An Oral Hypnotic Medication Does Not Improve Continuous Positive Airway Pressure Compliance in Men With Obstructive Sleep Apnea^* FREE TO VIEW

Obstructive sleep apnea (OSA) is a disorder that results from repeated upper airway narrowing or closure during sleep. In addition to sleep disruption and daytime sleepiness, OSA may cause neuropsychologic dysfunction, vehicular accidents, hypertension, metabolic disturbances, and cardiovascular disease.¹²³⁴⁵ Not surprisingly, increased health-care resource utilization has been reported in this patient population.⁶⁷

Continuous positive airway pressure (CPAP) stabilizes the vulnerable upper airway during sleep, relieves symptoms, and may prevent or mitigate the long-term complications of this condition.⁸⁹¹⁰ Although CPAP is widely regarded as effective treatment for OSA, there is no consensus as to what constitutes adequate usage.¹¹ It appears, for example, that many patients experience symptom relief by using CPAP < 4 h/night.¹²¹³ Whether this is enough to prevent the neuropsychologic and cardiovascular complications of OSA is unknown, and numerous strategies to increase compliance have been proposed.^1415161718

Current literature¹⁹²⁰ does not allow accurate prediction of individual long-term CPAP compliance, although it appears that the pattern of usage may be determined very early in the course of treatment. Therefore, efforts to optimize the initial experience with CPAP may be more rewarding than trying to salvage those patients who are failing or have given up. Many patients find CPAP intrusive and worry that they will not be able to sleep because of the air pressure, mask interface, or movement restriction. In the laboratory setting, we have found that a hypnotic medication often facilitates sleep during CPAP titration studies. Our experience also suggests that OSA patients who report significant insomnia are generally less tolerant of CPAP but sometimes increase usage when provided a hypnotic medication. For these reasons, we hypothesized that administration of a hypnotic agent would promote acclimation to CPAP by facilitating sleep during the critical period of treatment initiation. To our knowledge, the adjunctive use of a hypnotic medication to improve CPAP compliance has not been previously reported.

Consecutive patients referred for initiation of CPAP treatment at Naval Medical Center San Diego were screened and enrolled between April 2001 and June 2003. Inclusion criteria included age ≥ 18 years, an apnea/hypopnea index (AHI) ≥ 5, subjective daytime sleepiness manifest by an Epworth sleepiness scale (ESS) score ≥ 10, and willingness to try CPAP. Patients were excluded if they had previously used CPAP, had a known allergy or intolerance to the study medication, or were involved in an operational (ie, military) assignment that might preclude nightly use of CPAP or planned follow-up. Prior to referral for CPAP treatment, all patients had undergone full-night (n = 27) or split-night (n = 32) polysomnography or a cardiopulmonary sleep study (n = 13). Optimal CPAP pressure settings were determined in the sleep laboratory as part of a split-night protocol (n = 32) or overnight titration study (n = 40). Seventy-two patients with newly diagnosed OSA referred for CPAP treatment met the study criteria and agreed to participate. The mean age (± SD) of participants was 38 ± 7 years, and all patients were male. During this period, no women were referred for CPAP treatment, reflecting the relatively young population and male predominance of US military personnel, as well as the male predominance of the disorder. The study protocol was approved by the Naval Medical Center San Diego Scientific Review Committee and Committee for the Protection of Human Subjects. Informed consent was obtained from each participant.

This study was designed as a prospective, randomized, placebo-controlled clinical trial. Participants were assigned through a central randomization scheme to one of three groups: (1) zolpidem, 10 mg; (2) placebo pill; or (3) standard therapy (no hypnotic or placebo pill) [Fig 1] . At enrollment, each participant received a paper bag containing a bottle with 14 tablets (zolpidem or placebo pill) or an identical cotton-filled container (standard care group). Participants and investigators were thereby blinded to the contents of the bottle; however, subjects randomized to the standard therapy arm were no longer blinded after opening the empty bottle on the first night. Patients were told to take one pill each night 30 min prior to bedtime and not to lie down after taking the medication without initiating CPAP treatment. They were warned that the study medication could make untreated OSA worse and that they should not take the medication if they were not going to use CPAP that night.

Our sleep center performs overnight polysomnography, including split-night CPAP titrations (criteria for initiating CPAP: AHI ≥ 40 or ≥ 20 with oxygen desaturation ≤ 75% over the first 2 h of sleep) and single-night, six-channel (airflow, thoracic effort, abdominal effort, pulse oximetry, heart rate, and position) cardiopulmonary home studies in the evaluation of patients with sleep-disordered breathing. For this reason, the AHI of patients referred for CPAP treatment was based on slightly different methodologies. The AHI in patients undergoing overnight polysomnography and split-night studies was calculated by dividing the number of apneas and hypopneas by total sleep time. The AHI in patients undergoing a cardiopulmonary home study was calculated by dividing the number of apneas and hypopneas by total study time. Obstructive apneas and hypopneas were scored according to commonly accepted criteria: apneas were defined by cessation of oronasal airflow for at least 10 s with evidence of ongoing ventilatory effort, and hypopneas were defined by a reduction of oronasal airflow by at least one third with an associated oxygen desaturation ≥ 4%. All CPAP titration studies were performed in the laboratory and consisted of manipulation of pressure to eliminate evidence of upper airway obstruction including apnea, hypopnea, desaturation, snoring, and flow limitation.

All patients participated in a standardized “one-on-one” CPAP training session utilizing a checklist of topics that included nasal mask fitting, review of machine function and maintenance, and viewing of a commercially produced video (“Rise and Shine”; Respironics; Murrysville, PA) that explained the benefits of CPAP therapy. All patients were issued a CPAP unit (Aria LX; Respironics) that included internal usage (time at pressure) recording capability and passive humidification system. Patients were not informed of the recording capability of the CPAP machine but were told to bring it to the follow-up visit.

At the completion of the 4-week trial, CPAP usage data were downloaded into a personal computer using proprietary software (Encore; Respironics) provided by the manufacturer of the CPAP units. In addition to time spent at effective pressure (hours) during each day of treatment, the software provided the following data: days with device usage, days without usage, cumulative usage, average hours of use (all days), average hours of use (days used), percentage of days used < 4 h, and percentage of days used > 4 h.

Subjective sleepiness was assessed with the ESS at baseline and following treatment.²¹ Higher ESS scores indicate greater sleepiness, and a cut point of 10 is often used to distinguish normal (< 10) from excessive (≥ 10) daytime sleepiness. All subjects also completed the 26-item functional outcomes of sleep questionnaire (FOSQ) before and after treatment with CPAP.²² Scores were reported by subcategory (activity level, vigilance, intimacy, general productivity, and social outcomes) and summed to give a “global” score. Higher FOSQ scores indicate better function.

Compliance with the study medication was determined by having the patient bring the pill bottle to the follow-up appointment and performing a manual pill count. Patients were not contacted during the 28-day trial, although they were encouraged to call the clinic if they had any problems or questions.

In order to detect a 1-h difference in CPAP usage between groups, 23 subjects in each trial arm were required to achieve 80% power to exclude differences of > 1.2 SD, with α = 0.05. The three treatment groups were compared using a one-way analysis of variance or Kruskal-Wallis (rank) test for continuous data, and Fisher exact test for categorical data. Within-group changes in ESS, global FOSQ, and FOSQ subcategories following treatment were evaluated with either a paired t test or Wilcoxon matched-pairs test. Normality of all variables was assessed with the Shapiro-Wilk test. Statistical significance was taken as p < 0.05. All data analysis was performed using commercially available statistical software.

Seventy-two male patients (mean age, 38 ± 7 years; range, 22 to 59 years) were randomized to receive zolpidem (n = 24), placebo pill (n = 24), or standard care (n = 24) for the first 14 days of CPAP treatment. Comparison of baseline demographic parameters revealed no difference in age, body mass index, ESS, FOSQ, nadir oxygen saturation, or CPAP pressure setting. We also did not find a significant difference among the three groups based on the type of sleep study that had been performed prior to enrollment. The breakdown of studies was as follows: zolpidem group (polysomnography, n = 10; split-night study, n = 11; cardiopulmonary home study, n = 3); placebo pill group (polysomnography, n = 10; split-night study, n = 8; cardiopulmonary home study, n = 6); and standard care group (polysomnography, n = 7; split-night study, n = 13; and cardiopulmonary home study, n = 4) [p = 0.584]. Despite randomization, however, the standard care group had a higher AHI than either the zolpidem or placebo pill groups (54.75 ± 28.02 vs 32.61 ± 25.12 vs 38.09 ± 25.65, respectively; p = 0.012) [Table 1] .

Complete, internally recorded CPAP usage data for each of the 28 days were available for 69 of 72 patients. For one patient in the zolpidem group, only the total number of days used and the average use per night were available. For two patients in the placebo pill group, < 28 days of data were available (11 days and 7 days, respectively). These subjects were presumed not to have used CPAP on days with missing data and were included in the analysis according to the intention-to-treat principle. We arbitrarily chose a cutoff of 10 min to determine whether CPAP had been used on a particular day. Any usage < 10 min was counted as a day not used. All patients completed prestudy and poststudy questionnaires. Pill counts were available for 61 of 72 patients. Of these, only two bottles were not empty. In one case (from the zolpidem group), the patient reported stopping the medication after the second night because of a sleepwalking event witnessed by his wife. This was the only adverse medication effect reported by our patients. The other patient (from the placebo pill group) took only 4 of 14 tablets because he used CPAP only a few times during the trial and was following instructions not to take the pill when not using CPAP. All the other patients (including the 11 patients who failed to bring the bottle back for counting [3 patients from the zolpidem group, 4 patients from the placebo group, and 4 patients from the standard care group]) reported completing the medication as prescribed.

Compared to placebo pill and standard care groups, the zolpidem group did not show greater CPAP usage in terms of total days used (zolpidem, 20.58 ± 7.40 days; placebo pill, 17.83 ± 9.33 days; standard care, 22.92 ± 6.95 days; p = 0.198) or average time used per night (4.43 ± 1.16 h, 4.23 ± 2.14 h, and 4.94 ± 1.44 h; p = 0.361) over the course of 4 weeks (Table 2 ). When the initial 14 days of CPAP treatment were analyzed separately (the period when zolpidem was actually supposed to be taken), there was also no difference in the number of days used (zolpidem, 11.08 ± 3.66; placebo pill, 9.54 ± 4.60; standard therapy, 12.13 ± 3.23; p = 0.115) or average nightly use (zolpidem, 4.69 ± 1.43 h; placebo pill, 4.39 ± 2.12 h; standard care, 4.94 ± 1.44 h; p = 0.572). Although average nightly CPAP usage within the zolpidem group was higher in the initial 14 days than in the subsequent 14 days (4.69 ± 1.43 h/night vs 4.14 ± 1.43 h/night), this did not reach statistical significance (p = 0.197). We also compared the number of “regular users,” ie, patients who used CPAP on at least 70% of nights and averaged at least 4 h/night across the three groups, and found no difference (zolpidem, 58.3%; placebo pill, 50%; and standard therapy, 62.5%; p = 0.761).

There was a significant decrease in self-reported sleepiness and improved function in each of the groups. For the overall group, the ESS score dropped from 15.38 ± 3.51 to 8.94 ± 5.34 (p < 0.001) and the global FOSQ increased from 14.37 ± 2.82 to 17.10 ± 2.73 (p < 0.001) [Table 3] . Seventy-one of 72 patients agreed to continue nasal CPAP therapy at the conclusion of the 1-month trial.

Application of positive pressure through the nose or mouth maintains upper airway patency during sleep in patients with OSA. Numerous clinical studies,^{89232425262728293031} including placebo-controlled and sham trials, have demonstrated that CPAP relieves symptoms. Many patients, however, completely reject or use CPAP less than recommended and may forgo important treatment benefits. We hypothesized that administration of an oral hypnotic agent to a group of patients referred for CPAP initiation would facilitate acclimation to a relatively intrusive therapy and increase hours slept under effective treatment. Although all three of our CPAP treatment groups showed significant symptom improvement, as measured by the ESS and FOSQ, an oral hypnotic agent did not increase initial CPAP compliance.

There are several obvious limitations of our study that warrant consideration. First, our study was powered to detect a difference of 1 h in CPAP usage per night, and a smaller, yet clinically significant difference, could have been missed. Second, despite our randomization scheme, our standard care group had a significantly higher AHI than either the zolpidem or placebo pill groups. One could make the argument that the higher AHI in the standard care group might have lead to greater CPAP usage, thereby masking a beneficial effect of zolpidem in patients with less severe disease. Although this is a possibility, it should be noted that there was no difference in the baseline AHI between the zolpidem and placebo pill arms and CPAP usage was not significantly greater in those receiving zolpidem than in those receiving a placebo pill. Third, all of our subjects were male, and extrapolation of these findings to women may not be appropriate. It is generally accepted that women in both the general and OSA populations are more likely to report insomnia symptoms and might, therefore, be more likely to benefit from treatment with a hypnotic agent.³²³³ Fourth, our patients were somewhat younger than those generally reported in the adult OSA literature, and our results may not apply to an older population. Finally, we assumed that difficulty initiating sleep was a relatively common problem for patients beginning CPAP treatment. Many patients do complain of mask-related discomfort and excessive air pressure that inhibit relaxation and prolong sleep latency. Our goal to maximize the initial response to CPAP by providing a hypnotic medication on day 1 of CPAP treatment did not allow us to target this patient group. Clearly, not all new CPAP users have difficulty initiating sleep.

Zolpidem is a relatively short-acting (half-life, 1.5 to 2.4 h) nonbenzodiazepine (imidazopyridine) hypnotic agent widely used for treatment of insomnia. In normal individuals, doses ≤ 10 mg reduce sleep latency without effecting sleep architecture.³⁴ We selected a nonbenzodiazepine agent because the literature suggests less respiratory depression compared to benzodiazepines, although most of the data are from normal individuals or patients with chronic respiratory disease.^{35363738394041} There is a paucity of published data from the OSA population.⁴²⁴³ A recent placebo-controlled study⁴⁴ of another nonbenzodiazepine hypnotic agent (zalpelon) administered for 5 consecutive nights to patients receiving CPAP did not find a statistically significant difference in apnea indexes between groups, although the oxygen saturation nadir was slightly lower in the zalpelon group (79.2 ± 1.3% vs 82.1 ± 0.9%, p = 0.008). Nevertheless, we did strongly counsel our patients not to take the study medication without using CPAP.

It appears that long-term CPAP compliance may be determined or predicted by CPAP usage early in the treatment course, and interventions designed to optimize the initial experience may be critical.^19204546 The importance of the first night of sleep with CPAP was suggested by a study⁴⁷ that found that patients who experienced higher sleep efficiency during a titration study (compared with a baseline diagnostic study) used CPAP an average of 2 h more per night (mean, 46-day follow-up) than those patients who did not have an improvement in sleep efficiency during the CPAP titration study. Lettieri et al⁴⁸ recently reported their experience using zolpidem in the sleep laboratory to facilitate CPAP titration studies. Their retrospective data suggested a higher rate of successful CPAP titrations in patients who received premedication with zolpidem. This supports our anecdotal experience, although our laboratory policy has been to reserve hypnotics for those patients who have difficulty initiating sleep with CPAP. We have not routinely premedicated our patients undergoing CPAP titration in the laboratory.

Two studies⁴⁹⁵⁰ found a high prevalence of insomnia complaints among patients with OSA. Effective treatment of the underlying upper airway obstruction will likely ameliorate insomnia symptoms in some patients. Unfortunately, the intrusive nature of CPAP treatment may actually exacerbate insomnia in others. Tailored behavioral and pharmacologic strategies to manage insomnia while facilitating CPAP treatment are needed.

In conclusion, we were unable to increase initial CPAP compliance by providing a nonbenzodiazepine hypnotic medication to a population of CPAP-naïve male patients with OSA and cannot recommend empiric treatment. Further studies evaluating potential roles for hypnotic medications in subpopulations of OSA patients with preexisting insomnia or who have difficulty initiating or sustaining sleep while receiving CPAP may be more rewarding.