Advertisement

Obstructive Sleep Apnea Is a Risk Factor for Barrett's Esophagus

Published:September 13, 2013DOI:https://doi.org/10.1016/j.cgh.2013.08.043

      Background & Aims

      Common risk factors for obstructive sleep apnea (OSA) and Barrett's esophagus (BE) include obesity and gastroesophageal reflux disease (GERD). The aims of this study were to assess the association between OSA and BE and to determine whether the association is independent of GERD and body mass index (BMI).

      Methods

      Patients who had undergone a diagnostic polysomnogram and esophagogastroduodenoscopy were identified by using Mayo Clinic (Rochester, Minnesota) databases from January 2000–November 2011. They were randomly matched for age, sex, and BMI at time of polysomnogram into the following groups: BE but no OSA (n = 36), OSA but no BE (n = 78), both (n = 74), or neither (n = 74). Clinical and demographic variables were abstracted from medical records. The association between OSA and BE was assessed by using a multiple variable logistic model that incorporated age, sex, BMI, clinical diagnosis of GERD, and smoking history.

      Results

      Subjects with OSA had an 80% increased risk for BE compared with subjects without OSA (odds ratio, 1.8; 95% confidence interval, 1.1–3.2; P = .03). These findings were independent of age, sex, BMI, GERD, and smoking history. Increasing severity of OSA, measured by using the apnea-hypopnea index, was associated with an increased risk of BE (odds ratio, 1.2 per 10-unit increase in apnea-hypopnea index; 95% confidence interval, 1.0–1.3; P = .03).

      Conclusions

      In this case-control study, OSA was associated with an increased risk of BE, potentially through BMI and GERD independent mechanisms. Patients with OSA may benefit from evaluation for BE.

      Keywords

      Abbreviations used in this paper:

      AHI (apnea-hypopnea index), BE (Barrett's esophagus), BMI (body mass index), CI (confidence interval), CPAP (continuous positive airway pressure), EGD (esophagogastroduodenoscopy), GERD (gastroesophageal reflux disease), ICD-9 (International Classification of Diseases, 9th Revision), OR (odds ratio), OSA (obstructive sleep apnea), PSG (polysomnogram), SD (standard deviation)
      The prevalence of obesity and its associated conditions is rising worldwide, with significant impact on morbidity and mortality.
      • Murray C.J.
      • Vos T.
      • Lozano R.
      • et al.
      Disability-adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010.
      Obesity is considered a risk factor for gastroesophageal reflux disease (GERD), and a direct relationship exists between increasing body mass index (BMI) and GERD severity.
      • Jacobson B.C.
      • Somers S.C.
      • Fuchs C.S.
      • et al.
      Body-mass index and symptoms of gastroesophageal reflux in women.
      • Wajed S.A.
      • Streets C.G.
      • Bremner C.G.
      • et al.
      Elevated body mass disrupts the barrier to gastroesophageal reflux.
      Chronic GERD may lead to metaplasia of the distal esophagus, a premalignant condition known as Barrett's esophagus (BE). The pathogenesis of BE in obese patients is believed to be at least partially driven by mechanical disruption of the antireflux barrier by increased intra-abdominal pressure leading to worsening GERD.
      • Corley D.A.
      • Kubo A.
      • Levin T.R.
      • et al.
      Abdominal obesity and body mass index as risk factors for Barrett's esophagus.
      Obesity, in particular central obesity, may also lead to BE independent of GERD through inflammatory pathways associated with visceral adiposity.
      • Edelstein Z.R.
      • Farrow D.C.
      • Bronner M.P.
      • et al.
      Central adiposity and risk of Barrett's esophagus.
      • Nelsen E.M.
      • Kirihara Y.
      • Takahashi N.
      • et al.
      Distribution of body fat and its influence on esophageal inflammation and dysplasia in patients with Barrett's esophagus.
      Obesity is the strongest risk factor for obstructive sleep apnea (OSA), a common sleep-related breathing disorder characterized by repetitive upper airway narrowing or collapse, transient hypoxia, and arousal from sleep.
      • Caples S.M.
      • Gami A.S.
      • Somers V.K.
      Obstructive sleep apnea.
      OSA is associated with an increased risk of GERD as evidenced by symptomatology, esophageal pH monitoring, and endoscopic findings of esophagitis.
      • Demeter P.
      • Visy K.V.
      • Magyar P.
      Correlation between severity of endoscopic findings and apnea-hypopnea index in patients with gastroesophageal reflux disease and obstructive sleep apnea.
      • Ing A.J.
      • Ngu M.C.
      • Breslin A.B.
      Obstructive sleep apnea and gastroesophageal reflux.
      Studies looking at GERD in the morbidly obese with OSA suggest that the association of GERD and OSA is independent of BMI.
      • Sabate J.M.
      • Jouet P.
      • Merrouche M.
      • et al.
      Gastroesophageal reflux in patients with morbid obesity: a role of obstructive sleep apnea syndrome?.
      It is postulated that continued breathing efforts against the occluded upper airway during obstructive apnea creates more negative intrathoracic pressure, which is transmitted to the esophageal body and precipitates reflux.
      • Romero-Corral A.
      • Caples S.M.
      • Lopez-Jimenez F.
      • et al.
      Interactions between obesity and obstructive sleep apnea: implications for treatment.
      Esophageal manometry studies in patients with OSA show that a decline in esophageal body pressure is associated with OSA events. This is followed, however, by a compensatory increase in gastroesophageal junction and upper esophageal sphincter pressures that may actually help prevent reflux.
      • Kuribayashi S.
      • Massey B.T.
      • Hafeezullah M.
      • et al.
      Upper esophageal sphincter and gastroesophageal junction pressure changes act to prevent gastroesophageal and esophagopharyngeal reflux during apneic episodes in patients with obstructive sleep apnea.
      In fact, the temporal association between episodes of apnea and GERD has been inconsistent in studies that use concurrent polysomnography and esophageal pH monitoring.
      • Ing A.J.
      • Ngu M.C.
      • Breslin A.B.
      Obstructive sleep apnea and gastroesophageal reflux.
      • Penzel T.
      • Becker H.F.
      • Brandenburg U.
      • et al.
      Arousal in patients with gastro-oesophageal reflux and sleep apnoea.
      • Berg S.
      • Hoffstein V.
      • Gislason T.
      Acidification of distal esophagus and sleep-related breathing disturbances.
      Despite the lack of mechanical phenomena to explain GERD in patients with OSA, several studies demonstrate that GERD improves with the use of continuous positive airway pressure (CPAP), the standard therapy for OSA.
      • Kerr P.
      • Shoenut J.P.
      • Millar T.
      • et al.
      Nasal CPAP reduces gastroesophageal reflux in obstructive sleep apnea syndrome.
      • Friedman M.
      • Gurpinar B.
      • Lin H.C.
      • et al.
      Impact of treatment of gastroesophageal reflux on obstructive sleep apnea-hypopnea syndrome.
      • Tawk M.
      • Goodrich S.
      • Kinasewitz G.
      • et al.
      The effect of 1 week of continuous positive airway pressure treatment in obstructive sleep apnea patients with concomitant gastroesophageal reflux.
      Patients with OSA share several risk factors with patients with BE including obesity, GERD, older age, and male gender. The asymptomatic nature of BE makes it challenging to identify populations at risk for this condition that may benefit from screening. This study aims to determine the relationship between OSA and BE and to explore whether this relationship is independent of BMI and GERD.

      Methods

      This study was approved by the Mayo Clinic Institutional Review Board.

       Subject Identification and Data Collection

      We used institution-wide searchable databases that contain demographic and medical information for patients evaluated at the Mayo Clinic in Rochester, Minnesota. These databases were used to identify subjects who had undergone both a diagnostic polysomnogram (PSG) and esophagogastroduodenoscopy (EGD) from January 2000–November 2011. International Classification of Diseases, 9th Revision (ICD-9) codes were used to identify subjects from this cohort diagnosed with BE and/or OSA.
      Demographic information including age, gender, and race were abstracted for all study participants. Lifetime history of smoking was also recorded. BMI was calculated from measurement of weight and height by using the formula weight in kilograms divided by the square of height in meters within 1 year of the diagnostic PSG. A BMI ≥30 was used as a surrogate marker for central obesity.
      • Leggett C.L.
      • Nelsen E.M.
      • Tian J.
      • et al.
      Metabolic syndrome as a risk factor for Barrett esophagus: a population-based case-control study.
      The presence of GERD was identified by using the ICD-9 code 530.81. The use of acid suppression medications (eg, proton pump inhibitors, histamine-2 blockers) was recorded for all subjects, even for those who did not carry a diagnosis of GERD in their medical records.

       Barrett's esophagus subjects

      Patients with BE were initially identified by using the ICD-9 code 530.85. EGD reports were reviewed to confirm the diagnosis of BE based on the identification of at least 1 cm of columnar mucosa. Corresponding histopathology reports were reviewed for evidence of intestinal metaplasia (with or without dysplasia) on surveillance esophageal biopsies confirmed by an expert gastrointestinal pathologist. Endoscopy reports were also reviewed for the length of the maximal extent of the BE segment and presence and size of a hiatal hernia. The highest grade of dysplasia was determined from review of histology reports for all surveillance endoscopies. Endoscopic examinations were performed by board-certified gastroenterologists with expertise in the diagnosis of BE.

       Obstructive sleep apnea subjects

      Patients with OSA were identified by using the ICD-9 code 327.23. Data from PSG studies were abstracted including apnea-hypopnea index (AHI), arousal profile (arousal index, % movement related, % breathing related), and mean and minimal oxygen saturation. The AHI score represents the number of apneic (complete upper airway collapse with near-complete cessation of airflow) and hypopneic (partial upper airway collapse) episodes per hour. The presence of OSA was defined by an AHI score ≥5/h and confirmed by review of the diagnosis in the corresponding Sleep Medicine evaluation note. Severity of OSA was defined as mild when the AHI score was 5–15/h, moderate 15–30/h, and severe ≥30/h. Supportive evidence for the presence of OSA included arousal index >15 with a predominance of breathing-related arousals and a mean minimum oxygen saturation of <85%.
      Subjects were classified into 4 groups on the basis of the presence or absence of BE or OSA: with no diagnosis of BE or OSA, with a diagnosis of BE but not OSA, with no diagnosis of BE but with a diagnosis of OSA, and with a diagnosis of both BE and OSA. In patients who lacked a diagnosis of BE, endoscopy reports were reviewed for presence and severity of erosive esophagitis as a possible confounder. The severity of erosive esophagitis was defined by using the Los Angeles classification (grades A–D).
      • Armstrong D.
      • Bennett J.R.
      • Blum A.L.
      • et al.
      The endoscopic assessment of esophagitis: a progress report on observer agreement.
      The indication and findings for each EGD were also recorded in this group. In subjects who did not meet diagnostic criteria for OSA, PSG studies and Sleep Medicine evaluations were reviewed for an alternative diagnosis. The same parameters abstracted from PSG studies in subjects with OSA were also abstracted for subjects with no OSA.
      Groups were randomly matched in a 2:1 ratio for age (<60 vs ≥60 years), gender, and BMI (<30 vs ≥30 kg/m2) at time of PSG. The group with the smallest number of subjects (ie, group with diagnosis of BE but not of OSA) defined the number of matched subjects in the remaining groups.

       Statistical Analyses

      Data were summarized as mean (standard deviation [SD]) for continuous scale variables (eg, age, BMI) and as frequency (percent) for discrete scale variables. Univariate and multiple variable logistic regression models were used to assess the associations of demographic and clinical characteristics with BE. In particular, the odds ratios (ORs) (95% confidence intervals [CIs]) for BE in those with OSA were estimated from the model coefficients (and their standard errors). A separate model that used severity of OSA as a predictor of BE and a separate model to predict the presence of a diagnosis of GERD were also examined. A separate model that used BMI ≥30 (as a surrogate marker for central obesity) vs BMI <30 explored the role of visceral adiposity in the association between BE and OSA. The analyses were done by using SAS software (Version 9.3; SAS Institute, Cary, NC). A 0.05 alpha level was considered statistically significant.

      Results

       Baseline Characteristics

      A total of 7482 subjects were identified as having undergone both a PSG and EGD in the time frame outlined for this study (Figure 1 for study flow). Two thousand four hundred eighty subjects had a diagnosis of neither OSA nor BE, 83 subjects had a diagnosis of BE but not of OSA, 4641 subjects had a diagnosis of OSA but not of BE, and 278 subjects had a diagnosis of both BE and OSA. The medical records of subjects in the smallest group (diagnosis of BE but not of OSA) were reviewed to confirm the presence and absence of BE and OSA, respectively. Seven subjects lacked a formal diagnosis of BE based on endoscopic and histologic criteria, and 40 subjects carried a diagnosis of OSA based on PSG results. This resulted in 36 patients with a confirmed diagnosis of BE and absence of OSA. Subjects in this group were frequency matched by using a 2:1 ratio on age, gender, and BMI at the time of the diagnostic PSG. The presence or absence of BE and/or OSA was confirmed by review of individual medical records in each of the matched groups. A total of 74 subjects were identified as having neither OSA nor BE, 74 as having a diagnosis of both OSA and BE, and 78 as having a diagnosis of OSA but not BE.
      Figure thumbnail gr1
      Figure 1Flow-sheet of subject identification. A population-based patient database of 7482 subjects who underwent both PSG and EGD was screened for OSA and BE by using ICD-9 codes. Two thousand four hundred eighty subjects had a diagnosis of neither OSA nor BE, 4641 subjects had diagnosis of OSA but not BE, 278 subjects had a diagnosis of both BE and OSA, and 83 subjects had a diagnosis of BE but not OSA. A diagnosis of BE was confirmed by review of endoscopic records for evidence of ≥1 cm of columnar mucosa with histologic evidence of intestinal metaplasia. A diagnosis of OSA was confirmed by AHI ≥5/h and review of Sleep Medicine evaluation records. Patients with diagnosis of BE but not OSA accounted for the smallest group. In this group, 36 patients had confirmed diagnosis of BE and absence of OSA (7 subjects lacked formal diagnosis of BE, and 36 subjects carried diagnosis of OSA). These subjects were randomly matched in a 2:1 ratio on age, gender, and BMI to subjects in the remaining groups.
      Baseline characteristics of participants are summarized in Table 1. Overall, 188 participants (72%) were male, with a mean (SD) age of 61 (13) years. The majority (99%) of participants were non-Hispanic white. A total of 172 participants (66%) had a history of tobacco use with no association with BE/OSA group status. Overall, 148 subjects (56%) had GERD symptoms requiring treatment with acid suppression medications. A total of 86 of 110 BE patients (78.2%) and 101 of 152 OSA patients (66.4%) had a diagnosis of GERD in their medical records. Subjects with OSA and BE as well as subjects with no OSA and a diagnosis of BE had significantly increased odds for a diagnosis of GERD relative to subjects with neither diagnoses (OR, 3.6; 95% CI, 1.8–7.4, BE and OSA, and OR, 3.5; 95% CI, 1.4–8.7, BE only).
      Table 1Baseline Characteristics of Subjects
      BE/no OSA (N = 36)BE/OSA (N = 74)No BE/OSA (N = 78)No BE/no OSA (N = 74)P value
      Age, mean (SD) (y)58 (13)64 (11)64 (12)57 (17)
      Subjects matched on age, gender, and BMI at time of PSG.
      Male, n (%)25 (69)58 (78)56 (72)52 (70)
      Subjects matched on age, gender, and BMI at time of PSG.
      BMI, mean (SD) (kg/m2)30 (5)33 (8)32 (6)31 (7)
      Subjects matched on age, gender, and BMI at time of PSG.
      White race, n (%)32/32 (100)70/72 (97)76/77 (99)70/71 (99).91
      GERD on acid suppression medication, n (%)21 (58)56 (76)39 (50)32 (43)<.001
      Smoking history, n (%)23/36 (64)54/73 (74)51/76 (78)44/74 (59).31
      a Subjects matched on age, gender, and BMI at time of PSG.
      In patients with BE, the mean (SD) segment length was 3.1 (2.6) cm. Seventy-four (67.2%) had no dysplasia, 24 (21.8%) had low-grade dysplasia, 8 (7.3%) had high-grade dysplasia, and 4 (3.6%) had cancer. A hiatal hernia was identified in 90 of 110 patients (82%) with BE, with a mean (SD) size of 3.1 (1.8) cm. In patients with a diagnosis of both BE and OSA, the diagnosis of BE preceded or was made within a year of OSA diagnosis in 92% of subjects (range, 12 years before to 4 years after a diagnosis of OSA).
      The main indication for EGD in patients without BE was dyspepsia/abdominal pain (n = 37), followed by reflux symptoms (n = 28), dysphagia (n = 25), anemia/bleeding (n = 24), diarrhea/vomiting (n = 10), and other (n = 28). Overall, 81 subjects (74%) in this group had normal EGDs. The most common endoscopic finding was esophagitis (n = 22), followed by gastritis (n = 11), Schatzki ring (n = 12), and Cameron erosions (n = 3). Ninety-one percent of cases of erosive esophagitis in patients without a diagnosis of BE were mild (grade A, n = 13; grade B, n = 7). Among patients with erosive esophagitis, there was no significant association between the presence of OSA and the grade of esophagitis (P = .42).
      Patients with OSA had a mean (SD) AHI score of 24.7 (21.9), of which 89 (59%) had moderate-to-severe OSA. The mean (SD) minimum oxygen saturation was 81.5% (8.0), and the mean (SD) arousal index was 42.6 (25.1), with 64.2% (SD = 24.2%) induced by obstructive breathing, further supporting the diagnosis of OSA. Subjects with no OSA had a mean (SD) AHI score of 1.6 (1.3), and the majority were diagnosed with snoring (n = 58), followed by periodic limb movement disorder (n = 15) and insomnia (n = 12).

       Association Between Obstructive Sleep Apnea and Barrett's Esophagus

      The association between OSA and BE was explored by using univariate models assessing age, gender, BMI, clinical diagnosis of GERD, and history of smoking (Table 2). GERD and OSA were associated with BE (OR, 3.2; 95% CI, 1.8–5.6; P < .001 and OR, 2.0; 95% CI, 1.3–3.2; P = .01, respectively). A multiple variable model examined the association of OSA with BE, adjusting for age, gender, BMI, clinical diagnosis of GERD, and smoking history. Subjects with OSA had approximately 80% increase in odds for having BE as compared with subjects without OSA or BE (OR, 1.8; 95% CI, 1.1–3.2; P = .03) (Table 2). There was no change in the association between OSA and BE in an alternative multiple variable model that included a combined variable: a diagnosis of GERD and use of acid suppressive medications (OR, 1.8; 95% CI, 1.0–3.1; P = .04) (Supplementary Table 1). A model that used the same variables but replaced OSA with AHI as a continuous variable examined whether the odds for BE increased with severity of OSA. An increased risk for BE with every 10-unit increase in AHI was observed (OR, 1.2; 95% CI, 1.0–1.3; P = .03), suggesting a dose-response association between BE and OSA. In a model including BMI ≥30 (as a surrogate marker for central obesity) vs BMI <30, the risk of BE in OSA remained unchanged (OR, 1.8; 95% CI, 1.1–3.3; P = .02). There was no significant association between increasing severity of OSA (AHI) and the length of BE (P = .65).
      Table 2Multiple Variable Logistic Model for BE Risk Factors
      Univariate modelMultiple variable model
      OR (95% CI)P valueOR (95% CI)P value
      Age (per 5 y)1.0 (0.9–1.1).651.0 (0.9–1.1).69
      Gender
       Male1.2 (0.7–2.2).431.2 (0.6–2.2).62
       Female1.0 (reference)1.0 (reference)
      BMI (per 5 units)1.1 (0.9–1.3).481.0 (0.8–1.3).73
      Smoking history
       Yes1.4 (0.8–2.4).221.4 (0.8–2.5).23
       No1.0 (reference)1.0 (reference)
      GERD
       Yes3.2 (1.8–5.6)<.0013.4 (1.9–6.0)<.0001
       No1.0 (reference)1.0 (reference)
      OSA
       Yes2.0 (1.3–3.2).011.8 (1.1–3.2).03
       No1.0 (reference)1.0 (reference)

      Discussion

      We performed a case-control study to explore the association of OSA and BE in a cohort of patients who underwent both a diagnostic PSG and EGD. In this subset of patients the presence of OSA was associated with 80% increased risk of BE compared with subjects without OSA and BE. This association was dose-dependent, with an increase in severity of OSA (measured by AHI) being associated with an increased risk of BE.
      Because the association of BE and OSA could be confounded by gastroesophageal reflux, we aimed to determine whether the observed relationship between OSA and BE was independent of a clinical diagnosis of GERD.
      • Kuribayashi S.
      • Massey B.T.
      • Hafeezullah M.
      • et al.
      Upper esophageal sphincter and gastroesophageal junction pressure changes act to prevent gastroesophageal and esophagopharyngeal reflux during apneic episodes in patients with obstructive sleep apnea.
      • Samelson C.F.
      Gastroesophageal reflux and obstructive sleep apnea.
      • Kuribayashi S.
      • Kusano M.
      • Kawamura O.
      • et al.
      Mechanism of gastroesophageal reflux in patients with obstructive sleep apnea syndrome.
      Both GERD and OSA were associated with BE on univariate analysis (Table 2). A multiple variable model including both OSA and GERD as covariates showed that both OSA and GERD were independently associated with an increased risk of BE (Table 2). In this study, GERD was defined as a clinical diagnosis made by a physician. The association of BE with OSA remained robust in a second multiple variable model that included a variable combining clinical diagnosis of GERD with the use of acid suppressive medications (proton pump inhibitors and histamine-2 blockers) as a marker for symptomatic reflux treated with medications (Supplementary Table 1). From these observations, we conclude that reflux is involved in the pathogenesis of BE in patients with OSA but is likely potentiated by other mechanisms such as central obesity (increased visceral abdominal fat) and a generalized systemic inflammatory state, which have been reported to be prevalent in OSA.
      • Romero-Corral A.
      • Caples S.M.
      • Lopez-Jimenez F.
      • et al.
      Interactions between obesity and obstructive sleep apnea: implications for treatment.
      • Calvin A.D.
      • Albuquerque F.N.
      • Lopez-Jimenez F.
      • et al.
      Obstructive sleep apnea, inflammation, and the metabolic syndrome.
      In our study cohort we found no direct association between BMI and the risk of BE in patients with OSA. Visceral fat has been implicated in the pathogenesis of BE through metabolic and proinflammatory pathways.
      • Nelsen E.M.
      • Kirihara Y.
      • Takahashi N.
      • et al.
      Distribution of body fat and its influence on esophageal inflammation and dysplasia in patients with Barrett's esophagus.
      • El-Serag H.B.
      • Kvapil P.
      • Hacken-Bitar J.
      • et al.
      Abdominal obesity and the risk of Barrett's esophagus.
      To explore the role of visceral adiposity, we used BMI ≥30 as a surrogate marker for central obesity.
      • Leggett C.L.
      • Nelsen E.M.
      • Tian J.
      • et al.
      Metabolic syndrome as a risk factor for Barrett esophagus: a population-based case-control study.
      We found that the association between OSA and BE remained significant despite adjusting for central obesity, raising the possibility that OSA may increase the risk of BE through inflammatory pathways independent of visceral adiposity. Studies suggest that OSA is associated with a heightened state of systemic inflammation that is mediated in part by the generation of reactive oxygen species from repetitive hypoxia.
      • Suzuki Y.J.
      • Jain V.
      • Park A.M.
      • et al.
      Oxidative stress and oxidant signaling in obstructive sleep apnea and associated cardiovascular diseases.
      • Schulz R.
      • Mahmoudi S.
      • Hattar K.
      • et al.
      Enhanced release of superoxide from polymorphonuclear neutrophils in obstructive sleep apnea: impact of continuous positive airway pressure therapy.
      Increased levels of several inflammatory markers including cytokines, tumor necrosis factor-α, interleukin-6, and C-reactive protein have been reported in patients with OSA independent of BMI.
      • Calvin A.D.
      • Albuquerque F.N.
      • Lopez-Jimenez F.
      • et al.
      Obstructive sleep apnea, inflammation, and the metabolic syndrome.
      This state of inflammatory derangement can somewhat be reversed with treatment of OSA with CPAP.
      • Ishida K.
      • Kato M.
      • Kato Y.
      • et al.
      Appropriate use of nasal continuous positive airway pressure decreases elevated C-reactive protein in patients with obstructive sleep apnea.
      The design of our study limits our potential to further explore a specific mechanism for how OSA predisposes to BE. OSA may promote BE through proinflammatory pathways independent of BMI and perhaps central obesity, in addition to GERD. The lack of anthropometric measurements limits our definition of central obesity, although previous studies have demonstrated that BMI ≥30 is an appropriate predictor of increased visceral adipose tissue in both sexes.
      • Oka R.
      • Miura K.
      • Sakurai M.
      • et al.
      Comparison of waist circumference with body mass index for predicting abdominal adipose tissue.
      We recognize that a clinical definition of GERD that is not established through symptom questionnaires and confirmed by pH impedance is a limitation. We are consequently not able to account for the frequency and severity of gastroesophageal reflux events in our study population. This limits our ability to definitely assess the association of OSA with gastroesophageal reflux. However, the association of BE with OSA remained consistent with 2 definitions of reflux. Our study is also potentially limited by selection bias, because all subjects were referred for a PSG as a result of suspicion of a sleep disorder. Arousals during shallow sleep are associated with transient lower sphincter relaxation and a risk of GERD. We rigorously screened PSG studies and clinical notes of patients without a diagnosis of OSA for potential confounders and found that most of these patients suffered from snoring without any evidence of other sleep disorders. To our knowledge, snoring is not associated with an increased incidence of transient lower sphincter relaxation. Finally, the use of ICD-9 codes to diagnose BE and OSA may have initially overestimated the prevalence of these conditions in our population.
      • Corley D.A.
      • Kubo A.
      • DeBoer J.
      • et al.
      Diagnosing Barrett's esophagus: reliability of clinical and pathologic diagnoses.
      A rigorous review of patient medical records, including endoscopy and pathology reports, helped confirm the presence or absence of these diagnoses in our final cohort of 262 subjects.
      In this study, we describe that patients with untreated OSA are at an 80% increased risk of BE and that increased severity of OSA may heighten this risk. Because of the asymptomatic nature of BE and the associated higher risk of esophageal adenocarcinoma, patients with OSA may benefit from screening for BE. Prospective studies are necessary to confirm this association in the general population and to explore whether this risk is reversed with OSA treatment.

      Supplementary Material

      Supplementary Table 1Multiple Variable Logistic Model Replacing GERD With GERD + Use of Proton Pump Inhibitors
      OR (95% CI)P value
      Age (per 5 y)1.0 (0.9–1.1).55
      Gender
       Male1.3 (0.7–2.3).44
       Female1.0 (reference)
      BMI (per 5 units)1.1 (0.9–1.3).53
      GERD + acid suppression medication
      Acid suppression medications include proton pump inhibitors and histamine-2 blockers.
       Yes2.6 (1.5–4.4)<.001
       No1.0 (reference)
      OSA
       Yes1.8 (1.02–3.1).04
       No1.0 (reference)
      a Acid suppression medications include proton pump inhibitors and histamine-2 blockers.

      References

        • Murray C.J.
        • Vos T.
        • Lozano R.
        • et al.
        Disability-adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010.
        Lancet. 2010; 380: 2197-2223
        • Jacobson B.C.
        • Somers S.C.
        • Fuchs C.S.
        • et al.
        Body-mass index and symptoms of gastroesophageal reflux in women.
        N Engl J Med. 2006; 354: 2340-2348
        • Wajed S.A.
        • Streets C.G.
        • Bremner C.G.
        • et al.
        Elevated body mass disrupts the barrier to gastroesophageal reflux.
        Arch Surg. 2001; 136: 1014-1019
        • Corley D.A.
        • Kubo A.
        • Levin T.R.
        • et al.
        Abdominal obesity and body mass index as risk factors for Barrett's esophagus.
        Gastroenterology. 2007; 133 (quiz 311): 34-41
        • Edelstein Z.R.
        • Farrow D.C.
        • Bronner M.P.
        • et al.
        Central adiposity and risk of Barrett's esophagus.
        Gastroenterology. 2007; 133: 403-411
        • Nelsen E.M.
        • Kirihara Y.
        • Takahashi N.
        • et al.
        Distribution of body fat and its influence on esophageal inflammation and dysplasia in patients with Barrett's esophagus.
        Clin Gastroenterol Hepatol. 2012; 10 (quiz e61–e62): 728-734
        • Caples S.M.
        • Gami A.S.
        • Somers V.K.
        Obstructive sleep apnea.
        Ann Intern Med. 2005; 142: 187-197
        • Demeter P.
        • Visy K.V.
        • Magyar P.
        Correlation between severity of endoscopic findings and apnea-hypopnea index in patients with gastroesophageal reflux disease and obstructive sleep apnea.
        World J Gastroenterol. 2005; 11: 839-841
        • Ing A.J.
        • Ngu M.C.
        • Breslin A.B.
        Obstructive sleep apnea and gastroesophageal reflux.
        Am J Med. 2000; 108 (120S–125S)
        • Sabate J.M.
        • Jouet P.
        • Merrouche M.
        • et al.
        Gastroesophageal reflux in patients with morbid obesity: a role of obstructive sleep apnea syndrome?.
        Obes Surg. 2008; 18: 1479-1484
        • Romero-Corral A.
        • Caples S.M.
        • Lopez-Jimenez F.
        • et al.
        Interactions between obesity and obstructive sleep apnea: implications for treatment.
        Chest. 2010; 137: 711-719
        • Kuribayashi S.
        • Massey B.T.
        • Hafeezullah M.
        • et al.
        Upper esophageal sphincter and gastroesophageal junction pressure changes act to prevent gastroesophageal and esophagopharyngeal reflux during apneic episodes in patients with obstructive sleep apnea.
        Chest. 2010; 137: 769-776
        • Penzel T.
        • Becker H.F.
        • Brandenburg U.
        • et al.
        Arousal in patients with gastro-oesophageal reflux and sleep apnoea.
        Eur Respir J. 1999; 14: 1266-1270
        • Berg S.
        • Hoffstein V.
        • Gislason T.
        Acidification of distal esophagus and sleep-related breathing disturbances.
        Chest. 2004; 125: 2101-2106
        • Kerr P.
        • Shoenut J.P.
        • Millar T.
        • et al.
        Nasal CPAP reduces gastroesophageal reflux in obstructive sleep apnea syndrome.
        Chest. 1992; 101: 1539-1544
        • Friedman M.
        • Gurpinar B.
        • Lin H.C.
        • et al.
        Impact of treatment of gastroesophageal reflux on obstructive sleep apnea-hypopnea syndrome.
        Ann Otol Rhinol Laryngol. 2007; 116: 805-811
        • Tawk M.
        • Goodrich S.
        • Kinasewitz G.
        • et al.
        The effect of 1 week of continuous positive airway pressure treatment in obstructive sleep apnea patients with concomitant gastroesophageal reflux.
        Chest. 2006; 130: 1003-1008
        • Leggett C.L.
        • Nelsen E.M.
        • Tian J.
        • et al.
        Metabolic syndrome as a risk factor for Barrett esophagus: a population-based case-control study.
        Mayo Clin Proc. 2013; 88: 157-165
        • Armstrong D.
        • Bennett J.R.
        • Blum A.L.
        • et al.
        The endoscopic assessment of esophagitis: a progress report on observer agreement.
        Gastroenterology. 1996; 111: 85-92
        • Samelson C.F.
        Gastroesophageal reflux and obstructive sleep apnea.
        Sleep. 1989; 12: 475-476
        • Kuribayashi S.
        • Kusano M.
        • Kawamura O.
        • et al.
        Mechanism of gastroesophageal reflux in patients with obstructive sleep apnea syndrome.
        Neurogastroenterol Motil. 2010; 22 (611–e172)
        • Calvin A.D.
        • Albuquerque F.N.
        • Lopez-Jimenez F.
        • et al.
        Obstructive sleep apnea, inflammation, and the metabolic syndrome.
        Metab Syndr Relat Disord. 2009; 7: 271-278
        • El-Serag H.B.
        • Kvapil P.
        • Hacken-Bitar J.
        • et al.
        Abdominal obesity and the risk of Barrett's esophagus.
        Am J Gastroenterol. 2005; 100: 2151-2156
        • Suzuki Y.J.
        • Jain V.
        • Park A.M.
        • et al.
        Oxidative stress and oxidant signaling in obstructive sleep apnea and associated cardiovascular diseases.
        Free Radic Biol Med. 2006; 40: 1683-1692
        • Schulz R.
        • Mahmoudi S.
        • Hattar K.
        • et al.
        Enhanced release of superoxide from polymorphonuclear neutrophils in obstructive sleep apnea: impact of continuous positive airway pressure therapy.
        Am J Respir Crit Care Med. 2000; 162: 566-570
        • Ishida K.
        • Kato M.
        • Kato Y.
        • et al.
        Appropriate use of nasal continuous positive airway pressure decreases elevated C-reactive protein in patients with obstructive sleep apnea.
        Chest. 2009; 136: 125-129
        • Oka R.
        • Miura K.
        • Sakurai M.
        • et al.
        Comparison of waist circumference with body mass index for predicting abdominal adipose tissue.
        Diabetes Res Clin Pract. 2009; 83: 100-105
        • Corley D.A.
        • Kubo A.
        • DeBoer J.
        • et al.
        Diagnosing Barrett's esophagus: reliability of clinical and pathologic diagnoses.
        Gastrointest Endosc. 2009; 69: 1004-1010

      Linked Article

      • Obstructive Sleep Apnea and Environmental Contributors for Barrett's Esophagus
        Clinical Gastroenterology and HepatologyVol. 12Issue 6
        • Preview
          Leggett et al1 concluded that obstructive sleep apnea (OSA) is associated with an increased risk for Barrett's esophagus (BE) via body mass index (BMI) and gastroesophageal reflux disease (GERD)–independent mechanisms. The authors1 considered that increased levels of inflammatory cytokines, including tumor necrosis factor α and interleukin (IL) 6, have been reported in OSA patients independent of BMI, and this inflammatory derangement state can be reversed with OSA treatment. They suggested that OSA may promote BE through proinflammatory pathways independent of BMI and perhaps central obesity, in addition to GERD, although they mentioned that “our study limits our potential to further explore a specific mechanism for how OSA predisposes to BE.”1 Notably, other series also reported that a high proportion of BE patients had a high risk for OSA.
        • Full-Text
        • PDF