In this cross-sectional study, we evaluated the prevalence of RE using esophagogastroduodenoscopy and found that the non-obese OSAS group (31.0%) had a significantly higher prevalence than the non-OSAS group (control group, 9.5%). Moreover, compared to the control group, the obese OSAS group had a higher prevalence of RE (obese OSAS vs. control: 30.4% vs. 9.5%) and a higher prevalence of esophageal hiatal hernia (obese OSAS vs. control: 71.7% vs. 47.6%).
Although a 2019 meta-analysis study  reported a strong and clear association between GERD and OSA(H)S, it cannot be concluded that OSAS is a direct cause of GERD. Determining the presence of RE by pH/impedance monitoring analyses and EGD can be useful in diagnosing GER. However, many studies have small sample sizes due to the invasive nature of these tests. On the other hand, using a non-invasive questionnaire can increase the sample size, but it is inferior to EGD as proof of GER because it may include symptoms caused by other diseases such as esophageal cancer and functional dyspepsia. Although one previous study used EGD to evaluate RE in patients with OSAS who were diagnosed by polysomnography , this study is, to our knowledge, the first report that has used EGD to evaluate RE in patients with non-obese OSAS, as well as the first discussion of their potential association in non-obese patients with OSAS.
Association between RE and non-obese OSAS
In a study comparing the prevalence of GER by distal esophageal pH monitoring in 63 patients with OSA and in 41 individuals without OSA diagnosed using polysomnography, patients with OSA had significantly more total reflux within 8 hours than non-OSA patients (115 vs. 23; P < 0.001) and a significantly higher percentage of patients with pH < 4 hours which used in a diagnosis of GERD (21.4% vs. 3.7%; P < 0.001) . In a Chinese study of 177 patients diagnosed with GERD and healthy controls without GERD-related symptoms or endoscopic esophagitis (97 with GERD and 80 controls) which used the Berlin Questionnaire to diagnose OSAS, significantly more patients in the GERD group had a higher risk of OSAS than those in the control group (36.1% vs. 17.5%, P=0.005). In particular, patients with erosive reflux disease had a higher proportion of high-risk OSAS than the non-erosive reflux disease and healthy controls group (53.3% vs. 20.8% and 17.5%, respectively, P=0.001) . The only previous study that used EGD to evaluate RE in 243 patients who were divided according to polysomnography findings into OSA (98 patients) and non-OSA (145 controls) showed that endoscopy-confirmed esophagitis (p = 0.027) was significantly more prevalent among patients with OSA than among those without . In our study, both non-obese and obese patients with OSAS had a significantly higher prevalence of RE than sleep disordered controls without OSAS (non-obese OSAS vs. control: 31.0% vs. 9.5%; P<0.05, obese OSAS vs. control: 30.4% vs. 9.5%; P<0.05; Table 3), This result is consistent with the results of previous studies mentioned above. The finding of RE confirmed by EGD is a good indicator of GER [9–11]. GER is a cause of Barrett's esophageal cancer, which is become an increasing problem and requires close surveillance [1–8]. Thus, regular surveillance of EGD is necessary not only in obese patients at risk for GERD, but also in non-obese patients with OSAS.
Correlation between severity of OSA and RE in non-obese OSAS patients
In this study, multivariate analysis showed that the AHI, which correlates with the severity of illness in non-obese patients with OSAS, was not a significant independent risk factor for RE (Table 4). As the end-inspiratory intra-esophageal pressure decreases progressively during OSA, the mechanism most likely to underlie the development of GERD in patients with OSA is decreased intrathoracic pressure due to increased inspiratory effort in the face of upper airway occlusion. However, in a study of eight patients with OSA and RE, nine OSA patients without RE, and eight healthy controls, in which polysomnography with concurrent esophageal manometry and pH recording were performed, it was reported that in OSA patients, GER was mainly caused by transient lower esophageal sphincter relaxation, but not by negative intra-esophageal pressure due to OSA . In addition, there are many reports showing no association between GER and OSA severity [26, 27, 28]. The results of this study support these findings.
Risk factors for RE in non-obese patients with OSAS
Several studies have reported a high prevalence of RE in patients with OSAS, including some which used questionnaires to diagnose it [23, 29]. However, many studies have reported no association between OSA severity and GER. Many studies have reported that obesity is the cause of GERD in patients with SAS considering that obesity is a common risk factor [27, 28, 30–33]. Although this study was conducted on non-obese patients with OSAS, OSA severity (AHI) was not a significant independent risk factor associated with RE, similarly to what was found in previous studies. One of the aforementioned studies, which used EGD to evaluate RE in patients with OSAS who were diagnosed by polysomnography, pointed out that esophageal hiatal hernia is a risk factor for RE . However, since esophageal hiatal hernia is strongly associated with obesity , statistical adjustment is required if obesity and hernia are to be included as explanatory variables simultaneously in a multivariate analysis. In this study, when obese OSAS and non-obese patients with OSAS were separately compared with controls, obese patients showed a significantly greater prevalence of hiatal hernia than controls, while non-obese OSAS patients did not show a significant difference compared to controls (Table 3). Since obesity is a confounding factor, this study examined the association between OSAS and GER excluding the obesity factor by using a group non-obese patients with OSAS. Thus, the results showed that alcohol consumption was a greater risk factor than OSA severity in non-obese patients with OSAS complaining of sleep disturbances (Table 4).
The association between alcohol consumption and OSAS has been previously reported [35–37]. In a cohort of 775 males, for each increment of one drink per day, those who consumed more alcohol had 25% greater odds of sleep disordered breathing, defined as AHI ≥5 (OR = 1.25, 95% CI = 1.07–1.46, p = 0.006) .
There are several possible mechanisms by which OSA is exacerbated by alcohol intake. The central nervous system depressant effects of alcohol may affect the respiratory centers that control the tone of pharyngeal muscles, thereby promoting upper airway obstruction by increasing the likelihood of pharyngeal closure during sleep. Alcohol may promote OSA by directly causing pharyngeal irritation and edema . Since these mechanisms are not related to obesity and may well occur in non-obese patients, they may contribute to the risk of having RE is non-obese OSAS patients.
However, an association between alcohol consumption and GER has also been reported. Alcohol consumption may also lower esophageal sphincter pressure. In a study of 20 healthy volunteers, ingestion of white wine (8% alcohol) lowered the esophageal sphincter pressure. The fraction of time at pH < 4 in the first hour after ingestion of 300 ml white wine was significantly higher than after ingesting water . In a study of 24 healthy volunteers, beer (7% alcohol) also increased the median fraction time with pH < 4 compared to water . In a study of 12 healthy volunteers without GERD, the consumption of 180 ml of vodka significantly increased reflux as evaluated by pH measurements . In addition, basic studies in rabbits reported that exposure of the esophageal mucosa to 10% ethanol had a direct detrimental effect on the epithelium, making it more susceptible to acidic injury . These reports support the possibility that alcohol is a risk factor for RE and that it is a risk factor for both OSAS and RE.
This study has several limitations. Since this was a cross-sectional study, the pathogenetic causality between OSA, alcohol consumption, and reflux esophagitis could not be determined. Moreover, the association was determined by the presence or absence of alcohol consumption, but the amount of alcohol consumed by each individual was not considered. Furthermore, as endoscopy cannot evaluate non-erosive GER and esophageal motility, comprehensive studies combined with pH monitoring or high-resolution manometry are still necessary to gain a better understanding of GERD in patients with OSA. Moreover, we regrettably did not evaluate our patients for Helicobacter pylori infection. The inverse association between Helicobacter pylori infection and erosive esophagitis might have influenced the results of the multivariate analysis, especially with respect to RE. Finally, the study was performed in a single center. A multicenter study is needed to further validate these results.