Maxillomandibular Advancement and Upper Airway Stimulation for Treatment of Obstructive Sleep Apnea: A Systematic Review

This systematic review aimed to comparatively evaluate the efficacy and safety of maxillomandibular advancement (MMA) and upper airway stimulation (UAS) in obstructive sleep apnea (OSA) treatment. A MEDLINE and Embase database search of articles on MMA and/or UAS for OSA was conducted. Twenty-one MMA studies and nine UAS studies were included. All the MMA studies demonstrated a reduction in apnea hypopnea index (AHI) postoperatively, and success rates ranged from 41.1% to 100%. Ten MMA studies reported pre- and postoperative Epworth sleepiness scale (ESS), and all but one study demonstrated a reduction in ESS. In the UAS studies, all but one demonstrated a reduction in AHI, and success rates ranged from 26.7% to 77.8%. In the eight UAS studies reporting pre- and postoperative ESS, an ESS reduction was demonstrated. No studies reported any deaths related to MMA or UAS. The most common postoperative complications after MMA and UAS were facial paresthesia in the mandibular area and discomfort due to electrical stimulation, respectively. This systematic review suggests that both MMA and UAS are effective and generally safe therapies for OSA. However, due to the limitations of the included studies, there is no evidence yet to directly compare these two procedures in OSA treatment.


Introduction
Obstructive sleep apnea (OSA) is a prevalent sleep-related breathing disorder characterized by recurrent upper airway obstruction during sleep [1], and its overall prevalence ranges from 9% to 38% in the general adult population [2]. OSA is associated with considerable health risks, such as cardiovascular and cerebrovascular disease [3,4]. Continuous positive airway pressure (CPAP) is accepted as the first-line therapy for moderate to severe OSA, but poor compliance and suboptimal use of CPAP drive OSA patients to seek alternative therapies, including other non-invasive therapies and surgical treatment [5,6].

Materials and Methods
This systematic review was performed in accordance with the preferred reporting items for systematic review and meta-analysis (PRISMA) statement [19]. The protocol for this systematic review was registered at PROSPERO (PROSPERO ID: CRD42021261394; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=261394 (accessed on 14 November 2022)).
The exclusion criteria were: (1) sample size < 10 patients; (2) patients who underwent other adjunctive surgical procedures (e.g., uvulopalatopharyngoplasty) at the time of MMA or UAS; and (3) preliminary studies in which the findings had been nested in other studies with larger sample size and/or longer follow-up.

Literature Search
A literature search was performed with the help of an information specialist (RS) using MEDLINE and Embase databases on 14 December 2021. Search terms and search strategies used for each database are available in Supplementary Materials (Table S1 (a)).

Study Selection
After removal of duplicate articles, the remaining results were screened based on title and abstract by two independent reviewers (NZ and JH). The full texts of potentially relevant articles were retrieved and further evaluated by NZ and JH independently for compliance of studies with the eligibility criteria. Discrepancies were resolved by discussion. Reference lists of eligible studies were checked for additional studies.

Data Extraction
The extracted data included article title, year of publication, first author, study design, specific surgical technique, length of follow-up, sample size, age, gender, body mass index (BMI), preoperative and postoperative PSG data (AHI, respiratory disturbance index (RDI), and oxygen desaturation index (ODI)), preoperative and postoperative ESS score, preoperative and postoperative data on quality of life (QoL), surgical success rate and cure rate, and postoperative complications. According to the accordion severity grading system of surgical complications [20], the postoperative complications were classified as major or minor depending on the needs for endoscopic or interventional radiologic procedures or reoperation as well as failure of one or more organ systems.
Data were extracted by NZ and JH independently. Discrepancies were resolved through discussion. If RDI was reported by a study, it would be extracted as AHI, since these two respiratory parameters have been consolidated based on the 2013 American Academy of Sleep Medicine's manual for the scoring of sleep and associated events [21]. If there were multiple follow-up data in a study, the data with longest follow-up time were included. Surgical success was defined as "a postoperative AHI < 20 and at least 50% reduction in AHI after surgery" [22], and surgical cure was defined as "a postoperative AHI < 5" [23].

Quality Assessment
Methodologic quality assessment of each study was performed by NZ and JH independently, and any discrepancies were resolved by discussion.
The Methodological Index for Non-Randomized Studies (MINORS) quality assessment tool, a validated tool for the methodological assessment of non-randomized surgical studies [24], was used to assess the methodological quality of the included studies. The MINORS tool is composed of eight items applicable to all non-randomized studies and four additional items specifically for comparative studies. Each item was scored as 0 (not reported), 1 (reported but inadequate), or 2 (reported and adequate), giving a global ideal score of 24 for comparative studies and 16 for non-comparative studies. For comparative studies, the categorizations are as follows: 0-6, very low quality; 7-10, low quality; 11-15 fair quality; and ≥16, high quality. For non-comparative studies, the categorizations are as follows: 0-4, very low quality; 5-7, low quality; 8-12, fair quality; and ≥13, high quality [25].

Statistical Analysis
The collected parameters (age, BMI, AHI, ODI, and ESS) were pooled by weighted average and weighted standard deviation [26]. When there were RCTs or comparative studies between MMA and UAS, meta-analyses were performed to compare the overall effect of MMA and UAS in treating OSA. Heterogeneity of the studies was assessed using the I 2 statistic with a cutoff of 25% (low), 50% (moderate) and 75% (high) [27]. When moderate-to-high heterogeneity was present, a random effects model was adopted; otherwise, a fixed effects model was used. Because some patients may report multiple complications, the complication rate of each study was calculated by dividing the number of events by the number of patients.

Search Results
The flow diagram of study selection progress is summarized in Figure 1. A total of 2952 studies were screened after deduplication, and 212 were retrieved for full-text review.
Because there was no RCT or comparative study of MMA and UAS in treating OSA, a meta-analysis could not be performed to compare their overall effect sizes on OSA.

Quality Assessment
MMA group. One of the included studies was an RCT of MMA and autotitrating positive airway pressure (APAP), one was a retrospective quasi-experimental study, ten were prospective cohort studies, and nine were retrospective cohort studies. As only the MMA cohort of the RCT was included in the analyses, after omitting the unrequired APAP cohort, this study was regarded as a single-arm trial. The quality of the RCT was therefore assessed using the MINORS tool as per the other included studies. Of these studies, three studies were classified as "high quality", and the others were classified as "fair quality" (Supplementary Table S2 (a)).
UAS group. Six prospective studies and three retrospective studies were included. Of these, one study was classified as "high quality" and eight studies as "fair quality" (Supplementary Table S2 (b)).
Of four studies [11,32,36,43] (n = 78) reporting pre-and postoperative ODI, two demonstrated a significant reduction in ODI after MMA (p < 0.05), and the other two also reported an ODI reduction but without a p value. The weighted pre-and postoperative ODIs were 35.1 ± 22.8/h and 6.3 ± 6.4/h, respectively. UAS group. Of the selected studies, the study by Bachour et al. [55] did not show a significant reduction in AHI postoperatively. Five studies [48][49][50][51]54] demonstrated a significant reduction in AHI postoperatively (p < 0.05), and three studies [15,52,53] showed an AHI reduction but did not report a p value. The weighted pre-and postoperative AHIs in 1003 patients were 35.2 ± 14.7/h and 15.0 ± 16.1/h, respectively.

Subjective Parameters
MMA group. Of nine studies [11,34,36,[41][42][43][44][45]47] (n = 217) reporting pre-and postoperative ESS, the study from Lin et al. did not show an improvement in ESS after MMA, one study demonstrated a reduction in ESS but without a p value, and the others reported a significant reduction in ESS (p < 0.05). The weighted pre-and postoperative ESS values were 13.1 ± 5.5 and 6.7 ± 4.8, respectively.
Three studies [30,42,44] assessed pre-and postoperative QoL. Boyd et al. found that after MMA, there was a significant improvement in the Functional Outcomes of Sleep Questionnaire (FOSQ) (p < 0.05) [30]. Veys et al. assessed the subjective outcome of MMA using the OSA QoL questionnaire. They found that there was an improvement in all of the following six symptoms after MMA-daytime sleepiness, snoring, concentration, waking up at night, headache, and high blood pressure-while the influence of MMA on nocturia and sexual activity was variable [44]. Lin et al. found that there was no significant improvement in Short Form-36 quality of life (SF-36) after MMA [42]. UAS group. Of eight studies [15,[49][50][51][52][53][54][55] reporting pre-and postoperative ESS, seven demonstrated a significant reduction in ESS postoperatively (p < 0.05), and one reported a ESS reduction but did not report a p value. The weighted pre-and postoperative ESS values were 11.4 ± 5.4 (n = 1006) and 7.0 ± 4.6 (n = 1001), respectively.
Two studies reported pre-and post-UAS FOSQ scores. The STAR trial cohort demonstrated an increase in FOSQ score five years after surgery (14.3 ± 3.3 to 18.0 ± 2.2). Van de Heyning et al. also found a significant improvement in FOSQ score postoperatively (89.1 ± 23.5 to 100.8 ± 16.9, p < 0.05).
Since one study [54] did not report the count of minor complications, the safety outcomes of a subset of the study population (ADHERE cohort) reported in a previous study [56] were used to analyze the minor complication rate. In that study [56], the rates of minor surgery-related and device-related complications 137 ± 77 days after UAS implant were 6% (18/313) and 22% (69/313), respectively; 386 ± 136 days after UAS implant were 4% (8/217) and 24% (53/217), respectively. In the STAR trial cohort [15] consisting of 126 participants, the rates of minor surgery-related and device-related complication were both 136% (171/126) at the first year; at the fifth year, they were decreased to 1% (1/126) and 16% (20/126), respectively. Van de Heyning et al. [52] reported only minor surgery-related adverse events in their population, which yielded a minor complication rate of 57% (16/28). Philip et al. [49] and Steffen et al. [51] did not report any minor complications in their study populations. The most common minor surgery-related and device-related complications were incision discomfort [15,51,56] and discomfort due to electrical stimulation [15,56], respectively.

Discussion
This is the first systematic review aiming to comparatively evaluate MMA and UAS therapy in treating OSA. We reviewed 21 studies on MMA and 9 studies on UAS in treating OSA. Due to the fact that there is no RCT or comparative study of MMA and UAS, a meta-analysis cannot be performed to directly compare these two interventions. Separate analyses of studies on MMA and UAS were utilized for this review. In this review, the trials for MMA tended to be published earlier than those for UAS. Therefore, for some patients in the UAS group, MMA could have been considered at first as an alternative therapy to CPAP and not been chosen. It should be noted that UAS therapy has stricter and clearer inclusion criteria (e.g., 15/h ≤ AHI ≤ 65 /h, absence of CCCp during DISE) [14,17] for patients, especially in comparison to MMA. There is therefore discrepancy of patients' baseline characteristics between the MMA cohort and UAS cohort. In this review, the MMA cohort has younger age and higher baseline AHI compared to the UAS cohort. Moreover, it is impossible for us to compare other patients' characteristics associated with OSA, such as the size of tongue, retrolingual space, and jaw position. To obtain definitive results on the comparison of MMA and UAS, future studies should include comparative studies of these two therapies where participants would have comparable baseline characteristics and be qualified for both therapies. Another point to be noted is that the variations in MMA surgeries are probably greater than in UAS as the training and the lineage of potential variations are much higher in MMA than in UAS.

Objective Outcomes
Based on the separate analysis of studies on MMA and UAS, we reported that these two procedures are both effective treatment modalities for OSA. However, compared to UAS, MMA seems to be more effective in treating OSA with a more significant decrease in AHI and higher success rate. Through different mechanisms, MMA and UAS have been proven to be able to address multiple sites of collapse simultaneously [11,36]. MMA enlarges the entire pharynx and reduces the collapsibility of the upper airway by advancing the maxillomandibular complex and anterior pharyngeal tissues attached to the maxilla, mandible, and hyoid bone [39]. The mechanism by which UAS resolves multilevel col-lapse, is enlargement of the retropalatal airway associated with tongue protrusion, which is so called "palatoglossus coupling" phenomenon [48]. Safiruddin et al. found that the retropalatal enlargement in response to UAS was statistically significant only in the responders, while the responders and non-responders had similar degrees of retrolingual opening to stimulation [57]. Therefore, we are of the opinion that the superiority of MMA over UAS in OSA treatment may be associated with the ability of MMA to enlarge the retropalatal airway more significantly. To improve patient selection for MMA and UAS, the mechanism of action of these two surgical procedures and the role of pathogenesis of OSA on the outcome of both surgeries require clarification in future studies.

Subjective Outcomes
It is interesting to note that several studies [42,55] reported a discordance between objective outcome measures (e.g., AHI) and patient-reported outcome measures, which highlights the importance of subjective outcome evaluation for OSA patients. In contrast to published ESS data, there is a scarcity of evidence related to other subjective outcomes of surgical treatment for OSA. Boyd et al. [30] evaluated the impact of MMA on quality of life (QoL) using the Functional Outcomes of Sleep Questionnaire (FOSQ). Two years after MMA, a significant improvement in mean FOSQ scores of 4.7 was observed. In a study by Woodson et al. [15], the improvements in mean FOSQ scores following UAS were 3.0 at 1 year and 3.7 at 5 years, respectively. In addition to daytime sleepiness and QoL, patient satisfaction-an important measure of therapy quality-should be noted when evaluating treatment options for OSA. Currently, only a few studies have evaluated patient satisfaction with MMA or UAS for the management of OSA [56,[58][59][60][61][62]. In a study by Butterfield et al. [59], 95.5% of patients were satisfied with MMA surgery for OSA, 90.9% would repeat the procedure, and 86.4% would recommend MMA to others for OSA treatment. In the ADHERE registry, 94% of patients reported that they were satisfied with UAS therapy and would undergo UAS again, and 93% reported that they would recommend UAS to others [56]. According to the available evidence, both MMA and UAS could significantly improve the perception for OSA patients with high levels of patient satisfaction. However, the comparison of improvement in patient-perceived measures between the two therapies must be addressed in future studies.

Long-Term Outcomes
The long-term follow-up period of the included MMA studies ranges from 2 years to 12.5 years. Because of the small sample size, one study by Pottel et al. [63] reporting the longest follow-up result of MMA was excluded. In that study, the short term (within 2 years) success rate was 66.67% (8/12), and the long-term (median 19 years; range 14-20 years) success rate of MMA was 44.44% (4/9). Of the nine patients who attended long-term re-evaluation, the median ages at the time of MMA surgery and re-evaluation were 43 years (range 34-63 years) and 62 years (range 49-82 years), respectively. At the long-term follow up, two of the six patients who were initially successfully treated by MMA had relapse of OSA with AHI comparable to preoperative values. Both patients had significant weight gain (+4.1 and +7.9 kg/m 2 ). In a study of 29 OSA patients treated by MMA, Vigneron et al. [46] concluded that the success rate was 85.7% in the immediate postoperative period and 41.1% at 12.5 years. Additionally, they concluded that the good candidates for longterm success of MMA were the young patients (<45 years old) with BMI < 25 kg/m 2 , AHI < 45/h, SNB angle < 75 • , narrow retrolingual space (<8 mm), preoperative orthodontics, and without co-morbidity. It has been suggested that long-term failure of MMA might be attributed to weight gain [38,63,64], skeletal relapse [64], and ageing [63]. Given that UAS is an innovative therapy for OSA from the last decade, the longest follow-up period of the UAS studies was 5 years, from the STAR trial [15]. The success rates of UAS in the STAR trial cohort were 66% (83/126), 74% (73/98), and 75% (53/71) at 1, 3, and 5 years, respectively. In UAS therapy for OSA treatment, patients' adherence is necessary to guarantee clinical efficacy [65]. The STAR trial revealed a high adherence to UAS therapy in the long-term, with a patient-self-reported nightly device use of 80% at 5 years, which might partially explain the stability of treatment effect. In addition, lower baseline ODI was found to be predictive of 5-year response to UAS therapy. It is therefore concluded that both MMA and UAS were relatively stable treatments for patients with moderate-to-severe OSA. In order to maintain clinical efficacy, more effort is needed to provide continuous follow-up for OSA patients and to ascertain the factors associated with long-term stability of outcomes.

Safety
In terms of treatment safety, this systematic review revealed that both MMA and UAS were generally safe surgical procedures for OSA, with relatively low rates of major complication. In the included MMA studies, all but one of the major complications were reoperation for removal of hardware. Age has been shown to be a risk factor for increased need for hardware removal [66]. In addition, Passeri et al. found that patients who were active smokers or had a history of smoking had higher risk of complications, which included removal of hardware [67]. The most common minor complication of MMA detailed in the literature was paresthesia of the lower lip and chin. It has been suggested that age at the time of surgery and addition of a genioplasty increase the risk of facial paresthesia, and a large degree of advancement further increases the risk in older patients [68,69]. In the STAR cohort (n = 126), the rates of major complication requiring device explanation, reposition, or replacement were 4% at 4 years and 9.5% at 5 years, indicating that the reoperations after UAS may occur more often during the late time frame. The STAR cohort also suggested that the majority of minor complications after UAS were gradually resolved. Notably, Withrow et al. evaluated the impact of age on safety of UAS and found no significant difference between younger and older cohorts in complication rates [70]. Current evidence suggests that both MMA and UAS appear to be safe approaches in OSA treatment, and compared to MMA, treating OSA with UAS may lead to fewer complications for older patients.

Clinical Relevance
In patients with moderate to severe OSA and failure of CPAP treatment, a portion of them could qualify for both MMA and UAS therapy. Current evidence shows that MMA may have superior efficacy in OSA treatment. However, MMA is a more invasive intervention, exposing patients to longer recovery time and higher risk of postoperative complications. Overnight admission to the intensive care unit is required for OSA patients following MMA surgery, and the length of hospitalization after MMA reported previously ranged from <2 days to 5-8 days [69]. Additionally, MMA surgery often involves timeconsuming preoperative and/or postoperative orthodontic work. One notable potential problem with MMA has been the accompanying alteration in facial appearance; however, most patients undergoing MMA for OSA view the change in facial appearance as neutral or even positive [30,32,46]. In comparison to MMA, UAS surgery is less invasive and more patient-friendly and does not require extended recovery. The majority of patients are discharged the same day or one day after UAS surgery [71]. In addition to the information regarding treatment efficacy and safety, the cost of treatment options is important in assisting decision-making in OSA treatment. It has been indicated that UAS is cost-effective, with a lifetime incremental cost effectiveness ratio (ICER) of USD 39,471 per qualityadjusted life year (QALY) in the United States healthcare system [72] and EUR 44,446 per fails to respond to MMA or for a patient with relapse of OSA after previously successful MMA [75]. It is interesting to note that in a recent study [76], Sarber et al. evaluated the efficacy of UAS therapy in 18 OSA patients who did not meet all FDA criteria for UAS and found promising treatment outcomes. They suggested that future studies must consider the expansion of current FDA criteria for UAS, particularly in BMI and AHI criteria. Thus, to optimize surgical outcomes, reduce rates of mortality and morbidity, and improve quality of life and other subjective outcomes, further investigation is essential to clarify indications of each therapy for OSA.
In addition to MMA and UAS, there are other evidence-based therapeutic options for OSA, which include behavioral strategies (e.g., weight loss), medical therapy (e.g., CPAP), other surgical options, and adjuvant therapy (e.g., pharyngeal muscle training) [77,78]. Of the non-CPAP therapies for OSA, more invasive procedures, such as MMA, are not well accepted. Oral appliances offer a non-invasive option for managing OSA, the most common of which are mandibular advancement devices (MADs). MADs modify the position of the jaw, the tongue, and other supporting structures of the upper airway, thereby increasing upper airway volume and preventing collapse of the upper airway [79]. MADs are recommended as a first-line therapy for mild-to-moderate OSA and for severe OSA after CPAP failure, intolerance, or refusal [80]. Growing evidence suggests that MADs could achieve favorable outcomes regardless of the severity of OSA [81,82].
In the era of precision medicine, the interconnected risk factors for OSA must be considered in order to achieve precision medicine in OSA [78]. The combined modern therapies for OSA must be adjusted continuously in respect to recent scientific research in order to deliver the best results for patients, emphasizing their quality of life in addition to medical care. Therefore, any of the therapies may either have an important role as monotherapy in the treatment of OSA or could be used in combination with the other therapies. The greater the complexity of a clinical case, the greater the need for multidisciplinary collaboration.

Limitations
There are several limitations of the present review. Firstly, because of the inherent difficulty of randomizing patients to different surgical interventions or sham surgery [83], except for one RCT and one quasi-experimental trial, all the included studies were cohort studies, the majority of which demonstrated fair quality according to the MINORS tool. Due to the lack of RCT and comparative studies of MMA and UAS for OSA, a meta-analysis cannot be performed to directly compare these two procedures. Additionally, meta-analyses were not conducted to separately assess overall effect sizes of MMA and UAS therapy on OSA, as mean and SD of the difference between pre-and postoperative measures were absent in majority of the selected studies. In this review, we performed separate analyses for MMA and UAS studies, combined with noticeable differences between the two cohorts in age and OSA severity, which prevented us from generating a solid conclusion on the comparison of these two procedures. Due to the fact that some patients may fall between two stools, comparison of the two procedures is important. Future studies should include quasiexperimental trials and comparative cohort studies comparing MMA and UAS to better clarify which modality is superior in OSA treatment. These studies can be part of a future large international consortium, which is more likely to generate solid conclusions. Secondly, due to the implemented inclusion criteria, which included the presence of both preoperative and postoperative PSG data, some well-conducted studies reporting on only subjective outcomes and/or safety were excluded for this study. Therefore, the present analysis of subjective outcomes and safety may not be entirely representative of the population undergoing MMA or UAS in the current literature. Lastly, our review is exclusively based on studies published in English, which can introduce a language bias [84].

Conclusions
The results presented in this review suggest that both MMA and UAS are effective and generally safe surgical treatment modalities for patients with moderate-to-severe OSA.