Obstructive sleep apnea (OSA) is a common clinical condition that causes upper airway collapse and airway flow reduction by repeated episodes of apnea and hypopnea during sleep and shows a prevalence of 2%–10% in the adult population [1,2]. OSA is associated with several cardiovascular diseases such as hypertension, arrhythmia, cerebrovascular events, and even neurocognitive impairment and motor vehicle accidents, significantly impacting morbidity and mortality [3-5]. Laryngopharyngeal reflux (LPR) is also a common health problem characterized by acid from the gastrointestinal tract that rises into and damages the upper airway tract mucosa. This condition accounts for approximately 10% of otolaryngology outpatient clinic patients who complain of various bothersome symptoms like chronic cough, a globus sense in the throat, voice change, and dysphagia [6,7]. LPR is diagnosed on the basis of self-reported symptoms, laryngoscopic findings, multi-channel intraluminal impedance, and pH monitoring [8]. However, in clinical settings, the reflux symptom index (RSI) [9] and reflux finding score (RFS) [10], developed by Belafsky, are preferred for their high tolerability, applicability, and low cost.

Several studies have revealed a substantial concurrence between LPR and OSA, which are estimated to co-occur in 30.6%–89.2% of cases, as supported in a recent meta-analysis by Magliulo et al. [11]. Previous studies have suggested an association between these two disease entities based on their common risk factors of obesity, male sex, alcohol intake, and old age [12,13], as well as clinical improvement in LPR following continuous positive airway pressure (CPAP) treatment [14-16]. This improvement further suggests a possible causative relationship between LPR and OSA, even though it remains unclear whether a direct causal link exists [17].

Consistent with our previous study, which found RSI improvement after multilevel OSA surgery [18], another recent study showed LPR symptom relief after OSA surgery [19]. However, these studies have limitations because they did not evaluate clinician-examined laryngoscopic findings and only included self-reported symptom changes based on the RSI questionnaire. The aim of this study was to investigate the therapeutic effect of multilevel OSA surgery on LPR in terms of both symptoms and larynx findings by assessing both RSI and RFS changes following surgery.

Due to the high co-occurrence of LPR in OSA patients [11-13,20,21], many clinicians have investigated clinical improvements in LPR after OSA management such as CPAP or surgery [14-16, 18,19,22]. Those studies confirmed LPR improvement after treatment, but only in terms of self-reported symptoms, except for our previous study reporting both RSI and RFS improvement after CPAP treatment [16]. In contrast, in this study, we demonstrated the clinical efficacy of OSA surgery on LPR based on both symptom relief and changes in laryngoscopic findings. Both laryngeal reflux symptoms and laryngoscopic findings significantly improved after OSA surgery, demonstrating more advanced results than our previous study that only evaluated the post-surgical RSI response [18].

Although several studies have revealed a relationship between LPR and OSA, the exact mechanism and direct correlation are still unclear. Several authors have suggested that increased negative intrathoracic pressure generated by increased respiratory effort during apnea-hypopnea events induces relaxation of the lower esophageal sphincter and repeated arousal, and shallow sleep caused by apnea-hypopnea events could increase susceptibility to gastroesophageal reflux (GER) in OSA patients, but no causal relationship has been clearly shown [23]. Rather, the role of the upper esophageal sphincter that protects the upper airway mucosa is thought to be more important in LPR than in GER. One crucial pathophysiological condition is lower esophageal sphincter dysfunction, which is considered a distinct disease entity and not solely a manifestation of GER [24,25]. Eventually, esophageal sphincter dysfunction results in laryngeal mucosal inflammation and laryngeal reflux symptoms even with minimal esophageal reflux. Such repetitive inflammation could also induce exacerbation of OSA by direct narrowing of the upper airway due to inflammation-mediated tissue hypertrophy and by the impairment of reflexes in the upper airway tract induced by sensory dysfunction. These result in an upper airway that is vulnerable to collapse, eventually becoming part of a vicious cycle [17,26].

Herein, we pose several hypotheses for why LPR is relieved after multilevel OSA surgery. First, OSA surgery could directly reduce the mechanical trauma that could be induced during snoring and airway collapse by resecting and correcting the flexible structure of the laryngopharyngeal airway. It can also reduce respiratory effort, lower intrathoracic pressure, and protect against GER. Second, by improving mouth breathing, one of the main clinical manifestations of OSA, the humidification function of the upper airway mucosa is preserved to promote laryngeal tissue healing and mucosal homeostasis. An animal study showed that repetitive pressure changes and collapse of the upper airway tract could induce inflammatory changes that can lead to tissue damage [27].

In our study, the RSI and RFS positive rates decreased after surgery from 28.6% to 17.9% and 32.1% to 0%, respectively. Additionally, most subdomains of the RSI and RFS significantly improved after surgery. Throat clearing, throat mucus, breathing difficulty, troublesome cough, and heartburn sensation showed significant improvements among the RSI domains, while all subdomains in the RFS except granuloma, which was not observed in any patient at any time point, improved significantly after surgery. We offer several hypotheses to explain the higher negative conversion rate of the RFS and the significant improvements in each subdomain compared with the RSI. Because the patients included in this study mainly complained of snoring and apnea, not LPR itself, patients with mild LPR might have been preferentially included in our study, and it is highly likely that some people with severe LPR were excluded because they received other treatment. Therefore, even in patients with positive baseline RSI and RFS, the highest observed values were not very close to the maximum possible score, and most patients scored just above the cut-off value, as shown in Fig. 1B. Therefore, even only slight improvements could have led to a relatively high conversion rate, especially for RFS responses because the maximum preoperative RFS was only 13, corresponding to just half of the perfect score. Furthermore, postoperative discomfort such as swallowing difficulty and globus sensation after OSA surgery could be confounding factors because those complaints have similarities to several subdomains in the RSI, which is a self-reported scoring system. These factors may explain why the conversion rate of the RSI was not as high as that of the RFS. Lastly, the small number of patients in this study might have contributed to this result.

We also analyzed whether there were differences in therapeutic effects according to OSA severity, BMI, age, and surgical outcomes measured by PSG. Each subgroup showed improvements in the RSI and RFS after surgery, but no changes were statistically significant except for changes in the RSI in the subgroup analysis according to surgical outcomes. Some explanations can be proposed for these findings. First, because not all patients underwent follow-up PSG, this subgroup analysis included only 23 patients after excluding five patients. It is possible that these five patients reported good surgical outcomes regarding OSA symptoms and therefore refused a follow-up test. Additionally, many patients who underwent OSA surgery at our center tended to have early loss to follow-up, before the 3-month postoperative point, which is also presumed to have been due to dramatic symptom improvement after surgery. This is unlike CPAP treatment, which requires regular follow-up, but it nevertheless contributed to the small sample size in this study and limited the representation of real-world phenomena. Second, in terms of RSI changes according to surgical outcomes, we should take into account that the preoperative RSI values for the unsuccessful group were about 1.5 times greater than those in the successful group (15.70 vs. 10.38), which could account for these differences. However, both groups showed the almost same postoperative RSI (about 8.5), implying that LPR symptom alleviation could be achieved regardless of the OSA outcome determined by PSG values. Additionally, unlike the RFS, because the RSI is a self-reporting questionnaire, postoperative discomfort could act as a confounding factor because it could not be clearly distinguished from most RSI subdomains. Third, the use of PSG and AHI as the sole surgical outcome measures could be debated, and reproducibility was not addressed as we conducted only one follow-up PSG examination. An interesting finding is that the changes in both the RFS and RSI were as much as twice as large in more obese patients (BMI ≥25 kg/m²: 3.79 and 5.11, respectively) than in less obese patients (BMI <25 kg/m²: 2.22 and 2.56, respectively), although the differences were not statistically significant. The more obese patients had higher average preoperative RSI and RFS (12.95 and 6.74, respectively) than the less obese patients (9.89 and 6.00, respectively), which also implies a stronger positive correlation between OSA and LPR in obese patients than non-obese patients, as reported in a previous study [28].

Despite promising results and advances beyond previous studies about the clinical effect of OSA treatment on LPR, this study has several limitations. First, this was a retrospective study without a control group, and it was not conducted in a blinded fashion. Second, our results could have been weakened by patient selection bias in each surgical modality, the small number of patients that were treated at a single institution and who were operated on by a single surgeon, and a short follow-up period. Third, differences in anatomical positions and the severity of anatomical obstacles by position could affect LPR severity. Furthermore, surgical outcomes in terms of LPR improvement might have been affected by the combination of surgical methods. While all 28 patients received UPPP, 22 underwent tongue base reduction surgery, 16 underwent septoplasty, and three underwent partial epiglottectomy, all of which were performed as part of multilevel surgery. Therefore, large and comprehensive studies regarding whether and how these anatomical factors influence LPR and surgical outcomes are needed. Fourth, even though we added the RFS as a more reliable and objective method, we did not include more objective tests like multichannel intraluminal impedance and pH monitoring. Novel techniques detecting LPR have recently been developed, such as the pepsin salivary test [29] and airway pH monitoring using an oropharyngeal probe [30]. Hence, further studies with a longer follow-up period and large cohort studies with more objective tools are required to validate our results.

In conclusion, for OSA management, multilevel surgery has potential therapeutic effects on the LPR in terms of both laryngeal reflux symptoms and laryngoscopic examination findings. Future studies with more objective tools are required to establish the clinical impact of OSA surgery on LRP in greater depth.