Impaired Autoregulation in an Experimental Model of Chronic Cerebral Hypoperfusion in Rats
Background and Purpose To verify the hypothesis that impaired autoregulation may contribute to cerebral swelling or hemorrhage after a sudden recovery of perfusion pressure, we studied the chronic effects of cerebral hypoperfusion on the autoregulatory responses of the pial arterioles in situ.
Methods Eight to 12 weeks after a carotid-jugular fistula was created in rats, experiments were performed under α-chloralose and urethane anesthesia. Regional cerebral blood flow (rCBF) was determined by the hydrogen clearance method, and carotid pressure was measured. Using a closed cranial window, we determined the autoregulatory responses of the arterioles (30 to 50 μm) to both hypertension induced by norepinephrine and sudden fistula closure at various mean arterial pressures (MAPs).
Results rCBF on the fistula side was reduced by 27%. Carotid pressure was significantly lower than normal but was immediately increased by fistula closure. The pial arterioles showed marked elongation and enlargement. During induced hypertension, the arterioles in the fistula group started to dilate at an MAP lower than that of the control group (130 versus 180 mm Hg, respectively). The arterioles constricted when the fistula was occluded at normal MAP. However, when the fistula was occluded at an MAP higher than 130 mm Hg, the vessels dilated.
Conclusions It was demonstrated that (1) chronic hypoperfusion induced impairment of the upper limit of autoregulation and (2) sudden fistula closure under hypertensive conditions caused vasodilation of the arterioles. These findings suggest that rapid restoration of perfusion pressure is possibly followed by a pressure breakthrough phenomenon in a chronically hypoperfused cerebrovasculature.
Spetzler et al1 suggested that impairment of the cerebrovascular reactivity to an acute increase in perfusion pressure may lead to capillary breakthrough after ablation of an arteriovenous shunt, which is known as the normal perfusion pressure breakthrough theory. However, a carotid-jugular fistula model in the cat reported by Spetzler et al could not offer strong support for the existence of this phenomenon because they did not directly observe the intracranial cerebrovascular response to changes in perfusion pressure. Although the normal perfusion pressure breakthrough theory has been used to explain unusual brain swelling or hemorrhage associated with the excision of high-flow arteriovenous malformations2 3 4 and carotid endarterectomies,5 it has not been clarified and not all authors are in agreement.6 7 8
We therefore used a carotid-jugular fistula model in rats to ascertain whether chronic hypoperfusion could cause impairment of autoregulation and whether restoration of perfusion pressure by sudden closure of an arteriovenous fistula could induce hyperemia in the pial microcirculation.
Materials and Methods
Creation of Carotid-Jugular Arteriovenous Fistula
This experimental study was performed in accordance with the guidelines of Kitasato University for the care and use of laboratory animals. Male Sprague-Dawley rats (weight, 230 to 280 g), which were housed under diurnal lighting conditions and were allowed free access to food and water, were anesthetized with halothane (1.5%) and a mixture of oxygen and air. The right common carotid artery, the external carotid artery, and the external jugular vein were meticulously exposed under an operative microscope while the cervical sympathetic trunk was preserved. The right external carotid artery was ligated and resected at its origin. The ipsilateral common carotid artery and external jugular vein were ligated and divided above the ligature. According to the method reported by Morgan et al,9 an end-to-end anastomosis of the rostral ends of both vessels was completed, as illustrated in Fig 1⇓. The internal jugular veins are hypoplastic, and the cerebral venous blood drains mainly to the external jugular veins through the transverse sinus and the retroglenoid veins in rats.10 Therefore, it was expected that not only a decrease in arterial pressure on the same side as the arteriovenous fistula but also a slight elevation in the dural sinus pressure would result in this model.
To clarify the influence of possible damage to nerves, particularly the cervical sympathetic nerves, resulting from microsurgical procedures, sham operations were performed under the same conditions. The common carotid artery, the external carotid artery, and the external jugular vein of the right side were dissected in the same manner as in the fistula group but were neither ligated nor cut.
General Preparation for Experiments After Creation of the Fistula
The rats were anesthetized 8 to 12 weeks after creation of the fistula, initially with halothane (1.5%) in a mixture of oxygen and air. We examined the patency of the fistula by exposing the anastomotic site before each experiment. We confirmed a satisfactory degree of fistula patency in approximately 95% of the operated animals, which were then used for subsequent studies.
The femoral artery and vein were cannulated to monitor blood pressure and intravenous administration of drugs, respectively. The rats were subjected to tracheostomy and mechanically ventilated before being immobilized with tubocurarine chloride (1 mg/kg IV). After completion of the above procedures, α-chloralose (50 mg/kg) and urethane (500 to 750 mg/kg) were injected intraperitoneally, and then halothane was withdrawn gradually. Additional doses of α-chloralose (10 mg/kg) and urethane (100 mg/kg) were given as needed. End-tidal CO2 was continuously measured by a CO2 analyzer (Respina IH31, San-ei) and was kept at a constant level throughout each experiment by adjustment of the respiratory rate and volume. Arterial blood pH and gases were measured with a pH/blood gas analyzer (168 pH blood gas system, Corning). Rectal temperature was monitored and maintained at approximately 37°C with a heating mat.
To verify the cerebral hypoperfusion produced in this model, we measured regional cerebral blood flow (rCBF) on the fistula side using the hydrogen clearance method. The head of the rat was fixed in a frame, and a platinum electrode (0.2 mm in diameter) was inserted with the use of a micromanipulator, through a 2-mm-diameter burr hole that was made 1 mm anterior and 3 mm lateral to the bregma, into the cerebral tissue 4 mm deep to the dura mater. The tip of the electrode would therefore have been lying within the right caudoputamen11 ; the position of the electrode was verified in randomly selected animals after the animals were killed. A sintered Ag/AgCl reference electrode was implanted subcutaneously in the back. The electrode system was allowed to stabilize in the tissues for approximately one half hour before a stable baseline could be made. Ten percent hydrogen gas was fed directly into the endotracheal tube for 2 minutes. rCBF was calculated by initial slope analysis.12 13 In the same animal, rCBF changes associated with fistula closure were continuously monitored with a laser-Doppler flowmeter (ALF21N, Advance, Inc) over the right parietal cortex through a paper-thin layer of bone.
To estimate the effects of the carotid-jugular fistula on cerebral perfusion pressure, we measured the carotid arterial pressure on the fistula side, which was thought to reflect the pressure of the right internal carotid artery in this model. The rat was placed in a supine position, and the right common carotid artery was punctured with a 27-gauge needle. The needle was held in a micromanipulator and connected to a pressure transducer.
To raise the systemic arterial pressure, norepinephrine dissolved in physiological saline (5 μg/mL) was continuously injected into the femoral vein with an infusion pump. The mean arterial pressure (MAP) was gradually raised in steps of approximately 10 mm Hg by changing the infusion rate. The infusion rate of norepinephrine ranged from 0.4 to 2.4 μg/min (5 to 30 mL/h of the solution).
Closed Cranial Window
A closed cranial window was made over the right parietal cortex, and the pial microvessels were observed in situ with a microscope (Nikon) fitted with a video camera system (series CCD C2400 system, Hamamatsu Photonics). The method for a closed cranial window preparation was previously described in detail.14 In summary, the head of the anesthetized rat was fixed in a frame, and the skull was exposed by a longitudinal skin incision. Dental acrylic was applied to the skull in a doughnut shape, and three polyethylene tubes were embedded in the acrylic for use for intracranial pressure monitoring and artificial cerebrospinal fluid superfusion. Craniectomy (5×4 mm) was performed over the right parietal cortex (1 to 6 mm posterior and 1 to 5 mm lateral to the bregma) in the center of the dental acrylic, and the dura-arachnoid complex was incised while the pial surface was superfused with artificial cerebrospinal fluid. A cover glass was placed on the ring of dental acrylic and immediately sealed with cyanoacrylate. The diameter of the pial arterioles was continuously measured with a video dimension analyzer (model C1170, Hamamatsu Photonics).
The right rCBF was measured by the hydrogen clearance method in both rats with an open fistula (n=6) and normal rats (n=7) during steady state conditions. After baseline rCBF was determined by hydrogen clearance, in the same fistula group (n=6), changes in blood flow in the right parietal cortex were continuously monitored by laser-Doppler flowmetry during, before, and after sudden fistula closure at normal blood pressure and then at an MAP of approximately 130 mm Hg after elevation by norepinephrine infusion in the same animals.
Right carotid arterial pressure was measured before and after acute fistula closure at normal systemic arterial pressure. After fistula closure, hypertension was induced while carotid pressure was continuously monitored (n=5). Carotid pressure was also monitored during induced hypertension in the rats with an open fistula (n=4).
Observations of Morphological Changes in Pial Microvessels
The pial microvessels were observed through a closed cranial window and recorded on videotape. Morphological findings in the fistula group (n=8) were compared with those in the sham-operation rats and normal rats (n=8). The degree of dilation of the pial arterioles was determined by measuring the diameter with an image analyzer (Argus-10 image processor, Hamamatsu Photonics) at the first-order (1A), second-order (2A), and third-order (3A) branches of the middle cerebral artery, as defined by Harper and Bohlen.15 The arterial trees were seen through the cranial window, which was made in the same site in all animals.
Pial Arteriolar Response to Hypocapnia and Hypercapnia
When a stable baseline diameter was obtained after the closure of the fistula, changes in diameter during hyperventilation and 5% CO2 inhalation, both for 5 minutes, were recorded in six animals. CO2 reactivity was expressed as follows: ΔDiameter (%)/ΔPaco2 (mm Hg).
Pial Arteriolar Responses to Induced Hypertension
A stable baseline diameter was obtained after closure of the fistula. Hypertension was then induced as previously described. The diameter was continuously monitored and determined at each level of MAP sustained for at least 2 minutes. Changes in the diameter in the fistula group (n=7) were compared with those in the normal group (n=6) and the sham-operation group (n=4).
Pial Arteriolar Responses to Acute Fistula Closure
The pial arteriolar diameter was continuously recorded during acute fistula closure in 19 rats. Responses to fistula closure were measured under normal and also elevated systemic arterial pressure induced by the infusion of norepinephrine. Fistula closures was repeated one to four times at various systemic arterial pressures in a single rat with adequate intervals, during which time the vessel diameter returned to baseline.
Values are expressed as mean±SEM. The significant dilation from the control diameter during induced hypertension was determined by the Friedman test. Statistical differences among groups were examined by two-way repeated measures ANOVA. Polynomial regression analysis was used to examine the correlation between the pial arteriolar response to acute fistula closure and MAP. Values of P<.05 were considered significant.
rCBF on the Fistula Side
The rCBF values on the fistula side, as determined by the hydrogen clearance method, are indicated with physiological parameters in Table 1⇓. The rCBF with an open fistula was significantly lower than that in normal rats (P<.01). Changes monitored by laser-Doppler flowmetry in rCBF associated with sudden fistula closure showed a minimal increase (10±8%, P>.05) when the fistula was occluded at normal MAP (103±1 mm Hg). In contrast, a significant rCBF increase (40±14%, P<.05 determined by paired t test) was observed on fistula closure at an MAP of 130±1 mm Hg (Fig 2⇓).
Carotid Pressure on the Fistula Side
Carotid pressures with and without fistula closure during induced hypertension are shown in Fig 3⇓. The carotid pressure on the fistula side was very low (6±1 mm Hg) but immediately increased to 80±5 mm Hg after fistula closure at normal MAP (108±4 mm Hg). The degree of carotid pressure increase produced by acute fistula closure was estimated to be 65% of MAP. The pressure in the closed fistula group remained elevated as the systemic MAP was raised by norepinephrine. In the open fistula group, carotid pressure was not affected by MAP changes and remained at a low level.
Morphological Changes of Pial Microvessels
The pial arterioles showed considerable dilation and tortuosity, unlike the control sham-operation group (Fig 4⇓). The cortical veins were slightly dilated. These findings were observed in all animals with fistulas. The characteristic changes of the arterioles were elongation and enlargement associated with marked increases in the luminal diameter. The diameter of the pial arterioles was significantly larger in the fistula group than in the control group (for 1A, 2A, and 3A: fistula group, 70±3, 48±2, and 27±1 μm, respectively, versus control group, 55±2, 38±1, and 22±1 μm, respectively; P<.01 determined by unpaired t test).
Pial Arteriolar Response to Hypocapnia and Hypercapnia
CO2 reactivity measurements in the fistula group were −1.1±0.1%/mm Hg and 1.3±0.2%/mm Hg for hypocapnia and hypercapnia, respectively. The reactivity to changes in CO2 was preserved.14 Physiological parameters during hypocapnia and hypercapnia are shown in Table 2⇓.
Pial Arteriolar Response to Induced Hypertension
Fistula closure under normal MAP induced vasoconstriction of approximately 17±4% in the pial arterioles. Therefore, the baseline diameter obtained after closure of the fistula, which is shown in Fig 5⇓, was likely to be comparable to the mean diameter of arterioles of the control animals. MAPs before hypertension was induced were 102±10, 102±12 and 101±9 mm Hg for the fistula, normal, and sham-operation groups, respectively. The arterioles in the normal and the sham group did not show any significant changes until the MAP was elevated to 180 mm Hg, and they started to dilate rapidly at this pressure (Fig 5⇓). In contrast, in the fistula group the vessels began to dilate even when hypertension was slight (≈130 mm Hg) and continued to dilate as MAP increased (Fig 5⇓). The shift of the upper limit of autoregulation was demonstrated in the fistula group.
Pial Arteriolar Response to Sudden Fistula Closure
Baseline MAP was 104±3 mm Hg, and sudden fistula closure did not affect the MAP significantly. The arterioles constricted when the fistula was occluded when systemic arterial pressure was normal. However, these constrictor responses were replaced by dilation when the fistula was occluded at MAPs higher than 130 mm Hg (Fig 6⇓). Sudden fistula closure caused dilation in hypertension. A significant correlation between the response of the pial arterioles to sudden fistula closure and systemic arterial pressure was demonstrated (Fig 6⇓).
The major findings of this study are as follows: (1) chronic cerebral hypoperfusion achieved by carotid-jugular fistula induced the impairment of autoregulation at the upper limit; and (2) in hypertension, sudden fistula closure, which is associated with a rapid increase in perfusion pressure, caused vasodilation and hyperemia in the cortex.
To produce prolonged cerebral hypoperfusion, we used a model of carotid-jugular fistula in the rat described by Morgan et al9 16 for the following reasons: First, Morgan et al assessed cerebral blood flow quantitatively using [14C]iodoantipyrine and found approximately 30% reduction in rCBF in this model. Our study confirmed this result. In addition, we noted a marked reduction in the ipsilateral carotid pressure. Cerebral hypoperfusion maintained above the critical flow level has therefore been produced in this carotid-jugular fistula model. Second, we found that the carotid pressure on the fistula side increased immediately when the fistula was occluded, with the degree of this rapid change in perfusion pressure dependent on systemic arterial pressure. Accordingly, it was considered possible to produce chronic cerebral hypoperfusion and to rapidly restore perfusion pressure by occluding the fistula in this model.
Eight to 12 weeks after the creation of the fistula, the pial arterioles showed marked dilation and tortuosity. A few similar reports showing dilation and tortuosity in vessels after the reduction of cerebral perfusion pressure have already been published. Coyle17 observed increases in diameter and in length (by 24% to 29%), which led to greater tortuosity in collaterals 20 days after middle cerebral artery occlusion in rats. Lehman et al18 studied structural changes in the enlarged collateral artery and demonstrated an increased number of vascular smooth muscle cells in the dilated basilar artery after bilateral ligation of the carotid arteries in rabbits. Although the mechanism of those morphological changes is unknown, it is considered that they are clearly consequences of vascular behavior in response to chronic reduction of perfusion pressure or flow.
When systemic arterial hypertension was induced gradually, the pial arterioles in sham-operation and normal animals did not show any significant changes in the diameter in the MAP range of 100 to 170 mm Hg but dilated significantly at an MAP of approximately 180 mm Hg. Kontos et al19 demonstrated that the changes in arterial caliber in response to systemic arterial pressure changes were size dependent. Smaller pial arterioles did not constrict at MAPs of between 90 and 160 mm Hg in cats, which corresponds with our data from rats. In contrast, the pial arterioles in the fistula group started to dilate early, at an MAP of 130 mm Hg, and continued to dilate as the MAP increased, whereas they constricted in hypocapnic conditions. Therefore, this dilation in response to induced hypertension was considered to be specific for the impairment of autoregulation at the upper limit but was not due to a nonspecific decline of cerebrovascular reactivity.
Normal arterioles (resistance vessels) are supposed to constrict on sudden rises in perfusion pressure. In the present study, however, the pial arterioles dilated after fistula closure when the MAP was higher than 130 mm Hg but constricted when the systemic arterial pressure was normal. In addition, cortical hyperemia was observed by laser-Doppler flowmetry after fistula closure at an MAP of 130 mm Hg. These results indicate that the arterioles were able to constrict only in response to a moderate increase in perfusion pressure, but not to a larger increase, because the upper limit of autoregulation had been lowered. Changes in perfusion pressure induced by fistula closure at an MAP of 130 mm Hg seemed to be the threshold for pressure breakthrough in this model, in which hypoperfused cerebral vasculature was achieved for an extended period. This finding is consistent with results previously reported by others using the same model9 or a cat model,20 which found that carotid-jugular fistula occlusion in a state of induced hypertension 6 or 8 weeks after its creation resulted in disruption of the blood-brain barrier.
The normal perfusion pressure breakthrough theory1 is an important concept that may explain unexpected cerebral edema formation or hemorrhage after therapeutic procedures that improve chronic cerebral hypoperfusion. Our experimental results support the theory that the microcirculatory pressure breakthrough phenomenon can occur in chronically hypoperfused cerebrovascular beds when restored perfusion pressure exceeds the threshold, which is dependent on the efficiency of the autoregulatory response. The possibility is likely to be determined by a combination of the degree of impairment of autoregulation and the extent of sharp changes in perfusion pressure. Therefore, all aspects of these two major factors should be carefully assessed to allow prediction of the possibility of the breakthrough phenomenon. A systematic method of predicting this phenomenon in quantitative terms remains to be established.
The authors thank Shigeyoshi Maruyama for technical assistance and C.W.P. Reynolds for linguistic assistance.
Reprint requests to Katsumi Irikura, MD, Department of Neurosurgery, Kitasato University School of Medicine, 1-15-1 Kitasato, Sagamihara, Kanagawa 228, Japan.
Preliminary results were presented at the Microcirculatory Stasis in the Brain Conference, Tokyo, Japan, May 20, 1993.
- Received January 29, 1996.
- Revision received April 1, 1996.
- Accepted April 4, 1996.
- Copyright © 1996 by American Heart Association
Pellegrino LJ, Pellegrino AS, Cushman AJ. A Stereotaxic Atlas of the Rat Brain. 2nd ed. New York, NY: Plenum Press; 1979.
Olesen J, Paulson OB, Lassen NA. Regional cerebral blood flow in man determined by the initial slope of the clearance of intra-arterially injected 133Xe. Stroke. 1971;2:519-540.
Symon L, Pasztor E, Dorsch NWC, Branston NM. Physiological responses of local areas of the cerebral circulation in experimental primates determined by the method of hydrogen clearance. Stroke. 1973;4:632-642.
Harper SL, Bohlen HG. Microvascular adaptation in the cerebral cortex of adult spontaneously hypertensive rats. Hypertension. 1984;6:408-419.
Lehman RM, Owens GK, Kassel NF, Hongo K. Mechanism of enlargement of major cerebral collateral arteries in rabbits. Stroke. 1991;22:499-504.
Kontos HA, Wei EP, Navari RM, Levasseur JE Jr, Rosenblum WI, Patterson JL. Responses of cerebral arteries and arterioles to acute hypotension and hypertension. Am J Physiol. 1978;234:H371-H383.
The authors address the clinically important issue of whether the increase in perfusion pressure accompanying repair of an arteriovenous malformation is capable of causing an autoregulatory breakthrough phenomenon. Using an interesting model of carotid-jugular fistula for 8 to 12 weeks in the rat, they found that blood flow was 27% lower on the side with the fistula. With sudden closure of the fistula, carotid stump pressure increased and pial arterioles constricted if mean arterial pressure was normal. There was little change in blood flow indicative of active autoregulation. However, dilation and large increases in blood flow occurred if mean arterial pressure was greater than 130 mm Hg at the time of the closure. Furthermore, when closure was produced at normal arterial pressure and then norepinephrine was infused to increase arterial pressure, dilation began to occur at much lower arterial pressure than in controls. Therefore, pressure breakthrough occurs at lower autoregulatory pressure with chronic arteriovenous fistula formation. These results emphasize the need for careful blood pressure management during repair of arteriovenous malformations to prevent pressure-induced vascular injury manifested by edema formation.