NSAIDs and esophageal cancer: Ready for trials but not yet broad clinical application☆
Article Outline
Abstract
GASTROENTEROLOGY 2003;124:246-248
See article on page 47.
Research on the potential of aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs) to prevent esophageal cancer has been largely overshadowed by interest in the chemopreventive potential of NSAIDs against colorectal cancer. The relationship of NSAIDs to esophageal cancer is important, however, because of the nearly 7-fold increase in the incidence of esophageal adenocarcinoma among U.S. males since the mid-1970s,1 the high prevalence of gastroesophageal reflux disease (GERD) in the general population, the typically late stage at which esophageal cancers are diagnosed, the poor response to therapy, and low survival rates.2 In this issue of GASTROENTEROLOGY, Corley et al.3 provide the first systematic overview and meta-analysis of the epidemiologic studies of NSAIDs and esophageal cancer. The following paragraphs consider what a meta-analysis of observational studies can add to what is already known about NSAIDs and esophageal cancer, the biological basis whereby NSAIDs might inhibit the development of esophageal cancer, the current status of randomized clinical trials of NSAIDs in relation to adenocarcinoma and squamous cell carcinoma of the esophagus, and key issues that must be resolved to allow informed clinical decisions about the use of NSAIDs as anticancer agents.
One contribution of the meta-analysis is to provide a more precise estimate of the association between NSAID use and esophageal cancer than is obtained from any of the individual epidemiological studies. On average, persons who report any use of aspirin or other NSAIDs in these studies have approximately 40% lower annual risk of esophageal cancer (odds ratio [OR], 0.57; 95% confidence interval [95% CI], 0.47–0.71) than those who report no use. The inverse association observed here between NSAIDs and esophageal cancer is quantitatively similar to the association between NSAIDs and colorectal cancer seen elsewhere in epidemiological, clinical, and experimental studies.4 The proportionate reduction in risk is also similar for esophageal adenocarcinoma (OR, 0.67; 95% CI, 0.51–0.87) and squamous cell carcinoma (OR, 0.58; 95% CI, 0.43–0.78) and does not differ statistically between patients who used aspirin (OR, 0.50; 95% CI, 0.38–0.66) and those who use nonaspirin NSAIDs (OR, 0.75; 95% CI, 0.54–1.0). At the time that these studies were conducted, nonaspirin NSAIDs included a wide variety of traditional (nonselective) NSAIDs but not selective COX-2 inhibitors. The analysis by Corley et al.3 found no evidence of heterogeneity that would preclude combining studies of different design (prospective and case-control) and different sources of information about NSAIDs (questionnaires and pharmacy databases) in the meta-analysis.
The combined analyses by “dose” found lower risk of esophageal cancer among persons who reported frequent use of NSAIDs (OR, 0.54; 95% CI, 0.43–0.67) than in those who reported intermittent use (OR, 0.82; 95% CI, 0.67–0.99). However, the information on “dose” was limited, in that most studies collected information only on the frequency of NSAID use and not on the amount taken. Furthermore, the reduction in risk associated with “intermittent” or “occasional” use was larger than would be expected if the underlying mechanism of action required sustained inhibition of COX-2 activity. Even subjects classified as “frequent” users of NSAIDs would include many people who take a single antiplatelet dose of aspirin daily, a regimen that effectively inhibits COX-1 activity in platelets, but does not sustain inhibition of COX-2 activity in nucleated cells. While it is interesting that the published epidemiological studies demonstrate a gradient of lower risk with more frequent NSAID use, these studies do not presently define the optimal dose of NSAIDs to be tested in randomized clinical trials.
The subset of studies that could adjust for multiple risk factors for esophageal cancer in addition to age, found similar associations to the analyses adjusted only for age. Thus, there is no evidence that the inverse association between NSAIDs and esophageal cancer can be attributed to other known risk factors for esophageal cancer. The possibility remains, however, that patients with symptoms of esophagitis or dysplasia may avoid the use of NSAIDs and that reverse causation may account for the inverse association in these studies because most are case-control studies or cohort studies with relatively short follow-up. Meta-analyses of observational studies do not, in any case, provide randomized proof of efficacy.
The hypothesis that NSAIDs may inhibit the multistage development of esophageal cancer relates closely to evidence that chronic inflammation, and the attendant induction of COX-2 and increased expression of inflammatory mediators, contributes to tumor development. Inflammation involves a complex of host responses that recruit leukocytes (neutrophils, monocytes, macrophages, and eosinophils), release autocrine and paracrine inflammatory mediators, and induce interactions among chemokine ligand/receptor systems.5 Reactive oxygen and nitrogen species can directly damage the genes that control cell growth. Inflammatory mediators can stimulate cell proliferation, inhibit apoptosis, induce angiogenesis, and impair certain immune responses. Collectively, these factors can promote the clonal proliferation of mutated cells.5
Paradoxically, many of the factors that mediate inflammation are beneficial in the context of acute wound healing but potentially detrimental in settings where chronic inflammation may contribute to the multistage development of certain cancers. Factors that promote the survival and clonal proliferation of damaged cells increase the probability that a particular clone of cells will acquire the requisite genetic mutations to become an invasive and metastatic cancer. A critical enzyme that is induced during wound healing and chronic inflammation is cyclooxygenase (COX)-2, the inducible isoform of cyclooxygenase and rate-limiting step in prostaglandin production. COX-2 activity is inhibited by anti-inflammatory doses of both nonselective NSAIDs and by selective COX-2 inhibitors. Increased levels of COX-2 have been measured in Barrett's esophagus and associated adenocarcinoma.6 COX-2 expression can be induced in esophageal tissue exposed, ex vivo, to bile salts and gastric acid.7 Selective inhibition of COX-2 in mucosa isolated from Barrett's esophagus in vitro inhibits cell growth.8
To our knowledge, only 3 randomized clinical trials of NSAIDs for the prevention of esophageal cancer have been initiated. One study of squamous cell esophageal carcinoma has already been completed in a population in Linxian, China, where endemic nutritional deficiencies predispose to a high risk of this cancer. This involved a randomized, placebo-controlled, factorial design trial of celecoxib, 200 mg twice daily or selenomethionine, 200 mg daily of patients participating in periodic screening.9 After 10 months of treatment, no regression in the severity of mild or moderate squamous dysplasia was observed among 90 persons treated with celecoxib. Statistically significant regression was observed in patients with mild dysplasia treated with selenium.9 Further study of selenomethionine but not celecoxib is underway.
A second, NCI sponsored, multicenter phase II trial of celecoxib for the treatment of patients with Barrett's esophagus is just beginning at Johns Hopkins (NCI protocol NCI-P00-0145). This randomized, double blind, placebo-controlled trial of 200 patients will assess whether celecoxib treatment (200 mg twice daily for 48–96 weeks) induces the regression of high-grade to low-grade dysplasia, or low-grade dysplasia to metaplasia (A. Forastiere, personal communication). A third study is a Phase 2, pilot study to assess the safety and efficacy of treatment with celecoxib, following ablation of Barrett's metaplasia. It will compare the regeneration of squamous epithelium among patients randomized to treatment with a proton pump inhibitor alone or in combination with celecoxib (C. Lightdale, personal communication).
Patients with Barrett's esophagus provide a clinical opportunity for randomized clinical trials to assess the efficacy of NSAIDs, particularly selective COX-2 inhibitors, in preventing progressive dysplasia and malignant transformation. Each year, an estimated 0.5%–1% of persons with Barrett's esophagus progress to esophageal adenocarcinoma, a rate estimated to be 30–40 times higher than the general population.10 Nevertheless, the majority of persons with symptoms of GERD do not develop persistent esophageal dysplasia or adenocarcinoma.2 The degree of dysplasia is currently the strongest predictor of risk. Endoscopic screening can be used to identify high-risk populations where clinical trials are ethical and feasible. Such patients would benefit substantially if selective COX-2 inhibitors or other NSAIDs did in fact prevent progression or induce regression of cellular atypia. Treatment of Barrett's esophagus might plausibly involve a combination of antireflux agents and NSAIDs.6
Despite their promise, anti-inflammatory drugs are not yet recommended for the prevention or treatment of any cancers. Fundamental questions remain concerning their molecular and cellular targets of action, efficacy, safety, treatment regimen, indications, and the balance of risks and benefits from treatment in designated patient populations.
References
- . SEER Program Public Use Data Tapes, 1973–1999. DCCPS, Surveillance Research Program. 2002;2002:
- . Gastroesophageal reflux, Barrett esophagus, and esophageal cancer. JAMA. 2002;287:1972–1981
- . Protective association of aspirin/NSAIDs and esophageal cancer: a systematic review and meta-analysis. Gastroenterology. 2003;124:47–56
- . Nonsteroidal anti-inflammatory drugs as anticancer agents: mechanistic, pharmacologic, and clinical issues. J Natl Cancer Inst. 2002;94:252–266
- . Chronic inflammation, mutation, and cancer. In: Parsonet J editors. Microbes and malignancy—infection as a cause of human cancers. New York: Oxford University Press; 1999;p. 35–88
- . Cyclooxygenase-2 inhibitor therapy for the prevention of esophageal adenocarcinoma in Barrett's esophagus. J Natl Cancer Inst. 2002;94:406–407
- . Cyclooxygenase expression in Barrett's esophagus and adenocarcinoma: ex vivo induction by bile salts and acid exposure. Gastroenterology. 2000;118:487–496
- The effect of selective cyclooxygenase-2 inhibition in Barrett's esophagus epithelium: an in vitro study. J Natl Cancer Inst. 2002;94:422–429
- Chemoprevention of esophageal squamous cancer: randomized, placebo-controlled trial in a high-risk population. Digestive Disease Week. 2002; Abstract 100303
- Nonsteroidal anti-inflammatory drug use, body mass index, and anthropometry in relation to genetic and flow cytometric abnormalities in Barrett's esophagus. Cancer Epidemiol Biomarkers Prev. 2002;11:745–752
☆ Address requests for reprints to: Michael J. Thun, M.D., American Cancer Society, 1599 Clifton Road, Atlanta, Georgia 30329-4251. e-mail: mthun@cancer.org; fax: (404) 327-6450.
PII: S0016-5085(03)50036-1
doi:10.1053/gast.2003.50025
© 2003 American Gastroenterological Association. Published by Elsevier Inc. All rights reserved.
Refers to article:
- Protective association of aspirin/NSAIDs and esophageal cancer: A systematic review and meta-analysis
