Medroxyprogesterone in postmenopausal females with partial upper airway obstruction during sleep

The menopause has a major impact on sleep and breathing. Although the increase in breathing abnormalities after the menopause 1 is attributed to decreased progesterone secretion, the role of progesterone replacement therapy as a respiratory stimulant has not been established.

In addition to obesity and structural abnormalities in the upper airway, a decrease in the drive to breathe is likely to be involved in sleep-disordered breathing in postmenopausal females. Episodes of hypopnoea are relatively more common than apnoea in females 2. Furthermore, in females, the apnoea/hypopnoea index (AHI) may underestimate the severity of upper airway dysfunction, since females can be symptomatic with an AHI <5 3.

Conventional postmenopausal hormone replacement therapy aims to control climacteric vasomotor symptoms. Since the postmenopausal state predisposes to hypoventilation, some females might benefit from progesterone replacement therapy. The efficacy of short-term treatment with medroxyprogesterone acetate (MPA) on sleep and breathing in postmenopausal women with nocturnal hypoventilation related to partial upper airway obstruction during sleep was therefore evaluated. In a subgroup, the effects of MPA were compared to those of nasal continuous positive airway pressure (nCPAP), which is known to relieve partial upper airway obstruction without any immediate effect on respiratory drive.

Subjects

Initially, 71 healthy hysterectomized postmenopausal females were recruited for polygraphical sleep studies by announcements in newspapers 4. The static-charge sensitive bed (SCSB) recordings showed episodes of increased respiratory resistance (IRR) 5, with prolonged hypoxaemia in 11 subjects, of whom 10 agreed to participate in the present study. Their median age was 60 (range 54–68) yrs, median body mass index 31.3 (range 23.5–37.6) kg·m⁻², and the mean Epworth sleepiness score was 6.4±5.0. Five of 10 subjects reported snoring at least during three nights of the week. Their postmenopausal status was verified by measurements of serum follicle-stimulating hormone (FSH >30 IU·L⁻¹). None of the subjects were current smokers or suffered from any major disease.

Written informed consent was obtained from all participants. The protocol was approved by The Joint Commission on Ethics of Turku University and Turku University Central Hospital.

Methods

Results

Four out of eight subjects had a daytime oxygen tension in arterial blood (P_a,O2) <10 kPa. The carbon dioxide tension in arterial blood (P_a,CO2) was within reference range in all subjects. At baseline, mean arterial pH was 7.41±0.01 (mean±sd), mean P_a,CO2 5.4±0.30 kPa, mean P_a,O2 10.8±1.7 kPa.

At baseline, no episodes of OP-3 of OP-2 pattern were observed but partial upper airway obstruction was frequent (average IRR 23.4±1.5% and average OP-1 5.5±5.9% of time in bed). The Autoset® nasal pressure profile analysis showed significant upper airway flow limitation in all subjects. The median AHI determined by the Autoset® was 10.6 (interquartile range, (IQR) 6.2) and respiratory irregularity index 31.0 (IQR 17.4). The mean arterial oxygen saturation (S_a,O2) was 93.0±1.2%, maximum S_a,O2 98.7±1.4%, nadir S_a,O2 80.0±6.7% and the mean frequency of oxygen desaturation ⩾4% (ODI₄) per hour was 2.2±1.3. The percentage of sleep Stage 1 was 7.9±3.5, Stage 2 56.4±6.0, slow-wave sleep 10.4±5.0 and rapid eye movement (REM) 18.4±8.4. The median EEG-arousal index (arousals·h⁻¹) was 5.6 (IQR 1.4) and the median sleep efficiency 95.0 (IQR 4.6)%.

Discussion

The present study was based on a previous observation by the authors that 17% of healthy postmenopausal females spent a significant proportion of their sleep time with partial upper airway obstruction 10. Since partial obstruction predisposes to CO₂ retention 11, the effects of respiratory stimulants such as MPA are of interest. The present results show that MPA used at a dose of 60 mg daily for 14 days improves ventilation in the affected females. The improvement was maintained beyond 3 weeks post-treatment. nCPAP was more efficient than MPA in decreasing respiratory efforts but the P_et,CO2 decreased more with MPA.

The observations in postmenopausal females with partial upper airway obstruction should be extrapolated with caution to females with obstructive sleep apnoea. The current knowledge on the efficacy of progestins to control obstructive sleep apnoea is based on studies in males. In postmenopausal females with sleep-disordered breathing, only three studies have used MPA in combination with oestrogen 12–14 and one as a single therapy 15. Their common conclusion was that there was some decrease in AHI or shortening of apnoea duration. None of the studies measured partial upper airway obstruction or CO₂ and none compared the response with that of nCPAP.

As with previous observations by the authors, in postmenopausal females with chronic respiratory failure 16, the ventilatory improvement in the present study was sustained for at least 3 weeks after cessation of MPA. This contradicts an earlier observation that the ventilatory effects subside within 14 days of cessation of MPA 17. However, there is also evidence to support a prolonged ventilatory effect. In a study of the obesity hypoventilation syndrome, a normal P_a,CO2 was maintained for up to 4 months after cessation of intramuscular progesterone therapy 18.

There are at least three possible reasons for the sustained ventilatory improvement, which relate to obesity, female gender or abnormal breathing. In obese subjects, abundant adipose tissue may result in increased metabolism of endogenous steroids to oestradiol 19, which is needed for upregulation of progesterone receptors. Combining oestrogen with progesterone prolongs the ventilatory effects in healthy males. Obesity does not, however, explain the sustained ventilatory improvement observed earlier in nonobese females with chronic respiratory insufficiency 16. Therefore, other factors must also be involved.

Another possibility is related to the female sex. Although the oestrogen levels of postmenopausal females are comparable to those of males, the response to female hormone therapy could be more prolonged. The number of progesterone receptors might be greater or their distribution different in females. The progesterone receptors in females might also be more readily upregulated during treatment.

A third factor contributing to the prolonged respiratory stimulation seen in the present subjects with mild sleep-disordered breathing but not in healthy study subjects 17 might be the underlying ventilatory abnormality itself. The natural homeostatic mechanisms aim at correcting altered functions. In the case of sleep-disordered breathing, the altered function is hypoventilation; whereas, in healthy controls taking MPA, the altered function is the drug effect of excessive respiratory stimulation. In the former case, the homeostatic mechanisms act parrallel to the MPA effect when trying to correct hypoventilation. In the later case, the homeostatic efforts to correct the excessive ventilation oppose the action of MPA. The synergic action of two correcting mechanisms could explain the enhanced treatment effect of the MPA patients in the present study.

The sustained ventilatory effect of short-term MPA suggests that periodic administration might be sufficient to improve ventilation in females. In terms of sleep efficiency, periodic administration might even be superior to continuous therapy.

MPA did not improve the AHI. There was no significant difference between CPAP and MPA nights but this may be due to the small number of subjects. In previous studies, MPA has shown no consistent effect on AHI. No change 20, decreasing trend 21 or slight decrease of AHI 13, 14 has been reported with MPA alone or in combination with oestrogen. By decreasing P_et,CO2 MPA may also induce respiratory instability resulting in an increase in the number of apnoeic and hypopnoeic episodes if P_a,CO2 decreases below the apnoeic threshold. However, this was not supported by the present data, where the frequency of central apnoea did not increase with MPA. It is likely that the MPA-induced net effect on the number of apnoeic and hypopnoeic events depends on the subtle balance between sufficient or overdriven ventilation.

The presence of periodic breathing (OP-1) is in line with the observation that these females showed mainly partial upper airway obstruction. MPA had no effect on the frequency of periodic breathing but the high variance and small number of subjects weaken the power of this interpretation. CPAP decreased the OP-1 and IRR patterns. An important observation was that the IRR pattern, indicating increased respiratory efforts, remained unchanged on MPA but decreased on CPAP. This observation is in agreement with the hypothesis that the ventilatory drive remains similar with MPA, but it is achieved at lower levels of CO₂ 22. This may also explain why MPA did not increase the sensation of dyspnoea.

Whereas the AHI seems a surrogate variable to describe the severity of sleep-disordered breathing in postmenopausal females, the IRR phenomenon seems to be characteristic of sleep-disordered breathing in postmenopausal females. This pattern indicates increased respiratory efforts, corresponding to intrathoracic pressures <−20 cmH₂O 23. The pattern consists of high frequency respiratory components that appear when expiration becomes active. There is a high correlation between the IRR spike amplitude and the P_et,CO2 level 11.

MPA concentrations were unmeasurable or near the detection limit within 2 17 or 3 weeks (present study) of cessation of treatment. In postmenopausal females, FSH and luteinizing hormone remain decreased at least 3 weeks after MPA therapy 24. It is possible that the MPA-induced ventilatory effects outlast measurable serum concentrations.

Neither objective nor subjective sleep quality altered with MPA. Previously, MPA has been reported to have either no effect on sleep 15, to decrease REM sleep 25, or to decrease 13 or increase arousals 26. In the present subjects, sleep efficiency was surprisingly high at baseline and on MPA but improved still further during washout. This delayed MPA effect is in line with an earlier report of decrease in Stages 1 and 2 sleep and increase in REM sleep after MPA 20 and supports the view that short-term MPA therapy may have favourable rebound effects on sleep. The present findings of extremely high sleep efficiency and low arousal index are not, however, the first observations of this kind. In a previous study, by the authors, on healthy postmenopausal females, the median EEG arousal index was 3.2 (range 0.2–12.4, n=32) and the median sleep efficiency was 93.4% (range 74.4–98.6, n=63) 27. These findings were not scorer-related, since they were analysed by different scorers. Sleep efficiency decreases with age, although less so in females than in males. Whether there is an interaction between high CO₂ and low arousal index in these particular subjects, remains to be elucidated.

All of the postmenopausal females in the present study considered themselves healthy and were not seeking treatment. Although their mean S_a,O2 and P_et,CO2 during sleep were within reference ranges, they had frequent episodes of increasing P_et,CO2 and hypoxaemia with mean S_a,O2 nadir as low as 80%. This contrasts with their exceptionally good sleep quality, high sleep efficiency and low arousal frequency. Although the current study was experimental, the results prompt the question whether or not the females benefited from the intervention. Although, on direct questioning, some symptoms suggested upper airway obstruction (e.g. dreams of cats sitting on ones throat), excessive sleepiness was not reported. Therefore, any possible therapeutic benefit in such patients would certainly not be related to improving sleep quality and daytime performance.

Recent literature has produced controversial results about the possible cardiovascular effects of sleep-disordered breathing on one hand and female sex hormone replacement on the other. Not only sleep-disordered breathing but also atherosclerosis, rapidly increase after the menopause. In females, snoring increases the risk of stroke independent of other factors 28. Partial upper airway obstruction with nocturnal CO₂ retention offers a plausible cause of intracellular acidosis that might result in endothelial dysfunction and reduced nitric oxide (NO) production. Increasing evidence suggests that high CO₂ levels are involved in the pathogenesis of vascular disease. In postmenopausal females, resting P_et,CO2 is an independent determinant of systolic blood pressure and carotid artery intima-media thickness 29. Decreasing high or borderline CO₂ during sleep might therefore improve endothelial function and prevent cardiovascular sequelae. Increased endothelial NO production could be mediated by MPA-induced insulin-like growth factor 1 production 30, 31.

The development of a stroke in one of the study subjects is contradictory to the thinking that MPA might have immediate beneficial endothelial effects. The incidence of thromboembolic complications encountered during high-dose MPA treatment in breast cancer with or without chemotherapy is 2.1% 32. However, neoplastic diseases 33 may also enhance coagulability. MPA may change blood-clotting factors to favour thrombosis 32, but cerebral infarction during MPA treatment is an extreme rarity. A Medline database search found one case report of medullary infarction during MPA treatment 34. The role of MPA therapy in the development of cerebral infarction in the patient from this study, therefore, remains unclear. The patient was neither hypertensive nor a smoker, ethanol consumption was moderate and serum lipids and blood-clotting factors were within reference ranges. Oestrogen replacement therapy had previously been used without complications. However, the patient was a heavy snorer, which increases the risk of stroke 28.

To conclude, medroxyprogesterone acetate used at a dose of 60 mg daily for 14 days improves nocturnal ventilation in postmenopausal females with partial upper airway obstruction during sleep. This effect is maintained for at least 3 weeks. Medroxyprogesterone acetate is at least equally effective in reducing P_et,CO2 as nasal continuous positive airway pressure but it may be less effective in reducing episodes of apnoea or hypopnoea. Long-term intervention studies are needed to evaluate whether postmenopausal females would have any respiratory or cardiovascular benefit from lower nocturnal carbon dioxide levels, achievable with appropriate hormone replacement therapy.