1 CONCEPT OF MULTILEVEL TREATMENT – ESTABLISHING THE BASIS

1.1 INCIDENCE OF MULTILEVEL OBSTRUCTIONS IN OSAHS PATIENTS

The true incidence of multilevel obstruction is a subject of much debate. Fujita ¹ first described different anatomic levels of obstruction in OSAHS. He recognized that half of the patients who underwent uvulopalatopharyngoplasty (UPPP) were non-responders. Most of the non-responders were identified as having multilevel obstruction. Combined oropharyngeal and hypopharyngeal obstruction was noted in 54.5% (36/66) of patients in his study. Thus, it is clear that Fujita himself never intended to suggest that UPPP will cure most patients with OSAHS. In 1993, Riley et al.² reported their surgical experience, outlining a multilevel concept. Each patient was classified as having single level obstruction involving oropharynx only – type 1 – or the hypopharynx only – type 3. Multilevel obstruction was identified as type 2 and implied a combination of oropharyngeal and hypopharyngeal obstruction. Of the 239 patients, 93.3% (223 patients) were identified as having multilevel obstruction, type 2. Only 16 patients (6.7%) had single level obstruction. Of these, 10 patients had type 1 obstruction and six patients had type 3 obstruction.

This early classification by Fujita and Riley was based on physical examination of the patients with vague guidelines. Specific criteria for identifying unilevel versus multilevel obstruction were not reported. Subsequent development of the Friedman tongue position (FTP) allowed for a simplified method of staging the levels of obstruction.^3,⁴ The early data based on FTP indicated that approximately 25% of patients presenting with OSAHS had unilevel obstruction, while 75% had multilevel obstruction.

To more precisely identify the anatomic sites of obstruction during sleep, sleep endoscopy was proposed as a preferred method. den Herder et al.⁵ reported an unusually high number of single level obstructions. In their study of 127 patients, 63% had single level obstruction while only 37% had multilevel disease. The study, however, may have misidentified the level of obstruction; tongue base obstruction pushing the palate backward causing secondary palatal obstruction may have been classified as primary palatal obstruction.⁶ Another study by Abdullah van Hasselt confirmed the high incidence of multilevel disease and 87% of their 893 patient populations had multilevel obstruction.⁷

Many other techniques to identify levels of obstruction have been reported. The most popular are the Mueller maneuver and cephalometry. Both of these are used to identify the levels of obstruction. Other diagnostic aids such as sleep MRI have also been reported. The other techniques are reviewed in other chapters.

1.2 SINGLE LEVEL SURGERY CANNOT BE THE ONLY TREATMENT FOR MOST OSAHS PATIENTS

UPPP was designed to enlarge the oropharyngeal airway ⁸ and remains the most common surgical intervention for OSAHS. Review of the literature indicates the success rate (a 50% reduction in Apnea/Hypopnea Index (AHI) and a postoperative AHI of less than 20) for UPPP as an isolated procedure ranges from 25% to 80%. However, Sher et al.⁹ reported success rates for UPPP around 41% based on a meta-analysis of unselected cases. Although there are many reasons UPPP may fail, uncorrected retrolingual obstruction has been clearly identified as a major cause. This seems to support the concept that multiple levels of obstruction exist and explains why UPPP alone frequently results in failure.

1.3 MULTILEVEL SURGERY IS NOT LIMITED TO SEVERE DISEASE

Many otolaryngologists presume that although UPPP may not cure the patients with severe OSAHS, it is likely to be effective for patients with mild disease. There are however many studies indicating that the severity of disease is not a predictor of success with single level surgery.^3,^4,¹⁰ Senior et al.¹¹ studied a group of patients with mild OSAHS (AHI less than 15). These patients underwent UPPP and the success rate was only 40%. Friedman further studied a series of patients with mild disease and showed an overall success rate of approximately 40% as well.^3,⁴ If indeed most patients have multilevel disease, the success for the surgical treatment of mild OSAHS is not better than those for treating severe disease. In fact, the basis of the Friedman staging system is that anatomic findings are the most significant factors, rather than severity of disease. Multilevel surgery should not be reserved exclusively for the treatment of severe disease. It is therefore reasonable that multilevel treatment should be considered for most patients with mild disease as well as most patients with severe disease. Since the majority of patients suffering from OSAHS do have multilevel disease and directing the therapy to a single anatomic level has a high potential for failure, the need for multilevel therapy is evident.

2 HISTORY OF MULTILEVEL TREATMENT

Historically, surgical treatment for OSAHS was often based on trial and error. Patients would invariably undergo UPPP as a first stage. If the disease was not eliminated, they would go on to have hypopharyngeal surgery. Planned multilevel surgery at a single phase however has now become standard in many centers.

Published data on multilevel treatment can be divided into four groups.

1. The most commonly performed multilevel approach includes a UPPP as a basic technique with a second procedure designed to improve the hypopharyngeal airway. Most commonly this includes genioglossus advancement, thyrohyoid advancement, radiofrequency tissue volume reduction of the tongue base, and in some cases tongue base suspension.¹²^–¹⁵ The success rate for these procedures has been reported to be between 20% and 100% and was based on retrospective studies on small groups of patients. The largest series reported by a single group was by Riley et al.² who studied 239 patients who underwent what they describe as phase I surgery. In their study, 223 patients (93.3%) underwent multilevel surgery (UPPP + GAHM) in the initial phase. Their success rate based on single stage multilevel treatment for patients with mild, moderate and severe OSAHS was 60%.

2. The second group of patients studied who have undergone multilevel treatment include those who have had more invasive and more radical hypopharyngeal surgery such as open tongue base resection. Because of the aggressive nature of these procedures, most of these patients had a temporary tracheotomy and required significant hospitalization. There was significant postoperative morbidity as well. The success rate in this group varied between 44% and 100%.^2,¹⁵^–¹⁸

3. The third group of multilevel surgery for OSAHS included those patients undergoing bimaxillary advancement as part of the multilevel treatment program. Most of these patients had undergone a staged surgery often, with UPPP and genioglossus advancement as their primary procedure. This group was not included in the overall discussion and recommendation for treatment in this chapter.

4. ‘Multilevel minimally invasive treatment for mild/moderate OSAHS’ will be discussed in the next section. This section also lists minimally invasive as well as invasive procedures to address obstruction at the level of the nose, oropharynx and hypopharynx.

3 MULTILEVEL MINIMALLY INVASIVE TREATMENT FOR MILD/MODERATE OSAHS

The ideal procedure for OSAHS patients would have low morbidity, allow reasonable success, have low risk of alteration of the original upper aerodigestive tract function, and could be performed in a single stage. In our experience, patients with severe disease and obvious daytime sleepiness often have strong motivation for multilevel surgery with invasive procedures to address the obstructed sites. However, patients with mild or moderate obstructive sleep apnea are less likely to be willing to undergo aggressive surgical procedures.

Many options are available for minimally invasive treatment of palatal obstruction. Minimally invasive treatment options for retrolingual obstruction are limited, but have been in clinical use. There have been a few studies that looked at multilevel minimally invasive treatment. Steward ¹⁹ studied 22 patients who underwent combined radiofrequency reduction of the palate and the base of tongue and reported a success rate of 59%. None of the patients had concomitant nasal surgery. Fischer et al. presented a similar study about multilevel minimally invasive surgery with radiofrequency on the palate, tonsil and tongue base for 15 OSAHS patients.²⁰ The results showed that the AHI changed from 32.6±17.4 preoperatively to 22±15 for a success rate of 33%. Whereas the majority of Fischer’s patients had moderate disease, Steward’s patients had mild-to-moderate disease.

In 2004, Stuck et al.²¹ published their surgical results with radiofrequency on the palate and base of the tongue for 18 OSAHS patients with mild/moderate disease. They reported their success rate as 33.3% (by their definition of success, a postoperative AHI reduction of at least 50% and below a value of 15).

In 2007, we presented our experience on minimally invasive single-stage multilevel surgery (MISSMLS) for patients with mild/moderate OSAHS.²² Our patients underwent three-level treatment that included nasal surgery, palatal stiffening by pillar implant technique and radiofrequency volume reduction of the tongue base with a minimum follow-up of 6 months. The results in a retrospective review of a prospective dataset showed that both subjective and objective outcomes improved significantly, and treatment morbidity was minimal and only temporary. Mean bedpartner’s snoring visual analog scale (VAS) decreased from 9.4±0.9 preoperatively to 3.2±2.4 postoperatively. Epworth Sleepiness Scale (ESS) decreased from 9.7±3.9 preoperatively to 6.9±3.3 postoperatively. Mean AHI was reduced from 23.2±7.6 to 14.5±10.2. Classic success was achieved in 54 of 122 patients (47.5%).

Although there was a significant failure rate, patients that failed were candidates for additional procedures. The first stage procedures that were performed including pillar palatal implantation and radiofrequency tongue base tissue volume reduction in no way negatively affected their ability to undergo secondary UPPP or additional tongue base procedures. Thus, many of the patients in our study who failed went on to undergo classic UPPP combined with additional hypopharyngeal procedures. Presence of pillar implants in the palate did not affect the UPPP in any way that could be determined by the surgeons. Patients who underwent previous radiofrequency tongue base reduction were still candidates for additional radiofrequency treatment or other hypopharyngeal procedures. This MISSMLS approach resulted in a fairly high patient satisfaction rate even though many patients required secondary procedures.

MISSMLS offers reasonable improvement of both subjective symptoms and objective PSG findings in patients with mild/moderate OSAHS. Polysomnographic respiratory parameters, ESS and snoring VAS show statistically significant improvement in patients with mild/moderate OSAHS treated with MISSMLS.

The success rate of any protocol has to be weighed against associated morbidity and risks. The relatively low morbidity combined with reasonable success makes multilevel minimally invasive surgery for OSAHS worthwhile. It should be offered to patients when multilevel obstruction is identified in the context of mild/moderate disease and other non-surgical options have been excluded.

4 SYSTEMATIC REVIEW OF LITERATURE ON MULTILEVEL TREATMENT FOR OSAHS PATIENTS

There has been a significant increase in publications on the multilevel approach for OSAHS patients within the last 5 years. The procedures include modification of previously invasive procedures, combination of minimally invasive techniques and classic surgery, and even minimally invasive single-stage multilevel surgery.

We recently reported on a systematic review of all English language literature on multilevel surgery for OSAHS patients as of 31 March 2007.²³ Article titles and abstracts were reviewed to determine article eligibility. We also identified relevant publications from the lists of references in this subset of articles. The study design and its corresponding level of evidence were clarified as follows: (1) Level 1 – randomized controlled trials or a systematic review; (2) Level 2 – prospective cohort study; (3) Level 3 – case–control study; (4) Level 4 – case series; and (5) Level 5 – expert opinion.

There were 49 papers (58 groups) selected for final inclusion after detailed review (Table 17.1). There were 1978 subjects included in the study with a pooled mean age of 46.2 years. The mean minimal time period from multilevel surgery to postoperative PSG was 7.3 months (range, 1–100 months).

Table 17.1 Peer reviewed studies on multilevel surgery for OSAHS