INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS VENTILATORY STIMULANTS PSYCHOTROPIC AGENTS OTHER PSYCHOTROPIC AGENTS OTHER AGENTS SUMMARY FUTURE DIRECTIONS REFERENCES

Treatment of sleep-disordered breathing (SDB) today usually fits into one of five categories: weight loss, pharyngeal surgery, nasal continuous positive airway pressure (nCPAP), mandibular retention devices, or pharmacologic agents. From a convenience perspective, the use of medications is attractive. Presumably, treatment compliance would improve if one simply had to take a pill or two a day versus tolerating the discomfort and potential complications of surgery, wearing an external device over one's face, or keeping a foreign body within one's mouth throughout the night. Several different categories of pharmacologic agents have been used in the treatment of sleep-disordered breathing in studies, both poorly and well-controlled, with varying degrees of success. Agents studied include ventilatory drive stimulants, central nervous system stimulants, tricyclic antidepressants, serotonin reuptake inhibitors or antagonists, antihypertensive agents and even sedative-hypnotic agents. Specific treatment of endocrine disorders such as hypothyroidism has yielded interesting results.

The degree of certainty for recommending a given pharmacologic agent for the use in treatment of sleep-disordered breathing will be based on grading of the literature. New concepts that may be important in the future of pharmacologic treatment of sleep-disordered breathing will be presented.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS VENTILATORY STIMULANTS PSYCHOTROPIC AGENTS OTHER PSYCHOTROPIC AGENTS OTHER AGENTS SUMMARY FUTURE DIRECTIONS REFERENCES

Peer-reviewed, English-language manuscripts examining the pharmacologic treatment of obstructive sleep apnea (OSA), periodic breathing in sleep, and sleep-related hypoventilation, all considered sleep-disordered breathing (SDB), in humans were identified by computer searches of the National Library of Medicine databases and by examination of bibliographies of review articles. Case reports were cited only if they were deemed to be important to the field. Evidence tables were constructed in order to itemize methodology and results of specific studies (1).

Manuscripts were assigned a letter grade according to the methodologic design and the strength of the scientific evidence. The following grades were used: A, blinded, controlled trials; B, observational studies; C, case reports or expert opinion. Also, a numerical score was given for the literature of each drug category reviewed, so that the quality of the literature could be compared across the various groups of agents studied. An A study was given 10 points: a B study was given 5 points, and a C study was given 1 point. In order to weigh this grade for the number of subjects included, one tenth (0.1) of a point was added to the score of each study for each subject within a given study. To control for the number of studies within a category, the total score was divided by the number of studies within that category to derive a composite score. A score was not calculated if only three studies or fewer were reviewed for a particular drug category. The intent of the score was to indicate the level of scientific sophistication used in evaluating the usefulness of a given pharmacologic agent, or group of agents. The score does not reflect the clinical benefit of a given pharmacologic agent or group of agents.

	VENTILATORY STIMULANTS

TOP ABSTRACT INTRODUCTION METHODS VENTILATORY STIMULANTS PSYCHOTROPIC AGENTS OTHER PSYCHOTROPIC AGENTS OTHER AGENTS SUMMARY FUTURE DIRECTIONS REFERENCES

Medroxyprogesterone

Group score = 7.9 (Table 1). Medroxyprogesterone (MPA) has been known to be a ventilatory stimulant for some time (2, 3). It was reasoned that if MPA increased central responsiveness to CO₂ or hypoxia then the apnea cycle would be interrupted because of inhibition of the hypoventilatory portion of the apneic cycle. In addition, it was proposed that with increased drive to the pharyngeal muscles, pharyngeal patency would be improved. Six clinical trials with progestational agents have been conducted (4). As shown in Table 1, two placebo-controlled, blinded studies have been conducted (Grade A) with two doses of MPA (30 and 150 mg/d). These studies did not show an overall improvement in OSA. Strohl and colleagues (4) showed improvement in four of nine patients with OSA in an uncontrolled study (Grade B). Three of the four responders were hypercapnic. Kimura and colleagues (8) found a small, but significant, improvement in OSA with the progestational agent, chlormadinone (Grade B). In this study there was a correlation between the improvement in hypercapnic ventilatory drive and the improvement in OSA with therapy, but the changes seen were small and no assessment of clinical variables was conducted. Because the study was uncontrolled, it cannot be determined if the increase in ventilatory drive was the cause of the improvement in the OSA. Another study assessed the effect of MPA on upper airway muscles. Collop (10) found that MPA did not prevent the worsening of OSA produced by ethanol-induced pharyngeal muscle hypotonia.

In summary, four Grade B and two Grade A studies of MPA in sleep-disordered breathing have been conducted. Only 31 patients have been studied for no longer than 1 mo in a blinded, controlled fashion. Female subjects were well represented in the sample. With the limited data available, MPA has not been shown to be very useful in OSA, except in some hypercapnic patients. The hypercapnia in these patients may indicate the presence of a low hypercapnic ventilatory drive, which if stimulated by progesterone, may lead to improvement. Neither high nor low doses were effective (5, 9). Available literature indicated that progestational agents are not likely to be of value in the treatment of normocapnic patients with OSA. Individual hypercapnic patients with OSA may respond. Surely, a response should be documented objectively. Future studies should distinguish between normocapnic and hypercapnic patients and study adequate numbers of subjects for at least 6 mo with a high and low dose of progesterone, at doses that do and do not stimulate hypercapnic ventilatory drive, respectively.

Thyroxine

Group score = 3.6 (Table 2). Hypothyroidism has been associated with sleep-disordered breathing since the report of Massumi and Winnacker in 1964 (11) (Grade C). Although these investigators did not conduct polysomnograms, they documented the presence of obstructive apneas in two myxedematous patients. More formal polysomnograms were performed by Orr and colleagues (12) in three hypothyroid patients (Grade C). Repetitive obstructive apneas were seen during sleep. Rajagopal and colleagues (13) established this relationship more firmly (Grade B). Nine of 11 consecutive hypothyroid patients had OSA. Three nonobese hypothyroid patients had OSA. These data led to the conclusion that hypothyroidism and OSA were tightly linked in spite of the small size of the groups studied above. However, the data of Lin and colleagues (14) led to a different conclusion. These investigators found only a 3% prevalence of hypothyroidism in 65 patients in whom OSA was consecutively diagnosed, and a 25% prevalence of OSA in a group of 20 hypothyroid patients (Grade A). Those with OSA in the hypothyroid population were older and more obese than those hypothyroid patients without OSA.

Rajagopal and colleagues (13) found improvement in OSA occurred with l-thyroxine treatment (Grade B). In contrast Grunstein and Sullivan (15) (Grade B) did not find such a favorable response, in that six of eight obese hypothyroid patients with OSA failed to respond to thyroid replacement.

In summary, only Grades B and C studies, with small numbers of subjects, have raised the possibility of a pathophysiologic relationship between hypothyroidism and OSA. However, the only available prevalence study did not confirm the common existence of this association, as had been previously suspected. Age and obesity were identified as variables that increased the prevalence of OSA in hypothyroidism. Grade B observational studies demonstrated a response of OSA to thyroid replacement therapy, except in some obese patients. No distinction was made between myxedema and chemical hypothyroidism without myxedema in the manifestations of the disease and its treatment. No placebo-controlled groups were used in any of these investigations, and inadequate controls exist for sex, age, and body size. Thereby, conclusions made about the effectiveness of thyroid replacement in hypothyroid patients with OSA have to be made realizing the defects in the investigations published to date. Future studies should attempt to improve on these design flaws. Groups of both myxedematous and nonmyxedematous hypothyroid patients need to be studied with appropriate controls.

Acetazolamide

Group score = 7.5 (Table 3). Because the metabolic acidosis produced by acetazolamide (ACET) stimulates ventilation, it has been hypothesized that ACET would improve sleep-disordered breathing. Groups of patients with central, obstructive, and mixed sleep apnea have been studied. Four of six studies to be reviewed are uncontrolled and unblinded.

First, we will critique the studies where ventilatory control was assessed. Neither White and colleagues (16) (Grade B) nor DeBacker and colleagues (17) (Grade B) found a change in the gain of the hypoxic or hypercapnic ventilatory responses induced by ACET. In both studies the hypercapnic ventilatory response was moved to a lower CO₂ level. Both groups of investigators showed a significant improvement in patients with central sleep apnea during treatment with ACET. Sleep stage distribution was not altered by ACET. At high altitude Hackett and colleagues (18) demonstrated a decrease in periodic breathing and central apneas with three doses of 250 mg of ACET over 24 h compared with a controlled placebo period (Grade A). Interestingly, in this study, ACET did not increase hypoxic ventilatory drive (hypercapnic ventilatory drive was not tested). Almitrine, a hypoxic ventilatory drive stimulant, worsened the periodic breathing at altitude. In fact, the degree of periodic breathing was positively associated with the level of hypoxic drive, suggesting that a heightened hypoxic ventilatory drive led to unstable breathing during sleep at altitude.

Three studies have evaluated the effect of acetazolamide in OSA. In four such patients Sharp and colleagues (19) found no improvement (Grade B). Tojima and colleagues (20) (Grade B) found that ACET increased hypercapnic but not hypoxic drive. They showed a modest improvement in OSA and oxygenation in both NREM and REM sleep with symptomatic improvement. Because this study was not placebo-controlled or blinded, it cannot be certain that the improvement in symptoms was really due to ACET. In a blinded, placebo-controlled study, Whyte and colleagues (21) (Grade A) demonstrated that there was a physiologic, but not a clinical, improvement in 10 patients with OSA. Sharp and colleagues (19) (Grade B) found that the obstructive component of two of four of those with mixed apnea worsened with ACET administration. White and colleagues (16) and Tojima and colleagues (20) both reported patients who worsened while receiving ACET.

In summary, two of six studies of ACET were adequately controlled. One Grade A study with only four subjects, healthy climbers at altitude, and two Grade B studies showed that ACET may be helpful in those with central apnea. One Grade A study and two Grade B studies showed that ACET did not lead to much improvement in OSA. Some subjects may worsen while receiving ACET. Because the mild metabolic acidosis produced by ACET was shown not to stimulate ventilatory drive in all studies, the mechanism of action of ACET in central apnea is uncertain. Possibly, ACET lowers the apneic threshold for CO₂, as might be expected with a shift of the hypercapnic ventilatory response slope to a lower CO₂ level, so that ventilation is stimulated with minimal hypercapnia, thereby preventing the central apnea from occurring. Stimulation of hypoxic drive, as found with almitrine, may actually worsen obstructive apneas, possibly because of the generation of more chest wall muscle activity, higher subatmospheric pleural pressures, and, therefore, generation of a greater collapsing pressure within the upper airway. In total only 19 patients with OSA have been studied, 10 in a controlled fashion, for no longer than 2 wk. Six of these 19 patients were female. Future studies should examine the effect of ACET on both central apnea and OSA over an adequate duration in patients whose ventilatory drive status is well characterized before and after treatment.

Theophylline

Group score = 9.9. The effect of theophylline on central apnea, periodic breathing and OSA has been examined in some well-designed studies. It could be hypothesized that theophylline works, at least in part, by inhibiting the ventilatory depressant effect of adenosine, which has been found to be elevated in the peripheral blood of patients with OSA (22). In addition, theophylline may stimulate ventilation by other mechanisms: increasing metabolic rate, stimulating hypoxic and hypercapnic ventilatory drives, and by improving respiratory muscle performance. Theophylline also has a positive inotropic action on the cardiovascular system, which may indirectly improve stability of breathing by decreasing circulation time, at least in congestive heart failure. These properties have led to its use in sleep-disordered breathing.

Theophylline administered as a blinded, overnight intravenous infusion improved central, but not obstructive, apneas compared with placebo (23) (Grade A). However, the sleep pattern was disturbed by this acute administration of theophylline. In a group of 15 men with left ventricular systolic dysfunction, congestive heart failure and periodic breathing with central apneas during sleep, 5 d of oral theophylline, producing a serum level of 11 µg/ml, decreased central apneas from a mean value of 26 to 6 per hour (24) (Grade A). Apneic arousals were significantly decreased, but the total number of arousals was not changed, and the abnormal sleep stage distribution pattern present in these patients before treatment was not affected by theophylline. These findings suggest that because of its cortical stimulating action, theophylline prevented improvement in the abnormal sleep pattern present in these patients, even though the ventilation pattern was improved.

Theophylline has been used to treat OSA. Guilleminault and Hayes (25) found that theophylline did not improve OSA in an uncontrolled study (Grade B). Mulloy and McNicholas (26) examined the effect of theophylline on obstructive apneas and hypopneas in a double-blind, cross-over study (Grade A). In this study nine patients had an AM serum theophylline level of 14 µg/ml on a once-a-day long-acting oral theophylline preparation. The obstructive apnea/hypopnea index decreased from a mean value of 48 to 40 events/h while receiving therapy. Mean duration of apneas did not decrease, total sleep time and sleep efficiency decreased and the number of sleep stage changes per hour increased while receiving theophylline. Thus, the minimal improvement induced in these patients' sleep-disordered breathing was offset by worsening of the sleep pattern.

In summary, most of the studies of the effects of theophylline on sleep-disordered breathing were well controlled. Theophylline was shown to be useful for central apnea and periodic breathing, particularly in patients with heart failure. Theophylline does not appear to be helpful for OSA, although the number of patients with OSA examined is limited. Theophylline's primary drawback is sleep disruption.

Opioid Antagonists and Nicotine

Group score = 9.9. Through generalized cortical stimulation, opioid antagonists stimulate ventilation. Because increased opioid activity has been identified in the cerebral spinal fluid of patients with OSA (27), it was hypothesized that OSA would improve with opioid antagonist administration. In a placebo-controlled study, Suratt and colleagues (28) (Grade A) found that doxapram infusion decreased the length of apneas and therefore the extent of the arterial oxygen desaturation, but it did not change the apnea number or the number of arterial oxygen desaturation episodes in four male patients with OSA. Atkinson and colleagues (29) (Grade A) derived similar findings with naloxone infusion.

In one of two studies reviewed nicotine gum chewed before bedtime decreased apneas in the first 2 h of sleep (30) (Grade B). Transdermal nicotine did not improve apnea index and worsened sleep quality (31) (Grade A). Nicotine affected sleep ventilation to some degree in this study since apnea duration, arterial oxygen saturation nadir, and mean snoring intensity were improved. Gastrointestinal side effects were experienced by a large proportion of subjects in this study.

The quality of the science in this group of studies is good. Although oxygenation may be somewhat improved by generalized cortical stimulation, the number of apneas and the associated sleep disruption did not lead to major improvements in the disease process.

Carbon Dioxide

In a study designed to stabilize fluctuations in arterial CO₂ tension, and thereby to study the role of these PCO₂ swings in inducing OSA, CO₂ was administered in a closed mask system, and decreased obstructive apneas considerably (32) (Grade B). Use of CO₂ given through an open system did not duplicate this effect. Two case reports have been published using CO₂ for OSA and central sleep apnea (33, 34) (Grade C). In the later study a tent with a constant 3% CO₂ concentration nearly completely eliminated periodic breathing and associated central apneas. The ventilatory stimulation produced by CO₂ would eliminate the hypocapnia that leads to hypopnea and periodic breathing. If a single method of stabilizing PCO₂ could be administered, this form of therapy might be promising.

Section Conclusion

Although many of the studies utilizing ventilatory stimulants were not optimally controlled trials, in total they allow us to make some statements about the efficacy of these agents in the treatment of sleep-disordered breathing. Stimulation of ventilation may be helpful in some patients, especially those with central apnea, periodic breathing, or hypercapnia. However, in general these agents have not been shown to be useful in most patients with classic OSA, although some exceptions exist. As noted above in the section on acetazolamide, stimulation of ventilatory drive(s) actually may worsen sleep-disordered breathing because a heightened ventilatory drive state may predispose patients to intermittent upper airway collapse. In this regard, Gleeson and colleagues (35) showed that healthy, nonapneic subjects with heightened hypercapnic ventilatory drives had greater hyperventilation and subsequent hypoventilation after airway occlusion in NREM sleep than did subjects with lower hypercapnic ventilatory drives. Therefore, if a pharmacologic agent that is known to stimulate ventilatory drive is used in a patient with normal or heightened ventilatory drives, unstable breathing during sleep may be produced, possibly contributing to sleep-disordered breathing. These ventilatory stimulating drugs would best be used in patients with abnormally low ventilatory drives such as hypercapnic, obese patients.

Different mechanisms of sleep-disordered breathing have been proposed as rationales for the use of the different pharmacologic agents discussed above. If these different mechanisms actually exist and contribute to the cause of sleep-disordered breathing to different degrees in different patients, then it is invalid to anticipate that one specific pharmacologic agent addressing one of these mechanisms will resolve the abnormal ventilation in all patients with sleep-disordered breathing. The predominant mechanism contributing to the ventilatory abnormality in a given patient needs to be identified and then a medication used that is appropriate for that particular mechanism identified. However, we did not find studies where different pharmacologic agents had been tried with this rationale. Now that some information is available on which pharmacologic agent works in what type of patient, for instance in a patient with central versus an obstructive apnea, more controlled trials comparing the effectiveness of different agents in subsets of patients with sleep-disordered breathing would be in order.

	PSYCHOTROPIC AGENTS

TOP ABSTRACT INTRODUCTION METHODS VENTILATORY STIMULANTS PSYCHOTROPIC AGENTS OTHER PSYCHOTROPIC AGENTS OTHER AGENTS SUMMARY FUTURE DIRECTIONS REFERENCES

Protriptyline

Group score = 7.7 (Table 4). Protriptyline (PROT), a nonsedating tricyclic antidepressant was inadvertently found to improve the OSA that coexisted with narcolepsy in some of these patients. Six studies that have examined the value of PROT in OSA were reviewed (21, 36). Two of these studies are Grade A and four are Grade B. In most studies PROT decreased REM sleep time. OSA is often worse in REM sleep; and, therefore, by decreasing the REM time, the apnea time and severity of arterial oxygen desaturation were often improved (38) (Grade A). In studies where differentiation between sleep stages was made, apneas were not improved in NREM sleep in two studies (38, 39), whereas they were improved in one study (40) (Grade B). In one double-blind, placebo-controlled study, PROT did not improve apneas, oxygenation, or symptoms (21), whereas acetazolamide led to some improvement in these same patients. There is a potential problem with the methodology in this study in that the PROT treatment course of 2 wk possibly was too short for the full therapeutic effect of PROT to become evident, supported by the fact that REM sleep time was not decreased. A longer course of PROT treatment would have been more useful. However, this is often difficult because PROT causes aggravating anticholinergic side effects such as dry mouth, constipation, urinary hesitancy, and impotence. The latter two conditions can and often do result in the discontinuation of the drug by male subjects.

In summary, PROT was most effective in decreasing REM-associated OSA. A decrease in apnea length might also have occurred, thereby improving apnea-related arterial oxygen desaturation. From these data, it can be concluded that the best candidates for PROT are those patients with OSA isolated to or significantly worse in REM sleep. PROT has significant anticholinergic side effects that are often difficult for many men to tolerate because of interference in bladder and penile function. Therefore, if we were to propose that PROT primarily be used in female patients with OSA, this recommendation would have to be based on data from 12 patients, only two of whom were included in Type A randomized, adequately controlled studies. Obviously, more work will be required to establish PROT as an acceptable treatment for OSA in women.

	OTHER PSYCHOTROPIC AGENTS

TOP ABSTRACT INTRODUCTION METHODS VENTILATORY STIMULANTS PSYCHOTROPIC AGENTS OTHER PSYCHOTROPIC AGENTS OTHER AGENTS SUMMARY FUTURE DIRECTIONS REFERENCES

Serotonergic Agents

Group Score = 7.9. L-tryptophan, the serotonin precursor, was effective in decreasing the apneas in NREM sleep in a group of 12 patients with OSA, but it was not effective in three patients with central apnea (41) (Grade B). In a crossover, unblinded trial of fluoxetine and protriptyline in 12 patients with OSA, Hanzel and colleagues (40) (Grade B) found a decrease from 57 to 34 in the apnea/hypopnea index in NREM sleep to both agents, but fluoxetine was much better tolerated than protriptyline because of issues discussed in the paragraph above. A reduction of apneas in NREM sleep also was observed with paroxetine, another selective serotonin reuptake inhibitor, in a randomized, placebo-controlled, crossover study (42) (Grade A, abstract). Because this study is currently only available in abstract form, many details are missing. Buspirone, a nonsedating asapione anxiolytic agent, decreased the number of apneas by one-third in five patients with OSA in a randomized double-blind study (43) (Grade A). Imipramine was found to be helpful in 13 patients with central apnea, again primarily in NREM sleep (44) (Grade B). On the basis of the studies available to date, it is obvious that further investigation needs to be completed before serotonergic agents can be recommended for treatment of OSA. Subpopulations that would best respond to these agents need to be defined.

Benzodiazepines

It is hypothesized that if sleep continuity can be preserved, sleep-disordered breathing could be improved. Central sleep apnea was decreased in two patients with both apneas and periodic leg movements, for which clonazepam was administered (45) (Grade C). Better controlled studies are needed.

	OTHER AGENTS

TOP ABSTRACT INTRODUCTION METHODS VENTILATORY STIMULANTS PSYCHOTROPIC AGENTS OTHER PSYCHOTROPIC AGENTS OTHER AGENTS SUMMARY FUTURE DIRECTIONS REFERENCES

Antihypertensive Agents

A decrease in sympathetic tone or a decrease in baroreceptor activity may also improve SDB. Hydralazine decreased spontaneous apneas during sleep in naturally hypertensive rats (46). Weichler and colleagues (47) examined the effect of metoprolol, a beta adrenergic blocking agent, and cilazapril, an ACE inhibitor, in a randomized, double-blind parallel study in humans (Grade A). Apneas decreased from 40 to 27/h with each agent by undetermined mechanisms. Blood pressure and symptoms improved similarly with each agent. It is not known whether the improvement in SDB and hypertension that occurred with these agents are causally related. Possibly by its REM-suppressant quality, clonidine, an alpha₂ adrenergic agonist, decreased apneas in a group of eight OSA male patients (48) (Grade A).

Glutamate Antagonists

Glutamate may be the neuromediator responsible for ventilatory stimulation during acute hypoxia. Therefore, glutamate-induced ventilatory stimulation during the intermittent hypoxemia that exists in OSA may contribute to the ventilatory instability present during sleep in OSA. Therefore, it may be possible to decrease ventilatory variability in OSA by the use of a glutamate antagonist. Hedner and colleagues (49) used sabeluzole for this purpose in a randomized, double-blind, crossover study in 13 patients with OSA (Grade A). Apneas per se were not recorded in this study, but oxygenation was assessed. Compared with placebo, there was no improvement in oxygenation produced by sabeluzole. The extent of change in oxygenation was related to the blood level of sabeluzole, such that high blood levels were associated with improved oxygenation. Although patients preferred sabeluzole to placebo, there was no decrease in the symptoms of OSA or improvement in mood-related parameters. In another study, one dose of baclofen, a gamma-aminobutyric acid agonist, glutamine antagonist, improved sleep quality in 10 patients with mild OSA, but it did not decrease the degree of SDB (50) (Grade A). Further investigations will be required to examine the efficacy of these and other glutamate inhibitors more thoroughly.

	SUMMARY

TOP ABSTRACT INTRODUCTION METHODS VENTILATORY STIMULANTS PSYCHOTROPIC AGENTS OTHER PSYCHOTROPIC AGENTS OTHER AGENTS SUMMARY FUTURE DIRECTIONS REFERENCES

In this review we have attempted to not only review the subject matter of the pharmacologic treatment of sleep-disordered breathing but also to review the quality of the science that produced these data. By grading the literature we can see that for most pharmacologic agents, an equivalent mixture of Grades A and B studies, with an occasional Grade C study, contributed to the available literature. The best studied agents are theophylline and the opioid antagonist/nicotine group with scores of 9.9, indicating that more randomized, appropriately controlled, studies were used to evaluate these agents than others. Conclusions about the efficacy of other pharmacologic agents used in the treatment of SDB are less reliable when based on groups of articles that are dominated by non-randomized, uncontrolled studies. In this review the literature on thyroxine is a good example, with a score of 3.6. Case reports and small uncontrolled studies (11) suggested a tight relationship between hypothyroidism and sleep-disordered breathing, which was refuted when a more rigorous design was used to examine this association (14). Thus, the quality of the science of a group of studies evaluating a given therapy is important in assessing the validity of the conclusions drawn from these studies (1). The results of this literature grading is seen in Table 5, which weighs the conclusions derived from the literature on various pharmacologic agents used in the treatment of sleep-disordered breathing. The number of subjects studied is also indicated.

Mechanisms by which pharmacologic agents might improve sleep-disordered breathing are: changing sleep stage distribution, increasing the relative activity of upper airway muscles during sleep to enhance upper airway patency, and altering the control of breathing in order to improve or stabilize the pattern of breathing. It would be beneficial to decrease the amount of time a patient spends in a sleep stage where the breathing disorder is most severe, usually REM sleep, and to increase the proportion of restful sleep stages, namely, NREM Stages 3 and 4. Interestingly, most of the pharmacologic agents reviewed above did not change sleep-stage distribution, except the antidepressant agents, which, as anticipated, decreased REM sleep time. One study showed that clonidine also had this property. In addition, these agents may also improve sleep-disordered breathing in NREM sleep.

	FUTURE DIRECTIONS

TOP ABSTRACT INTRODUCTION METHODS VENTILATORY STIMULANTS PSYCHOTROPIC AGENTS OTHER PSYCHOTROPIC AGENTS OTHER AGENTS SUMMARY FUTURE DIRECTIONS REFERENCES

Both mechanistic and therapeutic studies would advance this field. For instance, regarding mechanisms, there are no data demonstrating that the pharmacologic agents reviewed improve upper airway muscle activity. Improvement in upper airway muscle tone and inspiratory activity during sleep would be a potential mechanism of action of these agents. Whether it is good to stimulate or inhibit the control of breathing needs to be evaluated. If a patient has evidence of central alveolar hypoventilation, it may be good to stimulate ventilation, as is done with medroxyprogesterone. Women with hypercapnic ventilatory disorders may be particularly helped by medroxyprogesterone. On the other hand, in the patient with classic OSA, it may be better to decrease ventilatory drive so as to dampen the wide fluctuations in ventilation that occur in this disease.

A major effort should be made to conduct more pharmacologic studies in classic OSA since this large group of patients has had the poorest response to pharmacologic intervention in studies conducted to date. Subgroups of patients with OSA need to be studied, such as those with REM-specific disease. Protriptyline may be especially helpful in women with this entity. Women with hypercapnic ventilatory abnormalities may be especially responsive to medroxyprogesterone. The effect of antihypertensive agents appears promising, although many patients with OSA are already receiving these agents when they initially present. Further studies on the effect of altering the glutamine/ gamma-aminobutyric acid system could be promising. Surely, it is conceivable that a combination of medications might be appropriate if more that one mechanism is thought to be contributing to the sleep-disordered breathing. Thus, there is much work to be done in exploring a pharmacologic treatment for SDB.