Traditional pharyngeal flap harvest places the lower edge of dissection below the soft palate. The modified technique is performed above this level so that the flap and donor site cannot be seen on the transoral view.
High, superiorly based flap is buried within a pocket created along the nasal surface of the soft palate. Note that the entire donor site lies above the free edge of the soft palate.
After wide undermining of the surrounding mucosa (shaded area), the donor site is closed vertically as far as possible to cover the raw surface along the posterior pharyngeal wall.
Nasopharyngoscopy of a well-healed flap. Distal view (A) and close-up view (B). Note that the flap is extremely high and extends downward toward its attachment with the palate.
Horizontal closure or contracture of the donor site results in narrowing of the pharyngeal airway. Such an effect may be complicated if tonsils are present, resulting in more severe obstructive symptoms.
Patient with obstructive sleep apnea after pharyngeal flap surgery. Lateral port obstruction by the tonsils was found to be total on the right and subtotal on the left. Symptoms resolved after tonsillectomy.
Closure of defects in the pharyngeal airway. A, Narrowing of the pharyngeal airway may be avoided by vertical closure of the defect. B, Decreased airway diameter is seen with the traditional design as a result of horizontal closure.
Chegar BE, Shprintzen RJ, Curtis MS, Tatum SA. Pharyngeal Flap and Obstructive ApneaMaximizing Speech Outcome While Limiting Complications. Arch Facial Plast Surg. 2007;9(4):252-259. doi:10.1001/archfaci.9.4.252
Author Affiliations: Center for Facial Plastics Head and Neck Surgery, Fayette Regional Health System, Connersville, Indiana (Dr Chegar); Communications Disorder Unit and the Center for the Diagnosis, Treatment, and Study of Velo-Cardio-Facial Syndrome, Department of Otolaryngology and Communication Science (Dr Shprintzen), and Departments of Otolaryngology and Pediatrics (Dr Tatum), State University of New York–Upstate Medical University, Syracuse; and Department of General Surgery, St Louis University School of Medicine, St Louis, Missouri (Dr Curtis).
Correspondence: Sherard A. Tatum, MD, Department of Otolaryngology, SUNY-Upstate Medical University, 750 E Adams St, Syracuse, NY 13210 (firstname.lastname@example.org).
Objective To assess speech results and rate of obstructive sleep apnea using a modified, superiorly based pharyngeal flap performed after staged adenotonsillectomy in a group with velopharyngeal insufficiency.
Methods In this nonrandomized, retrospective case series (July 1, 1996, through June 30, 2003), patients were mainly children referred to a multispecialty craniofacial clinic. Patients underwent staged adenotonsillectomy 2 months before width-customized pharyngeal flap surgery. Short flaps were created high above the level of the palate, just long enough to reach the nasal surface. Donor sites were closed by superior advancement of the inferior posterior pharyngeal wall tissue. Cardiopulmonary and oximetry data were analyzed for immediate obstructive apnea. Speech results and airway symptoms were assessed at 6-month and yearly follow-up examinations.
Results In the 54 consecutive patients who underwent staged adenotonsillectomy, no apnea occurred immediately after surgery. Long-term clinical examination revealed 4 cases of loud snoring. Polysomnographic results were negative in all cases. Complications included 3 cases of bleeding, 1 requiring transfusion. Velopharyngeal insufficiency was eliminated in 94% of patients.
Conclusion Complications related to obstructive sleep apnea have been significantly reduced while maintaining excellent speech results by a staged approach of removing tonsils and adenoids and by creating a short, high, wide, superiorly based pharyngeal flap with superior advancement of the inferior posterior wall to close the donor site.
Velopharyngeal insufficiency (VPI) is a disorder associated with cleft palate that persists in approximately 10% to 30% of cases after primary palatoplasty.1 Philip Gustav Passavant is credited with first describing, in 1862, the surgical attachment of the soft palate to the posterior pharyngeal wall to improve the hypernasal intonation in such patients.2 Twenty-five years later, Karl Schoenborn, MD, introduced the first true pharyngeal flap, but it was not until 1930 that Earl C. Padgett popularized the procedure.2 He found that the superiorly based flap was best at providing the necessary length, even in situations of significant soft palate deficiency.2 Although it has been regarded as an “unphysiological” procedure, most surgeons have agreed that the pharyngeal flap is the most effective technique, especially for management of severe cases.3
Despite its success, the safety of the pharyngeal flap has been a concern. Since a reported case of death related to airway obstruction after a pharyngeal flap, the incidence of obstructive sleep apnea (OSA) after this procedure has been a matter of increased focus.4 The largest studies5-7 have found evidence of OSA ranging from 2% to 10%, whereas smaller case series8 have documented obstruction as high as 90%. Although many of these reports have shown that this problem is limited mostly to the immediate postoperative period, there does appear to be a small group of patients who will continue to experience long-term difficulties, necessitating flap revision or even flap takedown.9
As a result of these findings, other operations have been used for treatment of VPI based on the presumption that the pharyngeal flap has a higher incidence of postoperative sleep apnea. Sphincter pharyngoplasty is commonly used, but the effectiveness of this technique has been variable.10-12 Additionally, rates of postoperative OSA may be just as high with sphincter pharyngoplasty as they are with the pharyngeal flap.13 Results from recent prospective comparison studies14-15 have found that rates of sleep apnea after surgery are essentially no different using either type of procedure.
Looking at ways of preventing OSA, several reports5, 7, 16 have addressed the need to contain the risk of airway obstruction after performance of the pharyngeal flap. Most problems appear to be limited to the first 24 to 48 hours after surgery, so it is recommended that immediate postoperative care be administered in an enhanced care setting. Additional suggestions have included limiting the number of concurrent head and neck procedures, postoperative use of a nasopharyngeal airway, and avoidance of pharyngeal flap placement in syndromic patients with a predisposition for airway collapse.7 Despite some apparent success at reducing the overall morbidity and mortality, there has yet to be a well-accepted guideline for managing this feared complication.
For the last 8 years at State University of New York–Upstate Medical University, a surgical protocol has been in place with the specific aim of limiting OSA after construction of the pharyngeal flap. This protocol has been particularly important because our patient population has required most flaps to be wide or very wide for the treatment of VPI. On the basis of previous observations of preventable sources of airway obstruction,5, 16 the main focus has been on removal of the tonsils and adenoids before flap surgery, shortening the length of the flap (and subsequently the length of the donor site), and superior advancement of the inferior posterior pharyngeal wall for closure of the donor site.
All patients were evaluated at the SUNY-Upstate Medical University's multidisciplinary Cleft Palate/Craniofacial Center from July 1, 1996, through June 30, 2003. Patients who were diagnosed as having VPI after perceptual speech analysis underwent multiview videofluoroscopy and nasopharyngoscopy. Individuals who demonstrated poor velopharyngeal closure and borderline patients in whom speech therapy failed were identified as candidates for surgery. Patients with a history of loud snoring, witnessed apnea, or other signs suggestive of OSA (eg, daytime hypersomnolence, enuresis, or irritability) were evaluated with nocturnal polysomnography before surgery. Those with a confirmed diagnosis of OSA (respiratory disturbance index >5 and oxygen saturation as measured by pulse oximetry nadir <90%) were recommended for alternative treatment such as a speech bulb prosthesis. Children with other comorbidities such as severe heart disease were carefully evaluated by their cardiologists; some were recommended for alternative treatment as well. Any patient with velocardiofacial syndrome (VCFS) underwent additional testing with magnetic resonance angiography to evaluate the presence of a medially displaced internal carotid artery along the posterior pharyngeal wall. All diagnoses of VCFS had been previously confirmed by chromosomal analysis that showed the 22q11.2 microdeletion.
Routine tonsillectomy was performed at least 8 weeks before the pharyngeal flap. In addition, because of the surgeon's technique of harvesting a superiorly based flap as high as possible along the posterior wall of the nasopharynx, an adenoidectomy was performed in any case in which adenoid tissue was present. Complete removal of adenoid tissue was performed by curettage. Patients and family members were counseled preoperatively that speech during this interval would likely be worse.
The pharyngeal flap was created similarly in all cases by 1 surgeon (S.A.T.). With the soft palate retracted by an assistant, the caudal edge of the flap was developed superiorly along the posterior wall typically at or just above the level of the velum (Figure 1). Dissection was performed using direct visualization superiorly in a plane just superficial to the prevertebral fascia. Flap width was customized to the degree of lateral wall movement that was previously documented endoscopically and fluoroscopically. Once sufficient length was obtained, a transverse incision was made along the nasal surface of the soft palate approximately 1 to 2 cm from the free edge and a pocket was created under the nasopharyngeal mucosa. The flap was then sutured into the pocket (Figure 2). With the aid of a dental mirror, the lateral ports were inspected and further adjusted as needed. Several 3-0 chromic sutures were placed through the oral surface of the soft palate, then through the flap edge, and then back through the soft palate. After securing the flap, the margins of the donor bed were widely undermined inferiorly and advanced superiorly, closing the defect (Figure 3). If the donor site could not be completely closed, the edges were advanced maximally so that the narrowest strip possible beneath the base of the flap was left to heal by secondary intention. In no patient was the donor site visible, on oral examination without palate retraction. The procedure was concluded after confirmation of the integrity of the sutures and proper hemostasis.
Postoperatively, all patients were observed in the intermediate care unit for at least 24 hours and then moved to the general floor unit if no respiratory difficulty or other complications occurred. Continuous oximetry and cardiac and apnea monitors were used at all times until discharge. After initial preoperative doses, patients were administered dexamethasone intravenously (0.25 mg/kg per dose 3 times daily) for 24 hours and antibiotics for 10 days. Pain control was usually managed initially with morphine sulfate followed by acetaminophen with codeine elixir, with attention to avoid oversedation and respiratory depression. Absolutely no intranasal instrumentation was allowed, and nasal suctioning was minimized. Humidified room air, topical nasal isotonic sodium chloride solution, and 0.05% oxymetazoline nasal spray were used to help avoid drying, crusting, and oozing within the airway that might contribute to obstruction. Patients were not allowed to take anything by mouth for the first 24 hours to prevent unnecessary tension to the flap from swallowing. A liquid diet was then started and advanced to a pureed consistency as tolerated. Discharge was determined by the adequacy of oral intake, pain control, and respiratory status after the patient had been observed for at least 24 hours. Instructions were given to the caregiver to observe the patient nightly for signs of breathing difficulty, including loud snoring or apnea.
Follow-up appointments were arranged at 1 week, 3 months, and 6 months after discharge. The structure and function of the flap were evaluated with nasopharyngoscopy. Speech results were gauged by a 3-member panel of speech language pathologists at least 1 month after surgery and again with subsequent follow-up visits. Judgments were rendered by consensus agreement of the panel, which was found to be a highly reliable method according to previous studies.17 Treatment of VPI was considered successful only if there was complete absence of both audible nasal emission and hypernasal resonance. The occurrence of hyponasality and other difficulties associated with nasal obstruction was noted. Initial determination of OSA was based on a reliable report of the sleep pattern and presence or absence of daytime hypersomnolence by the child's caregivers. Questionable histories were first evaluated by an audio sleep tape before deciding on polysomnography. Persistent obstructive symptoms were followed by polysomnography.
Data were prospectively collected on 54 consecutive patients (29 female and 25 male). Ages ranged from 4 to 28 years, with a mean age of 7.2 years and a median age of 5.8 years. The origin of the VPI was related to a diagnosis of VCFS in most cases (Table). Of the 54 patients, 7 had had their VPI unsuccessfully treated at an outside institution: 3 underwent sphincter pharyngoplasty, 3 underwent Furlow palatoplasty, and 1 underwent a pharyngeal flap. Forty-six patients (85%) required either a wide or near-total obstructing flap.
Three patients developed bleeding: 2 cases were localized to the interface of the flap and soft palate and occurred 1 week postoperatively, whereas the third case occurred in the recovery room and was caused by an unrelated nasal source. Hemostasis was obtained without injury to any of the flaps. One transfusion was required.
Postoperative airway monitoring showed no evidence of OSA. Transient events in the first 24 hours responded to minimal intervention, such as stimulation, positioning, or the administration of blow-by oxygen. No patients required reintubation or early return to the operating room for flap revision or take-down. One patient was noted to have persistent, loud snoring and nasal stuffiness immediately after surgery. Despite maintaining a stable airway, this individual was kept in the hospital for 6 days for monitoring because he required extended travel time back home overseas. No episodes of oxygen desaturation or apnea were recorded during his entire stay. All other patients were discharged on postoperative days 2 or 3 without evidence of respiratory obstruction.
Follow-up was obtained for all patients from a minimum of 9 months to a maximum of 7 years. One flap dehisced within the first postoperative week and underwent successful revision. The remainder of the flaps healed completely (Figure 4). No additional patients were found to have a breathing pattern suggestive of OSA according to the caregivers' reports at the 1-week postoperative visit. At 6 months, 1 patient had developed progressively loud snoring. Overnight polysomnography did not demonstrate evidence of OSA. Two additional patients were found to have snoring and associated sleep disturbance at the 1-year follow-up. Both of these patients also had a normal polysomnography result. Although the patient who had developed immediate postoperative snoring was not available for follow-up at the State University of New York–Upstate Medical University Cleft Palate/Craniofacial Clinic, correspondence was maintained through the patient's family and home physician. This patient continued to have persistent, loud snoring. Two polysomnograms were obtained, and both were negative for OSA. In none of these 4 patients was there evidence that this nighttime breathing pattern was negatively affecting the children's mental or physical health; therefore, flap modification was not performed.
Perceptual speech analysis showed complete elimination of VPI in 51 patients (94%), including all 7 patients who had a previous failed surgery. Although 2 cases had significant reduction in the amount of nasal air escape and hypernasality, resulting in improved speech intelligibility, these cases were considered operative failures by the speech pathologist. The third failure, which occurred in the patient who developed flap dehiscence, later achieved complete resolution after revision of the pharyngeal flap. If included, this case would have increased the success rate of the procedure to 96%.
Long-term complications related to nasal obstruction occurred in 8 patients (15%). Four patients had persistent hyponasal speech, 3 had chronic congestion and rhinorrhea, and 1 was found to have persistent hyposmia. All of these patients had presented initially with severe VPI and had required extremely wide flaps.
The superiorly based pharyngeal flap is an excellent choice for the treatment of VPI. The complete elimination of VPI obtained in 94% of patients in this series is similar to results found in other studies.5, 18 Although most individuals (85%) required wide flaps, a low frequency of long-term problems related to nasal obstruction was found (15%). This complication may be unavoidable in cases in which an extremely wide, customized flap is performed, but in no situation was flap takedown or modification considered desirable. Long-term follow-up has shown that the diameter of the lateral ports increases with time, with preservation of excellent speech results. This would suggest that the need for lining the raw surface of the flap, as suggested by Hogan,18 is unnecessary with this approach. Final flap width appears to depend as much on the width of the inset incision as on the width of the base of the flap. Additionally, palatal splitting, which is often required to allow lining of the flap, may lead to less reliable flap width at the inset site. By using appropriate preoperative measures to assess the degree of lateral wall motion, the surgeon is best served by tailoring wide flaps to even the most severe cases rather than by performing an alternative procedure that may produce suboptimal results.
Airway complications after pharyngeal flap construction can and should be avoided with appropriate preoperative planning and operative technique. Although it is not necessary to perform polysomnography on all patients,5, 19-20 patients should undergo careful screening to identify those at risk for airway complications. Our institution's protocol includes a detailed history of the patient's sleeping pattern (including the modified Epworth Sleepiness Scale), thorough head and neck examination, and auscultation of the upper airway. Any individual who has evidence suggestive of OSA or upper airway resistance syndrome has a sleep audiotape or videotape recorded during multiple representative nights to evaluate for interrupted breathing, arousals, and restlessness. Polysomnography is then performed as indicated. A confirmed diagnosis should be considered a relative contraindication for proceeding with the pharyngeal flap. Diagnosis and treatment of the cause of OSA should be undertaken before careful consideration of a flap.
Tonsils are 1 of the primary risk factors for the development of OSA associated with the pharyngeal flap. In a series of 571 patients who received a superiorly based pharyngeal flap, Ysunza et al6 found that 13 of 14 individuals who developed postoperative OSA had enlarged tonsils. Complete resolution was obtained in all cases after tonsillectomy and modified uvulopalatopharyngoplasty. Lesavoy et al20 found that, in a group of 29 patients who did not develop severe upper airway obstruction after performance of the pharyngeal flap, only 1 had evidence of enlarged tonsils.
Children who undergo pharyngeal flap construction with tonsils in place but do not develop postoperative OSA are still at risk for developing hypertrophy and obstruction even several years after surgery. This hazard may be compounded if the donor site is closed by horizontal or medial advancement of the lateral pharyngeal walls or if the same is allowed to happen by secondary intention, causing oropharyngeal contraction and posterior displacement of the tonsils into the oropharyngeal airway (Figure 5). Such a case was illustrated by a patient referred to our clinic who had developed late-onset OSA after performance of the pharyngeal flap at another institution. Examination revealed that the enlarged tonsils had been drawn posteriorly and were blocking both lateral ports (Figure 6). These symptoms completely resolved after tonsillectomy. Given that the results of our study show no evidence of acute or chronic OSA, we strongly believe that tonsillectomy should be considered routine before construction of the pharyngeal flap. Removal of tonsils has the additional benefit of improved exposure for the creation of a wider flap if necessary.
Although the adenoidectomy performed 8 weeks before the pharyngeal flap often worsens speech during the interval period, it is an important part of the creation of a high flap. Flap dissection is much easier without overlying lymphoid tissue, and adenoidectomy removes another potential source of airway obstruction. Given that the adenoid bed is relatively small with VCFS and the fact that some of the older patients who were enrolled did not have appreciable adenoid tissue, adenoid removal likely had a minimal impact on the airway results in this study. If necessary, it is preferable that adenoidectomy not be performed simultaneously with the pharyngeal flap because of the risk of creating a rent through the flap or diminishing its blood supply. Bleeding and cautery in this area would also limit visibility and increase the amount of postoperative airway edema.
The design of the extremely high flap and vertical advancement donor site closure method may have contributed to the lack of breathing complications in this series. Contraction during healing of the donor site can lead to narrowing of the pharynx. Flaps that are constructed as low as the hypopharynx can therefore lead to a compromise of the hypopharyngeal, oropharyngeal, and nasopharyngeal airways. By limiting the harvest site above the soft palate, such effects on the lower pharyngeal airway are prevented. The flaps created with this technique were relatively short, so vertical closure was possible and therefore avoided unwanted narrowing of the pharyngeal airway (Figure 7). Considering that 85% of the patients in this group received a wide or very wide flap, this finding would support previous assertions that flap width does not correlate with OSA incidence after pharyngeal flap surgery.5-6,21
Even more encouraging about the results of this study is that the lack of OSA was observed in a group of susceptible patients. Velocardiofacial syndrome has been associated with a high risk of OSA after construction of the pharyngeal flap because of poor neuromuscular control and associated cardiac anomalies.14 Fraulin et al7 found that airway obstruction occurred twice as frequently (12.5% vs 4.9%) after performance of the pharyngeal flap in patients with an associated medical condition, the most frequent of which was VCFS. Similarly, Shprintzen5 found that 43% of patients with VCFS developed OSA vs 6% of children with nonsyndromic cleft palate. Although most patients in this study had VCFS, we believe that this protocol may also be safe and effective for other syndromes that have previously been considered contraindications for the pharyngeal flap.7, 22
Although OSA was generally absent in this study, 4 children developed loud snoring at night. None of these patients had evidence of sleep apnea on polysomnography, but this does not rule out the possibility of upper airway resistance syndrome. Because these symptoms did not seem to have a negative health impact, further surgical intervention was not desirable, especially given their good speech results. Long-term, close evaluation is nonetheless recommended in this situation.
Other complications of this protocol were relatively uncommon. Flap dehiscence occurred in only 1 patient (2%) in this group. Even though these flaps are much shorter than those in classic descriptions,18 the high dissection into the nasopharynx permits a tension-free closure that has a lower likelihood of separation. The addition of the postoperative diet restrictions and attention to avoid flap manipulation may have contributed to our small complication rate.
Bleeding was encountered in 3 cases (6%), with 1 transfusion necessary. This finding does not vary significantly from bleeding complications found in other series.16, 23 No great vessel injuries occurred in any of the patients with VCFS. In all of these individuals, magnetic resonance angiography was valuable at identifying any abnormally positioned internal carotid arteries that were located in the area of the donor site. Although this condition is not considered a contraindication to performing a pharyngeal flap,24 magnetic resonance angiography permits the surgeon to plan a safe dissection to avoid catastrophic injury. In such cases, the flap width is not changed. Rather the flap is elevated off the carotid. The exposed vessel is bluntly lateralized, and the lateral pharyngeal wall tissue is sewn to the prevertebral muscle to cover the carotid.
Although the necessity for flap revision is low, the ability to reconstruct or modify the superiorly based pharyngeal flap was shown to be highly successful. Barone et al25 found that complete reconstruction was possible by harvesting a new flap from the scarred posterior pharyngeal wall donor site. In their study, 19 of 21 patients with residual hypernasality obtained normal nasal resonance after revision. The single salvage performed in this series was similarly successful at establishing normal speech. Additionally, the pharyngeal flap was successful in all 7 cases with a prior failed surgical repair at an outside institution. The simplicity of the procedure may explain the superior results compared with sphincter pharyngoplasty, which was shown to have a successful revision rate as low as 50%.11
In conclusion, the customized, superiorly based pharyngeal flap is a reliable and effective means of correcting hypernasality for both children and adults with VPI. Despite previous concerns, complications related to OSA have been significantly reduced by a staged approach of removing the tonsils and adenoids and by creating a short, extremely high, superiorly based pharyngeal flap with vertical advancement closure of the donor site.
Accepted for Publication: February 25, 2007.
Author Contributions:Study concept and design: Chegar, Shprintzen, and Tatum. Acquisition of data: Chegar and Curtis. Analysis and interpretation of data: Chegar and Tatum. Drafting of the manuscript: Chegar, Shprintzen, and Curtis. Critical revision of the manuscript for important intellectual content: Shprintzen and Tatum. Statistical analysis: Chegar and Curtis. Administrative, technical, and material support: Shprintzen and Curtis. Study supervision: Shprintzen and Tatum.
Financial Disclosure: None.
Previous Presentation: This study was presented at the Society for Ear, Nose, and Throat Advances in Children/American Academy of Pediatrics Joint Annual Meeting; November 1, 2003; New Orleans, Louisiana.