Hypoxia has a dual effect on ventilation: an initial period of hyperventilation, the acute hypoxic response, is followed after 3-5 min by a slow decline, the hypoxic ventilatory decline. Because of hypoxic ventilatory decline, subsequent acute hypoxic responses are depressed. In this study, the influence of a sedative concentration of propofol on ventilation was studied if hypoxia was sustained and intermittent.
Ten healthy young male volunteers performed two hypoxic tests without and with a target controlled infusion of propofol. The sustained hypoxic test consisted of 15 min of isocapnic hypoxia followed by 2 min of normoxia and 3 min of hypoxia. The test of hypoxic pulses involved six subsequent exposures to 3 min hypoxia followed by 2 min of normoxia. The bispectral index of the electroencephalogram was measured to obtain an objective measure of sedation.
Blood propofol concentrations varied among subjects but were stable over time (mean blood concentration 0.6 microg/ml). The sustained hypoxic test showed that propofol decreased acute hypoxic response by approximately 50% and that the magnitude of hypoxic ventilatory decline relative to acute hypoxic response was increased by > 50%. Propofol increased the depression of the acute hypoxic response after 15 min of hypoxia by approximately 25%. In control and propofol studies, no hypoxic ventilatory decline was generated during exposure to hypoxic pulses. The bispectral index-acute hypoxic response data suggest that subjects were either awake (with minimal effect on acute hypoxic response) or sedated (with 50-60% reduction of acute hypoxic response).
The depression of acute hypoxic response results from an effect of propofol at peripheral or central sites involved in respiratory control or secondary to the induction of sedation or hypnosis by propofol. The relative increase in hypoxic ventilatory decline is possibly related to propofol's action at the gamma-aminobutyric acid A (GABA(A)) receptor complex, causing increased GABAergic inhibition of ventilation during sustained (but not intermittent) hypoxia.
DURING monitored anesthesia care, in which anesthetics or analgesics are used as adjuvant to regional anesthesia, ventilatory control is predominantly metabolic or chemical in nature. Because hypoxia (especially episodic or intermittent hypoxia) is frequently associated with monitored anesthesia care, 1,2we investigated the influence of propofol, a popular hypnotic for sedation during such care, on the ventilatory responses to acute, sustained, and episodic isocapnic hypoxia.
Hypoxia has a dual effect on the ventilatory control system. A short episode of hypoxia (duration < 5 min) causes an increase in the hypoxic drive from the peripheral chemoreceptors of the carotid bodies. 3Hypoxia of longer duration causes depression of the respiratory centers in the brain stem. 3,4The net effect of these opposing phenomena is that the ventilatory response to sustained hypoxia is biphasic: An initial period of hyperventilation, the acute hypoxic response (AHR), is followed within 3 to 5 min by a slow hypoxic ventilatory decline (HVD). A steady state in ventilation (V̇i) is obtained after 15 to 20 min. 5,6The mechanism of the hypoxic depression of V̇i(HVD) is unknown. During moderate hypoxia (oxygen saturation 80–90%) the accumulation and release of inhibitory neurotransmitters or modulators, such as γ-aminobutyric acid (GABA) or adenosine, is thought to play a major role in the development of HVD. 5–9A consequence of the hypoxia-related central depression is that the recovery of the hypoxic response is not immediate (i.e. , after prolonged hypoxia subsequent hypoxic responses remain depressed). 5,6
Recent studies indicate that several general anesthetics, including propofol, interact with or modulate the GABAAreceptor complex. 10–12At clinical concentrations, propofol enhances GABA-evoked chloride currents and causes direct activation of the receptor in the absence of GABA. 11,12GABAAreceptors are thought to be involved in the generation of HVD, 3–6,8and hence there may be an important role for propofol in modulating or enhancing HVD.
Clinically, hypoxia is often episodic or periodic, especially during sleep or sedation. 2,13We therefore further examined the interaction of propofol and periodic hypoxia on V̇iand the development of HVD. Apart from assessing ventilatory responses, we measured heart rate responses and the bispectral index (BIS) of the electroencephalogram (EEG). The BIS informs us on the central nervous system arousal state of the participants, which is relevant because it may be an important factor in the study outcome.
Subjects and Methods
Ten healthy male volunteers (aged 18–25 yr) participated in the protocol after approval was obtained from the Leiden University Medical Center Human Ethics Committee. The subjects were healthy and did not have histories of tobacco or illicit drug use. They were instructed not to eat or drink for at least 8 h before the study.
After arrival at the laboratory, two intravenous catheters were inserted in the left and right cubital veins (one for propofol administration and one for blood sampling). Subsequently standard electrodes for EEG measurement (BisSensor, Aspect Medical Systems, Natick, MA) were placed and the subjects rested for 20 to 30 min. Next a face mask was applied over the mouth and nose. Subjects were in a comfortable semirecumbent position. The inspired and expired gas flows were measured with a pneumotachograph connected to a pressure transducer and electronically integrated to yield a volume signal. Corrections were made for the changes in gas viscosity resulting from changes in oxygen concentration of the inhaled gas mixtures. The pneumotachograph was connected to a T piece. One arm of the T piece received a gas mixture from a gas mixing system consisting of three mass-flow controllers (Bronkhorst High-Tec, Veenendaal, The Netherlands). A personal computer provided control signals to the mass-flow controllers so that the composition of the inspired gas mixtures could be adjusted to force end-tidal pressures of oxygen (PETO2) and carbon dioxide (PETCO2) to follow a specified pattern in time. 14The inspired and expired gas concentrations and the arterial hemoglobin–oxygen saturation (SpO2) were measured with a Datex Multicap gas monitor and Datex Satellite Plus pulse oximeter, respectively (Datex-Engstrom, Helsinki, Finland). The EEG and electromyelogram were recorded using an Aspect (Natick, MA) A-2000 EEG monitor. The monitor computed the BIS over 4-s epochs. We averaged the BIS values over 1-min intervals.
Two different hypoxic tests (sustained hypoxia and hypoxic pulses) were performed without and with propofol. Control studies preceded propofol studies. The order of hypoxic tests was randomized. Between tests there was ample time for resting. Control and drug hypoxic studies were performed at identical PETCO2s, 5–7 mmHg above awake resting values. This was done to balance an effect of the increase in PETCO2resulting from depression of V̇iby propofol.
Sustained Hypoxic Test.
The PETO2was forced as follows: (1) 10 min at 110 mmHg, (2) a rapid decrease to 50 mmHg, (3) 15 min at 50 mmHg, (4) a rapid increase to 110 mmHg, (5) 2 min at 110 mmHg, (6) a rapid decrease to 50 mmHg, (7) 3 min at 50 mmHg, and (8) at least 5 min at more than 300 mmHg.
The PETO2waveform was as follows: (1) 10 min at 110 mmHg, (2) a rapid decrease to 50 mmHg, (3) 3 min at 50 mmHg, (4) a rapid increase to 110 mmHg, and (5) 2 min at 110 mmHg. The hypoxic–normoxic sequence (steps 2–5) was repeated five times. This procedure yields six 3-min hypoxic pulses separated by 2 min of normoxia.
Propofol Administration, Sampling, and Assay.
A Psion (London, United Kingdom) palm-top computer programmed with a three-compartment propofol pharmacokinetic data set 15was used to control a Becton Dickinson infusion pump (St. Etienne, France) for the intravenous administration of propofol. The propofol target concentration was set at 1 μg/ml and was kept constant during both propofol hypoxic studies. After BIS values had reached a steady level, but at least 20 min after the target had been reached, the hypoxic studies started. Propofol blood samples (5 ml) were obtained 5 min before the first hypoxic study (T1), at the end of the first hypoxic study (T2) and at the end of the second hypoxic study (T3). The samples were collected in syringes containing potassium oxalate. The propofol concentrations were determined by reverse-phase high-performance liquid chromatography. 16
Sustained Hypoxic Test.
Mean values of the breath-to-breath data were chosen over identical time segments (see fig. 1). Period N is the 1-min period before the 15-min of hypoxia, period H1the third minute of hypoxia, period H2the 15th minute of hypoxia, and period H3the third minute of the second hypoxic bout. Differences in V̇ibetween periods N and H1were defined as the first AHR (AHR1), between periods N and H2as the sustained hypoxic response, and between periods N and H3as the second AHR (AHR2). The difference between AHR1and sustained hypoxic response was used as measure of the HVD. The V̇iresponses are expressed as the change in V̇iper percentage change in SpO2(unit: l · min−1·%−1).
We tested the occurrence of HVD by comparing the hypoxic response to the first hypoxic pulse with the response to the last hypoxic pulse. In order to do so, mean values of the breath-to-breath data of the last minute of normoxia before the first hypoxic pulse (period A) and the third minute of the first (period B) and last hypoxic pulse (period C) were calculated. Differences in V̇ibetween periods A and B were defined as the first AHR (AHRfirst), and between periods A and C as the last AHR (AHRlast).
A two-way analysis of variance was performed on the different periods (N, H1, H2, and H3) of the sustained hypoxic test. Because peak heart rate responses occurred at period H1+ 3 min (see Results), for hypoxic heart rate sensitivities a comparison was made among periods H1+ 3 min, H2, and H3. Differences between periods were tested with the Student–Newman–Keuls test. A paired t test was performed to compare ventilatory and heart rate responses to the first and sixth hypoxic pulses. To detect the significance of difference between the control and propofol studies, a paired t test was performed on individual parameters of the hypoxic studies. P values < 0.05 were considered significant. All values are mean ± SD.
Propofol Concentrations and Bispectral Index Values
Blood propofol concentrations and BIS values varied considerably among subjects but were constant over time within subjects. Over time, blood propofol concentrations were 0.61 ± 0.30 μg/ml at T1, 0.56 ± 0.23 μg/ml at T2, and 0.58 ± 0.23 μg/ml at T3(not significant). Mean coefficient of variation over time was 14.6% (range, 7–34%). Because of technical problems, the collection of BIS data was not achieved in one subject. Averaged control BIS values were 97 ± 2 and 96 ± 2 for the sustained hypoxic and hypoxic pulses studies, respectively. During propofol, the mean BIS values were 76 ± 14 for the sustained hypoxic study (range among subjects, 53–91; mean coefficient of variation over time, 7%), and 76 ± 10 for the hypoxic pulses study (range among subjects, 66–91; mean coefficient of variation over time, 8%).
Sustained Hypoxic Test.
Propofol reduced normoxic baseline V̇iby approximately 15% (P < 0.001;table 1). In figure 1, examples of a control and propofol hypoxic study of one subject are shown. In all subjects, in both control and propofol studies, the hypoxic ventilatory responses were biphasic and the recovery of the hypoxic response was not immediate (figs. 1 and 2, table 1). Propofol decreased AHR1by 50% from 1.74 ± 1.22 to 0.89 ± 0.70 l · min−1·%−1(P < 0.001), the sustained hypoxic response by 60% from 1.11 ± 0.80 to 0.33 ± 0.30 l · min−1·%−1(P = 0.002), and AHR2by 60% from 1.04 ± 0.48 to 0.39 ± 0.33 l · min−1·%−1(P < 0.001). The absolute magnitude of HVD did not differ between control and propofol (0.63 ± 0.51 vs . 0.56 ± 0.52 l · min−1·%−1, not significant). Propofol increased the ratio HVD/AHR1by > 50%, from 0.35 ± 0.16 to 0.54 ± 0.16 (P = 0.02) and caused more depression of the second hypoxic response: The ratio AHR2/AHR1was 0.67 ± 0.16 for control and 0.51 ± 0.22 for propofol (P < 0.05).
BIS values and control of end-tidal gas concentrations are listed in table 2. In figure 3, examples of a control and propofol study of one subject are shown. In the control study, V̇iin periods B and C increased to 197 ± 78% and 200 ± 63% of baseline, respectively. Corresponding values in the propofol study were 154 ± 27% (period B) and 144 ± 37% (period C). In control and propofol studies AHRfirstdid not differ from AHRlast: control, 1.35 ± 0.84 versus 1.35 ± 0.67 l · min−1·%−1(not significant;fig. 4); propofol, 0.64 ± 0.39 versus 0.58 ± 0.25 l · min−1·%−1(not significant).
Heart Rate Responses
Because of technical problems, the collection of heart rate data failed in one subject. Propofol decreased normoxic heart rate by 8–10 beats/min (table 1). In control and propofol sustained hypoxic studies, peak heart rate responses occurred at period H1+ 3 min (fig. 2). Control heart rate sensitivity decreased from its peak by 20% in period H2(not significant) and by 35% in period H3(P < 0.05). Propofol heart rate sensitivity decreased from its peak by 21% in period H2(P < 0.05) and by 38% in period H3(P < 0.05). Compared with control, propofol decreased hypoxic heart rate sensitivities in periods H1+ 3 min, H2, and H3by 27 ± 36%, 36 ± 22%, and 38 ± 34%, respectively (not significant). Heart rate responses during the hypoxic pulses test did not differ between the first and sixth hypoxic pulse and between control and propofol studies.
Influence of Propofol on the Ventilatory Response to Acute Hypoxia
In this study we observed that, in a healthy, young group of male volunteers, propofol, at a sedative concentration (mean BIS value 76), caused a ∼50% reduction of the ventilatory response to acute hypoxia. Two previous studies examined the influence of propofol on the hypoxic ventilatory response. Blouin et al. 17tested the influence of propofol at a blood concentration of 2 μg/ml on a ramp hypoxic test. They observed an 80% depression of the ventilatory response in eight male volunteers. However, because the ramp hypoxic test is a mixture of AHR and HVD, 18their finding is best compared with our observation of propofol-induced reduction of the sustained hypoxic response by about 60%. Nagyova et al. 19showed a reduction of the AHR by 21, 23, and 60% at respective propofol blood concentrations of 0.5, 1, and 2 μg/ml. This is less than our finding of 50% depression at 0.6 μg/ml. However, in the study by Nagyova et al. , differences between the level of consciousness in awake and drug studies were minimized by encouraging the subjects to watch television. There is now ample evidence that this induces behavioral control of breathing. As a consequence, additional (nonchemical) drives increase responses mediated via the carotid bodies. 20
Our study gives little information on the site of action of propofol with respect to its effects on normoxic resting ventilation and AHR. Propofol may affect ventilatory control at four possible sites: at the peripheral or central chemoreceptors; at the respiratory centers in the brain stem; at the neuromechanical link between brain stem and ventilation (this includes spinal motoneurons, respiratory muscles, lungs, and airways); at sites in the central nervous system involved in behavioral-state control (for example the brain stem reticular system). In cats and rabbits, Ponte and Sadler 21showed that a high-dose propofol infusion (18–35 mg/kg · h) into the carotid artery abolished chemoreceptor discharge in normoxia and prevented an increase during hypoxia. In humans, Dow and Goodman 22showed that during propofol anesthesia, sudden exposure to hyperoxia reduces V̇i. This suggests that the peripheral chemoreceptors remain active during propofol anesthesia but does not exclude some depression. In anesthetized cats, an inhibitory effect of propofol on cells in the dorsomedial and ventrolateral medulla mediating pressor effects has been observed. 23Both sites are involved in respiratory control and may contain central chemoreceptors. 24We suggest that, in our study, propofol decreased AHR by depression of the peripheral chemoreflex loop, either directly by enhancing GABAergic inhibition of respiratory centers within the brain stem or indirectly by changing the behavioral state of the subjects. At this time, we are unable to exclude an effect of propofol at the peripheral chemoreceptors.
The EEG data give some evidence for an effect of propofol on ventilatory control via sites involved in behavioral-state control. We used the BIS of the EEG as measure of the state of sedation or hypnosis. Among subjects we observed a large variability in blood propofol concentrations and BIS values. We relate this to the relatively large interindividual variability in pharmacokinetics and dynamics of intravenous anesthetics. For the BIS, we relate this further to the occurrence, at least in some subjects, of natural sleep on top of drug-induced sedation. Non–rapid eye movement and slow-wave sleep cause further decreases in BIS. 25In figure 5, we plotted the BIS values of individual subjects (x -axis) versus the depression of the AHR (y -axis, AHR as percentage of control and obtained from both hypoxic studies). The data suggest the existence of two (behavioral and respiratory) states in our subjects. Subjects were either awake (control studies [open squares in the right part of fig. 5] and propofol studies with little to no effect on ARH [open circles in the right part of fig. 5]) or sedated or asleep with 50–60% depression of AHR (closed circles in the right part of fig. 5). In the latter group, variations in the level of sedation or hypnosis (as measured by the BIS) had little further effect on AHR.
Influence of Propofol on HVD
In humans, the mechanism of the secondary decline in V̇iduring hypoxia of longer duration (> 3 min) is still under debate. The following observations with respect to HVD are of importance to our study and may provide some insight into the mechanism of HVD:
1. Moderate hypoxia causes an increase in cerebral blood flow (CBF). 26Because of the increase in CBF, the carbon dioxide tension within the central nervous system and hence the central ventilatory drive are decreased. As a consequence a significant part of HVD may be related to an increase in CBF.
2. When HVD develops, its influence on V̇idoes not dissipate immediately upon the termination of hypoxia. Subsequent hypoxic responses remain depressed (ratio AHR2/AHR1< 1), and 1 h of air breathing is necessary for a full recovery of AHR (AHR2/AHR1= 1). 5,6These observations have led to the hypothesis that the accumulation or release of neurotransmitters or modulators with net inhibitory influences on V̇iis quantitatively the more important contributor to HVD. The most important neurotransmitter candidates are GABA and adenosine. 3–6,8,9On the termination of hypoxia, their turnover is not immediate, and therefore the depressant effects of hypoxia wane slowly. Animal studies give further proof for a role of central inhibition from GABA. Bicuculline, a GABAAreceptor antagonist, reverses HVD in the anesthetized cat. 9
3. Among and within subjects, the ratio HVD/AHR1is constant with changing values of AHR1. 5,6This is explained by a modulatory effect of the hypoxic drive of the peripheral chemoreceptors on the central (i.e. , within the central nervous system) development of HVD. 6
In our study, propofol did not affect the absolute magnitude of HVD. However, relative to AHR1, HVD did increase, causing an increase in ratio HVD/AHR1. Furthermore, the ratio AHR2/AHR1decreased during propofol infusion. Because propofol concentrations and BIS values were constant over time, we argue that these observations indicate that propofol enhanced the development of HVD. Recent studies indicate that propofol acts at GABAAreceptors and enhances GABA-ergic transmission. 11,12,27We therefore suggest that the relative increase in HVD is related to propofol’s action at the GABAAreceptor complex. Other GABAAreceptor agonists such as midazolam cause similar effects on the ratio HVD/AHR1. 8Although these observations support a role for GABAergic neurons in the inhibition of brain stem respiratory centers during sustained hypoxia, they do not exclude the involvement of other receptor systems such as adenosine and opioid receptors. For example, Cartwright et al. 28showed an increase of the ratio HVD/AHR1by the μ receptor agonist alfentanil. Aminophylline, an adenosine-receptor blocker, reduces the ratio HVD/AHR1. 7These findings illustrate that HVD development is modulated by various agents commonly used in anesthesia. Studies in humans on the influence of propofol on CBF indicate that propofol reduces CBF concomitant with a reduction in cerebral metabolic rate and maintains the cerebrovascular response to changes in arterial PCO2. 29Although no studies on the influence of propofol on changes in arterial oxygen partial pressure are available, it seems improbable that propofol increased the CBF response to hypoxia.
In contrast to 15 min of sustained hypoxia, 15 min of intermittent hypoxia, without and with propofol, were unable to generate HVD. Because in our studies the level of hypoxia was relatively mild (inspired oxygen concentration 10%), a limited number of hypoxic pulses were applied, and the duration of the pulses was relatively short, we are unable to predict the impact of repeated episodes of deeper, longer, or more frequent periods of hypoxia on V̇i. Only limited data are available on ventilatory changes during repeated hypoxia. We recently studied the influence of 0.25% sevoflurane on the the ventilatory response to four to six 3-min hypoxic episodes (SpO284%) separated by 2 min of normoxia in humans. 30Hypoxic V̇iremained constant over time. In this respect, the behavior of sevoflurane is identical to propofol. In anesthetized pigs, repetition of 12 min of deep hypoxia (inspired oxygen concentration, 7%) was associated with cardiorespiratory depression during the second hypoxic episode. 31Further animal studies are required to study the interaction of repetitive (deep) hypoxia and anesthetics and hypnotics on V̇iand hemodynamics.
Influence of Propofol on the Heart Rate Responses to Hypoxia
The heart rate response to acute hypoxia depends on the effects of hypoxia at various sites in the body. Decreased heart rate may result from a direct effect of hypoxia on the sinoatrial node or hypoxic stimulation of the carotid bodies resulting in stimulation of the vagal centers in the medulla; increased heart rate may result from activation of aortic chemoreceptors, lung receptors (stimulated via an increase in V̇i), or sites within the central nervous system. These factors interact in a complex fashion. 31In normal breathing subjects, mild to moderate hypoxia is generally associated with tachycardia. Bradycardia is often observed if hypoxia coincides with apnea or at deeper levels of hypoxia. 32,33
In both control and propofol studies, the peak heart rate response to hypoxia occurred 6 min after the initiation of hypoxia (fig. 2). The reason for the relatively slow dynamics of the heart rate response (compared with the V̇iresponse) is unknown but may be related to the complex interaction of the various components responsible for the heart rate response, all with their own dynamics. The results of our study give little evidence for an important role of GABA-mediated hypoxic depression of centers in the brain stem involved in heart rate control. In contrast to the V̇idata, propofol did not enhance heart rate decline during sustained hypoxia or increase depression of the heart rate response to a hypoxic episode after 15 min of hypoxia. Furthermore, the decline of heart rate responses in period H2or H3(compared with period H1+ 3 min) may be caused by the decrease in V̇i. Further studies are needed (for example a protocol that allows for a fixed V̇ilevel) to elucidate the mechanism by which propofol affects heart rate responses to hypoxia.
It is not appropriate to simply extrapolate our data (obtained in healthy young male subjects, using imposed isohypercapnic hypoxic stimuli [mean SpO2, 86%], without the occurrence of obstructive apnea) to patients under monitored anesthesia care. Studies in patients with sleep apnea indicate that if hypoxia is related to obstructive apnea, initial bradycardia is followed, upon the relief of apnea, by brisk tachycardic responses. 34,35In light of the clinical importance of this issue, 36further studies are warranted to examine the influence of sedatives and analgesics on the chronotropic response to hypoxia, especially if upper-airway obstruction is involved.
The authors thank I. Olievier for her help with the experiments and F. Engbers, M.D., for providing the TCI computer program.