A variety of surgical procedures exist to treat obstructive sleep apnea in adults. Some, such as tracheotomy and maxillomandibular advancement, have very high cure rates, over 90%. These procedures have significant disadvantages, however, and there is a need to define the best combination of low morbidity procedures to provide similar success. To do this, better means to diagnose the sites of obstruction must be used and new procedures must be developed. This article reviews where we are in reaching these goals.
Currently, there is no ideal surgical procedure for obstructive sleep apnea (OSA), including the upper airway resistance syndrome (UARS). Surgical treatment of snoring in adults without evidence of obstruction has greater success; several reasonably successful effective procedures are available for office treatment of snoring, providing improvement or cure for the majority of patients treated. Although not yet ideal, they are closer to that goal than the procedures available for OSA.
Snoring without obstructive sleep apnea
For snoring alone, the ideal operation should be one stage, performed in the office under local anesthesia without sedation, and eliminate snoring in 90% or greater of cases. Furthermore, complications should be few and minor in nature. Postoperative pain should also be minimal and well-controlled with the usual oral pain medications. Pain medication should be required only for a few days. Because the procedure is not reimbursed by the usual health insurance policies, the costs should be reasonable. With an ideal operation, the results should be long-lasting, and any failures could be retreated with the expectation of a similar high incidence of success. This would bring the overall success rate to near 100%, a laudable goal. Uvulopalatoplasty (UPP) and uvulopalatopharyngoplasty (UPPP) are effective operations to reduce or eliminate palatal snoring, but are usually done in an operating room under general anesthesia to try and correct OSA associated with the snoring. The uvulopalatal flap is a variation that has the potential to be reversed if needed; this reversal is rarely indicated.
In one review of eight studies of UPPP for snoring , 29.8% of patients had no snoring and 43.1% had reduced snoring, for an overall “success” rate of 72.9%.
Procedures that come closest to achieving this ideal are discussed in the following sections.
Soft palate tightening
Soft palate tightening procedures use radiofrequency (RF) energy delivered submucosally into several sites by a needle electrode .
This consists of injection of a sclerosing solution into the palate to produce scarring and soft palate stiffening . The resultant stiffening or shortening of the palate prevents excessive vibration of the posterior edge of the soft palate, thought to produce the snoring noise in the majority of cases.
This procedure and the soft palate tightening procedure can be performed in the office under local anesthesia, without the need for sedation. Postoperative pain is usually mild and the complications minor. The cost of palate injection is extremely low; the cost of RF treatment is higher, and may require purchase of a disposable RF needle electrode, an additional expense. In addition, an RF generator is needed, and this cost must be amortized, adding to the expense. Over time, the effectiveness of the RF treatment decreases, and it may need to be repeated; the same will probably prove true for injection snoreplasty.
Uvulectomy in the office using an RF snare, standard cautery, or CO2 laser appears to be similarly effective, but has greater pain in the postoperative period, lasting a week or more in most patients . The same is true for extended uvulectomy procedures, such as UPP, that remove more of the free edge of the soft palate. These procedures may have the long-term complaint of a feeling of “mucous” or dryness on the nasopharynx side of the palate border; procedures that spare the free edge of the palate do not produce this symptom.
Laser assisted uvulopalatoplasty
Laser assisted uvulopalatoplasty (LAUP) is a multistage office procedure that requires a CO 2 or equivalent laser, but does not need a new handpiece for each treatment . Because the cost of the laser must be considered, the expense of the procedure is necessarily higher than nonlaser procedures. The associated pain is greater than with RF submucosal tightening, and the need for multiple procedures is a deterrent. Some surgeons perform LAUP as a one-stage procedure, which produces greater pain but eliminates the need for multiple office visits.
Stiffening the soft palate by insertion of three or more small barbed implants has been found to decrease snoring . It can be performed quickly in the office under local anesthesia. The major complication is postinsertion extrusion. One disadvantage is the relatively high cost of the implants.
Other office surgical procedures
Other procedures designed to shorten and tighten the palate, such as the cautery-assisted palatal stiffening operation (CAPSO), appear equally successful for snoring . All of these procedures seem to have some effectiveness in the treatment of mild OSA and UARS caused by obstruction at the palate level; over time, their success percentage decreases.
The role of nasal obstruction
It is unfortunate that opening an obstructed nose with standard nasal procedures is not regularly effective in improving snoring, although it should improve nasal breathing, including nocturnal nasal breathing. There are several reports that show subjective improvement; a review of eight papers found the average rate for cessation of snoring to be 41.9%, and reduction of snoring 85.3%, following correction of nasal obstruction . In patients who have nasal obstruction, whether just at night or both day and night, opening the nose is appropriate, independent of any effect on snoring. Assuming a straight nasal septum, a short trial of a long-acting, topical, alpha adrenergic spray, such as oxymetolazine, at bedtime to shrink the turbinates overnight will provide information to the surgeon and the patient regarding the role of nasal obstruction in producing snoring. If there is adequate snoring improvement, an office procedure to decrease the size of the inferior turbinates can be performed. Procedures to decrease the size of the turbinates can be performed at the same time as procedures that tighten or shorten the soft palate. Submucosal RF cautery to the turbinates to open the nose can be performed at the time the soft palate is treated in the office, and is regularly effective . Correction of symptomatic nasal septal deflections is usually performed in an operating room, commonly in combination with turbinate reduction. No information exists on the effectiveness of this combination of nasal and palate surgery on snoring or sleep apnea versus a soft palate procedure alone; however, it makes sense that opening an obstructed nose should make the patient feel better.
A trial of a device to open the nasal valves at night can be used in the same way as a nasal spray to shrink the turbinates. They can be used in combination to determine if collapse of the valves, turbinate engorgement, or both are contributing to nocturnal nasal obstruction, which results in snoring.
Several devices are available, sold over the counter, to open the nasal valves at night. Breathe Right strips (GlaxoSmithKline, Pittsburgh, Pennsylvania) taped to the lower nose at night are the best known, and some data indicate that they help . Such devices can be used indefinitely, or an operation to open the nasal valve area can be performed , either separately or in combination with inferior turbinate reduction or septoplasty.
Evaluation of snoring
Unfortunately, the nature of snoring is such that it is almost impossible for us to have hopes for an ideal snoring operation. There are three reasons for this, two which we can address and attempt to solve; the third is more difficult. The first is a diagnostic issue as to the exact site of the sound produced by the snorer. It is generally assumed that the majority of snoring comes from vibration of the soft palate free edge. The data for this are not as clear as we would like. Certainly, if the loud snoring that regularly accompanies OSA and upper airway resistance syndrome (UARS) is caused by the palate, it would appear logical that nocturnal airway obstruction has the same cause. This is usually not the case, with the tongue base or hypopharnx being a major component, with or without palate level obstruction. The nose can produce noise at night because of narrowing, and possibly the presence of thick mucus. This nasal snoring group are the patients whose snoring improves with nasal surgery. In addition, nasal obstruction leads to mouth breathing, and the presence of an open mouth during sleep increases the sound intensity of the snoring. Although it is possible to snore with the mouth shut, the sound levels achieved are not as great.
Some snorers snore during the exhalation stage rather than while inhaling, and some snore on both inhalation and exhalation; the significance of this in regard to treatment is unclear. Certainly, there are snorers in whom the tongue is a component of the snoring noise, and this may be the sole component after the palate has been shortened or tightened. Basically, our inability to identify the site or sites of the snoring noise hampers our ability to provide an appropriate operation. To operate on the palate of a snorer whose tongue is the major reason for the snoring makes no sense, but is probably done all the time.
The second problem is that even if we are sure that the snoring noise is coming from soft palate vibration, we do not know how much palate removal or tightening is required in a given case to totally eliminate the snoring. Because we do not want to produce velopharyngeal insufficiency (VPI), there is a limit to what can be done to surgically correct snoring from this site. This is common, as evidenced by the fact that snoring is rarely completely eliminated after UPP.
We need better tests to determine the site of the snoring, so that the effect of palatal stiffening or shortening can be evaluated before any surgery. Acoustic analysis of the snoring sounds has been performed by several investigators to assess the site, but is not used routinely in the pretreatment evaluation of snoring . Many home and hospital-based polysomnography systems include a measurement of the intensity and duration of snoring, which is helpful but not enough. One method of home analysis, snoring and apnea analysis (SNAP), claims to analyze the snoring sound so the sites of the snoring can be determined using a proprietary algorithm . It does provide objective information on snoring duration, loudness, and frequency. With practice, one can diagnose palatal snoring in the office by asking the patient to produce a snoring sound. The low frequency “motorboat” noise is distinctive when present. The presence of the bed partner during the examination is useful to confirm that the sound produced in the office is the same as that during sleep (except possibly louder).
Another approach is the temporary immobilization of the free edge of the palate and uvula, to validate the potential effectiveness of a palate procedure in eliminating snoring. The simplest test is to temporarily attach the uvula forward onto the soft palate, using a suture, to evaluate the effect on snoring for 1 or 2 nights, releasing the temporary attachment after the evaluation. The author and associates have used this technique on occasion to help predict the effect of soft palatal tightening/resection on snoring; it could be used to determine the effect on the respiratory disturbance index (RDI) in cases of OSA. In this case, the temporary attachment would be performed just before a nocturnal polysomnogram (PSG). More work is needed in this area.
The third problem in achieving an ideal snoring operation is that the evaluation of the snoring result is done by the bed partner, so that no matter what the objective data are, if the sleep partner says that snoring is still objectionable, the operation is considered a failure. Although forms and visual analog scales are available for the bed partner to fill out to analyze snoring, it is still difficult to objectively evaluate “improvement” in snoring. Anyone who has slept with a snorer knows it only takes two or three episodes to wake one up in the middle of the night, become angry, and move out of the bed (or make the snorer move out). There are few areas in surgery where the evaluation of success or failure of a procedure is not by the patient, but by an individual who the surgeon may not even know. What is the relationship between the snorer and the sleep partner? Is it hostile? Is the sleep partner an insomniac who wakes up at the slightest noise and then cannot go back to sleep? It is important that whenever possible, the surgeon meets the bed partner to assess the relationship before any snoring surgery.
Several nonsurgical snoring remedies are available. Three are mentioned:
Ear plugs. This solution is rarely effective, because the over-the-counter ear plugs usually tried are either uncomfortable, ineffective, or prevent awareness of important nocturnal sounds, such as a baby crying. Custom-fitted, deep, soft ear plugs are more likely to be successful.
Chin supports. Adjustable, elastic bands that go around the chin and vertex of the skull have been available for many years as a treatment for snoring. Termed “snoods,” they help by preventing the mouth from opening and the jaw from falling back. Soft, orthopedic neck collars have been used with some success to maintain mouth closure and an elevated chin, thus reducing snoring (and OSA). An open nose is essential.
Dental devices hold the teeth together during sleep, and can also advance the lower jaw forward. The most successful are those custom-fitted by a dentist.
Obstructive sleep apnea
The diagnosis of OSA is made with a PSG, and treatment initiated using nasal continuous positive airway pressure (CPAP). If CPAP cannot be tolerated, surgical treatment of OSA should be considered.
The problem of determination of sites of obstruction is a major issue in the surgical treatment of OSA and UARS. Preoperative assessment in the awake patient of the location of the obstruction during sleep is simply not good enough at this time. Physical examination, Mallampati or Friedman evaluation, lateral cephalometric radiographs, and fiberoptic endoscopy with Mueller maneuver all have limitations as to determining the site or sites of obstruction in a given patient. The OSA patient ideally needs to be evaluated during sleep.
Diagnosis of site
In the author’s opinion, the best objective test to determine the site of obstruction is multilevel pressure measurements during sleep . This is not routinely performed for a number of reasons. Sleep laboratories have no incentive to do such testing, because nasal CPAP corrects obstruction at all levels; site of obstruction determination is of interest primarily to the surgeon. Current reimbursement schemes would need to be modified to pay for such additional testing; sleep laboratories would need to be incentivised by surgeons to perform such tests, or the surgeons would need to support specialized testing systems that provide these data. Esophageal pressure measurements during sleep have allowed us to diagnose UARS; a single pressure measurement in the upper esophagus reflects the abnormally negative interthoracic pressure needed for inspiration during sleep. A similar test using multilevel pressure sensors would provide us with a measurement of obstruction at the palate and tongue base.
The use of artificial sleep using intravenous medications while the surgeon looks at the potential areas of obstruction with a fiberoptic endoscope (sleep endoscopy) is used routinely in several centers outside of the United States. It is expensive and time-consuming, unless performed at the time of surgery for OSA. This is not routinely done in the United States, and there are questions as to whether sleep produced by intravenous drugs is the same as normal sleep. Volume CT and MRI evaluations during artificial sleep also have limitations, with cost and risks similar to sleep endoscopy.
Definition of successful treatment
“Success” in surgical treatment is usually defined as meeting the criteria used by Sher and colleagues in their 1996 review : 50% improvement in the RDI, with a decrease in the RDI to below 20, or the apnea index (AI) to below 10. These objective criteria of success are less than ideal, because the normal RDI is 5 or less.
We must keep in perspective, however, that although nasal CPAP is more likely to come closer to the normal values, compliance and regular use is a major problem. Is reaching an RDI of less than 10, 60% of the sleep time better than reaching a RDI of less than 20, 100% of the sleep time?
The end result of eliminating excessive daytime sleepiness is determined by the patient. This may not always correlate with the results of a PSG, the standard to diagnose and quantitate OSA. How do we score a case in which the PSG shows marked improvement of the sleep apnea, whereas the patient reports no improvement, or may even claim he is worse? What about the reverse, when the patient is delighted at how she feels, but the PSG shows minimal change? Certainly, in explaining why some patients do not feel much better despite improvement in the PSG, the answer is that we have probably simply moved the patient from OSA to UARS, or there is another cause of the excessive sleepiness.
Surgical treatment: suspected palate level
In adults, UPPP appears to an appropriate and successful operation for the 25% or so of OSA cases in whom the obstruction is limited to the soft palate level, assuming we know who these patients are. It should also help, but to a lesser extent, the additional 25% of patients in whom the palate is a contributor but not the sole cause of the OSA. The overall result yields the 40.7% “success rate” reported with UPPP surgery alone, using the criteria described by Sher and colleagues . By eliminating patients who on clinical evaluation appeared to have retrolingual obstruction as the primary cause (large tongue, retrognathia, small or absent tonsils), postsurgical UPPP success was higher (52.3%) . The opposite is also true: UPPP in the suspected retrolingual group provides essentially no improvement in OSA. The morbidly obese (BMI>33) are not good UPPP candidates either.
Sleep endoscopy should provide improved UPPP results, because it can better identify the sole or significant soft palate etiology OSA patients who are the best candidates. In general, this is true; however, the difference is not as much as desired . Why is this? Although the answer is not totally clear, the author believes it is twofold: (1) analysis, even with this excellent method, is still a work in progress, and (2) UPPP, as usually performed, does not adequately correct palate obstruction in a sizeable percentage of cases. More aggressive approaches to palatal shortening may provide better postoperative results, but have longer recovery periods and are more involved. These include palatal Z-plasty and transpalatal advancement pharyngoplasty .
Combining preoperative sleep endoscopy or multilevel pressure measurements to confirm a palatal site of obstruction and a more aggressive palate shortening procedure should result in improved surgical results for OSA patients who have a palate cause. At this time, UPPP should not be the sole surgical treatment for the usual case of OSA—it is not good enough.
Other palatal shortening/tightening procedures
The palatal procedures previously described for snoring can improve (and even cure) OSA, particularly mild OSA. Overall, none stand out as a routine procedure for the usual case of OSA. Several are definitely less morbid than UPPP and can be done in the office, a major cost and convenience advantage.
In children, tonsillectomy and adenoidectomy have provided successful treatment of OSA in the majority of patients, because the obstruction is usually caused by enlarged tonsils or adenoids . In adults, the presence of large tonsils makes it more likely that a UPPP and tonsillectomy will be successful.
Surgical treatment: tongue base level suspected
These treatments are regularly combined with an UPPP, even when tongue base obstruction is suspected as the primary cause. They are really multilevel therapy.
The success of multiple RF treatments to the tongue base, in a review of 11 series, ranges from 20% to 83%, using the usual surgical success criteria . The usual number of insertions of the RF needle is four at one time, 750 J each site. Multiple sessions, weeks apart, are required. An average of 5.5 sessions was required in one series to produce a success rate of 46.7% . This is not adequate for the amount of treatments required, in the author’s opinion. Other RF protocols are being evaluated and show promise.
In combination with UPPP, geniotubercle advancement (GTA) produces a success rate of 39% to 70% .
Again, usually used in combination with UPPP, and in some series, GTA. By itself (but with UPPP) it appears to provide little advantage over UPPP alone in the author’s experience (17% success) . Others have found better results (53.3%) .The latter study used preoperative sleep endoscopy to better define the sites of obstruction. When combined with a GTA, there may be an advantage over a GTA alone, but this is not clear.
Suture suspension of the tongue (repose procedure)
Again, combined with UPPP, suture suspension of the tongue has a similar success rate to GTA, with a success range of 20 to 57% in six reports .
Maxillomandibular advancement (MMA), the current “gold standard” for OSA, has a success rate of 90% or greater , but it is technically demanding and has substantial morbidity and expense.
Tongue base resection (midline glossectomy)
These procedures can be effective, but have a high morbidity, particularly pain and postoperative dysphagia. Results vary from 25% to 83% success . Removal of large lingual tonsils is a different matter, and can be an effective OSA treatment, particularly in children.
Still very effective, but hard to convince patients to accept the side effects.
Bariatric surgery can be effective in reducing the RDI in grossly obese patients. Mandibular distraction in children who have poor mandibular development may be curative in selected cases. The use of a dental splint to further advance the lower jaw after failure of a GTA, repose, or RF tongue procedure may provide success after failure of the surgical procedure .