Association of Gastroesophageal Reflux With Malignancy of the Upper Aerodigestive Tract in Elderly Patients

Key Points

Question  Is gastroesophageal reflux disease (GERD) associated with malignancy of the upper aerodigestive tract (UADT)?

Findings  A case-control study of individuals in the United States aged 66 years and older was performed with data from the Surveillance, Epidemiology, and End Results (SEER)-Medicare database. GERD was associated with a greater odds of developing malignancy of the larynx, hypopharynx, oropharynx, nasopharynx, and paranasal sinuses.

Meaning  GERD is associated with the presence of malignancy of the UADT in the elderly population of the United States.

Importance  Chronic inflammatory states have been linked to the development of malignancy. Gastroesophageal reflux disease (GERD) is a known risk factor for esophageal adenocarcinoma as the end result of chronic inflammatory changes.

Objective  To investigate the association of GERD with the risk of malignancy in the upper aerodigestive tract (UADT).

Design, Setting, and Participants  We used the Surveillance, Epidemiology, and End Results (SEER)-Medicare database to conduct a case-control study of individuals in the United States who had been added from January 2003 through December 2011 and were 66 years or older. The study included patients diagnosed with malignancy of the larynx, hypopharynx, oropharynx, tonsil, nasopharynx, and paranasal sinuses. GERD was examined as an exposure. Controls were matched from a 5% random sample of Medicare beneficiaries without cancer. Multivariable unconditional logistic regression was performed.

Main Outcomes and Measures  Incidence of invasive malignancies of the UADT.

Results  A total of 13 805 patients (median [range] age, 74 [66-99] years; 3418 women [24.76%] and 10 387 men [75.24%]) with malignancy of the UADT were compared with 13 805 patients without disease and were matched for sex, age group, and year of diagnosis. GERD was associated with a greater odds of developing malignancy of the larynx (adjusted odds ratio [aOR], 2.86; 95% CI, 2.65-3.09), hypopharynx (aOR, 2.54; 95% CI 1.97-3.29), oropharynx (aOR, 2.47; 95% CI, 1.90-3.23), tonsil (aOR, 2.14; 95% CI, 1.82-2.53), nasopharynx (aOR, 2.04; 95% CI, 1.56-2.66), and paranasal sinuses (aOR, 1.40; 95% CI, 1.15-1.70).

Conclusions and Relevance  GERD is associated with the presence of malignancy of the UADT in the US elderly population. This epidemiological association requires further examination to determine causality and diagnostic utility.

Gastroesophageal reflux disease (GERD) has a significant association with the development of esophageal adenocarcinoma.^1-3 The natural history of this disease is thought to be due to chronic inflammation of the esophageal mucosa that results in cellular transformation.^4,5 Barrett esophagus represents metaplasia of the lower esophagus from a stratified squamous epithelium to a simple columnar epithelium, secondary to gastroesophageal reflux.^5,6 This metaplasia is a precursor to the development of esophageal adenocarcinoma and has been the impetus for screening esophagogastroduodenoscopy among patients with GERD complicated by dysphagia or hematemesis or in patients with multiple risk factors for Barrett esophagus.^7-11

There are other inflammatory conditions linked to the development of various malignancies.⁴ Inflammatory insults to the colonic mucosa in ulcerative colitis result in a high risk of developing colon cancer.¹² Hepatocyte inflammation in chronic hepatitis precedes progression into hepatocellular carcinoma.¹³ Furthermore, an association between previous sinonasal inflammatory disease and subsequent nasopharyngeal carcinoma and paranasal sinus cancer has recently been demonstrated in several world populations.^14-19

Malignancies of the upper aerodigestive tract (UADT) include cancers of the larynx, hypopharynx, oropharynx, tonsil, nasopharynx, and paranasal sinuses, and account for more than 800 000 cases of cancer and more than 360 000 deaths globally each year.²⁰ The etiology of these malignancies is thought to be multifactorial. Alcohol and tobacco use, viral infections, occupational exposures, and chronic inflammatory conditions of the upper airway mucosa have all been implicated in the development of UADT malignancies.^21-25 To this end, the association of GERD with malignancy beyond the esophagus has not been well defined.

To date, the studies examining the link between GERD and the development of UADT malignancies have been conflicting.^26-29 This association has never been examined in the United States elderly population, to our knowledge. Defining a potential association between GERD and subsequent malignancy of the UADT may contribute to the understanding of inflammatory disease as a risk factor for head and neck cancer, as well as identify an at-risk population that may benefit from increased surveillance.

The National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database was used for this study. SEER is an epidemiologic surveillance system consisting of population-based tumor registries designed to track cancer incidence and survival in the United States. The registries gather data such as demographics, incidence, and treatment information on newly diagnosed malignancies in geographically defined areas.³⁰ These areas represent approximately 28% of the population of the United States. The SEER-Medicare linked database provides information on medical services and disease status for Medicare beneficiaries. The database links cancer incidence and survival from SEER to Medicare claims from parts A (hospital, skilled nursing facility, and nursing home care, hospice, and home health services) and B (medically necessary and preventative services). Medicare is the primary insurer for 97% of people living in the United States older than 65 years.

The SEER-Medicare linked database was queried from January 2003 to December 2011 for cases under the “Oral Cavity and Pharynx” and “Other Respiratory System” cancer sites from the Patient Entitlement and Diagnosis Summary File (PEDSF), Medicare Provider Analysis and Review (MEDPAR), Carrier Claims (NCH), and Outpatient (OUTPT) databases. MEDPAR contains information about Part A services, containing 1 record per Medicare hospitalization; NCH contains information about Part B services with bills for services from physicians and other providers that occurred in the office or hospital. The OUTPT database also contains information about Part B services but contains Medicare outpatient claims including diagnosis, procedure codes, and demographic information. The PEDSF database contains records for each individual in the SEER database matched with the patient’s Medicare enrollment record. This study was approved by the institutional review board of Ochsner Clinic Foundation as required by SEER-Medicare data use policy.

The study cohort included patients ages 66 to 99 years diagnosed with malignancies of the larynx, hypopharynx, oropharynx, tonsil, nasopharynx, and paranasal sinuses using the International Classification of Diseases for Oncology, 3rd Edition topology and morphology codes from January 2003 to December 2011. Only patients who were alive at diagnosis and had at least 12 months of Part A and B non-HMO Medicare coverage prior to diagnosis were included to conclude that malignancy did not exist prior to diagnosis of GERD. GERD (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] diagnosis code 530.81) was examined as a comorbidity with carcinomas of the UADT, using inpatient and outpatient claims in the MEDPAR, OUTPT, and NCH data files. The diagnosis code for laryngopharyngeal reflux (ICD-9-CM code 478.79) was not adopted as an inclusion criteria in this study owing to the potential for unreliable reporting of that code. A 12-month exclusion period between diagnosis of GERD and diagnosis of UADT malignancy was used to reasonably conclude the malignancy did not exist prior to diagnosis of GERD.

Patients with morphology codes 8000 to 8576, 8940 to 8941, and topology codes C09.0 to C09.9 were categorized to tonsil; codes C10.0 to C10.9, oropharynx; codes C11.0 to C11.9, nasopharynx; C13.0 to C13.9; hypopharynx; C30.0 to C31.9, paranasal sinus; and C32.0 to C32.9, larynx. Our study excluded benign, in situ, and noninvasive cancers as determined by the fifth digit behavior code and only included invasive cancers.

Study controls were selected from the Summarized Denominator File (SUMDENOM), a 5% random sample of Medicare beneficiaries who live in the same geographic regions as those diagnosed with cancer. Controls were matched to each malignancy by sex, age category, and year of diagnosis. Only controls that were alive on July 1 (the midpoint) of the selected calendar year were included. For the cancer cases, only patients that had at least 12 months of Part A, Part B, and non-HMO Medicare coverage were included.

Patient demographic data including age at diagnosis, sex, race, urban or rural location, and geographical region were obtained from the PEDSF data files. Race was categorized as non-Hispanic whites (NHW), non-Hispanic blacks (NHB), or others, which include Asian/Pacific Islanders, Hispanic, American Indian, and Alaskan natives.

A case-control study was performed to examine the association between invasive cancers of the UADT and GERD. The cases and controls were matched using propensity score matching. The Pearson χ² test was used to determine statistical differences for bivariate analysis.

Unconditional logistic regression was performed to examine the association between malignancy of the UADT and GERD before and after adjustments for covariates. Covariates included in the adjusted model were race (NHW, NHB, and others), urban or rural status (metropolitan vs nonmetropolitan), SEER registries, and 3 variables used to match cases and controls (sex, age group, and diagnosis year). To account for the possibility that exposure lead time might affect outcomes, GERD exposure status and exposure time period was further examined and categorized as no GERD exposure before diagnosis, exposure less than 12 months, exposure 12 to 24 months, exposure 24 to 36 months, or exposure 36 months or longer. Overall survival for each group was calculated using the Kaplan-Meier method, and significant differences were compared using the log-rank test.

A secondary case-cohort analysis was conducted using a weighted sample of patients without cancer derived from the SUMDENOM file to calculate clinically relevant statistics including disease prevalence, attributable risk, relative risk, and number needed to harm. Patients without cancer obtained from the SUMDENOM file were weighted 20 times and analyzed with all UADT malignancies reported in the PEDSF file. Control patients were weighted 20 times because the SUMDENOM file represents a 5% random sample of Medicare beneficiaries in the same population as the cancer cases reported in the PEDSF file. The total population from which all cancer cases were derived could then be inferred for calculation of the cancer prevalence. The case-cohort study design maximizes the available data in this manner. In all analyses, significance was assessed at the .05 probability level.

A total of 27 610 patients were included from the SEER-Medicare linked database query; 13 805 patients had a diagnosis of carcinoma of the UADT and were included in the case group for analysis (Table 1). These cases included 60.5% laryngeal, 5.4% hypopharyngeal, 5.2% oropharyngeal, 14.4% tonsillar, 5.5% nasopharyngeal, and 9.0% paranasal sinus malignancies. The control group was comprised of 13 805 matched subjects. Overall, there were 20 774 men (75.24%) and 6836 women (24.76%).

GERD was associated with a greater odds of developing malignancy of the larynx (adjusted odds ratio [aOR], 2.86; 95% CI, 2.65-3.09), hypopharynx (aOR, 2.54; 95% CI, 1.97-3.29), oropharynx (aOR, 2.47; 95% CI, 1.90-3.23), tonsil (aOR, 2.14; 95% CI, 1.82-2.53), nasopharynx (aOR, 2.04; 95% CI, 1.56-2.66), and paranasal sinuses (aOR, 1.40; 95% CI, 1.15-1.70) (Table 2). Odds of a cancer diagnosis after GERD exposure were greatest for larynx and least for paranasal sinuses. Relative to NHWs, there was a higher odds of laryngeal cancer (aOR 1.30, 95% CI, 1.16, 1.46), and oropharyngeal cancer (aOR, 1.74; 95% CI, 1.18-2.57) than HNBs (Table 3). Patients in the others race group, which included Asians, Hispanics, American Indians, and Alaskan natives, had a higher odds of nasopharyngeal cancer (aOR, 2.38; 95% CI, 1.76-3.22) compared with NHWs.

There was a significant association between GERD and the presence of larynx, hypopharynx, oropharynx, tonsil, nasopharynx, and paranasal sinuses malignancy even when GERD exposure time is taken into account (Table 3). In addition, a survival analysis examining length of exposure to GERD in UADT malignancies showed that survival of patients remains similar regardless of exposure lead times (Figure). Patients with hypopharynx or oropharynx malignancies had worse survival rates than those having other types of UADT cancer.

In the secondary case-cohort analysis, the prevalence of malignancy in the nonexposed and exposed populations was examined across the UADT. Notably, laryngeal cancer was diagnosed in 0.72% of patients with GERD, compared with 0.21% of those without GERD. The relative risk of laryngeal carcinoma in patients with GERD was 3.47 (95% CI, 3.43-3.52), resulting in a number needed to harm of 195 (Table 4).

This longitudinal population-based study is the first to examine the association between GERD and the development of UADT malignancies in an elderly population in the United States 66 years and older. There is a significant association between GERD and the presence of malignancy at a variety of subsites in the head and neck. Chronic inflammation has been described as an etiology for malignancies of the esophagus,¹⁰ stomach,^31,32 liver,¹³ and colon,¹² and cellular mechanisms have been described for these processes.^4,31,32 An analogous mechanism may be operative in the UADT and would have implications for understanding the role of chronic inflammation in head and neck cancers. Regardless of causality, or lack thereof, the systematic nature of the association in this large, population-based study may be relevant for identifying an at-risk population, improving surveillance, and initiating earlier treatment.

The strongest association between GERD and the presence of malignancy was seen in the larynx. This intuitively makes sense owing to the proximity to the esophagus and the readily exposed mucosa that lines the larynx, resulting in reflux-related tissue injury, mucosal inflammation, and chronic laryngitis. GERD is thought to result in significant histopathologic changes.³³ Previous studies^34-36 have shown a link between the accumulation of bile acids in reflux and laryngeal carcinoma when compared with healthy control groups, as well as patients with biliary reflux in the setting of prior gastrectomy.^37,38 In the current study, as one moves from caudal to cranial, a persistent though decreasing association between the presence of malignancy and GERD was noted.

These data suggest that elderly patients with GERD in the United States are 3.47, 3.23, 2.88, and 2.37 times as likely as those without GERD to be diagnosed with laryngeal, hypopharyngeal, oropharyngeal, and tonsillar cancers, respectively (Table 4). Interestingly, 1 of every 195 patients with GERD had an associated diagnosis of laryngeal malignancy, though aside from causation this may have occurred through a variety of biases, including misdiagnosis, reverse causation, and surveillance bias. Additional study is necessary to examine this relationship. Still, this finding may have important potential clinical significance for those providing care to the aging population, as GERD is thought to be present in 10% to 30% of the population³⁹ and is more frequently encountered in patients who are obese and elderly.⁴⁰

A causal relationship between GERD and malignancies of the larynx and pharynx has been previously suggested though the association remains unclear. In one case-control study examining patients within the Veterans Health Administration (VHA), there was not an increased risk of laryngeal cancer among patients with GERD.²⁶ A second case-control study looking at patients in the VHA system found the prevalence of GERD was higher among hospitalized patients with laryngeal cancer (8.9% vs 4.0%; P < .001) and pharyngeal cancer (6.2% vs 3.8%; P < .001) when controlling for age, sex, smoking, and alcohol intake.²⁷ Others have suggested GERD may be either an independent carcinogenic factor or a cofactor associated with smoking and alcohol consumption in promoting the development of malignancies of the UADT.^28,29,35 We believe that our data support the hypothesis that the mucosa of the larynx and pharynx is susceptible to reflux-related tissue injury and GERD may be associated with the presence and development of malignancies at these subsites. The ideal study to investigate this relationship should evaluate GERD with regard to development of UADT malignancies in nonsmokers and nondrinkers.

Chronic inflammation of the esophageal mucosa in GERD leads to metaplasia, also known as Barrett esophagus. The diagnosis is a known intermediary to esophageal adenocarcinoma.^1,6,7 This is consistent with our hypothesis that GERD can be a precursor to malignancy of the UADT. It is unknown, however, if an intermediary pathologic state exists that might predict malignancy of the larynx, hypopharynx, oropharynx, tonsil, nasopharynx, or paranasal sinuses. Historically, gastroenterologists were able to demonstrate the importance of diagnosing and treating herald lesions. With this, a paradigm shift occurred as the screening esophagogastroduodenoscopy was implemented for patients at higher risk for developing Barrett esophagus and esophageal cancer.^9,11 A similar screening platform may benefit those patients at higher risk for the development of malignancy of the UADT, though further research is necessary.

The strengths of this study include its large sample size, population-based design, and use of validated cancer outcomes from the SEER cancer registries. These have been previously demonstrated to have high standards for completeness and accuracy. The patients were well matched for sex, age at time of diagnosis of malignancy, and year of diagnosis. As such, we believe the results of our analyses of Medicare beneficiaries may be generalizable to the elderly population of the United States, though causality and duration of exposure necessary to elicit carcinogenesis requires further study. This is corroborated in Figure in which exposure lead-time up to 36 months shows similar survival.

The data obtained from the SEER-Medicare-linked database may have inherent biases. It is unknown in these patients if the initial symptoms of the malignancy were misdiagnosed as GERD. Additionally, the means by which GERD was diagnosed is unknown, whether based on symptoms, endoscopic examination, or invasive testing. A diagnosis of laryngopharyngeal reflux was not adopted as an inclusion criterion for this study out of concern for the greater potential for unreliable reporting of that ICD-9-CM code compared with the code for GERD. Preliminary data modeling with both codes included was not substantially different than when the GERD code was used alone, and the decision was made to report the latter, more conservative data set. Symptoms such as change in taste, globus sensation, and postnasal drip are common complaints of GERD but may also be representative of early tumor development. Furthermore, elderly patients who present with symptoms of GERD may be more likely to undergo further medical evaluation, either with imaging, flexible scope, or surgery. This surveillance bias might result in identification of indolent tumors. Selection of this cohort included a 12-month exclusion period between the diagnosis of GERD and the diagnosis of malignancy of the upper aerodigestive tract. This reduces the likelihood of detecting a cancer during the surveillance period, though it is unknown how much time is necessary for GERD-induced inflammation to result in malignancy. Surveillance in this population may allow detection of malignancy and initiation of treatment earlier. Another limitation is lead time bias in the subjects diagnosed prior to 2005, as these patients did not have an exposure period greater than 36 months.

Other limitations of the SEER-Medicare database are well established. Importantly, data about tobacco and alcohol exposure, the most well-established risk factors for UADT malignancy, are not reported. These exposures would need to be taken into account to determine whether GERD is an independent risk factor rather than a cofactor in the onset of UADT carcinogenesis. Additional specific risk factors are also absent from the SEER-Medicare data. Known risk factors for the development of nasopharyngeal carcinoma includes viral infections such as Epstein-Barr virus,⁴¹ while environmental factors such as woodworking and metallurgy are associated with sinonasal malignancies. Human papilloma infection can lead to the development of oropharyngeal squamous cell carcinoma.⁴² Other variables that are not examined by the SEER-Medicare database include family history and dietary tendencies, as well as the severity and treatment patterns for GERD. Future directions for study would include examining the use of computed tomography or nasal endoscopy and flexible fiberoptic laryngoscopy in the screening and/or diagnoses of these malignancies in elderly patients with a diagnosis of GERD.

GERD is associated with the development of malignancy of the upper aerodigestive tract in an elderly population in the United States. The increased relative risk for laryngeal and pharyngeal cancers in this population suggests an opportunity for earlier detection and intervention. Future studies are necessary to determine this effect on a younger cohort and to investigate causation.

Corresponding Author: Edward D. McCoul, MD, MPH, Ochsner Clinic Foundation, Otorhinolaryngology, 1514 Jefferson Hwy, CT-4, New Orleans, LA 70121 (emccoul@gmail.com).

Accepted for Publication: October 5, 2017.

Published Online: December 21, 2017. doi:10.1001/jamaoto.2017.2561

Author Contributions: Dr McCoul had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: McCoul

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Riley, Wu, Marino.

Critical revision of the manuscript for important intellectual content: All authors.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.

Meeting Presentation: The abstract was presented at the Triological Society Annual Meeting at COSM; April 28 to 30, 2017; San Diego, California.

Disclaimer: All information and materials in this manuscript are original and have not been published elsewhere.