Abstract
Background & Aims:
Esophagogastroduodenoscopy (EGD) is frequently used to evaluate gastroesophageal reflux disease (GERD) without alarm symptoms, although the benefits are not clear. We aimed determine the prevalence of uncomplicated GERD as an indication for EGD, the demographic characteristics of these patients, and the endoscopic outcomes of these procedures.
Methods:
We collected endoscopy data from a large national database of 543,103 EGDs performed at 82 sites from 2003 through 2014. We identified patients with GERD without alarm symptoms (dysphagia, bleeding, vomiting, or weight loss). Endpoints included the prevalence of endoscopically suspected Barrett’s esophagus (sBE) and suspected long-segment BE (sLSBE), defined as ≥ 3cm.
Results:
We found that 73,535 EGDs (13.5%) were performed for a primary indication of GERD without alarm symptom: only 4122 patients (5.6%) had sBE, and of these, 24.2% had sLSBE. Significant risk factors for sBE and sLSBE included male sex, age, and white race. Other findings included peptic ulcers in 1337 patients (1.8%) and suspected tumors (47 esophageal, 42 gastric, 13 duodenal tumors, 2 other) in 101 patients.
Conclusion:
In an analysis of a large cohort of patients undergoing EGD, uncomplicated GERD accounted for almost 14% of EGDs; within this population, only 23.6% were white men older than 50 years, who have an increased risk of BE and esophageal adenocarcinoma. The prevalence of sBE is lower than in prior eras, and raises questions about the utility of EGDs to detect Barrett’s esophagus in patients with uncomplicated GERD. Guidelines for management of uncomplicated GERD should account for these observations.
Keywords: esophagus, diagnostic, cancer, detection
Introduction
Symptoms of gastroesophageal reflux disease (GERD) affect more than 25% of adults living in the United States [1]. Although GERD is a risk factor for developing Barrett’s esophagus (BE), the prevalence of BE in patients with GERD symptoms who do not have dysphagia is uncertain. The prevalence of BE in the proton pump inhibitor (PPI) era is also uncertain. In prior studies of patients with chronic GERD symptoms, 3-8% will have long-segment (≥3cm) Barrett’s esophagus, and 10-15% of these patients will have short-segment Barrett’s esophagus, in comparison with an estimated overall prevalence of 1.6-5.6% of BE of any length in the general population without reported GERD symptoms [2-6]. The role of upper gastrointestinal (GI) esophagogastroduodenoscopy (EGD) in the evaluation of GERD remains uncertain in the absence of alarm symptoms. The American Gastroenterological Association (AGA) strongly recommends against screening the general population with GERD for Barrett’s esophagus. The AGA also weakly recommends for screening patients with multiple risk factors associated with esophageal adenocarcinoma (age >50, male gender, white race, chronic GERD, hiatal hernia, elevated body mass index, intra-abdominal distribution of body fat), although is not clear how much weight to place on each of these factors [7].
In clinical practice, EGD is often used to evaluate GERD. Patients with concomitant alarm symptoms such as dysphagia, weight loss, and bleeding have clear indications for endoscopy regardless of the presence or absence of GERD. However, the benefit of using EGD for the evaluation of uncomplicated GERD in clinical practice is not known. Although our group has previously reported on the prevalence of GERD and BE across race, age, and gender, these studies largely preceded the widespread use of PPI, which became available over the counter in 2003 [8, 9]. We hypothesize that the utilization and outcomes of EGD for evaluation of uncomplicated GERD in the United States have changed over the period of 2003 to 2014.
The primary aims of this analysis were to determine 1) prevalence of GERD without alarm symptoms as an indication for EGD; 2) demographic characteristics and endoscopic outcomes in this population
The secondary aims of this analysis were to determine 1) whether there is an association between patient demographics and likelihood of Barrett’s esophagus; and 2) whether particular endoscopic findings, such as hiatal hernia or active esophagitis increase the likelihood of Barrett’s esophagus.
Methods
Clinical outcomes research initiative (CORI).
This project was developed in 1995, creating a consortium of clinical practice settings to determine utilization and outcomes of endoscopic procedures. Participating endoscopists use a structured computerized endoscopic report generator to produce endoscopic reports. The data that is transmitted from the local site to the National Endoscopic Database does not contain most patient or provider identifiers and qualifies as a Limited Data Set under 45 C.F.R. Section 164.514(e) [8]. After completion of quality control checks, data from all sites are merged in the data repository for analysis. Procedure counts are monitored on a weekly basis for atypical activity.
Practice Sites
Complete EGD reports received between January 1, 2003 and December 31, 2014, were included. The reports come from 82 practices. We have conducted several analyses to determine the representativeness of our data. We previously reported comparison of CMS data with the cohort of CORI subjects age 65 and older, demonstrating that the CORI cohort is representative of US endoscopy practice in this age group [10].
Patient Groups
Patients 18 years or older are included. Demographic characteristics of patients include age, sex, race and ethnicity. Race was recorded as white non-hispanic (White), black non-hispanic (Black), Hispanic, Asian, or Other. This analysis includes only the first exam for each patient during the study period. The endoscopist selected a procedure indication from a standard list available in CORI. The primary study group was defined as patients undergoing EGD with a primary indication of “reflux symptoms”, and allowing only secondary indications of “abdominal pain” and dyspepsia”.
Endoscopic endpoints
A key endpoint is “suspected” Barrett’s esophagus (sBE). Those with sBE 3cm were classified as suspected short-segment Barrett’s esophagus (sSSBE). Those with sBE ≥3cm were classified as suspected long-segment Barrett’s esophagus (sLSBE). The diagnosis of true BE depends on histological confirmation of intestinal metaplasia from biopsies in the tubular esophagus. Our prior work has shown that about 50% patients with sBE do not have biopsy-confirmed intestinal metaplasia, with longer segments of sBE being significantly more likely to have Barrett’s esophagus confirmed on histology [11]. This could be due to the absence of intestinal metaplasia or inadequate biopsies. We acknowledge that there may be misclassification of BE using endoscopic criteria alone. However, clinical decisions to obtain biopsies or perform follow-up exams are often based on the endoscopic appearance of the esophagus.
Other endoscopic outcomes recorded include presence of hiatal hernia, esophagitis, gastric ulcers, duodenal ulcers, and tumors. Depending on the user, esophagitis was graded either on a scale of mild, moderate, severe, or as Los Angeles (LA) Grade A, B, C, D[12]. For purposes of analysis, the grades were grouped as follows: mild/LA Grade A, moderate/LA Grade B, and severe/LA Grade C or D.
Analysis
The analysis considered the entire time period of 2003-2014. Analyses included the relationships between patient demographics and endoscopic findings; as well as relationships between findings of hiatal hernia and esophagitis with sBE.
Comparison of categorical data was performed by using Pearson’s chi-square test. Analysis of variance was performed to compare continuous variables. Univariate logistic regression analyses were performed to generate odds ratios for the outcomes of suspected BE and ulcer by age, gender and race/ethnicity. The models were limited to patients of White, Black, Hispanic, and Asian race/ethnicity due to small sample sizes for the remaining race/ethnic groups. All analyses were performed with version 9.4 of SAS software (SAS Institute Inc, Cary, NC).
Results
From 2003 to 2014, initial EGDs were performed on 543,103 patients. Of these, 73,535 (13.5%) were performed for a primary indication of uncomplicated GERD (Table 1). The majority were performed at community practices (79.3%), with the remainder at academic centers (10.2%) and VA medical centers (10.5%). The mean age was 53.1 years, with 38% of patients <50 years old. This population was 48% male and 81% Whites.
Table 1:
Demographics of Patients
| Total | 73,535 |
|---|---|
| Age | |
| <40 | 13219 (18.0) |
| 40-49 | 14753 (20.1) |
| 50-59 | 20640 (28.1) |
| 60-69 | 15725 (21.4) |
| >70 | 9198 (12.5) |
| Male Gender (%) | 35428 (48.2) |
| Male gender excluding VA/Military (%) |
29543 (44.9) |
| Race/Ethnicity | |
| Whites | 59492 (80.9) |
| Hispanic | 8532 (11.6) |
| Blacks | 3087 (4.2) |
| Asian/PI | 1341 (1.8) |
| Other | 1083 (1.5) |
Suspected BE was found in 4,122 (5.6%) patients with uncomplicated GERD (Table 2). Of these 4,122 patients with sBE, 3,159 (76.7%) had length recorded; 772 patients met criteria for sLSBE. Reflux esophagitis was found in 15,969 (21.7%) patients, with 9.2% of these cases classified as severe or LA Class C/D. A hiatal hernia was recorded in 31,489 (42.8%) patients.
Table 2:
EGD Findings
| Total | 73,535 |
|---|---|
| sBE | 4122 (5.6) |
| 0-3cm | 2392 (58.0) |
| >3cm | 767 (18.6) |
| Unknown | 963 (23.4) |
| Hiatal Hernia | 31489 (42.8) |
| Stricture | 2081 (2.8) |
| Ulcer | 1337 (1.8) |
| Gastric Ulcer | 876 (1.2) |
| Duodenal Ulcer | 361 (0.5) |
| Tumor | 101 (0.1) |
| Gastric tumor | 41 (0.1) |
| Esophageal tumor | 47 (0.1) |
| Duodenal tumor | 13 (0.0) |
| Reflux Esophagitis | 15969 (21.7) |
| LA Class/Severity Grade | 14920 |
| A or Mild | 8935 (59.9) |
| B or Moderate | 4611 (30.9) |
| C/D or Severe | 1374 (9.2) |
Endoscopic findings were stratified by gender (Table 3), age (Table 4), and ethnicity (Table 5). Adjusted odds ratios relative to patients without sBE are shown in Table 6. sBE was significantly more common in men compared to women (8.1% vs 3.3%, OR 2.61, 95% CI 2.44 – 2.79). The prevalence increased with age from 4.0% in patients <40 years to 6.3% in patients older than 60 years. Whites were more likely to have Barrett’s esophagus compared to all other races (6.1% vs 2.1%, OR 1.81, 95% CI 1.65 – 2.00). Among those with length of Barrett’s esophagus recorded (2228 men, 931 women), the prevalence of sLSBE was higher in men compared to women (OR 4.40, 95% CI 3.69 – 5.25) and in Whites compared to all other races (1.2% vs 0.5%, OR 2.49, 95%CI 1.94 – 3.20). Rates of sLSBE (as a proportion of all sBE with recorded length) increased with age from 13.1% in those <40 years to 22.5% in patients >60 years old.
Table 3:
Comparison of EGD findings by Gender
| Male | Female | p value |
|
|---|---|---|---|
| Total | 35,428 | 28,107 | |
| Barrett’s esophagus (suspected) | 2874 (8.1) | 1248 (3.3) | <.0001 |
| 0-3cm | 1618 (56.3) | 774 (62.0) | <.0001 |
| ≥3cm | 610 (21.2) | 157 (12.6) | <.0001 |
| Unknown | 646 (22.5) | 317 (25.4) | |
| Hiatal Hernia | 15438 (43.6) | 16051 (42.2) | <.0001 |
| Stricture | 1128 (3.2) | 953 (2.5) | <.0001 |
| Ulcer | 641 (1.8) | 696 (1.8) | 0.862 |
| Gastric ulcer | 370 (1.0) | 506 (1.3) | |
| Duodenal ulcer | 205 (0.6) | 156 (0.4) | |
| Tumor | 69 (0.2) | 32 (0.1) | <.0001 |
| Gastric tumor | 20 (0.1) | 21 (0.1) | |
| Esophageal tumor | 39 (0.1) | 8 (0.0) | |
| Duodenal tumor | 10 (0.0) | 3 (0.0) | |
| Reflux Esophagitis | 8313 (23.5) | 7656 (20.1) | <.0001 |
| LA Class/Severity Grade | 7850 | 7070 | |
| A or Mild | 4208 (53.6) | 4727 (66.9) | <.0001 |
| B or Moderate | 2731 (34.8) | 1880 (26.6) | |
| C/D or Severe | 911 (11.6) | 463 (6.6) |
Table 4:
Comparison of EGD findings by Age
| <40 | 40-49 | 50-59 | 60-69 | >70 | |
|---|---|---|---|---|---|
| Total | 13,219 | 14,753 | 20,640 | 15,725 | 9,198 |
| Barrett’s esophagus (suspected) |
534 (4.0) | 774 (5.3) | 1256 (6.1) | 1040 (6.6) | 518 (5.6) |
| 0-3cm | 336 (62.9) | 464 (59.7) | 755 (59.9) | 558 (53.7) | 279 (53.9) |
| ≥3cm | 70 (13.1) | 114 (15.0) | 233 (16.8) | 216 (20.7) | 134 (25.9) |
| Unknown | 128 (24.0) | 196 (25.3) | 268 (21.3) | 266 (25.6) | 105 (20.3) |
| Hiatal Hernia | 4794 (36.3) | 5880 (39.9) | 8909 (43.2) | 7206 (45.8) | 4701 (51.1) |
| Stricture | 142 (1.1) | 302 (2.1) | 587 (2.8) | 593 (4.7) | 457 (5.0) |
| Ulcer | 175 (1.3) | 229 (1.6) | 268 (1.8) | 360 (2.3) | 205 (2.2) |
| Gastric ulcer | 107 (0.8) | 140 (1.0) | 248 (1.2) | 246 (1.6) | 135 (1.5) |
| Duodenal ulcer | 51 (0.4) | 75 (0.5) | 102 (0.5) | 86 (0.6) | 47 (0.5) |
| Tumor | 6 (0.0) | 10 (0.1) | 31 (0.2) | 27 (0.2) | 27 (0.3) |
| Gastric tumor | 3 (0.0) | 7 (0.1) | 17 (0.1) | 7 (0.0) | 7 (0.1) |
| Esophageal tumor | 3 (0.0) | 2 (0.0) | 13 (0.1) | 16 (0.1) | 13 (0.1) |
| Duodenal tumor | 0 (0.0) | 1 (0.0) | 0 (0) | 4 (0.0) | 8 (0.1) |
| Reflux Esophagitis | 3301 (25.0) | 3356 (22.8) | 4460 (21.6) | 3051 (19.4) | 1801 (19.6) |
| LA Class/Severity Grade |
3091 | 3160 | 4174 | 2817 | 1678 |
| A or Mild | 1849 (59.8) | 1903 (60.2) | 2479 (59.4) | 1693 (60.1) | 1011 (60.3) |
| B or Moderate | 1004 (32.5) | 970 (30.7) | 1297 (31.1) | 845 (30.0) | 495 (29.5) |
| C/D or Severe | 238 (7.7) | 287 (9.1) | 398 (9.5) | 279 (9.9) | 172 (10.3) |
Table 5:
Comparison of EGD findings by Ethnicity
| White | Hispanic | Black | Asian/PI | Other | |
|---|---|---|---|---|---|
| Total | 59,492 | 8,532 | 3,087 | 1,341 | 1,083 |
| Barrett’s esophagus (suspected) | 3635 (6.1) | 308 (3.6) | 94 (3.1) | 39 (2.9) | 46 (4.2) |
| 0-3cm | 2121 (58.4) | 152 (49.4) | 66 (70.2) | 27 (69.2) | 26 (56.5) |
| ≥3cm | 699 (19.2) | 40 (13.0) | 10 (10.6) | 8 (20.6) | 10 (21.7) |
| Unknown | 815 (22.4) | 116 (37.7) | 18 (19.2) | 4 (10.3) | 10 (21.7) |
| Hiatal Hernia | 26163 (44.0) | 3368 (39.5) | 1240 (40.2) | 356 (26.6) | 362 (33.4) |
| Stricture | 1836 (3.1) | 110 (1.3) | 99 (3.2) | 10 (0.8) | 26 (2.4) |
| Ulcer | 995 (1.7) | 218 (2.6) | 69 (2.2) | 39 (2.9) | 16 (1.5) |
| Gastric ulcer | 652 (1.1) | 140 (1.6) | 44 (1.4) | 26 (1.9) | 14 (1.3) |
| Duodenal ulcer | 254 (0.4) | 71 (0.8) | 24 (0.8) | 11 (0.8) | 1 (0.1) |
| Tumor | 75 (0.1) | 14 (0.2) | 5 (0.2) | 2 (0.2) | 5 (0.5) |
| Gastric tumor | 24 (0.0) | 12 (0.1) | 2 (0.1) | 1 (0.1) | 2 (0.2) |
| Esophageal tumor | 41 (0.1) | 1 (0.0) | 2 (0.1) | 1 (0.1) | 2 (0.2) |
| Duodenal tumor | 10 (0.0) | 1 (0.0) | 1 (0.0) | 0 (0) | 1 (0.1) |
| Reflux Esophagitis | 12921 (21.7) | 1968 (23.1) | 590 (19.1) | 211 (15.7) | 279 (25.8) |
| LA Class/Severity Grade | 12153 | 1795 | 522 | 191 | 259 |
| A or Mild | 7186 (59.1) | 1131 (63.0) | 330 (63.2) | 132 (69.1) | 156 (60.2) |
| B or Moderate | 3842 (31.6) | 475 (26.5) | 163 (31.2) | 52 (27.2) | 79 (30.5) |
| C/D or Severe | 1125 (9.3) | 189 (10.5) | 29 (5.6) | 7 (3.7) | 24 (9.3) |
Table 6:
Risk of EGD findings (univariate)
| Odds Ratio of sBE (95% CI) |
Odds Ratio of sLSBE (95% CI) |
|
|---|---|---|
| Age | ||
| <40 | Reference | Reference |
| 40-49 | 1.32 (1.18 - 1.47) | 1.48 (1.10 - 1.99) |
| 50-59 | 1.54 (1.39 - 1.71) | 2.18 (1.67 - 2.85) |
| 60-69 | 1.68 (1.51 - 1.87) | 2.67 (2.03 - 3.49) |
| >=70 | 1.42 (1.25 - 1.61) | 2.80 (2.09 - 3.74) |
| Gender | ||
| Female | Reference | Reference |
| Male | 2.61 (2.44 - 2.79) | 4.40 (3.69 – 5.25) |
| Race/Ethncity | ||
| White | Reference | Reference (vs non-White) |
| Hispanic | 0.58 (0.51 - 0.65) | 0.39 (0.28 – 0.54) |
| Black | 0.48 (0.39 - 0.59) | 0.27 (0.14 – 0.50) |
| Asian/PI | 0.46 (0.33 - 0.63) | 0.49 (0.24 – 0.98) |
| Other | 0.68 (0.51 – 0.92) | 0.77 (0.41 – 1.44) |
| Hiatal Hernia | ||
| No | Reference | Reference |
| Yes | 1.60 (1.50 - 1.70) | 2.61 (2.25 – 3.04) |
| Reflux Esophagitis | ||
| No | Reference | Reference |
| Yes (Any) | 0.32 (0.29 - 0.36) | 0.45 (0.36 – 0.56) |
| LA Class/Severity Grade | (n= 14,920) | (n= 14,920) |
| A | Reference | Reference |
| B | 2.19 (1.72 - 2.78) | 3.62 (2.03 – 6.42) |
| C/D | 3.50 (2.59 - 4.73) | 13.10 (7.40 – 23.20) |
Esophagitis was more common in men compared to women (OR 1.22, 95% CI 1.18 – 1.26) and Hispanics compared to Whites (OR 1.08, 95% CI 1.02 – 1.14). While the prevalence of esophagitis decreased with age, severe or LA Class C/D esophagitis increased with age from 7.7% in patients <40 years to 10.0% in patients >60 years.
There were 104 tumors described in 101 patients. Of these, there were 47 esophageal tumors, 42 gastric tumors, 13 duodenal tumors, one tumor in the palate, and one patient did not have a specific tumor location recorded. Of the 47 patients with esophageal tumors, there were 39 (83.0%) men, 42 (89.4%) had age >50 years, and 41 (87.2%) were white. Of the 41 patients with gastric tumors, 20 (48.8%) were men, 31 (75.6%) had age >50 years, and 24 (58.5%) were white. Thirty-four of these patients were white, male, and had age >50 years. Of the 13 patients with duodenal tumors, 10 (76.9%) were men,12 (92.3%) had age >50 years, and 10 (76.9%) were white.
Endoscopy also detected non-esophageal pathology. The prevalence of gastric and duodenal ulcers was 1.8% and was similar in men and women, but increased with advancing age. Compared to Whites (1.7%), ulcers were more prevalent Blacks (2.2%, OR 1.34, 95% CI 1.05 – 1.72) and Hispanics (2.9%,OR 1.54, 95% CI 1.33 – 1.79).
Discussion
This is the largest analysis of upper endoscopy practice for uncomplicated GERD which has ever been reported from non-claims data sources in the United States, and shows that EGDs done for a primary indication of uncomplicated GERD account for more than 10% of all EGDs annually. In this population, 5.6% patients had sBE and an estimated 1.4% patients had sLSBE. Male gender, White race, and advancing age were associated with both increased sBE and sLSBE, Severe esophagitis was found in an estimated 2% patients and ulcers were found in 1.8% patients, while tumors were rare.
While the demographics largely mirrored that of patients undergoing EGD for other reasons in the CORI database, one notable exception is that the percentage of patients <50 years old undergoing EGDs was actually higher in this subset of patients undergoing EGD for uncomplicated GERD. Despite the AGA recommendation to focus screening on patients with multiple risks of adenocarcinoma, namely White males, age >50 years, it is notable that the majority of EGD-GERD are performed in women, 38% of EGDs are performed before age 50, and 25% performed in non-Whites. Only 23.6% patients met all three of the criteria of the AGA recommendation.
Among this group of patients undergoing EGD for uncomplicated GERD, the most common findings were esophagitis (21.7%) and hiatal hernias (43%). North American studies from the late 1990s to early 2000s described rates of esophagitis from 34% to 56% in patients undergoing EGD for uncomplicated GERD.[13, 14] The lower percentage of patients with esophagitis in our cohort may speak to the efficacy of acid reflux treatment among patients who receive endoscopy in the PPI era. .[15]
Patients with uncomplicated GERD and esophagitis can often be treated empirically, and may not necessarily benefit from EGD. Findings of notable pathology that warrant further management were found on 9.5% of EGDs in patients with uncomplicated GERD, with the majority of these being sBE (5.6%); other findings included severe or LA Grade C/D esophagitis (2%), gastric or duodenal ulcers (1.8%), and suspected tumors (<0.1%). The rate of sLSBE in our study is a little more difficult to calculate. Assuming the percentage of sLSBE remained consistent for those with unknown length of sBE, 1.4% of all patients undergoing EGD for uncomplicated GERD have sLSBE, with a range from 1.0% (none of the unknown lengths were sLSBE) to 2.4% (all of the unknown lengths were sLSBE). The true rates of BE (based on histology) in our study are likely lower than the rates of suspected BE, based on our prior work.[11]
These rates of both sBE (5.6%) and sLSBE (1.4%) are lower than prior estimates of Barrett’s esophagus in GERD patients. Studies from the 1970s and 1980s suggested rates of Barrett’s esophagus in patients with GERD symptoms ranging from 8.2% to 12.4%.[3, 16, 17] Many of these studies included patients with dysphagia or abnormal imaging, which could have inflated the rates of BE. Studies from the late 1990s to early 2000s in patients with GERD symptoms, found rates of Barrett’s esophagus ranging from 8.3% to 13.2%, and rates of LSBE ranging from 2.6% to 4.8%.[4, 5, 18] Lower rates of sBE in the current study could reflect the inclusion of patients with alarm symptoms in prior studies as well as changes in utilization of EGD for GERD. Since the mid-1990’s, EGD has often been used to screen patients with GERD for BE, with or without alarm symptoms. It is also possible that there is a true decline in rates of BE, over the past 20 to 30 years. However, since obesity is a risk factor for GERD and BE, we might have otherwise expected an increased, not a decreased rate of BE. Our finding of lower rates of BE coincides with increased prevalence of PPI use, which may have prevented the onset of BE in some patients.[19]
Our data support the hypothesis that rates of BE from 2003-2014 are lower than previously described in patients who have uncomplicated GERD. As shown by prior studies, age >50 years (OR 1.34, 95% CI 1.25 – 1.44), male gender (OR 2.61, 95% CI 2.44 – 2.79), and White race (OR 1.81, 95% CI 1.65 – 2.00), were strongly associated with increased risk of sBE.[8, 20-22] However, the absolute risk for these patient groups remains relatively low, with the following respective rates: age >50 (sBE 6.1%, LSBE 1.7%), male gender (sBE 8.1%, LSBE 2.2%), White race (sBE 6.1%, LSBE 1.5%). In the highest risk cohort of White men with age >50, prevalence of sBE was 9.8%% (number needed to screen is 10) and sLSBE was 1.4% (number needed to screen range 23-43). It has previously been shown in a model that screening for Barrett’s esophagus in White men, age>50, with GERD symptoms is likely cost-effective, although the model assumed a prevalence of 10% of Barrett’s esophagus in this population.[33] Our prevalence of histologically proven sBE would fall far short of this threshold. Given the even lower prevalence of sBE in women (3.3%), age <50 (4.6%), and non-White patients (4.5%), the value of endoscopy in these low-risk populations is particularly limited. If EGDs were limited to patients that met the AGA criteria of being male, white, and age >50, we would have detected 34/47 (72.3%) of esophageal tumors as well as finding sBE in nearly 10%, while reducing the burden of endoscopy by more than 75%.
The presence of a hiatal hernia was strongly associated with sBE (OR 1.6, 95% CI 1.5 – 1.7) and LSBE (OR 1.86, 95% CI 1.57 – 2.21). Although the presence of esophagitis was not associated with sBE, it was associated with the presence of sLSBE. It is possible that part of the lack of association is due to esophagitis masking underlying Barrett’s esophagus, which is more commonly short-segment in nature.[23] Severe or Grade C/D esophagitis was still strongly associated with both sBE and sLSBE. We suspect that many patients with uncomplicated GERD received some anti-secretory therapy prior to having EGD, and that patients with sSSBE may have milder forms of esophagitis or disease that is better controlled with medical treatment, compared to individuals with sLSBE.
Chronic GERD can cause Barrett’s esophagus, which can lead to esophageal adenocarcinoma (EAC).[24] The incidence of EAC has steadily risen since the 1970s, initially increasing at a rate of 8% per year until the late 1990s, then slowing to a rate of increase of less than 2% per year since then, with an incidence of 2.8 per 100,000 in 2012.[25, 26] Hypotheses for this deceleration of EAC risk include decline in smoking, better management of GERD or surveillance of Barrett’s esophagus. Since many patients with EAC do not have GERD symptoms, screening patients with GERD can only have a limited effect (27,28). Among patients with Barrett’s esophagus, the estimated annual incidence of EAC is only 0.1 to 0.4% per year and the estimated mortality from EAC is 3 per 1000 person-years [29-32].
The prevalence of ulcers in patients with uncomplicated GERD symptoms is not uncommon (1.8%), which may reflect some overlap of symptoms of GERD and dyspepsia. The demographic risk factors for ulcers are different than for sBE. While age >50 was also associated with increased risk of ulcers (OR 1.43, 95%CI 1.27 – 1.60), there was no difference in ulcer rates between men and women, and white patients were at lower risk compared to non-White patients (OR0 0.68, 95%CI 0.60 – 0.77).
Tumors were a very rare finding (0.1%) and shared similar risk factors with Barrett’s esophagus, namely male gender, increased age, and White race, with the exception of gastric tumors which were evenly split between male and female patients. This raises questions about the role of biology and genetic differences in the development of different types of upper GI pathology.
Limitations
The CORI database provides snapshots of endoscopic practice which depend on data provided in the endoscopic report. We believe these data are as accurate as possible because they become part of the official medical record and are used for quality assurance purposes. Misclassification of race or ethnicity is possible, because the endoscopist or nurse provided this information.
Procedure indications are provided by the endoscopists. In patients with multiple reasons for examination, some misclassification is possible. Key outcomes data are adjusted for age, gender and race to reduce the likelihood that practice changes may have altered the key outcomes. The selection of suspected BE and suspected malignancy as study endpoints is discussed in the methods. We acknowledge that many patients with “suspected” BE will not have histologically proven BE. Our prior work suggests that about 40% of patients with sSSBE and 75% of sLSBE will have histologic intestinal metaplasia consistent with Barrett’s esophagus [11]. Therefore, we believe that the detection of suspected BE overestimates the true prevalence of BE in this population. Certain findings such as hiatal hernia and esophagitis may be under-reported in patients who have other findings such as sBE or ulcer.
Conclusion
We report on data from more than 70,000 EGDs performed in diverse clinical practices for GERD without alarm symptoms. This accounted for 13.5% of all EGDs done during this time period and reflects a significant cost to the healthcare system. This is the largest report of EGD outcomes in the post-PPI era. In the absence of alarm symptoms, it is likely that most EGDs were performed to determine the presence or absence of BE.
Our study confirms previously suggested risk factors of older age, male gender, and White ethnicity for sBE and sLSBE. We find that low risk populations with uncomplicated GERD make up a significant number of EGDs done for uncomplicated GERD, with White men, age >50 comprising only 23.6% of this cohort.
Our study also demonstrates a lower prevalence of sBE and sLSBE compared to prior studies, which raises questions about the utility of EGDs to assess for Barrett’s esophagus in any population with uncomplicated GERD. In the absence of alarm symptoms, GERD is uncommonly associated with sLSBE. Future guidelines for management of uncomplicated GERD should account for these changes in endoscopic outcomes.
Acknowledgments
Grant Support:
R01 CA140574
Title: Improving Esophageal Adenocarcinoma Prevention, Screening and Treatment, PI: Chin Hur; Childrens Hospital Boston, NIDDK U01DK57132
Title: Clinical Outcomes Research Initiative
PI: David Lieberman; Oregon Health Sciences University
Abbreviations:
- (AGA)
American Gastroenterological Association
- (BE)
Barrett’s esophagus
- (CORI)
Clinical outcomes research initiative
- (EGD)
Esophagogastroduodenoscopy
- (GERD)
Gastroesophageal reflux disease
- (LA)
Los Angeles
- (PPI)
Proton pump inhibitor
- (sBE)
Suspected Barrett’s esophagus
- (sSSBE)
Suspected short-segment Barrett’s esophagus
- (sLSBE)
Suspected long-segment Barrett’s esophagus
Footnotes
Financial Disclosures:
The authors have no financial or personal disclosures relevant to this manuscript
Publisher's Disclaimer: 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 citable 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.
Background
EGDs are frequently used to evaluate GERD without alarm symptoms, although the benefits are not clear.
Findings
In an analysis of a large cohort of patients undergoing EGD, uncomplicated GERD accounted for almost 14% of EGDs; within this population, only 5.6% had suspected Barrett’s esophagus of any length, and an estimated 1.4% had suspected Barrett’s esophagus ≥3cm. Only 23.6% of EGDs for uncomplicated GERD were in white men older than 50.
Implications for patient care
The current prevalence of sBE is lower than in prior eras, and raises questions about the utility of EGDs to detect Barrett’s esophagus in patients with uncomplicated GERD.
Contributor Information
Emery C Lin, Division of Gastroenterology and Hepatology, Department of Medicine, Oregon Health Sciences University, 3181 SW Sam Jackson Park Road, Portland, OR 97239.
Jennifer Holub, Division of Gastroenterology and Hepatology, Department of Medicine, Oregon Health Sciences University, 3181 SW Sam Jackson Park Road, Portland, OR 97239.
David Lieberman, Division of Gastroenterology and Hepatology, Department of Medicine, Oregon Health Sciences University, 3181 SW Sam Jackson Park Road, Portland, OR 97239.
Chin Hur, Division of Gastroenterology and Hepatology, Department of Medicine, Massachusetts General Hospital, 55 Fruit Street, Boston, MA 02114.
References
- 1.Camilleri M, et al. , Prevalence and socioeconomic impact of upper gastrointestinal disorders in the United States: results of the US Upper Gastrointestinal Study. Clin Gastroenterol Hepatol, 2005. 3(6): p. 543–52. [DOI] [PubMed] [Google Scholar]
- 2.Cameron AJ, Epidemiology of Barrett’s esophagus and adenocarcinoma. Dis Esophagus, 2002. 15(2): p. 106–8. [DOI] [PubMed] [Google Scholar]
- 3.Sarr MG, et al. , Barrett’s esophagus: its prevalence and association with adenocarcinoma in patients with symptoms of gastroesophageal reflux. Am J Surg, 1985. 149(1): p. 187–93. [DOI] [PubMed] [Google Scholar]
- 4.Lieberman D, et al. , Endoscopic evaluation of patients with dyspepsia: results from the national endoscopic data repository. Gastroenterology, 2004. 127(4): p. 1067–75. [DOI] [PubMed] [Google Scholar]
- 5.Rex DK, et al. , Screening for Barrett’s esophagus in colonoscopy patients with and without heartburn. Gastroenterology, 2003. 125(6): p. 1670–7. [DOI] [PubMed] [Google Scholar]
- 6.Hayeck TJ, et al. , The prevalence of Barrett’s esophagus in the US: estimates from a simulation model confirmed by SEER data. Dis Esophagus, 2010. 23(6): p. 451–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Spechler SJ, et al. , American Gastroenterological Association medical position statement on the management of Barrett’s esophagus. Gastroenterology, 2011. 140(3): p. 1084–91. [DOI] [PubMed] [Google Scholar]
- 8.Rubenstein JH, Mattek N, and Eisen G, Age- and sex-specific yield of Barrett’s esophagus by endoscopy indication. Gastrointest Endosc, 2010. 71(1): p. 21–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wang A, et al. , Prevalence of complicated gastroesophageal reflux disease and Barrett’s esophagus among racial groups in a multi-center consortium. Dig Dis Sci, 2009. 54(5): p. 964–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sonnenberg A, et al. , Patterns of endoscopy in the United States: analysis of data from the Centers for Medicare and Medicaid Services and the National Endoscopic Database. Gastrointest Endosc, 2008. 67(3): p. 489–96. [DOI] [PubMed] [Google Scholar]
- 11.Wang A, et al. , The value of traditional upper endoscopy as a diagnostic test for Barrett’s esophagus. Gastrointest Endosc, 2008. 68(5): p. 859–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Lundell LR, et al. , Endoscopic assessment of oesophagitis: clinical and functional correlates and further validation of the Los Angeles classification. Gut, 1999. 45(2): p. 172–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Blustein PK, et al. , The utility of endoscopy in the management of patients with gastroesophageal reflux symptoms. Am J Gastroenterol, 1998. 93(12): p. 2508–12. [DOI] [PubMed] [Google Scholar]
- 14.Wo JM, et al. , Clinical impact of upper endoscopy in the management of patients with gastroesophageal reflux disease. Am J Gastroenterol, 2004. 99(12): p. 2311–6. [DOI] [PubMed] [Google Scholar]
- 15.Pottegard A, et al. , Use of proton-pump inhibitors among adults: a Danish nationwide drug utilization study. Therap Adv Gastroenterol, 2016. 9(5): p. 671–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Mann NS, Tsai MF, and Nair PK, Barrett’s esophagus in patients with symptomatic reflux esophagitis. Am J Gastroenterol, 1989. 84(12): p. 1494–6. [PubMed] [Google Scholar]
- 17.Winters C Jr., et al. , Barrett’s esophagus. A prevalent, occult complication of gastroesophageal reflux disease. Gastroenterology, 1987. 92(1): p. 118–24. [PubMed] [Google Scholar]
- 18.Westhoff B, et al. , The frequency of Barrett’s esophagus in high-risk patients with chronic GERD. Gastrointest Endosc, 2005. 61(2): p. 226–31. [DOI] [PubMed] [Google Scholar]
- 19.Friedenberg FK, et al. , Trends in gastroesophageal reflux disease as measured by the National Ambulatory Medical Care Survey. Dig Dis Sci, 2010. 55(7): p. 1911–7. [DOI] [PubMed] [Google Scholar]
- 20.Cook MB, Wild CP, and Forman D, 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(11): p. 1050–61. [DOI] [PubMed] [Google Scholar]
- 21.Corley DA, et al. , Race, ethnicity, sex and temporal differences in Barrett’s oesophagus diagnosis: a large community-based study, 1994–2006. Gut, 2009. 58(2): p. 182–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Abrams JA, et al. , Racial and ethnic disparities in the prevalence of Barrett’s esophagus among patients who undergo upper endoscopy. Clin Gastroenterol Hepatol, 2008. 6(1): p. 30–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Hanna S, et al. , Detection of Barrett’s esophagus after endoscopic healing of erosive esophagitis. Am J Gastroenterol, 2006. 101(7): p. 1416–20. [DOI] [PubMed] [Google Scholar]
- 24.Cook MB, et al. , Gastroesophageal reflux in relation to adenocarcinomas of the esophagus: a pooled analysis from the Barrett’s and Esophageal Adenocarcinoma Consortium (BEACON). PLoS One, 2014. 9(7): p. e103508. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Thrift AP, The epidemic of oesophageal carcinoma: Where are we now? Cancer Epidemiol, 2016. 41: p. 88–95. [DOI] [PubMed] [Google Scholar]
- 26.Hur C, et al. , Trends in esophageal adenocarcinoma incidence and mortality. Cancer, 2013. 119(6): p. 1149–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Bytzer P, et al. , Adenocarcinoma of the esophagus and Barrett’s esophagus: a population-based study. Am J Gastroenterol, 1999. 94(1): p. 86–91. [DOI] [PubMed] [Google Scholar]
- 28.Lagergren J, et al. , Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med, 1999. 340(11): p. 825–31. [DOI] [PubMed] [Google Scholar]
- 29.Hvid-Jensen F, et al. , Incidence of adenocarcinoma among patients with Barrett’s esophagus. N Engl J Med, 2011. 365(15): p. 1375–83. [DOI] [PubMed] [Google Scholar]
- 30.Sikkema M, et al. , Risk of esophageal adenocarcinoma and mortality in patients with Barrett’s esophagus: a systematic review and meta-analysis. Clin Gastroenterol Hepatol, 2010. 8(3): p. 235–44; quiz e32. [DOI] [PubMed] [Google Scholar]
- 31.Desai TK, et al. , The incidence of oesophageal adenocarcinoma in non-dysplastic Barrett’s oesophagus: a meta-analysis. Gut, 2012. 61(7): p. 970–6. [DOI] [PubMed] [Google Scholar]
- 32.Shakhatreh MH, et al. , The incidence of esophageal adenocarcinoma in a national veterans cohort with Barrett’s esophagus. Am J Gastroenterol, 2014. 109(12): p. 1862–8; quiz 1861, 1869. [DOI] [PubMed] [Google Scholar]
- 33.Inadomi JM, et al. , Screening and surveillance for Barrett esophagus in high-risk groups: a cost-utility analysis. Ann Intern Med, 2003. 138(3): p. 176–86. [DOI] [PubMed] [Google Scholar]