Treatment with leuprolide acetate decreases the threshold of the ventilatory response to carbon dioxide in healthy males

Authors


Corresponding author J. H. Mateika: John D. Dingell VA Medical Center, 4646 John R (11R), Room 4308, Detroit, MI 48201, USA. Email: jmateika@med.wayne.edu

Abstract

This investigation was designed to determine if suppression of testosterone alters the ventilatory response to carbon dioxide in the presence of high and low levels of oxygen. Eleven healthy male subjects completed a series of rebreathing trials during wakefulness, before and after treatment with a long-acting gonadotropin-releasing hormone agonist. Five subjects also completed studies during non-rapid eye movement (NREM) sleep. During wakefulness, subjects initially hyperventilated to reduce the partial pressure of carbon dioxide (PET,CO2) below 25 Torr. Subjects then rebreathed from a bag containing a normocapnic (42 Torr), low (50 Torr) or high oxygen (140 Torr) gas mixture. During each trial PET,CO2 increased while oxygen was maintained at a constant level. The threshold of the ventilatory response to carbon dioxide was considered to be the point at which minute ventilation began to rise in a linear fashion as PET,CO2 increased. The slope of the ventilatory response above the threshold was used as a measure of sensitivity to carbon dioxide. During NREM sleep, hypocapnia was induced via nasal mechanical ventilation. Several trials were completed until the cessation of mechanical ventilation resulted in a central apnoea which demarcated the threshold of the ventilatory response to carbon dioxide. In response to treatment with leuprolide acetate, the threshold measured in wakefulness decreased during carbon dioxide rebreathing in the presence of low (41.05 ± 0.77 versus 39.40 ± 0.83 Torr; P= 0.01) and high (46.32 ± 0.56 versus 44.78 ± 0.83 Torr; P= 0.01) oxygen levels. An increase in sensitivity (4.82 ± 0.61 versus 7.17 ± 1.20 l min−1 Torr−1; P= 0.02) was also observed during rebreathing in the presence of high but not low oxygen levels. The increase in sensitivity was accompanied by an increase in carbon dioxide production. The findings observed during NREM sleep were similar to those observed during wakefulness, since the PET,CO2 that demarcated the threshold was decreased after leuprolide treatment (42.1 ± 0.6 versus 39.6 ± 0.6 Torr; P= 0.002). Additionally, the decrease in PET,CO2 required to induce an apnoea was greater after treatment with leuprolide (2.56 ± 0.25 versus 4.06 ± 0.29 Torr; P= 0.004). We conclude that suppression of testosterone decreases the threshold of the ventilatory response to carbon dioxide during both wakefulness and sleep.

Sleep-disordered breathing is more prevalent in men then women (Young et al. 1993; Bixler et al. 2001). The difference in prevalence might be due in part to the impact that testosterone has on the neurochemical control of breathing. Testosterone may impact on the susceptibility to sleep apnoea via alterations in the threshold and/or the sensitivity of the ventilatory response to carbon dioxide. Studies completed during wakefulness have reported that the sensitivity of the response to hypercapnia (Saunders et al. 1972; Patrick & Howard, 1972; White et al. 1983; Aitken et al. 1986; Sebert et al. 1990; van Klaveren & Demedts, 1998) and isocapnic hypoxia (Patrick & Howard, 1972; White et al. 1983; van Klaveren & Demedts, 1998) is increased in males compared to females, and after the administration of testosterone in hypogonadal men (White et al. 1985). Moreover, studies completed during sleep have shown that the threshold of the ventilatory response to carbon dioxide is closer to resting measures of carbon dioxide and that the sensitivity to carbon dioxide is greater in males compared to females (Zhou et al. 2000), and in females after the administration of testosterone (Zhou et al. 2003).

Given these findings, we hypothesized that suppression of testosterone in males would lead to a reduction in the threshold and sensitivity of the ventilatory response to carbon dioxide in the presence of low and high oxygen levels. Moreover, we hypothesized that these alterations would manifest themselves during both wakefulness and sleep. To test our hypotheses, we measured the threshold and sensitivity in the presence of low and high oxygen levels during wakefulness before and after the suppression of testosterone in healthy men. Moreover, we measured the threshold during non-rapid eye movement (NREM) sleep before and after the suppression of testosterone.

Methods

Protocol overview

Wakefulness All the experimental protocols described were approved by the Human Investigation Committees of Wayne State University School of Medicine and Detroit Veterans Affairs Medical Center, and they conformed to the standards set by the Declaration of Helsinki. Eleven healthy males visited our laboratory during daytime hours on four occasions, after giving informed consent. The first visit was used to familiarize subjects with the experimental apparatus and the modified rebreathing protocol. On the second visit subjects completed six rebreathing trials. Immediately after completing the six trials, subjects received an injection of leuprolide acetate (7.5 mg depot; TAP Pharmaceuticals, Inc., IL, USA), which is a gonadotropin-releasing hormone agonist. One month later, subjects received a second injection (i.e. third visit). One month thereafter (i.e. fourth visit), subjects returned to the laboratory and repeated the six rebreathing trials. Testosterone, total oestrogen, oestradiol, luteinizing and follicle stimulating hormone levels were measured at baseline and 60 days after leuprolide injection. Prior to the second and fourth visits, subjects were asked to avoid eating or drinking caffeinated beverages 4 h prior to the experiment. In addition, during these visits, subjects completed the rebreathing trials at the same time of day.

NREM sleep Five subjects (3 of the initial 11 described above, in addition to 2 others) also completed studies during NREM sleep, before and after leuprolide treatment. Subjects were instructed to restrict their sleep the night prior to each sleep study (i.e. total sleep time was no more than 3 h). The sleep studies were conducted during regular sleep hours.

Experimental techniques

Modified rebreathing protocol (wakefulness) A modified rebreathing protocol (Mohan et al. 1999; Duffin et al. 2000; Mateika & Ellythy, 2003; Mateika et al. 2004) was employed during visit nos. 2 and 4 to measure the threshold and sensitivity of the ventilatory response to carbon dioxide during wakefulness. During each rebreathing trial, the subjects initially breathed room air for 5 min. Subsequently, subjects hyperventilated for 5 min while maintaining an end-tidal partial pressure of carbon dioxide (PET,CO2) between 20 and 25 Torr. After hyperventilating, the subjects were switched from room air to a rebreathing bag. The end-tidal partial pressure of oxygen (PET,O2) in the bag during three of the six experiments completed during the second and fourth visits was 50 Torr while the PET,O2 for the remaining three experiments was 140 Torr. The pressures were maintained throughout the rebreathing experiment (see below for further details) and the PET,O2 (50 versus 140 Torr) selected for the initial two experiments was random. The PET,CO2 in the bag at the start of the rebreathing experiment was 42 Torr. The trials from hereon will be referred to as the iso-oxic hypoxic (50 Torr) and hyperoxic (140 Torr) carbon dioxide rebreathing trials.

Rebreathing began at the end of expiration, and it was followed by three rapid and deep breaths that produced rapid equilibration of the carbon dioxide partial pressures in the bag, lungs and arterial blood to that of mixed venous blood. Rebreathing continued until PET,CO2 increased to a maximum of 10 Torr above the threshold. In all subjects, this point of termination occurred prior to the breathing frequency threshold that has been described (Mohan et al. 1999; Duffin et al. 2000).

During the rebreathing experiments, the subjects wore nose-clips and breathed through a mouthpiece that was connected to a pneumotachograph (Model RSS100-HR, Hans Rudolph, Inc., Kansas, MO, USA) that was used to monitor breath-by-breath changes in ventilation. The pneumotachograph was attached to one side of a three-way valve that allowed us to switch the subjects from room air to the rebreathing bag. End-tidal oxygen (Model 17518, Vacumed, Inc.,Ventura, CA, USA) and carbon dioxide (Model 17515, Vacumed, Inc.) were sampled from the pneumotachograph side of the three-way valve. The gas that was sampled for end-tidal monitoring was returned to the bag during rebreathing. The oxygen level in the bag during rebreathing was maintained by a flow of oxygen that was computer controlled. If oxygen decreased below the desired threshold (50 or 140 Torr), oxygen was immediately bled into the bag. Oxygen saturation was monitored using a pulse oximeter (Biox 3700, Ohmeda Corp., Boulder, CO, USA).

A 16-bit analog to digital converter (National Instruments, AT-MIO-16XE-50) digitized the analog signals for on-line computer analysis using software specifically designed for this purpose. The software calculated tidal volumes, breathing frequency, ventilation and end-tidal partial pressures of carbon dioxide and oxygen on a breath-by-breath basis.

Mechanical ventilation protocol (NREM sleep) To determine the threshold during NREM sleep, subjects were hyperventilated using a pressure support ventilator (Quantum PSV, Healthdyne Technologies, Marietta, GA, USA), as previously described (Zhou et al. 2000, 2003). Prior to and after mechanical hyperventilation, the ventilator was set at an expiratory positive airway pressure of 2.0 cmH2O. This was the minimum expiratory positive airway pressure allowed by the device. During periods of hyperventilation, inspiration was triggered by the subject, and inspiratory and expiratory times were matched to each subject's baseline eupnoeic rate. Hyperventilation was achieved by increasing the inspiratory pressure of the ventilator, with adjustments made during expiration. For each successive trial, the inspiratory pressure was increased in 1.0 cmH2O increments from the initial level of 2.0 cmH2O, which resulted in an increased tidal volume. Mechanical ventilation was continued for 3 min and was terminated during expiration. Each trial was repeated twice with trials separated by a minimum of 3 min. All trials were performed during stable NREM sleep. The post-mechanical ventilation period, which will be referred to hereafter as the recovery period, was examined for the presence of a central apnoea. Apnoea was defined as a period of no airflow for at least 5.0 s.

During the sleep studies, electroencephalograms (EEG), electrooculograms (EOG) and chin electromyogram were measured using the international 10–20 system of electrode placement (EEG: C3/A2, C4/A1, O1/A2; EOG: F7/A2 and F8/A1). A tight-fitting nasal continuous positive airway pressure (CPAP) mask (Respironics, Murrysville, PA, USA) was attached to the face of each subject and was connected to the ventilation circuit. Subjects were restricted to nasal breathing by placing tape over the mouth. Airflow was measured by a heated pneumotachometer (Model 3710, Hans Rudolph, Inc.) connected to the mask. Tidal volume was obtained by integrating the pneumotachograph flow signal (Model RSS100-HR, Hans Rudolph, Inc.). Inspiratory muscle activity was obtained by surface electromyogram electrodes placed 2–4 cm above the costal margin in the anterior axillary line. End-tidal carbon dioxide was measured with a gas analyser (Model CD-3 A, AEI Technologies, Pittsburgh, PA, USA). To confirm the central aetiology of apnoea and to ascertain upper airway resistance, supraglottic pressure was measured with a catheter (Model MPC-500, Millar Instruments, Houston, TX, USA). The catheter was positioned in the hypopharynx just below the base of the tongue. All physiological variables were analog to digitally converted at a sampling frequency of 200 Hz per channel and input into a microcomputer using a commercially available software package (Gamma Version 4.0, Astro-Medical, Inc., West Warwick, RI, USA).

Data analysis

Wakefulness Average values of minute ventilation, PET,O2 and PET,CO2 were determined from the last 5 min of a 15 min baseline period measured immediately prior to completion of the rebreathing trials. The data collected during the rebreathing experiments were analysed using a spreadsheet designed for this purpose. Prior to analysis, the three deep breaths that were required for gas equilibration, in addition to sighs or swallows that were detected by the software during the experiment were excluded from further analysis. Subsequently, breath-by-breath PET,CO2 was plotted against time and fitted with a least squares regression line. The equation for this line provided a predicted value of PET,CO2versus time, thereby minimizing interbreath variability associated with the measurement of this variable. Thereafter, ventilation was plotted against the predicted PET,CO2.

Subsequently, each of these plots was fitted with a model made up of the sum of two segments separated by one breakpoint. Model fitting was based on minimizing the sum of least squares for non-linear regressions using commercial software (Sigmaplot 7.0, SPSS). Figure 1 shows an example of the lines fitted to the responses of one subject who completed hyperoxic rebreathing experiments before and after leuprolide treatment. The first segment of the response was an exponential decline to a final value. The exponential decline was chosen to fit any waning of ventilatory ‘poststimulus potentiation’ (Fig. 1, open circles) that might have occurred after hyperventilation. However, poststimulus potentiation is often not observed (Fig. 1, filled squares) so that the time constant of the response may be less than 1 s.

Figure 1.

An example of the ventilatory response to iso-oxic hyperoxic CO2 rebreathing before and after treatment with leuprolide acetate in one subject
Note measurement of the threshold and sensitivity. Additionally, notice that the partial pressure of end-tidal CO2 (PET,CO2) that demarcates the threshold was decreased and the sensitivity of the ventilatory response to CO2 was increased after 2 months of treatment with leuprolide acetate (Lupron). See text for further details.

The second segment was characterized by a breakpoint (Fig. 1) followed by a linear increase in minute ventilation that occurred in conjunction with a rise in PET,CO2. The breakpoint was taken as a measure of the threshold of the ventilatory response to carbon dioxide. We previously referred to this point as the ventilatory recruitment threshold (Mateika & Ellythy, 2003; Mateika et al. 2004). The threshold measured during hyperoxia (140 Torr) was thought to originate from the central chemoreflex, while the threshold measured under hypoxic conditions was thought to derive from the sum of the central and peripheral chemoreflex. The slope of the line fitted to minute ventilation after the breakpoint was taken as a measure of the sensitivity to increases in PET,CO2 (Fig. 1). We previously referred to this measure as chemoreflex responsiveness (Mateika & Ellythy, 2003; Mateika et al. 2004). We assumed that the average slope recorded from the iso-oxic hyperoxic carbon dioxide rebreathing trials represented sensitivity of the central chemoreflex, while the slope recorded from the iso-oxic hypoxic carbon dioxide rebreathing trials represented the combined peripheral and central chemoreflex sensitivity.

For the data collected during the rebreathing experiments, a mean value for each measurement (i.e. threshold and sensitivity) was calculated for each subject from the trials completed during iso-oxic hyperoxic and hypoxic carbon dioxide rebreathing. Subsequently, mean values were calculated for a given treatment (before versus after leuprolide treatment) and condition (140 versus 50 Torr).

NREM sleep For each trial of mechanical ventilation during NREM sleep, PET,CO2 was measured immediately prior to (i.e. control period) and during the 3 min period of hyperventilation. The average of five breaths immediately preceding the onset of mechanical ventilation was determined from the control period. Moreover, we calculated the average of the last five mechanically ventilated breaths prior to the ventilator being turned back to an expiratory positive airway pressure of 2.0 cmH2O. The change in PET,CO2PET,CO2) was calculated as the difference between the average PET,CO2 recorded during the control period and the average PET,CO2 associated with the five breaths recorded during the period of mechanical ventilation. The breath associated with the lowest minute ventilation during recovery was recorded if a hypopnoea was induced after mechanical hyperventilation. The nadir breath occurred on the first recovery breath in the majority of trials and within the first three breaths in all trials. When a central apnoea, defined as a cessation of breathing for a minimum of 5.0 s, was induced after mechanical hyperventilation it was assigned a value of 0. Only those trials associated with stable sleep, characterized by the absence of arousal or ascent to a lighter sleep state, were analysed. The threshold was defined as the PET,CO2 at which apnoea occurred. ΔPmath formula threshold was defined as the difference in the PET,CO2 measured during the control and hyperventilation periods associated with the occurrence of the first apnoea.

Pressure-flow loops before and after leuprolide treatment were used to confirm the absence of inspiratory flow limitation and to determine upper airway resistance (Rowley et al. 2001) for the control breaths measured prior to mechanical ventilation.

Statistical analysis

Wakefulness Paired t tests were used to determine if differences in baseline measures of minute ventilation, carbon dioxide production and hormone levels existed before and after leuprolide treatment. A two-way analysis of variance with repeated measures was used to determine if the threshold or sensitivity was different before versus after treatment with leuprolide acetate. The levels of the main factors ‘treatment’ and ‘oxygen concentration’ were before versus after leuprolide treatment and 140 versus 50 Torr. A similar analysis was used to determine if the PET,CO2 that demarcated the threshold during the iso-oxic hypoxic carbon dioxide rebreathing trials during wakefulness was different than the threshold measured during NREM sleep before and after treatment with leuprolide acetate.

NREM sleep Paired student t tests were performed to compare: (1) PET,CO2 measured during NREM sleep, (2) PET,CO2 that demarcated the threshold and (3) ΔPmath formula threshold before and after leuprolide treatment. All statistical analyses was performed with Sigma Stat 3.0 (Jandel Scientific, San Rafael, CA, USA). All data are presented as means ± standard error. A value of P < 0.05 was considered significant.

Results

Wakefulness

The mean age, height and weight of the subjects were 40.1 ± 2.5 years, 1.8 ± 0.1 m and 87.7 ± 1.9 kg, respectively. Testosterone, total oestrogen, oestradiol, luteinizing hormone and follicle stimulating hormone levels were all suppressed after treatment with leuprolide (Table 1). Baseline measures of ventilation (P= 0.03) and carbon dioxide production (P= 0.01) increased after leuprolide treatment. In contrast, PET,CO2 was similar before and after leuprolide treatment (Table 1).

Table 1.  Baseline measures of respiratory parameters and hormone levels before and after treatment with leuprolide acetate
 Minute
ventilation
(l min−1; BTPS)

P ET,CO2
(Torr)

inline image
(ml min−1; STPD)

LH
(MIU ml−1)

FSH
(MIU ml−1)

Oestrogen
(pg ml−1)

Oestradiol
(pg ml−1)

Testosterone
(ng ml−1)
  1. *Significantly different from measures obtained before leuprolide treatment (P < 0.05). PET,CO2, partial pressure of end-tidal CO2; inline image, carbon dioxide production rate; LH, luteinizing hormone; FSH, follicle stimulating hormone; STPD, standard temperature and pressure, dry; BTPS, body temperature and pressure (saturated with water vapour).

Before leuprolide 13.9 ± 0.935.7 ± 1.0612.2 ± 96.76.0 ± 1.49.2 ± 3.260.9 ± 7.830.7 ± 5.7391.3 ± 68.7
After leuprolide 16.6 ± 1.9*35.2 ± 1.8  697.8 ± 129.0* 1.0 ± 0.2* 3.3 ± 0.4* 37.7 ± 2.4* 12.3 ± 1.6*  77.7 ± 28.9*

Figure 1 shows a typical example of alterations in the threshold and sensitivity of the ventilatory response to carbon dioxide in one subject during iso-oxic hyperoxic carbon dioxide rebreathing trials, before and after treatment with leuprolide. Note that the PET,CO2 that demarcated the threshold was decreased, and sensitivity to carbon dioxide was increased after treatment with leuprolide. This finding was similar to the average findings which showed that the threshold measured during both the iso-oxic hyperoxic (before treatment: 46.32 ± 0.56 versus after treatment: 44.78 ± 0.83 Torr; P= 0.01) and hypoxic rebreathing (before treatment: 41.05 ± 0.77 versus after treatment: 39.40 ± 0.83 Torr; P= 0.01) trials were significantly less after leuprolide treatment (P= 0.01 in both cases) (Fig. 2, top). The slope of the ventilatory response during the iso-oxic hyperoxic carbon dioxide rebreathing trials was significantly greater (P= 0.02) after leuprolide treatment (Fig. 2, bottom). In contrast, the slope measured during the iso-oxic hypoxic carbon dioxide rebreathing trials was similar before and after leuprolide treatment (Fig. 2, bottom).

Figure 2.

Bar graphs showing the average threshold (upper panel) and sensitivity (lower panel) of the ventilatory response to CO2 obtained from the iso-oxic hyperoxic and hypoxic CO2 rebreathing trials before and after treatment with leuprolide acetate
Note that the PET,CO2 that demarcated the threshold was decreased during both the hyperoxic and hypoxic rebreathing studies after leuprolide acetate. Moreover, note that the sensitivity of the ventilatory response during iso-oxic hyperoxic CO2 rebreathing was increased after treatment. *Significantly different from hypoxic rebreathing trials, P < 0.05; **Significantly different from before leuprolide acetate treatment, P < 0.05.

NREM sleep

A representative polygraph record of a trial completed during one subject's baseline study is shown in the upper panel of Fig. 3. Mechanical ventilation was initiated during expiration in stable NREM sleep (Fig. 3B). The increased tidal volume resulted in mild hypocapnia (ΔPET,CO2= 3.7 Torr from control), followed by a central apnoea after the termination of mechanical ventilation (Fig. 3C; note the lack of inspiratory effort in the PSG channel). The lower panel of Fig. 3 shows a representative polygraph from the same subject after leuprolide administration. In this polygraph, the increased tidal volume resulted in a larger degree of hypocapnia (PET,CO2= 4.1 Torr from control). In contrast to the baseline study, the termination of mechanical ventilation was followed by a hypopnea, indicating a decreased susceptibility to the disfacilitatory effects of hypocapnia.

Figure 3.

Polygraph recorded during mechanical ventilation trials completed before and after treatment with leuprolide acetate
A, eupnoeic breathing; B, the last several breaths during a period of mechanical hyperventilation; C, the immediate posthyperventilation breaths. The upper panel illustrates a trial prior to the administration of leuprolide. The mechanical ventilation decreased the PET,CO2 by 3.73 Torr, which resulted in a central apnoea after the termination of mechanical ventilation. The lower panel illustrates a trial during administration of leuprolide in the same subject. The mechanical ventilation decreased the PET,CO2 by 4.13 Torr, resulting in a hypopnoea. PP, supraglottic pressure; PP, mask pressure.

The PET,CO2 that demarcated the threshold during NREM sleep was not significantly different from the threshold measured during the iso-oxic hypoxic carbon dioxide rebreathing trials completed during wakefulness, both before and after leuprolide treatment. Additionally, the PET,CO2 that demarcated the threshold during NREM sleep significantly decreased after leuprolide treatment (42.1 ± 0.6 versus 39.6 ± 0.6 Torr, P= 0.002) (Fig. 4, top). To determine if changes in the threshold were secondary to differences in baseline PET,CO2, we measured NREM PET,CO2 before and after treatment with leuprolide. The PET,CO2 was not significantly different after as compared to before leuprolide treatment during NREM sleep (44.7 ± 0.8 versus 43.3 ± 0.6 Torr). Moreover, the change in PET,CO2 that was required to induce an apnoea was significantly increased (2.6 ± 0.3 versus 4.1 ± 0.3 Torr, P= 0.004) after leuprolide treatment (Fig. 4, bottom).

Figure 4.

Bar graphs showing the partial pressure of PET,CO2 that demarcated the threshold measured before and after treatment with leuprolide acetate
Additionally, the lower panel shows the average change in the partial pressure of PET,CO2, relative to baseline measures, that was required to induce an apnoea before (black bar) and after treatment (white bar) with leuprolide acetate. Note that the change in PET,CO2 required to induce an apnoea was greater after as compared to before treatment. *Significantly different from before leuprolide acetate treatment, P < 0.05.

Upper-airway resistance was measured from the linear portion of pressure–flow curves to determine if differences existed before as compared to after leuprolide treatment. No differences were found (5.0 ± 2.5 versus 6.1 ± 1.6 cmH2O l−1 s−1, P= 0.4, respectively).

Discussion

Our major findings were that the threshold of the ventilatory response to carbon dioxide was decreased during iso-oxic hyperoxic and hypoxic carbon dioxide rebreathing after treatment with leuprolide acetate. Moreover, in a manner similar to that observed for the threshold during wakefulness, we showed that the threshold during NREM sleep was reduced after leuprolide treatment. Lastly, we showed that the ventilatory response to iso-oxic hyperoxic but not iso-oxic hypoxic carbon dioxide rebreathing was increased after leuprolide treatment.

Critique of the methods

We chose not to randomize the order of the on-leuprolide studies because we postulated that alterations in chemoreflex control which might occur as a consequence of treatment may not be immediately abolished after the 2 month treatment period. Thus, our results may have been the consequence of an order effect. However, replication of our results during sleep argues against this possibility.

In the present investigation we chose to complete our studies after subjects received two injections of leuprolide acetate over a period of 8 weeks. This duration of time was selected because an initial increase in sex hormone concentration is often observed 1–2 weeks after the initial injection of leuprolide acetate. Beyond that time the continuous presence of leuprolide desensitizes gonadotropin releasing hormone receptors, leading to a suppression of the secretion of gonadotropins, testosterone and oestradiol, to levels approaching those seen in castrated men. Given that we were interested in examining changes in respiratory control after a prolonged suppression of testosterone, we elected to obtain our measures after 8 weeks.

Our primary aim was to examine the impact of testosterone on respiratory control during wakefulness and sleep. However, we cannot discount the possibility that testosterone mediates its impact on respiratory control via conversion to oestradiol, since the masculinizing influences of testosterone are mediated via a similar conversion (Behan et al. 2002). Moreover, we cannot discount the possibility that oestradiol and/or total oestrogen impacted on our findings independently, since this hormone was suppressed concomitantly with testosterone. Nonetheless, to date, the independent impact of oestrogen (i.e. independent of progesterone) as a modulator of respiratory control has not been well defined in humans (Behan et al. 2002). Likewise, our findings are in general agreement with the majority of the published work which suggests that testosterone impacts on respiratory control. Thus, we have restricted our discussion to the role that testosterone has in respiratory control.

The modified rebreathing technique has been discussed extensively in a number of prior publications (Mohan & Duffin, 1997; Mohan et al. 1999; Duffin et al. 2000; Mahamed & Duffin, 2001; Mahamed et al. 2001, 2003; Mateika et al. 2004). We refer the reader to our published work which comprehensively critiques the technique (Mateika et al. 2004). It is unlikely that the differences we observed after leuprolide treatment were a consequence of other factors unrelated to leuprolide treatment. Conditions conducive to maintaining quiet wakefulness were strictly adhered to during each rebreathing trial, both before and after leuprolide treatment. Thus, environmental factors that are known to influence the state of arousal and consequently ventilation were maintained throughout the experiment (Shea, 1996). It is also possible that the time of day could influence the ventilatory response during rebreathing (Stephenson et al. 2000). Thus, the rebreathing experiments were completed at the same time of day to eliminate this possibility.

It is possible that the effect of leuprolide on the threshold during sleep may have been due to changes in upper airway resistance, sleep state or baseline PET,CO2. However, upper airway resistance and sleep state conditions were not altered after leuprolide treatment. Moreover, baseline PET,CO2 was not altered after leuprolide treatment but the change in PET,CO2 required to induce an apnoea was greater after treatment. Thus, the most likely explanation for the decrease in the threshold is that administration of leuprolide impacted directly on chemoreflex function.

We defined an apnoea as an absence of flow for 5 s, instead of the standard 10 s used in most clinical sleep laboratories. We selected this criterion because it represents a doubling of expiratory time, a commonly used metric of apnoea in physiologic studies (Badr et al. 1992).

Central and peripheral chemoreflex threshold before and after treatment with leuprolide acetate

Given our assumptions (see Data analysis, Wakefulness, third paragraph) our findings suggest that the central chemoreflex threshold was reduced after treatment with leuprolide acetate. Our findings also imply that the peripheral chemoreflex threshold remained unchanged after treatment, since a similar decrease in the threshold occurred during both the iso-oxic hyperoxic and hypoxic carbon dioxide rebreathing trials after treatment with leuprolide acetate. The decrease in the threshold that we observed during wakefulness was replicated during NREM sleep. The suppression of testosterone after treatment with leuprolide acetate may have been responsible for the observed decrease in the threshold. This possibility is supported by our results, which showed that the threshold during NREM sleep was reduced in females compared to males (Zhou et al. 2000) and was increased in females after the administration of testosterone (Zhou et al. 2003).

Central and peripheral chemoreflex sensitivity before and after treatment with leuprolide acetate

The role of testosterone in increasing central chemoreflex sensitivity has been implicated in many studies which reported that the ventilatory response to carbon dioxide in the presence of high levels of oxygen was increased in males compared to females (Patrick & Howard, 1972; Saunders et al. 1972; White et al. 1983; Aitken et al. 1986; Sebert et al. 1990; van Klaveren & Demedts, 1998). However, testosterone administration in some studies failed to elicit increases in central chemoreflex sensitivity. Testosterone replacement in hypogonadal males had little impact on the ventilatory response to carbon dioxide in the presence of high oxygen levels (White et al. 1985). Additionally, the hypercapnic ventilatory drive was depressed after the administration of testosterone in sleeping primates (Emery et al. 1994). Moreover, androgen blockade had no impact on the hypercapnic ventilatory response (Stewart et al. 1992).

Likewise, based on our assumptions (see Data analysis, Wakefulness, third paragraph) our results do not support the hypothesis that elevated levels of testosterone are primarily responsible for increases in chemoreflex sensitivity, since suppression of testosterone was accompanied by an increase in central chemoreflex sensitivity. It is possible that increases in metabolic rate may have contributed to the increase in the ventilatory response to iso-oxic hyperoxic carbon dioxide rebreathing that we observed, given our results which showed that baseline minute ventilation and carbon dioxide production was increased while PET,CO2 remained constant after treatment with leuprolide acetate.

Thus, it is possible that facilitation of the ventilatory response via alterations in metabolic rate were more dominant than disfacilitation of the response that accompanied testosterone suppression. The manner in which treatment with leuprolide acetate leads to increases in metabolic rate has not been investigated previously. It is possible that administration of leuprolide acetate leads to alterations in other hormones (i.e. thyroid stimulating hormone) that impact on metabolic rate. This possibility is supported by case findings that have reported that transient thyrotoxicosis occurs following the administration of gonadotropin releasing hormone agonists (Kasayama et al. 2000; Amino et al. 2003). However, this speculation remains to be investigated.

The role of testosterone in increasing peripheral chemoreflex sensitivity has also been implicated in some (Patrick & Howard, 1972; White et al. 1983; van Klaveren & Demedts, 1998), but not all (Aitken et al. 1986), studies which reported that the ventilatory response to isocapnic hypoxia is greater in males compared to females. Likewise, the ventilatory response to isocapnic hypoxia increased after testosterone administration in hypogonadal men (White et al. 1985) and neutered male cats (Tatsumi et al. 1994). Our findings support these results if one accepts the assumption that central and peripheral chemoreflex inputs combine in an additive fashion to influence ventilation (Lloyd, 1966; Grodins et al. 1967; Khoo et al. 1991; Duffin et al. 2000). We showed that the ventilatory response to iso-oxic hypoxic carbon dioxide rebreathing (which based on our assumptions represented the combined central plus peripheral chemoreflex response) remained unchanged after suppression of testosterone. Given that central chemoreflex sensitivity increased after suppression of testosterone, our findings suggest that peripheral chemoreflex sensitivity was reduced after testosterone suppression, thereby supporting previous studies which suggested that testosterone enhances peripheral chemoreflex sensitivity.

Physiological significance

Alterations in the threshold of the ventilatory response to carbon dioxide were observed during both wakefulness and sleep after suppression of testosterone. Thus, the addition of behavioural effects during wakefulness does not appear to obscure some physiological alterations that may manifest during sleep.

The prevalence of obstructive sleep apnoea has been reported to be greater in males compared to females (Young et al. 1993; Bixler et al. 2001) and it is possible that this difference in occurrence is due in part to the impact of testosterone on the threshold and/or sensitivity of the ventilatory response to carbon dioxide (Zhou et al. 2000, 2003). Nonetheless, it remains to be determined whether these alterations exacerbate sleep apnoea. One possibility is that an unusual increase in the ventilatory response to an elevated level of carbon dioxide, induced possibly by an apnoea, could subsequently drive carbon dioxide levels to a point that would lead to a reduction or abolition of ventilatory motor output. This possibility would be enhanced if the PET,CO2 that demarcated the threshold was positioned closer to normal resting values. The decreases in ventilatory motor output could subsequently lead to partial or complete closure of the upper airway as reported previously (Onal et al. 1986; Badr & Kawak, 1996; Badr et al. 1997). This scenario is possible given that ventilatory instability may be induced within the respiratory system in response to increases in chemoreceptor sensitivity (Khoo, 2000).

Appendix

Acknowledgements

This work was supported by the Department of Veterans Affairs, the National Heart, Lung and Blood Institute, and the American Heart Association.

Ancillary