Symptomatic Gastroesophageal Reflux as a Risk Factor for Esophageal Adenocarcinoma

Reviewers: Kenneth Blank, MD
Source: The New England Journal of Medicine Volume 340, Number 11: March 18, 1999

Background

Adenocarcinoma of the esophagus has had the largest increase in incidencethan any other cancer since 1970. Specifically, from 1976 to 1987, theincidence of this cancer for men increased from 4% to 10% per year. Thereason for this increase remains unclear. While alcohol and cigarettesmoking are clearly associated with esophageal cancer and actsynergistically to promote malignant transformation, the rates of smokingand alcohol abuse have not changed dramatically and therefore cannotexplain the rise in incidence.

Symptoms of gastroesophageal reflux (heartburn) may be related to the onsetof esophageal cancer. Heartburn is known to play an important role in thedevelopment of Barrett?s esophagus (a pre-cancerous condition of theesophagus), but little evidence exists linking reflux symptoms directly toesophageal cancer. A March 18th report in the New England Journal ofMedicine examines the relationship between heartburn and the risk ofesophageal cancer and cancer of the gastric cardia.

Methods

All newly diagnosed cases of adenocarcinoma of the esophagus or gastriccardia and half of the newly diagnosed cases of squamous cell carcinoma ofthe esophagus occurring in Sweden between 1994 and 1997 were entered in thisstudy. A comprehensive catchment strategy was administered to ensure thatevery potential case throughout the country was identified soon afterdiagnosis. These patients were matched to a control group randomly selectedfrom persons of the same age and sex.

Tumors were carefully examined endoscopically and pathologically toclassify them properly as either gastric cardia cancers or esophagealcancers. Tumors that had their epicenters within two centimeters proximalor three centimeters distal to the gastroesophageal junction wereconsidered gastric cardia. The gastroesophageal junction was defined as thepoint where the proximal longitudinal mucosal folds begin in the stomach.

Professional interviewers using a computer assisted program obtained dataon reflux symptoms including heartburn and regurgitation. The interviewerswere not blinded to the patients diagnosis (case or control) but had noknowledge of the study hypothesis. To prevent collecting data on symptomscaused by the cancer, only symptoms that were present for at least fiveyears prior to study entry were included.

Results

618 patients and 820 controls were interviewed for the study. Eighty-fivepercent of the 618 patients participated in the study. The main reason fornonparticipation was physical or mental impediment or early death. Therisk of esophageal adenocarcinoma was eight times higher in personsreporting heartburn, regurgitation or both. The frequency of these symptomsand their severity correlated with the risk of adenocarcinoma. Persons withreflux symptoms greater than three times per week had a nearly 17 foldincrease in the risk of esophageal adenocarcinoma. Similarly, the longerthe symptoms were present, the higher the risk. Persons with reflux symptomsfor longer than twenty years were at higher risk compared to persons withsymptoms fewer than twenty years.

Reflux symptoms were associated with adenocarcinoma of the gastric cardia,but not as strongly as with esophageal adenocarcinoma. Symptoms were gradedby severity and frequency on a scale from 0-6.5, with higher scoresindicating worse and more frequent symptoms. Persons with high scores had anearly three fold increase in the risk of adenocarcinoma of the gastriccardia. Duration of symptoms also increased risk; persons with eitherheartburn or regurgitation greater than twenty years had a four foldincrease in the risk of gastric cardia adenocarcinoma.

Esophageal squamous cell carcinoma was not associated with heartburn orregurgitation regardless of the frequency, severity or duration ofsymptoms. Several confounding factors for esophageal cancer (adenocarcinomaand/or squamous carcinoma) were identified including Barrett?s esophagus,age, sex, body-mass index, smoking and alcohol use. However, none of theserisk factors significantly changed the risk estimates.

Conclusion

The authors documented a strong association between adenocarcinoma of theesophagus and reflux symptoms. A weaker association between reflux symptomsand adenocarcinoma of the gastric cardia was detected and there was noassociation between reflux symptoms and squamous cell carcinoma of theesophagus. These associations were independent of confounding variablesincluding Barrett?s esophagus, age, sex, smoking or alcohol abuse.

The clinical implications of this study are unclear. Should everyone withreflux symptoms be endoscoped to rule-out carcinoma? Since twenty percentof the adult population in the United States complains of heartburn weekly,such a policy would have a remarkably low yield and place a great burden ona medical system already strained financially. Instead, further studies arenecessary to define a group of patients with heartburn who are atsufficiently high risk of cancer that screening endoscopy is warranted.

Further studies are also necessary to determine the best treatment ofheartburn. Presently, many medications are effective at alleviating reflux symptoms including antacids, H2 blockers (cimetidine, ranitidine, famotidine, and nizatidine) andproton-pump inhibitors (such as lansoprazole and omeprazole).

However whether or not these medications will decrease the risk of canceris unknown. Clearly, heartburn is not a benign symptom and places patientsat risk for esophageal adenocarcinoma and gastric cardia adenocarcinoma.Patients with heartburn, and especially those with severe or long-standingsymptoms, should consult a physician for a complete evaluation.