Portal de Periódicos da CAPES

Influence of oral tramadol on the dynamic ventilatory response to carbon dioxide in healthy volunteers

2001; Elsevier BV; Volume: 87; Issue: 6 Linguagem: Inglês

10.1093/bja/87.6.860

1471-6771

Diederik Nieuwenhuijs, John Bruce, Gordon B. Drummond, P M Warren, Albert Dahan,

Heart Rate Variability and Autonomic Control

We tested the effect of tramadol on ventilatory control by quantifying its effect on the steady-state ventilatory carbon dioxide response and by locating its site of respiratory action within the ventilatory control system. We imposed square-wave changes in end-tidal carbon dioxide (∼1 kPa; end-tidal oxygen concentration kept constant at resting levels) in 10 healthy volunteers (six men, four women) before and after oral ingestion of 100 mg tramadol, and measured the ventilatory responses. Each hypercapnic response was separated into a fast, peripheral and a slow, central component. Two control and two tramadol carbon dioxide studies were performed in each subject. Tramadol reduced the total ventilatory carbon dioxide sensitivity by ∼30% from 12.8 (6.1) [lower (25%) and upper (75%) quartiles 7.4 and 16.6 litre min−1 kPa−1] to 9.1 (5.3) (5.3–14.1) litre min−1 kPa−1 (P<0.001). The fast and slow response gains were reduced by 23 (46) (3–54)% (P<0.05) and 30 (22) (15–54)% (P<0.01) respectively. The ratio of these carbon dioxide sensitivities and the apnoeic threshold were not significantly changed by tramadol. We suggest that tramadol affects the ventilatory control system by acting at the μ-opioid receptors in the respiratory integrating centres within the brainstem. We tested the effect of tramadol on ventilatory control by quantifying its effect on the steady-state ventilatory carbon dioxide response and by locating its site of respiratory action within the ventilatory control system. We imposed square-wave changes in end-tidal carbon dioxide (∼1 kPa; end-tidal oxygen concentration kept constant at resting levels) in 10 healthy volunteers (six men, four women) before and after oral ingestion of 100 mg tramadol, and measured the ventilatory responses. Each hypercapnic response was separated into a fast, peripheral and a slow, central component. Two control and two tramadol carbon dioxide studies were performed in each subject. Tramadol reduced the total ventilatory carbon dioxide sensitivity by ∼30% from 12.8 (6.1) [lower (25%) and upper (75%) quartiles 7.4 and 16.6 litre min−1 kPa−1] to 9.1 (5.3) (5.3–14.1) litre min−1 kPa−1 (P<0.001). The fast and slow response gains were reduced by 23 (46) (3–54)% (P<0.05) and 30 (22) (15–54)% (P<0.01) respectively. The ratio of these carbon dioxide sensitivities and the apnoeic threshold were not significantly changed by tramadol. We suggest that tramadol affects the ventilatory control system by acting at the μ-opioid receptors in the respiratory integrating centres within the brainstem. Tramadol is an analgesic with putative opioid and non-opioid modes of action.1Raffa RB Fridriechs E Reimann W et al.Opioid and nonopioid components independently contribute to the mechanism of cation of tramadol, an atypical opioid analgesic.J Pharmacol Exp Ther. 1992; 260: 275-285PubMed Google Scholar The respiratory effects of tramadol are not clear: some clinical studies indicate little or no respiratory depression2Houmes RJM Voets MA Verkaaik A Erdmann W Lachmann B. Efficacy and safety of tramadol versus morphine for moderate and severe postoperative pain with special regard to respiratory depression.Anesth Analg. 1992; 74: 510-514Crossref PubMed Scopus (262) Google Scholar, 3Tarkkila P Tuominen M Lindgren L. Comparison of respiratory effects of tramadol and pethidine.Eur J Anaesth. 1998; 15: 64-68Crossref PubMed Scopus (57) Google Scholar, 4Warren PM Taylor JH Nicholson KE Wraith PK Drummond GB. Influence of tramadol on the ventilatory response to hypoxia in humans.Br J Anaesth. 2000; 85: 211-216Crossref PubMed Scopus (25) Google Scholar but others find significant respiratory effects.5Seitz W Lubbe N Fritz K Sybrecht G Kirchner E. Eiffub von tramadol auf die ventilatorische CO2 -antwort un den mundokklusionsdruck.Anaesthesist. 1985; 34: 241-246PubMed Google Scholar 6Vickers MD Flaherty DO Szekely SM Read M Yoshizumi J. Tramadol: pain relief by an opioid without depression of respiration.Anaesthesia. 1992; 47: 291-296Crossref PubMed Scopus (360) Google Scholar Several case reports suggest that the respiratory depressant effect of tramadol is usually underestimated.7Rothe KF Brather R. Postoperative Atemdepressionen im Zusammenhang mit Tramadol-infusionsnarkose.Anaesthesist. 1983; 32: 88PubMed Google Scholar 8Barnung SK Treschow M Borgbjerg FM. Respiratory depression following oral tramadol in a patient with impaired renal function.Pain. 1997; 71: 111-112Abstract Full Text Full Text PDF PubMed Scopus (55) Google Scholar These differences could be caused by the different methods used to test ventilatory control, the state of arousal of the subjects or the doses and routes of administration. We set out to quantify the effect of oral tramadol on ventilatory control in healthy volunteers and to locate its site of action within the ventilatory control system. In order to do so, we used the dynamic end-tidal forcing technique.9Bellville JW Whipp BJ Kaufman RD Swanson GD Aqleh KA Wiberg DM. Central and peripheral chemoreflex loop gain in normal and carotid body-resected subjects.J Appl Physiol. 1979; 46: 843-853PubMed Google Scholar 10Dahan A DeGoede J Berkenbosch A Olivier ICW. Influence of oxygen on the ventilatory response to carbon dioxide in man.J Physiol (Lond). 1990; 428: 485-499Crossref Scopus (128) Google Scholar This technique measures the steady-state ventilatory carbon dioxide sensitivity and also estimates the relative contributions of the peripheral and central chemoreflex gains. We applied square-wave changes in end-tidal carbon dioxide concentration and divided the ventilatory response (measured on a breath-to-breath basis) into a fast, peripheral dynamic component and a slow, central component, using an empirical two-compartment model of the ventilatory controller. This mathematical model has been validated in cats.11DeGoede J Berkenbosch A Ward DS Bellville JW Olivier CN. Comparison of chemoreflex gains obtained with two different methods in cats.J Appl Physiol. 1985; 59: 170-179PubMed Google Scholar It has been used in humans to assess the effect of opioids,12Sarton E Teppema L Dahan A. Sex differences in morphine-induced ventilatory depression reside within the peripheral chemoreflex loop.Anesthesiology. 1999; 90: 1329-1338Crossref PubMed Scopus (93) Google Scholar anaesthetics13Dahan A van den Elsen MJLJ Berkenbosch A et al.Effects of subanesthetic halothane on the ventilatory responses to hypercapnia and acute hypoxia in healthy volunteers.Anesthesiology. 1994; 80: 727-738Crossref PubMed Scopus (76) Google Scholar and catecholamines14Ward DS Bellville JW. Effect of intravenous dopamine on hypercapnic ventilatory response in humans.J Appl Physiol. 1983; 55: 1418-1425PubMed Google Scholar on ventilatory control. Twelve volunteers (aged 22–56 yr; seven men) participated in the study after approval had been obtained from the local human ethics committee. The subjects were healthy and did not smoke tobacco. Subjects were asked not to take caffeine-containing drinks for 8 h beforehand, and fasted for at least 4 h before the studies. They were studied in a semirecumbent position in a well-lit room and listened to music via headphones. On the study day, after arrival at the laboratory, electrodes for ECG (Hewlett-Packard 78351A) and EEG electrodes (BisSensor; Aspect Medical Systems, Newton, MA, USA) were placed on the thorax and head respectively. Next, the subjects rested for 10–20 min. The subjects breathed through a facemask (Vital Signs, Totowa, NJ, USA). Expiratory gas flows were measured with a pneumotachograph (Fleisch no. 2) and a pressure transducer (Furness Micomanometer), and the signal was integrated electronically to obtain volume. The inspired gas mixture was set using mass-flow controllers (Bronkhorst High-Tec, Veenendaal, The Netherlands) controlled by a personal computer (Elonex PT-5120/1). This allows the forcing of end-tidal Po2 (Pe′o2) and end-tidal Pco2 (Pe′co2) according to a specified pattern in time. The inspired and expired oxygen and carbon dioxide concentrations were measured near the mouth with a mass spectrometer (VG Spectralab M, Winsford, UK) and the arterial haemoglobin oxygen saturation (Spo2) with a pulse oximeter (Ohmeda Biox 3700, Ohmeda, Helsinki, Finland) set to give a rapid response. End-tidal oxygen and carbon dioxide partial pressures, tidal volume, inspiratory time (Ti), expiratory time (Te), breathing frequency (f=60/[Ti+Te]), expired minute ventilation ( V˙e=f×Vt) and SpO2 were collected and stored on disk for further analysis. The EEG was recorded using an A-2000 EEG monitor (Aspect Medical Systems; software version 3.3). The monitor computed the bispectral index (BIS) over 5-s epochs. We averaged the BIS values over 1-min intervals. Oxygen consumption (litre min−1 standard temperature and pressure, dry (STPD) and carbon dioxide output (litre min−1 STPD) were measured from collections of mixed expired gas made over a 2-min period, and the gas exchange ratio was calculated. Concentrations of oxygen and carbon dioxide were measured using a Servomex oxygen analyser (model 570A, Servomex, Norwood, MA, USA) calibrated with air and 100% nitrogen, and carbon dioxide with a Datex analyser (Normocap 200, Datex, Helsinki, Finland) calibrated with four calibration gas mixtures. The experiments consisted of normoxic steps into and out of hypercapnia. After a period of steady-state breathing (assessed by stable ventilation) with Pe′co2 raised 0.1–0.2 kPa above resting values, Pe′co2 was increased by ∼1 kPa in a stepwise fashion and kept constant for 7 min. Subsequently, Pe′co2 was returned to its original value and kept constant for another 7 min. During the experiment, Pe′o2 was kept constant at resting values. In each subject, two control studies and two tramadol studies were performed. Control runs preceded the drug runs. The drug runs were started 30 min after the subject had taken 100 mg tramadol as two 50 mg tablets (Zydol; Searle). Females were studied within the first 10 days of a normal menses to ensure that they were not pregnant and to avoid any effect of progesterone on ventilation. The data were analysed by fitting the breath-by-breath ventilatory responses to a two-compartment model, as described previously.9Bellville JW Whipp BJ Kaufman RD Swanson GD Aqleh KA Wiberg DM. Central and peripheral chemoreflex loop gain in normal and carotid body-resected subjects.J Appl Physiol. 1979; 46: 843-853PubMed Google Scholar 10Dahan A DeGoede J Berkenbosch A Olivier ICW. Influence of oxygen on the ventilatory response to carbon dioxide in man.J Physiol (Lond). 1990; 428: 485-499Crossref Scopus (128) Google Scholar In short, the steady-state relationship of V˙e to Pe′co2 at constant Pe′o2 is described by the expression V˙e=(GP+GC) [Pe′CO2–B] where V˙e is minute ventilation, GP is the carbon dioxide sensitivity of the peripheral chemoreflex loop, GC is the carbon dioxide sensitivity of the central chemoreflex loop and B is the apnoeic threshold or extrapolated Pe′CO2 of the steady-state ventilatory response to carbon dioxide at zero V˙e. The sum of GP and GC is the total carbon dioxide sensitivity (GTOT). To describe the delay in effect and dynamics of the peripheral and central ventilatory responses to carbon dioxide, time delays and time constants are incorporated in the model. The deterministic model parameters are B, GC, GP, the time constant of the peripheral chemoreflex loop, the time constant of the central chemoreflex loop and a linear trend term.10Dahan A DeGoede J Berkenbosch A Olivier ICW. Influence of oxygen on the ventilatory response to carbon dioxide in man.J Physiol (Lond). 1990; 428: 485-499Crossref Scopus (128) Google Scholar The noise corrupting the data was modelled through an external pathway with first-order dynamics.10Dahan A DeGoede J Berkenbosch A Olivier ICW. Influence of oxygen on the ventilatory response to carbon dioxide in man.J Physiol (Lond). 1990; 428: 485-499Crossref Scopus (128) Google Scholar The parameters were estimated with a one-step prediction error method.15Ljung L. System Identification: Theory for the User. Prentice-Hall, Englewood Cliffs, NJ1987Google Scholar The estimated parameters of control and tramadol experiments were tested by two-way analysis of variance. P-values <0.05 were considered significant. All values are given as mean (sd) and the lower (25%) and upper (75%) quartiles. Control of the Pe′co2 was within 0.18 kPa (i.e. the sd of the breath-by-breath Pe′co2 of single periods was 0.18 kPa or less). Pe′o2 was controlled within 0.26 kPa. Data from two subjects were discarded because of consistently irregular breathing, which made estimates of the model parameters impossible. Mean age of the remaining subjects was 33 (7) (29–37) yr, weight 72 (14) (65–78.75) kg and height 170 (11) (160.0–180.75) cm. All subjects finished the measurements without side-effects. Examples of a control and a tramadol hypercapnic experiment and model fits of one subject are given in Fig. 1, which shows decreases in the fast and slow components ( V˙p and V˙c) after tramadol. The estimated model parameters are collected in Table 1. Tramadol had no effect on the position of the V˙co2 response curve relative to the x-axis (parameter B), but reduced the slope of the curve (i.e. the total carbon dioxide sensitivity) by 29 (20) (11–51)%. This was caused by a reduction in central and peripheral carbon dioxide sensitivities, by 30 (22) (15–54)% and 23 (46) (3–54)% respectively. Tramadol did not affect the ratio GP/GC. The trend term and the time constants and time delays of both chemoreflex loops were not affected (data not shown). In Fig. 2 the mean values of B, GP, GC and GP/GC of each subject for the control and tramadol experiments are shown in scatter diagrams.Table 1Estimated model parameters. Mean (sd). *Two-way analysis of varianceControlTramadolP*B (kPa)4.6 (0.8)4.2 (0.8)0.053GC (litre min−1 kPa−1)10.8 (5.0)7.8 (4.9)0.002GP (litre min−1 kPa−1)2.0 (1.1)1.3 (0.5)0.038GTOT (litre min−1 kPa−1)12.8 (5.7)9.1 (5.0)<0.001GP/GC0.19 (0.09)0.16 (0.16)0.88 Open table in a new tab Carbon dioxide output decreased significantly by a small amount [ V˙co2, control 0.19 (0.04) (0.15–0.22), tramadol 0.17 (0.04) (0.13–0.21) litre min−1 (P<0.043)]. The oxygen consumption and gas exchange ratio remained unaffected by tramadol [ V˙o2, control 0.22 (0.06) (015–0.26), tramadol 0.21 (0.04) (0.20–0.24) litre min−1 (not significant); gas exchange ratio, control 0.89 (0.18) (0.93–1.03), tramadol 0.82 (0.11) (0.69–0.94) (not significant)]. Tramadol did not effect the arousal level of the subjects as judged by the BIS of the EEG [control 96.2 (0.6) (95.8–96.9), tramadol 94.7 (4.1) (91.3–97) (not significant)]. Two subjects reported side-effects several hours after taking tramadol. One reported dizziness, the other nausea and vomiting. We found that an analgesic dose of tramadol depresses respiration in healthy volunteers. Respiratory control was assessed by measuring the ventilatory response to carbon dioxide. The V˙e–co2 response curve (i.e. its slope and its position relative to the x-axis) is a sensitive tool to assess the effects of pharmacological agents on ventilatory control and is superior to single measurements of V˙e, arterial, end-tidal or transcutaneous Pco2. Tramadol reduced the ventilatory carbon dioxide sensitivity by ∼30%. This effect could be caused at various sites: the peripheral or central chemoreceptors, the integrating respiratory centres in the brainstem, the neuromechanical link between the brainstem and ventilation (i.e. motor neurone, neuromuscular junction, respiratory muscles and lung tissue) or sites involved in the control of behavioural state. Our approach of using the dynamic end-tidal forcing technique,9Bellville JW Whipp BJ Kaufman RD Swanson GD Aqleh KA Wiberg DM. Central and peripheral chemoreflex loop gain in normal and carotid body-resected subjects.J Appl Physiol. 1979; 46: 843-853PubMed Google Scholar 10Dahan A DeGoede J Berkenbosch A Olivier ICW. Influence of oxygen on the ventilatory response to carbon dioxide in man.J Physiol (Lond). 1990; 428: 485-499Crossref Scopus (128) Google Scholar which allows us to quantify the relative contributions of the peripheral and central chemoreflex loops to total ventilation, in combination with measurement of the bispectral index of the EEG (an objective measure of sedation/hypnosis)16Rosow C Manberg PJ. Bispectral index monitoring.Anesthesiol Clin North Am. 1998; 2: 89-107Google Scholar during the experiments enabled us to differentiate between these sites and give an approximate location of tramadol's respiratory sites of actions. Tramadol did not materially affect V˙o2, V˙co2 or BIS, suggesting that tramadol did not depress carbon dioxide sensitivity by a decrease in the metabolic or arousal state of the subjects. An effect at sites common to both chemoreflex loops seems most probable because the outputs of the peripheral and central chemoreflex loops were decreased by about 25–30%. Tramadol probably affected the ventilatory control system by acting at the respiratory integrating centres within the brainstem. In this respect, tramadol does not differ from other agents acting at μ-opioid receptors. We performed control and drug experiments on one day. For each subject, the order of experiments was the same: first the control and then the tramadol experiment. There are several reasons for this approach. We did not want to perform control and drug experiments on separate days, because day-to-day variability of the ventilatory responses to hypercapnia is more significant than within-day variability.17Berkenbosch A Bovill JG Dahan A DeGoede J Olievier ICW. Ventilatory CO2 sensitivities from Read's rebreathing method and steady-state method are not equal in man.J Physiol (Lond). 1989; 411: 367-377Crossref Scopus (71) Google Scholar 18Sahn SA Zwillich CW Dick N et al.Variability of ventilatory responses to hypoxia and hypercapnia.J Appl Physiol. 1977; 43: 1019-1025Crossref PubMed Scopus (113) Google Scholar A randomized cross-over study on one day leads to excessively long sessions and discomfort of the subjects. Furthermore, because tramadol is not eliminated completely within a short time, an influence on subsequent control experiments cannot be excluded.19Lintz W Barth H Osterloh G Schmidt-Bothelt E. Bioavailability of enteral tramadol formulations.Arzneim-Forsch Drug Res. 1986; 36: 1278-1283PubMed Google Scholar Because the differences between treatments could have been small, we opted to use a protocol in which the run-to-run variability was minimal. Previous studies on the effect of tramadol on ventilatory control give conflicting results, which we relate to the various methods used to measure ventilatory effect and/or to the complexity of protocols. For example, Tarkkila and colleagues3Tarkkila P Tuominen M Lindgren L. Comparison of respiratory effects of tramadol and pethidine.Eur J Anaesth. 1998; 15: 64-68Crossref PubMed Scopus (57) Google Scholar and Vickers and colleagues6Vickers MD Flaherty DO Szekely SM Read M Yoshizumi J. Tramadol: pain relief by an opioid without depression of respiration.Anaesthesia. 1992; 47: 291-296Crossref PubMed Scopus (360) Google Scholar compared the respiratory effects of tramadol with meperidine or morphine on ventilation in anaesthetized patients breathing 0.3–1% halothane in 70% nitrous oxide before elective surgery. While meperidine and morphine caused significant respiratory depression, as observed by an increase in end-tidal Pco2, and decreases in minute ventilation and respiratory rate, i.v. tramadol seemed devoid of respiratory effects or had only a minor effect on respiratory rate. Such studies are hard to interpret, taking into account the respiratory effects of halothane (depression of peripheral and central carbon dioxide responses13Dahan A van den Elsen MJLJ Berkenbosch A et al.Effects of subanesthetic halothane on the ventilatory responses to hypercapnia and acute hypoxia in healthy volunteers.Anesthesiology. 1994; 80: 727-738Crossref PubMed Scopus (76) Google Scholar 20Stuth EA Tonkovic-Capin M Kampine JP Bajic J Zuperku EJ. Dose-dependent effects of halothane on the carbon dioxide responses of expiratory and inspiratory bulbospinal neurons and the phrenic nerve activities in dogs.Anesthesiology. 1994; 81: 1470-1483Crossref PubMed Scopus (29) Google Scholar) and nitrous oxide, which can stimulate ventilatory control, probably as a result of its sympathicomimetic properties.21Dahan A Ward DS. Effects of 20 percent nitrous-oxide on the ventilatory response to hypercapnia and sustained isocapnic hypoxia in man.Br J Anaesth. 1994; 72: 17-20Crossref PubMed Scopus (15) Google Scholar Interaction of these agents with the opioid and non-opioid actions of tramadol cannot be excluded. Our observations support those of Seitz and colleagues, who found that the V˙e–co2 response was depressed dose-dependently by 15–25% by tramadol 1 and 1.5 mg kg−1 i.v. in healthy awake volunteers.5Seitz W Lubbe N Fritz K Sybrecht G Kirchner E. Eiffub von tramadol auf die ventilatorische CO2 -antwort un den mundokklusionsdruck.Anaesthesist. 1985; 34: 241-246PubMed Google Scholar Warren and colleagues tested the effect of oral tramadol on the ventilatory response to short-term (7 min) hypoxia against the background of mild isohypercapnia.4Warren PM Taylor JH Nicholson KE Wraith PK Drummond GB. Influence of tramadol on the ventilatory response to hypoxia in humans.Br J Anaesth. 2000; 85: 211-216Crossref PubMed Scopus (25) Google Scholar Whereas hypercapnic ventilation was reduced, an observation in agreement with our findings, tramadol had no effect on the hypoxic V˙e response. This is surprising in view of our present observation of a depressant effect of tramadol on respiration, which is probably located in the respiratory integrating centres. The ventilatory response to hypoxia is biphasic: an initial hyperventilatory response, originating in the carotid bodies, is followed after 3–5 min by a slow decline, which originates centrally (i.e. within the central nervous system).22Dahan A Berkenbosch A DeGoede J van den Elsen M Olievier I van Kleef J. Influence of hypoxic duration and posthypoxic inspired O2concentration on short term potentiation of breathing in humans.J Physiol (Lond). 1995; 488: 803-813Crossref Scopus (25) Google Scholar The mechanism of this respiratory effect of central hypoxia remains unknown but may involve various neuromodulators. Tramadol may have reduced central hypoxic depression by its non-opioid modes of action, such as central serotonin release (see below), and thus offset the depression of the acute hypoxic response. Tramadol and the O-desmethyltramadol metabolite of its (+) enantiomer produce analgesia by an agonistic effect on the μ-opioid receptor.23Poulsen L Arendt-Nielsen L Brosen K Sindrup SH. The hypoalgesic effect of tramadol in relation to CYP2D6.Clin Pharmacol Ther. 1996; 60: 636-644Crossref PubMed Scopus (366) Google Scholar However, the antinociceptive effect of tramadol in the rat hotplate test is only partially antagonized by naloxone, and activation of opioid receptors appears to be responsible for only 50% of tramadol's analgesic effect.1Raffa RB Fridriechs E Reimann W et al.Opioid and nonopioid components independently contribute to the mechanism of cation of tramadol, an atypical opioid analgesic.J Pharmacol Exp Ther. 1992; 260: 275-285PubMed Google Scholar The remainder of its analgesic action may be by inhibition of norepinephrine and serotonin reuptake and by facilitation of serotonin release in descending neural antinociceptive pathways.1Raffa RB Fridriechs E Reimann W et al.Opioid and nonopioid components independently contribute to the mechanism of cation of tramadol, an atypical opioid analgesic.J Pharmacol Exp Ther. 1992; 260: 275-285PubMed Google Scholar 24Driessen B Reimann W. Interaction of the central analgesic tramadol, with the uptake and release of 5-hydroxytryptamine in the rat brain in vitro.Br J Pharmacol. 1992; 105: 13-17Crossref PubMed Scopus (229) Google Scholar The molecular mechanisms of the respiratory effects of tramadol remain unknown. Whereas the activation of the μ-opioid receptor is associated with respiratory depression,25Dahan A Sarton E Teppema L et al.Anesthetic potency and influence of morphine and sevoflurane on respiration in μ-opioid receptor knockout mice.Anesthesiology. 2001; 94: 824-832Crossref PubMed Scopus (120) Google Scholar the effects of monoamines on respiration are less evident.26Bianchi AL Denavit-Saubie M Champagnat J. Central control of breathing in mammals: neuronal circuitry, membrane properties, and neurotransmitters.Physiol Rev. 1995; 74: 1-45Google Scholar, 27Champagnat J Denavit-Saubie M Henry JL Leviel V. Catecholaminergic depressant effects on bulbar respiratory mechanisms.Brain Res. 1979; 160: 57-68Crossref PubMed Scopus (117) Google Scholar, 28Lundholm L Svedmyr N. Studies on the stimulating effects of adrenaline and noradrenaline on respiration in man.Acta Physiol Scand. 1966; 67: 65-75Crossref PubMed Scopus (10) Google Scholar Central release of serotonin may depress as well as stimulate breathing, depending on the type of respiratory neurone and 5-HT receptor subtype involved.26Bianchi AL Denavit-Saubie M Champagnat J. Central control of breathing in mammals: neuronal circuitry, membrane properties, and neurotransmitters.Physiol Rev. 1995; 74: 1-45Google Scholar Most studies indicate that central release of norepinephrine causes respiratory depression.27Champagnat J Denavit-Saubie M Henry JL Leviel V. Catecholaminergic depressant effects on bulbar respiratory mechanisms.Brain Res. 1979; 160: 57-68Crossref PubMed Scopus (117) Google Scholar To determine the relative effect of the μ-opioid receptor in tramadol-induced respiratory depression, Teppema and colleagues determined the ability of naloxone to reverse the depression by tramadol of the V˙e–co2 response in an anaesthetized cat model.29Teppema LJ Olivier CN Dahan A. Respiratory depression by tramadol: involvement of opioid receptors.Anesthesiology. 2000; 93: U233Crossref Google Scholar Respiratory depression by tramadol was reduced by 70–80% after naloxone pretreatment, suggesting that at least 70% of tramadol's respiratory effect is related to its action at opioid receptors, while the remainder could be by inhibition of serotonin and norepinephrine reuptake. Respiration in perioperative patients is related to the balance between stimulation from pain, stress and activated chemoreflexes and depression resulting from sedation and the direct effect of analgesics and anaesthetics on respiratory neurones.30Sarton E Dahan A Teppema L et al.Acute pain and central nervous system arousal do not restore impaired hypoxic ventilatory response during sevoflurane sedation.Anesthesiology. 1996; 85: 295-303Crossref PubMed Scopus (60) Google Scholar Our subjects were free of pain or surgical stress, which should be considered before extrapolating our findings to perioperative patients. The respiratory effects in the present study may have been overestimated. The effect observed in this study is equivalent to that found after morphine 0.13 mg kg−1 i.v. in healthy volunteers without pain.12Sarton E Teppema L Dahan A. Sex differences in morphine-induced ventilatory depression reside within the peripheral chemoreflex loop.Anesthesiology. 1999; 90: 1329-1338Crossref PubMed Scopus (93) Google Scholar In conclusion, tramadol reduces the hypercapnic ventilatory response. This depression probably acts through the respiratory integrating centres within the brainstem. This study was supported by a grant for international collaboration from the Wellcome Trust and by the Netherlands Organization for Pure Research (NWO), The Hague, The Netherlands (grant MW 902-21-211). The EEG monitor and electrodes were provided by Aspect Medical Systems International, Leiden, The Netherlands.