Acid Suppression Medications Reduce Risk of Oesophageal Adenocarcinoma in Barrett's Oesophagus

A Nested Case-control Study in US Male Veterans

M. C. Tan; H. B. El-Serag; X. Yu; A. P. Thrift


Aliment Pharmacol Ther. 2018;48(4):469-477. 

In This Article

Abstract and Introduction


Background: Proton pump inhibitors (PPIs) and histamine–2 receptor antagonists (H2RAs) may reduce the risk of oesophageal adenocarcinoma (OAC) in Barrett's oesophagus; however, current epidemiologic studies are inconclusive.

Aim: To evaluate the independent effects of PPIs and H2RAs on risk of OAC in patients with Barrett's oesophagus.

Methods: We conducted a nested case–control study of male veterans diagnosed with Barrett's oesophagus. Cases with incident OAC were matched by incidence density sampling on birth year and Barrett's diagnosis date to controls with Barrett's oesophagus who did not develop OAC. We identified prescription medication usage 1 year prior to Barrett's oesophagus diagnosis to 3 months prior to the OAC diagnosis. Odds ratios (OR) and 95% CI were estimated using conditional logistic regression.

Results: Compared with 798 controls, the 300 cases were less likely to use PPIs (90.0% vs 94.5%, P = 0.01) and H2RAs (19.7% vs 25.7%, P = 0.04). In the multivariable model including the use of statins, H2RAs, aspirin and nonsteroidal anti–inflammatory drugs, PPI use was associated with 41% lower risk of OAC (OR 0.59, 95% CI 0.35–0.99). While risk reduction of OAC was stronger for high–dose PPIs (omeprazole daily dose >40 mg, adjusted OR 0.11, 95% 0.04–0.36), we did not find a dose–response relationship with PPI duration (P trend = 0.45). Likewise, H2RA use was independently associated with 30% lower risk of OAC (OR 0.70, 95% CI 0.50–0.99).

Conclusion: Use of PPIs and H2RAs among patients with Barrett's oesophagus are associated with lower risk of OAC. Further clinical trials are needed to confirm this possible chemopreventive effect.


Patients with Barrett's oesophagus have a greater than 11–fold increased risk of oesophageal adenocarcinoma (OAC) compared with the general population.[1] Due to the continued rise in incidence of OAC and its high mortality to incidence ratio,[2,3] attention has turned to chemoprevention to delay or halt the progression of Barrett's oesophagus to neoplasia, including invasive cancer. For example, aspirin, nonsteroidal anti–inflammatory drugs (NSAIDs) and statins have been associated with reduced risk of progression to OAC in patients with Barrett's oesophagus.[4–8] Their combined use may be a cost–effective strategy for secondary prevention of OAC.[9]

Acid suppressive medications, such as proton pump inhibitors (PPIs) and histamine–2 receptor antagonists (H2RAs), are commonly used in patients with gastroesophageal reflux disease (GERD).[10] Chronic inflammation from GERD–related acid and bile acid exposure leads to Barrett's and OAC, possibly through increased cyclooxygenase (COX)–2 expression,[11] which inhibits apoptosis and increases cell proliferation.[12,13] PPIs and H2RAs decrease oesophageal acid and bile exposure and could have a role in mucosal healing and blocking the inflammation to neoplasia cascade.[14] Animal studies have shown acid suppression with PPIs decreases cellular proliferation, increases differentiation in Barrett's oesophagus[15] and decreases Barrett's and OAC development.[16] However, it has also been suggested that hypergastrinemia caused by PPIs may upregulate COX–2 expression and cell proliferation and may actually induce carcinogenesis in Barrett's patients.[17,18]

The independent effects of acid suppression medications on the risk of OAC in patients with Barrett's oesophagus remain unclear. In their meta–analysis of 7 observational studies (5 cohort and 2 case–control) published up to June 2013, Singh et al[19] found that the use of PPIs was associated with 71% lower risk of neoplastic progression to OAC or high–grade dysplasia (HGD) in patients with Barrett's oesophagus. A second meta–analysis that included 2 additional case–control studies[20,21] reported that the association with PPI use was attenuated towards the null and was not statistically significant.[22] However, considerable between–study heterogeneity was observed in both meta–analyses, and few studies controlled for concomitant use of H2RA, aspirin/NSAIDs and statins, or other known risk factors (including smoking and obesity) in their analysis.

Fewer studies have examined the chemopreventive effects of H2RAs independent of PPIs. Unlike the findings for PPI use, 2 cohort studies[23,24] and 1 case–control study[25] reported no association between H2RA use and risk of neoplastic progression in Barrett's oesophagus patients. However, analysis was limited by few patients using H2RAs exclusive of PPIs, small number of OAC cases developing in these patients (n = 156 from 3 studies), examining OAC alone as the endpoint, and limited adjustment for residual confounding from other medications.

Given the time–related biases in previous observational studies, we conducted a case–control study nested within a large national VA cohort of Barrett's oesophagus patients to examine the independent effects of PPIs and/or H2RAs on risk of OAC. In addition to being less susceptible to selection bias than the classic case–control study, nested case–control studies provide an unbiased estimate of the hazard ratio that would be obtained from a traditional time to event analysis of the full cohort.[26] Importantly, its inherent time–dependent nature ensures that it is also free of immortal time bias.