Previous studies indicate that the volatile anesthetics, halothane and isoflurane, produce dose-dependent increases in cerebral blood flow (CBF) [1,2]. It is generally agreed that inhaled anesthetics do not alter the total CBF response to arterial carbon dioxide tension (PaCO2) [3,4]. However, the effects of halothane and isoflurane on the cerebral microcirculation are less well elucidated.
Laser-Doppler flowmetry is a relatively new technique suitable for the continuous measurement of changes in perfusion of tissues. In a recent study  we compared the effects of halothane and isoflurane on cerebrocortical red blood cell flow and flow autoregulation using laser-Doppler flowmetry. We have also demonstrated the effects of CO2 inhalation on laser-Doppler flow (LDF) during barbiturate anesthesia . However, the effects of volatile anesthetics on the CO2 response of the cerebrocortical microcirculation have not been compared using the laser-Doppler technique. Furthermore, while some investigators have demonstrated similar results with laser-Doppler flowmetry and other CBF measurement modalities [7,8], others have demonstrated that laser-Doppler flowmetry and autoradiography yielded different results in an identical animal model .
Laser-Doppler flowmetry measures nondirectional perfusion of tissues by red blood cells in the microcirculation. Because cerebral capillary hematocrit is lower than systemic arterial hematocrit , it cannot be inferred that the influences on red blood cell flow mirror the effects of anesthetics on total or regional CBF. Thus, it is not known whether various anesthetics or other vasoactive substances have similar effects on total CBF and on red blood cell flow in the microcirculation. For example, it is possible that a larger fraction of red blood cells may be shunted away by arteriovenous anastomoses during hypercapnic vasodilation .
Since the red blood cell is the major oxygen carrier in the blood, the distinction between red blood cell flow in the microcirculation and overall blood flow in the brain is important. Hyperventilation-induced hypocapnia is used often during neurosurgical anesthesia to reduce cerebral bulk volume and intracranial pressure. Therefore, the specific effects of PaCO2 on cerebrocortical red blood cell flow during various modalities of anesthesia may provide useful information regarding cerebral oxygen availability.
The experimental procedures and protocols used in this investigation were reviewed and approved by our institutional animal care committee. Adult male Sprague-Dawley rats (250-400 g) were anesthetized with intraperitoneal pentobarbital (65 mg/kg body weight). The rats were then tracheotomized and the trachea intubated with polyethylene tubing (PE 240, Intramedic). One femoral artery and one femoral vein were then cannulated with polyethylene tubing (PE 50, Intramedic) to facilitate blood pressure measurement and drug administration. The rats' lungs were then ventilated (Model 707; Harvard Apparatus, South Natick, MA) with 30% oxygen in nitrogen. The inspired gases were mixed using a calibrated threetube flowmeter (Union Carbide Corp., South Plainfield, NJ). Body temperature was maintained at 37 +/- 0.2 degrees C with a water-circulated heating pad (YSI Model 73ATA and Model K-2-S; American Pharmaseal Co., Valencia, CA). When the rats were ventilated, anesthesia was maintained with either pentobarbital infusion (10-20 mg centered dot kg-1 centered dot h-1) or 0.5 minimum alveolar anesthetic concentration (MAC) of either halothane or isoflurane, depending on the experimental protocol. Halothane and isoflurane were vaporized into the breathing system using Ohio Liquid Vaporizers (Model F100; Ohio Medical Products, Airco Inc., Madison, WI). Inspired and expired oxygen, carbon dioxide, and volatile anesthetic concentrations were measured continuously (POET II Trademark; Criticare Systems, Inc., Milwaukee, WI). End-tidal carbon dioxide (ETCO2), mean arterial blood pressure (MABP), LDF, and concentration of anesthetic vapor were displayed and recorded on a Grass polygraph (Model 7; Grass Instruments Co., Quincy, MA).
In all experiments, the head was placed in a stereotaxic apparatus (Model 900; David Kopf, Tujunga, CA). The scalp and connective tissue were removed in a 1-cm2 area. A burr hole of 1-2 mm in diameter was drilled in the cranium over the right parietal cerebral cortex using a low-speed (0-8000 rpm) air drill (Model Rhino XP; Midwest Dental, DesPlaines, IL) with the aid of a stereomicroscope. The burr hole was deepened until only a thin translucent plate of skull remained, as previously described by Hudetz et al. . At the end of the experimental protocol the rats were killed by an intravenous (IV) injection of pentobarbital 50 mg + gallamine 20 mg followed by creation of a surgical pneumothorax.
Volatile anesthetics were obtained from Anaquest Inc. (Madison, WI). Pentobarbital, gallamine, and phenylephrine were obtained from Sigma Chemical Co. (St. Louis, MO). All IV drugs were reconstituted on the day of experiment with a balanced salt solution (Plasmalyte Registered Trademark; Baxter Healthcare Corp., Deerfield, IL).
Red blood cell flow was monitored throughout the experiments, using a laser-Doppler flowmeter (PF3; Perimed, Stockholm, Sweden) and a flow probe (PF316, dental probe) with a tip diameter of 1 mm. This probe measures LDF in a cone-shaped volume of approximately 1 mm3. The probe was lowered into the bottom of the cranial burr hole using a micromanipulator. A drop of mineral oil was applied to improve optical coupling between the flow probe and brain tissue. Once established, the probe position was not altered for the duration of the experimental protocol. Care was taken to avoid placing the probe over visible vessels so that the LDF signal was indeed detected from capillaries, small arterioles, and venules.
After placement of the LDF probe, the preparation was allowed to stabilize for 1 h on pentobarbital infusion (10-20 mg centered dot kg-1 centered dot h (-1)) or at 0.5 MAC halothane or isoflurane, and baseline LDF was measured at normocapnia (ETCO2 = 35-37 mm Hg). Hyperventilation to ETCO2 of 22 mm Hg was then performed and another LDF measurement taken 5 min after a steady state was achieved. Hyperventilation was discontinued, normocapnia reestablished, and a control measurement performed 5 min later. Then 5% CO2 was added to the inspired gas mixture to yield an ETCO (2) of 65-67 mm Hg and another measurement performed, again after 5 min of steady-state conditions. In the experiments performed at 1.5 MAC, the preparation was allowed to stabilize for 1 h at 0.5 MAC halothane or isoflurane, and baseline LDF was measured at 0.5 MAC at normocapnia (ETCO2 = 35-37 mm Hg). Anesthetic concentration was then increased to 1.5 MAC and, after 5 min equilibration, CO2 reactivity measured at hypo-, normo-, and hypercapnia as described above. The alpha1-adrenoreceptor agonist phenylephrine (0.5-5 micro gram centered dot kg-1 centered dot min-1) was administered IV, as required, to maintain MABP at control levels (i.e., MABP at 0.5 MAC at normocapnia).
Each LDF measurement was averaged over a 5-min test period during steady-state conditions. The sample sizes were 4, 6, 6, 5, and 5 for pentobarbital infusion, 0.5 MAC halothane, 0.5 MAC isoflurane, 1.5 MAC halothane, and 1.5 MAC isoflurane, respectively. The LDF response was, in all protocols, expressed as percentage of change from control. Control LDF was taken at commencement of the experiment, under steady-state conditions using either pentobarbital infusion or 0.5 MAC volatile anesthetic at normocapnia (ETCO2 = 35-36 mm Hg). Data were expressed as mean +/- SEM. Statistical comparisons were made using repeated-measures analysis of variance and Student-Newman-Keuls test for comparison of effects of different anesthetics. To check for a difference in CO2 reactivity between 1.5 MAC halothane and 1.5 MAC isoflurane, Fisher's test was used. CO2 reactivity was defined as the percentage of LDF change per mm Hg change in ETCO2. The relationship between ETCO2 and LDF was tested using linear regression analysis. P < 0.05 was considered significant. Each mean represents data from four to six rats.
Baseline LDF values, at an ETCO2 of 36 mm Hg were 159 +/- 8 perfusion units (PU) at 0.5 MAC halothane, 158 +/- 12 PU at 0.5 MAC isoflurane, and 143 +/- 13 PU during pentobarbital infusion (no difference). After testing the response to hypocapnia and restoration of normocapnia, the LDF was similar to the original baseline values. The corresponding mean arterial pressure values were 115 +/- 7, 125 +/- 5, and 117 +/- 14 mm Hg, respectively (no difference). MABP was maintained within 5 mm Hg of initial control blood pressure in all groups at all measurement points.
Significant responses of LDF to increases and decreases in arterial CO (2) were observed in all experiments. Figure 1 shows original tracings of continuously measured variables. In this example, the LDF response to changes in ETCO2 from 22 mm Hg to 36 mm Hg and then to 66 mm Hg is demonstrated at 1.5 MAC isoflurane. As can be seen from the tracing, the hyperemic response to hypercapnia occurs within seconds and is complete within minutes.
At all anesthetic concentrations, the relationship between ETCO2 and LDF was consistent with a linear response. That is, when we tested the relationship between ETCO2 and LDF using linear regression analysis, the r value varied between 0.92 and 0.97 and P value was 0.001 for each anesthetic. In addition, we tested to see whether there was a significant difference between the slopes of the regression lines in the hypocapnic to normocapnic range and their corresponding slopes in the normo- to hypercapnic range. We found no statistical difference for any of the anesthetic regimens used.
The hypercapnia-induced percentage of LDF increase during pentobarbital was similar to that at 0.5 MAC isoflurane and both were greater than that observed at 0.5 MAC halothane (P < 0.01) (Figure 2, top). No such difference could be detected in the hypocapnic response.
The CO2 response curves appeared similar at 1.5 MAC halothane and 1.5 MAC isoflurane (Figure 2, bottom). CO2 reactivities (percentage of LDF change/mm Hg ETCO2 change from hypo- to hypercapnia) at 1.5 MAC halothane and 1.5 MAC isoflurane suggested a difference but when tested, the calculated P value was P = 0.054. The CO2 reactivity was greater at 1.5 MAC than at 0.5 MAC for halothane (P < 0.05) but not for isoflurane Table 1.
LDF was greater at 1.5 MAC than at 0.5 MAC at each ETCO2 for both anesthetics (P < 0.01) suggesting the presence of volatile anesthetic-induced vasodilation in the whole ETCO2 range Figure 3.
The results from this study are consistent with a linear relationship between ETCO2 and cerebrocortical LDF during halothane and isoflurane anesthesia in the rat. Previous studies have shown that the CO2 response is preserved during isoflurane anesthesia in dogs  and during halothane  anesthesia in cats. Reivich  demonstrated that cerebral blood flow is linearly related to arterial CO2 tension within the range of 20-80 mm Hg in rhesus monkeys during barbiturate anesthesia. At less than 20 mm Hg, cerebral vessels are maximally constricted while near maximal vasorelaxation results at a PaCO2 of more than 80 mm Hg. Recently, Wang et al.  demonstrated a nonlinear response in rats anesthetized with halothane and nitrous oxide. In our study, the CO2 reactivity was similar in the hypocapnic to normocapnic and the normocapnic to hypercapnic range. This finding is consistent with, but does not prove, a linear relationship between cerebrocortical LDF and ETCO (2). Accurate assessment of the functional relationship between LDF and ETCO2 would require additional data points measured at several intermediate ETCO2 levels, which was not done in this study. We were satisfied with the results suggesting an approximately equal magnitude of LDF response at 22 and 66 mm Hg ETCO2, which is consistent with most previous studies on CBF.
For each millimeter of mercury change in PaCO2, CBF changes by 2%-4% [14,15]. In comparison, our LDF-ETCO2 slopes were low, varying between 1.19 +/- 0.14 and 2.67 +/- 0.35%/mm Hg, even though Cucchiara et al.  demonstrated CO2 reactivities of similar magnitude to these with halothane and isoflurane. Haberl et al. , using laser-Doppler flowmetry, demonstrated a 3.75% change in cortical red blood cell flow per mm Hg change in PaCO2. These data are not directly comparable to ours, since the investigators used an "open" window preparation in which the dura mater is opened and reflected. Open cranial window preparations are prone to the development of swelling and herniation of the brain, with the loss of cerebrospinal fluid and an unknown amount of vascular compression at the edge of the bone , especially, during global cerebral vasodilation. Our findings may suggest that the LDF signal may be damped as it traverses the intact dura and thinned cranium present in our experimental model. If this were true, however, the LDF signal would probably be equally affected at all CO2 concentrations and the CO2 reactivity would not be attenuated.
In this study, the CO2 response curves appear to be shifted upward (or to the left) at 1.5 MAC halothane and isoflurane compared with 0.5 MAC, indicating a direct vasorelaxant effect of these anesthetics independent of CO2. This was expected, since in all previous studies performed with halothane and isoflurane CBF increases at higher anesthetic concentrations. These results suggest that (i) the LDF measurement technique reflects a flow response consistent with a linear relationship between CBF and CO2, and (ii) CO2, in combination with volatile anesthetics, causes similar changes in red blood cell flow in the cerebrocortical microcirculation as it does on whole blood flow in brain. While we measured percentage of changes from control LDF and performed no quantitative measurements of CBF in these studies, Eyre et al.  demonstrated laser-Doppler flowmetry and microsphere-derived CBF in cats to be highly correlated (r = 0.92, P < 0.001). Haberl et al. , when comparing LDF with H2 clearance in rabbits, also demonstrated a very high correlation (r > 0.95) between the methods during chloralose and urethane anesthesia.
When comparing the influence of volatile anesthetics on the CO2 response, a possible dose-dependent difference between the effects of halothane and isoflurane was found. A greater effect of isoflurane than halothane on the LDF response to CO2 has not been reported. Whether this difference in CO2 reactivity during halothane and isoflurane anesthesia was due to the anesthetics' differing effect on the mechanisms of vascular relaxation to CO2, or to an indirect, cerebral metabolic effect, is not clear. Isoflurane is believed to cause a greater depression of cerebral metabolic requirement than halothane . This suggests that isoflurane would oppose, rather than augment, the CO2-induced increases in CBF. Baseline LDF was similar with all anesthetics tested in this study and, therefore, cannot be responsible for the between-anesthetic difference in the CO2 response.
We considered the possibility that the induction drug confounded the results through interactions with the volatile anesthetics. Sodium pentobarbital, a short-to-medium-acting barbiturate, was used for induction in all groups because it yielded approximately 30-45 min of surgical anesthesia, which was necessary to perform the surgical preparation. Animals in all three groups were then paralyzed and ventilated for a 1-h equilibration period on either 0.5 MAC halothane, isoflurane, or pentobarbital infusion. Due to the redistribution of pentobarbital to all body tissues and fluid compartments, it is unlikely that there was much of a residual effect of the initial, and only, intraperitoneal dose of pentobarbital at the time of testing. Pentobarbital would normally decrease CBF in response to decreased cerebral metabolism and, as the effect wears off, one would expect CBF to increase. In all of our experimental groups CBF was constant before testing and returned to baseline values on restoring normocapnia after the hypocapnia test. It is unlikely that halothane or isoflurane influenced the rate of elimination of sodium pentobarbital differently. We cannot exclude the possibility that a minimal constant background level of pentobarbital was present in the experiments, which may have been responsible for a lower CO2 response than previously reported. However, this should have influenced the experimental groups equally, and we have no evidence that a different interaction between pentobarbital and either of the volatile anesthetics influenced the CO2 reactivity.
In our experiments, phenylephrine was infused to support arterial blood pressure during inhalation of volatile anesthetics at high concentrations. As expected, the dose of phenylephrine was similar at 1.5 MAC halothane and 1.5 MAC isoflurane. Hypocapnia and hypercapnia had minimal effects on phenylephrine requirements. The effect of phenylephrine on CBF is negligible  and is not affected by halothane or isoflurane . ETCO2 is regarded as a valid reflection of PaCO2 in the healthy animal, except that, on average, ETCO2 is 3-5 mm lower than PaCO (2) . Factors that affect this include mismatching of pulmonary ventilation and perfusion or an error in the sampling method of the exhaled gas mixture. It could be argued that the alpha1-adrenoceptor agonist, phenylephrine, produced vasoconstriction, increasing dead space and resulting in a disparity between ETCO2 and PaCO2 levels. However, in recent control experiments, we found that phenylephrine infusion did not influence the relationship between ETCO2 and PaCO2 at 1.5 MAC halothane concentration . In further recognition of the well described difficulties in measuring ETCO2 in small animals, we performed experiments to test the relationship between ETCO2 and PaCO2 at hypocapnia and moderate hypercapnia Table 2. The relationship between ETCO2 and PaCO2 was maintained at a 3- to 5-mm difference at hypocapnia and moderate hypercapnia.
In the present study, CO2 reactivity was estimated from the percentage of change in LDF per mm Hg change in ETCO2 between the hypocapnic and the hypercapnic levels. Although the applied maneuvers to achieve hypocapnia and hypercapnia were different (i.e., hyperventilation and CO2 inhalation), the data demonstrate that the effects of volatile anesthetics on the LDF responses were similar during both hypocapnia and hypercapnia.
Reduction of cerebral blood volume by decreasing PaCO2 via hyperventilation is a routine procedure used to control acute increases in intracranial pressure in such states as severe head injury. Hyperventilation is also used to reduce normal brain bulk during intracranial aneurysm clipping in order to improve the surgical field and to reduce retractor pressure. Such hyperventilation, to decrease brain size during cerebral surgery, may adversely affect cerebral O2 supply, particularly if systemic arterial hypotension is also being used to improve operating conditions and to decrease bleeding if the aneurysm should rupture during operation. Therefore, if anesthesiologists possessed a volatile anesthetic that preferentially increased red blood cell flow in the microcirculation, while causing an equal or lesser increase in whole blood flow, it might be useful in clinical states where brain oxygenation is compromised. Whether isoflurane and, to a lesser degree, halothane have the potential to cause such a preferential increase in cortical red cell flow requires further experimental investigation, including the determination of local microvascular hematocrit.
In summary, the present study suggests that changes in cerebrocortical red blood cell flow with PaCO2 are greater with isoflurane than with halothane at least at 0.5 MAC and possibly at 1.5 MAC.
1. Todd MM, Drummond JC. A comparison of the cerebrovascular and metabolic effects of halothane and isoflurane in the cat. Anesthesiology 1984;60:276-82.
2. Hoffman WE, Edelman G, Kochs E, et al. Cerebral autoregulation in awake versus isoflurane-anesthetized rats. Anesth Analg 1991;73:753-7.
3. Cucchiara RF, Theye RA, Michenfelder JD. The effects of isoflurane on canine cerebral metabolism and blood flow. Anesthesiology 1974;40:571-4.
4. Drummond JC, Todd MM. The response of the feline cerebral circulation to PaCO2
during anesthesia with isoflurane and halothane and during sedation with nitrous oxide. Anesthesiology 1985;62:268-73.
5. Lee JG, Hudetz AG, Smith JJ, et al. The effects of halothane and isoflurane on cerebrocortical microcirculation as assessed by laser Doppler flowmetry. Anesth Anal 1994;79:58-65.
6. Hudetz AG, Roman RJ, Harder DR. Spontaneous flow oscillations in the cerebral cortex during acute changes in mean arterial pressure. J Cereb Blood Flow Metab 1992;12:491-9.
7. Haberl RL, Heizer ML, Marmarou A, et al. Laser-Doppler assessment of brain microcirculation: effect of systemic alterations. Am J Physiol 1989;25:H1247-54.
8. Eyre JA, Essex TJH, Flecknell PA, et al. A comparison of measurements of cerebral blood flow in the rabbit using laser Doppler spectroscopy and radionuclide labelled microspheres. Clin Phys Physiol Meas 1988;9:65-74.
9. Iadecola C. Does nitric oxide mediate the increases in cerebral blood flow elicited by hypercapnia. Proc Natl Acad Sci USA 1992;89:3913-6.
10. Cremer JE, Seville MP. Regional brain blood flow, blood volume and hematocrit values in the adult rat. J Cereb Blood Flow Metab 1983;3:254-6.
11. Tomita M, Fukuuchi Y, Tanahashi N, et al. Opening of shunts through the cerebral microvasculature indicated by tissue indicator dilution curves. In: Proceedings of the Fifth World Congress for Microcirculation. Louisville, KY: Microcirculatory Society, Inc., 1991:111.
12. Reivich M. Arterial PCO2
and cerebral hemodynamics. Am J Physiol 1964;206:25.
13. Wang Q, Paulson OB, Lassen NA. Effect of nitric oxide blockade by NG-nitro-L-arginine
on cerebral blood flow response to changes in carbon dioxide tension. J Cereb Blood Flow Metab 1992;12:947-53.
14. Ringaert KRA, Mutch WAC. Regional cerebral blood flow and response to carbon dioxide during controlled hypotension with isoflurane in the rat. Anesth Analg 1988;67:383-8.
15. Lauritzen M. Long-lasting reduction of cortical blood flow of the rat brain after spreading depression with preserved autoregulation and impaired CO (2
) response. J Cereb Blood Flow Metab 1984;4:546-54.
16. Lu GP, Kaul DK, Feldman SM, et al. Sodium nitroprusside (SNP) hypotension:intracranial pressure (ICP) and hemodynamics in pial arteriole in the rat. Microcirc Endothelium Lymphatics, 1990;6:315-41.
17. Drummond JD, Todd MM, Scheller MS, Shapiro HM. A comparison of the direct cerebral vasodilating potencies of halothane and isoflurane in the New Zealand White rabbit. Anesthesiology 1986;65:462-7.
18. Sokrab T-EO, Johansson BB. Regional cerebral blood flow in acute hypertension induced by adrenaline, noradrenaline and phenylephrine in the conscious rat. Acta Physiol Scand 1989;137:101-5.
19. Mutch WAC, Patel PM, Ruta TS. A comparison of the cerebral pressure-flow relationship for halothane and isoflurane at haemodynamically equivalent end-tidal concentrations in the rabbit. Can J Anaesth 1990;37:223-30.
20. Good ML. Principles and practice of capnography. In: 44th Annual Refresher Course Lectures. Park Ridge, IL: American Society of Anesthesiologists, Inc., 1993:222.