Accepted Manuscript
Gastroesophageal and Laryngopharyngeal Reflux Associated with Laryngeal Malignancy: A Systematic Review and Meta-Analysis
Sean M. Parsel, DO, Eric L. Wu, MD, Charles A. Riley, MD, Edward D. McCoul, MD, MPH
PII: S1542-3565(18)31150-9 DOI: https://doi.org/10.1016/j.cgh.2018.10.028 Reference: YJCGH 56150
To appear in: Clinical Gastroenterology and Hepatology Accepted Date: 8 October 2018
Please cite this article as: Parsel SM, Wu EL, Riley CA, McCoul ED, Gastroesophageal and Laryngopharyngeal Reflux Associated with Laryngeal Malignancy: A Systematic Review and Meta-Analysis, Clinical Gastroenterology and Hepatology (2018), doi: https://doi.org/10.1016/ j.cgh.2018.10.028.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT
Title : Gastroesophageal and Laryngopharyngeal Reflux Associated with Laryngeal Malignancy: A Systematic Review and Meta-Analysis
Sean M. Parsel, DO 1, Eric L. Wu, MD 1, Charles A. Riley, MD 2, and Edward D. McCoul, MD, MPH 1, 3, 4
1 Tulane University School of Medicine, Department of Otolaryngology—Head and Neck Surgery, New Orleans, LA 2 Weill Cornell Medical Center, Department of Otolaryngology—Head and Neck Surgery, New York, NY 3 Ochsner Clinic Foundation, Department of Otorhinolaryngology, New Orleans, LA 4 University of Queensland School of Medicine, Ochsner Clinical School, New Orleans, LA
Short title: Reflux and Laryngeal Malignancy
Grant support: none
Correspondence Edward D. McCoul, MD, MPH 1514 Jefferson Highway, CT-4 New Orleans, LA 70121 Phone: (504) 842-4080 Fax: (504) 842-3979 [email protected] MANUSCRIPT
Financial disclosures or conflicts of interest: No authors have any financial disclosures of conflicts of interest to report.
Abbreviations BE: Barret esophagus CI: confidence interval GERD: gastroesophageal reflux disease HNSCC: head and neck squamous cell carcinoma HPV: human papilloma virus ICD: International Classification of Disease LPR: laryngopharyngeal reflux MINORS: Methodological Index for Non-Randomized Studies NSND: nonsmokers and nondrinkers OR: odds ratio PICOS: Population,ACCEPTED Intervention, Control, Outcome, and Study Design PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analysis RSI: Reflux Symptom Index
ACCEPTED MANUSCRIPT
Author Contributions SMP: systematic review of literature, data analysis, manuscript preparation, critical revision of manuscript ELW: systematic review of literature CAR: manuscript preparation, data analysis EDM: study concept and design, manuscript preparation, critical revision of manuscript
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
Background & Aims: Gastric reflux could lead to chronic mucosal inflammation and contribute to development of laryngeal malignancies, although there is controversy over this association. We performed a systematic review and meta-analysis to assess this relationship and determine risk of laryngeal malignancy in patients with reflux disease.
Methods: We performed a systematic review and meta-analysis, searching MEDLINE, EMBASE, and Web of Science databases from 1900 through April 9, 2018 for observational studies of adults reporting associations between gastroesophageal reflux disease (GERD) and/or laryngopharyngeal reflux and the risk of having or developing laryngeal malignancies. An itemized assessment of the risk of bias was conducted for each study that met inclusion criteria. The meta-analysis was performed using the Mantel-Haenszel method with random effects to account for heterogeneity. We performed subgroup analyses to determine effect of reflux type, study design, diagnostic method, and confounding variables
Results: Of the 957 studies that were identified during systematic review, 18 case– control studies met the criteria for analysis. Our meta-analysis demonstrated that reflux disease significantly increased the risk of laryngeal malignancy (odds ratio, 2.47; 95% CI, 1.90–3.21; P<.00001; I 2 = 94%). This association remained when we controlled for patient smoking and drinking (odds ratio, 2.07; 95% CI, 1.26–3.41). There was no statistically significant difference in risk of laryngeal malignancies between patients with GERD vs laryngopharyngeal reflux ( P=.44).
Conclusions: In a systematic review and meta-analysis, we found a significant association between reflux disease and the presenceMANUSCRIPT of laryngeal malignancy. Prospective studies should be performed to examine this relatio nship.
KEY WORDS: LPR, PICOS, esophagus, laryngeal cancer, head and neck cancer
What You Need To Know Background: Gastroesophageal and laryngopharyngeal reflux can lead to chronic mucosal inflammation and increase risk for development of cancer. Researchers have reported a correlation between reflux and laryngeal carcinoma, but there is controversy over this association.
Findings: In a meta-analysis of the literature, we found an association between reflux disease and laryngeal carcinoma. We found patients with reflux disease to be 2.47-fold more likely to have a diagnosis of laryngeal cancer.
Implications for patient care: Although alcohol and tobacco use are established risk factors for laryngealACCEPTED cancer, other factors require identification. Awareness of a potential role for reflux in laryngeal carcinogenesis should prompt vigilance among clinicians to evaluate susceptible individuals and potentially mitigate risk through medical therapy.
ACCEPTED MANUSCRIPT
Background
Laryngeal cancer is the 14 th most common cancer worldwide with approximately
157,000 new cases annually. 1 These cancers are often detected at advanced stages. 2 The aggressive nature of these cancers portends a poor 5-year overall survival rate at around
53%, with better survival for earlier stages. 3, 4 In order to improve detection, a better understanding of the risk factors involved in the development of laryngeal carcinoma is necessary, which in turn might improve survival and outcomes.
The major identifiable risk factors implicated in the development of laryngeal malignancy are alcohol and tobacco use. However, other risk factors are thought to contribute to tumorigenesis. 1 In younger patients diagnosed with laryngeal cancer, tobacco and alcohol use are less commonly associated with the development of malignancy. 5 In the nonsmoking and nondrinking (NSND) population with head and neck cancers, there is a lesser degree of genetic mutati MANUSCRIPTons compared with smokers and drinkers, suggesting alternate etiologies; however, the incidence of laryngeal subsite cancer is significantly less in this population. 6, 7 Human papilloma virus (HPV) infection is hypothesized to account for some of these malignancies; however, the incidence of
HPV-induced laryngeal cancer remains substantially lower than with oropharyngeal cancer. 8 Other risk factors likely have a role in development of malignancy.
Gastroesophageal reflux disease (GERD) and laryngopharyngeal reflux (LPR) have been hypothesized to be risk factors for developing laryngeal carcinoma, though this association remainsACCEPTED incompletely defined. 9-11 Many of the prior studies have poor controls, leading to some controversy, and there is a paucity of literature examining the effects of reflux disease on the laryngeal mucosa while controlling for covariates. 12-14 15 ACCEPTED MANUSCRIPT
Since the overwhelming majority of laryngeal cancers arise from the effects of smoking and drinking, finding a cohort without this exposure is difficult. Despite the lack of controls, early investigation suggests that laryngeal injury can occur from reflux of acidic and nonacidic gastric contents. 16 Further, animal models of acid exposure to upper airway mucosa have shown an increase growth in the number of tumors when the mucosa is exposed to hydrochloric acid (HCl) and pepsin. 17 Although HCl may play a role, much of the risk appears to be secondary to the effects of pepsin on the mucosa, which may lead to cellular injury. 18, 19
GERD has been linked to other types of cancers involving the aerodigestive tract.
Specifically, there is a role in the development of esophageal adenocarcinoma where reflux is associated with development of Barrett esophagus (BE) and subsequent dysplasia and malignant transformation. 20-22 Study of hypopharyngeal mucosa exposed to gastric reflux in vitro has demonstrated inflammato MANUSCRIPTry changes similar to that found in BE. 23-26 Other conditions leading to chronic and/or recurrent inflammatory insults have been implicated in the development of malignancy within the gastrointestinal tract including chronic hepatitis and hepatocellular carcinoma 27 as well as colon cancer and ulcerative colitis.28 This trend is also present in the upper aerodigestive tract, as allergic rhinitis and chronic rhinosinusitis have been associated with nasopharyngeal carcinoma. 29-31
Given the potential link between inflammation and the development of malignancy, ACCEPTEDthe relationship between reflux—either GERD or LPR—and laryngeal carcinoma was investigated. A previous systematic review examining reflux and the association of laryngeal and pharyngeal malignancies suggested a correlation between ACCEPTED MANUSCRIPT
GERD and laryngeal cancer but not pharyngeal cancer. 12 Since the publication of that meta-analysis, there has been a significant increase in the number of studies assessing the association of reflux disease and laryngeal cancer necessitating further inquiry.
Furthermore, recent interest in this subject as reflected in the published literature indicates a need for contemporary synthesis of the available evidence relating specifically to the relationship between GERD and/or LPR and laryngeal malignancy.
Methods
Study Design
A comprehensive review of literature was performed on April 9, 2018 from the
MEDLINE (1966 to April 2018), EMBASE (1975 to April 2018), and Web of Science
(1900 to April 2018) databases. Search criteria included all occurrences in the title or abstract of the terms: reflux, gastroesophageal reflux,MANUSCRIPT gastroesophageal reflux disease, laryngopharyngeal reflux, or heartburn plus the term laryngeal carcinoma, laryngeal cancer, or laryngeal malignancy. Specific electronic search criteria can be found in Table
S1. Inclusion criteria for the literature search were defined using the Population,
Intervention, Control, Outcome, and Study Design (PICOS) approach (Table S2). The
Population included was male and female adults (age > 18 years). The Intervention was a history of GERD and/or LPR. The Control was patients without history of GERD and/or
LPR. The Outcome was patients with previous diagnosis of GERD and/or LPR assessed against controlsACCEPTED to obtain the odds ratio (OR) with 95% confidence intervals (CI) for association with laryngeal carcinoma. The Study Design included case–control and cohort studies and randomized controlled trials. A literature search for studies assessing ACCEPTED MANUSCRIPT
risk of extralaryngeal malignancies in addition to laryngeal malignancy was also performed to capture additional citations for review. The Preferred Reporting Items for
Systematic Reviews and Meta-Analyses (PRISMA) standard was used to ensure adequate reporting of this systematic review.
Two reviewers (SMP and ELW) independently performed eligibility assessment of data in a standardized manner. Duplicate records were removed. The abstract of each citation was then screened for relevance to the assessment of reflux, GERD, or LPR in adults and their relationship to laryngeal malignancies. Irrelevant citations were excluded, as were case reports, case series, and studies without a comparison group. The full text of the remaining citations were obtained along with additional records from the reference lists of the published articles. The full-text articles were reviewed and non-eligible studies were excluded. Studies not written in the English language and those that did not provide patient demographics, location of malignanc MANUSCRIPTy, type of reflux, diagnostic criteria, and quantitative patient data were excluded. Conference proceedings were excluded if they did not include the necessary information, including patient demographics and methods of diagnosis. Articles examining benign laryngeal pathology were excluded, as were articles examining malignancies of other subsites of the upper aerodigestive tract that did not include the larynx. The remaining studies were included for qualitative and quantitative synthesis.
Data ExtractionACCEPTED
Data gathered from full-text articles included criteria for patient selection, study design, source of individual patient data, type of reflux, number of cases of malignancy, ACCEPTED MANUSCRIPT
and history of laryngeal malignancy. The principal summary measures were exposure to reflux disease and the subsequent development of laryngeal cancers. Raw data for the number of patients with and without exposure, as well as with and without disease, were extracted for quantitative meta-analysis. The risk of bias was assessed at the study level by examining each study design, authors’ stated purpose for the study, and source of patient data collection.
Risk of Bias and Quality Assessments
The methodologic quality of the studies was assessed using the Methodological
Index for Non-Randomized Studies (MINORS) method.32 This is an externally valid instrument developed for observational and non-randomized studies in the surgical specialties that demonstrates high inter-reviewer agreement as well as high test-retest reliability and internal consistency. This instrume MANUSCRIPTnt was used to provide a semiquantitative assessment of quality for non-randomized studies. For purposes of analysis, all comparative studies were independently scored in ternary units over 12 items. A score of 0 was used if the information was not reported; a score of 1 was used for inadequate report of information; and a score of 2 was used for adequate report of information. The ideal overall score is 24.
Meta-Analysis and Statistical Analysis
Meta-analysisACCEPTED was performed using the Mantel-Haenszel method with random effects. Heterogeneity was assessed using I-squared statistics. The association between previous reflux condition and malignancy was evaluated across all included studies by ACCEPTED MANUSCRIPT
determining the OR with 95% CI using the mid-P method. Heterogeneity was considered present if I2 > 25%. Sources of heterogeneity were investigated by sensitivity analysis.
Specifically, subgroup analyses were performed to determine the effects of reflux type
(GERD or LPR), study design (retrospective or prospective), diagnostic method, and methodologic quality on the overall effect. Funnel plots were generated for evaluation of publication bias and sensitivity analysis was performed to determine the effects of individual studies. Generic inverse variance method with random effects was used to perform ad hoc sensitivity analysis on NSND populations. Analysis was performed using
Review Manager 5.3 software (Copenhagen, Denmark).
Results
Systematic Review The initial database query identified 957 MANUSCRIPT articles, which were screened for relevance to the assessment of reflux conditions and the presence and/or development of laryngeal malignancy (Figure 1). After removal of duplicate records, 642 articles were considered for abstract review. Of these, 54 were excluded as non-English language studies. There were 27 articles included for full-text review. Eleven studies were excluded from quantitative analysis because insufficient data was reported or there was no control group. Manual searching of reference lists of the 16 full-text articles resulted in 3 additional eligible studies. One study was removed as it did not distinguish between laryngeal andACCEPTED pharyngeal carcinoma. Additional search for studies assessing reflux disease and extralaryngeal malignancies did not produce further citations meeting inclusion criteria. In total, 18 articles were included for systematic review and meta- ACCEPTED MANUSCRIPT
analysis. These resulted in a large number of cases (n = 36,634) and controls (n = 74,623) and demonstrated a trend toward positive association between GERD and/or LPR and the presence of laryngeal malignancy.
Data from risk assessment of bias for individual studies were collected (Table 1).
All 18 included studies were case–control design, of which 8 were retrospective and 10 included prospective data. The source of retrospective patient data was varied across the included studies. Five studies included data from large-scale databases. Retrospective patient data was taken from the US Department of Veterans Affairs Health System in 2 of
18 studies. 33, 34 Other sources of retrospective patient data included the SEER database in
1 study 35 and the Clalit Health Services in Israel in 1 study.36 The remainder of the studies included patient data from both outpatient clinics and inpatient populations. A summary of the populations is represented in Table 1. Analysis of methodologic quality was performed, which demonstrated a mean MINORS MANUSCRIPT score of 12.61 (95% CI: 12.01- 13.21). For analysis, studies following below the lower 95% CI were considered low in quality while studies falling above the upper 95% CI were considered high in quality.
A summary of risk exposure and associated disease data extracted from each study is presented in Table 1. GERD was assessed as a risk exposure in 13 of 18 studies,
LPR alone was assessed in 3 of 18 studies, and LPR and GERD were assessed in 2 of 18 studies. These latter two studies assessed the association of LPR and GERD separately within the cases and controls. For the studies assessing GERD and LPR, meta-analysis was performedACCEPTED for each exposure separately. The method of diagnosis of GERD and LPR varied in all studies. Objective measurements that included 24-hour pH probe monitoring and findings on endoscopy. Subjective diagnosis included use of patient reported ACCEPTED MANUSCRIPT
questionnaires and patient reported symptoms. Furthermore, chart review-based studies utilized clinical history based on International Classification of Disease (ICD) codes for diagnosis. The diversity of diagnostic methods is reported in Table 1. Nonsmokers were assessed specifically in only 1 study. Six studies performed multivariate analysis controlling for smoking and drinking histories of patients. One of these 7 studies assessed the effects of LPR on laryngeal cancer while the remaining 6 studies assessed the effects of GERD.
Meta-Analysis
Meta-analysis was performed on the 18 studies using a random effects model.
Overall, the studies demonstrated an OR of 2.47 (95% CI: 1.90-3.21). The P-value for overall effect was < 0.00001 and I2 = 94% (Figure 2). Subgroup analysis conducted assessing the type of reflux (either GERD or LPR) MANUSCRIPT demonstrated an OR of 2.37 (95% CI: 1.79-3.14; P < .00001; I2 = 95%) for GERD and an OR of 3.09 (95% CI: 1.69-5.65; P =
.0003; I 2 = 9%) for LPR. There was no statistically significant difference between the two group ( P = .44 and I2 = 0%). Analysis of diagnostic method was performed through subgroup analysis comparing subjective, objective, and review-based methods (Figure 3).
For purposes of analysis, subjective methods included use of patient-reported complaints via the Reflux Symptom Index (RSI), self-reported reflux disease, and patient-reported questionnaires. Objective methods included use of endoscopic findings and 24-hour pH probe monitoring.ACCEPTED Chart review-based methods were those utilizing ICD codes and clinical history for diagnosis. This demonstrated an OR = 1.43 (95% CI: 0.93-2.20; P =
.10; I2 = 48%) for subjective diagnoses, an OR = 3.82 (95% CI: 2.61-5.59; P < .00001; I2 ACCEPTED MANUSCRIPT
= 32%) for objective methods, and an OR = 1.81 (95% CI: 1.19-2.78; P = .006; I2 = 98%) for review-based methods. There was a statistically significant difference between diagnostic methods (Chi 2 = 12.60; P = .002; I2 = 84.1%).
Further subgroup analysis was performed to evaluate the effects of study design type (prospective or retrospective) on the risk of laryngeal malignancy (Figure S1).
Prospective studies exhibited an OR = 2.46 (95% CI: 1.57-3.85; P < .0001; I2 = 68%) and retrospective studies exhibited an OR = 2.68 (95% CI: 1.86-3.85; P < .00001; I2 =
97%). There was no statistically significant difference between design types ( P = .77; I2
= 0%). The effect of methodologic quality on the risk association was also investigated
(Figure S2). For this analysis, any study with a MINORS score below the lower 95% CI was considered “low” quality, studies with score between the 95% CIs were considered
“moderate” in quality, and any study with a MINORS score above the upper 95% confidence interval was considered “high” quality. MANUSCRIPT Low quality studies demonstrated an OR = 1.93 (95% CI: 1.32-2.83; P < .00001; I2 = 97%), moderate quality studies demonstrated an OR = 3.09 (95% CI: 2.30-4.15; P < .00001; I2 = 36%), and high quality studies demonstrated an OR = 2.05 (95% CI: 0.80-5.25; P = .13; I2 = 62%). There was no statistically significant difference between the three groups (Chi 2 = 3.87; P = .14; I2 =
48.3%).
Sensitivity analysis was performed to determine the influence of each individual study on the overall pooled data (Figure 4). After excluding individual studies and assessing theACCEPTED pooled OR, the average OR was 2.47 with the average lower and upper
95% CI being 1.89 and 3.22, respectively. Regardless of study excluded, the pooled OR fell within the average 95% CI, suggesting that no study had major effects on the risk ACCEPTED MANUSCRIPT
association. Funnel plot analysis demonstrated slight asymmetry with lack of lower- powered negative effect studies (Figure S3). Sensitivity analysis performed on 7 studies that controlled for smoking status revealed pooled OR of 2.07 (95% CI: 1.26-3.41). The
P-value for overall effect was 0.004 and I2 = 96% (Figure 5). When comparing these studies to the remainder of the studies that did not control for covariates, there was no statistically significant difference between groups ( P = .24; I2 = 28.7%). The 11 studies without covariate control demonstrated an OR = 3.03 (95% CI: 2.07-4.43; P < .00001; I2
= 65%; Figure S4).
Discussion
In our analysis, patients with a diagnosis of reflux disease were 2.47 times as likely to have a diagnosis of laryngeal carcinoma. The majority of the reviewed studies examined the association between GERD and laryngealMANUSCRIPT malignancy, with patients being 2.37 times as likely to have a diagnosis of laryngeal cancer. Fewer studies examined the relationship between LPR and laryngeal carcinoma, though meta-analysis demonstrated these patients were 3.09 times as likely to have a prior diagnosis of malignancy; however, this difference was not statistically significant. Funnel plot analysis (Figure S3) showed publication bias with LPR and therefore the true difference may not yet be measurable.
By definition, LPR involves reflux of gastric contents into the laryngopharynx and subsequently onto the laryngeal mucosa, which is not necessarily present in GERD. 37
Prior study hasACCEPTED demonstrated that laryngeal mucosa is more sensitive to the effects of reflux compared to the esophagus. 38, 39 Given this, the direct insult to the larynx may increase the likelihood for dysplasia and development of malignancy to a greater extent ACCEPTED MANUSCRIPT
than with GERD. Despite this theory, further study with strict delineation between GERD and LPR using rigid diagnostic criteria is necessary to determine any significance.
While the data demonstrates increased association of laryngeal cancer with reflux disease, the majority of studies did not control for smoking and drinking status. This is problematic for analysis as these are known to be significant risk factors in development of laryngeal carcinoma and there is a paucity of data in the literature investigating the effects of reflux in NSND patients. To assess the association between reflux disease and laryngeal cancer in this population, we performed sensitivity analysis using 7 studies that controlled for smoking and found there to be a significant association between reflux disease and laryngeal carcinoma (Figure 5). Only 1 of these studies enrolled nonsmoking patients to assess the risk of GERD; however, that demonstrated a statistically significant increase in odds of laryngeal cancer compared to controls (OR = 7.63 [2.55, 22.86]). 40 The 6 other studies included in the sensitivity ana MANUSCRIPTlysis used multivariate logistic regression to control for smoking and drinking histories. Using generic inverse variance analysis on these 7 studies, we showed there was a positive risk association between reflux disease and laryngeal cancer (OR = 2.07, 95% CI: 1.26-3.41; P = .004). There was heterogeneity in the studies and, of the studies included, 3 were highly powered. These results may represent publication bias; however, it is noteworthy that one of these 3 studies showed no significant effect when controlling for smoking.33 In addition to these
7 studies, Langevin et al demonstrated that GERD was associated with laryngopharyngealACCEPTED malignancy regardless of smoking or drinking status through multivariate analysis, though this included tumors of the larynx and pharynx combined.41
Interestingly, there was no statistically significant difference ( P = .24) between the 7 ACCEPTED MANUSCRIPT
studies controlling for smoking and drinking histories and the remaining 11 studies that did not control for these covariates. Both subgroups demonstrated positive risk association between reflux disease and laryngeal cancer, regardless of smoking and drinking.
While our analysis supports an increased risk in NSND patients, controversy remains in the literature. In a small prospective case–control study, Geterud et al demonstrated no significant difference in laryngeal cancer rates in nonsmokers and healthy controls with GERD. 42 Similarly, Nilsson et al demonstrated no significant increase in risk of laryngeal cancer in patients with GERD after excluding those with alcoholism. 14 This same trend was demonstrated by Tae et al in which multivariate regression showed that only smoking and not reflux disease was significantly associated with laryngeal carcinoma when controlling for other variables.43 In general, there is a lack of literature investiga MANUSCRIPTting the etiology of laryngeal cancer in the NSND population. A recent Algerian laryngeal cancer study showed an incidence of 14.3% in NSND patients, though no causative factor was identified and only one patient was positive for HPV. 44 The remaining NSND patients had unidentifiable risk factors, which may be reflux disease. While smoking and drinking are strong risk factors, chronic reflux of acidic and nonacid gastrointestinal contents into the larynx may lead to inflammatory changes and molecular aberrations promoting the development of laryngeal carcinoma in NSND patients. 45, 46 This may be one explanation for the findings of our sensitivity analysisACCEPTED and has been demonstrated in vitro and in animal models.
Specifically, exposure of hypopharyngeal epithelial cells to acidic reflux has been shown to induce nuclear factor-kappa B, leading to deregulation of oncogenic mircoRNAs seen ACCEPTED MANUSCRIPT
in laryngeal carcinoma. 24, 47 While no definitive studies have been performed to date demonstrating these changes in human laryngeal mucosa, our meta-analysis supports this correlation when controlling for smoking and drinking history.
This study is not without limitations. While a prime strength of this study is the large number of cases and controls, the populations were largely heterogeneous and differences exist among the included studies regarding the method of diagnosis, whether by clinical history, physical examination findings, or diagnostic testing. Depending on the specific classification of reflux disease, a variety of testing methods are indicated, including proton pump inhibitor trial, ambulatory reflux monitoring, or upper gastrointestinal endoscopy. 48 Various quality-of-life and symptom instruments have been developed and used in practice as a surrogate for objective testing. While the RSI is a validated instrument, controversy exists over clinical practicality and effectiveness in diagnosis of LPR. 49, 50 In the present systematic MANUSCRIPTreview, there was a diversity of diagnostic methods used, including patient reported quality of life instruments (RSI and study-specific questionnaires). Two studies used the RSI for diagnosis of LPR and 3 studies utilized other study-specific questionnaires. In contrast, 5 studies utilized 24-hour pH probe monitoring, 2 studies used endoscopic findings on esophagoscopy, and 2 studies used history of gastric resection as a surrogate marker for diagnosis of GERD and
LPR. The remaining studies 4 studies used ICD codes and there was no information reported on how the diagnosis of GERD or LPR was made. This variation in diagnostic methods portendsACCEPTED some bias and may affect the accurate risk reporting for the association of reflux with laryngeal cancer. To control for this, we performed subgroup analysis on diagnostic method and demonstrated a statistically significant difference ( P < .00001) in ACCEPTED MANUSCRIPT
the risk association reported between studies using subjective and objective measures.
Specifically there was no association with reflux and laryngeal cancer when using subjective methods. Only 2 of the 5 studies using subjective methods employed the RSI.
Two of the remaining studies used study-specific questionnaires for diagnosis and 1 study used patient-reported symptoms of heartburn. These results may reflect the fact that symptomatic reflux may be a poor predictor of GERD and/or LPR or due to the use of nonvalidated methods in 3 of the 5 studies. Both objective diagnostic methods and chart review-based methods showed statistically significant risk association with laryngeal cancer.
This systematic review only examined manuscripts written in English, though excluded studies in non-English language were predominantly review articles and case reports of unrelated topics. Furthermore, there is heterogeneity present in the included studies, which likely reflects the clinical and met MANUSCRIPThodologic diversity in the included studies. In order to better account for this, we employed a random effects model for analysis. One major contributor to clinical diversity was the variable diagnostic methods used in the studies, as discussed above. Furthermore, the patient demographics varied considerably and, importantly, the effects of smoking and alcohol on the presence of reflux and the development of laryngeal cancer were not excluded. Though the majority of the queried studies in this review demonstrate an association between reflux and malignancy, it is unclear if this is due to the effects of tobacco and alcohol on the developmentACCEPTED of GERD, which is unknown. Additionally, the effect of tobacco and alcohol on the development of laryngeal malignancy in these studies is unknown. While this is a weakness of this study, it is inherent to the current state of the literature. To date ACCEPTED MANUSCRIPT
there have been no significant studies assessing the effects of reflux on laryngeal carcinogenesis in NSND populations. In order to account for this, we performed a sensitivity analysis to control for the effects of smoking and drinking and demonstrated an increased risk of laryngeal cancer with exposure to GERD and/or LPR.
Methodologic diversity likely had an effect on heterogeneity. All studies included were case–control by design and only varied in direction of inquiry. Ten studies were prospective and 8 studies were retrospective by design. Subgroup analysis was performed and showed no statistically significant difference ( P = .77) between the two study designs. Despite this, there are key limitations to case–control studies that must be considered. 51 As discussed above, the diagnostic methods were variable throughout the studies. While some studies used matched controls, this was not universal, which could promote heterogeneity. Furthermore, there is a possibility of either recall or selection bias affecting the results due to lack of randomization MANUSCRIPT as well as interviewer bias. Studies reported the odds of reflux disease in patients with and without laryngeal cancer, increasing the risk for recall bias. Further, in the retrospective studies, there is a risk of reverse causation as patients who were diagnosed with GERD or LPR may have been presenting with symptoms caused by their malignancy. Thus, it is feasible that a percentage of patients were included based on symptoms caused by an undiscovered malignancy masking as risk exposure. Additionally, the diagnosis of GERD and/or LPR may have erroneously been made, which was then included in the medical record. There is also the riskACCEPTED for confounding variables in case–control design, which alters the true risk association of reflux disease and laryngeal cancer. To further assess methodologic diversity, subgroup analysis was performed using MINORS criteria to determine if ACCEPTED MANUSCRIPT
methodologic quality had an effect on the overall risk association. This revealed no statistical difference between low, moderate, and high quality studies ( P = .14).
Bias may also have been introduced in this manuscript due to the nature of the review. Publication bias may be present, although, as detailed in Figure S3, the funnel plot is only slightly asymmetric with studies assessing LPR contributing more to the asymmetry. Higher powered studies may have contributed to a positive pooled OR; however, sensitivity analysis was performed suggesting that no single study had a significant influence on the overall pooled OR (Figure 3). In order to reduce bias, controlled prospective studies with properly defined cohorts are necessary to examine this relationship with special attention toward NSND populations. In future studies, one might also consider examining the time from diagnosis of risk exposure to the diagnosis of malignancy. MANUSCRIPT Conclusions
This systematic review and meta-analysis suggests a positive risk association between GERD and LPR and the presence of laryngeal malignancy regardless of smoking and drinking histories. These findings are consistent with the understanding that chronic or recurrent inflammatory tissue insult may lead to the development of malignancy. Larger prospective cohort studies controlling for known risk factors would better define the time to diagnosis and control for other known variables, which may identify patientsACCEPTED at increased risk for developing laryngeal carcinoma.
ACCEPTED MANUSCRIPT
References
1. Thompson LDR. Head and neck cancers. In: Stewart B, Wild CP, editors. World
Cancer Report 2014. Geneva: International Agency for Research on Cancer, WHO; 2014. p. 422-31.
2. Chen AY, Schrag NM, Halpern M, et al. Health insurance and stage at diagnosis of laryngeal cancer: does insurance type predict stage at diagnosis? Arch Otolaryngol
Head Neck Surg. 2007;133(8):784-90.
3. Ko HC, Harari PM, Chen S, et al. Survival outcomes for patients with T3N0M0 squamous cell carcinoma of the glottic larynx. JAMA Otolaryngol Head Neck Surg.
2017;143(11):1126-33.
4. Haapaniemi A, Koivunen P, Saarilahti K, et al. Laryngeal cancer in Finland: A MANUSCRIPT 5‐year follow ‐up study of 366 patients. Head Neck. 2016;38(1):36-43.
5. Toporcov TN, Znaor A, Zhang Z-F, et al. Risk factors for head and neck cancer in young adults: a pooled analysis in the INHANCE consortium. Int J Epidemiol.
2015;44(1):169-85.
6. Koch WM, Lango M, Sewell D, et al. Head and neck cancer in nonsmokers: A distinct clinical and molecular entity. Laryngoscope. 1999;109(10):1544-51.
7. HarrisACCEPTED SL, Kimple RJ, Hayes DN, et al. Never ‐smokers, never ‐drinkers:
Unique clinical subgroup of young patients with head and neck squamous cell cancers.
Head Neck. 2010;32(4):499-503. ACCEPTED MANUSCRIPT
8. Castellsagué X, Alemany L, Quer M, et al. HPV involvement in head and neck cancers: comprehensive assessment of biomarkers in 3680 patients. J Natl Cancer Inst.
2016;108(6).
9. Ward PH, Hanson DG. Reflux as an etiological factor of carcinoma of the laryngopharynx. Laryngoscope. 1988;98(11):1195-9.
10. Morrison MD. Is chronic gastroesophageal reflux a causative factor in glottic carcinoma? Otolaryngl Head Neck Surg. 1988;99(4):370-3.
11. Freije JE, Beatty TW, Campbell BH, et al. Carcinoma of the larynx in patients with gastroesophageal reflux. Am J Otolaryngol. 1996;17(6):386-90.
12. Zhang D, Zhou J, Chen B, et al. Gastroesophageal reflux and carcinoma of larynx or pharynx: a meta-analysis. Acta Otolaryngol (StocMANUSCRIPTkh). 2014;134(10):982-9. 13. Kuo C-L, Chen Y-T, Shiao A-S, et al. Acid reflux and head and neck cancer risk: a nationwide registry over 13 years. Auris Nasus Larynx. 2015;42(5):401-5.
14. Nilsson M, Chow W-H, Lindblad M, et al. No association between gastroesophageal reflux and cancers of the larynx and pharynx. Cancer Epidemiol
Biomarkers Prev. 2005;14(5):1194.
15. Qadeer MA, Colabianchi N, Strome M, et al. Gastroesophageal reflux and laryngeal cancer:ACCEPTED causation or association? A critical review. Am J Otolaryngol.
2006;27(2):119-28. ACCEPTED MANUSCRIPT
16. Koufman JA. The otolaryngologic manifestations of gastroesophageal reflux disease (GERD): a clinical investigation of 225 patients using ambulatory 24 ‐hour pH monitoring and an experimental investigation of the role of acid and pepsin in the development of laryngeal injury. Laryngoscope. 1991;101(S53):1-78.
17. Adams J, Heintz P, Gross N, et al. Acid/pepsin promotion of carcinogenesis in the hamster cheek pouch. Arch Otolaryngol Head Neck Surg. 2000;126(3):405-9.
18. Johnston N, Dettmar PW, Lively MO, et al. Effect of pepsin on laryngeal stress protein (Sep70, Sep53, and Hsp70) response: role in laryngopharyngeal reflux disease.
Ann Otol Rhinol Laryngol. 2006;115(1):47-58.
19. Johnston N, Wells CW, Samuels TL, et al. Pepsin in nonacidic refluxate can damage hypopharyngeal epithelial cells. Ann Otol Rhinol Laryngol. 2009;118(9):677-85. MANUSCRIPT 20. Spechler SJ, Goyal RK. Barrett's esophagus. N Engl J Med. 1986;315(6):362-71.
21. Rogers EL, Goldkind SF, Iseri OA, et al. Adenocarcinoma of the lower esophagus. A disease primarily of white men with Barrett's esophagus. J Clin
Gastroenterol. 1986;8(6):613-8.
22. Haggitt RC. Barrett's esophagus, dysplasia, and adenocarcinoma. Hum Pathol.
1994;25(10):982-93.
23. FitzgeraldACCEPTED R, Onwuegbusi B, Bajaj-Elliott M, et al. Diversity in the oesophageal phenotypic response to gastro-oesophageal reflux: immunological determinants. Gut.
2002;50(4):451-9. ACCEPTED MANUSCRIPT
24. Sasaki CT, Issaeva N, Vageli DP. In vitro model for gastroduodenal reflux – induced nuclear factor ‐kappaB activation and its role in hypopharyngeal carcinogenesis.
Head Neck. 2016;38(S1):E1381-E91.
25. O'riordan J, Abdel-Latif M, Ravi N, et al. Proinflammatory cytokine and nuclear factor kappa-B expression along the inflammation–metaplasia–dysplasia– adenocarcinoma sequence in the esophagus. Am J Gastroenterol. 2005;100(6):1257.
26. Gibson MK, Dhaliwal AS, Clemons NJ, et al. Barrett's esophagus: cancer and molecular biology. Ann N Y Acad Sci. 2013;1300(1):296-314.
27. Luedde T, Schwabe RF. NF-κB in the liver—linking injury, fibrosis and hepatocellular carcinoma. Nat Rev Gastroenterol Hepatol. 2011;8(2):108.
28. Castaño ‐Milla C, Chaparro M, Gisbert MANUSCRIPT J. Systematic review with meta ‐ analysis: the declining risk of colorectal cancer in ulcerative colitis. Aliment Pharmacol
Ther. 2014;39(7):645-59.
29. Riley CA, Marino MJ, Hawkey N, et al. Sinonasal tract inflammation as a precursor to nasopharyngeal carcinoma: a systematic review and meta-analysis.
Otolaryngl Head Neck Surg. 2016;154(5):810-6.
30. Beachler DC, Engels EA. Chronic sinusitis and risk of head and neck cancer in the US elderlyACCEPTED population. JAMA Otolaryngol Head Neck Surg. 2017;143(1):25-31. ACCEPTED MANUSCRIPT
31. Wu EL, Riley CA, Hsieh M-C, et al. Chronic sinonasal tract inflammation as a precursor to nasopharyngeal carcinoma and sinonasal malignancy in the United States.
Int Forum Allergy Rhinol. 2017;7(8):786-93.
32. Slim K, Nini E, Forestier D, et al. Methodological index for non ‐randomized studies (MINORS): development and validation of a new instrument. ANZ J Surg.
2003;73(9):712-6.
33. Francis DO, Maynard C, Weymuller EA, et al. Reevaluation of gastroesophageal reflux disease as a risk factor for laryngeal cancer. Laryngoscope. 2011;121(1):102-5.
34. El-Serag HB, Hepworth EJ, Lee P, et al. Gastroesophageal reflux disease is a risk factor for laryngeal and pharyngeal cancer. Am J Gastroenterol. 2001;96(7):2013-8. 35. Riley CA, Wu EL, Hsieh M-C, et al. Association MANUSCRIPT of gastroesophageal reflux with malignancy of the upper aerodigestive tract in elderly patients. JAMA Otolaryngol Head
Neck Surg. 2018;144(2):140-8.
36. Pines G, Dickman R, Niv Y, et al. Extraesophageal malignancies among patients with Barrett esophagus. J Clin Gastroenterol. 2014;48(1):e8-e11.
37. Koufman JA, Aviv JE, Casiano RR, et al. Laryngopharyngeal reflux: position statement of the committee on speech, voice, and swallowing disorders of the American Academy of ACCEPTEDOtolaryngology-Head and Neck Surgery. Otolaryngl Head Neck Surg. 2002;127(1):32-5. ACCEPTED MANUSCRIPT
38. Axford SE, Sharp N, Dettmar PW, et al. Cell biology of laryngeal epithelial defenses in health and disease: preliminary studies. Ann Otol Rhinol Laryngol.
2001;110(12):1099-108.
39. Johnston N, Bulmer D, Ross PE, et al. Cell biology of laryngeal epithelial defenses in health and disease: further studies. Ann Otol Rhinol Laryngol.
2003;112(6):481-91.
40. Bacciu A, Mercante G, Ingegnoli A, et al. Effects of gastroesophageal reflux disease in laryngeal carcinoma. Clin Otolaryngol. 2004;29(5):545-8.
41. Langevin SM, Michaud DS, Marsit CJ, et al. Gastric reflux is an independent risk factor for laryngopharyngeal carcinoma. Cancer Epidemiol Biomarkers Prev.
2013;22(6):1061-8. MANUSCRIPT 42. Geterud Å, Bove M, Ruth M. Hypopharyngeal acid exposure: an independent risk factor for laryngeal cancer? Laryngoscope. 2003;113(12):2201-5.
43. Tae K, Jin BJ, Ji YB, et al. The role of laryngopharyngeal reflux as a risk factor in laryngeal cancer: a preliminary report. Clin Exp Otorhinolaryngol. 2011;4(2):101.
44. Kariche N, Hortal MT, Benyahia S, et al. Comparative assessment of HPV, alcohol and tobacco etiological fractions in Algerian patients with laryngeal squamous cell carcinoma.ACCEPTED Infect Agent Cancer. 2018;13(1):8. ACCEPTED MANUSCRIPT
45. Kelly EA, Samuels TL, Johnston N. Chronic pepsin exposure promotes anchorage-independent growth and migration of a hypopharyngeal squamous cell line.
Otolaryngl Head Neck Surg. 2014;150(4):618-24.
46. Johnston N, Knight J, Dettmar PW, et al. Pepsin and carbonic anhydrase isoenzyme III as diagnostic markers for laryngopharyngeal reflux disease. Laryngoscope.
2004;114(12):2129-34.
47. Sasaki CT, Vageli DP. miR-21, miR-155, miR-192, and miR-375 deregulations related to NF-kappaB activation in gastroduodenal fluid-induced early preneoplastic lesions of laryngeal mucosa in vivo. Neoplasia. 2016;18(6):329-38.
48. Katz PO, Gerson LB, Vela MF. Guidelines for the diagnosis and management of gastroesophageal reflux disease. Am J Gastroenterol. 2013;108:308–28. MANUSCRIPT 49. Belafsky PC, Postma GN, Koufman JA. Validity and reliability of the Reflux
Symptom Index (RSI). J Voice. 2002;16(2):274-7.
50. Kim SI, Kwon OE, Na SY, et al. Association between 24-hour combined multichannel intraluminal impedance- pH monitoring and symptoms or quality of life in patients with laryngopharyngeal reflux. Clin Otolaryngol. 2017;42(3):584-91.
51. Song JW, Chung KC. Observational studies: cohort and case–control studies. Plast ReconstrACCEPTED Surg. 2010;126(6):2234. ACCEPTED MANUSCRIPT
52. Busch EL, Zevallos JP, Olshan AF. Gastroesophageal reflux disease and odds of head and neck squamous cell carcinoma in North Carolina. Laryngoscope.
2016;126(5):1091-6.
53. Cammarota G, Galli J, Cianci R, et al. Association of laryngeal cancer with previous gastric resection. Ann Surg. 2004;240(5):817.
54. Çekin E, Ozyurt M, Erkul E, et al. The association between Helicobacter pylori and laryngopharyngeal reflux in laryngeal pathologies. Ear Nose Throat J. 2012;91(3).
55. Cianci R, Galli J, Agostino S, et al. Gastric surgery as a long-term risk factor for malignant lesions of the larynx. Arch Surg. 2003;138(7):751-4.
56. Velanovich V, Vallance SR, Gusz JR, et al. Quality of life scale for gastroesophageal reflux disease. J Am Coll Surg. MANUSCRIPT 1996;183(3):217-24. 57. Da ğli S, Da ğli U, Kurtaran H, et al. Laryngopharyngeal reflux in laryngeal cancer. Turk J Gastroenterol. 2004;15(2):77-81.
58. Johnson LF, Demeester TR. Twenty-four-hour pH monitoring of the distal esophagus. Am J Gastroenterol. 1974;62(4).
59. Doustmohammadian N, Naderpour M, Khoshbaten M, et al. Is there any association between esophagogastric endoscopic findings and laryngeal cancer? Am J
Otolaryngol. ACCEPTED2011;32(6):490-3.
60. Galli J, Cammarota G, Calo L, et al. The role of acid and alkaline reflux in laryngeal squamous cell carcinoma. Laryngoscope. 2002;112(10):1861-5. ACCEPTED MANUSCRIPT
61. Ozlugedik S, Yorulmaz I, Gokcan K. Is laryngopharyngeal reflux an important risk factor in the development of laryngeal carcinoma? Eur Arch Otorhinolaryngol.
2006;263(4):339-43.
62. Sereg-Bahar M, Jerin A, Hocevar-Boltezar I. Higher levels of total pepsin and bile acids in the saliva as a possible risk factor for early laryngeal cancer. Radiol Oncol.
2015;49(1):59-64.
63. Vaezi MF, Qadeer MA, Lopez R, et al. Laryngeal cancer and gastroesophageal reflux disease: a case–control study. Am J Med. 2006;119(9):768-76.
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
Tables
Table 1. Summary of study characteristics, risk assessment, and quality of evidence.
Study Study Risk Type Population description Exposure Assessment MINORS Design Score † Bacciu 40 2004 ‡ Retrospective GERD NSND Patients treated “The Savary–Miller classification 13 case–control for laryngeal carcinoma system was used for the grading of at the University of esophagitis.” Patients had history of Parma EGD and FFL evaluation with rating of esophagitis” Busch 52 2016 ‡ Prospective GERD Patients from the “Gastroesophageal reflux disease 11 case–control CHANCE study with exposure was assessed via two history of laryngeal different questionnaire items cancer administered in-person by a nurse- MANUSCRIPTinterviewer within 2 months of diagnosis [of laryngeal SCC]” Cammarota 53 2004 Retrospective GERD Patients with history of The use of gastric resection for peptic 11 case–control laryngeal cancer and ulcer as a surrogate for reflux disease gastric resection in Italy Çekin 54 2012 Prospective LPR Patients with history of “Patients with an RSI of 14 or more 12 case–control laryngeal pathology at and/or an RFS of 8 or more were Haydarpasa Training considered to have LPR” Hospital in Istanbul Cianci 55 2003 Retrospective GERD Patients with history of History of gastric resection was used 14 case–control gastric resection > 5 yr as a surrogate marker for reflux. prior to study in Rome, Furthermore, “each patient received a ACCEPTEDItaly validated questionnaire 56 to evaluate typical symptoms of reflux disease” Da ğli 57 2004 Prospective GERD/LPR Patients with history of “De Meester's criteria 58 was used for 13 ACCEPTED MANUSCRIPT
case–control laryngeal cancer that the distal esophagus. […] Scores over underwent pH probe 15 are accepted as pathologic. For the testing and proximal esophagus, pH falls less esophagoscopy than 4 lasting more than 1% of the (Turkey) total time was accepted as pathologic” Doustmohammadian 59 Prospective GERD Tertiary care center Endoscopic findings from EGD 13 2011 case–control patients with history of laryngeal cancer El-Serag 34 2001 ‡ Retrospective GERD Outpatients US “GERD-related diagnoses, such as 10 case–control Veterans Affairs erosive esophagitis (ICD- code patients with history of 530.1), peptic esophagitis (ICD-code laryngeal or pharyngeal 530.19), esophageal stricture (ICD- cancer between 1991 to code 530.3), and gastroesophageal 1997 and 1999 reflux (ICD-code 530.81)” Francis 33 2011 ‡ Retrospective GERD Outpatient US Veterans “GERD was defined by ICD-9 codes 10 case–control Affairs patients MANUSCRIPT with (530.1–530.3, 530.81, 530.85)” history of laryngeal or pharyngeal cancer between 2000 to 20006 Galli 60 2002 Prospective GERD Patients in clinic with 24-hour pH probe monitoring was 12 case–control history of laryngeal used. “Patients were considered cancer examined with negative for pathological proximal pH probe monitoring reflux when showing < 10 episodes of reflux, with percent total time to acid exposure (pH < 4), and complete absence of reflux in supine position.” Koufman 1991 16 Prospective GERD 225 consecutive “pH probe values < 4.0 were 15 case–control ACCEPTEDpatients presenting to considered evidence of GER” outpatient otolaryngology clinics ACCEPTED MANUSCRIPT
Langevin 41 2013 Prospective GERD Patients in Boston with “Subjects completed a self- 15 case–control history of HNSCC administered epidemiologic questionnaire that provided detailed data on sociodemographics and personal characteristics, alcohol and tobacco use, personal and family cancer history, history of heartburn and other relevant dietary, occupational, residential, and medical exposures” Ozlugedik 61 2006 Prospective GERD/LPR Patients with history of pH probe values greater than: “5.3% 13 case–control treatment for laryngeal for the percent of total time the pH cancer (Turkey) was <4, 8.1% for the percent of upright time the pH was <4, 3.2% for the percent of supine time the pH was < 4 and 14.7 for the DeMeester MANUSCRIPTscore” Pines 36 2014 ‡ Retrospective GERD Patients within the “CHS database was queried for 12 case–control Clalit Health Service patients who were newly diagnosed (Israel) with history of with BE […] by at least 2 primary BE caregivers—a family physician and/or a gastroenterologist; or during hospitalization” Riley 35 2017 Retrospective GERD General population “ICD-9-CM diagnosis code 530.81 13 case–control aged 66 to 99 years was examined as a comorbidity with within the SEER carcinomas of the UADT, using database with history of inpatient and outpatient claims in the laryngeal or pharyngeal MEDPAR, OUTPT, and NCH data ACCEPTEDcancer files” Sereg-Bahar 62 2015 Prospective LPR Patients with T1 RSI score > 13 14 case–control laryngeal cancer in ACCEPTED MANUSCRIPT
tertiary care center (Slovenia) Tae 43 2011 ‡ Prospective LPR Patients with history of “LPR event was defined by an abrupt 13 case–control laryngeal cancer decrease in pH to below 4 in the examined with pH upper probe, with an accompanying probe monitoring or preceding decrease in pH to below 4 in the lower probe, except while eating food. Pathologic LPR is defined when more than three episodes of LPR occur” Vaezi 63 2006 ‡ Retrospective GERD Patients with history of “Diagnosis of GERD was established 13 case–control laryngeal cancer at by ICD-9 codes and presence of tertiary care center in symptoms (heartburn or acid Cleveland, Ohio regurgitation), esophagitis, Barrett’s epithelium, abnormal pH monitoring, hiatal hernia, reflux on barium MANUSCRIPTswallow, or presence of long-term acid-suppressive medications” Average 12.61 † The ideal global score is 24 for comparative studies ‡ Study controlled for smoking and drinking status in analysis BE: Barrett esophagus CHANCE: Carolina Head and Neck Cancer Epidemiology EGD: esophagogastroduodenoscopy GERD: gastroesophageal reflux disease HNSCC: head and neck squamous cell carcinoma LPR: laryngopharyngeal reflux NSND: nonsmoking nondrinking RFS: Reflux Finding Score ACCEPTED RSI: Reflux Symptom Index UADT: upper aerodigestive tract ACCEPTED MANUSCRIPT
Figure Legends
Figure 1. Flow diagram for literature selection process after application of inclusion and exclusion criteria.
Figure 2. Forest plot depicting the overall risk association of reflux disease and laryngeal cancer with subgroup analysis based on type of reflux (GERD or LPR). There is no statistically significant difference between the two subgroups ( P = .44)
Figure 3. Forest plot depicting the overall risk association of reflux disease and laryngeal cancer with subgroup analysis based on the type of diagnostic method used. Subjective, objective, and chart review-based methods are depicted. There is a statistically significant difference between the three groups (Chi 2 = 12.60; P = .002). Risk association with subjective methods showed no statistical significan MANUSCRIPTce ( P = .10).
Figure 4. Sensitivity analysis assessing the effect the omitted study has on the pooled OR for the combined data. The named study has been omitted.
Figure 5. Forest plot depicting the risk association of reflux disease and laryngeal malignancy after controlling for smoking and drinking covariates. There was a positive risk association after controlling for smoking and drinking (OR = 2.07 [1.26, 3.41]; P =
.004) and thereACCEPTED was no statistically significant difference between studies not controlling for covariates ( P = .24).
ACCEPTED MANUSCRIPT
Figures
Figure 1
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
Figure 2
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
Figure 3
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
Figure 4
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
Figure 5
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
Supplemental Tables
Table S1. Electronic search criteria for systematic review of literature. Database Search criteria EMBASE ((reflux) OR ('gastroesophageal reflux'/exp OR 'gastroesophageal reflux') OR ('laryngopharyngeal reflux'/exp OR 'laryngopharyngeal reflux') OR ('heartburn'/exp OR 'heartburn')) AND ((laryngeal AND ('carcinoma'/exp OR carcinoma)) OR (laryngeal AND ('cancer'/exp OR cancer)) OR ('larynx carcinoma'/exp OR 'larynx carcinoma') OR ('larynx cancer'/exp OR 'larynx cancer') OR (laryngeal AND ('malignancy'/exp OR malignancy)) OR (('larynx'/exp OR larynx) AND ('malignancy'/exp OR malignancy))) MEDLINE (reflux[All Fields] OR ("gastroesophageal reflux"[MeSH Terms] OR ("gastroesophageal"[All Fields] AND "reflux"[All Fields]) OR "gastroesophageal reflux"[All Fields]) OR ("laryngopharyngeal reflux"[MeSH Terms] OR ("laryngopharyngeal"[All Fields] AND "reflux"[All Fields]) OR "laryngopharyngeal reflux"[All Fields]) OR ("heartburn"[MeSH Terms] OR "heartburn"[All Fields])) AND (("laryngeal neoplasms"[MeSH Terms] OR ("laryngeal"[All Fields] AND "neoplasms"[All Fields]) OR "laryngeal neoplasms"[All Fields] OR ("larynx"[All Fields] AND "cancer"[All Fields]) OR "larynx cancer"[All Fields] OR ("laryngeal"[All Fields]MANUSCRIPT AND "cancer"[All Fields]) OR "laryngeal cancer"[All Fields]) OR (("larynx"[MeSH Terms] OR "larynx"[All Fields] OR "laryngeal"[All Fields]) AND ("carcinoma"[MeSH Terms] OR "carcinoma"[All Fields] OR "neoplasms"[MeSH Terms] OR "neoplasms"[All Fields] OR "malignancy"[All Fields]))) Web of Science (((Reflux) OR (Gastroesophageal Reflux) OR (Laryngopharyngeal Reflux) OR (Heartburn)) AND ((Laryngeal Carcinoma) OR (Laryngeal malignancy) OR (Laryngeal cancer) OR (Larynx carcinoma) OR (Larynx malignancy) OR (Larynx cancer)))
ACCEPTED ACCEPTED MANUSCRIPT
Table S2. Population, Intervention, Control, Outcome, Study Design (PICOS) inclusion criteria.
Description
Population Adult (age > 18 years) men and women
Intervention Adults with a history of GERD and/or LPR
Control Adults without a history of GERD and/or LPR
Outcomes: Risk Patients with previous diagnosis of GERD and/or LPR are assessed against controls to obtain the odds ratio with 95% CI for association with laryngeal carcinoma
Study Design: Case-control, cohort, randomized controlled trials Risk Assessment
CI: confidence interval GERD: gastroesophageal reflux disease MANUSCRIPT LPR: laryngopharyngeal reflux
ACCEPTED ACCEPTED MANUSCRIPT
Supplemental Figure Legends
Figure S1. Forest plot depicting the overall risk association of reflux disease and laryngeal cancer with subgroup analysis based the study method (retrospective versus prospective). There was no statistically significant difference between groups ( P = .77).
Figure S2. Forest plot depicting the overall risk association of reflux disease and laryngeal cancer with subgroup analysis based on the methodologic quality. Quality was separated into low, moderate, and high quality based on the 95% CI. There was no statistically significant difference between the three groups (Chi 2 = 3.87; P = .14; I2 = 48.3%).
Figure S3. Funnel plot indicating slight asymmetry suggesting possible publication bias.
Figure S4. Forest plot depicting the overall risk association MANUSCRIPT of reflux disease and laryngeal cancer with subgroup analysis based on studies that controlled for smoking and drinking and those that did not. There was no statistically significant difference between groups ( P = .24).
ACCEPTED ACCEPTED MANUSCRIPT
Supplemental Figures
Figure S1.
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
Figure S2.
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
Figure S3.
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
Figure S4.
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
MANUSCRIPT
ACCEPTED ACCEPTED MANUSCRIPT
MANUSCRIPT
ACCEPTED