Risk of histologic Barrett’s esophagus between African Americans and non-Hispanic whites: A meta-analysis

Abstract

Background

Barrett’s esophagus (BE) is rare in African Americans (AA). However, the risk difference magnitude in histologic BE prevalence between AA and non-Hispanic whites (nHw) has not been quantified to date.

Objective

The objective of this article is to determine the degree of histologic BE risk difference between AA and nHw.

Methods

PubMed, Web of Science and EMBASE were searched for studies reporting histologic BE in AA/nHw for inclusion. Pooled odds ratios (ORs) with risk estimates of histologic BE occurrence between AA/nHw were calculated along with 95% confidence intervals (CIs). Forest plots were used to quantify heterogeneity. Funnel plots and the Cochrane Collaboration Risk of Bias tool were used to assess bias risk.

Results

Eight studies reported BE histologic confirmation in AA/nHw. Analysis demonstrated a nearly 400% increased histologic BE risk in nHw patients compared to AA (OR 3.949, 95% CI 3.069–5.082). In the model without the case-control study, histologic BE risk remained elevated at approximately 360% in nHw compared to AA (OR 3.618, 95% CI 2.769–4.726). Heterogeneity was not present in either model. Risk of bias was significant.

Conclusions

Histologic BE risk is elevated in nHw by 3.6–4 times compared to AA. Investigation into understanding any clinical, molecular or genetic mechanisms underlying this risk disparity is warranted.

Keywords

Barrett’s esophagus ethnicity African Americans risk meta-analysis

Key summary

Summarize the established knowledge on this subject:

Barrett’s esophagus (BE) is associated with progression to esophageal adenocarcinoma (EAC).

Risk factors for BE include duration of gastroesophageal reflux disease (GERD), male sex, obesity, and cigarette smoking.

Helicobacter pylori colonization, regular aspirin, proton pump inhibitor or nonsteroidal anti-inflammatory use and African American (AA) ethnicity are associated with decreased risk.

Limited data exist regarding degree of risk difference by ethnicity of histologically confirmed BE.

What are the significant and/or new findings of this study?

Non-Hispanic whites (nHw) have a 3.6–4 fold increased risk of histologically confirmed BE compared to AA.

These findings quantify the observation of increased BE in nHw and can be used to further enhance cost-effectiveness studies in BE screening.

Continued exploration to determine factors explaining the disparity in BE formation and/or presence seen between United States ethnic groups is required.

Introduction

Barrett’s esophagus (BE) is a metaplastic alteration of the distal esophageal epithelium, following injury caused by reflux of acid and bile, in which one kind of fully differentiated cell replaces another.¹ This change, called specialized intestinal epithelium (SIM), is present in up to 1.6% of the general population and 5%–13% of patients having endoscopy for reflux symptoms.^2,3 BE is associated with a 0.12%–0.5% yearly risk of progression to esophageal adenocarcinoma (EAC), which is an aggressive cancer whose incidence has increased eight-fold in the United States (US) over the past several decades.^4–9 Risk factors for BE include non-Hispanic white race (nHw), male sex, obesity, and gastroesophageal reflux disease (GERD) while Helicobacter pylori colonization and regular aspirin or nonsteroidal anti-inflammatory drugs reduce risk.^2,10–21

Although investigators have recently increased their focus on BE in US multiethnic populations, the studies published to date have not featured large samples. In addition, the difference in degree of histologic BE prevalence risk between African Americans (AA) and nHw has not been determined to date. To quantify and confirm the magnitude of difference in histologic BE prevalence between AA and nHw, we conducted a meta-analysis combining the results of published studies reporting the prevalence of histologically proven BE in AA and nHw, thus providing a quantitative estimate of the risk reduction for histologically proven BE associated with AA race/ethnicity.

Methods

Search strategy

A systematic literature search was conducted of PubMed, Web of Science, and EMBASE databases from 1997 to 2014 to identify all relevant articles on histologically proven BE in AA and nHw. Key words used in the search included a combination of “Barrett esophagus,” “ethnicity,” “race,” “African American,” and “non-Hispanic white.” The search was restricted to studies in human participants published in the English language in peer-reviewed journals. Two authors (AA and KJV) independently reviewed the title and abstract of identified studies in the primary search to exclude those that did not address the research question of interest, based on preselected inclusion and exclusion criteria (details below). Full text of remaining articles were then examined to determine if it contained relevant information. Any discrepancy of article selection was resolved by consensus. Finally, bibliographies of selected articles as well as systematic and narrative review articles on the topic were manually searched for additional articles. Each included study was weighted according to its inverse variance.

Study inclusion criteria

All eligible studies had to satisfy the following inclusion criteria: (a) histologic confirmation of BE (presence of intestinal metaplasia either with or without goblet cells), (b) reported AA and nHw ethnicity for patients/participants, (c) adequate data available to provide an estimate of effect measure (odds or risk ratio), (d) was original, and (e) used the individual as the unit of analysis. We excluded studies that: (a) did not have histologic confirmation of BE and (b) consisted of highly selected cohorts (referral to a single individual).

Data extraction/statistical analysis

Data were extracted from the relevant studies specifically looking at histologic diagnosis of BE among nHw and AA. Each study was weighted according to its inverse variance. The distribution of effect measures was examined using visual and tabular displays as well as tests of homogeneity to reveal variation in the risk estimates of histologic BE occurrence between AA and nHw using a DerSimonian-Laird random-effects method. Odds ratio (OR) was calculated along with 95% confidence interval (CI) estimates. Forest plots were created and summary OR with 95% CI of histologic BE was reported. Heterogeneity was quantified using the I² statistic. A sensitivity analysis was performed comparing results with and without case-control studies. Software used to conduct the meta-analysis was the open source OpenMetaAnalyst platform. We used Excel to create a funnel plot according to number of patients and risk estimate for each study. Funnel plots are reported according to sample size and relative risk estimate. Funnel plots according to standard error and relative risk estimate were also created, yielding similar results. The Cochrane Collaboration Risk of Bias Tool was used to assess the quality of the studies.²²

Results

Search results

Our search strategy in PubMed, Web of Science, and EMBASE identified a total of 129 unique articles. On the basis of pre-specified inclusion and exclusion criteria, a total of eight studies were selected for review (Figure 1 and Table 1). These studies had a total of 76,574 patients for analysis.

Figure 1.

Study flow and selection.

Table 1.

Studies included in the analysis.

First author year^ref	Study design	Study period	Number of participants overall/nHw/AA	Histologic BE overall (%)	Histologic BE for nHw (%)	Histologic BE for AA (%)
Gerson 2001²³	Single-center, prospective cohort	NA	442/415/27	19.9	20.5	11.1
Spechler 2002²⁴	Single-center, retrospective cohort	1992–1994	2423/ 2174/249	8.2	8.9	2.4
Abrams 2008²⁵	Single-center, retrospective cohort	2005–2006	1040/792/248	5	6	1.6
Gilani 2008²⁶	Single-center, retrospective cohort	1998–2006	151/147/4	27.1	27.2	25
Wang 2008²⁷	Multicenter US, retrospective cohort	2000–2005	59895/55363/4532	1.9	2	0.4
Fan 2009²⁸	Single-center, retrospective cohort	2005–2007	3836//2908/928	1.9	2.2	1.1
Khoury 2012¹²	Single-center, retrospective cohort	2002–2007	6835/4138/2697	1.7	2.3	0.6
Nguyen 2014²⁹	Single-center, cross-sectional, case-control	2008–2013	1952/1248/704	15.4	21.3	5

BE: Barrett’s esophagus; AA: African American; nHw: non-Hispanic white; US: United States.

Overview of the included studies

The number of individuals evaluated in each investigation ranged from 151 to 59,895 (Table 1).^12,23–29 AA comprised 12.3% (N = 9389) of the overall patient population. Seven of eight studies reported the study period.^12,24–29 The oldest was conducted in 1992–1994 and published in 2002.²⁴ One study was from the national endoscopic database, comprising a combination of academic, community and Veterans Affairs (VA)/military centers.²⁷ Of the remaining seven reports, four were from academic medical centers in Massachusetts, New York, Florida and Texas, two from VA centers in Texas and Arizona with the final study a combination of an academic medical center and VA in California.^{12,23–26,28,29} The largest study overall was from the national endoscopy database and the largest single-center report had 6835 participants.^12,27

Prevalence of histologic BE

Overall prevalence of histologic BE was 2.6% in the eight studies analyzed. The lowest overall study prevalence was 1.7% (115 of 6835¹²) while the highest was 27.1% (41 of 151²⁶). For nHw, the overall study prevalence histologically confirmed BE was 2.9% (1924 of 67,185) while this was 1% for AA (95 of 9389). The lowest study prevalence for nHw was 2% (1132 of 55,363²⁷) and the highest 27.2% (41 of 147²⁶). For AA, the lowest study prevalence was 0.4% (20 of 4532²⁷) and highest was 25% (one of four²⁶).

Quantitative/qualitative data analysis

Analysis including the Nguyen study demonstrated a nearly 400% increased risk for nHw patients having histologic BE compared to AA (OR 3.995, 95% CI 3.031–5.266, Figure 2(a)). In the random-effects model without the case-control study, the risk of histologic BE remained elevated at approximately 360% in nHw compared to AA (OR 3.618, 95% CI 2.769–4.726, Figure 2(b)). Heterogeneity was not present in either model (case control included I²= 17%, p = 0.296, Figure 2(a); without case control I²= 0%, p = 0.42, Figure 2(b)). Funnel plots (Figure 3(a) and (b)) were constructed and asymmetry around the reported point estimates was noted. This suggests significant publication bias. In assessing the risk of bias using the Cochrane Collaboration Risk of Bias Tool, the currently available studies lacked a randomized, blinded design.

Figure 2.

Pooled odds ratios for histologic Barrett’s esophagus between non-Hispanic whites and African Americans. (a) all studies; (b) excluding Nguyen study.²⁹

Figure 3.

Funnel plots corresponding to Figures 2(a) and 2(b), respectively.

BE length, sex distribution, and age

Assessment of differences in length as well as age and sex distribution of AA and nHw BE patients was attempted. Unfortunately, these items were inconsistently reported in the included studies, preventing any opportunity to perform meta-analysis. Length of BE both in AA and nHw was reported in three included studies.^12,25,29 However, in two investigations, BE length was missing in 13 AA and three nHw patients, respectively.^25,29 Despite these missing data from the reports above, long-segment BE (LSBE) was present more frequently in nHw than AA (nHw 35.1% vs AA 14.5%, p < 0.002) while AA had short-segment BE (SSBE) more frequently that nHw (AA 85.5% vs nHw 64.9%, p < 0.002). Three studies reported sex distribution of BE cases.^12,26,29 However, two of these investigations were from VA centers limiting assessment by sex.^26,29 The single community-based report revealed more males then females with BE both in AA and nHw.¹² Interestingly, a trend was seen between AA and nHw regarding the degree of male predominance in histologic BE, with nHw have a greater male percentage than AA (p < 0.08). Finally, mean age of AA and nHw BE patients was reported by only two groups.^12,29 No difference in age between AA and nHw BE patients was observed in either investigation.

Discussion

It is well recognized that nHw ethnicity is a risk factor for histologically established BE compared to AA.^12,23–29 However, the degree of risk difference between AA and nHw has not been determined to date. The aim of the present investigation was to provide a quantitative estimate of the risk variation between these two racial/ethnic groups for histologically proven BE. The results indicate that histologically confirmed BE among more than 67,000 nHw and 9300 AA patients occurred 3.6 to 4 times more frequently in nHw than AA. This now quantifies the risk difference for histologically confirmed BE between these two ethnic groups as well as enhances future cost-effective analysis in screening for BE.

Reasons for such a risk variance between nHw and AA observed in the meta-analysis remain unclear at present. Clinical risk factors assessment between AA and nHw with BE was reported in two of the eight studies included in the analysis.^12,29 In one, from a multiethnic histologic BE cohort, only age, sex, and hiatal hernia were assessed between ethnic groups.¹² No difference was observed in any of those parameters between AA and nHw. Nguyen and colleagues evaluated clinical factors for BE at a Texas VA facility and reported that nHw with histologic BE had a higher waist-to-hip ratio than AA BE cases.²⁹ However, no difference was seen between AA and nHw histologic BE groups in age, sex, GERD symptoms, hiatus hernia, proton pump inhibitor (PPI)/nonsteroidal anti-inflammatory drug (NSAID)/aspirin/statin use as well as cigarette and alcohol use. Infection with H. pylori was assessed by only one group at a VA institution, with AA having higher rates of infection compared to nHw.²⁹ It is well known that H. pylori infection is associated with a decreased risk of BE.³⁰ The prevalence of H. pylori is higher among AA compared to nHw and this factor could affect differences observed in BE between ethnic groups.³¹

However, there are no data evaluating H. pylori prevalence in the overall endoscopic population simultaneously in any histologic BE study included in the meta-analysis.

Evaluations have recently occurred attempting to determine BE predisposition within the genome. At present, eight loci have been identified by genome-wide association studies associated with BE/EAC risk.³² Unfortunately, assessment by ethnicity has not occurred to date. If ethnic variation of the BE/EAC risk loci is found to exist, then improved BE-preventive and -therapeutic efforts can be targeted to those at greatest risk for development of EAC.

Our meta-analysis had a number of strengths. First, a systematic literature search was performed using well-defined inclusion criteria to eliminate studies that were redundant or had significant referral bias. Second, studies were included in the analysis only if all patients with BE had histologic confirmation, excluding any opportunity for misclassification bias. Third, minimal heterogeneity was observed in either analysis model. This was most evident in the simulation not using the case-control study but even with inclusion of that investigation, heterogeneity remained acceptable (Figure 2).

However, the present study had limitations. First, this meta-analysis, like many performed, is subject to the individual biases of each study included. Second, there was significant variation in the prevalence of histologically confirmed BE among the included studies both in AA and nHw. This occurred among both groups and ranged from 2% to 27.2% in nHw and 0.4% to 25% in AA. Of note, highest prevalence rates were seen in AA and nHw among studies with smaller cohorts and from a VA hospital.^23,26,29 Third, clinical (GERD symptom frequency, hiatal hernia and H. pylori infection rates, obesity) and social factors (smoking frequency or alcohol/medication use) known to affect BE prevalence were inconsistently reported, limiting any ability to address the effect from these items individually.^33–35 Fourth, items such as BE length, age, and sex distributions were inconsistently reported, limiting any ability to assess via meta-analysis. Finally, the funnel plots (Figure 3(a) and (b)) support the role of publication bias. In assessing the risk of bias using the Cochrane Collaboration’s tool, the studies available tend to be nonrandomized and non-blinded. This makes them subject to reporting bias. Thus, potential under-reporting of negative studies may be possible.

In summary, we have identified a large disparity in histologic BE prevalence between AA and nHw and quantified the risk difference (3.6–4 times higher in nHw than AA) between these ethnic groups. The results should be tempered by the limited number of studies available looking at the dimension of BE according to nHw and AA race. Future investigations should confirm these findings. In addition, continued assessment of clinical factors along with genetic mechanisms underlying this risk disparity are required to improve BE-preventive and -therapeutic efforts for maximization of benefit, reducing EAC incidence.

Footnotes

Acknowledgments

Author contributions are as follows: Ahmad Alkaddour: planning, collecting, and interpreting data and drafting and editing manuscript; Carlos Palacio: study design, interpreting data, statistical analysis, and editing manuscript for intellectual content; and Kenneth J. Vega: planning, collecting, and interpreting data and editing manuscript for intellectual content. All authors approved the final version of the manuscript.

Declaration of conflicting interests

None declared.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Ethics approval

This research did not require institutional review board approval as it was a meta-analysis of existing data in the literature.

Informed consent

Not required for this article.

References

Spechler

Sharma

Souza

et al.

American Gastroenterological Association technical review on the management of Barrett’s esophagus. Gastroenterology 2011; 140: e18–e52.

Ronkainen

Aro

Storskrubb

et al.

Prevalence of Barrett’s esophagus in the general population: An endoscopic study. Gastroenterology 2005; 129: 1825–1831.

Winberg

Lindblad

Lagergren

et al.

Risk factors and chemoprevention in Barrett’s esophagus—an update. Scand J Gastroenterol 2012; 47: 397–406.

Shaheen

Crosby

Bozymski

et al.

Is there publication bias in the reporting of cancer risk in Barrett’s esophagus?

Gastroenterology 2000; 119: 333–338.

Bhat

Coleman

Yousef

et al.

Risk of malignant progression in Barrett’s esophagus patients: Results from a large population-based study. J Natl Cancer Inst 2011; 103: 1–9.

Hvid-Jensen

Pedersen

Drewes

et al.

Incidence of adenocarcinoma among patients with Barrett’s esophagus. N Engl J Med 2011; 365: 1375–1383.

Wani

Falk

Hall

et al.

Patients with nondysplastic Barrett’s esophagus have low risks for developing dysplasia or esophageal adenocarcinoma. Clin Gastroenterol Hepatol 2011; 9: 220–227.

Desai

Krishnan

Samala

et al.

The incidence of oesophageal adenocarcinoma in non-dysplastic Barrett’s oesophagus: A meta-analysis. Gut 2012; 61: 970–976.

Thrift

Whiteman

. The incidence of esophageal adenocarcinoma continues to rise: Analysis of period and birth cohort effects on recent trends. Ann Oncol 2012; 23: 3155–3162.

10.

Cook

Wild

Forman

. A systematic review and meta-analysis of the sex ratio for Barrett’s esophagus, erosive reflux disease and nonerosive reflux disease. Am J Epidemiol 2005; 162: 1050–1061.

11.

Edelstein

Bronner

Rosen

et al.

Risk factors for Barrett's esophagus among patients with gastroesophageal reflux disease: A community clinic-based case-control study. Am J Gastroenterol 2009; 104: 834–842.

12.

Khoury

Chisholm

Jamal

et al.

African Americans with Barrett’s esophagus are less likely to have dysplasia at biopsy. Dig Dis Sci 2012; 57: 419–423.

13.

Lagergren

Bergström

Lindgren

et al.

Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med 1999; 340: 825–831.

14.

Hoyo

Cook

Kamangar

et al.

Body mass index in relation to oesophageal and oesophageal junction adenocarcinomas: A pooled analysis from the international BEACON consortium. Int J Epidemiol 2012; 41: 1706–1718.

15.

Kubo

Cook

Shaheen

et al.

Sex-specific associations between body mass index, waist circumference and risk of Barrett’s oesophagus: A pooled analysis from the international BEACON consortium. Gut 2013; 62: 1684–1691.

16.

Cook

Kamangar

Whiteman

et al.

Cigarette smoking and adenocarcinomas of the esophagus and esophagogastric junction: A pooled analysis from the international BEACON consortium. J Natl Cancer Inst 2010; 102: 1344–1353.

17.

Cook

Shaheen

Anderson

et al.

Cigarette smoking increases risk of Barrett’s esophagus: An analysis of the Barrett’s and Esophageal Adenocarcinoma Consortium. Gastroenterology 2012; 142: 744–753.

18.

Liao

Vaughan

Corley

et al.

Nonsteriodal anti-inflammatory drug use reduces risk of adenocarcinomas of the esophagus and esophagogastric junction in a pooled analysis. Gastroenterology 2012; 142: 442–452.

19.

Anderson

Johnston

Watson

et al.

Nonsteroidal anti-inflammatory drugs and the esophageal inflammation-metaplasia-adenocarcinoma sequence. Cancer Res 2006; 66: 4975–4982.

20.

Whiteman

Parmar

Fahey

et al.

Association of Helicobacter pylori infection with reduced risk for esophageal cancer is independent of environmental and genetic modifiers. Gastroenterology 2010; 139: 73–83.

21.

Thrift

Pandeya

Smith

et al.

Helicobacter pylori infection and the risks of Barrett’s oesophagus: A population-based case-control study. Int J Cancer 2012, pp. 2407–2416.

22.

Sterne JAC, Higgins JPT, Reeves BC on behalf of the development group for ACROBAT-NRSI. A Cochrane risk of bias assessment tool: for non-randomized studies of interventions (ACROBAT-NRSI), version 1.0.0, http://www.riskofbias.info (24 September 2014, accessed 6 January 2016).

23.

Gerson

Edson

Lavori

et al.

Use of a simple symptom questionnaire to predict Barrett’s esophagus in patients with symptoms of gastroesophageal reflux. Am J Gastroenterol 2001; 96: 2005–2012.

24.

Spechler

Jain

Tendler

et al.

Racial differences in the frequency of symptoms and complications of gastro-oesophageal reflux disease. Aliment Pharmacol Ther 2002; 16: 1795–1800.

25.

Abrams

Fields

Lightdale

et al.

Racial and ethnic disparities in the prevalence of Barrett’s esophagus among patients who undergo endoscopy. Clin Gastroenterol Hepatol 2008; 6: 30–34.

26.

Gilani

Gerkin

Ramirez

et al.

Prevalence of Barrett’s esophagus in patients with moderate to severe erosive esophagitis. World J Gastroenterol 2008; 14: 3518–3522.

27.

Wang

Mattek

Corless

et al.

The value of traditional upper endoscopy as a diagnostic test for Barrett’s esophagus. Gastrointest Endosc 2008; 68: 859–866.

28.

Fan

Snyder

. Prevalence of Barrett’s esophagus in patients with or without GERD symptoms: Role of race, age and gender. Dig Dis Sci 2009; 54: 572–577.

29.

Nguyen

Thrift

Ramsey

et al.

Risk factors for Barrett’s esophagus compared between African Americans and non-Hispanic whites. Am J Gastroenterol 2014; 109: 1870–1880.

30.

Fischbach

Nordenstedt

Kramer

et al.

The association between Barrett’s esophagus and Helicobacter pylori infection: A meta-analysis. Helicobacter 2012; 17: 163–175.

31.

Epplein

Signorello

Zheng

et al.

Race, African ancestry and Helicobacter pylori infection in a low-income United States population. Cancer Epidemiol Biomarkers Prev 2011; 20: 826–834.

32.

Palles

Findlay

Tomlinson

. Common variants confer susceptibility to Barrett’s esophagus: Insights from the first genome-wide association studies. Adv Exp Med Biol 2016; 908: 265–290.

33.

Andrici

Tio

Cox

et al.

Hiatal hernia and the risk of Barrett’s esophagus. J Gastroenterol Hepatol 2013; 28: 415–431.

34.

Singh

Sharma

Murad

et al.

Central adiposity is associated with increased risk of esophageal inflammation, metaplasia and adenocarcinoma: A systematic review and meta-analysis. Clin Gastroenterol Hepatol 2013; 11: 1399–1412.

35.

Beales

Dearman

Vardi

et al.

Reduced risk of Barrett’s esophagus in statin users: Case-control study and meta-analysis. Dig Dis Sci 2016; 61: 238–246.