Systematic Review of The Effect of Upper Airway Surgery Based on DISE Findings In Adults With Obstructive Sleep Apnea

The objective of this review is to evaluate the effectiveness of upper airway surgery in adults with OSA veried on Drug Induced Sedation Endoscopy (DISE) and evaluated by change in AHI with minimum 3 month´s follow-up. combination of upper 3 polysomnography home sleep apnea of 40 2019. All surgeries in upper airways including soft tissue of the retropharyngeal space, velum, tonsils and base of tongue were included. Surgeries on cartilage and bone as septoplasty, turbinoplasty, mandibular advancement surgery, epiglottoplasty and tracheostomy were included plus hypoglossal nerve stimulation implant. case reports and studies reporting treatment mandibular advancement Surgeries targeting other anatomical sites than upper airways with a known reduction in AHI as bariatric surgery were also excluded. Studies without pre-operative DISE were excluded. Publications in other language than English were excluded. and oropharynx, four studies addressing the base of tongue (BOT) and three studies in multi-level surgery. Velum and oropharynx surgery led to an AHI-decrease of 11.86, 95% CI (10.21; 13.51) event per hour. ESS was reduced 7.01 (5.99; 8.04). In BOT surgery AHI was reduced 19.31 (17.81;20.81) events/hour and ESS decreased with 7.03 (6.44; 7.63). Multilevel surgery reduced AHI with 28.65 (24.60, 32.69) events/hour and ESS with 8.55 (6.73; 10.38). or coblation group. The randomization was done in the operating room. One surgeon performed all robotic treatments, and another surgeon performed all coblation treatments. Clinical ndings were not reported; however, inclusion and exclusion parameters include clinical ndings. Previous tonsillar surgery, limited mouth opening, and severe comorbidities were exclusion parameters. Seventy-seven patients were included and randomized to the study. Four patients were excluded due to perioperative complications of anesthesia. Three patients (two in TORS group and 1 in coblation group) were lost to follow-up. Follow-up time was 6 months. Post-operative AHI, ESS and snoring VAS were reported and compared to preoperative data. Operation time, time to oral diet in days and the length of analgesia needed postoperative were also reported. The study reports a statistically signicant reduction in mean AHI with a 36% reduction in TORS group (19.00 (15.84; 22.16)) and 37.8% reduction in coblation group (16.90 (14.33; 19.47)). A statistically signicant reduction in ESS was also reported with 33.8% reduction in the TORS group (8.00 (6.94; 9.06) and 31.5% in the coblation group (7.80 (6.76; 8.84]). No statistically signicant differences were seen between the two groups in any outcome parameters, however coblation was associated with quicker recovery and lower complication risks. of the implant and thereby used as a control group. Primary outcomes were AHI and oxygen desaturation index (ODI) by PSG and secondary outcomes such as ESS and disease-specic quality of life assessed by the Functional Outcomes of Sleep Questionnaire (FOSQ) were reported. Patient baseline data and comorbidities, previous surgery such as uvulopalatopharyngoplasty and blood pressure were reported. Follow-up at two months and six months were performed but not stated. The implants were activated one month postoperatively. Surgical response was dened as Sher’s criteria at 12 month´s follow-up. Furthermore, surgical response was dened as at least 25% reduction in ODI and reported. Two patients were lost to follow-up and considered non-responders. signicant in mean AHI of 16.70 20.21. Surgical responders based upon AHI 83 of 126 patients (66%) upon ODI 126 patients Mean ESS and mean FOSQ were also statistically signicantly AHI beginning randomization AHI events events events AHI events with TORS in a prospective, non-randomized trial with historical controls, a study with 75 adult patients with moderate to severe OSA with AHI > 20, and who had failed CPAP. This study did not exclude patients with any prior airway surgery such as tonsillectomy or UPPP. The study did not have any maximum AHI or BMI. DISE was performed in all patients and patients with retro-lingual collapse underwent TORS. Patients with simultaneous retropalatal collapse were offered UPPP, if not performed previously. A historical cohort of previous UPPP surgery was used as comparison. All patients undergoing TORS and UPPP had both procedures as single-step surgery, TORS was performed rst. Postoperative follow-up including PSG and ESS evaluation was done at a minimum of 3 months. Detailed follow-up time was not reported. Surgical response and surgical success rates were dened as Sher's criteria like other studies and reported. Patient baseline data were not reported. Preoperative and postoperative AHI, ESS and SpO2 nadir were also reported. Group 1 were 45 patients with no prior surgery and received TORS and UPPP surgery. Group 2 were 31 patients with prior surgery (tonsillectomy or UPPP). 4.77;


Introduction
Obstructive sleep apnea (OSA) is a common condition among adults worldwide 34% in men aged 30-70 years and 17% in women aged 30-70 years [1]. The negative effects of OSA such as increased risk of cardiac and metabolic disease are well known-documented [2]. Continuous positive airway pressure (CPAP) is considered rst line treatment in preventing upper airway collapse, improving sleep, and reducing daytime sleepiness***. However, two systematic literature reviews found signi cant low level of compliance reported as non-adherence levels of 34.3% to CPAP therapy [3] and as high as 29% to 83% [4]. Hence alternatives for CPAP and lifestyle changes are explored. Various treatments are used from lifestyle changes through sleep position treatment, oral mandibular device, bariatric surgery to different types of upper airway surgery to relieve symptoms of OSA or to increase compliance to CPAP.
Drug induced sedation endoscopy (DISE), originally described in 1991 by Croft and Pringle [5], is an interesting examination in the clinical evaluation of the individual anatomical abnormalities before upper airway surgery to reduce OSA. DISE has been validated to be relevant prior any sleep surgery [6][7][8]. This review investigated the evidence of different types of upper airway surgeries to relieve OSA symptoms when DISE was included in the preoperative evaluation in the planning of the possible surgical approaches.
A preliminary search of PROSPERO, MEDLINE, the Cochrane Database of Systematic Reviews and the JBI Database of Systematic Reviews and Implementation Reports was conducted and no current or underway systematic reviews on the topic were identi ed. A systematic review regarding hypoglossal nerve stimulation was identi ed [9]. Two systematic reviews regarding transoral robotic surgery (TORS) was identi ed [10,11]. However, this review emphasizes on DISE as part of preoperative work up and it involves sleep related surgical principles of the upper airway with follow-up period of at least 3 months.
A Cochrane review on the effect of upper airway surgery for OSA measured by change in AHI from 2005 found a limited number of trials assessing diverse surgical techniques. The effects reported across the trials were inconsistent. The evidence from the small studies did not support the use of surgery in people with mild to moderate daytime sleepiness associated with sleep apnea [12].
Based on a protocol from 2017, an updated Cochrane Review on surgical treatment of obstructive sleep apnea is being constructed but it has not been published. This review seems to have a focus on change in AHI and in ESS but does not incorporate DISE prior to surgery [13].
The purpose of this review was to evaluate the effectiveness of upper airway surgery in adults with OSA veri ed on DISE and evaluated by change in AHI with a minimum of 3 months follow-up.

Review question(s)
The question of this review was: How is the effect of different types of upper airway surgery measured by change in AHI in adults with obstructive sleep apnea when Drug-induced Sedation Endoscopy is used for selecting the surgery? Inclusion criteria Participants Inclusion criteria for this review are adult patients older than 18 years with diagnosed OSA either by polysomnography or by home sleep test. Follow-up time is 3 months or longer. Surgery is performed after DISE and the surgery should address the obstruction con rmed by DISE.
Exclusion criteria are case reports, reviews reporting no data and studies with less than 40 participants or studies reporting treatment with oral appliance therapy (OAT) / mandibular advancement devices (MAD), position trainer or bariatric surgery.

Outcomes
Primary outcomes in this review are changes in apnea-hypopnea index (AHI), de ned by the number of apnea and hypopnea events per hour of sleep, before and minimum 3 months after surgery to evaluate surgical success and surgical response or success de ned by Sher's criteria [14] de ned as the reduction of AHI > 50% and to a value of less than 20 events per hour. Epworth Sleepiness scale (ESS) is used for measuring Patient related outcome.
Secondary outcomes such as oxygen desaturation index (ODI), nadir peripheral oxygen saturation (SpO2), snoring scales, social and functional questionnaires, and other outcomes are described in the results.

Types of studies
This review considers both experimental and quasi-experimental study designs including randomized controlled trials, non-randomized controlled trials, before and after studies and interrupted time-series studies. In addition, analytical observational studies including prospective and retrospective cohort studies, casecontrol studies and analytical cross-sectional studies are considered for inclusion. The systematic review includes English written randomized controlled trials, prospective and retrospective case-control studies, and cohort studies on one or a combination of surgeries on the upper airways in adults diagnosed with OSA and obstruction veri ed by DISE before surgery. Sample size was minimum 40 participants. AHI should be reported prior to and minimum 3 months after surgery by polysomnography or home sleep apnea test. Publications from January 1 st . 2000 to December 11 th . 2019 were included. Surgery of the upper airways, including soft tissue of the retropharyngeal space, velum, tonsils, epiglottis, and base of tongue are included. Septoplasty and turbinoplasty are also included as well as the novel implant techniques with hypoglossal nerve stimulation. Maxillomandibular advancement surgery and hyoid suspension surgery are also included in the review of the literature. This review excluded descriptive observational study designs including case series, individual case reports and descriptive cross-sectional studies.

Methods
The systematic review is conducted in accordance with the Joanna Briggs Institute methodology for systematic reviews of effectiveness evidence [15,16] Search strategy The electronic search strategy was designed and conducted by an experienced reference librarian based on input from the study investigators.
An initial limited search of MEDLINE and CINAHL was undertaken to identify articles on the topic. The text words contained in the titles and abstracts of relevant articles, and the index terms used to describe the articles were used to develop a full search strategy for Cochrane, MEDLINE, CINAHL and Embase. (see Appendix #). The search strategy, including all identi ed keywords and index terms, will be adapted for each included information source. The reference list of all studies selected for critical appraisal was screened for additional studies.

Study selection
Following the search, all identi ed citations were collated and uploaded into Rayyan and duplicates were removed [17]. Titles and abstracts was then screened by two independent reviewers for assessment against the inclusion criteria for the review.
Potentially relevant studies were retrieved in full and their citation details imported into the Joanna Briggs Institute System for the Uni ed Management, Assessment and Review of Information (JBI SUMARI) (Joanna Briggs Institute, Adelaide, Australia) [18] The full text of selected citations was assessed in detail against the inclusion criteria by two independent reviewers. Reasons for exclusion of full text studies that did not meet the inclusion criteria were noted and reported in the systematic review. Any disagreements that arise between the reviewers at each stage of the study selection process were resolved through discussion, or with a third reviewer. The results of the search were reported in full in the nal systematic review and presented in a Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) ow diagram [19].

Assessment of methodological quality
Eligible studies were critically appraised by two independent reviewers at the study level. Authors of papers was contacted to request missing or additional data for clari cation, where required. Any disagreements that arise was resolved through discussion, or with a third reviewer. Following critical appraisal, studies that do not meet a certain quality threshold were excluded. The included studies are reported in narrative form and in a table.

Data extraction
Data extracted from studies was included in the review by two independent reviewers using the standardized data extraction tool. The data extracted included speci c details about the populations, study methods, interventions, and outcomes of signi cance to the review objective.

Data synthesis
The meta-analysis was performed in a statistical software Cran-R-4.0.3 using the package 'metafor'. The estimates for change from pre to post-operative effect and their 95% con dence intervals were calculated using xed effect model with inverse of the variance as weights. Since majority of the studies have not reported the results from which standard deviation for the mean change in AHI from pre to post can be calculated, zero correlation between the pre and post measurements were assumed and hence the reported 95% con dence intervals could be broader than their actual width.
The data was divided into three groups, each speci c for the level of obstruction addressed, either velum/oropharynx (soft palate), base of tongue or multilevel surgery.

Soft palate surgery
Soft palate surgery aims to reduce and reconstruct the part of the soft palate which often collapses during obstructive sleep apnea. A procedure initially known as classical uvulopalatopharyngoplasty (UPPP), rst described by Fujita [19]. It involves the removal of both tonsils and the posterior soft palate and uvula, and the closure of the tonsillar pillars has evolved. Different techniques such as modi ed UPPP which involves removal of additional mucosa and submucosal adipose tissue [20], expansion sphincter pharyngoplasty (ESPL) rst described by Pang and Woodson [21] which includes the transection of the palatopharyngeal muscle, rotating the muscle superior-laterally to mobilize it and then suture it to the pterygomandibular raphe through a mucosal tunnel to create a broader widening of the pharyngeal space. Further techniques such as barbed wire stitch reposition [22], the use of radiofrequency ablation of the soft palate [23] or laser assisted uvulopalatoplasty [24] are also well known and used to treat retropharyngeal collapse of the upper airway. The use of implants to stiffen the soft palate is also used in different settings. A thorough description of these techniques is not within the scope of this paper. However, it is noticed that that some techniques are used in an out-patient clinical setup, others as full operations in general anesthesia and some techniques occurs in multi-level surgery, whereas some only occurs as single-step surgery.
Furthermore, some procedures are offered to patients with severe OSA where AHI is high, others in mild or moderate OSA. Clinical features such as comorbidity, BMI, tonsil size, the positioning of the soft palate, the narrowing of the pharyngeal wall, neck circumference, the collapses of upper airway observed during DISE are used, however not all available studies present these ndings, nor do many studies follow a uniform standardized approach. This review found three studies investigating the effect of soft palate surgery eligible for this systematic review.

Base of tongue
Besides the treatments offered to address the velopharyngeal and the lateral wall of the oropharynx collapses, a variety of procedures devised to address the hypopharynx and the base of the tongue have been investigated. Procedures such as genioglossus advancement [25], radiofrequency tissue ablation of the tongue base [26], suturing of the tongue base and hyoid suspension are described [27], however the studies are often case reports or retrospective studies with few subjects and a limited follow-up time. The introduction of transoral robotic surgery (TORS) and the advancements in tongue base resections initially in cancer surgery has evolved into also using TORS in sleep surgery [28]. Furthermore, the introduction of the hypoglossal nerve stimulation termed upper airway stimulation (UAS) is interesting [29]. The insertion of a unilateral stimulation lead attached to the branch of the hypoglossal nerve, thereby producing tongue protrusion synchronous with expiration during sleep. The impulse generator is implanted beneath the clavicle and a tunneled breathing sensing lead is placed between the external and intercostal muscles. One could advocate that UAS is a multilevel treatment modality, since postoperative DISE ndings indicates improvements in the retropalatal obstruction as well. However, the treatment is indicated in OSA with a primarily retro-lingual obstruction, this systematic review has categorized this treatment to the treatment of base of tongue.
This review found four studies eligible for a systematic review targeting treatment on the base of the tongue: Multi-level surgery As described previously, OSA is often associated with multi-level collapse of the upper airways, hence multilevel surgery is also necessary and introduced as a treatment opportunity. The bene ts and disadvantages of single-step and multi-step surgical intervention is beyond the scope of this review. Some surgeons advocate that nasal surgery is performed prior to any direct approach towards the soft palate, lateral wall of the oropharynx or the base of tongue is considered. This review found three eligible studies investigating multi-level surgery as a single-step procedure.
This review did not consider nasal surgery as a part of multi-level approach.

Results
Ten studies were included and divided in three subcategories covering the site of obstruction, which were retropharyngeal collapse, base of tongue and multilevel surgery. In total we included 697 adult patients older than 18 years with an average age of 50,4 years. Average BMI was 28,98. A summary of studies are presented in table 1.
Most studies included patients with moderate to severe OSA, however the de nition was either AHI above 15 or AHI above 20. Velum and oropharynx surgery led to a decrease in AHI of 11 Plaza et. al. presented a prospective non-comparative multi-center study of patients suffering from OSA selected to ESP as stand-alone procedure at three University Hospitals. Patients diagnosed with OSA with preoperative AHI 5-70, age between 18 and 70, BMI < 35, and predominant lateral wall collapse on oropharynx during DISE were included. If nasal procedures were necessary, this was performed before, at a different stage than pharyngeal surgery. Seventyve patients were included. Baseline characteristics including clinical ndings such as tonsil size, Friedman tongue position (FTP), lingual tonsil hypertrophy (LTH). Both preoperative and postoperative data as AHI and ESS were reported. Follow-up was at least 12 months with clinical exam, including PSG and ESS. Surgical success was reported using Sher criteria described as: Cured = AHI < 5 and ESS < 10 or AHI reduction >50%. The need for CPAP after surgery was de ned as AHI < 15 and ESS < 10 and reduction >50%.
The study showed a statistically signi cant reduction in mean AHI 13 Patient selection was based upon DISE ndings such as retropalatal circumferential narrowing above DISE grade II (more than 75% narrowing) and narrowed oropharynx due to lateral pharyngeal collapse of the bulky redundant soft tissue around the posterior pillars. The patients did not have obstruction at the tongue base or epiglottis. Clinical ndings such as tonsil sizes, Friedman tongue position scale or neck circumferential were not reported, nor were patient baseline characteristics described. Patients were carefully selected by inclusion criteria based upon previous feasibility studies. The rst 46 consecutive patients after 12 months were randomized 1:1 and 23 patients were selected to a therapy-withdrawal group with deactivation of the implant and thereby used as a control group.
Primary outcomes were AHI and oxygen desaturation index (ODI) by PSG and secondary outcomes such as ESS and disease-speci c quality of life assessed by the Functional Outcomes of Sleep Questionnaire (FOSQ) were reported. Patient baseline data and comorbidities, previous surgery such as uvulopalatopharyngoplasty and blood pressure were reported. Follow-up at two months and six months were performed but not stated. The implants were activated one month postoperatively. Surgical response was de ned as Sher's criteria at 12 month´s follow-up. Furthermore, surgical response was de ned as at least 25% reduction in ODI and reported. Two patients were lost to follow-up and considered non-responders.
The study showed a statistically signi cant reduction in mean AHI of 16.70 (13.19; 20.21. Surgical responders based upon AHI were 83 of 126 patients (66%) and based upon ODI was 94 of 126 patients (75%). Mean ESS and mean FOSQ were also statistically signi cantly improved.
The therapy-withdrawal study showed a signi cant difference between the withdrawal group and the therapy-maintenance group with respect to the change in AHI from the beginning of the randomization to the assessment 1 week later. In the withdrawal group, baseline AHI was 30,1 events per hour, before deactivation 7,6 events per hour and after deactivation 25,8 events per hour, whereas the maintenance group reported baseline AHI 31,3 events per hour, at randomization 7,2 events per hour and 1 week later 8,9 events per hour. Similar effects in ODI were reported in the study.
Adverse events were reported. Two patients had severe device-related adverse events which required repositioning and xation of the neurostimulator. Other mild adverse events are also described in the study.
The study concludes that UAS showed a reduction in the severity of obstructive sleep apnea, and the adverse-event pro le was acceptable. In 2018 5-year follow-up in the same group was published by Woodson et. al. (REF). Although only 71 patients had PSG at 5-year follow-up, the results were unchanged.
Since the study is based upon the same group, the data is not included in this review.
Huntley et al. 2017 [35] Upper airway stimulation for treatment of OSA was also investigated by   )) and nadir O2 saturations in both groups, although the nadir O2 saturation in group 1 was better than in group 2. Surgical success was illustrated in the study's gure 1 and reported around 90% of patients in both groups with no statistically signi cant differences between the two groups. The gure also illustrates the percentage of patients with AHI<15, AHI<10 and AHI<5 at each institution.
The study concluded that UAS is an effective alternative treatment in a subset of patients with moderate to severe OSA with base of tongue collapse, who were unable to comply with CPAP therapy. The study showed at statistically signi cant reduction in mean AHI 17.40 (12.38; 22.42) at 6-month follow-up and the improvement was unchanged at 12 month´s follow-up. Improvements in ODI, SpO2 nadir, mean ESS 6.30 (4.54; 8.06) and mean FOSQ was also statistically signi cant. At the end of the study, 30% of the patients were cured. 68% of the patients were considered as surgical responders. Patients lost to follow-up was regarded as non-responders.
Comparison of surgical responders and non-responders showed no differences in baseline data such as age, BMI, neck circumference, AHI, ODI, FOSQ or ESS. The results were consistent with previous reported outcome by the STAR trial. The study concludes that UAS is a safe and feasible therapy capable to reduce OSA severity and prove patient reported quality of life outcomes, and therapy adherence was high after 12 months of follow-up.
Multi-level surgery: showing promising results for TORS [39]. Although no correlations between the excised volume of the tongue base, the lymphatic/muscular ratio of removed tissue or total thickness of tongue base and AHI changes could be shown, the overall reduction in AHI in patients with moderate to severe OSA was reported and shown to be statistically signi cant 24.87 (19.31; 30.43).
The study consists of 51 patients who underwent single-step multi-level TORS including temporary tracheostomy, tonsillectomy, expansion sphincter pharyngoplasty and septoturbinoplasty and base of tongue excision. Patient characteristics such as age, gender, BMI, and preoperative and postoperative AHI were reported. Furthermore, mean tongue volume (cc), lymphatic/soft tissue ratio, total thickness of removed base of tongue and success rate were reported.
Surgical response was de ned accordingly to Sher's criteria.
No correlations were observed in postoperative AHI and anatomical measures of removed base of tongue tissue. Although the study did not show any differences in the primary measured outcomes, the observed reduction of AHI was found to be statistically signi cant and the rate of surgical success reported, concur with previous studies published.
El-Anwar et. al. 2018 [40] El-Anwar published a prospective non-randomized study with 40 patients scheduled for multilevel surgery. Surgery was multi-level surgery including hyoid suspension, tonsillectomy and pharyngeal suspension sutures described by El-Ahl and El-Anwar [41]. Group A consisted of patients with nasal obstruction and underwent inferior turbinate reduction), while group B did not. Postoperative follow-up was between 6 months and 14 months postoperatively. Patient baseline characteristics such as age, gender and BMI were reported. Primary outcome was AHI. Preoperative and postoperative data such as ESS and nadir oxygen saturation was also reported.
The preoperative and postoperative data was not tested within the groups. Between the groups, preoperative AHI was signi cantly higher in patients with nasal obstruction (group A) compared to patients with no nasal obstruction (group B) and the reduction in AHI was greater in group A than in group B ( Sundaram et. al reviewed surgical treatment compared to CPAP treatment or non-surgical appliance therapy and did not nd evidence that surgical intervention could improve OSA [12]. In comparison, this review indicates, that surgical intervention improves both clinical and subjective OSA parameters such as AHI and ESS, when treatment is based upon DISE ndings. However surgical intervention is limited to the procedures presented. This comparison also advocates that surgical intervention is suitable for patients non-compliant to CPAP treatment.
Each study varies in preoperative and postoperative evaluation of each patient and inclusion criteria, exclusion criteria, surgical outcome, and patient followup time. Hence, heterogeneity exists between the nal studies presented here. However, common parameters in the studies are also present. Surgical outcomes are primarily AHI and ESS, while terms such as surgical responders and surgical success is described using Sher's criteria, providing a suitable comparison between the studies. DISE ndings are essential, furthermore all studies use age, BMI, and the severity of OSA to screen patients before surgery.
Some studies include clinical observations related to the Friedman staging system [43]. Hence, the studies are comparable although risk of bias is present. Furthermore, all studies present in detail minor and severe adverse effects to surgery.
Weaver et al. showed that PSG parameters such as AHI does not correlate well to patient related outcome [44]. In sleep related surgery, patient related outcome is essential to address the bene ts of the surgical intervention. All studies except Cammaroto present ESS data. Thaler et al. report ESS, however only 31 patients of 75 patients completed ESS measurements, hence the data was not included in the meta-analysis. Some studies include additional patient related outcome measurement tools such as the SAQLI or FOSQ or merely VAS snoring scale. This review found an overall improvement in ESS. Lee et. al reported that ESS should not be used as a screening method in a population with sleep disorders since the method is unreliable [45]. Moffa et. al recently published a systematic review and found ve different approaches to address retropalatal airway collapse [47]. Although primary outcomes were improved after surgery, heterogeneity was high and therefore level of evidence for these procedures are low and further randomized studies are needed.
For base of tongue, two different approaches were addressed. Either the reduction and removal of lingual tissue with the use of coblation or TORS, or the use of hypoglossal nerve stimulation and thereby protrusion of the entire tongue synchronous with expiration during sleep. The study concludes that UAS showed a reduction in the severity of obstructive sleep apnea, and the adverse-event pro le was acceptable.       searchhistory.docx