- Email: [email protected]
Influence of tramadol on the ventilatory response to hypoxia in humans
British Journal of Anaesthesia 85 (2): 211±6 (2000)
In¯uence of tramadol on the ventilatory response to hypoxia in humans P. M. Warren1*, J. H. Taylor2, K. E. Nicholson2, P. K. Wraith3 and G. B. Drummond4 1
Respiratory Medicine Unit, Department of Medical and Radiological Science, University of Edinburgh, Royal In®rmary, Edinburgh, 2Department of Pharmacology, University of Edinburgh, 3Department of Medical Physics, Western General Hospital, Edinburgh and 4Department of Anaesthetics, University of Edinburgh, Royal In®rmary, Edinburgh, UK *Corresponding author We studied the effect of tramadol on the ventilatory response to 7 min acute isocapnic hypoxia (SpO 85.1 (SD 0.4)%) during steady mild hypercapnia (PE¢CO 0.7 kPa above normoxic baseline) in 14 healthy volunteers (seven male). The acute hypoxic response was measured before and 1 h after oral placebo or tramadol (100 mg). After tramadol, ventilation during mild hypercapnia (mean 11.28 litres min±1) was signi®cantly less (P<0.05) than during placebo baseline (13.93 litres min±1), tramadol baseline (14.63 litres min±1), or after placebo (14.95 litres min±1), con®rming that tramadol has a small depressive effect on the hypercapnic ventilatory response. There was no signi®cant difference in the hypoxic ventilation/SpO response (l min±1 %±1) measured during the placebo baseline (0.99), placebo (1.18), tramadol baseline (0.78) or tramadol (0.68) runs. These data suggest that tramadol does not depress the hypoxic ventilatory response. 2
2
2
Br J Anaesth 2000; 85: 211±6 Keywords: pharmacology, tramadol; ventilation, hypoxic response Accepted for publication: March 1, 2000
Tramadol hydrochloride, licensed in the UK in 1994, is used for moderate acute and chronic pain. Although developed as a synthetic opioid, it has weak opioid receptor af®nity1 with only ~30% of its antinociceptor and analgesic actions antagonized by naloxone.2 Much of its action is due to inhibition of the reuptake of the monoamines serotonin and noradrenaline at synapses in the descending neural pathways involved in analgesia.1 Unlike other opioids, tramadol appears to be relatively free of the unwanted side effect of ventilatory depression causing only 10±15% reductions in normoxic ventilation and the hypercapnic ventilatory response3 4 compared with the 30±50% decreases which occur following a dose of morphine or pethidine of similar potency.3 5±7 However, the effect on the hypoxic ventilatory response is not known. We have therefore studied the effect of oral tramadol on the ventilatory response to acute isocapnic hypoxia in normal subjects. The hypoxic ventilatory response was measured on a background of mild hypercapnia to mimic the postoperative situation.8 In addition, we studied both men and women as sex-related differences have been found in the effect of morphine on ventilatory control in humans.6
Subjects and methods Twenty healthy volunteers (nine males) gave written informed consent to participate in the study. None had a history of cardiorespiratory disease, or was receiving medication other than oral contraception at the time of study. The study was approved by the local Ethics Committees. Each subject attended the laboratory on three occasions. The ®rst visit was for familiarization. A medical history was obtained and the FEV1 and FVC measured to con®rm normal lung function. Subjects were then allowed to become accustomed to breathing through the facemask under normoxic, hypercapnic and hypoxic conditions. The second and third visits were study days and subjects were requested to fast for 3 h and to refrain from taking substances known to affect ventilation (e.g. caffeine) for a minimum of 8 h prior to study. On each study day the ventilatory response to hypoxia was measured before and 1 h after taking either tramadol 100 mg (ZYDOL, Searle) or matched placebo tablets. Drug and placebo study days were in random order and neither the subject nor assessor of ventilatory variable was aware of the treatment given.
Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2000
Warren et al.
Subjects sat and read during the 1-h period between studies. Experimental studies were scheduled for the same time of day in any individual subject to minimize the effects of diurnal variation on the hypoxic ventilatory response. Females were studied within the ®rst 10 days of a normal menses to ensure that they were not pregnant and to avoid the effect of increasing levels of progesterone in the luteal phase of the menstrual cycle. Subjects were studied in a semi-recumbent position in a well-lit room and listened to music via headphones. They were instructed to stay awake and were roused if they appeared to fall asleep. Subjects breathed through a facemask (Hans Rudolf, series 8930) which enclosed both nose and mouth, was sealed to the face with gel (Aqua Gel, Adams Healthcare, UK), and was connected to a lowresistance two-way valve. Inspiratory gas mixtures prepared from primary gases (oxygen, nitrogen, carbon dioxide) were delivered to the subject through a T-piece attached to the inspiratory port of the two-way valve. Ventilatory variables were recorded using methods described previously.9 Expiratory gas passed via a heated pneumotachograph (Fleisch No. 2) through mixing and drying chambers to a Parkinson Cowan CD4 dry gas meter modi®ed to give a digital signal. The integrated expiratory ¯ow signal, which gave breath-by-breath tidal volume, was calibrated against the output of the CD4 gas meter every 10 litres to correct the ¯ow signal. A mass spectrometer (VG Spectralab M), calibrated with four gas mixtures of known oxygen, carbon dioxide, nitrogen and argon concentration, measured inspiratory and end-tidal oxygen (PE¢O ; kPa) and carbon dioxide (PE¢CO ; kPa) partial pressures at the lips. Pulse oximetry (SpO , %; Ohmeda Biox 3700, set to a fast averaging time of 2 s), and electrocardiogram (HewlettPackard 78351A) were measured continuously throughout the study. Breath-by-breath values of inspiratory time (TI, s), expiratory time (TE, s), total breath time (TTOT=TI + TE, s), ventilatory frequency (f=60/TTOT, b.p.m.), tidal volume (VT, litres BTPS), instantaneous minute ventilation (VEinst=f 3 VT, litres min±1 BTPS), mean inspiratory ¯ow (VT : TI, litres s±1), and inspired and end-tidal partial pressures were digitized using an Olivetti PCS 286 computer and stored on disk for off-line analysis. Oxygen consumption (litres min±1 STPD) and carbon dioxide output (litres min±1 STPD) were measured from collections of mixed expired gas made over a 2-min period, and the gas exchange ratio calculated. Concentrations of oxygen and carbon dioxide were measured using a Servomex oxygen analyser (Servomex model 570A) calibrated with air and 100% nitrogen, and a Gould capnograph (Mark IV) calibrated with four gas mixtures of known carbon dioxide concentrations, respectively. Composition of the inspired gas was controlled by three mass-¯ow controllers (Bronkhorst Hi-Tech Types F-202AC and F-201AC with Bronkhorst E7200-AAA power supply and control unit). These supplied accurately controlled 2
2
2
¯ows in the range 0±10 litres s±1 for carbon dioxide, 0± 50 litres s±1 for oxygen and 0±100 litres s±1 for nitrogen. The gas input pressure for each controller was maintained at 2 bar by means of precision pressure regulator (0±4 bar, RS components) attached to the inlet to each mass-¯ow controller. The output of the three controllers was passed through a mixing chamber before being presented to the subject. A computer (Elonex PT-5120/l) was interfaced with the mass-¯ow controllers via a digital-to-analogue converter (Amplicon PC24) and also with the data-acquisition computer. A custom-written program displayed graded ¯ow scales representing the output of each of the mass-¯ow controllers and breath-by-breath values of end-tidal oxygen and end-tidal carbon dioxide on the Elonex PC monitor. Adjustment of the graded scales using the computer mouse altered the output of the mass-¯ow controllers so that the required gas mixture delivered to the subject could be adjusted to achieve rapid, accurate changes in end-tidal partial pressures. Each hypoxic ventilatory response measurement lasted 37 min. Subjects initially breathed room air for 15 min and duplicate measurements of oxygen consumption and carbon dioxide output were made between 8 and 13 min. The subjects then breathed a 21% oxygen mixture produced from the mass-¯ow controller system for a further 5 min to obtain a steady baseline. Mild hypercapnia was then induced for 10 min by increasing inspired carbon dioxide so that end-tidal PCO2 was raised by 0.7 kPa above the previous baseline level. Inspired oxygen concentration was then decreased so that SpO was reduced to 85% for 7 min. The end-tidal was maintained at 0.7 kPa above the initial baseline level throughout hypoxia. Mean values for all the ventilatory variables were calculated for each of the last 3 min of the baseline normoxic period, each of the last 3 min of the hypercapnic period, and each minute during hypercapnic hypoxia. In addition, mean values were calculated during the following 3-min periods of each run: normoxia (3 min before the onset of hypercapnia); hypercapnia (3 min before the onset of hypoxia): hypercapnic hypoxia (the last 3 min of hypoxia). The hypoxic ventilatory response was calculated as the ratio of the changes in VEinst and SpO occurring between the periods of hypercapnia and hypercapnic hypoxia (VEinst/ SpO ; litres min±1 %±1). Group means and SD for imposed ventilatory variables and group means and 95% con®dence limits (CL) for studied variables in the three measurement periods were calculated for each run. Student's t-test was used to compare anthropometric data in men and women, and a paired t-test for within-run comparison of end-tidal PCO2 during hypercapnia and hypoxia. Comparisons between runs for the group as a whole were made by one-way analysis of variance with post hoc comparisons made using the least signi®cant difference method. Comparisons between runs for males and females considered separately were made using Friedman's non-parametric analysis of variance.
212
2
2
2
Tramadol and the hypoxic ventilatory response
Signi®cance was set at P<0.05. Computations were made using SPSS version 9 for Windows.
Results
in the period after tramadol when end-tidal PCO2 was slightly but signi®cantly greater during hypercapnic hypoxia than hypercapnia alone (Table 2). The mean difference after tramadol was 0.04 and the maximum-recorded increase of 0.08 kPa is within the error of the measurement.
Of the 20 subjects recruited to the study, two subjects (one male) were withdrawn from the study after experiencing severe nausea after taking tramadol. Three female subjects were unable to return for day 3 of the study because of very irregular menstrual cycles. One further male subject completed the study but had a markedly irregular breathing pattern in the ®nal limb. The ventilatory variables could not be analysed accurately and the data from this subject were therefore excluded from the ®nal analysis. Anthropometric data on the 14 subjects included in the analysis are given in Table 1. The men were signi®cantly taller and had a higher BSA than the women, but there were no signi®cant differences in age, weight, or calculated drug dose in mg kg±1. The pattern of ventilatory response was similar in all four limbs of the study. Ventilation increased slightly during mild hypercapnia, with a further substantial increase and little secondary decline during the 7 min of hypoxia (Fig. 1). Within runs, end-tidal PCO2 was controlled, on average, to within 0.02 kPa of the hypercapnic level during hypercapnic hypoxia. End-tidal PCO2 did not differ signi®cantly between the periods of hypercapnia and hypercapnic hypoxia except Table 1 Means and SD or range for the subject details and drug dosage. *Signi®cant difference (P<0.05) between males and females Males
Females
7 20.6 (18±22) 179 (10) 73.7 (9.4) 1.9 (0.2) 1.4 (0.2)
7 21.3 (20±22) 168 (5)* 65.0 (8.6) 1.7 (0.1)* 1.6 (0.2)
Fig 1 Oxygen saturation (SpO ), instantaneous minute ventilation (VEinst) and end-tidal PCO2 (PE¢CO ) plotted against time for both placebo (lefthand side) and tramadol (right-hand side) studies for baseline (open circles) and placebo or tramadol (closed circles). Mean values 6 SD for each 1-min period are presented for the entire group. Mean values for each variable were calculated for the last 3 min of normoxia (A1), the last 3 min of hypercapnia (H1), and the last 3 min of hypercapnic hypoxia (H2) for each subject and used for comparison between runs. 2
n Age (yr) Height (cm) Weight (kg) BSA (m2) Drug dose (mg kg±1)
2
Table 2 Means (SD) for the imposed variables for the studied variables in the three measurement periods for baseline, placebo and tramadol runs. *Signi®cant difference between hypercapnic and hypercapnic hypoxia periods
End-tidal PCO2 (kPa)
End-tidal PCO2 (kPa)
SpO (%) 2
Normoxia
Hypercapnia
Hypercapnic hypoxia
Baseline Placebo Baseline Tramadol
5.27 5.14 5.08 5.39
5.97 5.96 5.83 5.82
5.99 5.97 5.86 5.86
(0.48) (0.48) (0.37) (0.39)*
Baseline Placebo Baseline Tramadol
12.87 13.72 13.77 13.49
6.15 6.16 6.11 6.07
(0.61) (0.65) (0.90) (0.98)
Baseline Placebo Baseline Tramadol
97.5 97.0 97.6 97.5
85.1 85.2 85.0 85.1
(0.3) (0.3) (0.5) (0.4)
(0.48) (0.46) (0.46) (0.34) (2.37) (0.57) (0.80) (0.75)
(0.7) (0.7) (0.7) (0.7)
213
(0.52) (0.51) (0.38) (0.39)
13.02 13.84 13.85 13.54 97.6 97.1 97.6 97.4
(2.17) (0.63) (0.81) (0.69)
(0.7) (0.5) (0.6) (0.8)
Warren et al. Table 3 Mean (95% con®dence limits) baseline metabolic values
Baseline Placebo Baseline Tramadol
Oxygen consumption (litres min±1)
Carbon dioxide output (litres min±1)
Gas exchange ratio
209 205 222 202
163 162 179 158
0.78 0.78 0.81 0.78
(183, (184, (190, (173,
234) 225) 254) 232)
(140, (140, (154, (132,
186) 183) 204) 180)
(0.74, (0.74, (0.78, (0.75,
0.82) 0.83) 0.84) 0.81)
Table 4 Means (95% con®dence limits) for the studied variables in the three measurement periods for baseline, placebo and tramadol runs. *Indicates signi®cant difference from placebo baseline, placebo and tramadol baseline values (P<0.05) Normoxia VEinst (litres min±1)
VT (litre)
f (b.p.m.)
Hypercapnia
Hypercapnic hypoxia
Baseline Placebo Baseline Tramadol
7.39 7.72 8.21 7.54
(6.43, (6.41, (7.31, (6.48,
8.33) 8.53) 9.11) 8.6)
13.93 (11.94, 14.95 (12.67, 14.63 (13.02, 11.28* (9.27,
Baseline Placebo Baseline Tramadol
0.50 0.48 0.52 0.45
(0.43, (0.42, (0.45, (0.40,
0.56) 0.55) 0.58) 0.49)
0.83 (0.72, 0.93) 0.85 (0.73, 0.97) 0.85 (0.73, 0.97) 0.65* (0.56, 0.73)
1.31 1.34 1.31 1.02
(1.10, (1.10, (1.11, (0.85,
1.52) 1.59) 1.50) 1.19)
Baseline Placebo Baseline Tramadol
15.3 16.8 16.7 17.2
(13.1, (14.2, (14.1, (14.5,
17.5) 19.4) 19.3) 20.0)
17.2 18.1 18.1 17.8
20.2 21.8 19.7 19.7
(16.9, (17.1, (16.8, (16.2,
23.4) 26.5) 22.5) 23.1)
(15.7, (15.8, (15.6, (14.9,
15.93) 17.23) 16.23) 13.29)
26.14 28.99 24.45 19.64
19.2) 20.4) 20.5) 20.6)
(20.00, (21.14, (21.25, (15.24,
32.27) 36.83) 27.65) 24.04)
Fig 2 Plots of the VEinst/SpO relationship for the placebo (upper panel) and tramadol (lower panel) limbs of the study for the entire group. Baseline studies are shown as open circles and studies after placebo or tramadol are shown as closed circles. Data are given as mean values and con®dence intervals. 2
Fig 3 Plots of the individual VEinst/SpO relationships measured before and after either placebo or tramadol. The data are grouped for males (upper panel) and females (lower panels). A lesser slope indicates a decreased hypoxic ventilatory drive. 2
There were no signi®cant differences in oxygen consumption, carbon dioxide output or the gas exchange ratio between the four runs (Table 3). During normoxia there were no signi®cant differences in the imposed variables between the four runs (Table 2), although end-tidal PCO2 tended to be greater and end-tidal PCO2 tended to be less after tramadol. Similarly, there were no signi®cant differences between runs in the studied
variables although VEinst and VT tended to be less during the tramadol run (Table 4). During hypercapnia there were no signi®cant differences between runs in the imposed variables (Table 2), but both VT and VEinst during the tramadol
214
Tramadol and the hypoxic ventilatory response
run were signi®cantly less than during either of the baseline runs or the placebo run (Table 4). During hypercapnic hypoxia there were no signi®cant differences in either imposed or studied variables, although again VEinst tended to be less after tramadol. There was no signi®cant difference between the VEinst/ SpO ratio for the placebo baseline (mean ±0.99, CL ±0.59, ±1.40 litres min±1 %±1), placebo (mean ±1.18, CL ±0.66, ±1.70 litres min±1 %±1), tramadol baseline (mean ±0.78, CL ±0.57, ±0.98 litres min±1 %±1), and tramadol (mean ±0.68, CL ±0.43, ±0.93 litres min±1 %±1) runs (Fig. 2). Similarly, there were no signi®cant differences in the VEinst/SpO ratios between the four runs when men and women were considered separately (Fig. 3). 2
2
Discussion We have shown that a standard analgesic dose of tramadol in normal subjects had no signi®cant effect on the acute ventilatory response to isocapnic hypoxia. During mild hypercapnia, VEinst was signi®cantly less after tramadol compared with either the baseline run or after placebo in spite of similar levels of end-tidal PCO2. This ®nding is consistent with a small effect of tramadol on the ventilatory response to hypercapnia.4 Although tramadol did not have a signi®cant effect on normoxic ventilation, the fact that VEinst and VT tended to be ~10% less and end-tidal PCO2 ~0.3 kPa greater after tramadol than during the tramadol baseline run is similar to changes reported previously in anaesthetized patients.3 7 Studies were performed on two separate days because the 5-h elimination half-life of tramadol10 meant that the order of the placebo and active drug runs could not be randomized if the studies were performed on the same day. However, the hypoxic ventilatory response is known to vary substantially within and between days.11 Individuals were studied at approximately the same time of day to minimize the effect of diurnal variation. To overcome the between-day variability, a baseline study was performed on each day and the effect of placebo or tramadol compared with the respective baseline measurements. We were not able to measure serum tramadol concentrations to ensure that adequate concentrations had been achieved. A single oral dose of tramadol 100 mg given to fasting normal subjects is absorbed rapidly after an initial lag time of ~30 min, with serum concentrations reaching the analgesic threshold of 100 ng litre±1 after ~40 min and reaching peak levels at 2 h.10 However, there is considerable individual variability in bioavailability, with times for peak serum concentration varying between 1 and 4 h.10 In this study, the repeated measurements were started 1 h after drug administration so that the repeated hypoxic challenge occurred after ~90 min when serum concentrations would be approaching maximum in most subjects. While this ensured that the hypoxic ventilatory response was unlikely to be depressed as a result of previous hypoxic exposure,12 it
is possible that peak serum concentrations had not been reached in some subjects. However, there was a reduction in hypercapnic VEinst of over 1.5 litres in nine out of 14 subjects after tramadol, whereas this only occurred in two subjects after placebo suggesting that active levels of tramadol had been achieved in most subjects. It is therefore unlikely that low drug concentrations account for the lack of depression of the hypoxic ventilatory response. Although 20 subjects were recruited, data were only obtained in 14 for a variety of reasons. Failure to detect a signi®cant effect of tramadol on the hypoxic ventilatory response might therefore be a type 2 statistical error. However, this is unlikely to be the explanation as the study had a power of 0.9 to detect a change in the VEinst/SpO ratio of 0.5. Dahan and colleagues6 found a sex-dependent effect on the hypoxic ventilatory response which was independent of differences in weight, lean body mass, body surface area and calculated fat mass. In their study, morphine signi®cantly reduced the initial ventilatory response to sustained hypoxia in women but not in men. We studied similar numbers of men and women. It was possible, therefore, that an effect in women was being masked by an absence of effect in the men. However, separate analysis of the results in seven men and eight women who completed the study showed no effect of tramadol on the initial hypoxic ventilatory response in either sex. The O-desmethyltramadol metabolite of the (+) enantiomer of tramadol, produced by hepatic phase I metabolism by cytochrome P-450 2D6 (CYP2D6), has an af®nity for the mopioid receptor which is ~200 times that of the parent compound.13 It is likely, therefore, that the bulk of any ventilatory depression would be mediated by this metabolite. Approximately 10% of the Caucasian population are phenotypically poor metabolizers14 and have a reduced analgesic effect of tramadol.13 If, by chance, most of our subjects were poor metabolizers, then this could explain the absence of effect on the hypoxic ventilatory response. This effect could contribute to our ®ndings but is unlikely to be the only explanation as tramadol suppressed the hypercapnic ventilatory response in nine out of 14 of our subjects. The lack of suppression by tramadol of the hypoxic ventilatory response contrasts with the 50±60% suppression produced by morphine. Unlike morphine, tramadol produces its analgesic effects by a combination of low but preferential activity at m-opioid receptors,1 by inhibiting both noradrenaline and 5-hydroxytrypamine (5-HT) uptake,1 and by facilitating 5-HT release.15 The effects of these non-opioid actions of tramadol on ventilation are dif®cult to assess. A recent study showed that the 5-HT reuptake inhibitor ¯uoxetine caused a slight depression in resting ventilation in goats.16 However, 5-HT can stimulate or depress ventilation depending on the subtype of receptor affected and the type of respiratory neurone activated.17 Catecholamines have a similar dual effect on ventilation depending in part on whether the action is central or peripheral. Catecholamines administered to the respiratory
215
2
Warren et al.
neurones in the medulla usually cause depression,18 whereas systemic administration stimulates breathing19 and increases the ventilatory response to hypercapnia.20 It is possible, therefore, that stimulation by 5-HT and noradrenaline offsets the opioid depressant effects of tramadol. The relative opioid receptor af®nity of tramadol may also explain the ventilatory effects. In cats, depression of peripheral chemosensory discharge is mediated via d receptors.21 Tramadol has modest af®nity for m-receptors but extremely weak d af®nity. Tramadol may therefore suppress the ventilatory response to hypercapnia4 via central m-receptor activity leaving the peripheral component relatively intact. Studies with selective opioid receptor blockers would be required to con®rm this. In conclusion, we have shown that tramadol, in a standard clinical dose, does not signi®cantly affect the ventilatory response to hypoxia. However, tramadol reduced resting tidal volume and insigni®cantly reduced ventilation, resembling the effects reported by Vickers and colleagues3 at a dose of 1 mg kg±1. Seitz and colleagues4 also showed that tramadol caused a moderate depression of the hypercapnic ventilatory response. The signi®cant reduction in hypercapnic ventilation after tramadol is in keeping with previous observations that tramadol causes a moderate depression of the hypercapnic ventilatory response,4 and suggests that this suppression can be signi®cant even at values which are only just above the normoxic resting level.
Acknowledgement This study was funded by the Moray Foundation, University of Edinburgh.
References 1 Collart L, Luthy C, Dayer P. Partial inhibition of tramadol antinociceptive effect by naloxone in man. Br J Clin Pharmac 1993; 35: 73P 2 Raffa RB, Friderichs E, Reimann W, Shank RP, Codd EE, Vaught JL. Opioid and nonopioid components independently contribute to the mechanism of action of tramadol, an `atypical' opioid analgesic. J Pharm Exp Therap 1992; 260: 275±85 3 Vickers MD, O'Flaherty D, Szekely SM, Read M, Yoshizumi J. Tramadol: pain relief by an opioid without depression of respiration. Anaesthesia 1992; 47: 291±6 4 Seitz W, Lubbe N, Fritz K, Sybrecht G, Kirchner E. Ein¯uû von Tramadol auf die ventilatorische CO2-Antwort und den Mundokklusionsdruck. Anaesthesist 1985; 34: 241±6
216
5 Weil JV, McCullough RE, Kline JS, Sodal IE. Diminished ventilatory response to hypoxia and hypercapnia after morphine in normal man. N Engl J Med 1975; 292: 1103±6 6 Dahan A, Sarton E, Teppema L, Olievier C. Sex-related differences in the in¯uence of morphine on ventilatory control in humans. Anesthesiology 1998; 88: 903±13 7 Tarkkila P, Tuominen M, Lindgren L. Comparison of respiratory effects of tramadol and pethidine. Eur J Anaesth 1998; 15: 64±8 8 Jones JG, Sapsford DJ, Wheatley RG. Postoperative hypoxaemia: mechanisms and time course. Anaesthesia 1990; 45: 566±73 9 Young CH, Drummond GB, Warren PM. Effect of a subanaesthetic concentration of halothane on the ventilatory response to sustained hypoxia in healthy humans. Br J Anaesth 1993; 71: 642±7 10 Lintz W, Barth H, Osterloh G, Schmidt-Bothelt E. Bioavailability of enteral tramadol formulations. Arzneim±Forsch/Drug Res 1986; 36: 1278±83 11 Sahn SA, Zwillich CW, Dick N, McCullough RE, Lakshiminaryan S, Weil JW. Variability of ventilatory responses to hypoxia and hypercapnia. J Appl Physiol 1977; 43: 1019±25 12 Easton PA, Slykerman LJ, Anthonisen NR. Recovery of the ventilatory response to hypoxia in normal adults. J Appl Physiol 1988; 64: 521±8 13 Poulsen L, Arendt-Nielsen L, Brosen K, Sindrup SH. The hypoalgesic effect of tramadol in relation to CYP2D6. Clin Pharmacol Ther 1996; 60: 636±44 14 Paar WD, Poche S, Gerloff J, Dengler HJ. Polymorphic CYP2D6 mediates O-demethylation of the opioid analgesic tramadol. Eur J Clin Pharmacol 1997; 53: 235±9 15 Driessen 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: 147±51 16 Henderson DR, Konkle DM, Mitchell GS. Effects of serotonin reuptake inhibition on ventilatory control in goats. Respir Physiol 1999; 115: 1±10 17 Bianchi AL, Denavit-Saubie M, Champagnat J. Central control of breathing in mammals: neuronal circuitry, membrane properties, and neurotransmitters. Physiol Rev 1995; 74; 1±45 18 Champagnat J, Denavit-Saubie M, Henry JL, Leviel V. Catecholaminergic depressant effects on bulbar respiratory mechanisms. Brain Res 1979; 160: 57±68 19 Lundholm L, Svedmyr N. Studies on the stimulating effects of adrenaline and noradrenaline on respiration in man. Acta Physiol Scand 1966; 67: 65±75 20 Cunningham DJC, Hey EN, Lloyd BB. The effect of intravenous infusion of noradrenaline on the respiratory response to carbon dioxide in man. Quart J Exp Physiol 1958; 43: 394±9 21 Kirby GC, McQueen DS. Characterization of opioid receptors in the cat carotid body involved in chemosensory depression in vivo. Br J Pharmac 1986; 88: 889±98
In¯uence of tramadol on the ventilatory response to hypoxia in humans P. M. Warren1*, J. H. Taylor2, K. E. Nicholson2, P. K. Wraith3 and G. B. Drummond4 1
Respiratory Medicine Unit, Department of Medical and Radiological Science, University of Edinburgh, Royal In®rmary, Edinburgh, 2Department of Pharmacology, University of Edinburgh, 3Department of Medical Physics, Western General Hospital, Edinburgh and 4Department of Anaesthetics, University of Edinburgh, Royal In®rmary, Edinburgh, UK *Corresponding author We studied the effect of tramadol on the ventilatory response to 7 min acute isocapnic hypoxia (SpO 85.1 (SD 0.4)%) during steady mild hypercapnia (PE¢CO 0.7 kPa above normoxic baseline) in 14 healthy volunteers (seven male). The acute hypoxic response was measured before and 1 h after oral placebo or tramadol (100 mg). After tramadol, ventilation during mild hypercapnia (mean 11.28 litres min±1) was signi®cantly less (P<0.05) than during placebo baseline (13.93 litres min±1), tramadol baseline (14.63 litres min±1), or after placebo (14.95 litres min±1), con®rming that tramadol has a small depressive effect on the hypercapnic ventilatory response. There was no signi®cant difference in the hypoxic ventilation/SpO response (l min±1 %±1) measured during the placebo baseline (0.99), placebo (1.18), tramadol baseline (0.78) or tramadol (0.68) runs. These data suggest that tramadol does not depress the hypoxic ventilatory response. 2
2
2
Br J Anaesth 2000; 85: 211±6 Keywords: pharmacology, tramadol; ventilation, hypoxic response Accepted for publication: March 1, 2000
Tramadol hydrochloride, licensed in the UK in 1994, is used for moderate acute and chronic pain. Although developed as a synthetic opioid, it has weak opioid receptor af®nity1 with only ~30% of its antinociceptor and analgesic actions antagonized by naloxone.2 Much of its action is due to inhibition of the reuptake of the monoamines serotonin and noradrenaline at synapses in the descending neural pathways involved in analgesia.1 Unlike other opioids, tramadol appears to be relatively free of the unwanted side effect of ventilatory depression causing only 10±15% reductions in normoxic ventilation and the hypercapnic ventilatory response3 4 compared with the 30±50% decreases which occur following a dose of morphine or pethidine of similar potency.3 5±7 However, the effect on the hypoxic ventilatory response is not known. We have therefore studied the effect of oral tramadol on the ventilatory response to acute isocapnic hypoxia in normal subjects. The hypoxic ventilatory response was measured on a background of mild hypercapnia to mimic the postoperative situation.8 In addition, we studied both men and women as sex-related differences have been found in the effect of morphine on ventilatory control in humans.6
Subjects and methods Twenty healthy volunteers (nine males) gave written informed consent to participate in the study. None had a history of cardiorespiratory disease, or was receiving medication other than oral contraception at the time of study. The study was approved by the local Ethics Committees. Each subject attended the laboratory on three occasions. The ®rst visit was for familiarization. A medical history was obtained and the FEV1 and FVC measured to con®rm normal lung function. Subjects were then allowed to become accustomed to breathing through the facemask under normoxic, hypercapnic and hypoxic conditions. The second and third visits were study days and subjects were requested to fast for 3 h and to refrain from taking substances known to affect ventilation (e.g. caffeine) for a minimum of 8 h prior to study. On each study day the ventilatory response to hypoxia was measured before and 1 h after taking either tramadol 100 mg (ZYDOL, Searle) or matched placebo tablets. Drug and placebo study days were in random order and neither the subject nor assessor of ventilatory variable was aware of the treatment given.
Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2000
Warren et al.
Subjects sat and read during the 1-h period between studies. Experimental studies were scheduled for the same time of day in any individual subject to minimize the effects of diurnal variation on the hypoxic ventilatory response. Females were studied within the ®rst 10 days of a normal menses to ensure that they were not pregnant and to avoid the effect of increasing levels of progesterone in the luteal phase of the menstrual cycle. Subjects were studied in a semi-recumbent position in a well-lit room and listened to music via headphones. They were instructed to stay awake and were roused if they appeared to fall asleep. Subjects breathed through a facemask (Hans Rudolf, series 8930) which enclosed both nose and mouth, was sealed to the face with gel (Aqua Gel, Adams Healthcare, UK), and was connected to a lowresistance two-way valve. Inspiratory gas mixtures prepared from primary gases (oxygen, nitrogen, carbon dioxide) were delivered to the subject through a T-piece attached to the inspiratory port of the two-way valve. Ventilatory variables were recorded using methods described previously.9 Expiratory gas passed via a heated pneumotachograph (Fleisch No. 2) through mixing and drying chambers to a Parkinson Cowan CD4 dry gas meter modi®ed to give a digital signal. The integrated expiratory ¯ow signal, which gave breath-by-breath tidal volume, was calibrated against the output of the CD4 gas meter every 10 litres to correct the ¯ow signal. A mass spectrometer (VG Spectralab M), calibrated with four gas mixtures of known oxygen, carbon dioxide, nitrogen and argon concentration, measured inspiratory and end-tidal oxygen (PE¢O ; kPa) and carbon dioxide (PE¢CO ; kPa) partial pressures at the lips. Pulse oximetry (SpO , %; Ohmeda Biox 3700, set to a fast averaging time of 2 s), and electrocardiogram (HewlettPackard 78351A) were measured continuously throughout the study. Breath-by-breath values of inspiratory time (TI, s), expiratory time (TE, s), total breath time (TTOT=TI + TE, s), ventilatory frequency (f=60/TTOT, b.p.m.), tidal volume (VT, litres BTPS), instantaneous minute ventilation (VEinst=f 3 VT, litres min±1 BTPS), mean inspiratory ¯ow (VT : TI, litres s±1), and inspired and end-tidal partial pressures were digitized using an Olivetti PCS 286 computer and stored on disk for off-line analysis. Oxygen consumption (litres min±1 STPD) and carbon dioxide output (litres min±1 STPD) were measured from collections of mixed expired gas made over a 2-min period, and the gas exchange ratio calculated. Concentrations of oxygen and carbon dioxide were measured using a Servomex oxygen analyser (Servomex model 570A) calibrated with air and 100% nitrogen, and a Gould capnograph (Mark IV) calibrated with four gas mixtures of known carbon dioxide concentrations, respectively. Composition of the inspired gas was controlled by three mass-¯ow controllers (Bronkhorst Hi-Tech Types F-202AC and F-201AC with Bronkhorst E7200-AAA power supply and control unit). These supplied accurately controlled 2
2
2
¯ows in the range 0±10 litres s±1 for carbon dioxide, 0± 50 litres s±1 for oxygen and 0±100 litres s±1 for nitrogen. The gas input pressure for each controller was maintained at 2 bar by means of precision pressure regulator (0±4 bar, RS components) attached to the inlet to each mass-¯ow controller. The output of the three controllers was passed through a mixing chamber before being presented to the subject. A computer (Elonex PT-5120/l) was interfaced with the mass-¯ow controllers via a digital-to-analogue converter (Amplicon PC24) and also with the data-acquisition computer. A custom-written program displayed graded ¯ow scales representing the output of each of the mass-¯ow controllers and breath-by-breath values of end-tidal oxygen and end-tidal carbon dioxide on the Elonex PC monitor. Adjustment of the graded scales using the computer mouse altered the output of the mass-¯ow controllers so that the required gas mixture delivered to the subject could be adjusted to achieve rapid, accurate changes in end-tidal partial pressures. Each hypoxic ventilatory response measurement lasted 37 min. Subjects initially breathed room air for 15 min and duplicate measurements of oxygen consumption and carbon dioxide output were made between 8 and 13 min. The subjects then breathed a 21% oxygen mixture produced from the mass-¯ow controller system for a further 5 min to obtain a steady baseline. Mild hypercapnia was then induced for 10 min by increasing inspired carbon dioxide so that end-tidal PCO2 was raised by 0.7 kPa above the previous baseline level. Inspired oxygen concentration was then decreased so that SpO was reduced to 85% for 7 min. The end-tidal was maintained at 0.7 kPa above the initial baseline level throughout hypoxia. Mean values for all the ventilatory variables were calculated for each of the last 3 min of the baseline normoxic period, each of the last 3 min of the hypercapnic period, and each minute during hypercapnic hypoxia. In addition, mean values were calculated during the following 3-min periods of each run: normoxia (3 min before the onset of hypercapnia); hypercapnia (3 min before the onset of hypoxia): hypercapnic hypoxia (the last 3 min of hypoxia). The hypoxic ventilatory response was calculated as the ratio of the changes in VEinst and SpO occurring between the periods of hypercapnia and hypercapnic hypoxia (VEinst/ SpO ; litres min±1 %±1). Group means and SD for imposed ventilatory variables and group means and 95% con®dence limits (CL) for studied variables in the three measurement periods were calculated for each run. Student's t-test was used to compare anthropometric data in men and women, and a paired t-test for within-run comparison of end-tidal PCO2 during hypercapnia and hypoxia. Comparisons between runs for the group as a whole were made by one-way analysis of variance with post hoc comparisons made using the least signi®cant difference method. Comparisons between runs for males and females considered separately were made using Friedman's non-parametric analysis of variance.
212
2
2
2
Tramadol and the hypoxic ventilatory response
Signi®cance was set at P<0.05. Computations were made using SPSS version 9 for Windows.
Results
in the period after tramadol when end-tidal PCO2 was slightly but signi®cantly greater during hypercapnic hypoxia than hypercapnia alone (Table 2). The mean difference after tramadol was 0.04 and the maximum-recorded increase of 0.08 kPa is within the error of the measurement.
Of the 20 subjects recruited to the study, two subjects (one male) were withdrawn from the study after experiencing severe nausea after taking tramadol. Three female subjects were unable to return for day 3 of the study because of very irregular menstrual cycles. One further male subject completed the study but had a markedly irregular breathing pattern in the ®nal limb. The ventilatory variables could not be analysed accurately and the data from this subject were therefore excluded from the ®nal analysis. Anthropometric data on the 14 subjects included in the analysis are given in Table 1. The men were signi®cantly taller and had a higher BSA than the women, but there were no signi®cant differences in age, weight, or calculated drug dose in mg kg±1. The pattern of ventilatory response was similar in all four limbs of the study. Ventilation increased slightly during mild hypercapnia, with a further substantial increase and little secondary decline during the 7 min of hypoxia (Fig. 1). Within runs, end-tidal PCO2 was controlled, on average, to within 0.02 kPa of the hypercapnic level during hypercapnic hypoxia. End-tidal PCO2 did not differ signi®cantly between the periods of hypercapnia and hypercapnic hypoxia except Table 1 Means and SD or range for the subject details and drug dosage. *Signi®cant difference (P<0.05) between males and females Males
Females
7 20.6 (18±22) 179 (10) 73.7 (9.4) 1.9 (0.2) 1.4 (0.2)
7 21.3 (20±22) 168 (5)* 65.0 (8.6) 1.7 (0.1)* 1.6 (0.2)
Fig 1 Oxygen saturation (SpO ), instantaneous minute ventilation (VEinst) and end-tidal PCO2 (PE¢CO ) plotted against time for both placebo (lefthand side) and tramadol (right-hand side) studies for baseline (open circles) and placebo or tramadol (closed circles). Mean values 6 SD for each 1-min period are presented for the entire group. Mean values for each variable were calculated for the last 3 min of normoxia (A1), the last 3 min of hypercapnia (H1), and the last 3 min of hypercapnic hypoxia (H2) for each subject and used for comparison between runs. 2
n Age (yr) Height (cm) Weight (kg) BSA (m2) Drug dose (mg kg±1)
2
Table 2 Means (SD) for the imposed variables for the studied variables in the three measurement periods for baseline, placebo and tramadol runs. *Signi®cant difference between hypercapnic and hypercapnic hypoxia periods
End-tidal PCO2 (kPa)
End-tidal PCO2 (kPa)
SpO (%) 2
Normoxia
Hypercapnia
Hypercapnic hypoxia
Baseline Placebo Baseline Tramadol
5.27 5.14 5.08 5.39
5.97 5.96 5.83 5.82
5.99 5.97 5.86 5.86
(0.48) (0.48) (0.37) (0.39)*
Baseline Placebo Baseline Tramadol
12.87 13.72 13.77 13.49
6.15 6.16 6.11 6.07
(0.61) (0.65) (0.90) (0.98)
Baseline Placebo Baseline Tramadol
97.5 97.0 97.6 97.5
85.1 85.2 85.0 85.1
(0.3) (0.3) (0.5) (0.4)
(0.48) (0.46) (0.46) (0.34) (2.37) (0.57) (0.80) (0.75)
(0.7) (0.7) (0.7) (0.7)
213
(0.52) (0.51) (0.38) (0.39)
13.02 13.84 13.85 13.54 97.6 97.1 97.6 97.4
(2.17) (0.63) (0.81) (0.69)
(0.7) (0.5) (0.6) (0.8)
Warren et al. Table 3 Mean (95% con®dence limits) baseline metabolic values
Baseline Placebo Baseline Tramadol
Oxygen consumption (litres min±1)
Carbon dioxide output (litres min±1)
Gas exchange ratio
209 205 222 202
163 162 179 158
0.78 0.78 0.81 0.78
(183, (184, (190, (173,
234) 225) 254) 232)
(140, (140, (154, (132,
186) 183) 204) 180)
(0.74, (0.74, (0.78, (0.75,
0.82) 0.83) 0.84) 0.81)
Table 4 Means (95% con®dence limits) for the studied variables in the three measurement periods for baseline, placebo and tramadol runs. *Indicates signi®cant difference from placebo baseline, placebo and tramadol baseline values (P<0.05) Normoxia VEinst (litres min±1)
VT (litre)
f (b.p.m.)
Hypercapnia
Hypercapnic hypoxia
Baseline Placebo Baseline Tramadol
7.39 7.72 8.21 7.54
(6.43, (6.41, (7.31, (6.48,
8.33) 8.53) 9.11) 8.6)
13.93 (11.94, 14.95 (12.67, 14.63 (13.02, 11.28* (9.27,
Baseline Placebo Baseline Tramadol
0.50 0.48 0.52 0.45
(0.43, (0.42, (0.45, (0.40,
0.56) 0.55) 0.58) 0.49)
0.83 (0.72, 0.93) 0.85 (0.73, 0.97) 0.85 (0.73, 0.97) 0.65* (0.56, 0.73)
1.31 1.34 1.31 1.02
(1.10, (1.10, (1.11, (0.85,
1.52) 1.59) 1.50) 1.19)
Baseline Placebo Baseline Tramadol
15.3 16.8 16.7 17.2
(13.1, (14.2, (14.1, (14.5,
17.5) 19.4) 19.3) 20.0)
17.2 18.1 18.1 17.8
20.2 21.8 19.7 19.7
(16.9, (17.1, (16.8, (16.2,
23.4) 26.5) 22.5) 23.1)
(15.7, (15.8, (15.6, (14.9,
15.93) 17.23) 16.23) 13.29)
26.14 28.99 24.45 19.64
19.2) 20.4) 20.5) 20.6)
(20.00, (21.14, (21.25, (15.24,
32.27) 36.83) 27.65) 24.04)
Fig 2 Plots of the VEinst/SpO relationship for the placebo (upper panel) and tramadol (lower panel) limbs of the study for the entire group. Baseline studies are shown as open circles and studies after placebo or tramadol are shown as closed circles. Data are given as mean values and con®dence intervals. 2
Fig 3 Plots of the individual VEinst/SpO relationships measured before and after either placebo or tramadol. The data are grouped for males (upper panel) and females (lower panels). A lesser slope indicates a decreased hypoxic ventilatory drive. 2
There were no signi®cant differences in oxygen consumption, carbon dioxide output or the gas exchange ratio between the four runs (Table 3). During normoxia there were no signi®cant differences in the imposed variables between the four runs (Table 2), although end-tidal PCO2 tended to be greater and end-tidal PCO2 tended to be less after tramadol. Similarly, there were no signi®cant differences between runs in the studied
variables although VEinst and VT tended to be less during the tramadol run (Table 4). During hypercapnia there were no signi®cant differences between runs in the imposed variables (Table 2), but both VT and VEinst during the tramadol
214
Tramadol and the hypoxic ventilatory response
run were signi®cantly less than during either of the baseline runs or the placebo run (Table 4). During hypercapnic hypoxia there were no signi®cant differences in either imposed or studied variables, although again VEinst tended to be less after tramadol. There was no signi®cant difference between the VEinst/ SpO ratio for the placebo baseline (mean ±0.99, CL ±0.59, ±1.40 litres min±1 %±1), placebo (mean ±1.18, CL ±0.66, ±1.70 litres min±1 %±1), tramadol baseline (mean ±0.78, CL ±0.57, ±0.98 litres min±1 %±1), and tramadol (mean ±0.68, CL ±0.43, ±0.93 litres min±1 %±1) runs (Fig. 2). Similarly, there were no signi®cant differences in the VEinst/SpO ratios between the four runs when men and women were considered separately (Fig. 3). 2
2
Discussion We have shown that a standard analgesic dose of tramadol in normal subjects had no signi®cant effect on the acute ventilatory response to isocapnic hypoxia. During mild hypercapnia, VEinst was signi®cantly less after tramadol compared with either the baseline run or after placebo in spite of similar levels of end-tidal PCO2. This ®nding is consistent with a small effect of tramadol on the ventilatory response to hypercapnia.4 Although tramadol did not have a signi®cant effect on normoxic ventilation, the fact that VEinst and VT tended to be ~10% less and end-tidal PCO2 ~0.3 kPa greater after tramadol than during the tramadol baseline run is similar to changes reported previously in anaesthetized patients.3 7 Studies were performed on two separate days because the 5-h elimination half-life of tramadol10 meant that the order of the placebo and active drug runs could not be randomized if the studies were performed on the same day. However, the hypoxic ventilatory response is known to vary substantially within and between days.11 Individuals were studied at approximately the same time of day to minimize the effect of diurnal variation. To overcome the between-day variability, a baseline study was performed on each day and the effect of placebo or tramadol compared with the respective baseline measurements. We were not able to measure serum tramadol concentrations to ensure that adequate concentrations had been achieved. A single oral dose of tramadol 100 mg given to fasting normal subjects is absorbed rapidly after an initial lag time of ~30 min, with serum concentrations reaching the analgesic threshold of 100 ng litre±1 after ~40 min and reaching peak levels at 2 h.10 However, there is considerable individual variability in bioavailability, with times for peak serum concentration varying between 1 and 4 h.10 In this study, the repeated measurements were started 1 h after drug administration so that the repeated hypoxic challenge occurred after ~90 min when serum concentrations would be approaching maximum in most subjects. While this ensured that the hypoxic ventilatory response was unlikely to be depressed as a result of previous hypoxic exposure,12 it
is possible that peak serum concentrations had not been reached in some subjects. However, there was a reduction in hypercapnic VEinst of over 1.5 litres in nine out of 14 subjects after tramadol, whereas this only occurred in two subjects after placebo suggesting that active levels of tramadol had been achieved in most subjects. It is therefore unlikely that low drug concentrations account for the lack of depression of the hypoxic ventilatory response. Although 20 subjects were recruited, data were only obtained in 14 for a variety of reasons. Failure to detect a signi®cant effect of tramadol on the hypoxic ventilatory response might therefore be a type 2 statistical error. However, this is unlikely to be the explanation as the study had a power of 0.9 to detect a change in the VEinst/SpO ratio of 0.5. Dahan and colleagues6 found a sex-dependent effect on the hypoxic ventilatory response which was independent of differences in weight, lean body mass, body surface area and calculated fat mass. In their study, morphine signi®cantly reduced the initial ventilatory response to sustained hypoxia in women but not in men. We studied similar numbers of men and women. It was possible, therefore, that an effect in women was being masked by an absence of effect in the men. However, separate analysis of the results in seven men and eight women who completed the study showed no effect of tramadol on the initial hypoxic ventilatory response in either sex. The O-desmethyltramadol metabolite of the (+) enantiomer of tramadol, produced by hepatic phase I metabolism by cytochrome P-450 2D6 (CYP2D6), has an af®nity for the mopioid receptor which is ~200 times that of the parent compound.13 It is likely, therefore, that the bulk of any ventilatory depression would be mediated by this metabolite. Approximately 10% of the Caucasian population are phenotypically poor metabolizers14 and have a reduced analgesic effect of tramadol.13 If, by chance, most of our subjects were poor metabolizers, then this could explain the absence of effect on the hypoxic ventilatory response. This effect could contribute to our ®ndings but is unlikely to be the only explanation as tramadol suppressed the hypercapnic ventilatory response in nine out of 14 of our subjects. The lack of suppression by tramadol of the hypoxic ventilatory response contrasts with the 50±60% suppression produced by morphine. Unlike morphine, tramadol produces its analgesic effects by a combination of low but preferential activity at m-opioid receptors,1 by inhibiting both noradrenaline and 5-hydroxytrypamine (5-HT) uptake,1 and by facilitating 5-HT release.15 The effects of these non-opioid actions of tramadol on ventilation are dif®cult to assess. A recent study showed that the 5-HT reuptake inhibitor ¯uoxetine caused a slight depression in resting ventilation in goats.16 However, 5-HT can stimulate or depress ventilation depending on the subtype of receptor affected and the type of respiratory neurone activated.17 Catecholamines have a similar dual effect on ventilation depending in part on whether the action is central or peripheral. Catecholamines administered to the respiratory
215
2
Warren et al.
neurones in the medulla usually cause depression,18 whereas systemic administration stimulates breathing19 and increases the ventilatory response to hypercapnia.20 It is possible, therefore, that stimulation by 5-HT and noradrenaline offsets the opioid depressant effects of tramadol. The relative opioid receptor af®nity of tramadol may also explain the ventilatory effects. In cats, depression of peripheral chemosensory discharge is mediated via d receptors.21 Tramadol has modest af®nity for m-receptors but extremely weak d af®nity. Tramadol may therefore suppress the ventilatory response to hypercapnia4 via central m-receptor activity leaving the peripheral component relatively intact. Studies with selective opioid receptor blockers would be required to con®rm this. In conclusion, we have shown that tramadol, in a standard clinical dose, does not signi®cantly affect the ventilatory response to hypoxia. However, tramadol reduced resting tidal volume and insigni®cantly reduced ventilation, resembling the effects reported by Vickers and colleagues3 at a dose of 1 mg kg±1. Seitz and colleagues4 also showed that tramadol caused a moderate depression of the hypercapnic ventilatory response. The signi®cant reduction in hypercapnic ventilation after tramadol is in keeping with previous observations that tramadol causes a moderate depression of the hypercapnic ventilatory response,4 and suggests that this suppression can be signi®cant even at values which are only just above the normoxic resting level.
Acknowledgement This study was funded by the Moray Foundation, University of Edinburgh.
References 1 Collart L, Luthy C, Dayer P. Partial inhibition of tramadol antinociceptive effect by naloxone in man. Br J Clin Pharmac 1993; 35: 73P 2 Raffa RB, Friderichs E, Reimann W, Shank RP, Codd EE, Vaught JL. Opioid and nonopioid components independently contribute to the mechanism of action of tramadol, an `atypical' opioid analgesic. J Pharm Exp Therap 1992; 260: 275±85 3 Vickers MD, O'Flaherty D, Szekely SM, Read M, Yoshizumi J. Tramadol: pain relief by an opioid without depression of respiration. Anaesthesia 1992; 47: 291±6 4 Seitz W, Lubbe N, Fritz K, Sybrecht G, Kirchner E. Ein¯uû von Tramadol auf die ventilatorische CO2-Antwort und den Mundokklusionsdruck. Anaesthesist 1985; 34: 241±6
216
5 Weil JV, McCullough RE, Kline JS, Sodal IE. Diminished ventilatory response to hypoxia and hypercapnia after morphine in normal man. N Engl J Med 1975; 292: 1103±6 6 Dahan A, Sarton E, Teppema L, Olievier C. Sex-related differences in the in¯uence of morphine on ventilatory control in humans. Anesthesiology 1998; 88: 903±13 7 Tarkkila P, Tuominen M, Lindgren L. Comparison of respiratory effects of tramadol and pethidine. Eur J Anaesth 1998; 15: 64±8 8 Jones JG, Sapsford DJ, Wheatley RG. Postoperative hypoxaemia: mechanisms and time course. Anaesthesia 1990; 45: 566±73 9 Young CH, Drummond GB, Warren PM. Effect of a subanaesthetic concentration of halothane on the ventilatory response to sustained hypoxia in healthy humans. Br J Anaesth 1993; 71: 642±7 10 Lintz W, Barth H, Osterloh G, Schmidt-Bothelt E. Bioavailability of enteral tramadol formulations. Arzneim±Forsch/Drug Res 1986; 36: 1278±83 11 Sahn SA, Zwillich CW, Dick N, McCullough RE, Lakshiminaryan S, Weil JW. Variability of ventilatory responses to hypoxia and hypercapnia. J Appl Physiol 1977; 43: 1019±25 12 Easton PA, Slykerman LJ, Anthonisen NR. Recovery of the ventilatory response to hypoxia in normal adults. J Appl Physiol 1988; 64: 521±8 13 Poulsen L, Arendt-Nielsen L, Brosen K, Sindrup SH. The hypoalgesic effect of tramadol in relation to CYP2D6. Clin Pharmacol Ther 1996; 60: 636±44 14 Paar WD, Poche S, Gerloff J, Dengler HJ. Polymorphic CYP2D6 mediates O-demethylation of the opioid analgesic tramadol. Eur J Clin Pharmacol 1997; 53: 235±9 15 Driessen 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: 147±51 16 Henderson DR, Konkle DM, Mitchell GS. Effects of serotonin reuptake inhibition on ventilatory control in goats. Respir Physiol 1999; 115: 1±10 17 Bianchi AL, Denavit-Saubie M, Champagnat J. Central control of breathing in mammals: neuronal circuitry, membrane properties, and neurotransmitters. Physiol Rev 1995; 74; 1±45 18 Champagnat J, Denavit-Saubie M, Henry JL, Leviel V. Catecholaminergic depressant effects on bulbar respiratory mechanisms. Brain Res 1979; 160: 57±68 19 Lundholm L, Svedmyr N. Studies on the stimulating effects of adrenaline and noradrenaline on respiration in man. Acta Physiol Scand 1966; 67: 65±75 20 Cunningham DJC, Hey EN, Lloyd BB. The effect of intravenous infusion of noradrenaline on the respiratory response to carbon dioxide in man. Quart J Exp Physiol 1958; 43: 394±9 21 Kirby GC, McQueen DS. Characterization of opioid receptors in the cat carotid body involved in chemosensory depression in vivo. Br J Pharmac 1986; 88: 889±98