The effects of adrenergic agonists, often used as local anesthetic additives or spinal analgesics, on spinal vessels have not been firmly established. The authors investigated the effects of alpha2- and alpha1-adrenergic agonists on spinal and cerebral pial vessels in vivo.
Pentobarbital-anesthetized dogs (n = 28) were prepared for measurement of spinal pial-vessel diameter in a spinal-window preparation. The authors applied dexmedetomidine, clonidine, phenylephrine, or epinephrine in three different concentrations (0.5, 5.0, and 50 microg/ml; [2.1, 1.9, 2.5, and 2.3] x [10(-6), 10(-5), and 10(-4)] M, respectively) under the window (one drug in each dog) and measured spinal pial arteriolar and venular diameters in a sequential manner. To enable the comparison of their effects on cerebral vessels, the authors also administered these drugs under a cranial window.
On topical administration, each drug constricted spinal pial arterioles in a concentration-dependent manner. Phenylephrine and epinephrine induced a significantly larger arteriolar constriction than dexmedetomidine or clonidine at 5 microg/ml (8%, 11%, 0%, and 1%, respectively). Spinal pial venules tended to be less constricted than arterioles. In cerebral arterioles, greater constrictions were induced by dexmedetomidine and clonidine than those induced by phenylephrine and epinephrine (14%, 8%, 0%, and 1%, respectively). Cerebral pial venules tended to exhibit larger constrictions than cerebral arterioles (unlike in spinal vessels).
Dexmedetomidine and clonidine constricted spinal vessels in a concentration-dependent manner, but such vasoconstrictions were smaller than those induced by phenylephrine and epinephrine.
FOR many years vasoconstrictors have been used to prolong the duration of action of various local anesthetics following subarachnoid administration. However, the effects of adrenergic agonists on spinal cord blood flow (SCBF) and spinal vessels have not yet been clearly defined. Kozody et al.  reported that subarachnoid epinephrine (0.2 mg) or phenylephrine (5 mg) did not decrease SCBF in dogs, whereas Dohi et al. also in the dogs, demonstrated that subarachnoid phenylephrine (2, 3, and 5 mg) decreased SCBF,  although subarachnoid epinephrine (0.1, 0.3, and 0.5 mg) did not significantly change SCBF.  Crosby et al.  reported that subarachnoid clonidine reduced both SCBF and glucose utilization in conscious rats; Gordh et al.  found that epidural clonidine (3 [micro sign]g/kg) did not affect regional SCBF in pigs. As far as dexmedetomidine is concerned, there have been several reports about cerebral blood flow [6–8] and vessels,  but little attention has been paid to its effects on SCBF and spinal vessels. However, our previous study indicated that vasoconstrictors affect both spinal and cerebral pial vessels. 
A better assessment of the direct spinal microvascular actions of adrenergic agonists might enable a better evaluation of their contribution to the prolongation of spinal anesthesia and their safety (in terms of their effects on the spinal microcirculation). There has been little attempt to evaluate the comparative effects of adrenergic agonists on the spinal and cerebral vascular beds in in vivo experiments. For this reason, we decided to investigate the concentration-related effects of dexmedetomidine, clonidine, phenylephrine, and epinephrine on spinal and cerebral pial vascular diameter using spinal and cranial window techniques.
Materials and Methods
The experimental protocols were approved by our Institutional Committee for Animal Care, and the experiments were performed in 52 anesthetized dogs weighing between 6 and 10 kg. Anesthesia was induced with pentobarbital sodium (20 mg/kg intravenously) and maintained with a continuous infusion of pentobarbital sodium (2 mg [middle dot] kg-1[middle dot] min-1). After tracheal intubation, each dog was mechanically ventilated with oxygen-enriched room air. The tidal volume and respiratory rate were adjusted to maintain an end-tidal carbon dioxide pressure of 35–40 mmHg. Polyvinyl chloride catheters were placed in the femoral vein for administration of drugs and fluids, and in the femoral artery for blood pressure monitoring and blood sampling. Rectal temperature was maintained between 36.5 [degree sign]C and 37.5 [degree sign]C with a warming blanket.
In the first set of experiments (n = 28), a closed spinal window was used for observation of the spinal pial microcirculation. The animal was placed in the sphinx position with the head immobilized in a stereotactic frame. After the skin was retracted following a longitudinal midline incision, the thoracolumbar paraspinal muscles were exposed from the lower thoracic level to the third lumbar vertebra. The periosteum and muscle attachments from the 12th thoracic to the 2nd lumbar vertebrae were separated from the laminae and spinous and articular processes with the aid of electrocautery. The block of paraspinal muscles in that region was removed. After the spinous process had been removed with a rongeur, a laminectomy was performed (5 x 10 mm) with an electric drill in the first lumbar vertebra. The surfaces of the laminae were ground flat using an electric grinder, and the dura and arachnoid membrane were cut carefully. A ring fitted with a cover glass was placed over the hole and secured with dental acrylic. The ends of four polyvinyl chloride catheters were inserted through the ring. The space under the window was filled with artificial cerebrospinal fluid (aCSF), the composition of which has been described elsewhere. [11,12] All solutions were bubbled with 5% CO2in air at 37.0 [degree sign]C. One catheter was attached to a reservoir bottle containing aCSF so as to maintain a constant intrawindow pressure of 5 mmHg. Two other catheters were used for infusion and drainage of aCSF and experimental drug solutions, and the final catheter was for continuous monitoring of intrawindow pressure. The volume below the window was between 0.5 and 1.0 ml.
In the second set of experiments (n = 24), a closed cranial window was used to observe the pial microcirculation. The scalp was retracted, the temporal muscle removed, and a hole 2 cm in diameter was made in the parietal bone. After coagulation of dural vessels with the aid of a bipolar electrocoagulator, the dura and arachnoid membrane were cut and retracted over the bone. A ring fitted with a cover glass was placed over the hole and secured with dental acrylic. The cranial window system used was similar to the spinal one described for the first set of experiments.
All in vivo experiments were carried out in the following manner. Dexmedetomidine, clonidine, phenylephrine, and epinephrine were freshly dissolved in aCSF, three different concentrations (0.5 [micro sign]g/ml, 5 [micro sign]g/ml, and 50 [micro sign]g/ml; 2.1 x [10-6, 10-5, and 10-4] M for dexmedetomidine, 1.9 x [10-6, 10-5, and 10-4] M for clonidine, 2.5 x [10-6, 10-5, and 10-4] M for phenylephrine, and 2.3 x [10-6, 10-5, and 10-4] M for epinephrine) being prepared for the spinal experiments, and two (0.5 [micro sign]g/ml and 5 [micro sign]g/ml) for the cranial. In practice, the highest concentration (50 [micro sign]g/ml) of phenylephrine and epinephrine was not used in the spinal experiments (see Results). The animals were allowed to recover from the surgical procedures for at least 30 min. Pial arteriolar and venular diameters, mean arterial pressure, heart rate, arterial blood-gas tensions, pH, blood sugar, and serum electrolytes were measured before and after topical application of each concentration of the drugs into the spinal window (n = 7 for each concentration of each agent) or cranial window (n = 6 for each concentration of each agent). To establish the baseline size of the various vessels, the window was continuously flushed with aCSF at the rate of 0.5 to 1.0 ml/min for 20 min after each measurement. Twenty minutes after the last administration of the study solutions, the spinal pial vascular diameter had returned to control values (except after the higher concentration [5 [micro sign]g/ml] of phenylephrine and epinephrine [see Results]).
The diameters of two pial arterioles and venules were measured sequentially after the administration of each solution of the drugs, the measurements being made using a videomicrometer (Olympus Flovel model, VM-20, Tokyo, Japan) attached to a microscope (Olympus model SZH-10). The data from each view were stored on videotape for later playback and analysis.
All variables used to assess the concentration-dependent effects of experimental drugs were tested by a one-way analysis of variance for repeated measurements, with Scheffe's test being used for post hoc comparisons. Differences among drugs at the same concentration were tested by a one-way analysis of variance followed by Scheffe's test. The difference between spinal and cerebral values was tested by a two-way analysis of variance, followed by an unpaired t test for comparing within the same drug and concentration. Significance was set at P < 0.05. All results are expressed as mean +/- SD.
Neither mean arterial pressure nor heart rate changed significantly following the topical administration of dexmedetomidine, clonidine, phenylephrine, or epinephrine at the two or three different concentrations used in experiments involving spinal or cranial preparations (Table 1 and Table 2). Moreover, arterial blood- gas tensions and pH, serum electrolytes, and blood sugar were all unchanged by any concentration of these drugs in either set of experiments (Table 3 and Table 4). The baseline diameters were 86 +/- 30 [micro sign]m for spinal arterioles and 122 +/- 66 [micro sign]m for spinal venules, and the corresponding values for cerebral vessels were 93 +/- 40 [micro sign]m and 107 +/- 39 [micro sign]m, respectively.
A concentration-dependent decrease in the diameter of spinal pial arterioles was observed after the topical administration of each of these drugs (Figure 1A). At 5 [micro sign]g/ml, phenylephrine and epinephrine induced significantly greater arteriolar constriction (8.1 +/- 5.4% and 10.6 +/- 7.8%) than dexmedetomidine and clonidine (0.0 +/- 0.0% and 0.8 +/- 2.2%). Dexmedetomidine and clonidine at 50 [micro sign]g/ml caused arteriolar constrictions (6.1 +/- 3.1% and 9.2 +/- 4.4%) that were similar in size to those induced by 5 [micro sign]g/ml of phenylephrine and epinephrine. Because the vessels had not recovered from the vasoconstrictions induced by 5 [micro sign]g/ml phenylephrine and epinephrine 2 h after drug administration, the 50-[micro sign]g/ml doses of these two agents were not tested. In spinal pial venules, the pattern of change was similar to that seen in spinal arterioles (Figure 1B).
A concentration-dependent decrease in diameter was also observed in cerebral arterioles after topical administration of the same drugs (Figure 2). In cerebral arterioles, the vasoconstrictions induced by 5 [micro sign]g/ml dexmedetomidine and clonidine (14.0 +/- 9.5% and 8.0 +/- 11.0%) were much larger than those induced by the same dose of phenylephrine and epinephrine (0.3 +/- 3.2% and 1.5 +/- 4.1%)(in contrast to the situation in spinal arterioles at 5 [micro sign]g/ml). Cerebral pial venules exhibited qualitatively similar but mostly larger changes than cerebral arterioles (in contrast to the situation in spinal venules).
The major findings in the present study were that, on topical application, dexmedetomidine, clonidine, phenylephrine, and epinephrine all constricted spinal pial arterioles and venules in a concentration-dependent manner, and that at the same dose phenylephrine and epinephrine induced significantly larger constrictions of spinal arterioles and venules than dexmedetomidine and clonidine. In contrast to the situation in spinal arterioles, dexmedetomidine and clonidine constricted cerebral arterioles to a more significant degree than phenylephrine and epinephrine. Cerebral venules were also constricted by these adrenergic agonists, and they were more sensitive than their spinal counterpart. Thus, as far as their effects on the spinal microcirculation are concerned, [Greek small letter alpha]2-adrenergicagonists, such as dexmedetomidine and clonidine, may have a wider margin of safety than phenylephrine and epinephrine (see subsequent sections).
In general, the prolongation of local anesthesia effect induced by adrenergic agonists might be expected to result from their vasoconstrictor action or an effect on nociception resulting from [Greek small letter alpha](2-adrenoceptor) stimulation. Thus, a proper understanding of the comparative effect of adrenergic agonists on spinal vessels would be expected to help clarify the interaction between local anesthetics and adrenergic agonists. However, the direct effects of adrenergic agonists on SCBF and spinal cord blood vessels have not yet been clearly defined. Previous studies found that subarachnoid epinephrine did not affect SCBF, [1,3], or that it attenuated the increase in SCBF induced by concomitantly administered subarachnoid tetracaine or lidocaine in dogs, [13,14] whereas subarachnoid phenylephrine in dogs either reduced SCBF dose-dependently  or did not affect SCBF at all.  Moreover, although Crosby et al.  reported that subarachnoid clonidine reduced SCBF in conscious rats, Mensink et al.  found that it increased regional SCBF in the dog. In the current study, although dexmedetomidine, clonidine, phenylephrine, and epinephrine all constricted spinal pial arterioles, the latter two agents induced significantly larger vasoconstrictions of spinal arterioles than dexmedetomidine and clonidine. However, previous reports have demonstrated that changes in superficial pial arterioles may not reflect changes in total SCBF. [15,16] The large contribution made by intraparenchymal arterioles to overall vascular resistance in the cerebral circulation suggests that spinal intraparenchymal arterioles may also play a greater role than their pial counterparts in the regulation of SCBF. Porter et al.  demonstrated that subarachnoid epinephrine did not induce demonstrable changes in SCBF during spinal anesthesia with lidocaine, mepivacaine, or tetracaine, although the absorption of such local anesthetics was reduced by added epinephrine. They suggested that such reduced vascular absorption of local anesthetic may result from vasoconstriction of the most superficial vessels in the spinal cord. On the basis of the published data discussed previously and the current results, it seems possible that the changes in pial-vessel diameter seen in the current study might not themselves lead to a critical decrease in total SCBF, even though regional SCBF may in fact be affected by such topically administered adrenergic agonists.
Previous studies have demonstrated that different combinations of local anesthetics and adrenergic agonists have different effect. The addition of phenylephrine or epinephrine seems to prolong hyperbaric tetracaine spinal anesthesia [18,19] but has a less obvious effect on the spinal anesthesia produced by lidocaine or bupivacaine. [20,21] It was postulated that such differences might result from differences in the vasodilator actions of the various local anesthetic drugs. Other studies have demonstrated that clonidine could be useful for prolonging the duration of action of local anesthetics whether given by the subarachnoid [22,23] or oral [24–26] route in similar doses. In the current study, [Greek small letter alpha]2-agonistsconstricted spinal pial arterioles less than epinephrine and phenylephrine. On this basis, the contribution of vasoconstriction to the prolongation of spinal anesthesia could be less with [Greek small letter alpha]2-agoniststhan with phenylephrine and epinephrine. Therefore, it is possible that [Greek small letter alpha]2-agonistsmay prolong the duration of local anesthetic effects during spinal anesthesia by a mechanism that is independent of their vasoactive action on spinal vessels.
As far as we know, there is little information as to differences between the responses of spinal and cerebral vessels to adrenergic agonists. To judge from the results of the current study, spinal arterioles seem to be more sensitive to [Greek small letter alpha]1-adrenergicstimulation than to [Greek small letter alpha]2-adrenergicsimulation, whereas cerebral arterioles are more sensitive to [Greek small letter alpha]2-adrenergicstimulation than to [Greek small letter alpha]1-adrenergicstimulation. In contrast to the situation in arterioles, spinal venules seemed to be less sensitive to [Greek small letter alpha]-adrenergic stimulation than their cerebral counterparts. Such differences between spinal and cerebral vessels in their responses to adrenergic stimulation might conceivably contribute to spinal ischemic damage during adrenergic agonist administration after cerebrospinal reperfusion.
It is well known that vessels of different sizes often respond quite differently to both vasodilators and vasoconstrictors. However, in the spinal-window preparation used in the present study, spinal vessels, especially arterioles, were few. We could not always find vessels of wide range in size. Thus, we chose sizes of vessels 86 +/- 30 [micro sign]m for spinal arterioles and 122 +/- 66 [micro sign]m for spinal venules, which were observed in every spinal window preparation. Moreover, in our pilot study, we could not find any size-dependent vasoreactivity within the range of size observed in the current study. Because we want to compare the spinal vessels with cerebral ones, we fit the size of cerebral vessels to that of spinal vessels. If we could choose larger arterioles or artery, different vasoreactivity may be induced by these agent, as previous reports indicated. [27,28]
Previous studies have indicated that the clinically effective doses of so-called vasopressors as local anesthetic additives are approximately 1–5 mg for phenylephrine, [19,21] 0.1–0.3 mg for epinephrine, [18,20] and 0.075–0.15 mg for clonidine. [22,23]. There is no information about the effective dose of dexmedetomidine as a local anesthetic additive in the clinical setting. To judge from the results of animal studies, [29,30] the required clinical dose of dexmedetomidine, if it were used as a local anesthetic additive, would be similar or smaller than the required dose of clonidine. When used as local anesthetic additives or analgesic agents by subarachnoid administration, alpha2-agonistssuch as dexmedetomidine and clonidine could be safer, in terms of their effects on spinal vessels than epinephrine or phenylephrine.
In conclusion, dexmedetomidine and clonidine constrict spinal pial vessels in a concentration-dependent manner. In this respect, their actions are essentially similar but less powerful than those of epinephrine and phenylephrine. The differences between these drugs' actions may contribute to the clinical variabilities seen when local anesthetics and vasoconstrictors are given together. In addition, [Greek small letter alpha]2-agonistscould be safer than [Greek small letter alpha]1-agonistswhen used as local anesthetic additives or analgesic agents, at least in terms of their effects on the spinal microcirculation.