Snoring is a common condition affecting 20% of the

Transcription

1 Temperature-controlled radiofrequency tissue volume reduction in the human soft palate AARON E. SHER, MD, PHILLIP B. FLEXON, MD, DAVID HILLMAN, MD, BRIAN EMERY, MD, JOHN SWIECA, MB, BS, FRACP, TIMOTHY L. SMITH, MD, MPH, ROSALIND CARTWRIGHT, PHD, ERIC DIERKS, MD, DMD, FACS, and LIONEL NELSON, MD, FACS Albany, New York, Savannah, Georgia, Baltimore, Maryland, Milwaukee, Wisconsin, Portland, Oregon, Nedlands, Western Australia, East Melbourne, Australia, Chicago, Illinois, and San Jose, California OBJECTIVE: To validate the use of temperaturecontrolled radiofrequency energy applied to the soft palate in a multicenter setting for reduction of snoring in a minimally morbid manner. METHODS: Prospective, nonrandomized multicenter study of 113 patients who had a respiratory disturbance index less than 15 and minimum oxygen saturation not less than 85% and who were seeking treatment for habitual disruptive snoring. Patients were given either single or multiple lesions to the soft palate during each treatment session. RESULTS: Patients received 1978 J on average with an overall average of 2.4 treatments. Snoring scores went from an average of 7.8 (visual analog scale [VAS], 0-10) pretreatment to 3.2 posttreatment. Pain was minimal, averaging 1.7 (VAS 0-10) on days 1 to 6. Complications were few and transient, and mild. CONCLUSIONS: The multiple lesion protocol was the most successful; reducing snoring from 7.6 to 2.7, on a VAS with an average of 1232 J delivered over 1.6 treatments. Temperature-controlled radiofrequency was found to be a minimally invasive, well-tolerated procedure that was safe and efficacious in this study group. (Otolaryngol Head Neck Surg 2001; 125:312-8.) From the Capital Region Otolaryngology (Dr Sher); Georgia Ear Institute (Dr Flexon); University of Maryland Medical Center (Dr Emery); Medical College of Wisconsin (Dr Smith); Oregon Health Services University (Dr Dierks); Sir Charles Gairdner Hospital (Dr Hillman); Mercy Private Hospital (Dr Swieca); Rush-Presbyterian-St Luke s Medical Center (Dr Cartwright); and Otolaryngology (Dr Nelson). Funding provided by Somnus Medical Technologies, Inc, Sunnyvale, CA. Reprint requests: Aaron E. Sher, MD, Capital Region Otolaryngology Head and Neck Surgery, 6 Executive Park Drive, Albany, NY 12203; , Aeshermd@aol.com. Copyright 2001 by the American Academy of Otolaryngology Foundation, Inc /2001/$ /1/ doi: /mhn Snoring is a common condition affecting 20% of the general adult population and up to 60% of men over the age of It results from narrowing and partial obstruction of the upper airway during sleep because of crowding or unfavorable positioning of the pharyngeal components, including uvula, soft palate, and tongue. Snoring has been associated with an increased risk of hypertension, angina pectoris, and cerebral infarction. 2 It is associated with significant social inconvenience and distress to the bed partners of snoring patients. It has recently been demonstrated that sleep efficiency, sleep continuity, and sleep architecture of the bed partner is disturbed. Elimination of snoring is associated with polysomnographic improvement in the bed partner s sleep manifested by significantly improved sleep efficiency and continuity. 3 More severe collapse of the pharyngeal airway may result in the morbid sequelae of upper airway resistance and obstructive sleep apnea syndrome (OSAS), including excessive daytime sleepiness. 4 The prevalence of OSAS is 4% of working men and 2% of women, almost all of whom are snorers. 5 The currently available treatment modalities for snoring range from noninvasive medical management to invasive surgical management. Medical management includes weight reduction, modification of sleep position, and dental appliances. Successful medical management depends to a large extent on patient compliance and tolerance. Surgical interventions, including pharyngeal debulking procedures that are often effective and not dependent on patient tolerance or compliance, are costly and can be associated with morbidity. Such surgical therapies, include various forms of uvulopalatopharyngoplasty (UPPP), which may be performed with or without electrosurgical instrumentation and with or without tissue flaps, 6,7 and laser-assisted uvulopalatoplasty LAUP. 8,9 A high percentage of snorers, although significantly troubled by their condition, decline surgical treatment because of associated invasiveness and morbidity. This multicenter clinical trial was conducted to validate previously published works and to see if the results could be reproduced by other practitioners. The 312

2 Volume 125 Number 4 SHER et al 313 approach involves application of temperaturecontrolled radiofrequency energy (TCRFe) to coagulate and ablate submucosal muscle of the soft palate via a needle electrode placed permucosally. TCRFe results in modest elevation of tissue temperature (45 to 90 C) and coagulative necrosis of tissue adjacent to the deep, uninsulated portion of the electrode. Mucosa adjacent to the proximally insulated needle electrode is protected. TCRFe disbursement is proportional to the reciprocal of the fourth power of the radius. 10 The rapid decrease in energy intensity with increasing distance from the source helps to target tissue to be precisely coagulated. Lesion size is controlled by the diameter and length of the needle electrode and by treatment duration and total energy delivered (measured in Joules or watts seconds). Further safety of soft tissue TCRFe ablation results from the intrinsic tissue/tcrfe interactive properties that limit tissue destruction. As tissues coagulate at the electrode-tissue interface, desiccation occurs. The use of temperature controls effectively limits tissue necrosis and prevents excessive tissue damage. These unique physical properties of TCRFe resulted in prior applications where precise tissue ablation was required and excess ablation was undesirable or dangerous. Examples include localized tissue ablation in the heart or central nervous system (for Wolfe-Parkinson-White syndrome and trigeminal neuralgia, respectively). 11,12 Radiofrequency tissue volume reduction (RFTVR), applied as treatment for benign prostatic hyperplasia (BPH) in a procedure named trans urethral needle ablation (TUNA ) represents the only efficacious treatment modality for BPH that can be performed as an office procedure with the patient under local anesthesia. 13 RFTVR has also been applied to treat malignant lesions of prostate cancer and liver cancer. 14,15 METHODS Protocol Design A prospective, nonrandomized clinical study was conducted at 8 institutions by investigators with experience in surgical treatment of snoring and sleep-disordered breathing (SDB). The study was approved by each institution s Institutional Review Board, and informed consent was obtained by the investigators before enrolling patients in this study. Outcome variables for this study using 10 cm visual analog scales (VAS) were: (1) a measure of the pain and discomfort associated with the procedure (VAS anchors at 0 for no pain to 10 for intense pain), and (2) a measure of snoring (anchors of 0 for no snoring to 10 for loud snoring, bed partner leaves the room). Both variables were scored by previously published indexes. 16,17 Sleep quality was also evaluated using daytime sleepiness by VAS (anchors at 0 for never sleeps during the day to 10 for would fall asleep if not stimulated) and the Epworth Sleepiness Scale (ESS) ESS is a patient questionnaire consisting of 8 questions assessing the subject s chance of dozing in specific situations. Patient Selection and Evaluation Study subjects were nonobese healthy individuals seeking treatment for habitual disruptive snoring. Subjects who agreed to participate were given a detailed physical examination of the upper airway and head and neck; a complete medical history was taken. Female patients without an effective contraceptive method were required to undergo a urine pregnancy test before each treatment session. Standard attended overnight polysomnography (PSG) was performed before the first treatment to rule out moderate or severe obstructive sleep apnea. To be eligible for the study, subjects were required to have a respiratory disturbance index (RDI) less than or equal to 15 apneas and hypopneas per hour and minimum oxygen saturation not lower than 85%. Before commencing treatment, subjects completed questionnaires to evaluate their degree of daytime sleepiness. Other baseline parameters, such as speech and swallowing were assessed subjectively by the patient using VAS scales. To assess snoring, the patient s bedpartner was asked to rank the loudness of snoring on the VAS. This instrument uses a scale of 0 to 10 (0, no snoring; 1-3, soft snoring not bothersome to the bedpartner; 4-6, snoring bothersome to the bedpartner; 7-9, snoring annoying to anyone nearby; and 10, snoring so bothersome that the bedpartner leaves the room). 18 Subjects were seen 24 to 72 hours, 1 week, and 4 weeks after each treatment for follow-up evaluation including physical examination and review of responses on patient questionnaires reporting postoperative pain and the effect of treatment on swallowing, speech, snoring, and sleep quality. Postoperative pain was further assessed by recording the type and duration of postoperative analgesia required for adequate pain control. Retreatment was offered to subjects if snoring was not improved to a level acceptable to the patient s bedpartner at each 4-week posttreatment follow-up visit. At least 8 weeks after the final treatment, subjects underwent a final PSG assessment. Radiofrequency Procedure RF energy (460 khz) was delivered using an RF generator and proprietary needle electrode that includes thermocouples in the active needle tip and at the tip of the insulation (Somnus Medical Technologies Inc, Sunnyvale, CA). The generator includes an algorithm that automatically modulates the power delivered to maintain the target temperature as set by the physician in the 60 to 90 C range. Development of the algorithm was previously reported. 18,20 The procedure was performed in an outpatient surgical suite or physician s office. The soft palate was anesthetized topically using a spray of 20% benzocaine followed by local injection of 1% or 2% lidocaine or 0.25% marcaine in the region of the soft palate to be

3 314 SHER et al October 2001 Table 1. Demographics and baseline characteristics Mean patient Mean body Mean neck age Sex mass index circumference Mean RDI Mean LSAT Mean Epworth Score Mean snoring index (±SD) (years) (% Males) (±SD) (kg/m 2 ) (±SD) (inches) (±SD) (±SD) (±SD) (±SD) 45.8 ± % 27.6 ± ± ± ± ± ± 2.1 (N = 105) (N = 105) (N = 104) (N = 68) (N = 105) (N = 105) (N = 103) (N = 105) Table 2. Pretreatment snoring index, overall mean pretreatment snore index for all patients (N = 105) was 7.8 ± 2.1 (range 2.0 to 10.0) Snoring severity category Mean snoring index all patients* (2.9%) (28.6%) (36.2%) (32.4%) Total 105 *As measured on a scale of 0 to 10. treated. Whether applied in the midline or in the paramedian positions, the needle electrode was placed below the mucosa into the muscle layer with the entry point located within 5 to 10 mm of the junction of the hard and soft palate, the exact site of entry determined by the length of the soft palate. The needle entry point was selected to assure that the active, distal portion of the electrode did not extend into the base of the uvula. In addition, the probe was placed at the proper depth level to avoid damage to the nasopharyngeal and palatal mucosa. Two treatment regimens were applied. Some patients were treated with a single midline energy dose at each treatment session, eg, the single lesion protocol. Other patients were treated with 3 energy doses, 1 in the midline and 2 paramedian, at each treatment session, eg, the triple lesion protocol. Some patients were treated with various combinations of these approaches, classified as other. Investigators were assigned to perform either single lesion or triple lesion procedures. An attempt was made to perform each ablation at a different site in the palate. Subjects were advised not to ingest food until local and topical anesthetic effects had resolved and not to participate in strenuous activities, but were otherwise not restricted from normal activities. After surgery, patients were advised to take nonnarcotic analgesic medication (without anti-inflammatory agents) as required to control pain, and to sleep with the head elevated to minimize the effects of postoperative edema. Patients were instructed to contact the investigator if the pain or edema were not adequately controlled by these measures. Statistical Analysis Descriptive statistics were computed for subject demographics and baseline characteristics. Results are expressed with standard deviation and were generated using a computerized statistical package (SAS, Cary, NC). Individual center data as well as pooled data were analyzed. Depending on the distribution of the variable, analysis of variance (ANOVA), or nonparametric methods were used to determine whether differences among centers and between baseline and final measurements were statistically significant. The incidence of acute or late complications for all subjects was recorded. RESULTS One hundred twelve subjects (24 female and 88 male) were enrolled in this multicenter prospective nonrandomized study between 1996 and One hundred five completed the study; 7 were withdrawn or lost to follow-up. These patients did not differ from the remainder of the study population. Demographics and baseline characteristics for the remaining patients at the time of treatment are shown in Table 1. Mean age was 45.8 ± 10.2 years (range, 23 to 74 years). Mean body mass index was 27.6 Kg/m 2 ± 3.6 (range, 20.5 to 39.6). Mean neck circumference was 16.0 ± 1.6 inches (range, 12.5 to 21.3). Mean RDI was 6.4 ± 4.3 apneas and hypopneas per hour (range, 0 to 16.7), and mean lowest oxygen saturation (LSAT) was 88.9 ± 4.0% (range, 77.0% to 97.0%). Mean Epworth Sleepiness Score (ESS) was 9.1 ± 4.9 (range, 0 to 20). Table 2 demonstrates the distribution of patients by snoring severity category. Mean pretreatment snoring VAS was 7.8 ± 2.1 (range, 2.0 to 10.0). Thirty-four (32.4%) patients had a pretreatment VAS of 10; 38 (36.2%) had a pretreatment VAS of 7 to 9; and 33 (31.5%) had a pretreatment VAS of 1 to 6. Tables 3 and 4 summarize treatment parameters including number of treatments, energy delivered, and different treatment regimens used. Combining all treatment groups, the mean number of treatment sessions per patient was 2.4 ± 1.2. The mean energy delivered per treatment was ± J, and the mean total energy delivered over all treatments was ± J. Fifty-seven patients were treated under the single lesion protocol. The mean energy delivered per treatment (and per lesion) was ± 69.3 J, and the mean total energy delivered over all treatments was ± J. The mean number of treatment sessions was 2.9 ± 1.1, with 79% having 3 or fewer sessions.

4 Volume 125 Number 4 SHER et al 315 Table 3. Summary number of treatments Number of treatments Single lesion (midline) Triple lesion (midline and laterals) Other All 1 Treatment 8 (14.0%) 18 (56.3%) 3 (18.8%) 29 (27.6%) 2 Treatments 7 (12.3%) 8 (25.0%) 8 (50.0%) 23 (21.9%) 3 Treatments 30 (52.6%) 6 (18.8%) 4 (25.0%) 40 (38.1%) 4 or More Treatments 12 (21.1%) 0 1 (6.3%) 13 (12.4%) Mean 2.9 ± ± ± ± 1.2 Table 4. Summary of mean energy delivered Treatment type Mean Joules per treatment* Mean Joules per patient (all treatments) Single lesion (Midline) (± SD) ± 69.3 (N = 167) ± (N = 57) Triple lesion (midline and laterals) (± SD) ± (N = 52) ± (N = 32) Combination (± SD) ± (N = 35) ± (N = 16) All (± SD) ± (N = 254) ± (N = 105) *Mean Joules per treatment is the calculated average of all lesions per patient per treatment. Mean Joules per patient is the sum of energy from all lesions per patient over all treatments. Thirty-two patients were treated under the triple lesion protocol. The mean energy delivered per treatment session (divided among 3 lesions) was ± J. The mean total energy delivered over all treatments was ± J. The mean number of treatment sessions was 1.6 ± 0.8; 56.3% had only 1 session, 81.3% had 2 or fewer sessions, and no patients had more than 3 sessions. Sixteen patients are classified in the other category combination; their treatment was conducted according to neither the single nor the triple lesion protocol, but a variable combination of both. The mean energy delivered per treatment session (divided among a variable number of lesions) was ± J. The mean total energy delivered over all treatment sessions was ± J. The mean number of treatment sessions was 2.2 ± 0.8; 18.8% had only 1 session, 93.8% had 3 or fewer sessions. Eight weeks after the final treatment session, the mean change in snoring for the entire treatment population was 4.6 ± 2.9, from preoperative mean of 7.8 ± 2.1 to a postoperative mean of 3.2 ± 2.2 (P < ) (Table 5). This represented a mean decrease in snoring index of 56.5 ± 31.3%. The mean percent change in snoring was 54.8 ± 33.7% (from 7.5 ± 2.0 to 3.2 ± 2.3) for single lesion protocol patients; ± 26.8% (from 7.6 ± 2.4 to 2.7 ± 1.7) for triple lesion protocol patients; and ± 31.7% (from 9.0 ± 1.3 to 4.0 ± 2.6) for other treatment patients. Four patients (3.8%) were worse after treatment, and 2 (1.9%) had no change. The body mass index of the study subjects did not change from the pretreatment values, 27.7 ± 3.5 Kg/m 2 (pre TCRFe) versus 27.5 ± 3.5 (N = 86, range 20.6 to 39.9 Kg/m 2, P = 0.43). A successful treatment outcome was defined as a 2- Table 5. Pretreatment and posttreatment snoring index (SI) and treatment type Triple lesion Single lesion (midline and (Midline) laterals) Other All Pretreatment 7.5 ± ± ± ± 2.1 Mean SI (±SD) (N = 57) (N = 32) (N = 16) (N = 105) Posttreatment 3.2 ± ± ± ± 2.2 Mean SI (±SD) (N = 57) (N = 32) (N = 16) (N = 105) month postoperative snore index of no more than 3, 66.7% of patients (70 of 105) had a successful outcome. The rate of success was 61.4% (35 of 57) for the single lesion protocol; 84.4% (27 of 32) for the triple lesion protocol; and 50.0% (8 of 16) for the combination patients. The difference in outcome between the 3 treatment types with respect to success is not statistically significant (P = 0.63). For the entire population achieving a 2-month postoperative snoring index of 3 or less (including single lesion, triple lesion, and combination categories), the number of treatments required was 1 (41.4%), 2 (22.9%), 3 (25.7%), or more than 3 (10.0%). The maximum number of treatments required was 6. Of those patients ultimately succeeding in the single lesion protocol (N = 35), 42.9% of the patients required 1 or 2 treatment sessions, 45.7% required 3 or 4 sessions and 11.4% required 5 sessions. Of those patients ultimately succeeding in the triple lesion protocol (N = 27), 85.2% required 1 or 2 treatment session and 14.8% required 3 sessions. The pretreatment mean RDI was 6.4 ± 4.3 (N = 105) range (0 to 16.7); the posttreatment RDI was 8.7 ± 8.9,

5 316 SHER et al October 2001 Table 6. Posttreatment pain assessment Mean value (±SD) Pain assessment All patients time point N (as measured on a scale of 0 to 10) Day ± 1.6 Weeks ± 2.2 Week ± 0.8 (n = 93, range 0 to 51.4), P < ). Twelve individuals refused to have a final PSG. The pretreatment mean Epworth Sleepiness Score of 9.1 ± 4.9 (n = 103, range 0 to 20) decreased to a mean posttreatment level of 5.9 ± 4.3, (n = 98, range 0 to 18), P < Postoperative pain was assessed by VAS (Table 6) as well as by evaluating the type of medication required and its duration of use; 41.9% required nonnarcotic medications and 6.7% required narcotic medications for one day or less. The remainder (51.4%) required no postoperative medication for control of pain. All patients experienced a mild sore throat and edema after the procedure. Edema reported after the procedure was adequately managed by having the patient sleep with his or her head elevated with pillows for 1 to 2 days after the treatment; 2 (3.2%) patients required postoperative steroids. Three (2.9%) patients developed 3 to 5 mm superficial ulcerations of the mucosa. Two (1.9%) patients had a small soft palate fistula (3 to 4 mm) that closed within several days without further sequelae. Two (1.9%) patients sloughed the uvula without further sequelae. Neither speech nor swallowing was adversely affected. The mean VAS score for speech at week 4 posttreatment was 0.1 ± 0.4 (range, 0 to 4) and for swallowing was 0.2 ± 0.7 (range, 0 to 5). DISCUSSION On the basis of feasibility, efficacy, safety, and reproducibility, a 2-phase study involving animal and human subjects was undertaken to evaluate the feasibility of applying TCRFe for RFTVR of pharyngeal tissue in patients with OSAS. 18,20 In the first phase of the trial, TCRFe was delivered to porcine tongues. It was shown to result in precise intramuscular lesions sparing surface mucosa at the electrode entry site. Lesion size correlated with the amount of energy delivered to the lesion site. Postoperative edema was minimal. Gross and histologic investigation of the tongue demonstrated that TCRFe lesions follow the usual course of wound healing. Histologic assessment over time documented wellcircumscribed lesions with no damage to nerves outside the core lesion. The findings contrasted favorably with those of soft palate electrocautery or laser assisted uvulopalatoplasty (LAUP), in which significantly higher temperatures (740 to 900 C) result in excessive mucosal, submucosal, and muscle damage. At 10 days after TCRFe treatment, a volumetric reduction equivalent to 0.84 g of tongue tissue was documented at each specific ablation site. By way of comparison, the weight of normal human tonsils averages 2 to 4 g. 20 The results of the animal phase of the study suggested that RFTVR might provide a minimally invasive, quick, and easy method for diminishing the volume of the human tongue for OSAS resulting from retrolingual airway collapse. The clinical phase of the study was designed to assess the feasibility and effect of TCRFe application in the human pharynx. The ultimate goal of this research was to apply TCRFe to the tongue for retrolingual collapse in OSAS. Nonetheless, it was felt that patient safety mandated that this first clinical trial of TCRFe application to the pharynx be carried out on the soft palate, not the tongue. 18 The goals of this portion of the study were evaluation of tissue shrinkage and evaluation of adverse effects (including postoperative edema, pain, and impairment of speech and swallowing). In effect, the study established safe and effective energy limits for TCRFe application for snoring and monitored adverse effects. Twenty-four subjects were selected for the primary complaint of snoring. In all cases, anatomic evaluation with fiberoptic endoscopy and lateral cephalometry suggested that the major site of obstruction was located at the palate. They were all assessed for sleep-disordered breathing (SDB) with overnight PSG, SDB more severe than mild OSAS disqualified the patient from participating in the trial. The impact of TCRFe on both snoring and sleep quality were evaluated. Resolution or satisfactory attenuation of snoring was the clinical endpoint for treatment for each patient. PSG was performed preoperatively and after 48 to 72 hours after the first TCRFe application, and finally at a minimum of 8 weeks after completion of the final TCRFe treatment. TCRFe treatments were delivered to the palate in an outpatient setting under local anesthesia. A series of consecutive low to mid level energy treatments were planned, with TCRFe delivered at 460 khz with an RF generator outfitted with custom-fabricated needle electrodes and delivery device (Somnus Medical Technologies, Inc, Sunnyvale, CA). Necessary feedback for temperature adjustment was provided by multiple thermocouples embedded along the electrode. Infrared thermography was used to document tissue thermal gradients to safely facilitate optimum maximum temperatures for treatment. A standardized questionnaire and visual analog scales (VAS) were administered before treatment, at regular postoperative intervals, and at termination of the protocol:

6 Volume 125 Number 4 SHER et al 317 Epworth Sleepiness Scale, visual analog scales for pain, speech problems, swallowing problems, and snoring levels. Once the subject and his/her bedpartner registered a satisfactory VAS score for snoring, the protocol was brought to termination, and the subject underwent final PSG evaluation. 18 The present study evaluates an application of temperature-controlled RFe to the palate in a large number of patients by 8 independent investigators. In the current study, mean pretreatment snoring VAS of 7.8 ± 2.1 was diminished to a mean posttreatment snoring VAS of 3.2 ± 2.2, a mean decrease of 56.5% (P < 0.001). Before treatment, 68.6% of patients had an index of 7 or greater, whereas after treatment, 66.7% of patients had a VAS of 3 or less. Although the rate of success was 61.4% (35 of 57) for the single lesion protocol compared with 84.4% (27 of 32) for the triple lesion protocol, the difference in outcome between these approaches with respect to success was not statistically significant. The results suggest that although the effect of TCRFe on the palate will vary among individual patients, an average of somewhat more than 2000 J will produce a substantial improvement in snoring. Note that this figure is in agreement with the outcome of the study by Powell et al, 18 in which the mean overall total energy delivered per patient at the end of treatment was 2377 ± 869 J. The TCRFe procedure most often requires 2 or more treatments to achieve a clinically meaningful reduction in the level of snoring. For the population achieving a snoring score (VAS) of 3 or less, 22.9% of the single lesion protocol patients required 1 treatment session and 66.7% of the successful triple lesion protocol patients required only 1 treatment. All patients experienced mild sore throat and edema after the procedure. Postoperative pain was modest and of relatively brief duration. Almost half of the patients required no postoperative analgesic medications, and fewer than 10% required narcotic medications. This is comparable to the findings in the report by Powell et al 18 and Coleman and Smith, 17 in which the mean VAS pain score was in the 0-3 category (mild pain not requiring prescription analgesic). This compares favorably with UPPP and LAUP, in which postoperative pain usually mandates postoperative narcotic analgesia in the early postoperative period and for up to 7 to 10 days. 19 The TCRFe procedure was found to be safe, with complications being few in number, transient, and mild in nature. Although a few patients presented with edema of the uvula in the early postoperative period and resulting disturbed sleep was reported in some cases on the first night or 2 after surgery, no patients demonstrated airway compromise. Any edema reported after the procedure was managed by having the patient sleep with head elevated with pillows for 1 to 2 days after the treatment. This is in agreement with the findings of Powell et al. 18 In that study, edematous changes noted at 1 to 3 days after treatment were reflected in increased RDI and decreased nadir SaO 2, but were not sufficient to obstruct the airway awake or during sleep. 18 None of the treatment-related complications reported in this study had long-term consequences. In the report of Powell et al 18 and Coleman and Smith, 17 fewer than 10% of patients experienced superficial mucosal erosion 2 to 4 days after treatment, and all resolved spontaneously without treatment. Neither speech nor swallowing were adversely affected. The mean VAS score for speech at week 4 posttreatment was 0.1 ± 0.43 (range, 0 to 4) and for swallowing was 0.2 ± 0.7 (range, 0 to 5). This is comparable to the report of Powell et al 18 and Coleman and Smith. 17 The anatomical alteration of the palate with TCRFe is minimal compared with that of UPPP or LAUP. Although the anatomical changes are subtle, reduction in the length of the soft palate (measured from the nasal spine to the tip of the soft palate) has been documented. 18 It is believed, however, that the impact of this procedure results primarily from diminished tissue volume and intramuscular tightening resulting from the lesions produced by the temperature-controlled RFe. It is important to note that the patient population for this experimental protocol was carefully selected: all subjects were nonobese and healthy with at most minimal OSAS objectively documented by PSG. Neither the safety nor the relative success of the procedure can be extrapolated for patients with different clinical parameters. Specifically, it should not be assumed that TCRFe is safe or efficacious for patients with more severe OSAS. Indeed, appropriate clinical evaluation of OSAS severity is mandatory for all patients considered for this procedure. The subjective snoring metric applied to assess response to therapy, response herein defined as decrease in the offensive quality of bedpartner snoring, was selected after consideration of potential alternate objective sound measurements. Objective parameters by which to define offensive snoring have not been adequately delineated. To date there are no studies validating the electronic measurement of a sound scored as a snore by polysomnographic technologist or by a computer and its perception as an offensive snore by a listener. In the absence of such validation and standardization, and considering the cost involved in objective measurements, snoring is frequently assessed by subjective means. 17,18 For these reasons, offensive snoring

7 318 SHER et al October 2001 is assessed subjectively in this article. Nonetheless, the lack of an objective outcome measure for snoring is recognized as a limitation of the study. The degree to which the clinical data may have been affected by placebo effect, by unrelated psychosocial factors in the relationship of each patient-bedpartner couple, and by unidentified and unquantifiable variation in each patient-physician interaction remain undefined. CONCLUSION The application of TCRFe of the palate for snoring, as demonstrated in this multicenter report, represents a continuum of a multistaged animal and human clinical trials to assess the safety and efficacy of temperaturecontrolled RFe as a treatment for significant OSAS in patients with retrolingual airway collapse. 17,18,21 Temperature-controlled RFe of the soft palate appears to be a minimally invasive and safe approach and can be performed easily in an office setting with minimal disruption of normal patient activities. Nominal postoperative discomfort and only very subtle anatomical alteration of the palate distinguish this procedure from LAUP, the current mainstay of surgical treatment for snoring in the absence of significant OSAS. The results of this multicenter study report 2-month follow-up. It is reported that the efficacy of radiofrequency tissue volume reduction of the palate decreases overtime to a long-term success rate comparable with UPPP or LAUP. 22 However, the lack of significant anatomic alterations of the palate and the high patient tolerance of the procedure allows for retreatment if necessary. We thank Denise Williamson, MPH, Jill Warren-Mordecai, MPH, and Lindsey Lian, MS, Quintiles, Inc, for providing direction and performing the statistical analyses. REFERENCES 1. Hoffstein V, Mateika JH, Mateika S. Snoring and sleep architecture. AM Rev Respir Dis 1991;143: Jennum P, Hein H, Suadicani P, et al. Cardiovascular risk factors in snorers: a cross-sectional study of 3,323 men aged 54 to 74 years: the Copenhagen male study. Chest 1992;102: Beninati W, Harris CD, Herold DL, et al. The effect of snoring and obstructive sleep apnea on the sleep quality of bed partners. Mayo Clinic Proceedings 1999;74: Cirignotta F, Lugaresi E. Some cineradiographic aspects of snoring and obstructive apneas. Sleep 1980;3: Young T, Palta M, Dempsey J, et al. The occurrence of sleep disordered breathing among middle-aged adults. N Engl J Med 1993;328: Fujita AS, Conway W, Zorick F, et al. Surgical correction of anatomic abnormalities in obstructive sleep apnea syndrome. Uvulopalatopharyngoplasty. Otolaryngol Head Neck Surg 1981; 89: Sher AE, Schechtman KB, Piccirillo JF. The efficacy of surgical modifications of the upper airway in adults with obstructive sleep apnea syndrome. Sleep 1996;19: Walker RP, Grigg-Damberger MM, Gopalsami C, et al. Laserassisted uvulopalatoplasty for snoring and obstructive sleep apnea: results in 170 patients. Laryngoscope 1995;105: Terris D, Clerk A, Norbash A, et al. Characterization of postoperative edema following laser-assisted uvulopalatoplasty using MRI and polysomnography: implications for the outpatient treatment of obstructive sleep apnea syndrome. Laryngoscope 1996;106: Smith TL, Smith JM. Radiofrequency electrosurgery. Oper Tech Otolaryngol Head Neck Surg 2000:11: Jackman W, Wang X, Friday K, et al. Catheter ablation of accessory atrioventricular pathways (Wolfe-Parkinson-White syndrome) by radiofrequency current. N Engl J Med 1991;324: Sweet W, Wepsic J. Controlled thermocoagulation of trigeminal ganglion and rootlets for differential destruction of pain fibers: I. trigeminal neuralgia. J Neurosurg 1974;3: Issa M, Oesterling J. Transurethral needle ablation (TUNA ): an overview of radiofrequency thermal therapy for the treatment of benign prostatic hyperplasia. Curr Opin Urol 1996;6: Prionas S, Kapp D, Goffinet D, et al. Thermometry of interstitial hyperthermia given as an adjuvant to brachytherapy for the treatment of carcinoma of the prostate. Int J Radiation Oncology Biol Phys 1993;28: McGahan J, Schneider P, Brock J, et al. Treatment of liver tumors by percutaneous radiofrequency electrocautery. Semin Interventional Radiol 1993;10: Powell NB, Riley RW, Guilleminault C, et al. Short report: surgical technique. a reversible uvulopalatal flap for snoring and sleep apnea syndrome. Chest 1996;19: Coleman SC, Smith TL. Midline radiofrequency tissue reduction of the palate for bothersome snoring and sleep disordered breathing: a clinical trial. Otolaryngol Head Neck Surg 2000;122: Powell NB, Riley RW, Troell RJ, et al. Radiofrequency volumetric tissue reduction of the palate in subjects with sleep disordered breathing. Chest 1998;113: Troell RJ, Powell NB, Riley RW, et al. Comparison postoperative pain between laser-assisted uvulopalatoplasty, uvulopalatopharyngoplasty, and radiofrequency volumetric tissue reduction of the palate. Otolaryngol Head Neck Surg 2000;122: Powell NB, Riley RW, Troell RJ, et al. Radiofrequency volumetric reduction of the tongue: a porcine pilot study for the treatment of obstructive sleep apnea syndrome. Chest 1997;111: Powell NB, Riley RW, Guilleminault C. Radiofrequency tongue base reduction in sleep-disordered breathing: a pilot study. Otolaryngol Head Neck Surg 1999;120: Li KK, Powell NB, Riley RW, el al. Radiofrequency volumetric reduction of the palate: an extended follow-up study. Otolaryngol Head Neck Surg 2000;122:410-4.