BACKGROUND. The effect of vasoconstrictors on intracerebral hemodynamics in anesthetized patients is controversial. The influence of phenylephrine and norepinephrine on the cerebral circulation was investigated in isoflurane- or propofol-anesthetized patients using transcranial Doppler ultrasonography.
Forty patients were randomly assigned to have vasoconstrictor tests with norepinephrine or phenylephrine during either isoflurane or propofol anesthesia. Blood flow velocities were simultaneously measured in the middle cerebral artery and ipsilateral extracranial internal carotid artery. Baseline recordings were done during stable anesthesia in a supine position (test 0). A second series of measurements were performed after norepinephrine or phenylephrine had increased mean arterial blood pressure by about 20% (test 1). With maintained norepinephrine or phenylephrine infusion, a final series of results were obtained after the increased mean arterial blood pressure was counteracted by a slightly head-up patient position (test 2).
Both vasoconstrictors significantly increased mean flow velocities in the middle cerebral artery (norepinephrine: 43 +/- 11 cm/s to 49 +/- 11 cm/s; phenylephrine: 43 +/- 8 cm/s to 48 +/- 9 cm/s; +/- SD) and internal carotid artery (norepinephrine: 27 +/- 7 cm/s to 31 +/- 8 cm/s; phenylephrine: 27 +/- 9 cm/s to 31 +/- 10 cm/s) in the isoflurane-but not in the propofol-anesthetized patients. In the head-up position, only small and insignificant flow velocity changes were observed in both cerebral arteries independent of the vasoconstrictor or background anesthetic.
The results of the present study indicate that norepinephrine and phenylephrine do not directly affect intracranial hemodynamics in anesthetized patients, but rather that hemodynamic changes observed with vasoconstrictors reflect the effect of the background anesthetic agents on cerebral pressure autoregulation.
BASED on studies in awake persons, vasoconstrictors are generally considered to have limited influence on the cerebrovasculature. [1–3]However, their effect in anesthetized patients is not well established, and most data available are derived indirectly from studies in which testing of vasoactive drugs was not the primary interest of the study. Studies in awake animals suggest controversial and inconsistent effects of vasoconstrictive agents on intracerebral hemodynamics. [4–7]Studies investigating the influence of selected vasoconstrictors in defined anesthetic procedures remain to be done.
The objective of this study was to investigate the effect of the potent vasoconstrictive agents phenylephrine and norepinephrine on the cerebral circulation of patients anesthetized with either propofol or isoflurane. To evaluate changes in cerebral hemodynamics induced by vasoconstrictors, we used transcranial Doppler ultrasonography, which provided a rapid and noninvasive assessment of relative changes in cerebral blood flow (CBF). [8,9]
The study protocol was approved by the Committee for Ethics in Human Research of the University of Basel, and written informed consent was obtained from 40 patients classified as American Society of Anesthesiologists physical status 1 or 2 who were scheduled for elective lumbar laminectomy or discectomy. Patients were excluded if they were known to have neurological or cardiovascular diseases.
All study procedures were done with patients in a supine position and completed before surgery. The patients were randomly allocated to one of four study groups: isoflurane-anesthetized patients treated with norepinephrine (group 1), isoflurane-anesthetized patients treated with phenylephrine (group 2), propofol-anesthetized patients treated with norepinephrine (group 3), and propofol-anesthetized patients treated with phenylephrine (group 4). Patients were fasted and unpremedicated. Electric activity of the heart, heart rate, end-tidal concentration of carbon dioxide (PETCO(2)) and isoflurane pulse oximetry, and esophageal temperature (Spacelabs Medical Inc., Redmond, WA) were routinely monitored. A radial artery catheter was inserted to measure continuous invasive mean arterial blood pressure (MABP) and to analyze blood gas samples.
Anesthesia was induced with isoflurane in groups 1 and 2 and with 2 mg/kg propofol in groups 3 and 4. If necessary, inhalational induction with isoflurane was supplemented with propofol titrated to facilitate mask ventilation with increasing gas concentrations. Fentanyl was administered in a dose range of 3–4 [micro sign]g/kg to provide stable hemodynamics. The tracheal intubation was facilitated by 0.5 mg/kg atracurium. After intubation, the lungs were mechanically ventilated (Sulla 808V, Drager Werker AG, Lubeck, Germany) with air and oxygen (FIO2= 0.5) adjusted to a PET (CO)(2) corresponding to a normocapnic arterial concentration of carbon dioxide (PaCO(2)). Anesthesia was maintained in groups 1 and 2 with 1 MAC isoflurane, considered to be 1.15%, and in groups 3 and 4 with 170 [micro sign]g [middle dot] kg-1[middle dot] min-1(10 mg [middle dot] kg-1[middle dot] h-1) propofol that was reduced to 130 [micro sign]g [middle dot] kg-1[middle dot] min-1(8 mg [middle dot] kg-1[middle dot] h-1) after 10 min and to 100 [micro sign]g [middle dot] kg-1[middle dot] min-1(6 mg [middle dot] kg (-1)[middle dot] h-1) after 20 min. 
The right middle cerebral artery (MCA) and the ipsilateral extracranial internal carotid artery (ICA) were insonated simultaneously according to standard criteria using 2-MHz transcranial Doppler transducers (Multidop MDX4 TCD8; DWL Elektronische Systeme GmbH, Sipplingen, Germany). The MCA was insonated through the temporal window with the transducer fixed in place by a custom-designed frame. The retromandibular and extracranial ICA were insonated high in the neck, with the transducer placed at the angle of the jaw and anchored by a self-retaining retractor system to the operating table. The depths of insonation in the MCA (30–50 mm) and in the ICA (50–60 mm) were adjusted to show simultaneously high and clear velocity signals at two different depths (separated by a 5-mm distance). Cerebral blood flow velocity in the MCA (V (MCA)) and the ICA (VICA) and the MABP were displayed on a video screen and recorded using the standard algorithm implemented on the instrument.
Cerebral blood flow velocity measurements were performed three times on each patient. Initially, during stable volatile or intravenous anesthesia (minimally 30 min after anesthesia induction), baseline blood flow velocity measurements were obtained (test 0). Next the allocated vasoconstrictor agent was titrated using a slow infusion until the MABP was increased 20% above the baseline value. Once MABP had reached a stable new plateau, the flow velocity measurements were repeated after steady-state parameters had persisted for 10 min (test 1). Under identical conditions, a final series of flow velocity data were obtained after having carefully placed the patient in a slightly head-up position while keeping the vasoconstrictor infusion constant (test 2). This maneuver was performed using an operating Table withthe rotation axis just under the patient's chest and allowing for a slow and smooth tilting procedure. The Table movementwas stopped when the vasoconstrictor induced MABP change, measured at the skull base level (external auditory canal), had decreased to baseline (i.e., test 0 values). To avoid a simultaneous MABP decrease, measured at the heart level, the patient's legs, if necessary, were slightly elevated. This tilting maneuver was performed to reduce the pharmacologically increased MABP at the brain level, representing cerebral perfusion pressure, but not at heart level.
Results are expressed as means +/- SD when not otherwise indicated. The two CBF velocity values per cerebral artery, simultaneously recorded at two different depths, were averaged before statistical analysis. Cerebral blood flow velocity and MABP changes were calculated as percentage changes of baseline values. Data were analyzed using commercially available statistical software (SPSS Inc., release 6.1 for windows, Chicago, IL). Probability values <0.05 were considered to be significant. To examine the vasoconstrictor effect on the two insonated cerebral arteries, the data from the 40 patients were pooled and subjected to linear correlation analysis. Analysis of variance for repeated measures was used to assess within-group effects (tests 0–2) and between-group effects (anesthetic agents, vasoconstrictors). In cases of a significant within-group effect, the means of the three stages of the intervention (tests 0–2) were compared using unpaired Student's t tests with Bonferroni's correction separately for each group. If a significant effect of the anesthetic agent or the vasoconstrictor was found, means were compared using unpaired paired Student's t tests with Bonferroni's correction as appropriate.
There were no significant differences among the four groups with respect to age, weight, and hemoglobin (Table 1).
During the study, initial body temperature, ranging from 36.2 to 37.0 [degree sign]C, was kept constant. The PETCO(2), corresponding to a normocapnic baseline PaCO(2), was maintained at a constant level during the study procedures (Table 2). Changes of hemodynamics and intracranial dynamics are also summarized in Table 2. During stable isoflurane anesthesia, both norepinephrine and phenylephrine significantly increased MABP and both VMCAand VICA. When MABP was returned to normal levels by tilting the table, both flow velocities returned to baseline. During propofol anesthesia, a similar increase in MABP was not accompanied by any significant changes in VMCAand VICA. Similarly, decreasing MAP to baseline by Table tilthad no influence on flow velocities (Figure 1). Analysis of variance for repeated measures of the CBF velocity data revealed a significant within-group effect (tests 0–2) and a significant effect of the anesthetic agent (F sub [1–37]= 68.8; P < 0.001). However, no significant effect of the vasoconstrictor was noted (F sub [1–37]= 0.6), indicating that norepinephrine and phenylephrine did not differ with respect to their effect on CBF velocities.
A correlation analysis plotting change in MABP versus change in relative flow velocities (combining tests 0, 1, and 2) for the two anesthetics is shown in Figure 2. This analysis revealed a significant correlation between velocity changes during test 1 (r = 0.82, P < 0.001) and test 2 (r = 0.83, P < 0.001).
Our results indicate that the vasoconstrictive agents norepinephrine and phenylephrine do not directly affect CBF in normal anesthetized patients, and that CBF changes observed with these vasoconstrictive agents reflect background anesthetic-induced impairment of cerebral pressure autoregulation rather than intrinsic effects of the vasoconstrictors on the cerebrovasculature. Because these results were obtained in patients without intracranial disease, caution should be used in applying these data to brain-injured patients.
Norepinephrine is a endogenous catecholamine with mixed [Greek small letter alpha]- and [Greek small letter beta]-agonist properties but with predominant [Greek small letter alpha]1-adrenergiceffects that can produce intense arterial and venous vasoconstriction. Phenylephrine is a selective [Greek small letter alpha]1-adrenergicagonist with essentially no [Greek small letter beta]-adrenergic activity. Both of these vasoconstrictors are generally considered to have no significant influence on intracerebral hemodynamics. Studies performed in animals [5,15]or humans [1–3]indicate that, in themselves, vasoconstrictive agents do not alter intracerebral hemodynamics. However, animal investigations have suggested that during anesthesia their effect on intracerebral dynamics may depend on the background anesthetic used. With volatile anesthetics cerebral vasoconstriction, vasodilation, and no change in vasomotor tone have been demonstrated. As yet no data are available for the effect of vasoconstrictive agents in the presence of intravenous anesthetics. Furthermore, animal studies have shown that CBF may be markedly different with different vasopressors. [4,5,17,18]
The absence of changes in CBF during exogenous sympathomimetic vasoconstrictors is not explained by the lack of sympathetic receptors or innervation in the cerebrovasculature. Cerebral vessels have a rich neuronal innervation and several neurotransmitters have been identified. Failure of vasoactive drugs to cause direct effects on the cerebral vasculature may be related to the blood-brain barrier, which prevents them from directly interacting with the cerebrovascular smooth muscle. In vivo studies with intact blood-brain barrier have shown that circulating catecholamines produce modest or no increases in cerebrovascular resistance, which in turn does not affect CBF. 
Limitations and methodologic considerations that need to be addressed when interpreting transcranial Doppler results are described elsewhere. Briefly, the validity of the assumption that the CBF velocities we obtained in the current study are proportional to flow volume depends on the premise that the cross-sectional areas of the two insonated vessels do not significantly change during MABP changes (i.e., norepinephrine or phenylephrine infusion). However, because the ICA is a large conductance vessel and variations in cerebrovascular resistance are largely effected by changes in resistance vessels, it is unlikely that vasoconstrictors would have direct effects on the ICA diameter. In addition, direct observations of the MCA during craniotomy indicate that the diameter of this artery changes only slightly during vasoconstrictor-induced MABP changes in the range tested in this study. Further, the proportional change of the CBF velocity in both series-connected arteries throughout the current study makes a relevant vasoconstrictor influence on the MCA or ICA diameter, or both, unlikely. Therefore, we consider the CBF velocity changes measured in the current study as valid indicators for CBF variations during the norepinephrine- or phenylephrine-induced MABP alterations.
Because the intracranial vasculature can autoregulate CBF during blood pressure manipulations, the effect of the vasoactive medication on the cerebral circulation could have been influenced by such autoregulatory mechanisms. In an attempt to eliminate pressure autoregulation effects, a hydrostatic gradient was created by carefully placing the patient in a slightly head-up position to counteract the pharmacologically increased blood pressure. By performing this maneuver while maintaining the vasoconstrictor infusion constant, autoregulatory effects on the cerebral hemodynamics were excluded.
A potential criticism in this approach to eliminate pressure-induced effects on intracranial hemodynamics is that we monitored MABP but not jugular venous pressure, which could be another crucial determinant of cerebral perfusion pressure and CBF. Although both tested vasoconstrictors exert their chief vascular action on arterioles and precapillary sphincters, a venous vasoconstriction cannot be entirely excluded. In contrast, the head up-position might well have reduced jugular venous pressure by improving cerebral venous outflow. However, there were only moderate changes in the pressure increase and head elevation, and we believe that these opposing effects on venous outflow pressure did not significantly influence our results.
Based on the previous assumption of constant ICA and MCA diameters, the significant increase in flow velocity with norepinephrine and phenylephrine that we observed during isoflurane but not during propofol anesthesia suggests an anesthetic-induced rather than a direct vasoconstrictor-induced change in intracranial hemodynamics. A direct cerebrovascular effect of norepinephrine and phenylephrine only during volatile but not during intravenous anesthesia is unlikely. On the other hand, impaired cerebral pressure autoregulation has been demonstrated using volatile anesthetics; this has not, however, been confirmed with intravenous anesthetics. Therefore, we assume that cerebral pressure autoregulation mechanisms are responsible for the observed hemodynamic differences between the volatile and intravenous anesthetic regimens. This hypothesis might be supported by flow velocities returning to baseline value in the isoflurane groups after restoration of baseline MABP at the skull base (i.e., restoration of baseline cerebral perfusion pressure) by placing the patient's head in an elevated position and eliminating autoregulatory mechanisms as described before.
In conclusion, norepinephrine and phenylephrine infused to increase MABP in anesthetized patients failed to show any significant vasoconstrictor effects on cerebral circulation. The observed increases in V (MCA) during isoflurane anesthesia might be suggestive of an anesthetic-induced cerebral pressure autoregulation impairment caused by the volatile anesthetics rather than of an intrinsic effect of the vasoconstrictor agents. However, our study was performed in patients without intracranial diseases. Extrapolation of these data to patients affected with cerebral injury, hypoxia, or ischemia, and most often a defective blood-brain barrier, might be inappropriate, because, as in these cases, the effect of vasoconstrictors might be substantially different from those described here under physiologic conditions.