Patients with Adenocarcinoma of the Esophagus or Esophagogastric Junction Frequently Have Potential Screening Opportunities

RESEARCH LETTER

Symptoms of gastroesophageal reflux disease (GERD) are risk factors for Barrett’s esophagus (BE) and esophageal adenocarcinoma (EAC). Guidelines recommend considering endoscopic screening for patients with GERD and additional risk factors.¹ Screening programs can fail due to failure of uptake of initial screening or of surveillance among those with known BE. Prior studies have demonstrated that fewer than 20% of patients with EAC had an esophagogastroduodenoscopy (EGD) before diagnosis of cancer, but were typically defined as an EGD within just a few years before.^2–5 Those studies also did not address whether there were potential opportunities to undergo screening. Such opportunities could include encounters for GERD, prescription of acid-reducing medications, or for colorectal cancer screening. We quantified the proportion and types of potential opportunities for screening and surveillance among individuals who later developed EAC. Because esophagogastric junction adenocarcinoma (EGJAC) shares many risk factors with EAC we also explored the same opportunities for EGJAC.

We performed a case-control analysis of US Veterans. Cases were Veterans with EAC or EGJAC identified in the Veterans Health Administration (VHA) Central Cancer Registry diagnosed between 2010 and 2015. Age-matched controls were Veterans receiving services within VHA during the same time period. Diagnoses, procedures, prescriptions, and laboratory tests were extracted for the 10 years prior to the diagnosis of cancer or match date. We defined overdue for surveillance as a prior diagnosis of BE with most recent EGD greater than 5 years prior. We defined patients with a potential opportunity for screening as those with no prior EGD, but with one of the following: GERD diagnosis, acid-reducing medications, colonoscopy, or fecal occult blood testing (FOBT). Finally, we calculated Kunzmann scores, which predict the risk of incident EAC and EGJAC.^{8, 9}

We identified 3,781 patients with EAC and 1,427 with EGJAC (eTable 1). Even among cases without GERD or acid-reducing medications, Kunzmann scores were high (mean = 10.8). Of those with continuous care, 18.6% of EAC cases had undergone prior EGD (odds ratio [OR] vs. controls = 1.34; 95% confidence interval [CI] = 1.09, 1.64; eTable 2). Among EAC patients who had a prior EGD, 63.3% had a billing diagnosis of BE 6 months to 10 years prior to cancer diagnosis, and 21.7% had a billing diagnosis of BE within 6 months prior, indicating that EGD is sensitive for BE that may progress to EAC. Fewer EGJAC patients with a prior EGD were diagnosed with BE (32.5%); an additional 27.5% had a billing diagnosis for BE within 6 months prior. Among the 216 EAC patients with a prior EGD and BE diagnosis, 28.7% were overdue for surveillance, similar to controls (30.1%).

Among cancer cases without prior EGD, 85.4% and 86.0% of EAC and EGJAC cases, respectively, had at least one opportunity in which screening EGD might have been offered (eTables 1 and 2). EAC and EGJAC cases were only slightly more likely to have a GERD diagnosis than controls, and no more likely to have acid-reducing medications than controls. The most common potential opportunity to offer EGD was at the time of FOBT.

Overall, cancer cases with prior BE who were overdue for surveillance made up only 3.3% and 1.3% of EAC and EGJAC cases, respectively (Figure 1). The most common mechanism of failure to control cancer was missed potential opportunities to undergo screening (69.5% and 71.6% of EAC and EGJAC cases, respectively), including among those with a high predicted cumulative lifetime risk of EAC or EGJAC.

Figure 1. Proportional Mechanisms of Failure of Screening and Surveillance.

Panel A: Esophageal adenocarcinoma cases. Panel B: Esophagogastric junction adenocarcinoma cases. Panel C: Esophageal adenocarcinoma cases with Kunzmann score ≥ 8.5 indicating predicted lifetime cumulative incidence of either cancer ≥ 5.0%. Panel D: Esophagogastric junction adenocarcinoma cases with Kunzmann score ≥ 8.5

BE Dx: Barrett’s esophagus billing diagnosis, EAC: esophageal adenocarcinoma, EGD: esophagogastroduodenoscopy, EGJAC: esophagogastric junction adenocarcinoma, Mos.: months, Opp.: opportunity, Yrs: years

In summary, we conducted case-control analyses in the largest integrated health system in the US finding that the vast majority of patients with EAC or EGJAC had not undergone prior EGD, but very often had potential opportunities at which screening EGD could have been offered. Only a small minority of cases had been diagnosed with BE and were overdue for surveillance. Efforts directed at improved uptake of screening in individuals at increased risk has potential to substantially decrease the burden of EAC.

Our definition of a potential opportunity for screening included opportunities in those without a diagnosis of GERD, which is not explicitly endorsed by most published guidelines.¹ But our study confirms that those who ultimately developed EAC had a baseline elevated predicted risk of EAC even in the absence of documented GERD or acid-reducing medications. Given the very high prevalence of GERD diagnosis and acid-reducing medications among controls, there is likely substantial over-diagnosis of GERD and overuse of such medications, making them less useful predictors in actual clinical practice as opposed to research studies with validated GERD questionnaires. The discussions with patients about colorectal cancer screening or diagnostic colonoscopy are particularly opportune moments to consider EAC screening since EGD can be performed at the same encounter as colonoscopy, and at considerably lower cost.

Our study was limited to Veterans receiving care within the VHA, so the results might not be generalizable to other populations. However, patients managed in VHA are more likely to be up to date with colorectal cancer screening,^{8, 9} we found a similar proportion of individuals with a prior diagnosis of BE who were overdue for surveillance as has been demonstrated in a prior non-VHA study,¹⁰ and studies in other settings found similar low proportions of cancer cases having prior EGD.^{2, 4, 5} Our study was limited in that we cannot determine whether patients with potential screening opportunities were offered EGD but failed to adhere, or whether the EGD was not offered at all. Finally, there is scant high quality evidence regarding the effectiveness of screening and surveillance, including among those without GERD symptoms.

In conclusion, we found that more than 80% of individuals diagnosed with EAC or EGJAC had not undergone prior EGD, and the vast majority of those individuals had potential opportunities at which screening could have been offered. Further studies are needed to determine the magnitude of the benefits (morbidity, quality of life, mortality, stage shift, etc.) from screening, and if substantial, how best to leverage those opportunities to implement endoscopic or non-endoscopic screening among individuals at increased risk.

Abbreviations:

BE

Barrett’s esophagus

BMI

body mass index

CI

confidence interval

CCR

Central Cancer Registry

CDW

Corporate Data Warehouse

CPT

Current Procedural Terminology

EAC

esophageal adenocarcinoma

EGD

esophagogastroduodenoscopy

EGJAC

esophagogastric junction adenocarcinoma

FOBT

fecal occult blood testing

GERD

gastroesophageal reflux disease

H2RA

histamine type 2 receptor antagonists

ICD

International Classification of Diseases

LOINC

Logical Observation Identifiers Names and Codes

OR

odds ratio

PPI

proton pump inhibitors

SD

standard deviation

VHA

Veterans Health Administration