Upper Airway Stimulation for Obstructive Sleep Apnea: The Surgical Learning Curve

Transcription

1 The Laryngoscope VC 2015 The American Laryngological, Rhinological and Otological Society, Inc. Upper Airway Stimulation for Obstructive Sleep Apnea: The Surgical Learning Curve Alexander W. Murphey, MD; Andrew B. Baker, BS; Ryan J. Soose, MD; Tapan A. Padyha, MD; Shaun A. Nguyen, MD, MA; Christopher C. Xiao, BA; M. Boyd Gillespie, MD, MSc Objectives/Hypothesis: To determine the effect of surgeon experience with an upper airway stimulation (UAS) system on surgical time and complication rates. Study Design: Retrospective review. Methods: Surgical procedure times and complication rates observed in patients implanted at 22 study centers as part of a phase III, multicenter surgical trial of upper airway nerve stimulation therapy for obstructive sleep apnea were reviewed. Results: The study included 126 subjects who were predominantly male (83%), with a mean age of 54.5 years (range years), and the mean body mass index was There were an average of 5.7 (range ) surgical implants per site, with an average surgical time of hours (range hours). The surgical implant time decreased significantly with surgeon experience, from hours for a surgeon s first implant (n 5 22) to hours for the fifth implant (n 5 10, P 5.025). Surgical time was inversely correlated with the site implant number (rho , P <.001). Procedure-specific complications were uncommon and self-limited and did not decrease appreciably with increasing experience. Conclusions: Surgical time for implantation of the UAS system decreased significantly after the first five implants and then stabilized. The rate of surgical complications did not decrease with surgeon experience, although this may be attributable to the low overall rate of serious surgical complications and low number of implants at some centers. Key Words: Obstructive sleep apnea, upper airway stimulation, hypoglossal nerve, snoring. Level of Evidence: 4 Laryngoscope, 126: , 2016 INTRODUCTION Obstructive sleep apnea (OSA) is a common disorder characterized by narrowing and collapse of upper airway tissues leading to reductions and cessations of airflow during sleep. OSA is associated with numerous comorbidities including daytime somnolence, mood alterations, impaired cognition, and increased risk of cardiovascular events. 1 4 Continuous positive airway pressure (CPAP) is the gold standard therapy for patients with moderate to severe sleep apnea, defined as an apneahypopnea index (AHI) 15. CPAP uses air pressure as a From the Department of Otolaryngology Head and Neck Surgery (A.W.M., A.B.B., S.A.N., C.C.X., M.B.G.), Medical University of South Carolina, Charleston, South Carolina; Division of Sleep Surgery (R.J.S.), Department of Otolaryngology, University of Pittsburgh, Pittsburgh, Pennsylvania; Department of Otolaryngology Head and Neck Surgery (T.A.P.), University of South Florida, Tampa, Florida, U.S.A. Editor s Note: This Manuscript was accepted for publication July 2, Presented at the Triological Society 118th Annual Meeting at COSM, Boston, Massachusetts, U.S.A., April 24 25, Dr. Gillespie has received research support from Inspire Medical Systems. Dr. Soose has received research and grant support from Inspire Medical Systems and has consulted for Inspire Medical Systems and Philips-Respironics. The authors have no other funding, financial relationships, or conflicts of interest to disclose. Send correspondence to M. Boyd Gillespie, MD, Department of Otolaryngology Head and Neck Surgery, 135 Rutledge Avenue MSC 550, Charleston, SC gillesmb@musc.edu DOI: /lary pneumatic splint to prop open the upper airway and prevent tissue collapse. When used appropriately by motivated patients, CPAP effectively reduces airway collapse, improves quality of life, and reduces cardiovascular risks. 5,6 Despite the benefits, treatment is limited by poor adherence to therapy. Only 17% to 54% of patients are adherent to >4 hours use/night with significant reductions after 1 year of therapy. 7,8 Surgical therapy is an option for symptomatic patients who fail a trial of CPAP therapy. Sleep surgery traditionally has treated specific areas of collapse in the upper airway by tissue removal or repositioning. Many patients, however, avoid sleep surgery due to its morbidity, side effect profile, and variable success rates. Therefore, there has been a push to find alternative therapy options for patients with moderate to severe sleep apnea. Electrical stimulation of the upper airway for OSA was first reported in the early 1990s. 9,10 As opposed to altering upper airway anatomy, upper airway stimulation increases upper airway dilator muscle activity to maintain airway patency normally reduced during sleep. 11 Direct stimulation of the specific branches of the hypoglossal nerve increases genioglossus muscle activity and can provide multilevel enlargement and stabilization of the upper airway. 12 Early clinical studies showed that hypoglossal nerve stimulation (HNS) provides significant reduction in AHI, daytime somnolence, and sleep-related quality of life in patients with OSA

2 A recent multicenter, prospective, phase III trial with the implantable HNS (Stimulation Therapy for Apnea Reduction [STAR] trial; Inspire Medical Systems, Minneapolis, MN) found significant reductions in AHI, oxygen desaturation index, and sleep-related quality of life in 126 patients with moderate to severe sleep apnea. 16 As HNS systems transition from clinical trials to US Food and Drug Administration approved commercial implants, it is important to evaluate the technical aspects of the various implant procedures. Surgeon performance typically improves with increasing surgical experience correlating with faster procedure times and fewer adverse events, 17 but the point at which surgical proficiency is optimized varies from one procedure to another. The goal of the present study was to describe the surgical learning curve of the Inspire Upper Airway Stimulation (UAS) system as observed during the STAR trial to determine the effect of surgeon experience on surgical procedure time and complication rates. MATERIALS AND METHODS Study Design The STAR trial is a phase III, prospective, multicenter clinical trial to demonstrate the safety and efficacy of an implantable upper airway cranial nerve XII stimulator (Inspire II; Inspire Medical Systems, Minneapolis, MN) for adults with moderate to severe OSA who are nonadherent to positive pressure therapy. Participants who met specific polysomnographic, clinical, and anatomical inclusion criteria were eligible for device implantation. Details of the study design have been previously published. 16 The Medical University of South Carolina received research support from Inspire Medical Systems as a participating site in the STAR trial. The current project is not funded by Inspire. This project is an investigator-initiated effort conceived, designed, analyzed, and written solely by the authors. Inspire provided access to the aggregate data of the STAR trial to make this research project possible. This article focuses on the surgical issues of upper airway stimulation therapy including surgical procedural time and surgically related complications. Procedural Safety Assessment All adverse clinical events were recorded from initial enrollment through the most recent follow-up, and were adjudicated by an independent clinical events committee that consisted of two otolaryngologists, one sleep medicine physician, and one biostatistician. Specifically for this article the focus was procedural safety; therefore, adverse events were limited to those occurring intraoperatively or in the immediate postoperative period to hospital discharge. Adverse events were categorized as serious or nonserious, and whether they were specifically related to upper airway stimulation therapy or generic to surgical procedures. Serious adverse events were those resulting in death, permanent injury or illness, requiring invasive intervention to prevent death or injury, or requiring an additional inpatient hospitalization or prolongation of existing hospitalization. Nonserious adverse events included any event not requiring invasive intervention such as incision-related numbness or hypertrophic scar, incision site pain, intubation effects, headache, or other nonserious postoperative problems. 502 Postoperative Pain Assessment Postoperative pain was self-assessed by the patient using the visual analog scale (VAS) at implant and each subsequent study visit. Patients marked their pain level on a 10-cm line, with 0 mm indicating no pain, and 10 mm indicating intense pain. Postoperative pain medications were defined as any pain medication given to the patient between surgical implant and the first-month visit. Pain medications were classified into their various drug categories, and start and stop dates were calculated from the patient s medication record. Surgical Time Surgeries were performed at one of 22 sites participating in the trial. At each site, there was only one surgeon performing the procedure. All the surgeons were new to the implantation procedure and underwent the same training and cadaveric dissections prior to first implant. Surgical time was measured from the time of initial incision to the time of final wound closure, and was collected systematically throughout the course of the trial. Surgical implant number refers to the numerical order of implants performed by the surgeon at each site during the time course of the STAR trial. Statistical Analysis All analyses were performed with SPSS 22.0 (IBM Corp., Armonk, NY) and SigmaPlot 12.5 (Systat Software, San Jose, CA). Categorical variables are presented as frequency and percentage, and continuous variables are presented as mean 6 standard deviation, range, or median with interquartile range (IQR) in text and tables. All continuous variables were assessed for normality using the Kolmogorov-Smirnov test. If these variables were not normally distributed, descriptive measurements such as median and IQR were calculated. Comparisons of patient characteristics and outcomes were performed using a v 2 test or Fisher exact test for categorical variables and an independent t test or a Mann-Whitney rank sum test for continuous variables. A Spearman rank correlation model was used to determine the relationships among all outcome variables and surgical implant number. A regression model was utilized to predict surgical time by surgeon implant number. A P value of <.05 was considered to indicate a statistically significant difference for all statistical tests. RESULTS Between February 2011 and March 2012, implantation was performed in 126 patients (87 United States, 39 Europe). There were 22 participating implant sites (15 United States, three Germany, two France, one the Netherlands, and one Belgium). Study sites included none academic centers and 13 community hospitals. There was a mean of 5.7 implants per site (range ). Implantation data are exhibited in Table I. Demographic information is contained within Table II. Procedural Safety The device was successfully implanted on the first attempt in 100% of the patients. Adverse events are summarized in Table III. There were no cases of pneumothorax, lead dislodgement, permanent nerve damage, or device-related infections associated with the implantation of the upper airway stimulation system. A total of 115 nonserious adverse events occurred in 72 (58.1%)

3 TABLE I. Implantation Data. TABLE II. Demographics. Implantation Data Value Variable Value Successful implantation 126 (100%) on first attempt (%) Mean implant time, hr (range ) Mean implants/center 5.7 (range ) Follow-up length, yr (range ) Discharged same day (%) 20 (16%) Discharged next day (%) 99 (79%) Discharged 2 days 7 (5%) postoperatively (%)* The average implant time was 2.52 hours. *Seven European subjects were discharged on day 2 secondary to local medical policy. patients. Of those 115, 84 (73.0%) were generic adverse events related to surgical procedures in general. The only complication specifically related to upper airway stimulation was temporary tongue weakness in 21 (17%) patients. The majority of cases of tongue paresis resolved after a period of observation. Two patients had additional treatment with steroid therapy, one required pain control, one had fiberoptic laryngeal examination, and one underwent electromyography testing. Tongue weakness most commonly resulted in temporary mild dysarthria with no cases of significant dysphagia. There were no observed device-related infections. There was no correlation between site implant number and complication rate (rho ; P 5.291). Also, the proportion of patients experiencing a complication did not differ when comparing the groups with implant numbers less than five versus five or more (P 5.511). Postoperative Length of Stay The majority of patients were discharged on or before postoperative day 1 (n 5 119, 95%), whereas seven (5%) patients were discharged on postoperative day 2 due to specific hospital policies in those countries. Patients who were discharged on an opioid 6 nonsteroidal anti-inflammatory drug stayed on these medications for a median of 4.5 days (IQR ). Age, yr, mean (range) 54.5 (31 80) Male gender (%) 105 (83%) BMI (kg/m 2 ) Baseline AHI (events/hr) AHI 5 apnea-hypopnea index; BMI 5 body mass index. Postoperative Pain Patients reported a mean pain VAS (pvas) of mm for initial postoperative pain within the first 24-hours postimplantation, and mm at 1 week postoperatively. Neither initial postoperative pvas nor 1-week postoperative pvas correlated with site implant number (rho , P 5.424; rho , P 5.443) (Fig. 1A). In addition, neither postoperative pvas nor 1-week postoperative pvas differed when comparing the groups with implant number less than five versus five or more (P and.557, respectively) (Fig. 1B). Surgical Learning Curve The average implant time was hours (range hours). The surgical implant time decreased significantly with surgeon experience, from hours for a surgeon s first implant (n 5 22) to hours for the fifth implant (n 5 10, P 5.025). Surgical time was inversely correlated with the site implant number (rho ; P <.001). The residual plot allows for the visual evaluation of the goodness of fit of the selected regression model (Fig. 2A). This full model explains 8.7% of the variance in surgical time with surgeon implant number (F (1,125) , P <.001). With an r and a regression coefficient of 0.296, this highlights a good model fit for this type of novel surgery, but also shows variations for certain transect. Also, surgical time decreased from a mean of hours for implants <5 to hours for five or more implants (P <.001) (Fig. 2B). TABLE III. Adverse Events. Unique Events Unique Patients (%) Events Resolved by Observation (%) Events Resolved by Medications (%) Resolved by Other Methods (%) Expected general surgery observations (40%) 58/84 (69%) 20/84 (24%) 6/24 (7%) Events related to postoperative incision (27%) 32/47 (68%) 13/47 (28%) 2/47 (4%) Postoperative discomfort independent (23%) 26/37 (70%) 7/37 (19%) 4/37 (11%) of an incision Observations unique to implantation procedure Temporary tongue weakness (17%) 26/31 (84%) 3/31 (10%) 2/31 (6%) Permanent tongue weakness 0 0 (0%) Device infections 0 0 (0%) The majority of adverse events were resolved by time alone, without requiring intervention on the part of the healthcare team. 503

4 Fig. 1. Pain visual analog scale (pvas). (A) Scatterplot of mean postoperative and 1-week pvas over the number of implants the surgeon had currently performed. There was no correlation between implant number and pvas. (B) Mean postoperative and 1-week pvas examined by implant number less than five and five or more. There was no significant difference between these two groups. DISCUSSION Increased surgical experience for most surgical procedures correlates with improved surgical outcomes and reduced operative time. 18 Early results from UAS in the STAR trial, despite a small sample size, show similar findings. The UAS implant procedure using the Inspire system is performed under general anesthesia. The surgeon uses three incisions, as previously described, 16 to insert the hypoglossal nerve stimulation lead, the respiratory sensing lead, and the implantable pulse generator. Intraoperative testing using electromyography and visual confirmation is required to confirm correct placement of the cuff electrode on the distal branch of the hypoglossal nerve to selectively stimulate the genioglossus (protrusor) branches while excluding the hyoglossus and styloglossus (retrusor) branches. As stated earlier, the average time for all surgical implants throughout the STAR trial was around 2.5 hours. Due to the low number of surgical cases, a surgical time by implant number curve was created instead of a traditional procedure of time curve (Fig. 2A). The surgical implant time decreased significantly with surgeon experience, from hours for a surgeon s first implant (n 5 22) to hours for the fifth implant (n 5 10, P 5.043). Possible factors for improved surgical time include better recognition of the surgical landmarks necessary for proper placement of the stimulation and sensory leads, as well as an increasing familiarity with the device and the implantation procedure. In addition, there appears to be continued reduction in surgical times with increased operative experience, as the physician with the largest single experience (n 5 22) of the study had an average implant time of hours in the surgeon s last six implants. When compared to the vagal nerve stimulator (VNS) procedure, which is performed in around 1 hour in experienced hands, UAS takes twice as long. 19,20 It is important to note that the 1-hour estimation is taken from surgeons with much more experience with VNS implants than any of our trial surgeons. Additionally, UAS requires an additional Fig. 2. Surgical time. (A) Scatterplot with overlaid regression equation of surgical time by surgeon implant number. Surgical time decreased as implant number increased. (B) Mean surgical time comparison for less than five implant number was significantly higher compared to five or more implant number (P <.001). [Color figure can be viewed in the online issue, which is available at 504

5 Fig. 3. Postoperative pain visual analog scale (VAS) (millimeters). *Troell et al. 2000; Rombaux et al. 2003). UPPP 5 uvulopalatopharyngoplasty. [Color figure can be viewed in the online issue, which is available at incision in the chest wall to place the sensing lead for respiration. Although this study demonstrated a 27% reduction in operative time from the first to the fifth implant, this study was unable to determine at which point the learning curve for the surgical time reached a plateau. Further studies with larger surgeon volumes are likely needed to determine when optimum surgical proficiency is achieved. More important than surgical time, the safety profile and perioperative morbidity for UAS was acceptable and compares favorably to other implantable device procedures. In a study of patients who underwent any cardiac implantable electronic device, there was a 9.9% (432/4,355) complication rate for patients undergoing new implants. 21 These included a 0.2% (10/4,355) rate of local infection causing reintervention, 0.2% (10/4,355) pocket revisions due to pain, and 1.1% (47/4,355) wound infections treated by antibiotics. In a study of 105 patients undergoing VNS, there was a 5.7% (6/105) rate of surgery-related complications, which included wound infection or poor wound healing (four patients) and transiency vocal cord palsy (two patients). 22 In the STAR trial, 72 patients experienced an adverse event. However, 51 of those were considered generic surgical side effects, which were not considered adverse events in the other studies. The remaining 21 were related to temporary tongue weakness. There were no wound or device infections in the trial. The protocol training for the trial used the latest guidelines to prevent infections in implantable pacemaker devices including preoperative surgical site wash, intravenous antibiotics prior to incision, iodine draping of the surgical site, and thorough irrigation of the wound with antibiotic-infused saline solution. Of note, two devices required repositioning and were identified as serious device-related events; additionally, there were two unrelated deaths during the trial. Minor wound pain is expected after device implantation and is also common with other pacemaker devices. 23 Wound pain typically resolves on its own but oral analgesic medications can be considered postoperatively. Postoperative pain after cardiac device implantation has been reported to be mm immediately after the procedure, decreasing to mm at 1 month. 24 Postoperative pain from UAS shows a similar postoperative profile with pain a VAS of mm, which decreases to mm at 1 week. Compared to historical data for traditional airway reconstructive surgery for OSA, such as uvulopalatopharyngoplasty (UPPP), postoperative pain was substantially lower in severity and duration for the UAS implant procedure (Fig. 3) Although the patient populations and surgical technique of UPPP and UAS surgery are dissimilar, this comparison remains important, as higher postoperative opioid requirements have been shown to increase perioperative risk and length of stay in obstructive sleep apnea patients undergoing anesthesia. The possibility of reduced postoperative pain with the UAS implant procedure may suggest the likelihood of decreasing perioperative risk for OSA patients such as the need for overnight hospitalization and associated costs. Limitations Although utilization of data from a prospective clinical trial enabled the capture of complete outcomes, limitations include that the STAR trial involved a relatively small sample size and lack of a control group. A higher sample size is needed to determine maximal surgical proficiency and to decrease the risk of a type II error. CONCLUSION Surgeon experience has been shown to improve surgical times in most common procedures, and upper airway stimulation surgery is no different. Implant time for upper airway nerve stimulation decreased significantly after the first five Implants. Complication rates and postoperative pain after UAS surgery are comparable to other implantable neurostimulation devices, and surgeon experience does not appear to improve these outcomes. The UAS procedure can be performed safely by properly trained otolaryngologists, who on average demonstrate improving proficiency with the procedure as their surgical experience increases. Average surgical implant time for the UAS procedure decreases commensurate with increased surgeon experience. Serious procedure-related or device-related complications were uncommon and self-limited, and postoperative pain was substantially less than traditional anatomy-altering surgeries such as UPPP. BIBLIOGRAPHY 1. Montplaisir J, Bedard MA, Richer F, Rouleau I. Neurobehavioral manifestations in obstructive sleep apnea syndrome before and after treatment with continuous positive airway pressure. Sleep 1992;15:S17 S Peker Y, Hedner J, Norum J, Kraiczi H, Carlson J. Increased incidence of cardiovascular disease in middle-aged men with obstructive sleep apnea: a 7-year follow-up. Am J Respir Crit Care Med 2002;166: Shahar E, Whitney CW, Redline S, et al. Sleep-disordered breathing and cardiovascular disease: cross-sectional results of the Sleep Heart Health Study. Am J Respir Crit Care Med 2001;163: Young T, Peppard PE, Gottlieb DJ. Epidemiology of obstructive sleep apnea: a population health perspective. Am J Respir Crit Care Med 2002;165: Giles TL, Lasserson TJ, Smith BH, White J, Wright J, Cates CJ. Continuous positive airways pressure for obstructive sleep apnoea in adults. Cochrane Database Syst Rev 2006;(3):CD McNicholas WT. Cardiovascular outcomes of CPAP therapy in obstructive sleep apnea syndrome. Am J Physiol Regul Integr Comp Physiol 2007; 293:R1666 R

6 7. Chai-Coetzer CL, Luo YM, Antic NA, et al. Predictors of long-term adherence to continuous positive airway pressure therapy in patients with obstructive sleep apnea and cardiovascular disease in the SAVE study. Sleep 2013;36: Weaver TE, Grunstein RR. Adherence to continuous positive airway pressure therapy: the challenge to effective treatment. Proc Am Thorac Soc 2008;5: Decker MJ, Haaga J, Arnold JL, Atzberger D, Strohl KP. Functional electrical stimulation and respiration during sleep. J Appl Physiol 1993;75: Edmonds LC, Daniels BK, Stanson AW, Sheedy PF III, Shepard JW Jr. The effects of transcutaneous electrical stimulation during wakefulness and sleep in patients with obstructive sleep apnea. Am Rev Respir Dis 1992;146: Deegan PC, McNicholas WT. Pathophysiology of obstructive sleep apnoea. Eur Respir J 1995;8: Eisele DW, Smith PL, Alam DS, Schwartz AR. Direct hypoglossal nerve stimulation in obstructive sleep apnea. Arch Otolaryngol Head Neck Surg 1997;123: Eastwood PR, Barnes M, Walsh JH, et al. Treating obstructive sleep apnea with hypoglossal nerve stimulation. Sleep 2011;34: Mwenge GB, Rombaux P, Dury M, Lengele B, Rodenstein D. Targeted hypoglossal neurostimulation for obstructive sleep apnoea: a 1-year pilot study. Eur Respir J 2013;41: Van de Heyning PH, Badr MS, Baskin JZ, et al. Implanted upper airway stimulation device for obstructive sleep apnea. Laryngoscope 2012;122: Strollo PJ Jr, Soose RJ, Maurer JT, et al. Upper-airway stimulation for obstructive sleep apnea. N Engl J Med 2014;370: Hopper AN, Jamison MH, Lewis WG. Learning curves in surgical practice. Postgrad Med J 2007;83: Treasure T. The learning curve. BMJ 2004;329: Binnie CD. Vagus nerve stimulation for epilepsy: a review. Seizure 2000;9: Fox DJ, Davidson NC, Royle M, et al. Safety and acceptability of implantation of internal cardioverter-defibrillators under local anesthetic and conscious sedation. Pacing Clin Electrophysiol 2007;30: Kirkfeldt RE, Johansen JB, Nohr EA, Jorgensen OD, Nielsen JC. Complications after cardiac implantable electronic device implantations: an analysis of a complete, nationwide cohort in Denmark. Eur Heart J 2014;35: Spuck S, Tronnier V, Orosz I, et al. Operative and technical complications of vagus nerve stimulator implantation. Neurosurgery 2010;67: Pavia S, Wilkoff B. The management of surgical complications of pacemaker and implantable cardioverter-defibrillators. Curr Opin Cardiol 2001;16: Bollmann A, Marx A, Sathavorn C, et al. Patient discomfort following pectoral defibrillator implantation using conscious sedation. Pacing Clin Electrophysiol 1999;22: Clarke RW, Yardley MP, Davies CM, Panarese A, Clegg RT, Parker AJ. Palatoplasty for snoring: a randomized controlled trial of three surgical methods. Otolaryngol Head Neck Surg 1998;119: Rombaux P, Hamoir M, Bertrand B, Aubert G, Liistro G, Rodenstein D. Postoperative pain and side effects after uvulopalatopharyngoplasty, laser-assisted uvulopalatoplasty, and radiofrequency tissue volume reduction in primary snoring. Laryngoscope 2003;113: Troell RJ, Powell NB, Riley RW, Li KK, Guilleminault C. Comparison of postoperative pain between laser-assisted uvulopalatoplasty, uvulopalatopharyngoplasty, and radiofrequency volumetric tissue reduction of the palate. Otolaryngol Head Neck Surg 2000;122: