This study gathers information in humans on the sites of sex-related differences in ventilatory depression caused by the mu-opioid receptor agonist morphine.
Experiments were performed in healthy young men (n = 9) and women (n = 7). Dynamic ventilatory responses to square-wave changes in end-tidal carbon dioxide tension (7.5-15 mmHg) and step decreases in end-tidal oxygen tension (step from 110 to 50 mmHg, duration of hypoxia 15 min) were obtained before and during morphine infusion (intravenous bolus dose 100 microg/kg, followed by 30 microg x kg(-1) x h(-1)). Each hypercapnic response was separated into a fast peripheral and slow central component, which yield central (Gc) and peripheral (Gp) carbon dioxide sensitivities. Values are mean +/- SD.
In carbon dioxide studies in men, morphine reduced Gc from 1.61 +/- 0.33 to 1.23 +/- 0.12 l x min(-1) x mmHg(-1) (P < 0.05) without affecting Gp (control, 0.41 +/- 0.16 and morphine, 0.49 +/- 0.12 l x min(-1) x mmHg(-1), not significant). In carbon dioxide studies in women, morphine reduced Gc, from 1.51 +/- 0.74 to 1.17 +/- 0.52 l x min(-1) x mmHg(-1) (P < 0.05), and Gp, from 0.54 +/- 0.19 to 0.39 +/- 0.22 l x min(-1) x mmHg(-1) (P < 0.05). Morphine-induced changes in Gc were equal in men and women; changes in Gp were greater in women. In hypoxic studies, morphine depressed the hyperventilatory response at the initiation of hypoxia more in women than in men (0.54 +/- 0.23 vs. 0.26 +/- 0.34 l x min(-1) x %(-1), respectively; P < 0.05). The ventilatory response to sustained hypoxia (i/e., 15 min) did not differ between men and women.
The data indicate the existence of sex differences in morphine-induced depression of responses mediated via the peripheral chemoreflex pathway, with more depression in women, but not of responses mediated via the central chemoreflex pathway. In men and women, morphine did not change the translation of the initial hyperventilatory response to short-term hypoxia into the secondary decrease in inspired minute ventilation (Vi) caused by sustained hypoxia.
IN humans, we recently showed important sex differences in the influence of morphine on the steady state ventilatory response to carbon dioxide and the semi-steady state or peak ventilatory response to acute hypoxia (i.e., 3 min).  Compared with morphine in men, morphine in women caused significantly more depression of both responses (in either phase of their menstrual cycle). We argued that these findings are analogous to sex-related differences in opioid-mediated analgesia [2–4] and that the bases of our findings are biologic differences in the pharmacodynamics of morphine rather than pharmacokinetic differences. 
The main aim of this study was to gain information regarding the sites of the sex-related ventilatory effects of morphine within the ventilatory control system. To that purpose, we studied the influence of morphine on the dynamic ventilatory response to square-wave changes in end-tidal carbon dioxide pressure (PETCO(2)) and to step decreases in end-tidal oxygen pressure (PETO(2)) in healthy young men and women. Using a mathematical model of the ventilatory controller, we partitioned the breath-to-breath ventilatory responses to carbon dioxide into a fast component from the peripheral chemoreflex loop and a slow component from the central chemoreflex loop.  Both components are characterized by a carbon dioxide sensitivity, time constant, time delay, and off-set. This model has been used successfully to study the influence of catecholamines, anesthetics, and opioids on the respiratory control system in humans and animals. [6–9]
Short-term hypoxic input to the ventilatory control system, as used in our previous study, causes brisk increases in ventilation (the acute hypoxic response). However, if hypoxia is maintained for periods longer than 3 min, a slow decrease in ventilation (hypoxic ventilatory decrease) becomes apparent.  A new steady state inspired minute ventilation (VI) is reached within 15–20 min. To compare the influence of morphine on the development of the hypoxic ventilatory decrease in men and women, the ventilatory response to isocapnic hypoxia was evaluated over a 15-min period.
Subjects and Methods
Eighteen volunteers (9 men, 9 women, aged 18–35 yr) were recruited to participate in this protocol, which was approved by the Leiden University Medical Center Human Ethics Committee. The subjects were healthy and did not have a history of illicit substance abuse. Because all subjects participated in previous respiratory studies in our laboratory, there was no need for familiarization with the apparatus or procedures. The women reported normal menstrual cycles (they did not use oral contraceptives). Because we did not study the influence of menstrual cycle per se, we did not prospectively control for the phase of the menstrual cycle.
After arrival in the laboratory, an intravenous catheter was inserted in a vein of the arm or the hand, through which saline was given by infusion pump (Becton Dickinson, St. Etienne, France) at a rate of 6 ml/h. Subsequently, the subjects rested for 30–45 min. Next, a mask (Vital signs, Totowa, NJ) was fitted over the nose and the mouth and the experiments were started. The inspired and expired gas flows were measured using a pneumotachograph, electronically integrated to yield a volume signal, connected to a pressure transducer. Corrections were made for the changes in gas viscosity caused by 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 DEC PDP 11/23 microcomputer (Digital Equipment, Maynard, IA) provided control signals to the mass-flow controllers so the composition of the inspired gas mixtures could be adjusted to force PETO(2) and PETCO(2) to follow an specified pattern in time. The oxygen and carbon dioxide concentrations of inspired and expired gases and the arterial hemoglobin-oxygen saturation (Sp (O)(2)) were measured using a Datex Multicap gas monitor and Datex Satelite Plus pulse oximeter, respectively (Datex-Engstrom, Helsinki, Finland).
Control (carbon dioxide and hypoxic) studies preceded morphine studies. In both studies, carbon dioxide studies preceded hypoxic studies. Initially resting PETCO(2) and Vivalues (i.e., without inspired carbon dioxide) were measured for 40 min in control and morphine studies. Between studies there was ample time for resting.
Carbon Dioxide Studies. PETCO(2) was increased by approximately 3 mmHg from the resting value. This value was maintained for 5 min, after which a step increase in PETCO(2) of 7.5–15 mmHg was applied. This high level was kept constant for 7–9 min. Subsequently, PETCO(2) was decreased rapidly to its original value and maintained for another 7–9 min. This sequence was repeated two to three times in each treatment (control or morphine). The PETO(2) was maintained at 110 mmHg throughout the studies.
Hypoxic Studies. PETCO(2) was maintained at approximately 8 mmHg above resting control values in control and morphine hypoxic studies (i.e., isocapnia). This allows hypoxic responses to be performed at identical PETCO(2) levels. The PETO(2) was forced as follows:(1) 10 min at 110 mmHg, (2) a rapid decrease to 50 mmHg, (3) 15 min at 50 mmHg, and (4) at least 5 min at more than 300 mmHg.
Morphine Administration. Morphine was administered from an adjacent room via a 4-m line to prevent disturbance of the subjects. After completion of the control studies and a 20-min resting period, the saline in the syringe was replaced by morphine. The initial morphine dose was 100 [micro sign]g/kg, followed by a continuous infusion of 30 [micro sign]g [middle dot] kg-1[middle dot] h-1. The morphine study started 40 min after the bolus was administered.
Resting Values. Before clamping of PETO(2) and PETCO(2), resting Viand PETCO(2) values were determined after at least 10 min of steady state Viby averaging the breath-to-breath Viand PET (CO)(2) data over a 1-min period.
Carbon Dioxide Studies. The steady state relation between Viand PETCO(2) at constant PETO(2) is described by Vi=(Gp+ Gc)(PETCO(2)- B) where Gcand Gpare the central and peripheral ventilatory carbon dioxide sensitivities, and B represents the apneic threshold or extrapolated PETCO(2) at zero Vi. To estimate G (c), Gp, and B, we fitted the breath-to-breath data to a two-compartment model, as described previously.  We report all parameters of the model: time constant of the peripheral chemoreflex loop ([Greek small letter tau]P); time constant of the on-response of the central chemoreflex loop, i.e., at high PETCO(2)([Greek small letter tau]ON); time constant of the off-response of the central chemoreflex loop, i.e., at low PETCO(2)([Greek small letter tau]OFF); time delays of the central and peripheral chemoreflex loops (TCand TP); and Gc, Gp, and B (Figure 1).  Parameter estimates were obtained using a least-squares method.
Hypoxic Studies. Mean values of the breath-to-breath data were chosen over identical time segments. Normoxia was the 1-min period before the hypoxic step; early hypoxia was min 3 of hypoxia; late hypoxia was min 15 of hypoxia. We defined the following responses: acute hypoxic response (AHR), which was early hypoxia - normoxia and sustained hypoxic response (SHR), which was late hypoxia - normoxia. The Viresponses are expressed as the change in Viper percentage change in SpO(2)(units: 1 [middle dot] min-1[middle dot]%-1). We calculated the difference between acute and sustained hypoxic responses as a measure of the hypoxic ventilatory decrease (HVD).
Morphine versus Control. To detect the significance of difference between the control and morphine measurements, a two-way analysis of variance was performed on individual parameter values of the carbon dioxide studies, and a paired t test was performed on parameter values of the hypoxic studies. Separate analyses were performed in men and women.
Men versus Women. Because each morphine study was preceded by control measurements, Student t tests were performed on the morphine-induced parameter differences (i.e., Delta s). Finally, Student t tests were performed on SpO(2), PETCO(2), and PETO(2) at normoxia and early and late hypoxia. Probability levels less than 0.05 were considered significant. Values are the mean +/- SD.
Two women did not complete the protocols because of the occurrence of nausea and vomiting. These data were discarded. The demographic data show that the remaining men and women did not differ with respect to age (men: 27 +/- 5 yr vs. women: 27 +/- 4 yr). Compared with the women, the men were heavier (78 +/- 7 vs. 61 +/- 8 kg, t test, P < 0.05) and taller (181 +/- 4 vs. 171 +/- 3 cm, P < 0.05). Three women were observed in the follicular phase and four were observed in the luteal phase of their menstrual cycles.
Resting Values. An example of the changes in resting Viand PETCO(2) is given in Figure 2for one subject. In both sexes, morphine reduced Viand increased PETCO(2) to the same extent. In men, resting Videcreased from 10.9 +/- 1.7 to 9.0 +/- 1.5 l/min (P < 0.001, paired t test), with a concomitant increase in PETCO(2) from 39.5 +/- 1.4 to 43.5 +/- 1.8 mmHg (P < 0.0001). In women, Videcreased from 10.5 +/- 1.4 to 8.0 +/- 1.3 l/min (P < 0.01) and PETCO(2) increased from 38.0 +/- 1.5 to 42.0 +/- 2.3 mmHg (P = 0.001).
Carbon Dioxide Studies. The model fits of a control and morphine study in a female subject are shown in Figure 3. They exemplify the general observation in women that, after administration of morphine, the contributions of the peripheral (VPin Figure 3) and central (VC) chemoreflex loops to total Viare both decreased. Mean values of the estimated parameters are given in Table 1. In men, morphine increased the apneic threshold (B) and decreased the central carbon dioxide sensitivity, but did not affect the peripheral sensitivity. As a result Gp/Gcwas increased. In women, morphine did not change the apneic threshold, but reduced both the central and the peripheral carbon dioxide sensitivities, causing similar values of Gp/Gcin control and morphine studies.
Hypoxic Studies. The mean values of ventilatory variables are presented in Table 2 and Table 3. Although a small trend toward lower Sp (O)(2) values at late hypoxia was apparent in women, the magnitude of the early and late hypoxic challenges did not differ between men and women in control and morphine studies (P > 0.05). In Figure 4, an example of a male control and morphine response is given. It shows the biphasic nature of the ventilatory responses to sustained isocapnic hypoxia in control and morphine studies.
Morphine-induced reduction of the acute hypoxic response was greater in women (60% depression) than in men (20% depression). The response to sustained hypoxia was affected equally in both sexes (60% depression in men and women). The latter was related to a smaller hypoxic ventilatory decrease after morphine in women than in men (see Figure 5and Table 4).
Because there is ample evidence that the magnitude of the HVD is related to that of the acute hypoxic response, we plotted HVD against AHR on the pooled data sets (Figure 6) and performed linear regression. In the control study, the slope was 0.5; the y intercept was 0.01 l [middle dot] min (-1)[middle dot]%-1(r2= 0.5). In the morphine study, the slope was 0.6 and the y intercept was 0.00 l [middle dot] min-1[middle dot]%-1(r2= 0.6). These data show that morphine did not alter the relation between AHR and HVD.
The results of this study confirm our previous findings and conclusions that sex differences exist in the respiratory pharmacology of the [micro sign]-opioid receptor agonist morphine.  Our current study is an extension of our previous work. We observe that the respiratory effects of morphine differ in men and women because of the opioid action at sites within the peripheral chemoreflex loop, but not within the central chemoreflex loop. Furthermore, our data suggest, not only that the influence of morphine on the ventilatory response to acute hypoxia shows sex-dependency, but also that this applies to the influence of morphine on the magnitude of the hypoxic ventilatory decrease.
Morphine reduced the central ventilatory carbon dioxide sensitivity in men and women equally. This indicates that the result of the morphine-induced [micro sign] receptor activation within the central chemoreflex loop did not differ between the men and women. This observation is important because it gives further evidence that morphine concentrations at [micro sign]-opioid receptors within the central nervous system in general and the central chemoreflex loop in particular do not differ between men and women. We have no reason to doubt that this also would hold for morphine concentrations at sites within the peripheral chemoreflex loop (see also  for a discussion of possible sex differences in the pharmacokinetics of morphine).
Peripheral Chemoreflex Loop
The absence of an effect of morphine on the peripheral ventilatory carbon dioxide sensitivity in men, but an appreciable depressant effect in women, and the finding of significantly greater depression of the AHR in women indicates profound sex differences in those parts of the peripheral chemoreflex pathway that are not common to the peripheral and central chemoreflexes. Sex-related differences in responses mediated via the peripheral chemoreflex loop or in responses originating at the baroreceptors of the carotid bodies have been observed before, but none involved [micro sign]-opioid receptors. For example, physiologic adaptation to hypoxia of high altitude shows sex differences that are related to differences in catecholamine activity in certain areas of the afferent peripheral chemoreflex pathway.  Recent studies show sex dependency of baroreflex control in healthy humans. [12,13] Although we are unable to rule out an opioid receptor-independent sex difference in the functioning of the peripheral chemoreflex loop, the observation that control carbon dioxide and acute hypoxic responses did not differ between men and women strongly suggests that our findings are related to [micro sign]-opioid receptors at specific sites within the peripheral chemoreflex loop. This may involve [micro sign]-receptor density, binding affinity, response to activation, or interaction with other receptors at the carotid bodies or at specific sites within the central nervous system.
Animal studies show that carotid body stimulation, caused by hypoxia or hypercapnia, causes the increase of glutamate in the central nervous system (for example, in the nucleus tractus solitarius). [14–16] This indicates the importance of N-methyl-D-aspartate receptors-among other receptors-in the process of peripheral chemoafferent impulse transmission. [14–16] Studies on endogenous and exogenous opioid analgesia show that N-methyl-D-aspartate receptors modulate [micro sign]-opioid receptors and that this process is sex dependent. [17,18] An interesting hypothesis, analogous to the animal studies of opioid-induced analgesia, is that, in the nucleus tractus solitarius, activated [micro sign]-opioid receptors may interact with N-methyl-D-aspartate receptors activated by glutamate in a different fashion in men and women. As a consequence, more depression of carotid body responses after administration of morphine is observed in women than in men. Clearly, additional studies on the mechanisms of the observed sex differences in our studies are warranted. Our scenario may serve as an initial working hypothesis.
Ventilatory Response to Sustained Hypoxia
The origin of the biphasic response to sustained isocapnic hypoxia is only partly understood. The acute hypoxic response originates at the carotid bodies.  The mechanism and site of generation of the slow ventilatory decrease remain controversial. [20,21] Recent studies suggest a prominent role for the central accumulation or release of inhibitory neurotransmitters and modulators (for instance, adenosine, [Greek small letter gamma]-aminobutyric acid, dopamine, and, as recently shown, neuronal nitric oxide synthase). [22–25] Our finding that the magnitude of HVD after morphine administration differed in men and women suggests that activated [micro sign]-opioid receptors influenced the development of HVD in a sex-dependent fashion. However, before any such conclusion can be drawn, we need to discuss recent findings of the relation between AHR and HVD.
In human studies, it is observed that the magnitude of the HVD is related proportionally to the size of the acute or peak hypoxic response. Subjects with large AHRs have a larger magnitude of HVD and vice versa (Figure 6). Furthermore, if the AHR is increased with almitrine, the magnitude of HVD is increased.  The reverse is true for isoflurane, somatostatin, and dopamine. [9,27,28] The recent finding that an intact peripheral drive is necessary for the development of HVD led to the suggestion that the increased afferent input from the carotid bodies during hypoxia activates the buildup of inhibitory modulators in the brain stem and that, as a consequence, Videcreases. 
Our observation that the magnitude of AHR is related to that of HVD without and with morphine (see Figure 6) is in agreement with the previously mentioned observations and concepts. Figure 6shows further that the relation between AHR and HVD was not changed by morphine. This indicates that the translation of the peripheral drive into hypoxic ventilatory decrease was not affected-quantitatively or qualitatively-by morphine. Taking this into account, the difference in magnitude of HVD after the use of morphine in men and women is primarily related to the difference in magnitude of the acute response to hypoxia and not to sex per se.
At this point, we are unable to exclude a small interactive effect of sex and morphine on HVD, independent of the AHR. This may account for our inability to observe a small reduction of HVD related to the 20% reduction of the AHR in men. Conversely, the power of our study may have been inadequate to observe a small reduction of HVD in the nine men. However, further indications for an interactive sex-morphine effect on HVD come from recent data of Cartwright et al.  They showed in a predominantly male population that the [micro sign]-receptor agonist alfentanil, at a blood concentration of 37 ng/ml, caused an approximately 25% depression of the AHR, but did not change the magnitude of the HVD. Their findings with alfentanil are in agreement with our data regarding morphine in men. Evidently, more studies are needed to fully understand the complex interactions of opioid receptors, sex, AHR, and HVD. For now, we conclude that the influence of morphine on the development of HVD was relatively unimportant in both sexes. Furthermore, in awake humans, Kagawa et al.  showed no crucial role for endogenous opioids in the development of HVD, because high-dose naloxone did not alter the biphasic response to sustained hypoxia. These findings, together with ours, indicate that the decrease of Viduring sustained hypoxia is not mediated by (endogenous or exogenous) opioid peptides, at least not in healthy adults. 
In conclusion, we studied the influence of an analgesic dose of morphine on the dynamic ventilatory responses to carbon dioxide and sustained isocapnic hypoxia in healthy young men and women. Our data showed the existence of sex differences in morphine-induced depression of responses mediated via the peripheral chemoreflex pathway, with more depression found in women. No sex differences were observed in morphine-induced depression of responses mediated via the central chemoreflex pathway. Furthermore, morphine did not change the translation of the acute hypoxic response into the secondary depression of ventilation caused by sustained hypoxia. We argue that our findings are related to sex differences in [micro sign]-opioid receptors, their responses to activation, or interaction with other receptors at sites within the peripheral chemoreflex pathway.
The authors thank Dr. Ben Kest for his critical reading of the manuscript.