Cardiovascular Effects of Sevoflurane Compared with Those of Isoflurane in Volunteers

Methods: Twenty-one subjects were randomized to receive sevoflurane, isoflurane, or sevoflurane: 60% N²O. Anesthesia was induced and maintained by inhalation of the designated anesthetic. Hemodynamic measurements were performed before anesthesia, during controlled ventilation, during spontaneous ventilation, and again during controlled ventilation after 5.5 h of anesthesia.

Results: A few subjects became excessively hypotensive at high anesthetic concentrations (2.0 minimum alveolar concentration [MAC] sevoflurane, 1.5 and 2.0 MAC isoflurane), preventing data collection. Sevoflurane did not alter heart rate, but decreased mean arterial pressure and mean pulmonary artery pressure. Cardiac index decreased at 1.0 and 1.5 MAC, but in subjects with mean arterial pressure greater or equal to 50 mmHg returned to baseline values at 2.0 MAC when systemic vascular resistance decreased. Sevoflurane did not alter echocardiographic indices of ventricular function, but did decrease an index of afterload. Sevoflurane caused a greater decrease in mean pulmonary artery pressure than did isoflurane, but the cardiovascular effects were otherwise similar. Administration of sevoflurane with 60% N²O, prolonged administration or spontaneous ventilation resulted in diminished cardiovascular depression.

Conclusions: At 1.0 and 1.5 MAC, sevoflurane was well tolerated by healthy volunteers. At 2.0 MAC, in subjects with mean arterial pressure greater or equal to 50 mmHg, no adverse cardiovascular properties were noted. Similar to other contemporary anesthetics, sevoflurane caused evidence of myocardial depression. Hemodynamic instability was noted in some subjects at high anesthetic concentrations in the absence of surgical stimulation. The incidence was similar to that with isoflurane. The cardiovascular effects of sevoflurane were similar to those of isoflurane, an anesthetic commonly used in clinical practice since 1981.

SEVOFLURANE is a new inhalational anesthetic with the desirable properties of a low blood-gas partition coefficient and nonpungent character. Understanding the cardiovascular properties of a new anesthetic is important when evaluating its cardiovascular safety and when applying the agent to clinical situations in which cardiovascular function is an important consideration. The cardiovascular properties of sevoflurane in dogs have been shown to be similar to those of isoflurane, except that heart rate was lower with 1.2 MAC sevoflurane than with isoflurane. In surgical patients, heart rate was lower with sevoflurane than with isoflurane from 3–5 min before incision to 60 min after incision, but the cardiovascular properties of the agents were otherwise similar. .

In this article, we report the hemodynamic effects of sevoflurane administered to healthy male volunteers. We report the effects of sevoflurane administered in oxygen with ventilation controlled to maintain normocarbia and compare them with the hemodynamic effects of isoflurane administered under identical conditions. Previously studied volatile anesthetics have been associated with less cardiovascular depression during spontaneous than during controlled ventilation. We therefore compare the cardiovascular properties of sevoflurane during spontaneous ventilation with those during controlled ventilation. Similarly, studies of contemporary volatile anesthetics have shown that when each is compared to an equipotent mixture of volatile anesthetic and N²O, the latter produces less cardiovascular depression. Therefore, we examined the cardiovascular properties of sevoflurane administered in 60% N²O:40% Oxygen². Because studies of halothane, enflurane, and desflurane have demonstrated attenuation of cardiovascular effects or cardiovascular stimulation with time, we repeated cardiovascular measurements during controlled ventilation after 5.5 hr of anesthesia.

Twenty-one subjects were enrolled in a randomized (1:1:1), parallel trial. Using a computer-generated schedule 7 volunteers were randomized to receive sevoflurane in 100% Oxygen², 7 isoflurane in 100% Oxygen², and 7 sevoflurane in 60% N²O:40% Oxygen². Hemodynamic measurements were performed for each subject before anesthesia, at three anesthetic concentrations during controlled ventilation, at three anesthetic concentrations during spontaneous ventilation and at two anesthetic concentrations during controlled ventilation after 5.5 hr anesthesia.

After approval from The University of Arizona Human Subjects Committee, informed consent was obtained from 21 healthy, young male volunteers aged 19–34 years. They were of normal weight for height and were healthy by history and physical examination. Specific exclusion criteria included: evidence of cardiovascular, pulmonary, hepatic, or renal disease; use of chronic medications; personal or family history of unusual response to halogenated anesthetics; receipt of a general anesthetic within 6 months; previous use of investigational drugs; use of alcohol or medications within 7 days; and food or drink within 8 hr. Each subject had normal complete blood count, blood chemistry (including sodium, potassium, glucose, blood urea nitrogen, creatinine, aspartate aminotransferase, alanine aminotransferase, and alkaline phosphatase), urinalysis, electrocardiogram, and diagnostic echocardiogram.

Each volunteer was placed in the left lateral decubitus position. After 20 min rest, a left ventricular short axis two-dimensional echocardiogram was recorded at the mid-papillary muscle level using an echocardiography system (Model CFM 775, Vingmed Sound, Inc., Milpitas, CA) with a 5-MHz transthoracic echocardiography probe (Vingmed Sound, Inc.). The volunteer was then placed supine and peripheral intravenous, radial artery and pulmonary artery catheters were inserted using local anesthesia without sedation. The position of the pulmonary artery catheter was established by the pressure waveform. After 20 min rest, preanesthetic hemodynamic values were measured.

Hemodynamic pressures were measured using transducers (Model T 4812AD-R, Viggo-Spectramed, Oxnard, CA), which had been calibrated using a manometer, and a monitor (Spacelabs 90305 PC2 Bedside Monitor, Spacelabs, Inc., Redmond, WA) equipped with modules for electrocardiogram, invasive pressure, and cardiac output measurement. Transducers were zeroed at the mid-axillary line. Hemodynamic measurements were made at end-expiration. Pulmonary artery occlusion pressure was not measured for reasons of safety. Cardiac output was determined at end-expiration in triplicate by thermodilution, using a pulmonary artery thermodilution catheter (Edwards Model 93A-931H7.5F, Baxter Healthcare Corp., Irvine, CA) and room-temperature saline.

After induction of anesthesia and intubation of the trachea, a 5-MHz multiplane transesophageal echocardiography probe (Vingmed Sound, Inc.) was positioned in the esophagus at the mid-papillary muscle level and used to record subsequent two-dimensional echocardiograms. Echocardiographic measurements were made at end-expiration.

Customary formulas were used to calculate derived hemodynamic variables. A video analysis system (Vingmed Sound, Inc.) was used to measure left ventricular cross-sectional areas and circumferences. Area ejection fraction was calculated as described by Abel et al. Velocity of circumferential fiber shortening was calculated from two-dimensional measurements of left-ventricular circumference as described by Ruschaupt et al. and corrected for heart rate as described by Colan et al. Meridonal left ventricular end-systolic wall stress was calculated as described by Douglas et al. .

Anesthesia was induced by inhalation of the designated anesthetic during spontaneous ventilation. Tracheal intubation was facilitated with vecuronium (0.1 mg/kg). Ventilation was controlled to maintain normocarbia. Cardiovascular measurements were obtained at 1.0 minimum alveolar concentration (MAC), 1.5 MAC, and 2.0 MAC end-tidal anesthetic concentrations after the desired end-tidal anesthetic concentration had been maintained stable for 20 min. Measurements were made in order of increasing anesthetic concentration. If mean arterial pressure (MAP) decreased to less than 50 mmHg, anesthetic concentration was decreased by 0.5 MAC for reasons of safety. Minimum alveolar concentration of sevoflurane was taken as 2.05%, MAC of isoflurane was taken as 1.15%, and MAC of N²O was taken as 104%. Minimum alveolar concentration multiple of sevoflurane was assumed to be additive to that of nitrous oxide. End-tidal anesthetic concentrations were measured at the endotracheal tube orifice by infrared spectroscopy (Datex Capnomac Ultima, Datex Medical Instrumentation, Inc., Tewksbury, MA). Body temperature was maintained at normothermia with a forced-air warming device (Bair Hugger, Augustine Medical, Inc., Eden Prairie, MN). Intravenous fluids were infused as lactated Ringer's solution, using an estimate of basal metabolic rate of 42 kcal *symbol* m sup -2 *symbol* h sup -1 and assuming a water requirement of 1 ml/kcal expended. The calculated fluid deficit from overnight fasting was replaced before hemodynamic or echocardiographic measurements. Additional saline was administered at three times the amount of blood lost by sampling (approximately 150 ml). Additional fluid was administered if the left ventricular end-diastolic area, as measured using the video analysis system, decreased compared to values obtained after replacement of the calculated deficit and before induction of anesthesia.

A one-way analysis of variance was used to compare demographic and baseline variables of treatment groups. Cardiovascular data were analyzed to compare anesthetic effects between treatment groups, to examine effects of anesthetic in each treatment group, to compare spontaneous and controlled ventilation and to compare initial measurements with those after 5.5 hr of anesthesia. The change from preanesthesia to each intra-anesthetic measurement was calculated. At each evaluation time, these changes were compared among the three treatment groups with least significant difference tests within a one-way analysis of variance. The effects of anesthetics within each treatment group were tested using a pooled estimate of error from the analysis of variance. Repeated-measures analysis of variance could not be used because measurements could not be made for all subjects at all anesthetic concentrations. The differences between controlled and spontaneous ventilation or between early and late measurements were analyzed using a paired t test.

Of the 21 subjects enrolled in this protocol, 7 received sevoflurane in Oxygen², 7 sevoflurane in N²O:Oxygen², and 6 isoflurane in Oxygen². One subject randomized to receive isoflurane withdrew from the study prior to anesthetic induction.

The treatment groups were comparable with regard to age, weight, and height (Table 1). They were also comparable with regard to baseline values of cardiovascular measurements, except that subjects receiving sevoflurane had higher preanesthetic mean pulmonary artery pressure than did subjects receiving sevoflurane: N²O or isoflurane and heart rate was lower at baseline in subjects receiving sevoflurane:N sub 2 O than in subjects receiving sevoflurane.

In the 7 subjects receiving sevoflurane during controlled ventilation, MAP was greater or equal to 50 mmHg during 1.0 and 1.5 MAC anesthetic, whereas in 5 of 7 MAP was greater or equal to 50 mmHg during 2.0 MAC sevoflurane. All subjects receiving sevoflurane in N²O tolerated 1.0 and 1.5 MAC, while 6 of 7 tolerated 2.0 MAC. All subjects receiving isoflurane tolerated 1.0 MAC, 4 of 6 tolerated 1.5 MAC, and in 3 of 6 MAP was greater or equal to 50 mmHg with 2.0 MAC. When the desired MAC level could not be maintained, data for that subject at that MAC multiple were not included in the analysis.

After replacement of the calculated fluid deficit from overnight fasting, left ventricular end-diastolic area was maintained constant by administration of additional fluid, if necessary. The mean total amount of fluid administered, including replacement of the calculated deficit, through the controlled ventilation phase of the protocol (data shown in Figure 1, Figure 2, Figure 3) was 1857 plus/minus 181 ml for sevoflurane, 2015 plus/minus 304 ml for isoflurane, and 1757 plus/minus 137 ml for sevoflurane:60% N²O.

The effects of sevoflurane on selected hemodynamic variables are shown in F1-4. Sevoflurane did not alter heart rate compared to awake values. Sevoflurane decreased MAP, mean pulmonary artery pressure, and stroke volume index at all anesthetic concentrations. Cardiac index decreased compared to awake measurements at 1.0 and 1.5 MAC, but not at 2.0 MAC where systemic vascular resistance decreased. Central venous pressure decreased slightly compared to awake values at 1.5 MAC, but not at 1.0 or 2.0 MAC. Mixed venous hemoglobin saturation increased at all anesthetic concentrations, while total body oxygen consumption decreased at all anesthetic concentrations. Arterial base-excess was unchanged. Sevoflurane did not alter left ventricular area ejection fraction, or velocity of circumferential fiber shortening (F2-4). Systolic wall stress decreased at all anesthetic concentrations. Left ventricular end-diastolic area was unchanged, as dictated by the experimental protocol.

The effects of isoflurane on cardiovascular variables are shown in F1-4and F2-4. Isoflurane increased heart rate at 1.5 and 2.0 MAC. Cardiac index decreased compared to awake measurements at 1.0 MAC, but not at 1.5 and 2.0 MAC where systemic vascular resistance decreased. Mean arterial pressure, stroke volume index, and systolic wall stress decreased at all anesthetic concentrations. Mixed venous hemoglobin saturation increased at all isoflurane concentrations, while total body oxygen consumption decreased at 1.5 and 2.0 MAC. Other measured variables were unchanged.

When treatment groups were compared directly, sevoflurane caused a greater decrease in mean pulmonary artery pressure at 1.0 and 1.5 MAC than did isoflurane. No other differences between treatment groups were noted.

The cardiovascular effects of sevoflurane administered in 60% N²O:40% O²during controlled ventilation were compared directly to those of sevoflurane in 100% Oxygen²(F1-4and F2-4). Mean arterial pressure decreased less at 1.0 MAC and mean pulmonary artery pressure decreased less at 1.0 and 1.5 MAC with sevoflurane:N²O than with sevoflurane. Mixed venous oxyhemoglobin saturation increased less with sevoflurane:N²O than with sevoflurane at all anesthetic concentrations.

During sevoflurane, sevoflurane:N²O, or isoflurane anesthesia in spontaneously ventilating subjects, arterial partial pressure of carbon dioxide increased compared to preanesthetic values (F3-4). At each MAC multiple, PaCO²values did not differ among treatment groups.

All subjects receiving sevoflurane during spontaneous ventilation tolerated 1.0 and 1.5 MAC, while in 1 of 7 subjects MAP fell below 50 mM Hg at 2.0 MAC. Heart rate and stroke volume index were higher during spontaneous than during controlled ventilation at 1.0 and 1.5 MAC (Table 2). Mean arterial pressure was higher during spontaneous ventilation at 1.0 MAC. Cardiac index was higher and systemic vascular resistance lower during spontaneous ventilation at all anesthetic concentrations. Velocity of circumferential fiber shortening was higher during spontaneous ventilation at 1.5 MAC.

All subjects receiving sevoflurane after 5.5 hr of anesthesia tolerated 1.0 MAC, while 3 of 7 subjects developed intolerable hypotension (MAP < 50 mmHg) at 2.0 MAC. At 1.0 MAC heart rate, cardiac index, stroke volume index, and mixed venous oxygen saturation were higher after 5.5 hr of sevoflurane anesthesia than during the first 150 min of anesthesia (Table 3). Systemic vascular resistance was less in the later period at both 1.0 and 2.0 MAC.

We evaluated the hemodynamic effects of sevoflurane in healthy volunteers not undergoing surgery. The use of volunteers allows detailed measurement of cardiovascular variables under controlled conditions without confounding factors common in surgical patients.

Cardiovascular measurements could not be performed on some subjects receiving high anesthetic concentrations because of unacceptable hypotension. These data were not available for inclusion in the statistical analysis. This introduces a bias in the analysis of the response of MAP and possibly of other cardiovascular variables to 1.5 or 2.0 MAC anesthetic. This bias is significant because data from subjects with the greatest hemodynamic response to anesthetic were not available for analysis. In addition to creating a possible bias in averaged data, inability to include data from some subjects decreases statistical power and may cause cardiovascular effects of higher anesthetic concentrations to go undetected. It was possible, however, to collect data for all subjects at 1.0 MAC sevoflurane, sevoflurane:N²O, and isoflurane. Therefore, comparisons are valid at this anesthetic concentration. It is likely that the effects of isoflurane on MAP at higher anesthetic concentrations were underestimated to an equal or greater extent than were those of sevoflurane and sevoflurane:N²O because a greater number of measurements were unable to be performed during isoflurane anesthesia than during anesthesia induced by the other agents. The effects on other hemodynamic variables of the loss of data from some subjects at higher anesthetic concentrations cannot be accurately predicted.

We administered anesthetics in increasing concentrations so that if adverse hemodynamic events occurred at lower concentrations, higher concentrations with more adverse consequences would not be attempted. This also raises a possible confounding factor (the effect of anesthetic duration) when different anesthetic concentrations are compared.

Examination of our data reveals some apparent physiologic inconsistencies. For example, left ventricular end-diastolic volume and left ventricular area ejection fraction were unchanged, whereas stroke volume decreased. Also, at 2.0 MAC sevoflurane, cardiac index and heart rate were unchanged, whereas stroke volume decreased. This is probably because changes in some related variables did not achieve statistical significance owing to the small number of subjects enrolled. In addition, two-dimensional echocardiographic indices of ventricular volume are likely to be relatively insensitive to small changes in volume. Therefore, caution should be used in drawing detailed physiologic conclusions from our data. These data are more appropriately used descriptively to anticipate the changes in the same variables when patients are anesthetized with these agents.

Systematic administration of fluid is important when evaluating the cardiovascular effects of a medication, because fluid administration can affect preload. Preload can, in turn, affect other measures of cardiovascular performance. In this study, we administered fluid based on metabolic formulas. The limitation of this approach is that it does not account for variation in urine output or in insensible losses. We minimized respiratory losses by humidifying inspired gasses; however we did not measure urine output and were unable to measure other insensible losses. We therefore administered additional fluid to maintain constant left ventricular end-diastolic volume, our best measure of left ventricular preload. Our results therefore most accurately apply under conditions of constant preload.

Our results, which are measured in healthy, young, non-premedicated male subjects not undergoing surgery have limitations when applied to surgical patients. Cardiovascular disease, age, gender, physical conditioning, concomitant medications, and surgical stimulation potentially affect cardiovascular responses to sevoflurane. However, because cardiovascular properties of the anesthetic are a significant determinant of cardiovascular responses during surgery, understanding these properties enables the clinician to anticipate possible cardiovascular responses to the variety of clinical situations faced. Most importantly, the detailed cardiovascular responses to sevoflurane are nearly identical to those of isoflurane, an anesthetic with which most clinicians are familiar.

The decrease in cardiac index or systemic vascular resistance, combined with the unchanged filling pressures and heart rate observed with sevoflurane indicates decreased myocardial contractility. The pattern of unchanged echocardiographic ejection-phase indices of left ventricular function (left ventricular area ejection fraction and velocity of circumferential fiber shortening) with unchanged preload (left ventricular end-diastolic area) and decreased afterload (systolic wall stress) also indicates decreased myocardial contractility. The ejection-phase indices would be expected to increase in the face of decreasing afterload if contractility were maintained. Sevoflurane-induced myocardial depression is consistent with the work of Pagel et al., who showed that sevoflurane administered to chronically instrumented dogs decreases preload-recruitable stroke work slope, a reliable index of myocardial contractility. In addition, isoflurane, an anesthetic known to decrease contractility, produced a similar pattern of changes in preload, afterload, cardiac index, and echocardiographic indices of contractility.

In our comparative study, the cardiovascular effects of sevoflurane were qualitatively and quantitatively similar to those of isoflurane. When cardiovascular measurements during anesthesia were compared to baseline values, some differences between anesthetics were suggested. However, when the effects of sevoflurane and isoflurane were compared directly, sevoflurane differed from isoflurane only in that mean pulmonary artery pressure decreased at 1.0 and 1.5 MAC sevoflurane but not with equipotent concentrations of isoflurane. This may have been because preanesthetic mean pulmonary artery pressure was higher in subjects receiving sevoflurane than in those receiving isoflurane.

Our measurements during isoflurane anesthesia correspond well to those of Stevens et al. Four differences were noted, however. First, in our study, heart rate increased only at 1.5 and 2.0 MAC, not at 1.0 MAC as observed by Stevens et al. Second, in our study, systemic vascular resistance was maintained at 1.0 MAC and decreased only at higher end-tidal isoflurane concentrations. Third, our data show isoflurane's effect on cardiac index to be biphasic; in subjects with MAP greater or equal to greater or equal to 50 mmHg at all anesthetic concentrations, cardiac index decreased at 1.0 MAC, but returned to preanesthetic values at higher isoflurane concentrations as systemic vascular resistance decreased. Fourth, the increased right atrial pressure observed by Stevens et al. was not found. The reasons for these differences are not clear. The subjects of Stevens et al. displayed a higher baseline systemic vascular resistance, possibly contributing to the observed differences in this variable between the two studies. It is possible that the subjects of Stevens et al. were relatively hypervolemic compared to our subjects, leading to increased right atrial pressure. Stevens et al. did not describe the protocol used for fluid administration, whereas our echocardiographic data demonstrate constant preload.

The cardiovascular effects of sevoflurane are similar to those of previously studied volatile anesthetics, with a few exceptions. During sevoflurane anesthesia, cardiac index is maintained as systemic vascular resistance decreases, whereas during halothane anesthesia, cardiac index falls as systemic vascular resistance is maintained. The response of cardiac index to sevoflurane differed from that with halothane, enflurane, and desflurane. With sevoflurane, in subjects who tolerated all three anesthetic concentrations, cardiac index decreased at 1.0 and 1.5 MAC, but returned to preanesthetic values at 2.0 MAC as systemic vascular resistance fell. In this study, we observed this biphasic response of cardiac index with isoflurane as well, suggesting methodologic differences between our protocol and those of previous investigators. Sevoflurane also differed from previously studied volatile anesthetics in that central venous pressure did not increase. Echocardiographic measurements during sevoflurane anesthesia yielded similar results to those obtained during desflurane anesthesia by Weiskopf et al. .

Sevoflurane administered in 60% N²O:40% Oxygen²caused a smaller decrease in MAP at 1.0 MAC than did an equipotent concentration of sevoflurane in 100% Oxygen². Better maintenance of arterial pressure with N²O mixtures has also been observed with previously studied volatile anesthetics. Although neither cardiac index nor systemic vascular resistance independently was higher with sevoflurane:N sub 2 O than with sevoflurane alone, both tended to be higher at 1 MAC sevoflurane:N²O than at 1 MAC sevoflurane and may have contributed to better maintenance of MAP. Of previously studied volatile anesthetics, isoflurane:N²O and desflurane:N²O caused higher systemic vascular resistance, while halothane:N²O and enflurane:N²O caused higher cardiac index than did the volatile anesthetics alone. Sevoflurane:N²O decreased mean pulmonary artery pressure less than did sevoflurane alone. No difference in mean pulmonary artery pressure was noted with desflurane:N²O compared to desflurane alone. The increased pulmonary artery pressures observed with desflurane:N²O, were not observed with sevoflurane:N²O.

Differences between sevoflurane:N²O and sevoflurane alone may have been due to the specific cardiovascular properties of N²O or may have been caused when sevoflurane was replaced in part by an agent with fewer cardiovascular effects. Our study design does not allow us to distinguish between these possibilities. Forty percent N²O alone, when administered to healthy volunteers, did not affect MAP. It did increase total peripheral resistance, one factor that may have contributed to the observed increase in MAP when N²O was added to sevoflurane. Forty percent N²O alone did cause a decrease in heart rate and in central venous pressure. We did not observe these changes when sevoflurane:N²O was compared to sevoflurane alone.

Sevoflurane in 60% N²O:40% Oxygen²resulted in lower values of mixed venous oxyhemoglobin saturation than did sevoflurane in 100% Oxygen². This may have resulted from a lower inspired oxygen concentration and lower arterial oxygen content leading to decreased venous oxygen content. Even during sevoflurane:N²O anesthesia this index was higher than before anesthesia.

Kikura and Ikeda studied the cardiovascular properties of sevoflurane with 60% N²O in surgical patients. In contrast to our results, they found no increase in heart rate compared to preanesthetic values, although they did find a decrease in blood pressure at all anesthetic concentrations. Systolic wall stress decreased from the lowest (0.9 MAC) to the highest (1.6 MAC) sevoflurane concentration used. Velocity of circumferential fiber shortening, fractional area change, and end-diastolic area did not change with anesthetic concentration. It is not possible to determine changes in echocardiographic variables from the preanesthetic state, because preanesthetic values were not measured in the group of patients studied during anesthesia.

The cardiovascular effects of sevoflurane administered with 60% N²O are similar enough to those of equipotent concentrations of sevoflurane that the decision to add nitrous oxide to sevoflurane should be based on factors other than hemodynamic considerations.

We observed diminished cardiovascular depression when sevoflurane was administered during spontaneous ventilation compared to controlled ventilation. The cardiovascular effects of sevoflurane administered during spontaneous ventilation are likely to be a combination of direct anesthetic effects and effects of anesthetic-induced hypercapnia. In addition, comparison of cardiovascular effects of anesthetics in normocapneic and hypercapneic subjects is typically complicated by the hemodynamic effects of mechanical ventilation. Our study design does not allow separation of these confounding effects. The cardiovascular effects of sevoflurane during spontaneous ventilation are generally similar to those during normocarbic controlled ventilation, suggesting that direct anesthetic effects play a major role in determining cardiovascular performance in both conditions.

Many of the quantitative differences between the cardiovascular effects of sevoflurane administered during spontaneous ventilation and those during mechanical ventilation may be due to the effects of hypercarbia. In conscious, healthy, mechanically ventilated volunteers, hypercarbia increased heart rate, cardiac index, and stroke volume. Hypercarbia decreased systemic vascular resistance. These results parallel the observed differences in cardiovascular properties of sevoflurane during controlled and spontaneous ventilation.

Weiskopf et al. suggested that spontaneous ventilation improves the safety of inhaled anesthetic administration, because the concentration of previously studied volatile anesthetics producing cardiovascular collapse exceeds that producing apnea. Spontaneous ventilation during sevoflurane anesthesia results in higher MAP, cardiac index, and stroke volume index than does controlled ventilation. This diminished cardiovascular depression may result in increased cardiovascular safety.

We observed evidence of cardiovascular stimulation during prolonged sevoflurane anesthesia. This was manifested by increased heart rate and cardiac index and decreased systemic vascular resistance and stroke volume index at 1.0 MAC after 5.5 hr of sevoflurane compared to the earlier period. At 2.0 MAC, the only difference noted with prolonged anesthesia was that systemic vascular resistance was lower during the later period. This pattern of changes was similar to those observed with halothane, enflurane, and desflurane. The most notable difference in our study compared to the referenced studies was a decrease in stroke volume index with prolonged anesthesia, whereas stroke volume was unchanged over time with desflurane and increased over time with 1% halothane and 1 MAC enflurane. We cannot exclude the possibility that the hemodynamic changes noted with prolonged anesthesia occurred because the later period followed a period of spontaneous ventilation in which changes in heart rate, cardiac index, systemic vascular resistance, and stroke volume index were also noted.

Sevoflurane was well tolerated by healthy volunteers at 1.0 and 1.5 MAC end-tidal anesthetic concentrations. Hemodynamic instability (MAP < 50 mmHg) was noted in some subjects at 2.0 MAC end-tidal concentrations in the absence of surgical stimulation. The incidence of intolerable hypotension with sevoflurane was similar to that seen with isoflurane. At 2.0 MAC, in subjects with MAP greater or equal to 50 mmHg no adverse cardiovascular effects of sevoflurane were noted. When administered with 60% N²O, sevoflurane caused a smaller decrease in MAP than did sevoflurane alone. Spontaneous ventilation during sevoflurane anesthesia or prolonged sevoflurane administration resulted in diminished cardiovascular depression. The cardiovascular effects of sevoflurane are similar to those of isoflurane, an anesthetic commonly used in clinical practice since 1981.

The authors thank Burnell R. Brown, Jr., M.D., Ph.D., F.R.C.A., and Robert G. Merin, M.D., for advice and editorial assistance. They also thank Tailiang Xie, Ph.D., for advice regarding statistical analysis, and Suzanne Causbie, for administrative and clerical assistance.