract/Free Full Text]
  3. Olson R.M. In vivo blood viscosity and hindrance. Am J Physiol (1964) 206:955–961.[Abstract/Free Full Text]
  4. Murray J.F., Karp R.B., Nadel J.A. Viscosity effects on pressure-flow relations and vascular resistance in dogs' lungs. J Appl Physiol (1969) 27:336–341.[Free Full Text]
  5. Benis A.M., Usami S., Chien S. Effect of hematocrit and inertial losses on pressure-flow relations in the isolated hindpaw of the dog. Circ Res (1970) 27:1047–1068.[Abstract/Free Full Text]
  6. Djojosugito A.M., Folkow B., Öberg B., White S. A comparison of blood viscosity measured in vitro and in a vascular bed. Acta Physiol Scand (1970) 78:70–84.[Web of Science][Medline]
  7. Messmer K., Sunder-Plassmann L., Klövekorn W.P., Holper K. Circulatory significance of hemodilution: Rheological changes and limitations. Adv Microcirc (1972) 4:1–77.
  8. Messmer K., Lewis D.H., Sunder-Plassmann L., Klövekorn W.P., Mendler N., Holper K. Acute normovolemic hemodilution. Changes of central hemodynamics and microcirculatory flow in skeletal muscle. Eur Surg Res (1972) 4:55–70.[Web of Science][Medline]
  9. Benis A.M., Chien S., Usami S., Jan K.M. Inertial pressure losses in perfused hindlimb: a reinterpretation of the results of Whittaker and Winton. J Appl Physiol (1973) 34:383–389.[Free Full Text]
  10. Bucens D., Pain M.C.F. Influence of hematocrit, blood gas tensions, and pH on pressure-flow relations in the isolated canine lung. Circ Res (1975) 37:588–596.[Abstract/Free Full Text]
  11. Öberg B., Little R., Rippe B., Folkow B. The influence of rheological and collapse factors on pre- and post-capillary flow resistances in the skeletal muscle vascular bed of the cat. Acta Physiol Scand (1975) 95:179–190.[Web of Science][Medline]
  12. von Restorff W., Höfling B., Holtz J., Bassenge E. Effect of increased blood fluidity through hemodilution on general circulation at rest and during exercise in dogs. Pflügers Arch (1975) 357:25–34.[CrossRef][Web of Science][Medline]
  13. Nihill M.R., McNamara D.G., Vick R.L. The effects of increased blood viscosity on pulmonary vascular resistance. Am Heart J (1976) 92:65–72.[CrossRef][Web of Science][Medline]
  14. Jan K.M., Chien S. Effect of hematocrit variations on coronary hemodynamics and oxygen utilization. Am J Physiol (1977) 233:H106–H113.[Web of Science][Medline]
  15. Chapler C.K., Cain S.M., Stainsby W.N. Blood flow and oxygen uptake in isolated canine skeletal muscle during acute anemia. J Appl Physiol (1979) 46:1035–1038.[Free Full Text]
  16. Gaehtgens P., Kreutz F., Albrecht K.H. Optimal hematocrit for canine skeletal muscle during rhythmic isotonic exercise. Eur J Appl Physiol (1979) 41:27–39.[CrossRef][Web of Science]
  17. Gustafsson L., Appelgren L., Myrvold H.E. Blood flow and in vivo apparent viscosity in working and non-working skeletal muscle of the dog after high and low molecular weight dextran. Circ Res (1981) 48:465–469.[Abstract/Free Full Text]
  18. Shepherd A.P., Riedel G.L. Optimal hematocrit for oxygenation of canine intestine. Circ Res (1982) 51:233–240.[Abstract/Free Full Text]
  19. Riedel G.L., Scholle J.L., Shepherd A.P., Ward W.F. Effects of hematocrit on oxygenation of the isolated perfused rat liver. Am J Physiol (1983) 245:G769–G774.[Web of Science][Medline]
  20. Julien M., Hakim T.S., Vahi R., Chang H.K. Effect of hematocrit on vascular pressure profile in dog lungs. J Appl Physiol (1985) 58:743–748.[Abstract/Free Full Text]
  21. Crystal G.J. Coronary hemodynamic responses during local hemodilution in canine hearts. Am J Physiol (1988) 254:H525–H531.[Web of Science][Medline]
  22. Sutera S.P., Tilton R.G., Larson K.B., Kilo C.J., Williamson J.R. Vascular flow resistance in rabbit hearts: "apparent viscosity" of RBC suspensions. Microvasc Res (1988) 36:305–313.[CrossRef][Web of Science][Medline]
  23. Kiel J.W., Shepherd A.P. Optimal hematocrit for canine gastric oxygenation. Am J Physiol (1989) 256:H472–H477.[Web of Science][Medline]
  24. Raj J.U., Ramanathan R., Chen P., Anderson J. Effect of hematocrit on microvascular pressures in 3- to 5-wk-old rabbit lungs. Am J Physiol (1989) 256:H766–H771.[Web of Science][Medline]
  25. Sutton D.W., Mead E.H., Schmid-Schönbein G.W. A high precision dual feedback pump for unsteady perfusion of small organs. Ann Biomed Eng (1989) 17:269–278.[CrossRef][Web of Science][Medline]
  26. Tateishi N., Suzuki Y., Soutani M., Maeda N. Flow dynamics of erythrocytes in microvessels of isolated rabbit mesentery: cell-free layer and flow resistance. J Biomech (1994) 27:1119–1125.[CrossRef][Web of Science][Medline]
  27. Pries A.R., Secomb T.W., Gaehtgens P. Structure and hemodynamics of microvascular networks: Heterogeneity and correlations. Am J Physiol (1995) 269:H1713–H1722.[Web of Science][Medline]
  28. Sutton D.W., Schmid-Schönbein G.W. The pressure-flow relation in resting rat skeletal muscle perfused with pure erythrocyte suspensions. Biorheology (1995) 32:29–42.[Web of Science][Medline]
  29. Pries A.R., Secomb T.W., Gaehtgens P., Gross J.F. Blood flow in microvascular networks — experiments and simulation. Circ Res (1990) 67:826–834.[Abstract/Free Full Text]
  30. Pries A.R. A versatile video image analysis system for microcirculatory research. Int J Microcirc Clin Exp (1988) 7:327–345.[Web of Science][Medline]
  31. Pries A.R., Secomb T.W., Gessner T., Sperandio M., Gross J.F., Gaehtgens P. Resistance to blood flow in microvessels in vivo. Circ Res (1994) 75:904–915.[Abstract/Free Full Text]
  32. Pries A.R., Neuhaus D., Gaehtgens P. Blood viscosity in tube flow: dependence on diameter and hematocrit. Am J Physiol (1992) 263:H1770–H1778.[Web of Science][Medline]
  33. Warnke K.C., Skalak T.C. Leukocyte plugging in vivo in skeletal muscle arteriolar trees. Am J Physiol (1992) 262:H1149–H1155.[Web of Science][Medline]
  34. Pries AR, Secomb TW, Jacobs H, Sperandio M, Osterloh K, Gaehtgens P. Microvascular blood flow resistance: Role of the endothelial surface layer. Am J Physiol 1997;273:H2272–H2279.
  35. Braide M., Blixt A., Bagge U. Leukocyte effects on the vascular resistance and glomerular filtration of the isolated rat kidney at normal and low flow states. Circ Shock (1986) 20:71–80.[Web of Science][Medline]
  36. Copley A.L. Hemorheological aspects of the endothelium-plasma interface. Microvasc Res (1974) 8:192–212.[CrossRef][Web of Science][Medline]
  37. Copley A.L., Staple P.H. Hemorheological studies on the plasmatic zone in the microcirculation on the cheek pouch of Chinese and Syrian hamsters. Biorheology (1962) 1:3–14.[Medline]
  38. Duling B.R., Desjardins C. Capillary hematocrit — what does it mean? NIPS (1987) 2:66–69.[Abstract/Free Full Text]
  39. Desjardins C., Duling B.R. Heparinase treatment suggests a role for the endothelial cell glycocalyx in regulation of capillary hematocrit. Am J Physiol (1990) 258:H647–H654.[Web of Science][Medline]
  40. Vink H., Duling B.R. Identification of distinct luminal domains for macromolecules, erythrocytes and leukocytes within mammalian capillaries. Circ Res (1996) 79:581–589.[Abstract/Free Full Text]
  41. Born G.V.R., Palinski W. Unusually high concentrations of sialic acids on the surface of aortic endothelium. Br J Exp Pathol (1985) 66:543–549.[Web of Science][Medline]
  42. Silberberg A. The physiological role of a thick gel-like layer over the endothelial surface. Int J Microcirc Clin Exp (1990) 9(Suppl. 1):204. (Abstract).
  43. Klitzman B., Duling B.R. Microvascular hematocrit and red cell flow in resting and contracting striated muscle. Am J Physiol (1979) 237:H481–H490.[Web of Science][Medline]
  44. Haldenby K.A., Chappell D.C., Winlove C.P., Parker K.H., Firth J.A. Focal and regional variations in the composition of the glycocalyx of large vessel endothelium. J Vasc Res (1994) 31:2–9.[Web of Science][Medline]
  45. Luft J.H. Fine structure of capillary and endocapillary layer as revealed by ruthenium red. Fed Proc (1966) 25:1773–1783.[Web of Science][Medline]
  46. David G. Integral heparan sulfate proteoglycans. FASEB J. (1993) 7:1023–1030.[Abstract]
  47. Schnitzer J.E., Carley W.W., Palade G.E. Specific albumin binding to microvascular endothelium in culture. Am J Physiol (1988) 254:H425–H437.[Web of Science][Medline]
  48. Delvos U., Preissner K.T., Müller-Berghaus G. Binding of fibrinogen to cultured bovine endothelial cells. Thromb Haemost (1985) 18:26–31.
  49. Reinhart W.H., Boulanger C.M., Lüscher T.F., Haeberli A., Straub P.W. Influence of endothelial surface on flow velocity in vitro. Am J Physiol (1993) 265:H523–H529.[Web of Science][Medline]
  50. Witte S. The influence of the fibrinolytic system on the affinity of fibrinogen for the endothelial-plasma interface. Thromb Res (1988) 52:111–117.[CrossRef][Web of Science][Medline]
  51. Schneeberger E.E., Lynch R.D., Neary B.A. Interaction of native and chemically modified albumin with pulmonary microvascular endothelium. Am J Physiol (1990) 258:L89–L98.[Web of Science][Medline]
  52. Schneeberger E.E., Hamelin M. Interaction of serum proteins with lung endothelial glycocalyx: its effect on endothelial permeability. Am J Physiol (1984) 247:H206–H217.[Web of Science][Medline]
  53. Yarnell J.W.G., Baker I.A., Sweetnam P.M., Bainton D., O'Brien J.R., Whitehead P.J., Elwood P.C. Fibrinogen, viscosity, and white blood cell count are major risk factors for ischemic heart disease: The Caerphilly and Speedwell Collaborative Heart Disease Studies. Circulation (1991) 83:836–844.[Abstract/Free Full Text]
  54. Koenig W., Sund M., Ernst E., Keil U., Rosenthal J., Hombach V. Association between plasma viscosity and blood pressure. Results from the MONICA-project Augsburg. Am J Hypertens (1991) 4:529–536.[Web of Science][Medline]
  55. Koenig W., Sund M., Lowe G.D., Lee A.J., Resch K.L., Tunstall-Pedoe H., Keil U., Ernst E. Geographical variations in plasma viscosity and relation to coronary event rates. Lancet (1994) 344:711–714.[CrossRef][Web of Science][Medline]
  56. Scarabin P.Y., Bara L., Ricard S., Poirier O., Cambou J.P., Arveiler D., Luc G., Evans A.E., Samama M.M., Cambien F. Genetic variation at the beta-fibrinogen locus in relation to plasma fibrinogen concentrations and risk of myocardial infarction. The ECTIM Study. Arterioscler Thromb (1993) 13:886–891.[Abstract/Free Full Text]

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 PerfusionHome page
H Vermeer, S Teerenstra, R. de Sevaux, H. van Swieten, and P. Weerwind
The effect of hemodilution during normothermic cardiac surgery on renal physiology and function: a review
Perfusion, November 1, 2008; 23(6): 329 - 338.
[Abstract] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow E-letters: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when E-letters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Pries, A. R
Right arrow Articles by Gaehtgens, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Pries, A. R
Right arrow Articles by Gaehtgens, P.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?