JOURNAL
OF SURGICAL
RESEARCH
53,625-630
Improvement JOSEPHL.UNTHANK,
(19%)
of Flow Through Arterial Stenoses by Drag Reducing Agents
PH.D., STEPHENG.LALKA,
M.D.,J.
CRAIGNIXON, MS., ANDALAN P. SAWCHUK, M.D.
Departments of Surgery and Physiology and Biophysics, Indiana University Richard L. Roudebush Veterans Affairs Medical Center, Indianapolis, Submitted
for publication
November
School of Medicine, Indiana 46202
1, 1991
ing (see review by Morgan and McCormick, [l] ), several studies have also indicated that DRP have potential for biomedical applications. In a brief report in 1971 [2], Greene et al. observed that Separan AP-30, a long-chain polyacrylamide (PA), reduced by 30% the pressure drop of blood flowing through glass tubing during turbulent flow. Subsequent studies by this group found PA to dampen flow instabilities measured by hot wire anemometry in slightly constricted canine descending aortas [3]. Based upon this observation, Mostardi et al. [3] proposed that DRP might have antiatherogenic properties. A study performed in an atherogenic pigeon model [4] provided experimental evidence for this hypothesis and was confirmed in the rabbit model of atherogenesis by Faruqui et al. [5]. Under some conditions, drag reducing polymers also appear to be capable of increasing cardiac output. In rats in which cardiac output was decreased approximately 50% by anesthesia and opening of the thoracic cavity, Coleman et al. [6] found aortic blood flow to increase more than 100% and total peripheral resistance to decrease by more than 50% after intravenous injection of the polyacrylamide Separan AP-273. Similar results were observed with a long-chain polymer of polyethylene oxide (PEO) [7]. Coleman et al. [6] and Polimeni and Ottenbreit [7] also reported preliminary observations that there was little effect of DRP in large animals with normal cardiac output, but “sometimes striking” hemodynamic effects were observed in animals in cardiogenic or hemorheologic shock. More recently, Hutchison et al. [8] reported a reduction in poststenotic tlow disturbance by PA in canine carotid arteries with an artificial stenosis (41% diameter reduction). However, neither aortic nor carotid artery blood flow were significantly altered. In the most recent study with DRP, Sumpio et al. [9] observed that the addition of PA could improve renal function in isolated, perfused rat kidneys. The studies cited above have demonstrated the antiatherogenic potential of DRP and DRP’s ability to decrease flow disturbances distal to an arterial stenosis and improve hemodynamics in cardiac depressed animals. If DRP are primarily effective under conditions of
The potential of drag reducing polymers (DRP) to selectively improve blood flow through clinically significant arterial stenoses was investigated. An artificial stenosis of the left common iliac artery in dogs decreased left femoral artery pressure by 25%. High-molecular-weight polyacrylamide (PA) or polyethylene oxide (PEO) were infused at a slow constant rate while we measured left and right common iliac artery blood flows and left and right femoral artery and vein pressures. As DRP were infused, left iliac artery flow (QL) increased early and then decreased to baseline values as flow began to increase in the right iliac artery. The peak increase in QL was 24 f 9% for PA and 46 f 19% for PEO and occurred before right iliac artery flow (Qn) increased. As additional polymer was infused, Qn increased to a maximum of 41 f 12 and 131-+ 40% with PA and PEO, respectively. Femoral artery pressures and hindlimb resistances tended to decrease in both limbs but the only significant differences occurred in the right (nonstenosed) side when Qz was elevated. This study provides the first evidence that low concentrations of DRP might be capable of improving blood flow through stenotic blood vessels without altering flow in normal vessels. Although DRP might represent a new class of compounds that could be utilized in the treatment of cardiovascular diseases, the degreeof variation in individual responses is a concern, the exact mechanism of action is unclear, and information on pharmacodynamics is lacking. 0 1992 Academic Press, Inc.
INTRODUCTION Drag reducing polymers (DRP) are a distinct class of chemical compounds that reduce resistance to turbulent flow in pipes. DRP are characterized by extraordinary molecular weights (> 106) or the capability to form highmolecular-weight aggregates [ 11. Under conditions of turbulent flow, these agents can reduce the frictional resistance to flow by as much as 75%. Although DRP have been utilized primarily in nonbiological applications to reduce the energy requirements for fluid pump625
0022-4804/92
$4.00
Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
626
JOURNAL
OF SURGICAL
RESEARCH:
turbulent flow, as is generally thought [ 1, lo], they might also be capable of selectively improving flow through stenotic vessels associated with turbulence. The purpose of this study was to determine if the administration of DRP would selectively improve blood flow through clinically significant arterial stenoses. Two DRP were evaluated: a PA that is industrially similar to the commonly used Separan AP-273 and AP-30 [2-6,8,9,11,12] (that are no longer manufactured) and a high-molecularweight PEO. METHODS
Eight greyhound or foxhound dogs (27 f 0.7 kg) were successfully studied. All animals were sedated with intravenous xylazine (1 mg/kg) and anesthetized with 30 mg/ kg pentobarbital sodium. Endotracheal intubation was performed for volume-controlledventilation. An intravenous drip line was inserted into the jugular vein and approximately 500 ml/hr of a physiological ringers solution was delivered to maintain blood volume. Anesthesia was maintained by the constant delivery of pentobarbital(3 mg/kg/hr) with a syringe pump connected to the iv line. A midline abdominal incision was made and the left and right iliac arteries were isolated. Ultrasonic, perivascular flow probes (Transonic Systems, Inc., Ithaca, NY) were positioned on both proximal common iliac arteries. The middle sacral artery was cannulated with a 20-gauge angiocatheter to measure aortic pressures and the vessel was ligated distally. Catheters were placed in left and right femoral arteries and veins for pressure measurements. An artificial stenosis was created by encircling the left distal common iliac artery with a i-in. umbilical tape and progressively circumferentially compressing the vessel with a Rommel tourniquet system. Thirty minutes were allowed for the hemodynamic parameters to stabilize after creation of the stenosis. Two drag reducing polymers were used in this study; a polyacrylamide with an approximate molecular weight of 14 X lo6 (Praestol 2515 TR, Stockhausen, Inc., Greensboro, NC) and a polyethylene oxide of approximately 4 x lo6 MW (Polyox WSR 301, NF grade, Union Carbide, Danbury, CT). Both polymers were prepared similar to the method recommended by the manufacturers. A mechanical stirrer (Model 700-4510) with a turbine propeller (No. 4544-10, Barnant Co., Barrington, IL) was used to create a vortex in 100 ml of phosphate-buffered saline in a 500-ml beaker into which 0.50 g of DRP was sprinkled. The solution was then transferred to a 1 liter polypropylene bottle and placed on a laboratory rocker (Buchler Instruments, Fort Lee, NJ) for 16-20 hr. The solutions were stored at 4°C. During the experiments, the DRP solutions were delivered by a syringe pump (Model 22, Havard Apparatus, South Natick, MA) for 30 min at a rate of 0.80 or 0.16 mg/kg/min for PA and PEO. These slow, continuous infusions of
VOL.
53, NO. 6, DECEMBER
1992
DRP were made while monitoring hemodynamic parameters in order to estimate the dosages that might be effective in producing selective improvement of blood flow through stenosed arteries.
Measurements Superficial femoral artery and vein pressures were measured in both hindlimbs using pressure transducers (Model 156PCO6 GW12, Microswitch, Freeport, IL) connected to a general purpose amplifier (Model GPA-4, Stemtech, Inc., Houston, TX). Left and right iliac artery blood flows were measured with 6-mm perivascular flow probes and an ultrasonic bloodflow meter (Transonic Systems, Inc., Ithaca, NY). Flows and pressures were measured from approximately 10 min before creation of the artificial stenosis until 30 min after DRP infusion was stopped. Averaged flow and pressure measurements were acquired by a data acquisition system consisting of an IBM AT-compatible computer (HP Vectra ES, Hewlett-Packard Co., Palo Alto, CA), analog input board (DAS-8PGA, Metrabyte Corp., Taunton, MA), and software (LabTech Notebook, V5.0, Laboratory Technologies Corp., Wilmington, MA) at a rate of 1 Hz. Ten-second block averages of the digitized data were recorded on floppy diskettes for off-line computer analysis. Calculations
and Statistics
From the pressure and flow data, right and left hindlimb resistances were calculated from the equation R = (P, - Pv)/Q, where R is resistance, P, is femoral artery pressure, P, is femoral vein pressure, and Q is iliac artery blood flow. These calculations were performed for each animal in a data spreadsheet (Lotus 123 r2.01, Lotus, Cambridge, MA) containing all the computer-acquired data. The time points at which maximum left and right arterial blood flows occurred during DRP infusion were identified from line plots of each animal and expressed relative to the time the infusion began, as illustrated in Fig. 1. To reduce variation that existed between animals in baseline blood flows, the responses observed during DRP infusion were expressed as percentage of change from baseline. Statistical comparisons were made between left and right arterial pressure and arterial flow and hindlimb resistance measurements, made approximately 30 min after creation of the arterial stenosis and immediately before DRP infusion, with one-way analysis of variance. During the infusion of DRP, the percentage of change from baseline for each of these parameters at the time points when left and right flows were maximum were compared by a two-way ANOVA (DRP X time). A statistical analysis software program (StatPackets, Walonick Associates, Minneapolis, MN) was used with the data spreadsheet for these statistical evaluations at P G 0.05.
UNTHANK
ET AL.: FLOW
IMPROVEMENT cI
200
I
I”
150
-
P 4
50
-
J
01 -15
_ 200 m = 150
--
E
-
n. a
100 50
t?z
Relative
Time
I
t
I 15
I 30
I 45
I
I
I
1
I 0
I 15
I 30
I 45
I
,
a 0
t
01 -15
Relative
(min)
FIG. 1. A representative record of the effect of polymer infusion on arterial blood flows in an individual animal is illustrated by line plots for the left (stenosed) and right common iliac arteries. An intravenous infusion of PEO was started at 0 min and stopped at 30 min. The time points where left and right blood flows peaked during polymer infusion are indicated by arrows and designated as TLMAX and
627
BY DRP
Time
(min)
FIG. 2. The effects of polymer infusion on left superficial femoral artery pressure (LAP) distal to the stenosed common iliac artery and right superficial femoral artery pressure (RAP) are illustrated for the same animal selected to represent the response to flow in Fig. 1. A decrease in pressure followed by a recovery was the typical response observed during polymer infusion.
TRMW
RESULTS
The average measurements of iliac artery blood flows, femoral artery pressures, and hindlimb resistances made at least 30 min after creation of the artificial stenosis of the left iliac artery and immediately prior to DRP infusion are reported in Table 1. Pressure in the left femoral artery averaged 33 mm Hg, or 25% less than right femoral artery pressure, verifying the existence of a clinically significant stenosis. Blood flow through the left iliac artery was 35% less than the 190 ml/min flow through the right iliac artery. Total hindlimb resistance averaged 15% higher on the left than on the right side. Figures l-3 illustrate, from data collected from the same dog, the effect of DRP infusion on arterial flow, arterial pressure, and hindlimb resistance. In the top graph of Fig. 1, blood flow through the left (stenosed) iliac artery peaked early during DRP infusion and then returned to baseline or slightly above. Blood flow in the
right iliac artery (Fig. 1, bottom) initially decreased and then later increased and peaked at a much higher level than left flow. This early decrease in blood flow in the right iliac artery was not observed in all records. The times at which left and right flow peaked are designated as TLMA~ and TRM~, respectively. Figure 2 represents the effect of DRP infusion on the blood pressure in the left (top) and right (bottom) femoral artery. On both sides, pressure tended to increase initially during DRP infusion and then decreased rapidly followed by a gradual recovery. The plots of hindlimb resistance in Fig. 3 show an initial rise in right resistance (bottom) and then
-15
TABLE
1
Left and Right Common Iliac Artery Superficial Femoral Artery Pressures, Resistances; Poststenosis, Prepolymer Pressure (mm Hg) Right Left (stenosed)
134 + 5.8 101 f 10.1*
* Left # Right; P < 0.05.
Flow (ml/min) 190 f 46.7 129 t 36.5*
Blood Flows, and Hindlimb
Resistance (mm Hg - min/ml) 0.81 + 0.10 0.90 -t 0.12
I 0
I 15
I 30
I 45
30
45
: 2.0 ! 1.5 .‘D g 1.0 of 2
ul0.5 *z0.0 -15
0
15 Relative
Time
(min)
FIG. 3. Resistance was calculated from the experimental data collected from each animal. The line plots above illustrate the changes that occurred in hindlimb resistances on the artificially stenosed (left) and control (right) sides for the same animal as Figs. 1 and 2.
628
JOURNAL
TABLE
OF SURGICAL
RESEARCH:
VOL.
53, NO. 6, DECEMBER
LEFT
2
1992
RIGHT
Time during DRP Infusion when Maximum (Peak) Blood Flow Occurred in Left (Stenosed) and Right (Control) Common Iliac Arteries Time (min)
Control Stenosed * Stenosed # Control;
PA
PEO
3.6 f 0.77 1.5 k 0.16*
7.5 -t 2.65 2.1 + 0.60*
P =S0.05.
a marked decrease in both left and right resistance. As the infusion of DRP continued, left resistance (top) returned toward baseline more quickly than right resistance. The average times for peak left and right iliac artery blood flows are reported in Table 2 for PA and PEO. TLm occurred significantly earlier than TRMAX for both PEO and PA. Although both TLMAx and TRMAX tended to be greater for PEO than those for PA, there was a difference in the infusion rates of PA and PEO (0.80 vs 0.16 mg/kg/min, respectively). The average changes in arterial flows, pressures, and hindlimb resistances at TLMAx and TRMAX are depicted in Figs. 4-6. Values are expressed as percent change from baseline to reduce variations between individual dogs. Responses to PA and PEO in both the stenosed (left)
LEFT
FIG. 5. The average percentage change in left (LAP) and right (RAP) femoral arterial pressures at the time of maximum flow on the left ( TWAx) and right ( TaMnx ) sides is depicted for both PA and PEO. Statistically significant differences (P < 0.05) are denoted for changes from baseline (0) and between TLMAX and TaMAx (*).
and control (right) limbs are shown in separate panels for each figure. At TIMAX, blood flow through the stenosed (left) artery was increased from baseline during infusion of both PA (24 + 9.3%; range 8-43%) and PEO (46 + 19.4%; range 22-104%). Blood flow through the right iliac ar-
RIGHT LEFT
RIGHT
20
0
-20
-40
l& T LNAX n=PEO
T RNAX
T LMAX
-60 T RHAX
m=PA
T LMAX
T RMAX
n=PEO FIG. 4. The average percentage change in left (stenosed, left) and right (control, right) arterial blood flows at the time of maximum flow sides is depicted for both PA and on the left ( TLMAx ) and right ( Tm) PEO. Statistically significant differences (P < 0.05) are denoted for changes from baseline (0), between TLMAx and TRMAX (*), and between PA and PEO (+).
TLMAX
T RMAX
m=PA
FIG. 6. The average percentage change in left and right hindlimb resistances at the time of maximum flow on the left ( TLmx) and right signifi(TRMm ) sides is depicted for both PA and PEO. Statistically cant differences (P < 0.05) are denoted for changes from baseline (0), between TLMAx and TRIMAx (*), and between PA and PEO (+).
UNTHANK
tery
ET AL.: FLOW
was not
significantly altered from baseline at flow through the stenosed artery had T LMAX.A~TRMAX, decreased and was no longer significantly different than pre-DRP values. Flow through the right iliac artery was increased for both PA (40 + 12%; range 16-73%) and PEO (131 + 39.8%; range 73-249%); however, only the increase during PEO administration was statistically significant. There was no significant change in arterial pressures in either limb at TLMAX.At TRm arterial pressures in both limbs tended to be decreased but the only significant change was a 15 & 5.7% decrease in right femoral artery pressure during the infusion of PEO (Fig. 5). Although blood flow through the stenosed artery was increased at TLMAX(Fig. 4), this occurred without a significant decrease in left hindlimb resistance (Fig. 6). The only changes in hindlimb resistances during DRP infusion occurred on the right side at TRMAX.The right resistances were decreased 32 + 6.5% (range 22-51%) by PA and 60 + 6.7% (range 48-80%) by PEO. The decrease in right hindlimb resistance with PEO was statistically greater than with PA. DISCUSSION
The most important finding of this study was that during a slow, constant intravenous infusion of DRP, blood flow through an iliac artery with a clinically significant stenosis was significantly elevated before blood flow through the contralateral vessel was increased (Figs. 1 and 4 and Table 2). The hemodynamic data reported in Table 1 verify the existence of a clinically significant arterial stenosis in our experimental model. At the time point when blood flow through the stenosed vessel was maximum, arterial pressure distal to the stenosis was not altered (Fig. 5) and hindlimb vascular resistance tended to be decreased (Fig. 6) but the changes were not statistically significant. Because no dose-response data have been reported in the biomedical studies that have used DRP, we had no indication of the threshold concentration of DRP that might produce changes in blood flow through arterial stenoses. For this reason, we made slow, continuous infusions of both PA and PEO while monitoring arterial blood flows and pressures. The changes in hemodynamics that occurred during the polymer infusion were assumed to be concentration dependent. Coleman et al. [ 61 has reported that for PA in rats, a gradual decrease in the hemodynamic effects occurs over a period of l-2 hr postinjection but little if any change takes place during the first 15-30 min. Since our maximum increases in blood flow through a stenosed vessel occurred in less than 2.5 min after initiation of DRP administration (Fig. 4) and because the increase in blood flow on the stenosed side (where greater turbulence should have existed) occurred significantly earlier than maximum effects on the right side (Figs. 1 and 5 and Table 2), we believe the difference observed between the left and
IMPROVEMENT
BY DRP
629
right sides was due to a concentration-dependent, rather than temporal, effect. If the changes in hemodynamics that occur during polymer infusion are concentration dependent, these data would suggest that the maximum effect on blood flow through a critical arterial stenosis in a canine iliac artery would occur at a concentration of approximately 0.3 mg/kg for PEO and 1.2 mg/kg for PA (calculated from TLMAXvalues in Table 2 and infusion rates). These conct,,trations are lower than those used in earlier studies which produced alterations in central hemodynamics [3, 71 or reduced flow disturbances in subcritical arterial stenoses [6, 81. Although previous studies have demonstrated that flow disturbances produced by subcritical stenoses in dog aorta [3] and carotid artery [8] are reduced by DRP without affecting aortic blood flow [8] and that DRP may have antiatherogenic potential [4,5], this is the first study to demonstrate that DRP might be capable of improving blood flow through clinically significant arterial stenoses without altering blood flow through other vessels. Even though the results of this and earlier studies suggest that the class of compounds known as drag reducing polymers may have therapeutic potential in cardiovascular diseases, we have several major concerns regarding the use of DRP and, clearly, there is a need for much additional information regarding mechanisms of action and pharmacodynamics. One concern relates to the variability of the responses observed with administration of DRP. Coleman et al. reported [6] that in some cases in large animals the effects of DRP on hemodynamics were negligible but in other cases “striking.” Although the reports of the effects of DRP on flow disturbances in dogs [3, 81 do not make mention of variability, considerable differences in the percentage of increase in iliac artery blood flow between individual dogs were observed in our studies. During the infusion of DRP, we observed the maximum increase in blood flow through the stenosed artery to vary from 8 to 104%. Furthermore, although increases in blood flow occurred early, in some cases blood flows had dropped to a fraction of baseline levels before DRP administration was stopped. Polimeni has reported that concentrations of DRP not much higher than those used to produce maximum effects can be lethal [6, 111. The exact mechanism that mediates the effect of DRP in uiuo has not been established. Several theories have been proposed to explain drag reduction in pipes [ 1, 10, 131. Both in pipes and in the cardiovascular system the ability of DRP to alter fluid dynamics is thought to be dependent on turbulent or disturbed flow [ 1, 61. For DRP to decrease resistance in uiuo, either blood viscosity or vascular tone must be decreased. The initial report of the effect of DRP on blood flow through glass tubing [3] would suggest that the apparent viscosity of blood is reduced when turbulence is present. However, we have been unable to confirm this original finding with an in vitro perfusion loop. Furthermore, Greene et al. found
630
JOURNAL
OF SURGICAL
RESEARCH:
that with in vitro perfusion loops either drag reduction or drag enhancement could occur depending upon the geometry of the stenosis [12]. Measurements made in viscometers have not found blood viscosity to be decreased by the addition of DRP [ 21. With these considerations in mind, it is not apparent why we observed blood flow through the arterial stenosis to increase and then decrease as blood flow through the normal vessel continued to increase (Fig. 1) or why a relatively large drop in arterial pressure occurs early in the period when DRP is infused (Fig. 2). Such results may be more consistent with peripheral vasodilation than a decrease in apparent blood viscosity under conditions of turbulent blood flow. If DRP do indeed act to decrease resistance to disturbed blood flow by some mechanism other than peripheral vasodilation, much additional information will be needed regarding dose dependency, mechanisms of action, pharmacodynamics, and toxicology before the clinical potential of DRP can be determined. However, the handful of studies that have examined the cardiovascular effects of DRP indicate this class of compounds can have dramatic effects on increasing cardiac output in cardiac-depressed animals, improving blood flow through arterial stenoses, and inhibiting atherogenesis.
VOL.
theories
structure.
Prog. Polymer Sci.
implications of drag reducing agents. Biorheology 7: 221, 1971. 3. Mostardi, R. A., Greene, H. L., Nokes, R. F., Thomas, L. C., and Lue, T. The effect of drag reducing agents on stenotic flow disturbances in dogs. Biorheology 13: 137, 1976. 4. Greene, H. L., Mostardi, R. F., and Nokes, R. F. Effects of drag reducing polymers on initiation of atherosclerosis. Polymer Eng. Sci. 20: 499, 1980. 5. Faruqui, F. I., Otten, M. D., and Polimeni, P. I. Protection against atherogenesis with the polymer drag-reducing agent Separan AP-30. Circulation 75: 627, 1987. 6. Coleman, P. B., Ottenbreit, B. T., and Polimeni, P. I. Effects of a drag-reducing polyelectrolyte of microscopic linear dimension (Separan AP-273) on rat hemodynamics. Circ. Res. 61: 787,
1987. 7. Polimeni,
8.
9.
11. 12.
REFERENCES Morgan, S. E., and McCormick, C. L. Water-soluble copolymers XxX11: Macromolecular drag reduction. A review of predictive
and the effects of polymer
2. Greene, H. L., Nokes, R. F., and Thomas, L. C. Biomedical
LO.
1.
1992
15:507,1990.
ACKNOWLEDGMENTS This work was supported in part by a Department of Veterans Affairs Merit Review Grant (S.G.L.) and a grant from the Indiana University Research Investment Fund (J.L.U. and S.G.L.).
53, NO. 6, DECEMBER
13.
P. I., and Ottenbreit, B. T. Hemodynamic effects of a Poly(ethylene oxide) drag reducing polymer, polyox WSR N-60K, in the open-chest rat. J. Cardiouasc. Pharmacol. 14: 374, 1989. Hutchison, K. J., Campbell, J. D., and Karpinski, E. Decreased poststenotic flow disturbance during drag reduction by polyacrylamide infusion without increased aortic blood flow. Microuasc. Res. 38: 102, 1989. Sumpio, B. E., Upchurch, G. R., and Johnson, G. The influence of perfusate viscosity, RBC deformability and drag of function of an isolated perfused rat kidney. J. Surg. Res. 46: 4, 1989. Sellin, J., Hoyt, J. W., and Scrivener, 0. The effect of drag-reducing additives on fluid flows and their industrial applications. J. Hydraul. Res. 20: 29, 1980. Polimeni, P. I., Ottenbreit, B., and Coleman, P. Enhancement of aortic blood flow with a linear anionic macro-polymer of extraordinary molecular length. J. Mol. Cell. Cardiol. 17: 721, 1985. Greene, H. L., Nokes, R. F., and Thomas, L. C. Drag reduction phenomena in pulsed blood flow. In R. B. Dowdell (Ed.), Flow, Its Measurements and Control in Science and Industry. Ann Arbor, MI: Book on Demand, 1974. Pp. 1459-1464. Berman, N. S. Drag reduction by polymers. Ann. Reu. Fluid Mech. 10: 47, 1978.
OF SURGICAL
RESEARCH
53,625-630
Improvement JOSEPHL.UNTHANK,
(19%)
of Flow Through Arterial Stenoses by Drag Reducing Agents
PH.D., STEPHENG.LALKA,
M.D.,J.
CRAIGNIXON, MS., ANDALAN P. SAWCHUK, M.D.
Departments of Surgery and Physiology and Biophysics, Indiana University Richard L. Roudebush Veterans Affairs Medical Center, Indianapolis, Submitted
for publication
November
School of Medicine, Indiana 46202
1, 1991
ing (see review by Morgan and McCormick, [l] ), several studies have also indicated that DRP have potential for biomedical applications. In a brief report in 1971 [2], Greene et al. observed that Separan AP-30, a long-chain polyacrylamide (PA), reduced by 30% the pressure drop of blood flowing through glass tubing during turbulent flow. Subsequent studies by this group found PA to dampen flow instabilities measured by hot wire anemometry in slightly constricted canine descending aortas [3]. Based upon this observation, Mostardi et al. [3] proposed that DRP might have antiatherogenic properties. A study performed in an atherogenic pigeon model [4] provided experimental evidence for this hypothesis and was confirmed in the rabbit model of atherogenesis by Faruqui et al. [5]. Under some conditions, drag reducing polymers also appear to be capable of increasing cardiac output. In rats in which cardiac output was decreased approximately 50% by anesthesia and opening of the thoracic cavity, Coleman et al. [6] found aortic blood flow to increase more than 100% and total peripheral resistance to decrease by more than 50% after intravenous injection of the polyacrylamide Separan AP-273. Similar results were observed with a long-chain polymer of polyethylene oxide (PEO) [7]. Coleman et al. [6] and Polimeni and Ottenbreit [7] also reported preliminary observations that there was little effect of DRP in large animals with normal cardiac output, but “sometimes striking” hemodynamic effects were observed in animals in cardiogenic or hemorheologic shock. More recently, Hutchison et al. [8] reported a reduction in poststenotic tlow disturbance by PA in canine carotid arteries with an artificial stenosis (41% diameter reduction). However, neither aortic nor carotid artery blood flow were significantly altered. In the most recent study with DRP, Sumpio et al. [9] observed that the addition of PA could improve renal function in isolated, perfused rat kidneys. The studies cited above have demonstrated the antiatherogenic potential of DRP and DRP’s ability to decrease flow disturbances distal to an arterial stenosis and improve hemodynamics in cardiac depressed animals. If DRP are primarily effective under conditions of
The potential of drag reducing polymers (DRP) to selectively improve blood flow through clinically significant arterial stenoses was investigated. An artificial stenosis of the left common iliac artery in dogs decreased left femoral artery pressure by 25%. High-molecular-weight polyacrylamide (PA) or polyethylene oxide (PEO) were infused at a slow constant rate while we measured left and right common iliac artery blood flows and left and right femoral artery and vein pressures. As DRP were infused, left iliac artery flow (QL) increased early and then decreased to baseline values as flow began to increase in the right iliac artery. The peak increase in QL was 24 f 9% for PA and 46 f 19% for PEO and occurred before right iliac artery flow (Qn) increased. As additional polymer was infused, Qn increased to a maximum of 41 f 12 and 131-+ 40% with PA and PEO, respectively. Femoral artery pressures and hindlimb resistances tended to decrease in both limbs but the only significant differences occurred in the right (nonstenosed) side when Qz was elevated. This study provides the first evidence that low concentrations of DRP might be capable of improving blood flow through stenotic blood vessels without altering flow in normal vessels. Although DRP might represent a new class of compounds that could be utilized in the treatment of cardiovascular diseases, the degreeof variation in individual responses is a concern, the exact mechanism of action is unclear, and information on pharmacodynamics is lacking. 0 1992 Academic Press, Inc.
INTRODUCTION Drag reducing polymers (DRP) are a distinct class of chemical compounds that reduce resistance to turbulent flow in pipes. DRP are characterized by extraordinary molecular weights (> 106) or the capability to form highmolecular-weight aggregates [ 11. Under conditions of turbulent flow, these agents can reduce the frictional resistance to flow by as much as 75%. Although DRP have been utilized primarily in nonbiological applications to reduce the energy requirements for fluid pump625
0022-4804/92
$4.00
Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
626
JOURNAL
OF SURGICAL
RESEARCH:
turbulent flow, as is generally thought [ 1, lo], they might also be capable of selectively improving flow through stenotic vessels associated with turbulence. The purpose of this study was to determine if the administration of DRP would selectively improve blood flow through clinically significant arterial stenoses. Two DRP were evaluated: a PA that is industrially similar to the commonly used Separan AP-273 and AP-30 [2-6,8,9,11,12] (that are no longer manufactured) and a high-molecularweight PEO. METHODS
Eight greyhound or foxhound dogs (27 f 0.7 kg) were successfully studied. All animals were sedated with intravenous xylazine (1 mg/kg) and anesthetized with 30 mg/ kg pentobarbital sodium. Endotracheal intubation was performed for volume-controlledventilation. An intravenous drip line was inserted into the jugular vein and approximately 500 ml/hr of a physiological ringers solution was delivered to maintain blood volume. Anesthesia was maintained by the constant delivery of pentobarbital(3 mg/kg/hr) with a syringe pump connected to the iv line. A midline abdominal incision was made and the left and right iliac arteries were isolated. Ultrasonic, perivascular flow probes (Transonic Systems, Inc., Ithaca, NY) were positioned on both proximal common iliac arteries. The middle sacral artery was cannulated with a 20-gauge angiocatheter to measure aortic pressures and the vessel was ligated distally. Catheters were placed in left and right femoral arteries and veins for pressure measurements. An artificial stenosis was created by encircling the left distal common iliac artery with a i-in. umbilical tape and progressively circumferentially compressing the vessel with a Rommel tourniquet system. Thirty minutes were allowed for the hemodynamic parameters to stabilize after creation of the stenosis. Two drag reducing polymers were used in this study; a polyacrylamide with an approximate molecular weight of 14 X lo6 (Praestol 2515 TR, Stockhausen, Inc., Greensboro, NC) and a polyethylene oxide of approximately 4 x lo6 MW (Polyox WSR 301, NF grade, Union Carbide, Danbury, CT). Both polymers were prepared similar to the method recommended by the manufacturers. A mechanical stirrer (Model 700-4510) with a turbine propeller (No. 4544-10, Barnant Co., Barrington, IL) was used to create a vortex in 100 ml of phosphate-buffered saline in a 500-ml beaker into which 0.50 g of DRP was sprinkled. The solution was then transferred to a 1 liter polypropylene bottle and placed on a laboratory rocker (Buchler Instruments, Fort Lee, NJ) for 16-20 hr. The solutions were stored at 4°C. During the experiments, the DRP solutions were delivered by a syringe pump (Model 22, Havard Apparatus, South Natick, MA) for 30 min at a rate of 0.80 or 0.16 mg/kg/min for PA and PEO. These slow, continuous infusions of
VOL.
53, NO. 6, DECEMBER
1992
DRP were made while monitoring hemodynamic parameters in order to estimate the dosages that might be effective in producing selective improvement of blood flow through stenosed arteries.
Measurements Superficial femoral artery and vein pressures were measured in both hindlimbs using pressure transducers (Model 156PCO6 GW12, Microswitch, Freeport, IL) connected to a general purpose amplifier (Model GPA-4, Stemtech, Inc., Houston, TX). Left and right iliac artery blood flows were measured with 6-mm perivascular flow probes and an ultrasonic bloodflow meter (Transonic Systems, Inc., Ithaca, NY). Flows and pressures were measured from approximately 10 min before creation of the artificial stenosis until 30 min after DRP infusion was stopped. Averaged flow and pressure measurements were acquired by a data acquisition system consisting of an IBM AT-compatible computer (HP Vectra ES, Hewlett-Packard Co., Palo Alto, CA), analog input board (DAS-8PGA, Metrabyte Corp., Taunton, MA), and software (LabTech Notebook, V5.0, Laboratory Technologies Corp., Wilmington, MA) at a rate of 1 Hz. Ten-second block averages of the digitized data were recorded on floppy diskettes for off-line computer analysis. Calculations
and Statistics
From the pressure and flow data, right and left hindlimb resistances were calculated from the equation R = (P, - Pv)/Q, where R is resistance, P, is femoral artery pressure, P, is femoral vein pressure, and Q is iliac artery blood flow. These calculations were performed for each animal in a data spreadsheet (Lotus 123 r2.01, Lotus, Cambridge, MA) containing all the computer-acquired data. The time points at which maximum left and right arterial blood flows occurred during DRP infusion were identified from line plots of each animal and expressed relative to the time the infusion began, as illustrated in Fig. 1. To reduce variation that existed between animals in baseline blood flows, the responses observed during DRP infusion were expressed as percentage of change from baseline. Statistical comparisons were made between left and right arterial pressure and arterial flow and hindlimb resistance measurements, made approximately 30 min after creation of the arterial stenosis and immediately before DRP infusion, with one-way analysis of variance. During the infusion of DRP, the percentage of change from baseline for each of these parameters at the time points when left and right flows were maximum were compared by a two-way ANOVA (DRP X time). A statistical analysis software program (StatPackets, Walonick Associates, Minneapolis, MN) was used with the data spreadsheet for these statistical evaluations at P G 0.05.
UNTHANK
ET AL.: FLOW
IMPROVEMENT cI
200
I
I”
150
-
P 4
50
-
J
01 -15
_ 200 m = 150
--
E
-
n. a
100 50
t?z
Relative
Time
I
t
I 15
I 30
I 45
I
I
I
1
I 0
I 15
I 30
I 45
I
,
a 0
t
01 -15
Relative
(min)
FIG. 1. A representative record of the effect of polymer infusion on arterial blood flows in an individual animal is illustrated by line plots for the left (stenosed) and right common iliac arteries. An intravenous infusion of PEO was started at 0 min and stopped at 30 min. The time points where left and right blood flows peaked during polymer infusion are indicated by arrows and designated as TLMAX and
627
BY DRP
Time
(min)
FIG. 2. The effects of polymer infusion on left superficial femoral artery pressure (LAP) distal to the stenosed common iliac artery and right superficial femoral artery pressure (RAP) are illustrated for the same animal selected to represent the response to flow in Fig. 1. A decrease in pressure followed by a recovery was the typical response observed during polymer infusion.
TRMW
RESULTS
The average measurements of iliac artery blood flows, femoral artery pressures, and hindlimb resistances made at least 30 min after creation of the artificial stenosis of the left iliac artery and immediately prior to DRP infusion are reported in Table 1. Pressure in the left femoral artery averaged 33 mm Hg, or 25% less than right femoral artery pressure, verifying the existence of a clinically significant stenosis. Blood flow through the left iliac artery was 35% less than the 190 ml/min flow through the right iliac artery. Total hindlimb resistance averaged 15% higher on the left than on the right side. Figures l-3 illustrate, from data collected from the same dog, the effect of DRP infusion on arterial flow, arterial pressure, and hindlimb resistance. In the top graph of Fig. 1, blood flow through the left (stenosed) iliac artery peaked early during DRP infusion and then returned to baseline or slightly above. Blood flow in the
right iliac artery (Fig. 1, bottom) initially decreased and then later increased and peaked at a much higher level than left flow. This early decrease in blood flow in the right iliac artery was not observed in all records. The times at which left and right flow peaked are designated as TLMA~ and TRM~, respectively. Figure 2 represents the effect of DRP infusion on the blood pressure in the left (top) and right (bottom) femoral artery. On both sides, pressure tended to increase initially during DRP infusion and then decreased rapidly followed by a gradual recovery. The plots of hindlimb resistance in Fig. 3 show an initial rise in right resistance (bottom) and then
-15
TABLE
1
Left and Right Common Iliac Artery Superficial Femoral Artery Pressures, Resistances; Poststenosis, Prepolymer Pressure (mm Hg) Right Left (stenosed)
134 + 5.8 101 f 10.1*
* Left # Right; P < 0.05.
Flow (ml/min) 190 f 46.7 129 t 36.5*
Blood Flows, and Hindlimb
Resistance (mm Hg - min/ml) 0.81 + 0.10 0.90 -t 0.12
I 0
I 15
I 30
I 45
30
45
: 2.0 ! 1.5 .‘D g 1.0 of 2
ul0.5 *z0.0 -15
0
15 Relative
Time
(min)
FIG. 3. Resistance was calculated from the experimental data collected from each animal. The line plots above illustrate the changes that occurred in hindlimb resistances on the artificially stenosed (left) and control (right) sides for the same animal as Figs. 1 and 2.
628
JOURNAL
TABLE
OF SURGICAL
RESEARCH:
VOL.
53, NO. 6, DECEMBER
LEFT
2
1992
RIGHT
Time during DRP Infusion when Maximum (Peak) Blood Flow Occurred in Left (Stenosed) and Right (Control) Common Iliac Arteries Time (min)
Control Stenosed * Stenosed # Control;
PA
PEO
3.6 f 0.77 1.5 k 0.16*
7.5 -t 2.65 2.1 + 0.60*
P =S0.05.
a marked decrease in both left and right resistance. As the infusion of DRP continued, left resistance (top) returned toward baseline more quickly than right resistance. The average times for peak left and right iliac artery blood flows are reported in Table 2 for PA and PEO. TLm occurred significantly earlier than TRMAX for both PEO and PA. Although both TLMAx and TRMAX tended to be greater for PEO than those for PA, there was a difference in the infusion rates of PA and PEO (0.80 vs 0.16 mg/kg/min, respectively). The average changes in arterial flows, pressures, and hindlimb resistances at TLMAx and TRMAX are depicted in Figs. 4-6. Values are expressed as percent change from baseline to reduce variations between individual dogs. Responses to PA and PEO in both the stenosed (left)
LEFT
FIG. 5. The average percentage change in left (LAP) and right (RAP) femoral arterial pressures at the time of maximum flow on the left ( TWAx) and right ( TaMnx ) sides is depicted for both PA and PEO. Statistically significant differences (P < 0.05) are denoted for changes from baseline (0) and between TLMAX and TaMAx (*).
and control (right) limbs are shown in separate panels for each figure. At TIMAX, blood flow through the stenosed (left) artery was increased from baseline during infusion of both PA (24 + 9.3%; range 8-43%) and PEO (46 + 19.4%; range 22-104%). Blood flow through the right iliac ar-
RIGHT LEFT
RIGHT
20
0
-20
-40
l& T LNAX n=PEO
T RNAX
T LMAX
-60 T RHAX
m=PA
T LMAX
T RMAX
n=PEO FIG. 4. The average percentage change in left (stenosed, left) and right (control, right) arterial blood flows at the time of maximum flow sides is depicted for both PA and on the left ( TLMAx ) and right ( Tm) PEO. Statistically significant differences (P < 0.05) are denoted for changes from baseline (0), between TLMAx and TRMAX (*), and between PA and PEO (+).
TLMAX
T RMAX
m=PA
FIG. 6. The average percentage change in left and right hindlimb resistances at the time of maximum flow on the left ( TLmx) and right signifi(TRMm ) sides is depicted for both PA and PEO. Statistically cant differences (P < 0.05) are denoted for changes from baseline (0), between TLMAx and TRIMAx (*), and between PA and PEO (+).
UNTHANK
tery
ET AL.: FLOW
was not
significantly altered from baseline at flow through the stenosed artery had T LMAX.A~TRMAX, decreased and was no longer significantly different than pre-DRP values. Flow through the right iliac artery was increased for both PA (40 + 12%; range 16-73%) and PEO (131 + 39.8%; range 73-249%); however, only the increase during PEO administration was statistically significant. There was no significant change in arterial pressures in either limb at TLMAX.At TRm arterial pressures in both limbs tended to be decreased but the only significant change was a 15 & 5.7% decrease in right femoral artery pressure during the infusion of PEO (Fig. 5). Although blood flow through the stenosed artery was increased at TLMAX(Fig. 4), this occurred without a significant decrease in left hindlimb resistance (Fig. 6). The only changes in hindlimb resistances during DRP infusion occurred on the right side at TRMAX.The right resistances were decreased 32 + 6.5% (range 22-51%) by PA and 60 + 6.7% (range 48-80%) by PEO. The decrease in right hindlimb resistance with PEO was statistically greater than with PA. DISCUSSION
The most important finding of this study was that during a slow, constant intravenous infusion of DRP, blood flow through an iliac artery with a clinically significant stenosis was significantly elevated before blood flow through the contralateral vessel was increased (Figs. 1 and 4 and Table 2). The hemodynamic data reported in Table 1 verify the existence of a clinically significant arterial stenosis in our experimental model. At the time point when blood flow through the stenosed vessel was maximum, arterial pressure distal to the stenosis was not altered (Fig. 5) and hindlimb vascular resistance tended to be decreased (Fig. 6) but the changes were not statistically significant. Because no dose-response data have been reported in the biomedical studies that have used DRP, we had no indication of the threshold concentration of DRP that might produce changes in blood flow through arterial stenoses. For this reason, we made slow, continuous infusions of both PA and PEO while monitoring arterial blood flows and pressures. The changes in hemodynamics that occurred during the polymer infusion were assumed to be concentration dependent. Coleman et al. [ 61 has reported that for PA in rats, a gradual decrease in the hemodynamic effects occurs over a period of l-2 hr postinjection but little if any change takes place during the first 15-30 min. Since our maximum increases in blood flow through a stenosed vessel occurred in less than 2.5 min after initiation of DRP administration (Fig. 4) and because the increase in blood flow on the stenosed side (where greater turbulence should have existed) occurred significantly earlier than maximum effects on the right side (Figs. 1 and 5 and Table 2), we believe the difference observed between the left and
IMPROVEMENT
BY DRP
629
right sides was due to a concentration-dependent, rather than temporal, effect. If the changes in hemodynamics that occur during polymer infusion are concentration dependent, these data would suggest that the maximum effect on blood flow through a critical arterial stenosis in a canine iliac artery would occur at a concentration of approximately 0.3 mg/kg for PEO and 1.2 mg/kg for PA (calculated from TLMAXvalues in Table 2 and infusion rates). These conct,,trations are lower than those used in earlier studies which produced alterations in central hemodynamics [3, 71 or reduced flow disturbances in subcritical arterial stenoses [6, 81. Although previous studies have demonstrated that flow disturbances produced by subcritical stenoses in dog aorta [3] and carotid artery [8] are reduced by DRP without affecting aortic blood flow [8] and that DRP may have antiatherogenic potential [4,5], this is the first study to demonstrate that DRP might be capable of improving blood flow through clinically significant arterial stenoses without altering blood flow through other vessels. Even though the results of this and earlier studies suggest that the class of compounds known as drag reducing polymers may have therapeutic potential in cardiovascular diseases, we have several major concerns regarding the use of DRP and, clearly, there is a need for much additional information regarding mechanisms of action and pharmacodynamics. One concern relates to the variability of the responses observed with administration of DRP. Coleman et al. reported [6] that in some cases in large animals the effects of DRP on hemodynamics were negligible but in other cases “striking.” Although the reports of the effects of DRP on flow disturbances in dogs [3, 81 do not make mention of variability, considerable differences in the percentage of increase in iliac artery blood flow between individual dogs were observed in our studies. During the infusion of DRP, we observed the maximum increase in blood flow through the stenosed artery to vary from 8 to 104%. Furthermore, although increases in blood flow occurred early, in some cases blood flows had dropped to a fraction of baseline levels before DRP administration was stopped. Polimeni has reported that concentrations of DRP not much higher than those used to produce maximum effects can be lethal [6, 111. The exact mechanism that mediates the effect of DRP in uiuo has not been established. Several theories have been proposed to explain drag reduction in pipes [ 1, 10, 131. Both in pipes and in the cardiovascular system the ability of DRP to alter fluid dynamics is thought to be dependent on turbulent or disturbed flow [ 1, 61. For DRP to decrease resistance in uiuo, either blood viscosity or vascular tone must be decreased. The initial report of the effect of DRP on blood flow through glass tubing [3] would suggest that the apparent viscosity of blood is reduced when turbulence is present. However, we have been unable to confirm this original finding with an in vitro perfusion loop. Furthermore, Greene et al. found
630
JOURNAL
OF SURGICAL
RESEARCH:
that with in vitro perfusion loops either drag reduction or drag enhancement could occur depending upon the geometry of the stenosis [12]. Measurements made in viscometers have not found blood viscosity to be decreased by the addition of DRP [ 21. With these considerations in mind, it is not apparent why we observed blood flow through the arterial stenosis to increase and then decrease as blood flow through the normal vessel continued to increase (Fig. 1) or why a relatively large drop in arterial pressure occurs early in the period when DRP is infused (Fig. 2). Such results may be more consistent with peripheral vasodilation than a decrease in apparent blood viscosity under conditions of turbulent blood flow. If DRP do indeed act to decrease resistance to disturbed blood flow by some mechanism other than peripheral vasodilation, much additional information will be needed regarding dose dependency, mechanisms of action, pharmacodynamics, and toxicology before the clinical potential of DRP can be determined. However, the handful of studies that have examined the cardiovascular effects of DRP indicate this class of compounds can have dramatic effects on increasing cardiac output in cardiac-depressed animals, improving blood flow through arterial stenoses, and inhibiting atherogenesis.
VOL.
theories
structure.
Prog. Polymer Sci.
implications of drag reducing agents. Biorheology 7: 221, 1971. 3. Mostardi, R. A., Greene, H. L., Nokes, R. F., Thomas, L. C., and Lue, T. The effect of drag reducing agents on stenotic flow disturbances in dogs. Biorheology 13: 137, 1976. 4. Greene, H. L., Mostardi, R. F., and Nokes, R. F. Effects of drag reducing polymers on initiation of atherosclerosis. Polymer Eng. Sci. 20: 499, 1980. 5. Faruqui, F. I., Otten, M. D., and Polimeni, P. I. Protection against atherogenesis with the polymer drag-reducing agent Separan AP-30. Circulation 75: 627, 1987. 6. Coleman, P. B., Ottenbreit, B. T., and Polimeni, P. I. Effects of a drag-reducing polyelectrolyte of microscopic linear dimension (Separan AP-273) on rat hemodynamics. Circ. Res. 61: 787,
1987. 7. Polimeni,
8.
9.
11. 12.
REFERENCES Morgan, S. E., and McCormick, C. L. Water-soluble copolymers XxX11: Macromolecular drag reduction. A review of predictive
and the effects of polymer
2. Greene, H. L., Nokes, R. F., and Thomas, L. C. Biomedical
LO.
1.
1992
15:507,1990.
ACKNOWLEDGMENTS This work was supported in part by a Department of Veterans Affairs Merit Review Grant (S.G.L.) and a grant from the Indiana University Research Investment Fund (J.L.U. and S.G.L.).
53, NO. 6, DECEMBER
13.
P. I., and Ottenbreit, B. T. Hemodynamic effects of a Poly(ethylene oxide) drag reducing polymer, polyox WSR N-60K, in the open-chest rat. J. Cardiouasc. Pharmacol. 14: 374, 1989. Hutchison, K. J., Campbell, J. D., and Karpinski, E. Decreased poststenotic flow disturbance during drag reduction by polyacrylamide infusion without increased aortic blood flow. Microuasc. Res. 38: 102, 1989. Sumpio, B. E., Upchurch, G. R., and Johnson, G. The influence of perfusate viscosity, RBC deformability and drag of function of an isolated perfused rat kidney. J. Surg. Res. 46: 4, 1989. Sellin, J., Hoyt, J. W., and Scrivener, 0. The effect of drag-reducing additives on fluid flows and their industrial applications. J. Hydraul. Res. 20: 29, 1980. Polimeni, P. I., Ottenbreit, B., and Coleman, P. Enhancement of aortic blood flow with a linear anionic macro-polymer of extraordinary molecular length. J. Mol. Cell. Cardiol. 17: 721, 1985. Greene, H. L., Nokes, R. F., and Thomas, L. C. Drag reduction phenomena in pulsed blood flow. In R. B. Dowdell (Ed.), Flow, Its Measurements and Control in Science and Industry. Ann Arbor, MI: Book on Demand, 1974. Pp. 1459-1464. Berman, N. S. Drag reduction by polymers. Ann. Reu. Fluid Mech. 10: 47, 1978.
