Transoral Robotic Surgery for Obstructive Sleep Apnea

Abstract

Objective

To perform a systematic review of the international biomedical literature evaluating the effectiveness, complications, and safety of transoral robotic surgery (TORS) for the treatment of obstructive sleep apnea (OSA).

Data Sources

PubMed/MEDLINE, Embase, and EMB Reviews databases were searched up to November 27, 2015.

Review Methods

Two authors systematically and independently searched for articles on TORS for the treatment of OSA in adults that reported either outcomes for the apnea-hypopnea index (AHI), lowest oxygen saturation percentage (LSAT) or changes in the Epworth Sleepiness Scale (ESS), and/or rates and types of complications associated with the operation.

Results

In total, 181 records were identified and 16 articles met inclusion criteria. Transoral robotic surgery was almost always combined with other sleep surgery procedures. The summary estimate of the decrease in AHI using TORS as part of a multilevel surgical approach was 24.0 (95% confidence interval [CI], 22.1-25.8; P < .001, I² = 99%). The summary estimate of a decrease in ESS score was 7.2 (95% CI, 6.6-7.7; P < .001, I² = 99%) and of the overall surgical “success” (defined as AHI <20 and 50% reduction) was 48.2% (95% CI, 38.8%-57.7%; P < .001, I² = 99%). Three large studies reported on their total complication rates with an average of 22.3% (range, 20.5%-24.7%).

Conclusions

The initial results for the use of TORS as part of a multilevel surgical approach for OSA are promising for select patients. However, the cost and morbidity may be greater than with other techniques offsetting its advantages in visualization and precision. More prospective studies are needed to determine the optimal role of this tool.

Keywords

obstructive sleep apnea transoral robotic surgery TORS sleep-disordered breathing base of tongue

Obstructive sleep apnea (OSA) remains a serious public health epidemic with mild cases (apnea-hypopnea index [AHI] ≥5) occurring in an estimated 3% to 28% and moderate cases (AHI ≥15) occurring in 1% to 14% of the American populace.¹ An estimated 5% of the Western global populace is undiagnosed.¹ An association exists between OSA and hypertension,¹ stroke,^2,3 heart failure,³ coronary heart disease,³ cognitive dysfunction,^4,5 motor vehicle accidents,⁶ and death.² Undiagnosed OSA causes an estimated $3.4 billion in US health care costs annually.⁷

Continuous positive airway pressure (CPAP) is the standard of care for medical treatment of OSA with demonstrated improvement in sleep quality, cognitive impairment, and hypertension.⁸ However, for those patients who fail CPAP due to nontolerance and/or noncompliance, surgical treatment provides a logical yet debatable option for certain patients.⁹ The radiofrequency base-of-tongue (RFBOT) reduction and coblation (submucosal minimally invasive lingual excision [SMILE])¹⁰ employ the technique of resecting a portion of the base of tongue (BOT) and provide 2 potentially effective surgical options. Transoral robotic surgery (TORS) provides another novel technique for surgical resection of the BOT and lingual tonsil. Transoral robotic surgery was initially used for resection of base-of-tongue neoplasms in 2006, and it provided improved magnification, surgical precision, avoidance of critical vessels and nerves, hemostatic control, and precise handling of tissues.¹¹

In 2010, Vicini et al¹² introduced TORS for BOT and lingual tonsil resection for OSA, with good functional results of reasonable pain, good swallowing, and improved quality of life that were considerably encouraging. Since then, several studies have been published looking at TORS for OSA. To our knowledge, this is the first systematic review of TORS as a treatment for OSA and fills a gap in knowledge for this particular area of research. Our objective was to perform a systematic review of the international biomedical literature evaluating the effectiveness, complications, and safety of TORS for the treatment of OSA and then to provide recommendations for future research.

Materials and Methods

Search Strategy

A review of the literature identified articles related to TORS in patients with OSA. We used the Preferred Reporting Items for Systematic Reviews and Meta-Analyses checklist and recommendation for this systematic review.¹³ The search was initiated on April 21, 2015, and was completed November 27, 2015. We systematically and independently searched PubMed, Embase, and the EBM Reviews, which reviews the Cochrane DSR, ACP Journal Club, DARE, CCRCT, CMD, HTA, and NHSEED. We conducted searches for articles containing TORS and OSA. For PubMed, the search strategy used the terms ((transoral[tiab] AND robot*[tiab]) OR (robot*[tiab] AND surg*[tiab]) OR TORS[tiab] OR Robotic Surgical Procedures [mh] OR robotics[mh] OR “da vinci”[tiab]) AND (Sleep Apnea Syndromes[mh] OR Sleep Apnea* OR (“Sleep-Disordered” AND Breathing [tiab]) OR hypopnea OR hypopnoea OR sleep apnoea* OR sleep[tiab]). For Embase, the search was (“robotic surgical procedure”/exp OR (transoral AND robot*) OR tors OR (robot* AND surg*) OR “robotics”/exp OR “da vinci”) AND (“sleep disordered breathing”/exp OR “sleep apnea” OR “sleep apneas” OR (“sleep-disordered” AND breathing) OR “sleep apnoea” OR “sleep apnoeas” OR hypopnea OR hypopnoea) AND [embase]/lim. For EBM, we used the following search terms: (sleep or apnoea or apnea or hypopnea or hypopnoea).mp. (transoral or (robot* and surg*) or tors).mp.

Inclusion and Exclusion Criteria

We independently reviewed the selected articles that studied TORS for OSA. We identified changes in AHI, lowest pulse oxygen saturation (LSAT), Epworth Sleepiness Scale (ESS), and the rate of successful surgery of TORS for OSA and/or those articles that reported rates and types of complications. In addition, all study designs met inclusion criteria except case reports and small cases series with fewer than 5 patients. We studied articles addressing both single and multilevel operations. Only English-language articles with a title, abstract, and full text were included. After an initial review of the titles and abstracts, the reviewers culled the list to 16 full publications. Two investigators (G.A.J. and E.T.C.) reviewed the articles for inclusion and exclusion criteria.

Study Extraction, Categorization, and Analysis

Articles were analyzed for variables including study size, intervention performed, sleep study and operative data, complications, and outcomes. Data on surgical success (defined as AHI ≤20 and reduction of AHI by 50%) and cure (identified as AHI ≤5) were collected. Two investigators (G.A.J. and E.T.C.) both reviewed the articles independently and reached a consensus on the data and the outcomes of the various studies. Articles were then split into 2 categories: (1) those that addressed sleep study data and the rate of surgical success and cure and (2) those that discussed complications. Articles were then subcategorized by the type of operations performed.

Level and Quality of Evidence Assessments

Both the level of evidence and the quality of each study were evaluated. The 2011 Oxford Centre for Evidence-Based Medicine Levels of Evidence table was used to assess the level of evidence.¹⁴ Level 1 symbolized a systematic review of randomized trials or n-of-1 trials; level 2 represented a randomized trial or observational study with dramatic effect; level 3 denoted a nonrandomized controlled cohort/follow-up study; level 4 characterized a case series, case-control, or historically controlled study; and level 5 described a mechanism-based reasoning study.¹⁴

A checklist from a recent article¹⁵ was adapted and altered as needed to assess the quality of evidence. A 14-item checklist was developed that took into account study design and demographics, intervention performed, sleep study and operative data, complication rate, and type and overall surgical success ( Table 1 ). Studies that had 11 to 14 points were deemed high quality; 6 to 10, moderate quality; and 0 to 5, poor quality.¹⁵

Table 1.

Criteria for Assessing Methodological Quality of Studies.

1. Clear description of research design (eg, observational, prospective, or retrospective)
2. Well-defined description of consecutive samples
3. Well-formulated inclusion and exclusion criteria
4. Demographics of population described
5. Appropriate control or comparison group
6. Description of procedures conducted
7. Preoperative and postoperative AHI, ESS, LSAT
8. Surgical success discussed
9. Follow-up rate and length described
10. Total number of complications addressed
11. Type of complications discussed
12. Operating room times addressed
13. Resected volume addressed
14. Length of hospital stay addressed

Abbreviations: AHI, apnea-hypopnea index; ESS, Epworth Sleepiness Scale; LSAT, lowest oxygen saturation percentage.

Statistical Methods

Statistical analysis was performed with statistical software (STATA 8.2; StataCorp LP, College Station, Texas). Random-effects modeling (standard error estimate = inverse of the sample size) was used to calculate summary effect measures with corresponding 95% confidence intervals (CIs), and forest plots were generated. The I² statistic was used to assess between-study heterogeneity. In some cases, the between-study heterogeneity was not significant, and therefore the random-effect modeling estimate equaled the fixed-effects estimate. Possible publication bias was assessed using graphical funnel plot analysis. A P value of less than .05 was considered significant.

Results

Search Results for TORS for OSA

In total, 181 records were identified using the search strategy detailed above. A flowchart ( Figure 1 ) illustrates how the selected studies were included. Articles were excluded for numerous reasons: duplicates (n = 45), unrelated to TORS for OSA (n = 106), presentation abstract (n = 5), review article (n = 2), no sleep and/or complications data (n = 4), description of surgical technique (n = 2), and cadaveric study (n = 1).

Figure 1.

Flowchart of article selection. OSA, obstructive sleep apnea; TORS, transoral robotic surgery.

Vicini et al¹² published the first article in 2010, and since then, there have been 15 other articles published on TORS in OSA.^10,16
-29 Overall, there has been an increase in the number of publications on TORS for OSA, with 1 each in 2010 and 2011, 2 each in 2012 and 2013, and 8 in 2014 but only 2 in 2015. The number of patients included in these studies ranged from 6 to 285.

Quality and Level of the Evidence

All 16 articles were evaluated for the quality and level of evidence presented ( Table 2 ). Ten studies were of moderate quality^{12,16
-20,24,25,27,28} and 6^{10,21
-23,26,29} were of high quality. There were no randomized control studies identified. There were 2 prospective nonrandomized studies,^20,25 1 retrospective study with historical controls,¹⁰ and 13 retrospective case series. All were level 4 with regard to the 2011 Oxford Centre for Evidence-Based Medicine Levels of Evidence table.

Table 2.

Study Design, Quality, and Level of Evidence.

Study	Study Design, Surgeon/Institution, Study Patient Inclusion Dates	Quality and Level of Evidence
Eesa et al¹⁶	Retrospective review, multiple/single, December 2011 to February 2014	9/14, 4
Friedman et al¹⁰	Historical cohort study compared with 2 matched cohorts of patients, RFBOT + z-plasty and SMILE + z-plasty, single/single, October 2010 to June 2011	11/14, 4
Glazer et al¹⁷	Retrospective cohort study, single/single, December 2010 to November 2013	10/14, 4
Hoff et al¹⁸	Retrospective case series, single/single, December 2010 to October 2013	8/14, 4
Hoff et al¹⁹	Multicenter, single-arm, retrospective case series, 3/3, January 2010 to October 2013	10/14, 4
Lee et al²⁰	Prospective nonrandomized trial using historical controls, unspecified/single, unspecified	10/14, 4
Lin et al²¹	Retrospective case series, single/single, June 2010 to May 2012	12/14, 4
Lin et al²²	Retrospective case series, unspecified/single, June 2010 to May 2014	13/14, 4
Muderris et al²³	Retrospective review of patient records, single/single, September 2012 to December 2013	11/14, 4
Richmon et al²⁴	Retrospective cohort, unspecified/single, September 2009 to February 2013	7/14, 4
Thaler et al²⁵	Prospective, nonrandomized with historical cohorts, unspecified/single, 2011 to 2014	9/14, 4
Toh et al²⁶	Retrospective, multiple/single, 2011 to 2013	12/14, 4
Vicini et al¹²	Retrospective, multiple/single, unspecified	9/14, 4
Vicini et al²⁷	Retrospective, multiple/single, unspecified	10/14, 4
Vicini et al²⁸	Retrospective, multiple/single, May 2008 to December 2011	10/14, 4
Vicini et al²⁹	Retrospective cohort, multiple/7, June 2008 to November 2012	12/14, 4

Abbreviations: RFBOT, radiofrequency base of tongue; SMILE, submucosal minimally invasive lingual excision.

Patient Demographics, Interventions and Operative/Hospital Data

Various patient demographic data, including mean age and body mass index (BMI), were collected. In addition, descriptions of the various interventions performed, operative time, resected volumes, and length of hospital stay were reviewed ( Table 3 ). In the 16 studies, the grand mean age reported was 49.39 years (Richmon et al²⁴ was not included since this study reported age as a median and included patients with cancer). The range was 43.8 to 54.6 years. Preoperative BMI was reported in 11 studies.^{10,17
-23,25,26,29} The grand mean preoperative BMI was 30.44 kg/m² with a range of 26.5 to 34.5 kg/m². Most studies had more male than female patients, with an average of 74.25% (range, 25%-90%; Muderris et al²³ was all female and Richmon et al²⁴ had patients with cancer; both were not included).

Table 3.

Inclusion Criteria, Surgical Operation, Demographics, and Hospital Data.

Publication	N, Study Inclusion Criteria	Age, Mean (SD), y	BMI, Mean (SD), kg/m²	Mean Follow-up	Intervention	Operative Time, Mean (SD), min	Resected Volume, Mean (SD)	Hospital Stay, Mean (SD), d
Eesa et al¹⁶	78, NR	48	NR	20 (7.12) mo	23 TORS TBR alone, 55 TORS + EPI (78 [100%]) or TORS + TRA (64 [82%])	39 (11)	12.35 mL BOT and epiglottic	8.5 (2.63)
Friedman et al¹⁰	27, SS, AHI ≥15, FTP III or IV, PO ≥6 mo, CPAP F, ≥18 yo	43.8 (9.2)	32.3 (3.3)	NR	27 TORS with TBR and ZPA	1-h increase in operative time	2.28 (0.43) g	1.6 (0.7), significantly greater than RFBOT
Glazer et al¹⁷	166, AHI ≥20, CPAP F, preoperative PSG, ESS, DISE	54.6 (12.3)	29 (4.6)	2 wk and then 3 mo	166 TORS LT. Mean number of procedures was 3.5. Most common: EPI (79), TON (70), GLO (65)	75.5	7.9 mL	1.5 total, 1 SICU, 2 hospital
Hoff et al¹⁸	121, moderate to severe OSA, CPAP F, TBH-DISE, TBE, PO ≥3 mo	54.5	Preoperative 28.5, postoperative 27.5, P < .001	3 mo	121 TORS TBR. Mean number of procedures was 3.5. Completed nasal, palatal, and pharyngeal surgery. EPI in 45%	NR	7.92 mL	NR
Hoff et al¹⁹	285, OSA, AO, TBH, D, ≥18 yo	51.5 (11.1)	30.5 (4.9)	NR	293 TORS LT or TORS LT + TBR, UPPP (66.2%), lateral pharyngoplasty (12.3%), TON (32.1%), ZPA (9.6%)	86.7 (faster than operating time of nonrobotic surgery by 30-90 min)	8.3 mL (lingual tonsil + BOT)	1.8
Lee et al²⁰	20, ≥18 yo, indications for surgical management of BOT, SIC	45.7	32.6	4-19 mo	20 TORS + UPPP	NR	NR	NR
Lin et al²¹	12, exclusion, PP or lack of postoperative PSG	46.5 (13.3)	34.5 (7.3), postoperative 33.5 (6.7)	4-6 mo	12 TORS TBR	76.3 (20)	27.6 (16.9) mL	2.7 (0.8)
Lin et al²²	39, PO ≥4 mo, completed clinical information, standard PSG	46.5 (13.2)	32.9 (7), postoperative 32.4 (77.3)	4-6 mo	11 TORS TBR, 2 TBR + UPPP, 7 TBR + EPI, 19 TBR, UPPP + EPI	RS 24.2 (7.7), TORS 59.5 (20.6)	22.2 (11.7) mL	3.5 (1.5)
Muderris et al²³	6, histologically proven TBH, and D, SS, CO, GS, severe OSA	49 (range, 43-55)	27.3 (range, 24-30)	8.5 mo (range, 4-14)	6 TORS LT, 1 patient had additional EPI	87 total, (RS 43, TORS 44)	4.45 (range, 3.9-5.3) cm	5.3 (range, 4-9 )
Richmon et al²⁴	91 (7 were for OSA), NR	Median 59 years (range, 27-88)	NR	NR	13 TORS LT	NR	NR	1.51 (range, 1-5)
Thaler et al²⁵	75, >18 yo, CPAP F, AHI >20, peoperative PSG, MRI of neck, DISE, PO ≥3 mo	49.7	32.3	NR	45 no prior surgery TORS TBR + UPPP; those with prior TON or UPPP underwent TORS TBR and UPPP if needed	NR		NR
Toh et al²⁶	20, CPAP F, PO ≥6 mo, DISE, PP	47.1	26.5	8.2 (3.2) mo	20 TORS TBR, EPI, UPPP	RS 20.9 (16), TORS 26.8 (7.3)	9.15 (4.1) mL	4
Vicini et al¹²	10, PO ≥3 mo, AHI ≥20 or ESS ≥11, CPAP F, TBH, AE, no CI	51.8 (range, 35-64)	Unchanged	6 (3-10) mo	10 TORS TBR, TRA, various multistep surgeries	RS 41.4 (19.7), TORS 45.5 (16.5)	NR	11.4 (8-16)
Vicini et al²⁷	20, PO ≥10 mo, AHI ≥20 or ESS ≥11, CPAP F, TBH, AE, no CI	50.15 (range, 20-65)	NR	≥10 mo	20 TORS TBR, TRA, various multistep surgeries	RS 31.2 (18.2), TORS 42.5 (15.2)	NR	NR
Vicini et al²⁸	24, PO ≥6 mo	Group A: 54.17 (range, 32-65); group B: 49.58 (range, 29-64)	NR	≥6 mo	Two groups of 12 patients: group A, TORS TBR + UPPP; group B, TORS TBR + expansion sphincter pharyngoplasty	Group A: TORS 35 (20-50), RS 105.33 (43-185)Group B: TORS 36.68 (20-63), RS 89.3 (44-150)	7-51 cc	Group A: 7.25 (range, 5-10)Group B: 8.33 (range, 6-13)
Vicini et al²⁹	243, DISE, PSG, ESS, PO ≥3 mo	50 (12)	28.53 (3.87)	≥3 mo	243 TORS TBR, EPI (14%) nasal (59%, either during DISE or TORS), UPPP (60%), TON (60%), ZPA, expansion sphincter pharyngoplasty	Total 98.2 (49.9), overall robotic time 55.7 (17.1), RS 19.5 (10.5), TORS TBR 26.5 (9.2), EPI 10.4 (4.9)	BOT 10.29 (4.12) mL, tonsils 7.4 mL (3.3)	3.5 (3.18)

Abbreviations: AE, adequate exposure; AHI, apnea-hypopnea index; AO, airway obstruction; BOT, base of tongue; CI, contraindications; CO, cough; CPAP F, continuous positive airway pressure failure; D, dysphagia; DISE, drug-induced sleep endoscopy; EPI, epiglottectomy; ESS, Epworth Sleepiness Scale; FTP, Friedman tongue position; GLO, nonrobotic midline glossectomy; GS, globus sensation; LT, lingual tonsillectomy; MRI, magnetic resonance imaging; NR, not reported; OSA, obstructive sleep apnea; PO, postoperative follow-up with PSG; PP, prior nasal palatal, tongue base led to exclusion; PSG, polysomnography; RFBOT, radiofrequency base of tongue; RS, robot setup time; SIC, signed informed consent; SICU, surgical intensive care unit; SS, significant snoring; TBE, adequate tongue base exposure; TBH, tongue base or lingual tonsillar hypertrophy; TBR, tongue base reduction; TON, tonsillectomy; TORS, transoral robotic surgery; TRA, tracheostomy; UPPP, uvulopalatopharyngoplasty; yo, years old; ZPA, z-palatoplasty.

There was wide diversity among the 16 studies with regard to the types of procedures performed. All studies reported the use of TORS in base-of-tongue resection or lingual tonsillectomy with or without simultaneously performing additional sleep surgery. Only 1 study looked simply at TORS BOT resection,²¹ another explored TORS BOT resection and uvulopharyngopalatoplasty (UPPP),²⁰ and a different article looked at TORS BOT resection with partial epiglottectomy and UPPP.²⁶ An additional interesting study was a comparison of TORS BOT resection and UPPP vs TORS BOT resection and extension sphincter pharyngoplasty.²⁸ Friedman et al¹⁰ published a retrospective cases series with historical cohorts comparing TORS with BOT resection and z-palatoplasty compared with RFBOT with z-palatoplasty and SMILE with z-palatoplasty. All other studies^{12,16-19,22-25,27,28} consisted of TORS lingual tonsillectomy or BOT resection and multiple other operations, with epiglottectomy, UPPP, and tonsillectomy being the most common. In addition, 4 of the reviewed studies provided data on patients who received a tracheostomy, all of whom were European.^12,16,27,29

The various studies revealed a degree of heterogeneity in the amount of time needed to perform TORS BOT resection or lingual tonsillectomy, with some studies including epiglottectomy and other procedures. Thus, we split our analysis into those reporting total operative times in which the specific amount of time of the actual TORS procedure as part of a multilevel operation was vague vs those reporting TORS docking and operative times. Eleven studies^{12,16,17,19,21
-23,26
-29} reported their results for an average operative period of 80.68 minutes with a range of 39 to 132.5 minutes. Seven studies^{12,22,23,26
-29} broke their times into docking and operative times. The average docking time was 41.6 minutes with a range of 20.9 to 97.5 minutes, and operative time was 44.15 minutes with a range of 26.8 and 55.7 minutes. Finally, Friedman et al¹⁰ noted an average of a 1-hour increase in operative time compared with SMILE and RFBOT.

There was wide diversity in the reporting of the resected volume of tissue, with some including just BOT and others BOT and epiglottic tissues ( Table 2 ). In addition, multiple studies failed to report any description of the resected tissue. Furthermore, 1 study reported the volume of resected tissue in grams, while another reported the thickness of the tissue in centimeters. For studies that reported the mean volume of resected tongue tissue, the summary estimate using random-effects modeling was 12.7 cm³ (k = 9 studies; 95% CI, 11.3-14.1 cm³, P < .001, I² = 99%).

Hospital Stay

Twelve studies reported the average length of hospital stay ( Table 3 ). The summary estimate of hospital stay was 4.9 days (k = 12 studies; 95% CI, 3.6-6.2; P < .001, I² = 99%). In addition, there was a wide range of locations for the hospital stay, with some patients placed in the intensive care unit (ICU)^{17,18,21,22,26} and others extubated directly after surgery^10,25 and placed in medical/surgery wards.

Nasogastric Tube Requirement and Time to Oral Feeding

Nasogastric tube requirement (NGT) requirement was discussed in 3 studies,^17,19,23 with 1 to 2 patients per study requiring NGT placement due to dysphagia. Vicini et al^12,27 published initially that they used NGT in their first 3 patients but realized quickly that they were not routinely necessary. These data were insufficient for meta-analysis.

In addition, multiple studies published on the time to oral feedings and return to normal diet. One study discussed return to oral diet as early as 5 hours,²³ although most published around 24 hours.^12,16,26,29 In addition, 3 articles^10,22,26 discussed the time to normal diet, which ranged from 2 to 4 weeks. These data were also insufficient for meta-analysis.

Outcomes for AHI, LSAT, and ESS and Surgical Success

Twelve studies reported at least 2 of the preoperative and postoperative AHI, ESS, and LSAT ( Table 4 ). Almost all cases reported a significant improvement in AHI, ESS, and LSAT. Using random-effects modeling, the summary estimate of the decrease in AHI using TORS as part of a multilevel approach was 24.0 (k = 13 studies; 95% CI, 22.1-25.8; P < .001, I² = 99%) ( Figure 2 ). The change in LSAT was an increase of 5.3% (k = 9 studies; 95% CI, 4.0-6.7%; P < .001, I² = 99%). The summary estimate of a decrease in ESS was 7.2 points (k = 12 studies; 95% CI, 6.6-7.7; P < .001, I² = 99%) ( Figure 3 ). Finally, the estimate of surgical “success” was 48.2% (k = 9 studies; 95% CI, 38.8-57.7%; P < .001, I² = 99%) ( Figure 4 ).

Table 4.

Outcomes for AHI, LSAT, and ESS and Surgical Success.

		AHI		ESS		LSAT
Publication	Intervention	Pre-operative	Post-operative	Pre-operative	Post-operative	Pre-operative	Post-operative	Surgical Success
Friedman et al¹⁰	27 TORS with TBR and ZPA	54.6 (21.8)	18.6 (9.1), P < .001	14.4 (4.5)	5.4 (3.1), P < .001	78.5 (7.4)	86.5 (6.3), P < .001,	SS 66.7%
Hoff et al¹⁸	121 TORS TBR. Mean number of procedures was 3.5. Completed nasal, palatal, and pharyngeal surgery. EPI in 45%	42.7	22.2, P < .001	NR	NR	NR	NR	SS 51.2%, SC 14%
Lee et al²⁰	20 TORS + UPPP	55.6	24.1, P < .001	13.4	5.9, P = .003	75.8	81.7, P = .013	SS 45%, SR 65%
Lin et al²¹	12 TORS TBR	43.9 (41.1)	17.6 (16.2), P = .007	13.7 (5.2)	6.4 (4.5), P < .001	83.3% (5.5%)	84%, (6.4%)	SS 50%
Lin et al²²	11 TORS TBR, 2 TBR + UPPP, 7 TBR + EPI, 19 TBR, UPPP + EPI	43.9 (32.3)	21.9 (23.5), P < .001	15.6 (5.4)	5.7 (4.3), P < .001	81.6 (8.1)	83.4 (7.3)	SS 53.8%
Muderris et al²³	6 TORS LT, 1 patient had additional EPI	27.5 (range, 22.1-35.3)	6.3 (range, 0.5-18.0), P < .05	14.1 (range, 9-21)	7.1 (range, 4-15), P < .05	NR	NR	NR
Thaler et al²⁵	45 no prior surgery TORS TBR + UPPP; those with prior TON or UPPP underwent TORS TBR and UPPP if needed	57.5 (23.9)	31.4 (28.6), P < .0001	Obtained in only 31 patients; 12.8 (6.5)	5.8 (4.9), P < .0001	78.8 (9.6)	83.1 (7.3), P < .002	SS 45%
Toh et al²⁶	20 TORS TBR, EPI, UPPP	41.3 (22.1)	13.5 (17.1),P < .001	13 (2.8)	5.6 (4.4), P < .001	72.9 (19.3)	84.5 (7.1), P = .009	SC 35%, SS 55%
Vicini et al¹²	10 TORS TBR, TRA, various multistep surgeries	38.3 (23.5)	20.6 (17.3), P < .05	12.4 (3.5)	6.9 (2.8), P < .05	72.7	80.2	SC 20%, SS 20%
Vicini et al²⁷	20 TORS TBR, TRA, various multistep surgeries	36.3 (21.1)	16.4 (15.2)	12.6 (4.4)	7.7 (3.3)	77.3 (9.7)	81.9 (7.3)	SC 60%
Vicini et al²⁸	Two groups of 12 patients: group A, TORS TBR + UPPP; group B, TORS TBR + expansion sphincter pharyngoplasty	Group B: 38.5 (14.3), group A: 38.4 (19.7)	Group B: 9.9 (8.6), group A: 19.9 (14.1)	Group A: 13.75 (4), group B: 12 (4.9)	7.6 (4.4), 4.4 (4.1)	NR	NR	NR
Vicini et al²⁹	243 TORS TBR, EPI (14%) nasal (59%, either during DISE or TORS), UPPP (60%), TON (60%), ZPA, expansion sphincter pharyngoplasty	43.21 (22.60) (range, 5.8-112.0)	17.54 (16.48), P < .001	12.5 (5.78) (range, 0-24)	5.78 (3.60), P < .001	79.50% (9.09%) (range, 48%-95%)	83.83% (6.42%) (P < .001)	SC 22.9%, SS 30.9%

Abbreviations: AHI, apnea-hypopnea index; DISE, drug-induced sleep endoscopy; EPI, epiglottectomy; ESS, Epworth Sleepiness Scale; LAST, oxygen saturation; LT, lingual tonsillectomy; NR, not reported; ROC, receiver operating characteristic; SC, surgical cure; SR, surgical response; SS, surgical success; TBR, tongue base reduction; TON, tonsillectomy; TORS, transoral robotic surgery; TRA, tracheostomy; UPPP, uvulopalatopharyngoplasty; ZPA, z-palatoplasty.

Figure 2.

Summary estimate of the decrease in apnea-hypopnea index (AHI) using random-effects modeling.

Figure 3.

Summary estimate of the decrease in Epworth Sleepiness Scale (ESS) using random-effects modeling.

Figure 4.

Summary estimate of surgical success using random-effects modeling.

Complications

Three large studies reported on their total complication rates with an average of 22.3% and range of 20.5% to 24.7%.^17,19,29 A review of the complications included in 14 articles can be studied in Table 5 . Complications found included bleeding requiring cauterization in the operating room, pulmonary embolism/deep venous thrombosis, aspiration, dysphagia, pneumonia, dehydration, uncontrolled pain, globus sensation, minor bleeding, and dysgeusia. These data were too diverse to allow for a formal meta-analysis.

Table 5.

Complications.

Study	Intervention	Complications Requiring Operative Intervention	Complications Not Requiring Operative Intervention
Eesa et al¹⁶	23 TORS TBR alone, 55 TORS + EPI (78 [100%]) or TORS + TRA (64 [82%])	NA	1 patient with persistent dysgeusia of >6 mo
Glazer et al¹⁷	166 TORS LT. Mean number of procedures was 3.5. Most common: EPI (79), TON (70), GLO (65)	7 (4.2%) major bleeds requiring intervention	43 in 41 patients (24.7%), 5 (3%) minor bleeds, 2 (1.2%) DVT/PE, 1 (0.6%) aspiration, 16 (9.6%) dehydration/pain control, 8 (4.8%) globus sensation, 2 lip burn (1.2%), 1 pharyngeal laceration during intubation (0.6%)
Hoff et al¹⁹	293 TORS LT or TORS LT + TBR, UPPP (66.2%), lateral pharyngoplasty (12.3%), TON (32.1%), ZPA (9.6%)	12 (4.1%) bleeding requiring cauterization	59 patients (20.7%) with a total of 77 complications, 15 dysphagia, 14 dehydration, 6 pneumonia, 6 hypoxemia, 2 reintubation, 3 pain, 1 gastrointestinal complications, 1 cardiac arrhythmias, 15 other, including deep vein thrombosis, urinary retention, tongue swelling, fever, aspiration pneumonitis
Lee et al²⁰	20 TORS + UPPP	1 patient (4.2%) bleeding required cauterization	20.8% dysphagia, 12.5% transient dysgeusia, 8.3% transient globus sensation
Lin et al²¹	12 TORS TBR	NA	4 total (33%), 1 dysphagia (8%), 3 dysgeusia (25%)
Lin et al²²	11 TORS TBR, 2 TBR + UPPP, 7 TBR + EPI, 19 TBR, UPPP + EPI	NA	3 (7.7%) dysphagia, 3 (7.7%) dysgeusia at >1 year
Muderris et al²³	6 TORS LT, 1 patient had additional EPI	NA	Oropharyngeal synechiae in 1 patient (who had previous tonsillectomy and also required the epiglottoplasty) and weight loss in 1 patient
Thaler et al²⁵	45 no prior surgery TORS TBR + UPPP; those with prior TON or UPPP underwent TORS TBR and UPPP if needed	4 (5.7%) postoperative bleeding requiring cautery in OR	8 total (17%), 4 (5.7%) dehydration/pain, 2 intubations postoperatively (2.7%)
Toh et al²⁶	20 TORS TBR, EPI, UPPP	NA	All had tongue numbness and soreness, 7 dysgeusia (55%), 1 dysphagia (5%), 1 tonsillar bleed (5%)
Vicini et al¹²	10 TORS TBR, TRA, various multistep surgeries	NA	4: 1 case of pharyngeal edema, 3 small bleeds not requiring cautery
Vicini et al²⁷	20 TORS TBR, TRA, various multistep surgeries	NA	6 complications: 2 cases of subcutaneous emphysema, 1 of pharyngeal edema, 3 cases of small bleeds
Vicini et al²⁸	Two groups of 12 patients: group A, TORS TBR + UPPP; group B, TORS TBR + expansion sphincter pharyngoplasty	NA	3 cases: 1 pharyngeal edema, 1 subcutaneous emphysema, 1 pneumonia
Vicini et al²⁹	243 TORS TBR, EPI (14%) nasal (59%, either during DISE or TORS), UPPP (60%), TON (60%), ZPA, expansion sphincter pharyngoplasty	1.7% late bleeding managed in OR	Complications in 20.5% of patients (0.4% intraoperative bleeding, 2.9% postoperative self-limited bleeding, 0.4% transient globus sensation, 0.4% transient pharyngeal edema, 14.2% transient hypogeusia less than 8 mo in all cases, 0.4% pharyngeal stenosis)

Abbreviations: DISE, drug-induced sleep endoscopy; DVT/PE, deep venous thrombosis/pulmonary embolism; EPI, epiglottectomy; GLO, nonrobotic midline glossectomy; LT, lingual tonsillectomy; NA, not applicable; OR, operating room; TBR, tongue base reduction; TON, tonsillectomy; TORS, transoral robotic surgery; TRA, tracheostomy; UPPP, uvulopalatopharyngoplasty; ZPA, z-palatoplasty.

Discussion

Transoral robotic surgery for OSA is a relatively new tool for BOT resection and lingual tonsillectomy first described by Vicini et al¹² in 2010. The advantages described of TORS for BOT resection and lingual tonsillectomy compared with non-TORS methods of tongue reduction are reported to be improvement in magnification, surgical precision, avoidance of critical vessels and nerves, hemostatic control, and precise handling of tissues.¹¹ However, these advantages come at the expense of both longer total operative times of about an hour and greater disposable costs of as much as $500 per case.¹⁰ In addition, there is the increase in cost associated with purchasing the robot itself and the maintenance expenses, but the results of sleep and complications data collected have been encouraging for TORS for OSA as part of a multilevel surgery.

The AHI, ESS, and LSAT data were reviewed in this study with an approximate summary estimate decrease in AHI of 24 and ESS of 7.2, as well as improvement of LSAT of 5.3%. Almost all studies showed a significant difference in AHI, ESS, and LSAT. The results from the Friedman et al¹⁰ study were particularly encouraging, with an AHI and ESS reduction of approximately 36 and 9, respectively, both of which were significantly greater than RFBOT and SMILE.

As most of these studies contain multilevel operations, it is difficult to ascertain whether this change in AHI, ESS, and LSAT data is specifically due to the use of TORS for BOT resection or lingual tonsillectomy. Lin et al²¹ was the only study to look at TORS BOT resection alone. The results were promising, with a significant mean (SD) preoperative and postoperative reduction in AHI and ESS of 43.9 (41.1) to 17.6 (16.2) (P = .007) and 13.7 (5.2) to 6.4 (4.5) (P < .001), respectively. Interestingly, these results are similar to other studies in the review, in which patients underwent multilevel surgery. As this was a retrospective study of only 12 patients and they were not able to find any demographic or clinical data that correlated with surgical success, it is hard to make any judgments on whether TORS BOT resection could be completed as a standalone procedure, but this should become an interesting area of future research. In addition, although its effect in isolation appears promising, it is difficult currently to pinpoint its effect in multilevel surgery. Future studies will need to ascertain its value in multilevel surgery relative to cost and complexity.

With regard to surgical success, the summary estimate was 48.2%. In the Friedman et al¹⁰ study, TORS for OSA was found to have a significant surgical success rate compared with RFBOT but not SMILE. A couple articles^18,22 looked at factors that predicted surgical success. Hoff et al¹⁸ observed the effect of BMI on surgical success and found that patients with a BMI ≤30 kg/m² vs those with a BMI ≥30 kg/m² had a difference in success of 69.4% and 41.7% (P = .004), respectively. There was not a significant difference in the cure rate. Lin et al²² found that BMI ≤30 kg/m², AHI ≤60, and absence of lateral pharyngeal wall collapse had a significant difference in surgical success, with excellent rates of 88.2%, 67.9%, and 66.7%, respectively. In the same study, Friedman tongue position; tonsil size; BOT collapse; concomitant UPPP, epiglottectomy, or UPPP and epiglottectomy; or amount of tissue removed did not correlate with surgical success. Lin et al³⁰ published a systematic review of multilevel surgery for OSA and found a surgical success rate of 66.4% in all patients and 59.2% in those who did not undergo maxillomandibular advancement. It appears that with the selection of proper patients, the use of TORS may lead to improved results.

In addition, multiple studies discussed the results of the improvement in AHI, LSAT, and ESS in relationship to the amount of tissue removed.^10,16,22 Eesa et al¹⁶ found that the removal of between 10 and 20 cc of tissues resulted in greater surgical success than <10 cc or >20 cc. Friedman et al¹⁰ and Lin et al²² did not find a correlation. Future randomized studies should be conducted in which comparisons of more vs less aggressive tongue base resection are compared.

The 3 largest studies^17,19,29 had an average complication rate of 22.3%, which is comparable to other studies. A recent systematic review of glossectomy for OSA found a complication rate of 16.7%,³¹ while Woodson and Fujita³² studying lingualplasty found a 27% complication rate and Hou et al³³ reviewing UPPP and midline glossectomy found a 24% complication rate. Glazer et al¹⁷ looked as factors that correlated with complications. They found that American Society of Anesthesiologists (ASA) score (P = .003) and number of adjunct procedures (P = .004) correlated with postoperative complications, while age, BMI, AHI, LSAT, number of comorbidities, and anticoagulation therapy were dissociated. Since there was no standard in reporting of complications between all of the studies and because of the heterogeneity of procedures completed during the operation, it is difficult to observe whether a particular procedure increased the rates of certain complications.

An important observation from this review was that dysphagia requiring NGT was incredibly rare and reported in only 4 total cases. Critically, the use of NGT was not required routinely. A previous study³⁴ on the use of TORS for partial pharyngectomy and partial glossectomy for neoplastic disease found higher rates of dysphagia than in the non-TORS operations (19.5% vs 8.0%, P < .001). Other studies^32,33,35
-39 using various techniques for glossectomy for OSA have had varied rates of dysphagia ranging from 0% to 33%. However, in these studies and in the articles included in this review, there was a wide variety in the methods and definitions of dysphagia. Eesa et al¹⁶ studied the effect of TORS for OSA on swallowing and found an insignificant difference in the preoperative and postoperative MD Anderson Dysphagia Inventory questionnaire. In addition, there was no correlation between the amount of tissue resected and the risk of aspiration.¹⁶ It is encouraging that multiple studies demonstrated a return to oral feeding in 24 hours.^{12,16,23,26,29} However, the time to normal diet was at least 2 to 4 weeks,^10,22,26 and Friedman et al¹⁰ found this to be significantly longer than SMILE and RFBOT.

A few articles^10,21,27 disagreed about the requirement of an overnight ICU stay and length of stay in the hospital postoperatively. Friedman et al¹⁰ advocated for extubation directly after the operation with a mean (SD) hospital stay of 1.6 (0.7) days. Vicini et al²⁷ completed routine tracheostomy on all patients with a mean (SD) hospital stay of 9.5 (2.9) days. Lin et al²¹ took the middle ground by keeping their patients intubated overnight in the ICU with a mean (SD) hospital stay of 2.7 (0.8) days. Due to the number of TORS for OSA studies completed with no difference in complications with or without tracheostomy, this suggests that tracheostomy may not be a necessary adjunct procedure.^{17,19,20
-23,25,26} However, the difference between maintaining a patient intubated in the ICU for 1 night vs extubating after the surgery is still up to debate. The amount of tissue removed in the Friedman et al¹⁰ study was reported as 2.28 (0.43) g, which is approximately 2.28 mL since the density of the tissue removed is about 1 g/mL. In comparison, the amount of tissue removed in the Lin et al²¹ study was 27.6 (16.9) mL. Most of the studies^17,18,22,26 in which larger amounts of tissue were removed than in Friedman et al¹⁰ advocated an overnight stay in the ICU. Unfortunately, Thaler et al²⁵ reported on direct extubation of patients after surgery but did not report on the amount of tissue resected. In future studies, whether removal of a larger amount of tissue and extubation directly after surgery lead to an increase in airway complications after surgery should be analyzed.

The heterogeneity of the data in these studies due to variation in reporting and its use as a tool in multilevel surgery makes it difficult to compare studies of TORS for OSA and note its true effect in isolation. However, due to the increase of the use of TORS for OSA, we felt that a review of the studies would point out areas of improvement in future research. For future publications, more homogeneous reporting of data is necessary so that better comparisons to other studies using TORS and/or other surgical operations for OSA can be made.

For example, there should be a uniform method of identifying surgical response (AHI ≥50% reduction), success (AHI ≥50% reduction and AHI ≤20), and cure (AHI ≤5). In addition, consistent methods of describing the volume of BOT resected are needed; we would recommend solely reporting the volume of tongue resected and not including the tissue from the epiglottectomy. There need to be more prospective studies of TORS alone as a tool for BOT resection or lingual tonsillectomy in comparison to multilevel procedures using TORS for part of the procedure. Due to the heterogeneity of the data, it is currently difficult to analyze the differences between glossectomy and lingual tonsillectomy with regard to surgical success and complication rates. In addition, future studies should look at Friedman tongue position and lingual tonsil grade prior to TORS BOT resection or lingual tonsillectomy, as well as the effect of epiglottectomy in isolated studies, since their effect on surgical success or complications remains unclear. Differences in the volume of tongue resected should be reviewed. Although retrospective case series are critical, to demonstrate the true benefit, complications, and financial cost, more studies such as Friedman et al¹⁰ need to be published. Finally, a cost benefit analysis of TORS for BOT resection should be published as Friedman et al¹⁰ demonstrated that it is more costly and time-consuming but has better efficacy than SMILE and RFBOT.

There is significant risk of bias in this systematic review. It was difficult to get a true number of cases that had actually been completed since many of the authors were collaborating with each other. They published patients from the same institution and time periods using different inclusion criteria for different studies. As such, there may have been overlap of patients, in the sleep and complications data, and surgical success between multiple studies. In addition, many of the articles were retrospective case series without control arms. Finally, many of these studies used multilevel operations, and as such, it is hard to judge if BOT resection alone had an influence on the surgical results.

Conclusion

The initial data from these 16 studies are encouraging, with a statistically significant reduction in AHI and ESS and an increase in LSAT. Good rates of surgical success and cure were reported in certain patients, especially those with a BMI ≤30 kg/m². The complication rate is comparable to previous studies addressing surgical operations for OSA. Two critical findings from this systematic review are that tracheostomy and NGT are not necessary after surgery. New studies are necessary to review the effects of the amount of tongue volume resected, the need for an ICU stay, and the cost-benefit of TORS for OSA vs other methods of glossectomy.

Author Contributions

Grant A. Justin, conception and design of the work, acquisition and analysis of data, drafting and revising the work, final approval of publication, agreement to be accountable for all aspects of the work; Edward T. Chang, conception and design of the work, acquisition and analysis of data, revising the work, final approval of publication, agreement to be accountable for all aspects of the work; Macario Camacho, conception and design of the work, interpretation of the data, revising the work, final approval of publication, agreement to be accountable for all aspects of the work; Scott E. Brietzke, conception and design of the work, interpretation of the data, revising the work, final approval of publication, agreement to be accountable for all aspects of the work.

Disclosures

Competing interests: None.

Sponsorships: None.

Funding source: None.

Disclaimer

The views expressed in this manuscript are those of the authors and do not reflect the official policy or position of the Department of the Army, Department of Defense, or the US Government.

Footnotes

Acknowledgements

The authors thank Rhonda Allard, MLS, for her extensive help with designing the literature search and retrieving various documents.

No sponsorships or competing interests have been disclosed for this article.

References

Young

Peppard

Gottlieb

. Epidemiology of obstructive sleep apnea: a population health perspective. Am J Respir Crit Care Med. 2002;165:1217-1239.

Yaggi

Concato

Kernan

. Obstructive sleep apnea as a risk factor for stroke and death. N Engl J Med. 2005;353:2034-2041.

Shahar

Whitney

Redline

. Sleep-disordered breathing and cardiovascular disease: cross-sectional results of the Sleep HeartHealth Study. Am J Respir Crit Care Med. 2001;163:19-25.

Lau

Choi

Lai

. Working memory impairment and its associated sleep-related respiratory parameters in children with obstructive sleep apnea. Sleep Med. 2015;16:1109-1115.

Daulatzai

. Evidence of neurodegeneration in obstructive sleep apnea: relationship between obstructive sleep apnea and cognitive dysfunction in the elderly. J Neurosci Res. 2015;93:1778-1794.

Basoglu

Tasbakan

. Elevated risk of sleepiness-related motor vehicle accidents in patients with obstructive sleep apnea syndrome: a case-control study. Traffic Inj Prev. 2014;15:470-476.

Kapur

Blough

Sandblom

. The medical cost of undiagnosed sleep apnea. Sleep. 1999;22:749-755.

Basner

. Continuous positive airway pressure for obstructive sleep apnea. N Engl J Med. 2007;356:1751-1758.

Kotecha

Hall

. Role of surgery in adult obstructive sleep apnoea. Sleep Med Rev. 2014;18:405-413.

10.

Friedman

Hamilton

Samuelson

. Transoral robotic glossectomy for the treatment of obstructive sleep apnea-hypopnea syndrome. Otolaryngol Head Neck Surg. 2012;146:854-862.

11.

O’Malley

Jr Weinstein

Snyder

. Transoral robotic surgery (TORS) for base of tongue neoplasms. Laryngoscope. 2006;116:1465-1472.

12.

Vicini

Dallan

Canzi

. Transoral robotic tongue base resection in obstructive sleep apnoea-hypopnoea syndrome: a preliminary report. ORL J Otorhinolaryngol Relat Spec. 2010;72:22-27.

13.

Moher

Liberati

Tetzlaff

; PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Int J Surg. 2010;8:336-341.

14.

OCEBM Levels of Evidence Working Group. The Oxford levels of evidence 2. http://www.cebm.net/index.aspx?o=5653. Accessed December 14, 2015.

15.

Kozin

Gulati

Kaplan

. Systematic review of outcomes following observational and operative endoscopic middle ear surgery. Laryngoscope. 2015;125:1205-1214.

16.

Eesa

Montevecchi

Hendawy

E, et al

. Swallowing outcome after TORS for sleep apnea: short- and long-term evaluation. Eur Arch Otorhinolaryngol. 2015;272:1537-1541.

17.

Glazer

Hoff

Spector

. Transoral robotic surgery for obstructive sleep apnea: perioperative management and postoperative complications. JAMA Otolaryngol Head Neck Surg. 2014;140:1207-1212.

18.

Hoff

D’Agostino

Thaler

. Transoral robotic surgery in benign diseases including obstructive sleep apnea: safety and feasibility. Laryngoscope. 2015;125:1249-1253.

19.

Hoff

Glazer

Spector

. Body mass index predicts success in patients undergoing transoral robotic surgery for obstructive sleep apnea. ORL J Otorhinolaryngol Relat Spec. 2014;76:266-272.

20.

Lee

Weinstein

O’Malley

Jr BW

. Transoral robot-assisted lingual tonsillectomy and uvulopalatopharyngoplasty for obstructive sleep apnea. Ann Otol Rhinol Laryngol. 2012;121:635-639.

21.

Lin

Rowley

Badr

. Transoral robotic surgery for treatment of obstructive sleep apnea-hypopnea syndrome. Laryngoscope. 2013;123:1811-1816.

22.

Lin

Rowley

Folbe

. Transoral robotic surgery for treatment of obstructive sleep apnea: factors predicting surgical response. Laryngoscope. 2015;125:1013-1020.

23.

Muderris

Sevil

Bercin

. Transoral robotic lingual tonsillectomy in adults: preliminary results. Acta Otolaryngol. 2015;135:64-69.

24.

Richmon

Feng

Yang

. Feasibility of rapid discharge after transoral robotic surgery of the oropharynx. Laryngoscope. 2014;124:2518-2525.

25.

Thaler

Rassekh

Lee

. Outcomes for multilevel surgery for sleep apnea: obstructive sleep apnea, transoral robotic surgery, and uvulopalatopharyngoplasty. Laryngoscope. 2016;126(1):266-269.

26.

Toh

Han

Tay

. Transoral robotic surgery for obstructive sleep apnea in Asian patients: a Singapore sleep center experience. JAMA Otolaryngol Head Neck Surg. 2014;140:624-629.

27.

Vicini

Dallan

Canzi

. Transoral robotic surgery of the tongue base in obstructive sleep apnea-hypopnea syndrome: anatomic considerations and clinical experience. Head Neck. 2012;34:15-22.

28.

Vicini

Montevecchi

Campanini

. Clinical outcomes and complications associated with TORS for OSAHS: a benchmark for evaluating an emerging surgical technology in a targeted application for benign disease. ORL J Otorhinolaryngol Relat Spec. 2014;76:63-69.

29.

Vicini

Montevecchi

Pang

. Combined transoral robotic tongue base surgery and palate surgery in obstructive sleep apnea-hypopnea syndrome: expansion sphincter pharyngoplasty versus uvulopalatopharyngoplasty. Head Neck. 2014;36:77-83.

30.

Lin

Friedman

Chang

. The efficacy of multilevel surgery of the upper airway in adults with obstructive sleep apnea/hypopnea syndrome. Laryngoscope. 2008;118:902-908.

31.

Murphey

Kandl

Nguyen

. The effect of glossectomy for obstructive sleep apnea: a systematic review and meta-analysis. Otolaryngol Head Neck Surg. 2015;153:334-342.

32.

Woodson

Fujita

. Clinical experience with lingualplasty as part of the treatment of severe obstructive sleep apnea. Otolaryngol Head Neck Surg. 1992;107:40-48.

33.

Hou

Yan

Wang

. Treatment of obstructive sleep apnea-hypopnea syndrome with combined uvulopalatopharyngoplasty and midline glossectomy: outcomes from a 5-year study. Respir Care. 2012;57:2104-2110.

34.

Chung

Rosenthal

Magnuson

. Transoral robotic surgery for oropharyngeal and tongue cancer in the United States. Laryngoscope. 2015;125:140-145.

35.

Fujita

Woodson

Clark

. Laser midline glossectomy as a treatment for obstructive sleep apnea. Laryngoscope. 1991;101:805-809.

36.

Mickelson

Rosenthal

. Midline glossectomy and epiglottidectomy for obstructive sleep apnea syndrome. Laryngoscope. 1997;107:614-619.

37.

Friedman

Soans

Gurpinar

. Evaluation of submucosal minimally invasive lingual excision technique for treatment of obstructive sleep apnea/hypopnea syndrome. Otolaryngol Head Neck Surg. 2008;139:378-385.

38.

Suh

. Evaluation of open midline glossectomy in the multilevel surgical management of obstructive sleep apnea syndrome. Otolaryngol Head Neck Surg. 2013;148:166-171.

39.

Gunawardena

Robinson

MacKay

. Submucosal lingualplasty for adult obstructive sleep apnea. Otolaryngol Head Neck Surg. 2013;148:157-165.