Previous studies have examined the efficacy of lingual tonsillectomy alone, but none have analyzed the outcomes of simultaneous lingual tonsillectomy and epiglottopexy. In addition, there are no published data related to epiglottopexy as a standalone procedure, as retrodisplacement of the epiglottis may be secondary to tongue base anatomy. Given these limitations, it is difficult for pediatric otolaryngologists to appropriately counsel patients regarding the likelihood of success with sleep surgery directed at the tongue-base/epiglottic complex.

Lingual tonsillectomy is one of many surgical options for the treatment of OSA in the pediatric population, specifically for patients with hypertrophy of this tissue. Lingual tonsillectomy may be performed with any number of ablative techniques, such as monopolar suction cautery, laser, radiofrequency ablation (“Coblation”), or microdebridement. The lingual tonsil tissue can also be removed en bloc through the use of monopolar cautery, bipolar cautery, laser, robotic cautery, or cold steel [5,6,7,8,9, 12]. Epiglottopexy techniques vary; however, most agree that the procedure should involve some degree of mucosal disruption through superficial trauma or mucosal removal, followed by fixation of the epiglottis to the base of tongue (Fig. 1).

Variability exists in the role of aryepiglottic fold trimming, mechanism of suture fixation, suture selection, suture placement, and treatment of the hyoepiglottic, or lingual-epiglottic, ligament [12,13,14,15]. Within the study institution, a non-absorbable suture is placed in a horizontal mattress format between the lingual surface of the epiglottis and the base of tongue. The hypoepiglottic ligament is foreshortened by cauterization with Coblation prior to suture placement. The hyoepiglottic ligament plays an important role in suspending the epiglottis to the base of tongue (Fig. 2). Depending on the technique used, lingual tonsillectomy can disrupt its structural integrity, leading to epiglottic retroflexion and secondary obstruction. Many surgeons anecdotally feel that epiglottopexy prevents this collapse by recreating the tension vector of the ligament and improving epiglottic stability, however studies regarding this have not been published.

Photodocumentation of lingual tonsillectomy and epiglottopexy operative technique. a Preoperative lingual tonsillar hypertrophy and epiglottic prolapse, b Removal of lingual tonsil tissue with coblation, c Removal of lingual tonsil tissue with microdebrider, d Endoscopic suturing of the base of tongue to the lingual surface of the epiglottis, e Postoperative repositioning of the lingual tonsil-epiglottis complex with a widely patent airway

The relationship of the hyoepiglottic—or lingual-epiglottic—ligament (arrow) to the base of tongue makes it an important suspensory support for the epiglottis

In this cohort, concurrent lingual tonsillectomy and epiglottopexy led to a statistically significant reduction in oAHI (p < 0.05) and a reduction in severity of obstructive sleep apnea (p < 0.01). Additional subgroup analysis showed that patients with a preoperative diagnosis of moderate OSA had a significantly (p = 0.011) greater relative reduction in oAHI compared to the severe group. The most likely reason for this is that patients with severe sleep apnea often have significant multilevel obstruction, even if the base of tongue is the most notable. Additionally, as severe sleep apnea has no upper limit, even a large reduction in oAHI can result in a final number that is considered severe. Conversely, a similar absolute oAHI decrease in a patient with moderate OSA could downgrade the patient to the mild category.

The success rate in this cohort was 84.6%, which is higher than the range reported in studies of lingual tonsillectomy alone (51–62%) [5,6,7,8,9]. Additionally, of the 9 patients who did not have postoperative PSGs performed, 5 did not due to significant clinical improvement, while 2/5 patients without preoperative PSGs had postoperative tests showing oAHI < 5 events/hour. Rather than suggesting epiglottopexy should be performed on all patients undergoing lingual tonsillectomy, these results demonstrate concurrent surgery can be safe and effective when indicated based on DISE, congruent with previous publications on lingual tonsillectomy. These results further support the role of DISE-directed pediatric airway surgery in the management of refractory or complex airway obstruction.

The included cohort includes a large percentage of patients who have Trisomy 21 or Trisomy 18, making up 54.2% of patients. Although this is much higher than a representative portion of the general pediatric population, the risk of multilevel airway obstruction and base of tongue obstruction in patients with Trisomy 21 has been well-established [2, 4, 6, 8, 9]. Although this does limit generalizability of the data to all pediatric patients, the authors do not believe this rate of patients with chromosomal abnormalities significantly differs from the patient population of children with dynamic airway collapse on DISE which is amenable to repair with lingual tonsillectomy and epiglottopexy. Within our institution, a PSG is preformed when patients have symptoms of SDB without adenotonsillar hypertrophy; and if OSA is present a DISE is performed. Additionally, patients who have persistent OSA after an adenotonsillectomy undergo a DISE.

One patient did note an increase rather than a decrease in AHI from 1.6 to 2.0 events per hour. This increase in AHI of 25%, however, represents a variance of 1 breath every 2.5 h, leading the authors to believe this may be an insignificant change. Additionally, this particular patient noted a decrease in REM oAHI of 73.2%, from 11.2 to 3.0. Night-to-night variability in PSG within adult patients with OSA is well-documented [16,17,18], although data specifically geared towards pediatric patients are somewhat varied in their conclusions [19,20,21,22,23]. Based on levels of success from other patients and a small absolute change in oAHI, this particular patient’s results were believed to be due to PSG variability rather than a worsening of sleep parameters in the postoperative setting, especially given the dissonance between oAHI and REM oAHI.

Two patients with valid pre- and postoperative PSG data did not meet criteria for successful surgery (Patients 2 and 5 in the below tables). Other than severe obstructive sleep apnea, these patients had little in common precluding definitive conclusions as to possible risks of unsuccessful surgery. Patient 2 had many risk factors for OSA not amenable to surgical treatment and generalized hypotonia including trisomy 21, hypothyroidism, non-alcoholic steatohepatitis, severe asthma, and obesity. Patient 5 had cerebral palsy with hypotonia, significant developmental delay, seizures, scoliosis, and dihydropyrimidine dehydrogenase deficiency. Although both of these patients do have hypotonia, there were many patients within the cohort with similar risk factors (> 50% with trisomy 21 within this population) who reached surgical success. Interestingly, these two patients were entirely separate from the three patients who underwent subsequent sleep surgeries. Two of these patients did not have a preoperative PSG, while the other achieved a reduction in oAHI of 55.7 events per hour (69.9%) from 79.7 to 24.0.

There were no incidences of bleeding or dysphagia identified with this simultaneous procedure in the examined cohort. Only one previous study—a series of five patients—has addressed this, finding no increased incidence of postoperative dysphagia, consistent with the results therein. [13].

With relatively uncommon procedures, such as the one described here, some differences in surgical technique and instrumentation invariably exist [5,6,7,8,9, 12]. Previously described with the use of coblation and laser, within this cohort three surgeons performed the procedure entirely with coblation and the fourth with microdebrider and coblation. Given the relatively small cohort and multiple other confounders, it is difficult to draw any conclusions regarding the optimal surgical method. However, analysis showed no statistical difference in sleep outcomes when comparing the two methods. Larger studies with more surgeons would be needed to sufficiently power an analysis of optimal technique.

Major limitations of this study are inherent to its retrospective nature and the relative rarity of this procedure leading to a small patient cohort. Surgeries were performed by a variety of attending physicians and trainees, leading to variability in the technical steps of the surgery. Additionally, only a bit more than half of the patients (13/24) underwent both a pre- and postoperative sleep study. Although this was due to many factors, such as resolution of clinical symptoms or transition of care outside of the hospital system, it draws attention to the need for standardized care in these complex patients, which can drive opportunities for clinical research. These patients were included in general information on the demographics, and indications and complications of surgery to lessen risk of a type 2 error, but excluded from objective sleep analysis for completeness. There is a large component of patients within the study population with chromosomal abnormalities, which limits applicability to all pediatric patients. Additionally, while multiple surgeons performed the procedures, they were all performed at one institution, which may limit generalizability in separate practice environments. Patients were followed for less than 5 years, which does introduce the possibility of long-term sequelae from the surgery, or a lack of preservation of results.

As a relatively large cohort for an uncommon procedure, this study represents a significant contribution to the existing literature on tongue base surgery for refractory pediatric OSA. This study provides further data for targeted upper airway intervention for OSA, contextualized with previously existing literature documenting success of lingual tonsillectomy as a standalone procedure. While algorithms in management have been established previously, this expands on those to include information on steps that can be taken after levels of obstruction are identified to successfully treat OSA surgically in children [24]. Certainly, more research is required to better elucidate optimal candidacy for and clinical results of DISE-directed sleep surgery. However, for this cohort of patients—with lingual tonsillar hypertrophy and epiglottic prolapse—simultaneous correction of these levels was both safe and effective. This provides further data to support targeted upper airway intervention for complex OSA.