INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Treatment of the obstructive sleep apnea (OSA) syndrome aims at alleviating symptoms such as excessive daytime sleepiness and reduced cognitive performance (1), and at preventing potential consequences of OSA such as traffic accidents (2) and impairment of social relationships and professional activities (3). Several randomized, placebo-controlled studies have shown that sleep-associated breathing disturbances and symptoms of OSA are alleviated by nocturnal application of continuous positive airway pressure (CPAP) (4). However, nightly wearing of the nasal mask and use of the pressure generator required for CPAP therapy, in addition to the inconvenience and occasional side effects of such therapy make it unacceptable to many patients. Compliance with this treatment is therefore less than ideal (8).

To overcome these drawbacks, alternative treatment options have been evaluated. One alternative nonsurgical modality consisting of intraoral appliances (IOAs) that advance the mandible, has been found to reduce breathing disturbances and snoring, improve sleep quality, and ease daytime symptoms in several studies (11). Randomized crossover trials suggest that the objectively measured beneficial effects of an IOA on breathing disturbances are less pronounced than the results achieved with CPAP, particularly in moderate to severe sleep apnea (11, 13). Nevertheless, many patients in these trials preferred an IOA over CPAP because of its greater convenience (11, 13).

A goal of research in the design and use of IOAs is therefore to improve the effectiveness and patient comfort of these devices by optimizing their design. This point was addressed in the present study. Our purpose was to perform a randomized, controlled, crossover study to allow direct comparisons among patients with untreated OSA and those fitted with two types of IOA, an OSA-Herbst appliance similar to that used by other groups (13, 15), and an appliance developed at our center and known as the OSA-Monobloc (Figure 1). The OSA-Herbst appliance consists of two independent parts, snapped onto the upper and lower dental arches, respectively. The two parts are connected by lateral telescopic rods that force the mandible into an anterior position while allowing limited lateral, protrusive, and opening motions of the jaw that may enhance comfort during use of this appliance. The OSA-Monobloc, on the other hand, incorporates an upper and a lower part, which are held firmly together by steel wire. Potential advantages of this design are its robustness and simplicity of application.

View larger version (78K): [in this window] [in a new window]   Figure 1.   Each patient was individually fitted with an OSA-Herbst (left panel ), and an OSA-Monobloc (right panel ) appliance. In an individual patient, protrusion was set identically for the two IOAs. Since the upper and lower parts of the OSA-Monobloc device are firmly held together by stainless steel wire, the relative position of the mandible and maxilla are fixed. In contrast, the lateral telescopic rods of the OSA-Herbst device allow limited opening and protrusive and lateral motion of the mandible.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

We enrolled successive patients with OSA in whom this diagnosis was based on a typical history including complaints of excessive daytime sleepiness, habitual snoring, nocturnal apneas observed by the patient's bed partner (where applicable), and either a sleep study with an apnea/ hypopnea index (AHI) of at least 5/h or a breathing pattern with signs of sleep-disruptive snoring (18, 19) combined with an increased arousal index of > 20/h. A trial of CPAP therapy was proposed first to all patients. It was explained that this was the standard treatment modality for OSA. Information on the option of IOA therapy was given to patients who refused to use CPAP from the beginning, or to those who declined continuation of CPAP therapy after an initial trial. It was emphasized that use of an IOA was a relatively new treatment modality for OSA, and was still under investigation, and that our purpose was to evaluate the effectiveness and side effects of two types of IOA.

Exclusion criteria for this study were sleep disorders other than OSA that can interfere with sleep quality (such as significant internal medical, neurologic, or psychiatric disorders). Active dental or gingival disease was treated first if necessary. Patients had to agree with regular use of the IOA and with medical and dental follow-up examinations according to the study protocol.

Twenty-four patients gave written informed consent to participate in the study, which was approved by the Ethics Committee of the University Hospital of Zurich. The patients' characteristics at the time of study enrollment are outlined in Table 1.

Measurements

Clinical evaluation and questionnaires. A general medical history was recorded and a clinical examination was performed on each patient. A German translation of the Sleep Disorders Questionnaire developed by Douglass and coworkers (20), the German version of the Epworth Sleepiness Scale (21), and part of the sleep symptom questionnaire described by Kump and colleagues (22), modified for our needs (APPENDIX I), were administered. Side effects and patient preference of IOA, and reasons for this preference, were evaluated in a structured interview.

Sleep studies. Polysomnography was performed as described previously (23). It included recordings from two electroencephalogram channels (C4/A1, C3/A2), left and right electrooculograms, submental and bilateral tibialis anterior electromyograms, recording of body position, temperature recording with an oral-nasal thermistor, and calibrated respiratory inductive plethysmography (Respitrace PT; Noninvasive Monitoring Systems Inc., Miami Beach, FL) (18). Snoring sounds were recorded with a miniature microphone taped to the patient's skin on the lateral aspect of the larynx. Data were recorded and analyzed on a computerized system (Alice 3; Healthdyne Technologies, Marietta, GA). Sleep stages and arousals were scored manually according to standard criteria (24, 25) through a review of 30-s epochs on a high-resolution computer video screen. An episode of apnea or hypopnea was defined as a decrease in the sum signal of the respiratory inductive plethysmograph to  25% of baseline for more than 10 s (18). An episode of oxygen desaturation or arousal was not required by this definition. A snore was defined by an excursion in the microphone channel that exceeded baseline noise by more than 50% and that was periodic with respiration. The snoring index was defined as the number of snores per hour of sleep (26).

Dental evaluation and IOA. Dental evaluation included a clinical examination, and cephalography. Wax-bite impressions and plaster dental models were used to individually mold the two types of IOA for each patient. The IOAs were manufactured of a methylmethacrylate material and a stainless steel alloy (Figure 1). The IOAs provided full coverage of the upper and lower dental arches. The initial protrusion was set at 75% of maximal protrusion (corresponding to a protrusion of 10 ± 0.4 mm [mean ± SE]), and was made identical for both IOAs. The vertical interincisal gap was fixed at 5 to 10 mm for the OSA-Monobloc device. For the OSA-Herbst device, the minimal vertical interincisal gap was 4 to 6 mm and the range of opening was greater than 15 mm.

Patients were encouraged to use both IOAs over a period of at least 1 mo to get accustomed to each device. If both devices were well tolerated and subjectively successful, the series of studies subsequently described were scheduled. This was the case with 16 of the 24 patients. If treatment was subjectively unsatisfactory, protrusion was increased by 1 to 3 mm (up to maximal protrusion) in one or several steps over the course of a few weeks until symptoms improved. If the IOA was not well tolerated, protrusion was reduced or other adjustments were made as necessary. In eight of the 24 patients, adjustments had to be made to the IOA (one change in protrusion in five patients, and two, three, and four changes, respectively, in three patients). The follow-up studies were scheduled thereafter.

Protocol

After initial evaluation and adaptation to the use of the two IOAs, patients underwent a sequence of treatment regimens and follow-up studies over a period of three consecutive weeks. During this period, they followed three different regimens, each lasting 1 wk. At the end of each week, symptoms and side effects were evaluated with questionnaires referring to the regimen of the previous 7 d. Nocturnal polysomnography was also performed. The three regimens consisted of treatment with the OSA-Herbst or OSA-Monobloc device or no treatment. Patients were randomized to one of the six possible sequences of the three treatment regimens. The randomization procedure assured a balanced design (four patients with each of the six possible sequences). The sequence of treatments and evaluations is illustrated graphically in Figure 2.

View larger version (12K): [in this window] [in a new window]   Figure 2.   Schematic representation of the study protocol. After initial baseline evaluation and an adaptation period during which they grew accustomed to the use of the two IOAs, patients underwent a sequence of three different regimens over the course of a 3-wk period. The regimens consisted in 1 wk of treatment with the OSA-Herbst device, 1 wk of treatment with the OSA-Monobloc device, and 1 wk without an IOA. The order of the three successive regimens was assigned through a randomization table that provided a balanced design. Evaluation at baseline and at the end of each 1-wk period during follow-up (Eval BL and Eval A, B, and C, respectively) included collection of data on symptoms and side effects by questionnaire, and a sleep study.

Statistics

Data were tested for normality of distribution with the Komolgorov- Smirnov statistic (27). Normally distributed data are summarized as means ± SE. Nonnormally distributed data are summarized as medians and quartile ranges. Data obtained during study phases with different treatment regimens were compared through analysis of variance (ANOVA) for repeated measures, followed by the Newman- Keuls multiple comparisons procedure (normally distributed data), or by Friedmann's ANOVA and Wilcoxon's matched-pairs test (nonnormally distributed data). A potential order effect of the different regimens was investigated through ANOVA. Expected versus observed frequencies were compared with the chi-square statistic, applying Yates's correction. A statistically significant difference among means was defined by a value of p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The randomized follow-up studies were performed after a period of 156 ± 14 d (mean ± SE) after initiation of IOA therapy. The patients' body weight had not changed during this time (86.3 ± 2.3 kg at baseline, versus 86.2 ± 2.4 kg at the time of the follow-up studies, p = NS). All patients stated that treatment with IOA had positive effects on their well-being, and all wished to continue this treatment. Treatment adherence could not be verified objectively, but the patients stated that they used one or the other IOA for at least four to seven nights per week. Inspection of the IOAs at the time of the follow-up studies showed signs of regular use of at least one of the two devices by every patient. The majority of patients (15 of 24) preferred treatment with the OSA-Monobloc over that with the OSA-Herbst device, eight patients had no clear preference, and only one patient preferred the OSA-Herbst device (p < 0.008 versus OSA-Monobloc). The following reasons were given for preference of the OSA-Monobloc device: greater effectiveness in relieving symptoms (11 of 15 patients); greater simplicity of application and robustness (five of 10 patients); and lack of side effects (four of 10 patients). The reason for preference of the OSA-Herbst device by the one patient who preferred it was a greater effectiveness in relieving symptoms.

Subjective ratings of sleepiness and several other aspects of daytime performance were significantly improved during the weeks with IOA treatment as compared with the week without treatment (Table 2). As compared with the OSA-Herbst device, the OSA-Monobloc device was slightly more effective in relieving OSA symptoms and snoring as subjectively perceived by the patients themselves. For 20 of the 24 patients, snoring had also been assessed by partners or family members. During the week without IOA therapy, 19 of the 20 relevant partners or family members were disturbed by the patient's snoring, as compared with nine of the 20 relevant partners who were disturbed during use of the OSA-Herbst device (p = 0.0007), and five of the 20 who were disturbed during use of the OSA-Monobloc device (p < 0.00001) (p = NS for chi-square test for comparison of OSA-Herbst and OSA-Monobloc devices).

The objective effects of IOA therapy on sleep and respiration are detailed in Table 3. Total sleep time and sleep efficiency did not differ during treatment and control nights. There was a relative increase in slow-wave sleep and a decrease in sleep fragmentation (arousal index) during nights with IOA treatment. However, the differences were significant only for the OSA-Monobloc device. Both IOAs improved sleep-disordered breathing and measured snoring. These effects were even more pronounced for the OSA-Monobloc than for the OSA-Herbst device, but the differences were not significant. Individual effects of the OSA-Monobloc and OSA-Herbst devices are shown in Figure 3. In every patient, the AHI was improved by one or the other IOA, with a reduction of from 2.4 to 43.7 apneas/hypopneas/h (mean ± SE improvement: 16.8 ± 2.4 apneas/hypopneas/h). In one patient, the use of the OSA-Herbst device and in another patient the use of the OSA-Monobloc device was associated with a small increase in the AHI (5.3 and 6.0/h, respectively), but in these patients the use of the other IOA was associated with a decrease in the AHI of 9.0 and 13.1/h, respectively.

View larger version (25K): [in this window] [in a new window]   Figure 3.   Plot of individual effects of OSA-Herbst and OSA-Monobloc devices on AHI in comparison with no treatment during randomized follow-up studies.

Treatment success, defined as a reduction in the AHI to less than 10/h and satisfaction with treatment, was achieved in 21 (88%) of the patients with either the OSA-Monobloc or OSA-Herbst device, in 18 patients (75%) with the OSA-Monobloc device, and in 16 patients (67%) with the OSA-Herbst device (p = NS for the OSA-Monobloc versus the OSA-Herbst device).

The AHI during treatment with the OSA-Herbst device but not with the OSA-Monobloc device was significantly correlated with the baseline AHI (R = 0.50, p < 0.01, and R = 0.49, p = NS). There was no significant correlation of the improvement in the AHI with the two IOA and either the AHI or the body mass index at baseline.

Statistical comparisons among the various radomized sequences of study regimens did not reveal any order effect. Furthermore, there were no significant differences in subjective symptoms and objective sleep and breathing disturbances among evaluations at baseline (Table 1 [sleep symptom questionnaire results not shown]) and corresponding data at the end of 1 wk without treatment (Tables 2 and 3).

Fourteen of the 24 patients experienced side effects of IOA treatment. These effects were mild and transient, lasting from a few minutes to 2 h after removal of the appliance in the morning. Seven patients complained of temporomandibular joint pain, four of muscle discomfort, and three of dental discomfort. The prevalences of these side effects were identical for the OSA-Herbst and OSA-Monobloc appliances, and the effects could not be relieved by switching from one appliance to the other. In contrast, three patients experienced transient mild mucosal erosions exclusively during use of the OSA-Herbst appliance, caused by its lateral attachments.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The results of our prospective, randomized and controlled trial demonstrate that symptoms and measured sleep and breathing disturbances of patients with OSA were significantly alleviated by both types of appliances examined in the trial as compared with the untreated condition. The study was the first to directly compare the effectiveness of IOAs with different designs. We found a clear patient preference for the OSA-Monobloc device related to its greater effectiveness in relieving symptoms and its simplicity of application. Subjectively perceived reductions in snoring and in certain aspects of impairment in daily activities were more pronounced with the OSA-Monobloc than with the OSA-Herbst device. In addition, there was a trend toward greater improvement in several objective variables of breathing and sleep disturbance with the OSA-Monobloc device.

Several groups have reported successful treatment of OSA with removable two-piece mandibular advancement devices with Herbst attachments. In three case series, one of them involving patients who had undergone uvulopalatopharyngoplasty (17), treatment success, defined as a reduction in the AHI to less than 10/h, was achieved in 11 of 15 (73%) (15), 10 of 19 (51%) (16), and 10 of 24 (42%) (17) patients, respectively. In one randomized prospective trial (13), the effect of the OSA-Herbst device was compared with nasal CPAP in patients with OSA. Reduction of the AHI to less than 10/h was achieved in only three of 21 patients (14%) with the OSA-Herbst device and in eight of 21 (38%) with CPAP. A direct statistical comparison of the results with IOA and CPAP treatment was not provided (13). A comparison of these published data with the success rate obtained with the OSA-Herbst device in the current study (i.e., reduction of AHI to less than 10/h and satisfaction with treatment in 16 of 24 patients [67%]) is hampered by differences in study design.

Two randomized crossover trials have evaluated the relative efficacy of oral appliances other than the OSA-Herbst device, and of nasal CPAP (11, 12). An anterior mandibular positioner (Snore-Guard; Hays & Meade Inc., Albuquerque, NM) was less effective than nasal CPAP in terms of reducing the mean AHI and in its success rate (reduction of the AHI to less than 10/h in 48% of patients with the Snore-Guard versus 62% with CPAP) (12). In another study performed in the same center, the success rate with an adjustable anterior mandibular positioner was 55%, as opposed to 70% with nasal CPAP (11). Again, comparison of these data with our results must take differences in study design into account. In particular, our protocol required crossover from one regimen to the other after 1 wk, and sleep studies were performed in the laboratory. This contrasts with the 4-mo periods to crossover and to home-sleep studies, respectively, in the two cited studies (11, 12). Furthermore, the untreated condition was not randomized in the study using the Snore-Guard (12).

Since polysomnographic and questionnaire data obtained at enrollment in our study and at the end of the 1 wk period without treatment were identical, the effect of an IOA seems to be short lived (i.e., less than 1 wk). On the other hand, the beneficial effects documented 1 wk after resumption of IOA therapy, and the lack of an order effect of the sequence of randomized regimens, suggest a relatively rapid onset of action of the IOAs examined. This may be particularly relevant during initiation of treatment and for patients who would like to transiently switch from nasal CPAP to an IOA during periods of impaired nasal passage, during travel, during transient deficiency of electrical power, or under other conditions.

According to patient preference, which was mainly related to subjectively perceived benefit (Table 2), and according to the trends of polysomnographic data (Table 3), the OSA-Monobloc device was more effective, with a similar prevalence of side effects, than the OSA-Herbst device. Both theoretically and on the basis of preliminary data from published studies (28, 29), improvements effected by IOAs in sleep- related breathing disturbances in OSA patients appear to be related to increases in upper airway caliber through protrusion of the mandible. Our results obtained with OSA-Monobloc and OSA-Herbst devices fabricated with identical protrusion for an individual patient suggest that factors in the design of the appliances other than protrusion are also important for patient comfort and objective success. We suspect that the amount of mouth opening, which differed with the OSA-Herbst and OSA-Monobloc devices, may at least in part account for the differences in their effectiveness. It is conceivable that downward rotation of the mandible with jaw opening may result in a retrusive motion of the tongue base at a given amount of protrusion measured at the level of the incisor cusps. Since the amount of mouth opening is not or is only loosely fixed by interarch elastics during application of the OSA-Herbst device, the effect of this appliance on airway caliber may vary in an individual patient with time, and among patients. This may explain in part the difference in the relative effectiveness of the OSA-Herbst and OSA-Monobloc devices. Other mechanisms that remain to be identified may also play an important role.

In the current study, the protrusion of the IOA was initially set at 75% of maximal protrusion. With this approach, two thirds (16 of 24) of the patients were successfully treated and did not require subsequent adjustments. The potential benefits and costs of routinely using appliances that incorporate components to facilitate adjustments of protrusion (such as Herbst attachments, a hinge [11], or a screw mechanism [30]) warrant further study.

Limitations of our own and other studies of IOAs published to date are the lack of objective assessment of the effect of IOAs on cognitive performance, and the lack of a placebo control. However, we are confident that the median reduction of 4.5 points in the Epworth Sleepiness Scale score achieved with the two IOAs examined in our study (Table 2) exceeds a placebo effect, since the latter corresponded to a reduction of only 2 points in the Epworth scale in two recent studies (5, 7).

In conclusion, our findings show that IOAs that advance the mandible are an effective alternative for treatment of selected OSA patients who are unwilling or unable to use nasal CPAP. If the symptomatic or objective improvement achieved with one IOA is not satisfactory, a different type of IOA may provide better results, even with an identical protrusion. Further research is needed to identify those specific characteristics of an IOA that provide maximal airway patency and comfort during application of this treatment modality for OSA.