© 2002 American Thoracic Society
Bipolar Radiofrequency Thermal Ablation of the Soft Palate in Habitual Snorers without Significant Desaturations Assessed by Magnetic Resonance ImagingDepartment of Otorhinolaryngology—Head & Neck Surgery and Department of Radiology, Helsinki University Central Hospital; and Department of Clinical Neurosciences, Haaga Neurological Research Center, University of Helsinki, Helsinki, Finland Correspondence and requests for reprints should be addressed to Leif J. J. Bäck, M.D., Department of Otorhinolaryngology—Head & Neck Surgery, University Central Hospital, Haartmaninkatu 4 E, P.O. Box 220, Helsinki 00290, Finland. E-mail: leifback{at}hotmail.com
Sleep-disordered breathing, including habitual snoring, is a major health problem. Treatment of primary habitual snoring should be individualized using both conservative and active treatment methods. Active surgical interventions are, however, associated with significant morbidity. Therefore, procedures causing far less morbidity should be preferred. Our aim was to assess the efficacy and acceptability of bipolar radiofrequency thermal ablation of the soft palate in habitual snorers without significant desaturations associated with excessive daytime sleepiness. We treated 20 nonobese habitual snorers (median age, 43 years, range 35–63). All the patients had the major site of obstruction at the level of the soft palate, and they were treated on an outpatient basis in two treatment sessions separated by 1 week. The pretreatment and post-treatment symptoms and findings as well as the overall efficacy of the procedure were evaluated by questionnaires, visual analogue scales, and magnetic resonance imaging. All the questionnaires showed a statistically significant change, indicating decreased snoring and daytime sleepiness. The magnetic resonance studies showed that the procedure induced notable T1-signal alterations in the treated tissue, and when compared with the pretreatment images, certain dimensions of the soft palate were significantly changed. Bipolar radiofrequency thermal ablation of the soft palate seems to be well tolerated and effective regarding primary habitual snorers without significant desaturations associated with excessive daytime sleepiness.
Key Words: radiofrequency thermal ablation • habitual snoring • magnetic resonance imagine
Habitual snoring is a common condition affecting 9–35% of the general adult population, and its prevalence increases up to the age of 60–65 years and decreases slightly thereafter (1). Snorers without obstructive sleep apnea syndrome may be at a risk of having a decrease in daytime alertness if they have the upper airway resistance syndrome (2). In addition, snoring with related daytime sleepiness was associated with an increased risk of occupational accidents, according to a population-based follow-up study done by Lindberg and coworkers (3) over a period of 10 years. An association between snoring and cardiovascular risk factors exists, but there are several confounding effects that need to be considered in future studies on the predictive value of snoring regarding cardiovascular diseases (4, 5). Habitual snoring is strongly associated with the development of sleep apnea, and an early intervention to control this symptom may be important. Snoring is a noisy inspiratory sound produced by vibrations and partial obstruction in the oropharynx. Although in some cases it can be associated with structural abnormalities, the obstruction is mostly the consequence of a functional pharyngeal hypotonia (possibly due to the Bernoulli effect [6]) in relation to a relatively narrow pharyngeal cross-sectional area (7). In a study by Kimoff and coworkers (8), snorers and patients with obstructive sleep apnea syndrome, compared with normal subjects, have impaired sensation in the upper airway, which may predispose them to the development of apneas of longer duration. Treatment of habitual snoring should be individualized, including conservative measures such as weight loss, alcohol and sedative avoidance, smoking cessation, and sleep-position training. Active treatment methods of habitual snoring include mandibular advancement appliances (9) and nasal continuous positive airway pressure respiration. However, these methods cannot be considered as good choices for a habitual snorer to use for the rest of his life. The surgical interventions of the soft palate share a common goal, i.e., reduction, stiffening, and/or stabilization of the soft palate. Surgical manipulations are effective in patients with sleep-disordered breathing (SDB) resulting in a successful relief from snoring in 80% of the subjects (10) and from obstructive sleep apnea syndrome in 50% (11) of the subjects. However, surgical procedures are associated with significant morbidity, and life-threatening complications have been encountered (12, 13). In addition, the frequencies of long-term side effects on swallowing, regurgitation, taste, smell, and voice have been reported to be up to 40–60% (14). Therefore, procedures in which a variable amount of normal tissue of the soft palate is removed cannot be regarded as an optimal solution for the treatment of habitual snoring. It is clear that, in the future, procedures causing far less morbidity should replace these surgeries. Radiofrequency currents, using heat as the mechanism of action to modify tissue, have been used for various applications in medicine to provide cutting, coagulation, or necrosis of tissue. Radiofrequency energy is delivered in a monopolar or bipolar fashion, with low heat energy sufficient to denature protein. Monopolar radiofrequency energy is considered effective and safe, and has been associated with low morbidity in treatments of the soft palate (15, 16), nasal turbinates (17), and the tongue base (18). Monopolar radiofrequency energy has been shown to produce less postoperative pain compared with laser-assisted uvulopalatoplasty and uvulopalatopharyngoplasty due to reduced coagulation temperature and the insulation of the mucosa from tissue injury (19). Bipolar radiofrequency thermal ablation (bRFTA) produces an ionized saline layer that disrupts molecular bonds without using heat as a primary mechanism. At sufficiently high voltage gradients, the particles in the layer gain adequate energy to cause dissociation of the molecular bonds within structures. This induces volumetric removal of tissue using a thermal effect of 45–85°C, which is significantly lower than that in traditional radiofrequency techniques (20, 21). The possible advantages of bRFTA over conventional procedures include less pain, faster healing, and reduced postoperative care. Furthermore, the procedures can be performed under local anesthesia in the outpatient office rather than in the operating room. Our aim was to assess the efficacy and morbidity of bRFTA in patients with habitual snoring without significant desaturations associated with excessive daytime sleepiness and to evaluate the possible changes in different dimensions of the soft palate using magnetic resonance imaging (MRI).
This prospective, nonrandomized study was approved by the Research Ethical Committee of the Helsinki University Central Hospital. The patients had been referred according to their history to the ear, nose, and throat ward because of habitual snoring associated with excessive daytime sleepiness, and they were waiting for a surgical procedure of the soft palate. The preoperative evaluation included a physical examination and cephalometric radiographs. Fiberoptic nasopharyngoscopy with Muller maneuver was done when a direct visual estimation of the obstruction level was considered necessary. Each patient underwent nocturnal polysomnography to document severity of sleep-disordered breathing. For inclusion in the study the oxygen desaturation index 4% had to be below five, and the oxygen desaturation index 10% had to be below one. A maximum oxygen saturation decrease to 85% was allowed during total nocturnal sleep time. All patients with micrognathia and those with obstruction at the level of the posterior airspace of hypopharynx were excluded. Markedly obese patients (body mass index > 35 kg/m2) and patients with coagulopathy, psychiatric/neurologic disorder, or implanted pacemakers were excluded. Patients with a history of previous palatal surgery, speech or swallowing disorders and "too sensitive throats" were also not accepted. Treatment was delivered on an outpatient basis. We used the ENTec Coblator Plasma Surgery System with the ReFlex 55 wand. The bipolar electrode was inserted into the mucosa in the midline of the soft palate, 1 cm below the junction of the hard and soft palate, and two additional entries were made laterally on both sides. We used a power level of 6 (216–234 Voltage root-mean-square value) keeping the wand in position for 15 seconds. The patients were treated in two treatment sessions separated by 1 week. Each subject filled out visual analogue scales for 2 weeks for sore throat (0–100), swelling sensation of the palate (0–100), speech (0–100), and eating (0–100). The postoperative use of medication for pain during the 2-week period was registered using a different questionnaire. A slightly modified Finnish version of the Epworth sleepiness scale score (ESS) (22, 23) and a snoring score (SNRSs, answered by the patient) (0p = no snoring, 10p = heavy snoring, the bed partner leaves the room) questionnaires were filled before treatments, at 3 months, and at the median of 9.5 (range 9–11.5) months after the treatments. An additional snoring score (SNRSo) (24) was given to the bed partner. A successful treatment outcome was defined as a reduction of SNRSs from at least 7 to 0–3 points or a reduction of SNRSo from at least 6 to 0–3 points. An improvement of 3 points in the scores was considered significant. The relapses in the treatments were defined as a change of more than 2 points in the scores between 3 and 9.5 (9–11.5) months after treatment. MRI was done before and 3 months after treatment. To ensure the use of exactly the same measurement points in the MRIs, the different oropharyngeal dimensions in one session were measured by the same radiologist, before and after the treatments. Evaluations of changes in T1-signal intensity were done by two radiologists who arrived at a consensus value. The assessors were blinded to the condition of the subjects. Paired Wilcoxon's signed ranks test and Friedman's repeated measures analysis of variance on ranks were used to determine whether changes from the baseline to the final measurements were significant. Nonparametric Spearman's correlation coefficients were used. Linear regression analyses were used to study relationships between dependent and independent variables. Results are expressed as medians with 95% confidence limits and range, and they were generated using a computerized statistical package (SPSS 9.0, SigmaStat 2.0, Medstat 2.01). A p value below 0.05 was considered statistically significant.
Twenty patients, 19 men and one woman, fulfilled the initial inclusion criteria for enrolment in this study. One patient did not answer his questionnaires at all, and one did not submit his bed partners questionnaire results. However, their MRI results and visual analogue scales were included. The median age of the patients at the time of the first treatment was 43 (range 35–63) years. The median body mass index was 25.7 (95% confidence interval 24.0–27.0, range 22.6–34.6) kg/m2 and 26.2 (95% confidence interval 24.0–26.0, range 22.6–34.6) kg/m2, 9.5 (range 9–11.5) months after treatment; the change was not significant (p = 0.789). The questionnaires on ESS (n = 19), SNRSs (n = 19), and SNRSo (n = 18) showed a significant change, indicating decreased snoring and daytime sleepiness (Figures 1 and 2) . Significant improvements, successful outcomes, and relapses, according to our criteria are shown in Table 1 . The changes in the SNRSs and SNRSo questionnaires from before treatments to 3 and 9.5 (9–11.5) months after the treatments correlated significantly (SNRSs versus SNRSo, 3 months: r = 0.792, p = < 0.001; SNRSs versus SNRSo, 9.5 months (9–11.5): r = 0.529, p = 0.029). In the regression analyses, the changes in the SNRSs and SNRSo questionnaires from before treatments to 3 and 9.5 (9–11.5) months after the treatments were of the same kind, indicating similar subjective improvement by the patient and the bed partner. SNRSs explained 70.2% of the variance of the SNRSo at 3 months (R2 = 0.702, F = 15.561, p = 0.001) and 54.5% of the variance at 9.5 (9–11.5) months (R2 = 0.545, F = 17.988, p = 0.001).
The second MRI was performed 13 weeks after the first treatment. There was a statistically significant change in the distance between the tip of the uvula and the bottom of the sella, indicating a retraction of the uvula (p = 0.005). However, the length of the soft palate was not significantly shortened (Table 2) . The change in the T1-signal intensity, reflecting fibroid scar formation of the soft palate showed major changes in nine cases, moderate changes in four cases, and no/mild changes in six cases. These changes did not correlate significantly with the changes in the questionnaires before and 3 months after the treatment (T1-signal–change versus SNRSs-change, 0–3 months: r = -0.196, p = 0.421; versus SNRSo-change, 0–3 months: r = -0.156, p = 0.536; versus ESS-change, 0–3 months: r = 0.064, p = 0.795). In a linear regression analyses, SNRSs-change explained only 2.3% of the variance of the T1-signal–change. There was no relationship between SNRSs-change or ESS-change and T1-signal–change (SNRSs: R2 = 0.023, F = 0.404, p = 0.534, SNRSo: R2 = 0.005, F = 0.083, p = 0.777, ESS: R2 = 0.011, F = 2.633, p = 0.699). The T1-signal intensity before the treatments was not a prognostic factor for a successful treatment outcome (T1-signal intensity versus SNRSs-change, 0–3 months: R2 = 0.116, F = 2.222, p = 0.154; versus SNRSs-change, 0–9.5 [9–11.5] months: R2 = 0.007, F = 0.111, p = 0.793; versus SNRSo-change, 0–3 months: R2 = 0.072, F = 1.241, p = 0.282; versus SNRSo-change, 0–9.5 [9–11.5] months: R2 = 0.000, F = 0.007, p = 0.933). There were no significant changes in the other dimensions of the soft palate (Table 2, Figure 3) .
All the patients tolerated the bRFTA well and were willing to undergo additional treatment if necessary. The most common complaint was swelling of the soft palate, which gradually decreased during the 2-week follow-up period. One patient was given corticosteroids because of the swelling. Four patients received antibiotics, three of them after the first week and one after the second week, because of development of a small abscess in the insertion point of the hand piece. Pain scores were the highest on the night after the treatment and then gradually decreased during the 2-week follow-up period. Minor transient changes in speech and swallowing difficulties were encountered (Figure 4) . The use of medications for pain was low, with a median of two (95% CL 1–7, range 0–27) dosages during the 2-week period.
Mucosal blanchings 5 and 10 mm in size were developed by six and two patients, respectively. The blanchings correlated with the use of medications for pain (r =0.882, p = 0.036) but not with the use of antibiotics (r =0.357, p = 0.882). However, in a linear regression model the occurrence of blanchings did not cause a significant increase in the use of medications for pain (R2 = 0.030, F = 0.553, p = 0.467) or in the need for antibiotics (R2 = 0.128, F = 2.633, p = 0.122). The postoperative morbidity, i.e., visual analogue scales scores, showed no significant relationship with the occurrence of blanchings after the treatments (data not shown). There were no signs left of these blanchings at 3 months after treatment. Four patients complained of a swelling sensation in the throat at 3 months after treatment, two of them had the sensation left at 9.5 (range 9–11.5) months after the treatments.
Radiofrequency thermal ablation is a minimally invasive treatment. It leads to necrosis and fibrosis with contracture of tissue without actual removal (25). The bRFTA induced objectively detectable changes in certain dimensions and in the T1-signal intensity of the soft palate. Simultaneously, there was an advantageous outcome regarding snoring and daytime somnolence. However, there was no statistical relationship between the results of the questionnaires and the MRI analyses. The degree of the T1-signal intensity 3 months after treatment reflects, most probably, scar tissue formation, supporting the theoretical background of radiofrequency treatment. We presume that the increased rigidity of the palate caused by tissue scarring is the basis of the relief from symptoms and improved sleep quality. Thus, bRFTA seems more reasonable compared with the removal of the noise-causing vibrating portion of the snorers soft palate. MRI provides indirect information about the changes in the stability of the soft palate. The soft tissue resolution capability is superior when using MRI; it avoids radiation exposure and yields both transverse and sagittal sections of the upper airways. Our hypothesis was that MRI would be the only radiologic way to find possible objective changes caused by bRFTA. To our knowledge, there is no previous report using MRI on treatment evaluation of habitual snorers. In the evaluation of the upper airways in patients with sleep-disordered breathing, MRI has been shown to be a useful research tool in two review articles (26, 27). In addition, as shown in our previous report, the cephalometric measurement of the palate length was not a good parameter to determine the treatment outcome (16). Snoring is first and foremost a subjective perception by a listener. There is no general agreement on how snoring should be assessed. Neither are there are any standard measurements of snoring in the sleep laboratory. In addition, other qualities of snoring, apart from the loudness of snoring, such as its frequency, may contribute to the annoyance caused by snoring (28). Therefore, in most studies reliance is placed on subjective assessment. Assessment of snoring by self-report is potentially limited because persons naturally are unaware of their behavior during sleep. Many clinicians recommend that spouses should be routinely queried concerning the sleep-related symptoms of their partners. The overall agreement between self- and spouse-related symptoms has been found to be generally high. Furthermore, patients who report that they snore habitually do seem to snore objectively (29). However, the results of the questionnaires may be misleading in part, which can affect estimates of treatment effects (30). In addition, the assessment tools based on self-report can be influenced by the placebo effects. ESS was chosen to evaluate daytime sleepiness. The pretreatment scores of our study group (median 4.5, range 1–16) did not clinically significantly differ from the ESS scores of snorers as reported by Johns (22, 23). The patient himself evaluated SNRSs to obtain subjective assessment of snoring. SNRSo was given to the bed partner, and it has been shown to be a uniform, comprehensive, and simple-to-understand method for the evaluation, grading, and reporting of snoring (24). The subjective and successful treatment outcomes according to these questionnaire results (snoring and daytime somnolence) from the bRFTA were satisfactory. Our results are in agreement with the results previously reported using monopolar radiofrequency energy (15, 16, 31). The short-term results of the present study are compatible with those of uvulopalatopharyngoplasty (32), laseruvulopalatoplasty, (33), and laser-assisted uvulopalatoplasty (34). We chose our treatment regimen for three reasons. First, we had positive experiences previously with a 1-week interval between treatments (16). There were no refusals, and the treatment results were promising. Second, the intention of the second treatment is to induce more scarring on regions that are intact after the first procedure and subsequently achieve a larger ablation area. Third, with this regimen, we had no dropouts, which is an important factor in the statistical calculations. Two treatment sessions separated by 1 week is not standard treatment with bRFTA, and longer intervals between the treatments are justifiable. Histopathologic studies reveal that collagen is not deposited in sufficient amounts for at least 10–12 days after the monopolar radiofrequency tissue injury, and spacing the treatment session farther apart might allow more collagen formation and more tissue contraction (25). To our knowledge, histopathologic studies with bipolar radiofrequency tissue injury have not been done, and the possible farther spacing of the treatments cannot be evaluated. The safety and effectiveness of bRFTA have been proven through its extensive use in arthroscopic surgery (35). bRFTA has been adapted to operate in dry field environments, and the first clinical applications have been in cosmetic surgery (36). None of our patients reported bRFTA channeling to be a painful procedure. Almost all the patients experienced restless sleep on the first night after treatment because of mild discomfort and feeling of swelling of the soft palate. Only mild or moderate speech or swallowing problems were reported with bRFTA. The overall postoperative discomfort was low, even though there were a few blanchings. The occurrence of blanchings did not increase the postoperative discomfort, which is in accordance with the studies of Pazos and Mair (37) and Terris and Chen (38). bRFTA treatment with the Coblation hand piece, ReFlex 55, caused more blanchings than we experienced using monopolar Vidamed (Vidamed Inc., Fremont, CA) (16); the same has been reported in treatments using monopolar Somnoplasty (Somnus Medical Technologies Inc., Sunnyvale, CA) (15, 39). There were also four patients who needed antibiotics during their recovery because of the development of a small abscess in the insertion point of the wand. This complication has not been reported previously. The size of the hand piece used in the Coblation technology is of a greater than that used in Vidamed or Somnoplasty, but the treatment time was only 15 seconds per site, which the patients accepted easily. The ongoing acute and chronic inflammatory response, due to our treatment regimen, may contribute to an increased risk of infection. It cannot be evaluated in this study how great a role these blanchings and need of antibiotics had on postoperative morbidity. It appears that the efficacy of bRFTA treatment and that of monopolar radiofrequency energy treatment, decreases with time as with other surgical procedures of the soft palate. Scar tissue matures and softens over 12–18 months, which might explain the relapses (40). However, the relapses after monopolar radiofrequency energy treatments can be reduced with additional treatments (41). The minimal invasiveness of this treatment modality provides a high acceptance of retreatment, which was also shown in this study. In our series, additional bRFTA channeling is scheduled for those patients with relapses or unsuccessful treatment outcomes; we are going to continue our follow-up and will report the results. In conclusion, bRFTA of the soft palate in patients with sleep-disordered breathing on an outpatient basis seems to be an effective treatment for habitual snoring in patients without significant desaturations associated with excessive daytime sleepiness. The MRI showed that the procedure caused significant changes in the dimensions of the soft palate and in the T1-signal intensity. The advantageous outcome in the questionnaires showed no association with the results of the MRI analyses. However, the positive result of the intervention is obviously due to increased rigidity of the palate caused by the tissue scarring. bRFTA is safe and associated with low morbidity. If future studies on larger patient samples show similar efficiency, bRFTA might well be a good, minimally invasive alternative for the treatment of patients with habitual snoring. The promising results need to be confirmed in a randomized, controlled study.
The authors thank ENTec® (division of ArthroCare Corp., local representative Canella Medical OY, Espoo, Finland) for allowing them to use the Radiofrequency generator free of charge.
Supported by the Helsinki University Central Hospital Research Fund and the Instrumentarium Research Foundation, Helsinki, Finland. This article has an online data supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Received in original form April 26, 2001; accepted in final form April 30, 2002
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
