British Journal of Anaesthesia 1992; 68: 360-364

HEART RATE PERIODICITIES DURING INDUCTION OF PROPOFOL-NITROUS OXIDE-ISOFLURANE ANAESTHESIA D. C. GALLETLY, T. CORFIATIS, A. M. WESTENBERG AND B. J. ROBINSON SUMMARY This study examined the variation in cardiac interbeat interval during induction of anaesthesia with propofol and subsequent inhalation anaesthesia with nitrous oxide and isoflurane. In comparison with preoperative control values, heart rate variability was reduced by anaesthesia and there was a complex, but consistent, pattern of R-R interval change during induction. Immediately after propofol 2 mg kg~\ high frequency heart rate oscillations were replaced by low frequency oscillations (0.05 Hz). Subsequently, with decreasing propofol and deepening nitrous oxide-isoflurane anaesthesia, high frequency components returned towards, although not reaching, control values; low and, to a greater extent, mid range (0.1-Hz) frequency components remained reduced. We postulate that these alterations are consistent with an immediate but transient post-induction ventilator/ depression, subsequent decrease in efferent sympathetic activity and reduction in baroreflex responsiveness. R-R interval analysis is suggested to be a useful tool in the evaluation of autonomic function during anaesthesia. KEY WORDS Anaesthetic techniques: induction. Anaesthetics, intravenous: propofol. Anaesthetics, volatile: isoflurane. Heart: R-R interval analysis.

The main determinants of heart rate are the sympathetic and parasympathetic efferent input to the atria! sinus pacemaker. Because this autonomic control is the product of complex feedback oscillations within the cardiovascular pressure control system, and because of interacting periodicities caused by ventilation, autonomic activity at the sino-atrial node is changing constantly. In consequence, heart rate shows a marked beat-to-beat variability. Although the full extent of this variability is unlikely to be appreciated from the averaged heart rate values given by standard operating room ECG monitoring systems, it is readily apparent if individual R—R intervals are plotted against time (fig. 1). In the past 5—10 years, important advances have increased our understanding of heart rate variability (HRV) and its possible clinical applications. Early time domain analyses (measures such as the mean or variance of HRV) have been supplemented by frequency domain analyses by which periodicities

within the R-R interval series are examined quantitatively using fast Fourier [1] and autoregressive methods of spectral analysis [2], and the interactions between different physiological variables are examined using transfer function analysis [3]. As a consequence, it has been demonstrated that discrete Fourier spectral components of R-R interval time series reflect different aspects of cardiovascular and autonomic control. At least three spectral components have been identified: two low frequency fluctuations (< 0.15 Hz) reflecting aspects of arterial pressure control and a superimposed high frequency " ventilatory " component in phase with ventilation [4]. During anaesthesia, autonomic function is influenced by the effects of drugs and surgical stimulation in addition to changes in posture, temperature and blood volume. It would seem probable therefore that in addition to the gross alterations in heart rate which are apparent from standard ECG monitoring systems, there may exist more subtle fluctuations in R-R interval variability characteristic of different techniques of anaesthesia and which reflect the changing state of autonomic and cardiorespiratory activity. In this study we have examined the effects of a standard general anaesthetic technique on die spectral components of R-R interval variability and attempted to explain how such variability may be caused by known autonomic effects of the anaesthetic agents used. PATIENTS AND METHODS

We studied 10 fasting, unpremedicated ASA I patients (mean age 26.4 yr, range 18-35 yr; three male) undergoing urological or gynaecological surgery under general anaesthesia. None had any evidence of cardiac or respiratory pathology and none was taking regular medications. The study was approved by the regional Ethics Committee and all patients gave written informed consent. One hour before operation, a continuous 5-min ECG (CM5) recording was made with the subject supine and resting quietly. The voltage output from die ECG monitor (Hewlett-Packard) was passed to a purpose built R wave detector which generated an D . C. GALLETLY, F.F.A.R.A.C.S., F.C.ANAES.J A. M . WESTENBERG, B.SC. ; B. J. ROBINSON, M.SC.; Section of Anaesthesia, Wellington

School of Medicine, Private Bag Wellington, New Zealand. T. CORFIATIS, M.sc., Department of Physics, Victoria University,

Wellington. Accepted for Publication: November 1, 1991.

HEART RATE PERIODICITY

361

100 -

1 Propofol 2 mg kg

-

90 _ T

1.5 % isoflurane + nitrous oxide

- llU

80

_c

E 70

-° fif)

1 min

1 min

50 -

-

40 -

Control

1 min Induction

Post-induction

FIG. 1. Heart rate (HR) variability during the preanaesthetic (control), induction and post-induction periods.

accurate 50-ms pulse synchronous with the peak of each R wave. The R wave was detected by half wave rectifying the electronically filtered (high pass to remove P and T waves) and differentiated ECG signal and passing this output to a peak detector which generated the 50-ms pulse when the slope of the signal changed from positive to negative. This pulse was input to the parallel port of a microcomputer (IBM-compatible) which computed the interval between successive R-R intervals and stored these as a series for later analysis. Anaesthesia was induced with propofol 2 mg kg"1 over 15 s into an antecubital vein. A facemask was applied and patients breathed 66% nitrous oxide and 1.5 % isoflurane in oxygen via a circle absorption system. Transient apnoea was managed, if necessary, by gentle IPPV in order to maintain an oxygen saturation of > 95 %. An R-R interval series was recorded throughout the induction sequence and for 15 min after induction, before the onset of surgical stimulation. Episodes of apnoea, IPPV and patient movement were recorded. Spectral analysis of HRV was performed using a method based on that of Akselrod and colleagues [5,6]. The raw R-R series were examined for artefacts and any discrete, abnormal intervals caused by ectopic beats were removed using linear interpolation. Artefact-free, stationary segments of 128 s of R-R interval data were then taken from the preoperative recording and from the final minutes of the post-induction recording while patients were breathing nitrous oxide and isoflurane. These segments were sampled at a rate of 4 Hz to produce a series of 512 (4x 128 s) discrete values of instantaneous heart rate. This heart rate series was then high pass filtered to remove fluctuations less than 0.015 Hz and low pass filtered at 2 Hz to remove components of greater than Nyquist frequency. A fast Fourier analysis was performed using a Harming window. The total spectral power (area under the curve) was normalized for the square of the heart rate and the proportion of this power calculated for three frequency ranges, chosen as representative of previously published work [1,5,7-9] because individual authors quote slightly different values for

each frequency component: low frequency ("vasomotor"), 0.02-0.08 Hz; mid frequency ("baroreflex"), 0.08-0.15 Hz; high frequency ("ventilatory") 0.15-0.45 Hz. The SD of the instantaneous heart rate series and total power were calculated as a measure of overall HRV. Differences in spectral power for the three frequency bands were analysed using Student's t test for paired data. Differences were considered significant at P < 0.05. Analysis was performed using a Statview II statistical package on an Apple Macintosh Ilex microcomputer. RESULTS

Visual examination of the R-R interval recordings showed a consistent pattern of HRV during the induction sequence. At least three phases were apparent (fig. 1). (1) Preinduction variability was replaced by a transient smooth acceleration in heart rate. The mean time to onset of this tachycardia was 22 (SEM 1.8) s after the end of propofol injection, with peak heart rate occurring at mean 35 (2.3) s. Heart rate then returned rapidly towards preinduction value. (2) Over the next approximately 3 min, the most frequent pattern was of a slow heart rate fluctuation. In six patients in whom this periodicity could be measured accurately over five or six peaks, the mean frequency was 0.056 Hz (range 0.049-0.059 Hz). (3) These low frequency fluctuations in heart rate became less apparent as the high frequency (ventilatory) components returned and increased in amplitude over 5-10 min. A comparison of spectral components during the resting preoperative period with that 10 min after induction is shown in table I. During anaesthesia, although heart rate was unchanged, heart rate variability decreased. This reduction in HRV was seen as a decrease in total spectral power ( — 59%) and, to a lesser extent, as a reduction in R-R interval SD (— 21 %). The power in each of the three frequency bands decreased, although there was a greater reduction in the power of low and more especially the

BRITISH JOURNAL OF ANAESTHESIA

362

TABLE I. Mean (SEM) values of HRV data before operation and 10-15 min after the induction of propofol—nitrous oxide-•isoflurane anaesthesia. Absolute power is given in arbitrary units. Proportional power is calculated as absolute power for each band divided by total power

1

HR (beat min" ) HR SD (beat min-1) Total power Low band power Mid band power Low + medium (absolute) power High band power Ratio abs. power (Hi (Lo + Med)) Proportion of low band power Proportion of mid band power Proportion of high band power

1.2-

1.2 -i

Control

Pre-anaesthesia

Anaesthesia

Change (%)

67.20 (2.78) 4.05 (0.94) 0.260 (0.058) 0.049 (0.008) 0.055 (0.01) 0.10(0.02)

67.87 (2.29) 3.21 (0.57) 0.107(0.018) 0.019 (0.002) 0.014(0.002) 0.03 (0.01)

+1 -21 -59 -61 -74 -70

ns ns 0.02 0.005 0.0018 0.002

0.157(0.045) 1.4(0.24)

0.073 (0.016) 2.1 (0.34)

-53 + 50

0.07 0.05

0.22 (0.03)

0.217(0.028)

1

0.235 (0.025)

0.134(0.01)

-43

0.005

0.545 (0.044)

0.648 (0.032)

+ 19

0.04

Anaesthesia

I

biti ary unit

V)

0.9-

0.9 -

0.6-

0.6 -

0.3-

0.3 -

Power

3-

0-0

0.1 0.2 0.3 Frequency (Hz)

0.4

0

0.1 0.2 0.3 0.4 Frequency (Hz)

FIG. 2. Power spectrum of heart interval variability during the preoperative period (control) and during nitrous oxide-isoflurane anaesthesia.

mid range frequency components when compared with the high frequency band. The proportion of the total spectral power within the high frequency band increased by 19 %, the mid band decreased by 43 % and the proportion of the power in the low band was unchanged. During anaesthesia, alteration in the morphology of the high frequency spectral component was apparent also, with the ventilatory peak becoming narrower and more clearly denned (fig. 2). This was considered to reflect a decrease in the variability of ventilatory rhythm during inhalation anaesthesia. DISCUSSION

We have observed that the recording of R—R interval series during the induction of anaesthesia revealed a complex but consistent pattern of heart rate fluctuations which are presumed to represent the interaction between the changing plasma concentrations of the anaesthetic agents with cardiovascular and ventilatory control mechanisms. Although a comparison of preoperative with intra-anaesthetic R-R interval series showed no significant difference when time domain analysis (heart rate and SD) were performed, differences were revealed clearly using spectral analysis.

P

ns

In the normal human subject, at least three spectral components of HRV are described [1, 10]: (1) A low frequency component (centred between 0.04 and 0.08 Hz), is said to represent thermoregulatory fluctuations in vasomotor tone [11] or compensatory baroreflex oscillations in heart rate, for arterial pressure fluctuations at the same frequency, which are caused by peripheral vasomotor variability, and which are normally damped by the renin—angiotensin system [5]. The cardiac efferent component is mediated by parasympathetic and beta-sympathetic activity and may be enhanced (or "undamped") by angiotensin converting enzyme inhibitors [1,5]. (2) Mid range frequencies centred as 0.12 Hz (the "10-s rhythm" or the Mayer waves of arterial pressure [12]) have been ascribed to the activity of a 10-s brain stem oscillator and modulating effect of the baroreflex or its frequent response [13, 14]. In the supine position, the efferent control is predominantly parasympathetic, while a greater sympathetic component is present with standing or during tilt [10, 15, 16]. (3) High frequency components, in phase with ventilation, are ascribed to oscillating baroreflex responsiveness during breathing [17], a central coupling of respiratory generator and cardiac vagal efferents, a central vagal reflex arc stimulated by pulmonary stretch receptor activation and, since small fluctuations still occur in denervated heart transplants and IPPV, changes in right ventricular preload and myocardial wall stretch [18]. The ventilatory component is thought to be controlled by parasympathetic vagal efferents (the sympathetic nervous system is too sluggish to operate at this high frequency) and, apart from small mechanically mediated fluctuations, is abolished by atropine [15, 16]. Donchin, Feld and Porges [19], examining the effect of isoflurane-nitrous oxide anaesthesia on high frequency (0.12-0.4 Hz) parasympathetic spectral components, observed a diminution of the ventilatory heart rate variability and related this to a reduction in vagal tone. These authors postulated that changes in high frequency variability might be

HEART RATE PERIODICITY used as a measure of depth of anaesthesia, although subsequent work from Yli-Hankala and colleagues [20] found no relationship between anaesthetic depth and the ventilatory component of HRV for more than 1 MAC of isoflurane and enflurane. Two major effects of propofol-isoflurane-nitrous oxide anaesthesia on HRV were observed in the present study. First, immediately after induction with propofol, high frequency "ventilatory" fluctuations were almost abolished, being replaced by a low (0.056-Hz) frequency heart rate periodicity. Second, after approximately 3 min, and presumably as the isoflurane—nitrous oxide inhalation anaesthesia deepened and the initial effects of propofol decreased, the fast ventilatory component assumed an increasing dominance over the low and, more especially, the mid range frequencies; however, its absolute power returned only to approximately 50 % of the preoperative control value. In comparison with the unanaesthetized state, there was an overall reduction in HRV, seen as a reduction in total spectral power. The initial acceleration in heart rate after propofol occurred at approximately the same time as the onset of CNS depression and within one arm—brain circulation time. This acceleration may be a result of a parasympathetic, sympathetic or direct sino-atrial node action. Subsequent, low frequency waves at 0.05 Hz (with a period of 18 s) are too fast to be caused by recirculation and therefore are most likely to be the low frequency components of the cardiovascular control system. The dominance of these low frequency waves immediately following propofol could come about through a number of mechanisms. First, it is probable that the "ventilatory" high frequencies are so markedly depressed that normal background low frequency components are unmasked. Such high frequency depression may have been caused by: (1) A parasympatholytic effect of propofol reducing vagal efferent activity. Although this is supported by animal studies, in which a dose-dependent reduction in the vagal component of the baroreflex arc has been demonstrated [21] and would explain the observed initial tachycardia, in humans such an effect of propofol has not been shown [22]. Furthermore, at times when low frequencies dominate the Fourier spectrum in the period immediately after induction, the heart rate in some patients was little different from preinduction values. (2) Reduced ventilatory depth or effort after propofol injection. Ventilatory depression is a transient but well described feature of propofol induction. In conscious subjects, a reduction in tidal volume is associated with a reduction in high frequency spectral energy and may, through a complex dynamic interaction, cause a corresponding increase in lower frequency components [18, 23, 24]. (3) Peripheral vascular dilatation or increased capacitance [25-27] from propofol would be expected to reduce venous return and reduce the activity of atrial and arterial stretch receptors. This reduces the parasympathetic dominance of sino-atrial control and thereby reduces the parasympathetic ventilatory component of HRV.

363

In addition to depression of ventilatory fluctuations, it is also possible that the low frequency waves are increased by a sympathetically mediated response to sudden changes in capacitance induced by propofol [27]. Such an effect is observed in dogs during infusion of nitroglycerine [28] or haemorrhage [29]. These observations might, therefore, be consistent with a peripheral vascular action of propofol [26]. During subsequent isoflurane-nitrous oxide inhalation, the power in all frequency bands was reduced compared with preoperative values. However, as a proportion of total spectral power, high frequency fluctuations dominated over the mid frequencies and, to a lesser extent, over low frequency vasomotor activity. In addition to a return of normal ventilatory depth, the following mechanisms could help explain these changes: (1) The reduction in low and mid range frequencies during isoflurane anaesthesia might reflect both a relative decrease in adrenergic activity from a preoperative stressed condition and an absolute decrease in normal background sympathetic tone as anaesthesia deepened (isoflurane-nitrous oxide anaesthesia reduces efferent sympathetic activity by 56% at 1.0 MAC [30]). (2) Low and mid range frequencies are under the control of parasympathetic and sympathetic efferents; a reduction in vagal efferent activity (consistent with the ability of isoflurane to increase heart rate) could help explain the reduced low and mid range power and explain the smaller reduction in absolute power of the high frequency component. (3) Low and mid range frequency bands are baroreflex mediated. As isoflurane, as with all general anaesthetics, depresses baroreflex activity (30% depression at 1.0 MAC [31]), a corresponding reduction in these spectral components would be expected. Despite their diminution, however, at least one component (0.05-Hz) remains active throughout the induction sequence and subsequent inhalation anaesthesia. This is consistent with other studies which have shown the presence of intact baroreflex sympathetic control under both isoflurane and propofol anaesthesia [21, 30, 32]. Because the baroreflex frequency (0.1 Hz) is depressed to a considerably greater extent than vasomotor fluctuations at 0.05 Hz, it is possible that the inhalation anaesthetic agents or residual effects of the propofol are in some way depressing the source of the "10-s" baroreflex oscillator, while leaving intact the baroreflex mechanism itself. This preliminary study, examining the brief period of anaesthetic induction and deepening inhalation anaesthesia cannot reveal in detail the mechanisms that underlie the observed complexities of HRV. Under more steady state conditions, however, and with additional recording of ventilation and other cardiovascular indices, it is possible that HRV analysis could be a useful tool in the haemodynamic investigation of anaesthetic drugs. Previous reports on the cardiovascular actions of propofol have shown inconsistent changes for heart rate in the first few minutes after bolus injection [23]. This is

BRITISH JOURNAL OF ANAESTHESIA

364

explained readily by examining the complex fluctuations in HRV seen in figure 1, by researchers' use of ECG heart rate recorders which average multiple R-R intervals and by the standard research procedure of averaging heart rates from the group of patients under study. Except for the period immediately after injection of propofol, clear differences between pre- and intra-anaesthetic heart rate series (which had the same average heart rate) were apparent only when spectral analysis was performed. Averaged heart rate measurements are, therefore, only an approximation to the underlying complexity of instantaneous heart rate. We therefore suggest that HRV spectral decomposition may be a valuable quantitative tool for the analysis of autonomic fluctuations in awake and anaesthetized patients. ACKNOWLEDGEMENTS Mr T. Corfiatis is supported by grants from the Wellington Medical Research Foundation, the Kelvin Day Trust and the Wellington Anaesthesia Trust. Mr Westenberg was supported by a New Zealand Medical Council Summer Studentship. Equipment was purchased from a New Zealand Lotteries Board Grant.

REFERENCES 1. Aksclrod S, Gordon D, Ubel F, Shannon D, Barger A, Cohen R. Power spectrum analysis of heart rate fluctuation: a quantitative probe of beat to beat cardiovascular control. Science 1981; 213: 220-222. 2. Basclli G, Cerutti S, Civardi S, Lombardi F, MaHiani A, Merri M, Pagani M, Rizzo G. Heart rate variability signal processing: a quantitative approach as an aid to diagnosis in cardiovascular pathologies. International Journal of BioMedical Computing 1987; 20: 51-70. 3. Appel M, Berger R, Saul J, Smith J, Cohen R. Beat to beat variability in cardiovascular variables: Noise or music? Journal of the American College of Cardiology 1989; 15: 1139-1148. 4. Akselrod S. Spectral analysis of fluctuations in cardiovascular parameters: a quantitative tool for the investigation of autonomic control. Trends in Pharmacological Science 1988; 9: 6-9. 5. Akselrod S, Gordon D, Madwed J, Snidman N, Shannon D, Cohen R. Hemodynamic regulation: investigation by spectral analysis. American Journal of Physiology 1985; 249: H867-875. 6. Berger R, Akselrod S, Gordon D, Cohen R. An efficient algorithm for spectral analysis of heart rate variability. IEEE Transaction on Biomedical Engineering 1986; 33: 900-904. 7. Parati G, Castiglioni P, Di Rienzo M, Omboni S, Pedotti A, Mancia G. Sequential spectral analysis of 24 hour blood pressure and pulse interval in humans. Hypertension 1990; 16: 414-^21. 8. Weise F, Hcydcnrcich F, Runge U. Heart rate fluctuations in diabetic patients with cardiac vagal dysfunction: A spectral analysis. Diabetic Medicine 1988; 5: 324-327. 9. Saul J, Arai Y, Berger R, Lilly L, Colucci W, Cohen R. Assessment of autonomic regulation in chronic congestive heart failure by heart rate spectra] analysis. American Journal of Cardiology 1988; 61: 1292-1299. 10. Pagani M, Lombardi F, Guzzetti S, Rimoldi O, Furlan R, Pizzinelli P, Sandrone G, Malfatto G, Dell'Orto S, Piccaluga E, Turiel M, Baselli G, Cerutti S, Malliani A. Power spectral analysis of heart rate and arterial pressure variabilities as a marker of sympatho—vagal interaction in man and conscious dog. Circulation Research 1986; 59: 178-193.

11. Kitney RI. Entrainmcnt of the human R-R interval by thermal stimuli. Journal of Physiology (London) 1975; 252: 37P-38P. 12. Penaz J. Mayer waves: History and methodology. Automedica 1978; 2: 135-141. 13. Zweiner U. Pathophysiologie und Funktionsdiagnostik der nervosen Herzkreislaufregulation. Ergebmsse ExperimenteUe Medizin 1985; 46: 90-105. 14. Kitney R. An analysis of the non-linear behaviour of the human thermal vasomotor control system. Journal of Theoretical Biology 1975; 52: 231-248. 15. Pomeranz B, Macaulay R, Candill M, Kutz I, Adam D, Gordon D, Kilborn K, Barger A, Shannon D, Cohen R, Benson H. Assessment of autonomic function in humans by heart rate spectral analysis. American Journal of Physiology 1985; 248: H151-153. 16. Weise F, Heydenreich F, Runge U. Contributions of sympathetic and vagal mechanisms to the genesis of heart rate fluctuations during orthostatic load: a spectral analysis. Journal of the Autonomic Nervous System 1987; 21: 127-134. 17. Eckbcrg D, Kifle Y, Roberts D. Phase relationship between normal human respiration and baroreflex responsiveness. Journal of Physiology (London) 1989; 304: 489-502. 18. Bcrnardi L, Keller F, Sanders M, Reddy P. Griffith B, Meno F, Pinsky M. Respiratory sinus arrhythmia in the denervated human heart. Journal of Applied Physiology 1989; 67; 1447-1455. 19. Donchin Y, Fcld JM, Porges SW. Respiratory sinus arrhythmia during recovery from isoflurane-nitrous oxide anesthesia. Anesthesia and Analgesia 1985; 64: 811-815. 20. Yli-Hankala A, Porkkala T, Kaukinen S, Hakkinen V, Jantti V. Respiratory sinus arrhythmia is reversed during positive pressure ventilation. Acta Physiologica Scandinavica 1991; 141: 399-407. 21. Blake D, Jover B, McGrath B. Haemodynamic and heart rate reflex responses to propofol in the rabbit. British Journal of Anaesthesia 1988; 61: 194-199. 22. Cullen P, Turtle M, Prys-Roberts C, Way W. Effect of propofol anesthesia on baroreflex activity in humans. Anesthesia and Analgesia 1987; 66: 1115-1120. 23. Kitney R, Byrne S, Edmonds M, Watkins P, Roberts V. Heart rate variability in the assessment of autonomic diabetic neuropathy. Automedica 1982; 4: 155—167. 24. Kitney RI. A non-linear model for studying oscillation in the blood pressure control system. Journal of Biomedical Engineering 1979; 1: 89-99. 25. Grounds RM, Twigley A, Carli F, Whitwam J, Morgan M. The haemodynamic effects of intravenous induction. Comparison of the effects of thiopentone and propofol. Anaesthesia 1985; 40: 735-740. 26. Claeys M, Gepts E, Camu F. Haemodynamic changes during anaesthesia induced and maintained with propofol. British Journal of Anaesthesia 1988; 60: 3-9. 27. Goodchild C, Serrao J. Cardiovascular effects of propofol in the anaesthetized dog. British Journal of Anaesthesia 1989; 63: 87-92. 28. Rimoldi O, Pierini S, Ferrari A, Cerutti S, Pagani M, Malliani A. Analysis of short-term oscillations of R—R and arterial pressure in conscious dogs. American Journal of Physiology 1990; 258: H967-976. 29. Madwed J, Albrccht P, Mark R, Cohen R. Low frequency oscillations in arterial pressure and heart rate: a simple computer model. American Journal of Physiology 1989; 256: H1573-H1579. 30. Sellgren J, Ponten J, Wallin G. Percutaneous recording of muscle nerve sympathetic activity during propofol, nitrous oxide, and isoflurane anesthesia in humans. Anesthesiology 1990; 73: 20-27. 31. Kotrly KJ, Ebert T, Vucins E, Igler FO, Barney J, Kampinc J. Baroreceptor reflex control of heart rate during isoflurane anesthesia in humans. Anesthesiology 1984; 60: 173-179. 32. Stevens W, Cromwell T, Halsey M, Eger E, Shakespeare T, Banlman S. The cardiovascular effects of a new inhalational anesthetic, Forane, in human volunteers at constant arterial carbon dioxide tension. Anesthesiology 1971; 35: 8-16.