Abrupt changes in systemic hemodynamics occur during and after abdominal aortic cross-clamping and unclamping performed to facilitate abdominal aortic aneurysmectomy. Various vasoactive substances exert effects on the cerebrovascular microcirculation while also causing systemic hemodynamic changes (1). We previously reported that unclamping after a 20-min abdominal aortic cross-clamp caused biphasic changes in cerebral pial arteriolar diameter: specifically, an initial transient vasodilation and a subsequent sustained vasoconstriction (as assessed using the closed cranial window technique in rabbits) (1). Further, we noted that this vasoconstriction was at least partly induced via a washout of thromboxane (Tx) A2 that is produced in the ischemic region while the clamp is in place (and probably also after cross-clamp release). Aadahl et al. (2), using a laser Doppler technique, also observed in pigs that cerebral blood flow decreased after unclamping of the thoracic aorta. The cerebral vasoconstriction observed in these studies seems to imply that cerebral ischemia may occur as a result of microcirculatory failure after unclamping, and this would be especially critical in the clinical setting for patients with endothelial damage, such as those with atherosclerosis or hypertension. Although prevention or attenuation of vasoconstriction ought to be favorable in such patients, ways of achieving this have not been well studied.
Many drugs used to improve cardiovascular disorders affect blood vessels within the central nervous system (CNS) (3,4). Milrinone and colforsin daropate, new positive-inotropic drugs, also have a vasodilator action, and both have been used in critical conditions, including abdominal aortic aneurysmectomy. In the present study, we hypothesized that both of these drugs would attenuate the sustained constriction of pial vessels seen after unclamping of an abdominal aortic cross-clamp. We therefore examined the effects of clinical doses of each drug on the cerebral microcirculation in a rabbit model of aortic cross-clamping. Also, we evaluated the effects on regional cerebral blood flow (rCBF) and tissue oxygen tension (PBrO2).
The procedures used in the present study conformed to the Guiding Principles in the Care and Use of Animals approved by the Council of the American Physiologic Society, and the experimental protocols were approved by our Institutional Committee for Animal Care. The experiments were performed on 47 anesthetized rabbits weighing 2.0 to 2.2 kg. Each animal was initially anesthetized with pentobarbital sodium (25 mg/kg body weight, IV). Anesthesia was maintained using a continuous infusion of the same drug (5 mg · kg−1 · h−1). Mechanical ventilation was achieved through a tracheotomy tube using oxygen-enriched room air (arterial oxygen content, 14–17 Vol%). The tidal volume and respiratory rate were continually adjusted to maintain end-tidal carbon dioxide tension (Petco2) between 35 and 40 mm Hg; Petco2 was monitored throughout the experiment. Polyvinyl chloride catheters were placed in the femoral vein for administration of fluid and in the right axillary and left femoral arteries for the continuous monitoring of proximal and distal aortic pressures (PrAP and DiAP), heart rate (HR), and blood sampling (from the right axillary artery). Rectal temperature was maintained between 38.5°C and 39.5°C by a heating blanket and warming lamp. A skin incision was made in the lateral abdomen. The abdominal aorta was freed from surrounding tissues, and tapes were passed around the aorta to permit tightening in preparation for clamping just distal to the renal arteries.
In the first experiment, a closed cranial window was used to observe the cerebral pial microcirculation (n = 35). Each animal was placed in the sphinx posture, the scalp was retracted, and a 10-mm diameter hole was made in the parietal bone. The dura and arachnoid membranes were opened carefully, and a polypropylene ring with a glass coverslip, placed over the hole, was secured with dental acrylic. The space under the window was filled with artificial cerebrospinal fluid (aCSF), the composition of which was Na+ 151 mEq/L, K+ 4 mEq/L, Ca2+ 3 mEq/L, Mg2+ 1.3 mEq/L, Cl− 134 mEq/L, HCO3 − 25 mEq/L, urea 40 mg/dL, and glucose 67 mg/dL (pH adjusted to 7.48). This solution was freshly prepared each day and bubbled with 5% CO2 in air at 39.0°C for 15 min just before use. Three polyethylene catheters were inserted through the ring: one was attached to a reservoir bottle containing aCSF to maintain the desired level of intra-window pressure (5 mm Hg), whereas the second was used to monitor intra-window pressure and the third for draining the fluid. The temperature within the window was monitored using a thermometer (Model 6510; Mallinckrodt Medical, St. Louis, MO) and was between 38.5°C and 39.5°C.
The diameters of 2 large (75–120 μm) and 2 small (40–75 μm) pial arterioles were measured in each cranial window using a videomicrometer (Olympus Flovel videomicrometer Model VM-20; Flovel, Tokyo, Japan) on a television monitor attached to a microscope (Model SZH-10; Olympus, Tokyo, Japan). We defined which vessels would be used to collect data before drug administration. This was done to eliminate a bias that might occur if vessels were selected after drug administration. Data from the pial views were stored on videotape for later playback and analysis. The percentage changes recorded for individual arterial segments were averaged for each type (large or small) of vessel in each rabbit, and this average value was used in the statistical analysis.
Rabbits were assigned to one of five groups. All experiments were performed after at least 30 min recovery from the surgical preparation. After baseline measurements had been made, each rabbit was infused IV with one of the following: saline (control group, n = 7), milrinone (M-0.05 group, 0.05 μg · kg−1 · min−1, n = 7; M-0.5 group, 0.5 μg · kg−1 · min−1, n = 7), colforsin daropate (C-0.05 group, 0.05 μg · kg−1 · min−1, n = 7; or C-0.5 group, 0.5 μg · kg−1 · min−1, n = 7) (5,6). All infusions were continued throughout the experiment. At 15 min after the start of the IV infusion, aortic clamping was performed (duration, 20 min). The clamping and unclamping were done gradually (each taking approximately 30 s to perform) to minimize hemodynamic changes. Measurements of cerebral pial arteriolar diameter, hemodynamic variables (PrAP, DiAP, and HR), and various physiologic variables (rectal temperature, intra-window temperature, arterial blood gas tensions, electrolytes, blood glucose, and blood pH) were taken at the following time points: just before the start of IV administration (baseline), after 15 min IV administration (Pre-Clamp), just after aortic clamping (After Clamp), 20 min after clamping (Pre-Unclamp), and at 0, 2, 5, 15, 30, and 60 min after unclamping (the time-point “0 min after unclamping” was actually 30 s after the start of unclamping, which took about 30 s to perform, as above).
In a second experiment, rabbits (n = 12) were surgically prepared similarly to those in the first experiment, except without the closed cranial window. These 12 rabbits were prepared for measurement of rCBF and PBrO2 using laser Doppler flowmetry using a tissue microprobe (Model OXY LAB Po2 and LDF; OXFORD OPTRONIX, England), positioned at a 2-mm depth below the parietal surface. rCBF, PBrO2, hemodynamic values (PrAP, DiAP, and HR), rectal temperature, and intra-cranial temperature were measured at baseline, 15 min after IV infusion of either 0.9% saline (control group, n = 6) or milrinone (milrinone group: 0.5 μg · kg−1 · min−1, n = 6), just after aortic clamping (after clamp), at 1, 2, 5, 10, 15, and 20 min after aortic clamping (pre-unclamp), and at 0, 1, 2, 3, 4, 5, 10, 15, 20, 30, 45, and 60 min after aortic unclamping. Arterial blood gas tensions, electrolytes, blood glucose, and blood pH were measured at the same points as in the first experiment.
All variables used to assess time-dependent effects within groups were tested by one-way analysis of variance for repeated measurements followed by the Scheffé F test for post hoc comparisons. Differences among groups and between large and small arterioles were examined by a two-way analysis of variance and then by a one-way analysis of variance for factorial measurements followed by the Scheffé F test. In the second experiment, all variables used to assess time-dependent effects within groups were tested by one-way analysis of variance for repeated measurements followed by the Scheffé F test for post hoc comparisons, and differences among groups were examined by an unpaired Student’s t-test. Significance was considered to be demonstrated at P < 0.05. All results are expressed as mean ± sd.
In the first experiment, there were no significant differences in baseline hemodynamic or physiological variables among the groups nor did HR vary significantly throughout the experiments in any group. In addition, rectal and intra-window temperatures did not alter at any stage of the experiments in any group. Pao2, Na+, K+, and blood glucose were stable at all stages of the experiments in each group. In every group: a) PrAP was decreased significantly at time point 0 min after unclamping (P < 0.05), b) DiAP was decreased significantly after clamping (P < 0.05) but then recovered after unclamping (Table 1), c) arterial pH was decreased significantly at 0 (P < 0.05) and 2 min after unclamping (P < 0.05), and d) Paco2 was significantly increased at 0 (P < 0.05) and 2 min after unclamping (P < 0.05) (Table 2).
There were no significant differences among the groups in the baseline diameters of the 2 sizes of arterioles (for large or small arterioles, respectively: control, 91 ± 14 μm and 54 ± 8 μm; M-0.05, 92 ± 15 μm and 54 ± 9 μm; M-0.5, 89 ± 15 μm and 55 ± 9 μm; C-0.05, 91 ± 15 μm and 54 ± 8 μm; C-0.5, 90 ± 16 μm and 53 ± 9 μm). In the following paragraphs, all percentage values represent changes in diameter with respect to baseline.
In the control group, neither large nor small pial arterioles showed significant changes in diameter after clamping, but both types of arterioles dilated significantly just after unclamping (maximum increase, 7% and 11%, respectively). They then constricted significantly, starting at 5 min after unclamping (−5% and −6%, respectively). The constrictions were still significant (and, indeed, appeared still to be increasing) at 60 min after unclamping (−15% and −24%, respectively), as in previous studies (1) (Figs. 1 and 2).
In the M-0.05, C-0.05, and C-0.5 groups, baseline pial arteriolar diameters (large and small) did not change after IV administration of drug, but in the M-0.5 group both sizes of pial arterioles dilated significantly after IV milrinone administration (Fig. 1). In all groups, large and small pial arterioles showed significant dilation just after unclamping; the maximum increases in diameter for these 2 sizes of arterioles were by 12% and 14% for M-0.05, by 15% and 19% for M-0.5, by 12% and 13% for C-0.05, and by 12% and 15% for C-0.5, respectively (P < 0.05 versus baseline). Except in the M-0.5 group, these dilations were not significantly different from those seen in the corresponding control group.
The pial arteriolar constriction observed at 5 min or more after unclamping in the control group was completely inhibited by milrinone (in the M-0.05 and M-0.5 groups) in both large and small arterioles. Indeed, the arterioles were instead dilated in the M-0.5 group (11% and 12% at 5 min and 8% and 9% at 60 min after unclamping for large and small arterioles, respectively) (Figs. 1 and 2). Although vasoconstriction was still present in the C-0.05 group, it was significantly attenuated in the C-0.5 group in both large and small arterioles (7% and 9% at 5 min, and −3% and −4% at 60 min after unclamping for large and small arterioles, respectively). For each of the five groups, the small arterioles tended to be more reactive than the large ones but not significantly so (Figs. 1 and 2).
In the second experiment, hemodynamic and physiological variables in both groups were similar to those in the first experiment. There were no significant differences between groups for baseline Pao2 or PBrO2 (control group, 134 ± 19 mm Hg and 22 ± 5 mm Hg; milrinone group, 129 ± 17 mm Hg and 20 ± 2 mm Hg). In the control group, rCBF decreased significantly from 15 min after unclamping and PBrO2 decreased significantly from 10 min after unclamping (Fig. 3). In the milrinone group, rCBF increased significantly after IV administration of milrinone and significant differences between those groups were observed throughout the experiments (at 60 min after unclamping, −16% for control group and 5% for milrinone group) (Fig. 3a). Decrease of PBrO2 after unclamping observed in the control group was significantly attenuated in the milrinone group (at 60 min after unclamping, −35% for control group and 4% for milrinone group) (Fig. 3b).
The major findings of the present study were that in rabbits: 1) IV milrinone, but not colforsin daropate, caused a significant pial arteriolar dilation (at the larger dose tested), 2) with milrinone (larger dose tested), but not with colforsin daropate, the pial arteriolar dilation seen just after unclamping of an abdominal aortic cross-clamp was larger than in the control (saline) group, 3) both milrinone and colforsin daropate attenuated the sustained pial arteriolar vasoconstriction seen in the control group after unclamping of an abdominal aortic cross-clamp, and 4) unclamping induced sustained rCBF and PBrO2 decreases in the control group, but these decreases were attenuated by IV infusion of milrinone.
Milrinone acts both as a powerful inotrope and as a vasodilator because it increases cyclic 3′,5′-adenosine monophosphate (cAMP) levels through inhibition of type 3 cAMP-specific phosphodiesterase in both cardiac and vascular smooth muscle. Increased cAMP concentrations may produce smooth muscle relaxation via several mechanisms, including protein kinase A, a direct effect on vascular Ca2+, Mg2+-adenosine triphosphatase, or Na+, K+- adenosine triphosphatase activity (7,8).
We (1) previously reported that, in rabbits, unclamping of an abdominal aortic cross-clamp causes a transient dilation of cerebral pial arterioles (for approximately 2 minutes), followed by a sustained (for at least 60 minutes) vasoconstriction mediated, at least in part, by a washout of the TxA2 produced in distal tissues during the period for which the aorta is clamped and probably after cross-clamp release (for at least 60 minutes). Later (9), using the closed cranial window technique to observe the pial microcirculation, we found that IV infusion of milrinone produced significant pial arteriolar dilations at a dose of 0.5 μg/ · kg−1 · min−1. Khajavi et al. (10) and Arakawa et al. (11) reported that milrinone is a vasorelaxant that can prevent delayed cerebral ischemia resulting from vasospasm in both animals and humans. Moreover, milrinone is a potent inhibitor of TxA2 synthesis (12,13), and it completely reverses U46619 (a stable TxA2 analog)-induced contraction in the human internal mammary artery (14). Our present findings are consistent with the results noted in these previous reports. Further, the clinical doses of 0.05 and 0.5 μg/ · kg−1 · min−1 of milrinone used in the present study significantly attenuated the sustained pial arteriolar vasoconstriction seen in the control group after unclamping of the abdominal aortic cross-clamp.
Colforsin daropate, a water-soluble forskolin derivative, (+)-(3R, 4aR, 5S, 6S, 6aS, 10S, 10aR, 10bS-5-acetoxyl-6-(3-dimethylaminopropionyloxy)-dodecahydro-10, 10b-dihydroxy-3, 4a, 7, 7, 10a-pentamethyl-3-vinyl-1H-naphtho[2, 1-b]pyran-1-one monohydrochloride (Adehl®; Nihonkayaku, Tokyo, Japan), exerts a positive inotropic influence by directly activating adenylate cyclase (15). It thereby increases cardiac performance in patients with acute heart failure (16). Moreover, forskolin, a water-insoluble adenylate cyclase stimulant, produces cerebral vasodilation both in vivo and in vitro (17–19) with topical application at 10−9 or 10−6 mol/L producing cerebral vasodilation (as assessed using the cranial window technique) in both mice and rats (18,19). Cracowski et al. (20) reported that U46619 inhibited both the forskolin-induced relaxation and the forskolin-induced production of cAMP in the human internal mammary artery. The results reported here are consistent with these earlier observations.
The differences in vascular responses to aortic unclamping between the milrinone and colforsin groups could be explained as follows. Cracowski et al. (20), who examined the human internal mammary artery, suggested that U46619 exerts an inhibitory effect on the cAMP pathway for relaxation but not on the nitric oxide/cyclic guanosine monophosphate (cGMP) pathway for relaxation. Milrinone, but not colforsin daropate, may produce vascular relaxation by a mechanism not clearly related to changes in cAMP (8). For example, in rat and guinea pig aortic smooth muscle, milrinone increases cGMP concentrations; cGMP is a second messenger that causes vasodilation in vascular beds in some species (6). In addition, the difference between milrinone and colforsin daropate in blood-brain barrier permeability, which could be related to molecular weight (milrinone, 211; colforsin daropate, 546), might contribute to their differential effects on cerebral vasoreactivity after aortic unclamping.
Both rCBF and PBrO2 decreased significantly from 10 or 15 min to at least 60 min after aortic unclamping in the second experiment. Although the incidence of neurological complications in the CNS related to abdominal aortic surgery seems not to be particularly frequent, such complications are potentially serious (21,22). Cerebral vasoconstriction and decreased rCBF and PBrO2 after aortic unclamping increase the risk for neurological complications caused by cerebral ischemia and tissue hypoxia. Therefore, partial or complete prevention of such decreases of rCBF and PBrO2 by a concomitantly administered drug such as milrinone can be expected to be favorable in the clinical setting.
In patients who have a damaged endothelium (such as those with atherosclerosis or hypercholesterolemia), the presumed TxA2-induced responses of cerebral vessels to aortic unclamping could be different from those seen in healthy subjects, possibly more pronounced than those seen in healthy subjects. In fact, it has been reported that endothelial damage induces arterial thrombosis via an increase in TxA2 (23,24). If so, it might be advantageous to use milrinone or colforsin daropate during abdominal aortic aneurysmectomy in patients with damaged endothelium to prevent or ameliorate ischemic events in the CNS.
Because the basal anesthetic state with pentobarbital might affect the tone of the cerebral arterioles, we cannot exclude the possibility that the observed effects on pial arteriolar tone could have been modulated by pentobarbital.
In conclusion, pial arteriolar vasoconstriction was induced in pentobarbital-anesthetized rabbits after unclamping the aorta after a 20-min aortic cross-clamp. This unclamping-induced vasoconstriction was prevented or attenuated by IV infusion of milrinone and colforsin daropate. Although both rCBF and PBrO2 decreased significantly after unclamping, milrinone attenuated those reactions.
1. Uchida M, Iida H, Iida M, Dohi S. Changes in cerebral microcirculation during and after abdominal aortic cross-clamping in rabbits: role of thromboxane A2 receptor. Anesth Analg 2003;96:651–6.
2. Aadahl P, Saether OD, Stenseth R, et al. Cerebral haemodynamics during proximal aortic cross-clamping. Eur J Vasc Surg 1991;5:27–31.
3. Ishiyama T, Dohi S, Iida H, et al. Mechanisms of vasodilation of cerebral vessels induced by the potassium channel opener nicorandil in canine in vivo experiments. Stroke 1994;25:1644–50.
4. Dohi S, Matsumoto M, Takahashi T. The effects of nitroglycerin on cerebrospinal fluid pressure in awake and anesthetized humans. Anesthesiology 1981;54:511–4.
5. Cuffe MS, Califf RM, Adams KF Jr, et al. Short-term intravenous milrinone for acute exacerbation of chronic heart failure: a randomized controlled trial. JAMA 2002;287:1541–7.
6. Wajima Z, Yoshikawa T, Ogura A, et al. Intravenous colforsin daropate, a water-soluble forskolin derivative, prevents thiamylal-fentanyl-induced bronchoconstriction in humans. Crit Care Med 2002;30:820–6.
7. Silver PJ, Lepore RE, O’Connor B, et al. Inhibition of the low Km cyclic AMP phosphodiesterase and activation of the cyclic AMP system in vascular smooth muscle by milrinone. J Pharmacol Exp Ther 1988;247:34–42.
8. Kauffman RF, Schenck KW, Utterback BG, et al. In vitro
vascular relaxation by new inotropic agents: relationship to phosphodiesterase inhibition and cyclic nucleotides. J Pharmacol Exp Ther 1987;242:864–72.
9. Iida H, Iida M, Takenaka M, et al. The effects of alpha-human atrial natriuretic peptide and milrinone on pial vessels during blood-brain barrier disruption in rabbits. Anesth Analg 2001;93:177–82.
10. Khajavi K, Ayzman I, Shearer D, et al. Prevention of chronic cerebral vasospasm in dogs with milrinone. Neurosurgery 1997;40:354–62.
11. Arakawa Y, Kikuta K, Hojo M, et al. Milrinone for the treatment of cerebral vasospasm after subarachnoid hemorrhage: report of seven cases. Neurosurgery 2001;48:723–30.
12. Barradas MA, Jagroop A, O’Donoghue S, et al. Effect of milrinone in human platelet shape change, aggregation and thromboxane A2 synthesis: an in vitro
study. Thromb Res 1993;71:227–36.
13. Jeremy JY, Gill J, Mikhailidis D. Effect of milrinone on thromboxane A2 synthesis, cAMP phosphodiesterase and 45Ca2+ uptake by human platelets. Eur J Pharmacol 1993;245:67–73.
14. Huraux C, Makita T, Montes F, et al. A comparative evaluation of the effects of multiple vasodilators on human internal mammary artery. Anesthesiology 1998;88:1654–9.
15. Hosono M, Kanbe E, Noguchi M, et al. Effects of NKH477 (colforsin daropate), a novel forskolin derivative, in isolated cardiac muscles. Clin Pharmacol Ther 1996;6:1061–72.
16. Hosoda S, Motomiya T, Katagiri T, et al. Acute effect of colforsin daropate, a novel forskolin derivative, in patients with acute heart failure: a multicenter placebo-controlled double-blind trial. Jpn J Clin Pharmacol Ther 1997;28:583–602.
17. Suzuki Y, Huang M, Lederis K, Rorstad OP. The role of adenylate cyclase in relaxation of brain arteries: studies with forskolin. Brain Res 1988;457:241–5.
18. Rosenblum WI. In vivo
evidence that an adenylate cyclase-cAMP system dilates cerebral arterioles in mice. Stroke 1988;19:888–91.
19. Ibayashi S, Ngai AC, Meno JR, Winn HR. Effects of topical adenosine analogs and forskolin on rat pial arterioles in vivo
. J Cereb Blood Flow Metab 1991;11:72–6.
20. Cracowski JL, Stanke-Labesque F, Chavanon O, et al. Thromboxane A(2) modulates cyclic AMP relaxation and production in human internal mammary artery. Eur J Pharmacol 2000;387:295–302.
21. Millar SM, Alston RP, Andrews PJD, Souter MJ. Cerebral hypoperfusion in immediate postoperative period following coronary artery bypass grafting, heart valve, and abdominal aortic surgery. Br J Anaesth 2001;87:229–36.
22. Sakka SG, Huttemann E. Ischemia brain infarct and rupture of an infrarenal aortic aneurysm. Anaesthesist 2003;52:801–4.
23. Imura Y, Terashita Z, Nishikawa K. The role of thromboxane (TX) A2 in rabbit arterial thrombosis induced by endothelial damage. Thromb Res 1990;59:195–205.
© 2005 International Anesthesia Research Society
24. Naruse K, Shimizu K, Muramatsu M, et al. Long-term inhibition of NO synthesis promotes atherosclerosis in the hypercholesterolemic rabbit thoracic aorta. PGH2 does not contribute to impaired endothelium-dependent relaxation. Arterioscler Thromb 1994;14:746–52.