Complications of Obstructive Sleep Apnea Surgery
Obstructive sleep apnea (OSA) can be treated by multiple medical and surgical methods. Continuous positive airway pressure (CPAP), oral appliances, and weight loss serve as the foundation of medical therapy for most patients. For those patients who are unable to tolerate or unable to achieve treatment success with medical therapy options, surgical therapy can be beneficial—successfully improving symptoms and quality of life as well as reducing cardiovascular risks. Surgery for OSA can be used as an adjunct to improve adherence and success with medical device therapy (CPAP or oral appliance). Surgery can also be employed as a sole treatment strategy, most commonly as a staged multilevel plan.
Most OSA surgical procedures are aimed at enlarging and stabilizing the narrow and collapsible portions of the upper airway. Proper airway phenotyping, thorough examination and endoscopy techniques, and proper procedure selection are critical to successful surgical therapy and are beyond the scope of this chapter. In general terms, OSA surgery encompasses procedures of the nasal airway, pharynx, and craniofacial skeleton. Bariatric surgery, hypoglossal nerve stimulation, and tracheotomy also fall under the category of surgical therapy for OSA. This chapter focuses specifically on the soft tissue surgical procedures of the pharynx and the associated perioperative complications, whereas nasal surgery and craniofacial procedures are discussed elsewhere ( Figure 13.1 ) Complications associated with office-based procedures for nonapneic snoring or mild OSA are also described, as well as the perioperative management of any patient with OSA undergoing general anesthesia.
Since its introduction by Fujita in 1981, uvulopalatopharyngoplasty (UPPP) has been one of the most commonly performed operations for the treatment of OSA.1 Incidence and severity of some complications have changed with modifications of the procedure; however, the most commonly reported complications specific to UPPP remain bleeding, velopharyngeal insufficiency, nasopharyngeal stenosis, and globus sensation. As with any surgical therapy for OSA, airway obstruction/respiratory compromise, vascular complications, and persistent OSA are also potential risks.
Complications in the Perioperative Period
Anatomical factors and comorbidities associated with OSA predispose patients to complications in the perioperative patient. Incidence of serious complications following UPPP in the early postoperative period, however, is relatively low in most studies.2–7 Most of the data come from case series with a limited number of patients and varying descriptions of the complications. Kezirian et al2 prospectively collected data on the incidence of serious complications and 30-day mortality rate in over 3,000 adults undergoing inpatient UPPP from the national Veterans’ Affairs database from 1991 to 2001 and found an overall incidence of only 1.5%. This included death, respiratory; cardiovascular, including cardiac arrest, myocardial infarction, cerebrovascular accidents, pulmonary embolism, hemorrhage greater than 3 units of packed erythrocytes, coma, wound infection, systemic sepsis, deep venous thrombosis, and renal failure.
The most common and potentially severe complication is respiratory compromise. The severity of respiratory compromise varies and may include oxygen desaturation, reintubation, pneumonia, prolonged ventilation (> 48 hours), emergent tracheotomy, or pulmonary edema. The incidence of respiratory compromise varies widely between studies, with reported rates between 1.1% and 11%.2,8
Studies evaluating whether UPPP can be safely performed as an outpatient procedure showed that serious airway complications including airway obstruction and postoperative pulmonary edema occurred in the immediate postoperative period, usually evident in the recovery room within the first few hours after surgery. Oxygen desaturation may occur at any time during the postoperative period, but because oxygen desaturation is part of the underlying disease process, it may be difficult to distinguish between the disease itself and a postoperative respiratory complication. Riley et al7 reported that six of 182 patients had desaturation to the upper 80% range in the first two postoperative days. In a retrospective review by Hathaway and Johnson,8 3% of patients undergoing UPPP with or without septoplasty had oxygen desaturation in the recovery room. Studies have shown that postoperative oxygen desaturation is comparable to that observed during preoperative polysomnography. The polysomnography, as well as patients’ body mass index, cardiopulmonary status, and other clinical factors, should be considered when choosing patients who can safely have surgery as an outpatient.8,9
Careful patient selection and postoperative monitoring for several hours, avoids perioperative cardiopulmonary complications in most situations, even if the surgery is performed as an outpatient procedure. Reports of risk factors for serious complications following UPPP are conflicting in the literature. Haavisto and Suonpaa4 found that previous cardiac disease contributed. Esclamado et al3 suggested that weight, apnea–hypopnea index and lowest oxygen saturation on the polysomnogram, and preoperative narcotics were all associated with perioperative risks. Concurrent nasal procedures may temporarily worsen sleep-disordered breathing in the immediate postoperative period, particularly if nasal packing is used, by interfering with the nasal receptor-mediated control of breathing.3–5 More recently, Kezirian et al10 analyzed the records of 3130 veterans and concluded that apnea–hypopnea index, body mass index, medical comorbidities, and concurrent tongue-base procedures were each associated with an increased risk of serious complications.
Hemorrhage and Wound Dehiscence
Bleeding following UPPP has been estimated between 2 and 14% and appears to be similar to the bleeding rates following adult tonsillectomy.9,11 One cohort study showed that the incidence of a substantial hemorrhage (> 4 units packed red blood cells) was 0.3%.2 Bleeding may occur in the immediate postoperative period and likely represents a technical error. More commonly, bleeding after palatal surgery, with or without tonsillectomy, occurs in a delayed fashion and can occur anytime in the first 2-weeks after surgery.12 Delayed or secondary hemorrhage is likely related to wound healing factors, including granulation tissue formation, the inflammatory process, and eschar disruption. Although the data are conflicting, dietary and activity restrictions as well as postoperative medical therapy to reduce secondary inflammation and promote mucosalization, such as corticosteroids and topical sucralfate, may reduce the risk of postoperative bleeding. Exposed muscle in the tonsillar fossa may also contribute to the bleeding risk. Typically, the tonsils are removed concomitantly, and the tonsillar pillars are closed. Tension on the closure often leads to dehiscence of the closure, particularly at the tonsillar poles, leading to exposed muscle and potential for bleeding.
Velopharyngeal insufficiency (VPI) is the inability of the velopharyngeal sphincter to sufficiently separate the oropharynx and nasopharynx during swallowing or speech. Closure of the velopharynx depends on the musculus uvula and the action of the levator palatine muscles.
Following traditional UPPP with aggressive resection of the uvula and velum, temporary postoperative VPI is common; however, more permanent or long-term VPI may also occur with potentially devastating impact on quality of life. Some studies have reported that up to 10 to 24% of patients continued to complain of intermittent nasopharyngeal regurgitation 1 year after surgery.4,13 In his survey of 72 centers over 9 years, however, Fairbanks reported only a small number of patients with this complication.14
The amount of soft palate that can be safely resected varies among patients because the point of contact of the elevating palate and posterior pharyngeal wall varies, as does palatal length. A traditional resection of 1 to 2 cm will compromise the velopharyngeal function in some patients. Preoperative evaluation of the kneepoint, which corresponds to the distal point of the levator sling, is important. Intraoperatively it can be evaluated by retroplacement of the velum to the posterior pharyngeal wall. Surgical resection should not violate the levator sling. Extensive cautery and subsequent scarring can further shorten the palate and decrease velopharyngeal function.13,15
As VPI has potentially serious, long-term consequences, the best treatment of VPI is avoidance. VPI is best avoided altogether by employing current reconstructive procedures that preserve mucosa, uvular structure and function; rather than destructive or excisional, palatal procedures. Surgical correction of VPI is difficult because of shortening of the palate and scarring. It also carries the risk of worsening OSA. A superior pharyngeal flap or sphincter palatoplasty has been reported to result in improvement. Palatal pushback procedures, obturators, and Teflon paste injections have also been described.16–18
Nasopharyngeal stenosis (NPS) is the partial or complete obstruction between the nasopharynx and oropharynx due to concentric scar contracture of the tonsillar pillars and posterior pharyngeal wall ( Figs. 13.2 and 13.3 ).19 NPS can cause nasal airway obstruction, hyponasal speech, rhinorrhea, and worsening of OSA. Fortunately, it is a rare problem estimated to be less than 1%.20 In the prevention of nasopharyngeal stenosis, it is important to avoid forming a contiguous raw surface about the nasopharyngeal isthmus.15 Technical errors that have been shown to increase the incidence of NPS include excessive removal or cauterization of the posterior tonsillar mucosa or undermining the posterior pharyngeal wall mucosa, as well as closure under tension. A concomitant adenoidectomy should also be avoided.13,21
As with VPI, the best treatment of NPS is avoidance by using UPPP modifications that maximally preserve mucosa of the velopharyngeal isthmus. Once it has occurred, the treatment of NPS is difficult.21 Krespi and Kacker22 reported 18 patients with NPS following UPPP. All patients were treated with the CO2 laser with or without a nasopharyngeal obturator resulting in adequate nasopharyngeal lumens following treatment. Though severity ranged from mild to severe, it should be noted that results are typically better in less severe forms of stenosis. Jones et al had success with treatment of patients with laser followed by Mitomycin C and an obturator.23 Other treatment options include dilation, scar revision with or without skin grafting, scar excision with maxillofacial prosthesis, local flaps, pedicled flaps, and free-tissue transfer.23–28
Though most of the above complications of UPPP have been well documented and studied, less attention has been given to postoperative taste disturbances. A review of the literature revealed a reported rate of 7 to 10%.29,30 A longitudinal intervention study designed specifically to investigate taste disturbances following UPPP revealed a 4.6% rate 3 months after surgery. Nine months after the operation all but one of five patients had a restoration of their taste sensation. Gustatory function tests showed that deficiency of sweet sense was the most common type of taste disturbance.31 Following lateral pharyngoplasty, which is discussed further in a following section, Cahali32 also noted one patient with a taste loss for chocolate, which recovered after 6 months.
Causes of taste dysfunction may include damage to the lingual branch of the glossopharyngeal nerve in the tonsillar fossa, excision of taste receptors on the soft palate, or mechanical pressure on the tongue. It is speculated that the decreased use of electrocautery for tonsillectomy, particularly at the inferior tonsillar pole region, may decrease the incidence of taste dysfunction. The palate receives its taste sensation from the geniculate ganglion via the great petrosal nerve. Sweet sensation is the most sensitively perceived taste of the palate. In the study by Li et al,29 the incidence of taste disturbances was significantly higher for patients with electrocautery used during development of the uvulopalatal flap. Avoiding excision of the soft palate during palatopharyngeal surgery and decreasing the use of electrocautery on the soft palate may further decrease postoperative hypogeusia, particularly to sweet tastes.29–31
The results of voice disturbances following UPPP have varied in the literature. Powell pointed out that interpretation of speech problems are difficult; he suggested that the redundant mucosa may have led to abnormal speech patterns before surgery. Van Lierde et al33 performed a prospective study of nasalance, nasality, voice, and articulation of 26 patients after UPPP and found that there was no impact on nasality, voice, and articulation. Regarding nasalance, the only change that occurred involved the sound “i”.
Brosch et al34 prospectively studied patients undergoing muscle sparing UPPP and tonsillectomy and found a significant raising of the fundamental frequency. Tewary and Cable35 found that the fundamental frequency was lower in patients undergoing UPPP compared with those undergoing tonsillectomy. Most patients will not notice such changes, but professional voice users should be counseled preoperatively.
Though often considered minor when compared with the other reported complications of UPPP, globus, or a foreign body sensation, has been the most commonly reported complaint following UPPP. It has been reported in 22 to 31% of patients after 1 year and up to 60% of patients after 3 years. Some patients even complain of persistent long-term discomfort. This potentially very bothersome complication may result from scarring of the soft palate and excision of the uvula. The uvula contains the highest concentration of serous glands in the oral cavity/oropharynx and plays an important role in lubricating the posterior pharyngeal wall, which explains why removal of the glands may result in pharyngeal dryness and globus sensation.15,36–38
Therefore, preservation of the uvula may substantially reduce the incidence of this complication. Kwon et al39 designed a study to evaluate the postoperative effect of uvular-preserving palatopharyngoplasty on OSA and globus sensation. Using a VAS, they showed no change in globus sensation following surgery compared with preoperative values. No treatment has been shown to be curative, but injection of scarred tissue with depot corticosteroids in the intermediate postoperative period has been attempted. For long-term treatment, oral lubricants may provide modest symptomatic relief.
Complications Related to Specific Palate Procedures
The following procedures have not been as widely researched, and the actual incidence of the complications is not known. The reported rates of complications, as well as the potential complications, and avoidance and treatment strategies are discussed that have been reported in the literature.
Transpalatal Advancement Pharyngoplasty
This technique was first described in the literature in 1993 and was developed to address the limitations of traditional UPPP techniques. This technique enlarges the velopharynx and retropalatal segment by excising a portion of the posterior hard palate and advancing the soft palate anteriorly. Like most mucosa-sparing reconstructive techniques, the overall incidence of complications, particularly VPI, NPS, and globus sensation, may be dramatically reduced compared with traditional UPPP. However, the transpalatal advancement pharyngoplasty, with its osteotomy technique, introduces the possibility of a new complication: oronasal fistula.
In the literature, the fistula rate increased when the technique was modified and the tensor tendon was incised to increase mobilization. The largest series to date showed the occurrence of an oronasal fistula in 12.7% (6 of 47 patients). Subsequent modifications have led to a reduction in fistula rates. Shine and Lewis40 compared effectiveness of the procedure using two incision types, traditional Gothic arch incision and a modified propeller incision. They noted decreased fistula rates in the patients who had undergone the propeller soft tissue approach.
Expansion Sphincter Pharyngoplasty
Lateral pharyngeal wall collapse has been demonstrated to be an important component of OSA and is not addressed by standard UPPP techniques. In 2003, Cahali32 first described the lateral pharyngoplasty, which involved sectioning of the superior pharyngeal constrictor muscle and creating a laterally based flap. In addition, a palatopharyngeal Z-plasty was performed. After the procedure most patients reported dysphagia especially to dry, solid foods lasting from 8 to 70 days, with a median time 14.5 days. Pharyngeal swallow returned without further treatment.
The expansion sphincter pharyngoplasty (ESP) was developed, as a modification of the lateral pharyngoplasty, to enlarge the lateral dimension of the retropalatal space in patients with a large lateral wall component. Bilateral rotation flaps comprising the palatopharyngeus muscles characterize the ESP. In a prospective controlled trial comparing ESP with traditional UPPP, ESP provided more effective treatment of the OSA and was not associated with any significant complications.43
For patients with a primarily anterior–posterior pattern of palatal collapse at the level of the levator sling and velum, anterior palatoplasty may provide successful management of the palatal flutter and obstruction while limiting complications by preserving the structure and function of the uvula and free edge of the soft palate. In this technique, a strip of mucosa and submucosal tissue is removed from the proximal soft palate followed by closure of the horizontal defect and anterior advancement of the soft palate. In a prospective series of 77 patients, there was no VPI, NPS, or other reported complications.44
A variety of oropharyngeal/hypopharyngeal procedures are available for the treatment of OSA. Although genioglossus advancement, other skeletal procedures, and hypoglossal nerve stimulation address this portion of the airway as well, this section focuses on the soft tissue procedures only. These procedures primarily involve either (1) volumetric reduction or (2) advancement/suspension of the tongue base.
Hypopharyngeal procedures carry the risk of significant perioperative complications, which may be more difficult to treat because of their location lower in the airway. Identification and control of hemorrhage and airway obstruction may be more challenging, and potentially dangerous, complications in this area. Data suggest that most of these patients should be kept overnight though the perioperative complications following hypopharyngeal surgery have not been as well studied as those following UPPP. In the study by Kezirian et al,2 they showed an increased risk of perioperative morbidity when performing tongue-base procedures with UPPP, though it was not possible to separate the cumulative risk versus the individual risk of the procedures.
Volumetric Reduction Procedures
Glossectomy involves decreasing the size of the soft tissues of the tongue base and hypopharynx. Initially, this procedure was associated with a high morbidity as the result of aggressive resections.45–47 Submucosal minimally invasive lingual excision (SMILE), transoral submucosal endoscopy-assisted glossectomy, and lingual tonsillectomy have all been described as modifications developed to treat tongue-base-associated airway obstruction in adult and pediatric patients. The SMILE technique and transoral endoscopy-assisted glossectomy both involve removal using a plasma-mediated radiofrequency device (coblation) while sparing the taste buds and mucosa.47,48 Potential complications include bleeding, hypoglossal nerve injury, edema, and airway compromise.49 Actual incidence of complications with the transoral endoscopy-assisted glossectomy technique is not known, as there is little in the literature, but surgeon reports are rare.