1 INTRODUCTION

Obstruction of the airway at the retropalatinal and retropharyngeal airway is the key factor in the pathophysiology of obstructive sleep apnea syndrome (OSAS). Patient selection is crucial in successful surgery of the upper airway in OSAS. In addition to polysomnography, topical diagnostic work-up is of paramount importance in this regard. We routinely perform polysomnography first, and after that in cases of an Apnea/Hypopnea Index (AHI) below 30, schedule patients for sedated endoscopy (‘sleep endoscopy’) with midazolam (without an anesthetist present), in those patients in whom surgery is considered. In patients with an AHI > 30, who refuse NCPAP treatment upfront, or in patients who cannot accept NCPAP for whatever reason, sedated endoscopy is performed as well, but by an anesthetist, with propofol. In the study period (March 2000 to June 2004), in the case of mainly or only retrolingual obstruction, as assessed by sleep endoscopy and a low AHI (arbitrarily <15–20, snoring up to mild sleep apnea) we usually started with radio-frequent ablation of the tongue base. In this situation, in the case of mild or moderate OSAS, oral devices were offered as an alternative. In cases of a relatively higher AHI (moderate to severe OSAS), the effect of MRA treatment is less efficacious. In the case of an index of 15–30, and mainly retrolingual obstruction, we performed hyoid suspension (a.k.a. hyoidthyroidpexia) as the only procedure. In higher AHI patients we perform multilevel surgery (hyoid suspension, radiofrequency ablation of the tongue base, uvulopalatopharyngoplasty (UPPP), with/or without genioglossal advancement (see Chapter 17, Multilevel surgery). These patients usually have more severe and multilevel obstruction, which explains the higher AHI.

Hyoid suspension involves stabilization of the hyoid bone inferiorly by attachment to the superior border of the thyroid cartilage. The underlying principle for altering the hyoid is that anatomically, the hyoid complex is an integral part of the hypopharynx. Anterior movement of the hyoid complex increases the posterior airway space and neutralizes obstruction at the tongue base. This concept has been supported by several reports. In the United States, hyoid suspension is often performed in combination with genioglossus advancement (GA) and followed by maxillomandibular osteotomy (MMO) as phased surgery in case of failure. The rationale for using hyoid suspension only in our series was to avoid more radical and more extensive unnecessary surgery in well-selected patients. In this chapter we report our experiences with this procedure.

2 PATIENTS AND METHODS

Surgery was offered to symptomatic patients with moderate to severe OSAS for whom UPPP was unsuccessful or who rejected or could not accept CPAP and preferred surgical therapy.

All patients had full polysomnography and underwent upper airway examination using physical examination and sleep endoscopy under midazolam or propofol.

2.2 UPPER AIRWAY ASSESSMENT

Hyoid suspension was performed in the case of obstruction at the base of tongue (Fig. 49.1), assessed by physical examination and flexible sleep endoscopy. A high suspicion of mainly retropalatal obstruction (large tonsils and long uvula) excluded patients for hyoid suspension. Candidates for surgery were categorized into two groups: those who did not have prior surgery at oro- or hypopharyngeal level (primary hyoid suspension); and those for whom UPPP was inadequate or detrimental (secondary hyoid suspension). In the latter group hyoid suspension was offered as salvage treatment. Multilevel obstruction (Fujita II) occurred in 20 patients; slight obstruction at retropalatal level in nine patients (primary hyoid suspension, n=14) and residual retropalatal obstruction after UPPP in 12 patients (secondary hyoid suspension, n=17). Only four patients who underwent primary hyoid suspension showed simple tongue base obstruction (Fujita III). Prior to UPPP, 15 patients showed multilevel obstruction (Fujita II), with emphasis on thepalatal level; two patients showed retropalatal obstruction only (Fujita I). Sleep endoscopy work-up according to the Fujita classification is shown in Figure 49.2.

Figure 49.1 Obstruction at tongue base level during sleep endoscopy with midazolam.

Figure 49.2 Endoscopic work-up using Fujita’s classification and detailed synopsis of surgical success. Type I: Palate obstruction (normal base of tongue), Type II: Palate and base of tongue obstruction. Type III: Base of tongue obstruction with normal palate. HS, hyoid suspension.

3 SURGICAL TECHNIQUE

3.1 HYOIDTHYROIDPEXIA AND POSTOPERATIVE MANAGEMENT

Under general anesthesia, with the head in a slightly extended position, a horizontal incision of approximately 5 cm is made in a relaxed skin tension line at the level between hyoid and thyroid cartilage (Fig. 49.3). Excessive fat tissue is excised, if helpful for better visualization. In the case of a further posterior positioned hyoid, removal of fat is recommended also, since otherwise the anterior placement of the hyoid will result in a somewhat turkey-like neck contour.

Figure 49.3 A, Anterior view. After exposure via a horizontal incision at the level of the thyroid membrane, the strap muscles (sternohyoid, omohyoid and thyrohyoid) are divided just below the hyoid and the tendons of the stylohyoid are divided from the hyoid bone, superior to the hyoid. B, Before hyoid suspension. C, After hyoid suspension. Permanent sutures are inserted through the thyroid cartilage and around the hyoid bone with subsequent antero-caudalmobilization and permanent fixation of the hyoid to the thyroid.

Second, the strap muscles are severed just below the attachment to the hyoid. Partial removal of the severed strap muscles at the level between hyoid and thyroid cartilage is sometimes also considered for the same cosmetic reasons, while there is no point in leaving the non-functional cut strap muscles in situ.

The tendon of the stylohyoid muscle is cut only if after release of the strap muscles insufficient mobilization is gained. Otherwise the stylohyoid tendon is preserved. By mobilizing the hyoid bone in an anterocaudal direction and fixing it to the thyroid with two permanent sutures per side through the thyroid cartilage and around the hyoid bone, more space is created retrolingually. Although with increasing age ossification of the thyroid will take place, in more than 80 cases we have never needed to make drill holes. A sharp cutting needle has so far always been sufficient to pierce the thyroid cartilage.

Antibiotics are not routinely applied. A surgical drain is placed and usually removed after 24–72 hours postoperatively if drainage was less than 10 ml per 24 hours. Nocturnal oximetry is monitored throughout the first postoperative night in the intensive care unit and non-opioid analgesics are used for pain relief, if necessary.

4 POSTOPERATIVE EVALUATION

All patients had a second polysomnography 3 months after surgery. The AHI, Epworth Sleepiness Scale (ESS), desaturation index and visual analogue scores (VAS) for hypersomnolence and snoring and pain sensation were compared pre- and postoperatively, per patient and between the two groups. Success was defined as when AHI indexes decreased at least 50% or dropped below the threshold of 20, or AI indexes decreased below 10 and responders as when AHI has decreased.

5 RESULTS

Between March 2000 and June 2004, 31 patients: (29 males and two females) underwent hyoid suspension. Secondary hyoid suspension was performed in 17 patients. Fourteen patients underwent primary hyoid suspension: 12 had difficulties using NCPAP, two refused NCPAP. Patient baseline characteristics are shown in Table 49.1. No differences were found preoperatively for AHI, Body Mass Index (BMI), age, ESS, desaturation indexes and VAS scores. The BMI did not change significantly postoperatively and all polysomnographies were considered valid.

Changes in AHI score are shown in Table 49.2. Twenty-five of the 31 (81%, 95% confidence interval 63–93%) patients showed a decrease in AHI; after secondary hyoid suspension, the mean AHI decreased from 31.5 to 26.2 (P=0.059) (Fig. 49.4) and after primary hyoid suspension the AHI decreased from 32.9 to 17.5 (P=0.0067). The decrease was significantly greater in the primary hyoid suspension group compared to the secondary hyoid suspension group (P=0.037) (Fig. 49.5). In terms of success, 16 of the 31 (52%, 95% confidence interval 33–70%) patients showed a reduction of the AHI of more than 50%, or a post-AHI value below 20 and post-AI below 10. Although almost twice as many patients showed success in the primary hyoid suspension group (35% versus 71%), the difference was not significant (P=0.073).

Figure 49.4 AHI post-UPPP and post-HTP.

Figure 49.5 AHI pre- and post-primary HTP.

In contrast to the primary hyoid suspension group, the desaturation index in the secondary hyoid suspension group decreased significantly from 23.6 to 17.5 (P=0.017).

Reduction of snoring in the primary hyoid suspension group (VAS: 8.5 to 5.6, P=0.009) and secondary hyoid suspension group (VAS: 8.8 to 5.6, P=0.002) was significant. No difference in snoring reduction was observed between both groups (P=0.80). VAS scores for hypersomnolence showed a significant decrease from 6.6 to 4.3 (P=0.027) and 5.8 to 3.4 (P=0.002) after primary and secondary hyoid suspension respectively; again, no difference was found between the groups (P=0.88). All patients experienced less morbidity after hyoid suspension than after UPPP. The VAS pain score after UPPP (with paracetamol 1000 mg four times daily) was 7.0; after hyoid suspension patients’ marked (without painkiller) 2.2. ESS scores significantly dropped for both primary and secondaryhyoid suspension (P=0.001 and P=0.031 respectively). No difference was found between the two groups (P=0.13). Statistical details are shown in Table 49.3.

No significant correlations between ESS scores and AHI indexes pre- and postoperatively, ESS scores and VAS scores for hypersomnolence preoperatively and AHI indexes and VAS scores for hypersomnolence preoperatively were seen. However, a positive correlation between postoperative VAS scores for hypersomnolence and AHI indexes as well as ESS scores postoperatively (Pearson correlation coefficient 0.48, P=0.006 and 0.52, P=0.004) was detected.

6 COMPLICATIONS

Up to July 2006, >80 patients underwent hyoid suspension, alone or in combination with other procedures. Tracheotomy (in the third patient) was performed once because of a compromised airway due to postoperative bleeding. There is a definite learning curve here. A few patients had swallowing problems for a few days; re-intervention for post-operative bleeding was necessary in two patients; fistulas occurred in three patients. In these cases one or two stitches were removed with good result, and in one case (multilevel treatment) a lesion of the hypoglossal nerve occurred (see Chapter 17, Multilevel surgery). No other serious complications were observed. Overall, hyoid suspension was better tolerated than UPPP.

7 DISCUSSION

Short-term experiences with hyoid suspension only, especially as primary treatment, in this small series of patients with retrolingual obstruction as assessed by sedated endoscopy, with moderate to severe OSAS, are favourable.Our results show that hyoid suspension only can be efficacious: 71% of the patients did not need further treat-ment. Only four patients were not successfully treated. Alternative treatment was offered to hyoid suspension failures, including a new attempt with NCPAP treatmentor oral device and revision surgery of relative palatal stenosis after UPPP. A floppy epiglottis caused airway obstruction at hypopharyngeal level in two other cases; these patients were considered non-responders. One patient underwent partial epiglottectomy. Clinical improvementin those patients, in terms of reduced snoring andimprovement in general well-being, was such that further therapy was not regarded as necessary by both patient and physician.

Secondary hyoid suspension showed a reduction in AHI, but was occasionally unsatisfactory and definitely inferior to primary hyoid suspension, although the improvement of desaturation index was significant. Relative stenosis at soft palate level after UPPP, objectified after sleep endoscopy, might explain the lack of improvement of OSAS. Improvement of the airway at the base of tongue level by secondary hyoid suspension is less effective or sometimes ineffective. It goes without saying that this group is complex and might not at all be comparable to patients undergoing primary hyoid suspension. Twelve patients with UPPP failure (n=17) had multilevel obstruction, with emphasis at the base of the tongue, in addition to the preoperatively assessed obstruction at palatal level with sleep endoscopy prior to UPPP. In the primary hyoid suspension group, nine of the 14 patients also had a slight obstruction at palatal level. The latter patients clearly responded better in terms of a stronger decline in AHI. A strong reduction in the amount of soft palate tissue (UPPP) might cause the tongue to drop back more easily, especially in cases of multilevel obstruction, because of lack of support by the soft palate. These patients might be better candidates for more invasive and aggressive treatments; in the case of a very high AHI, MMO should be considered. The overall success rate was 52%; overall the AHI significantly decreased (P=0.0005). Some patients with UPPP failure were cured by hyoid suspension, but it remains hard to predict which patients will do well in the future.

Comparison with other series is difficult, because in almost all other series hyoid suspension was combined with other procedures, with the exception of a series from Stuck et al. from Mannheim, Germany, who found an AHI reduction of 35 to 27 in a group of 14 patients. Treatment outcome of surgery is further negatively affected by an increasing BMI. Patients with BMI > 25 are considered overweight, in the case of BMI > 30, obese. Bowden et al. in a paper with a somewhat misleading title (‘Outcomes of hyoid suspension for the treatment of obstructive sleep apnea’) in fact included UPPP in most patients. Their results were very disappointing, probably due to the very high BMI (34) in their series. It is our belief that performing sedated endoscopy would probably have prevented surgery in these patients, because in the case of a very high AHI usually a circumferential collapse (anterior–posterior as well as lateral wall collapse) is seen, making such patients unsuitable for hyoid suspension.

Although our experiences are encouraging, we have changed our policy after the reported series. Since it has been shown that radiofrequency of the tongue base as the only procedure has a moderate effect, and since we have found that radiofrequency of the tongue base is almost without adverse events and side effects (applying antibiotics, but no steroids) we now more or less routinely (‘why not’ intervention) combine it with hyoid suspension. Results will be reported elsewhere.

We conclude that hyoid suspension, in particular as primary treatment in cases of obstruction at tongue base level, is a valuable addition to the therapeutic armentarium of moderate to severe OSAS. It is clear that experience with more patients is needed, while long-term results are lacking. In patients with UPPP failure or patients who refuse, or have difficulties using NCPAP, hyoid suspension should be considered. Selection criteria are: moderate to severe OSAS with preferably a BMI < 28, major obstruction at the base of the tongue, small tonsils and normal uvula, without a floppy epiglottis or a palatal stenosis after UPPP.