Antacid Drug Use and Risk of Esophageal and Gastric Adenocarcinomas in Los Angeles County

Histamine2-receptor antagonists and proton pump inhibitors (PPI) were introduced to the market, in the late 1970s and 1980s, correspondingly, and became widely used in the treatment of acid-peptic disorders such as peptic ulcer disease, gastresophageal reflux disease (GERD), and acute stress ulcers. H₂-receptor antagonists are competitive inhibitors of histamine binding at the parietal cell H₂ receptor. Gastric acid secretion is suppressed if the parietal cell H₂ receptor is not bound with histamine in the stomach (1). PPIs suppress gastric acid secretion by irreversibly blocking the hydrogen/potassium adenosine triphosphatase enzyme system of the gastric parietal cell (2). Because PPIs target the terminal-step in acid production, they are significantly more effective than H₂-receptor antagonists in reducing gastric acid secretion. These drugs do not permanently change the acid milieu; although they raise gastric pH and provide rapid relief of symptoms; evidence of this impermanence is that erosive esophagitis recurs in most patients within 6 months after discontinuing the drug therapy (3). Therefore, chronic maintenance therapy is generally required. Although today, some doses of these drugs are available without prescription, initially all required prescription.

Some concerns have been expressed that long term gastric acid suppression, particularly by H₂-receptor antagonists, may be associated with increased risk of esophageal and gastric adenocarcinomas (4). Some epidemiologic studies, both prospective studies and retrospective studies, have attempted to evaluate associations of H₂-receptor antagonists and PPIs with esophageal and gastric adenocarcinoma risk (5-15), but the results have not been consistent. Although some investigators have reported positive associations between antacid drug use and increased risk of gastric cancer (8, 11, 14) or esophageal cancer (12, 15), others find no effect (5, 7, 9, 13). Only one study stratified participants according to whether or not they had severe GERD (7). It is important to consider the effect of GERD when examining the association between antacid use and risk of esophageal and gastric adenocarcinomas because GERD itself is a strong independent risk factor for adenocarcinomas of the esophagus and gastric cardia (5, 7, 9, 16).

We have explored the association between use of antacid drugs and risk of esophageal and gastric adenocarcinomas using data from a large population-based case-control study conducted among residents of Los Angeles County. To understand any observed association better and to assess confounding of results by indication for use of the drug, we have investigated the relationship separately in groups with a history of upper gastrointestinal (UGI) disorders and those without such a history.

The details of the study population and study design have been described elsewhere (16-21). Basically, incident cancer cases for this study were identified by the University of Southern California Cancer Surveillance Program. This cancer registry covers the ethnically diverse Los Angeles County population of >9.5 million people. Case patients who were newly diagnosed with first incident esophageal adenocarcinoma (GCA; International Classification of Disease for Oncology code C15.0-C15.9), gastric cardiac adenocarcinoma (GCA; International Classification of Disease for Oncology code C16.0), or distal gastric adenocarcinoma (DGA; International Classification of Disease for Oncology codes C16.1-C16.6 and C16.8-C16.9) between 1992 and 1997 were identified and contacted for participation. Control participants were matched individually to each case patient on sex, race, age (±5 y), and neighborhood of residence. To increase the statistical power, we sought two control participants for each case patient whenever possible. Of the 938 case patients interviewed, 523 had one control participants, 381 had 2 or more control participants, and 34 had no eligible control participants identified for an overall total of 1,356 control participants. An experienced interviewer conducted the in-person interviews at participants' homes or other convenient locations. Next-of-kin were interviewed when case patients were unable to be interviewed due to death or illness. Although it was not feasible to blind the interviewers to case (or next-of-kin) or control status, interviewers and study participants were not aware of the study hypotheses. Written informed consent was obtained from each study participant before interview. A total of 938 case patients (220 incident EA/277 GCA/441 DGA; numbers for the three tumor sites are shown separately with these abbreviations) and 1,356 control participants are included in the current study. Next-of-kin interviews accounted for 269 of the 938 interviews with case patients (65 EA/85 GCA/119 DGA).

We developed a structured questionnaire specifically for this study that was administered during the in-person interview. The questionnaires obtained data up to a specified reference date, defined as the date that was 1 y before the date of diagnosis of the case patient; this same reference date was used for each case patient's matched control participants. The interview queried general background information such as ethnicity, marital status, birthplace, highest level of education, smoking history, lifetime use of all types of alcoholic beverages, usual diet, weight at ages 20 and 40 y and on the reference date, and height. In addition, we asked detailed questions regarding personal and family history of various nonmalignant diseases.

We explicitly asked about 26 over-the-counter (OTC) and 19 prescription brand name antacid and related drugs in the questionnaire. For each of the listed medications, we first asked the participants whether they had ever used the drug before their reference date. If the answer was “no”, the participant was classified as a nonuser. If the answer was “yes”, the participant was asked if he or she had ever taken the drug two or more times a week for one month or longer. If the answer was no, the participant was classified as an “irregular user”. Otherwise, the participant was defined as a “regular user” and was asked further about ages at first and last use, duration of use, usual frequency and dosage of use, and the primary reason for each use. We also asked the participants if they had used any medications that were not on our list and recorded the drug name and details of use if the participant had used the medication “regularly”. All of the antacid medications in the study were categorized as requiring a prescription or as nonprescription (OTC). At the time the study was conducted, nonprescription (OTC) antacid drugs were limited to acid neutralizing agents (e.g., Tums, Pepto-Bismol, and Maalox); the majority of prescription antacids at that time were H₂-receptor antagonists. Only a very small number of participants reported use of the PPI, which also required a prescription.

To assess medical history, we provided a list of diseases, including heart disease, gastrointestinal tract disorders, and diabetes, and asked if the participant had any of those conditions diagnosed by a physician before the reference date. UGI-related conditions that we queried included gastric ulcer, duodenal ulcer, gastritis, hiatal or diaphragmatic hernia, esophagitis, Barrett's esophagus, gastresophageal reflux disease, excess acid, or gastric hyperacidity. If the response was yes to any of the conditions, the participant was then asked the age at first diagnosis with the condition. In addition, participants were asked if they had experienced various symptoms including gas pain (in the stomach), sour stomach including acid indigestion and regurgitation, heartburn after meals, and trouble swallowing (feeling that solid food was sticking to their throats as it went down) before their reference date. If the response was positive, the frequency (daily, weekly, monthly, less than monthly) with which each symptom occurred and the age that the participant first experienced the symptom on a regular basis were recorded.

Participants were grouped as never users, irregular users, or regular users of nonprescription acid neutralizing agents and of prescription acid suppressive drugs, separately. Among participants who reported regular use, duration of use variables were created for nonprescription acid neutralizing agents and for prescription acid suppressive drugs by summing all durations of use of the same class of medication (prescription or nonprescription) for each person (All antacid drugs with any reported use are listed in Appendix A). Categories for duration of regular antacid use (prescription or nonprescription) were <1 y, 1 to 3 y, and >3 y of use. Because early symptoms of an UGI cancer may influence the use of antacid drugs, we have excluded from the calculation of total duration of use any drugs that were first taken in the year before a case or control participant's reference date (for case patients, this therefore excludes drug use that was initiated in the 2 y before the date of diagnosis). Additional analyses were conducted extending this exclusionary period to 3 y and to 5 y, respectively.

Polychotomous logistic regression was used to compute the odds ratios (OR), as estimates of the relative risk, and corresponding 95% confidence intervals (CI) for adenocarcinomas of the esophagus, gastric cardia, and distal stomach simultaneously in relation to use of nonprescription acid neutralizing agents and to use of prescription acid suppressive drugs. In a previous study (19), this approach provided more precise estimates of the ORs, and the magnitude of the ORs was consistent with those obtained in separate conditional logistic regression analyses that preserved the original case-control match within each cancer site. Therefore, for the purpose of maximizing statistical power, we report results based on unconditional polychotomous logistic regression models with adjustments made for the matching variables in this analysis; this allowed us to use all control participants in a single analysis.

Forward stepwise logistic regression was used to select confounders from a list of variables: age (≤39, 40-49, 50-59, 60-69, 70+ y), sex (male/female), race (Non-Latino White versus other), birthplace (US born, non-US born), smoking status (never smoker, ex-smoker, current smoker), body mass index at reference age (<25, 25-29.9, 30+ kg/m²), history (no/yes) of UGI disorders diagnosed by physician (including gastric ulcer, duodenal ulcer, unspecified type of ulcer, gastritis, hiatal hernia, esophagitis, Barrett's esophagus, gastresophageal reflux disease, excess acid or gastric hyperacidity, and other diseases of the stomach). A test for trend across ordinal categories of duration of antacid use was done for each type of cancer.

We constructed a 1 degree of freedom likelihood ratio test to assess homogeneity of trends in duration of nonprescription acid neutralizing agent use among individuals who had a history of physician-diagnosed UGI disorders and those who did not. With adjustment for prescription acid suppressive drug use and other confounding factors, we conducted the analyses separately for each type of cancer using unconditional logistic regression comparing a multivariable model that fit two trend variables (one for each category of UGI disorder history) with a multivariable model that fit a single trend variable for all participants. The P values reported for trend tests and for the test for homogeneity of trends are two sided.

For validity purposes, we repeated all statistical analyses excluding data collected from next-of-kin participants. Risk estimates were not materially different from the results based on all participants combined (i.e., self-respondents and next-of-kin participants).

Demographic characteristics of the study sample have been described previously (17). Briefly, the mean ages of cases at the reference date (1 year before diagnosis) were 60.0 years (SD, 9.4) for patients with EA, 59.8 years (SD, 10.2) for patients with GCA, and 59.4 years (SD, 11.5) for DGA patients; control participants were, on average, 58.7 years (SD, 11.5) at their reference dates. Among control participants, 74% were male, compared with 91% of patients with EA, 83% of patients with GCA, and 59% of the DGA patients. Non-Latino Whites represented 77% of those with EA, 76% of those with GCA, and 30% of those with DGA. For all three tumor sites, case patients tended to have lower education and a higher percentage were current smokers than control participants. Forty percent of case patients reported a history of UGI disorders diagnosed by physicians, compared with 27% of control participants. These proportions became larger when self-reported UGI symptoms were considered. Participants who reported UGI symptoms accounted for 56% of case patients and 40% of control participants in this study sample (Table 1).

Regular use of nonprescription acid neutralizing agents for 1 to 3 years was associated with increased risk of EA (adjusted OR, 3.16; 95% CI, 1.39-7.20) after adjusting for prescription acid suppressive drug use and other confounding factors (Table 2). Among those who took nonprescription acid neutralizing agents for >3 years, the OR for EA increased to 6.32 (95% CI, 3.14-12.69; P_trend<0.01). With duration of nonprescription acid neutralizing agent use in the statistical model, use of prescription acid suppressive drugs was unrelated to risk of EA. No strong effect on risk of cardiac or DGAs was observed for duration of use of either prescription acid suppressive drugs or nonprescription acid neutralizing agents. Conversely, irregular use of nonprescription acid neutralizing agents was associated with lower risk of GCA (adjusted OR, 0.47; 95% CI, 0.30-0.72) and DGAs (adjusted OR, 0.48; 95% CI, 0.34-0.68).

Because prior diagnosis of UGI disorders was also strongly associated with increased risk of adenocarcinomas for esophagus and gastric cardia (16, 17), and because presence of UGI disorders might influence frequency of antacid drug use, we conducted analyses separately among participants with and without a history of physician-diagnosed UGI disorders (Table 3). After stratification of UGI disorders, we observed a modest difference in the dose-response effects for nonprescription drug use between persons with UGI disorders and those without UGI disorders (homogeneity of trends P = 0.07), although both trends for duration of use were statistically significant (P_trend <0.01). The increase in EA risk among long-term (>3 year) users of nonprescription acid neutralizing agents was limited to participants without a history of physician-diagnosed UGI disorders (adjusted OR, 10.89; 95% CI, 4.73-25.05). Although the adjusted ORs for 1 to 3 years and >3 years of nonprescription antacid use among participants with a positive history of physician-diagnosed UGI disorders were also >1, the CIs were wide and did not exclude 1.0 (adjusted OR, 2.01; 95% CI, 0.35-11.40 and adjusted OR, 3.19; 95% CI, 0.62-16.48). No significant increases in risk or dose-response effects were observed for adenocarcinomas of the gastric cardia or distal stomach. Irregular use of nonprescription acid neutralizing agents was associated with reduced ORs for GCA and DGAs, but the risk reductions were statistically significant for persons who had no history of physician-diagnosed UGI disorders.

All of the results presented in the current study exclude exposures that were initiated during the 1 year interval before each participant's reference date. We also evaluated exposure using a longer lag time, excluding drugs that were first used 2 years before the reference date (3 years before diagnosis for a case) and 5 years before the reference date. The results of the 1, 2, and 5 year lags did not differ in any meaningful way (results for the 2 and 5 year lag times not shown). Similarly, restricting the analyses to self-reported data only, that is, excluding data collected through next-of-kin interviews, did not differ from those presented, which used all participant data, although among those who participated in the interview directly, the ORs increased slightly when next-of-kin interviews were excluded (data not shown).

In this large population-based case-control study, risk of EA increased with increasing duration of regular use of OTC acid neutralizing agents but not with increasing duration of regular use of prescription acid suppressive drugs. No association with either type of drug was observed for adenocarcinomas of the gastric cardia or distal stomach. The substantially lower risks of GCA and DGAs observed among persons who used acid neutralizing agents irregularly was not explained by the number of such drugs that they used irregularly (data not shown).

Our findings of increased EA risk among regular users of nonprescription acid neutralizing agents and reduced risk of adenocarcinomas of the gastric cardia and distal stomach among irregular users of OTC acid neutralizing agents were restricted to persons who did not report any physician-diagnosed UGI disorders. None of the tests for homogeneity of trends in duration of regular use comparing those with and without UGI disorders were statistically significant; these analyses excluded persons with irregular use. We did not observe that persons with UGI disorders had greater risk of EA with increasing duration of risk, which would have suggested that our observed association reflected confounding by indication. Nevertheless, it is likely that persons who did not have a physician-diagnosed UGI disorder were taking nonprescription acid neutralizing agents for various UGI symptoms.

In the absence of acid neutralization or acid suppression, the degree of esophageal damage has been found to correlate with the degree of esophageal acid exposure through gastro-esophageal reflux (GERD) with esophageal 24-hour pH monitoring in humans (22, 23). However, gastric acid is not the only substance that refluxes from the stomach into the esophagus. Animal studies have suggested that exposure to bile acid, a major component of duodenal juice that originates in duodenum and may reflux into the esophagus through duodenogastro-esophageal reflux (DGER), increases the risk of EA (24-26). In a rat model, duodenal juice refluxed into esophagus although DGER induced histologic changes, from squamous-cell epithelium into columnar-cell epithelium, in the lining of the distal esophagus (25). Moreover, another animal study showed that the prevalence of EA increased as the amount of acid gastric juice that was permitted to reflux with duodenal juice into the esophagus decreased (26), suggesting that acid suppression therapy may be detrimental by encouraging esophageal metaplasia and tumorigenesis in participants with DGER reflux. In humans, it has been shown that the degree of esophageal damage increases with increasing amount of DGER, with the highest DGER levels found in patients with Barrett's esophagus (27). With 24-hour ambulatory esophageal pH and bilirubin and gastric pH monitoring, Marshall et al. (28) have observed that most DGER takes place at night when the pH is between 4 and 7. Based on previous laboratory findings, Todd et al. (27) concluded that at a pH of 6 or greater, deconjugated bile acids, if refluxed into the esophagus, cause not only short-term damage to the esophageal mucosa but also long-term damage to the cellular DNA.

Although acid suppressive drugs in the form of H₂-receptor antagonists or PPIs allow symptom relief and promote endoscopic mucosa healing in GERD patients, they do not guarantee normalization of pH level at the gastro-esophageal junction nor do they prevent GERD or DGER (29). There is some disagreement about how well OTC acid neutralizing agents work in changing gastric pH. On the one hand, OTC acid neutralizing agents are believed to cause an immediate short-term increase in the pH, which differs from the long-lasting and profound effect of H₂-receptor antagonists. On the other hand, OTC acid neutralizing agents are taken at the time of maximum symptoms when reflux is probably most intense in the patients, and OTC drugs may be taken at relatively large doses to provide the most rapid relief (30, 31), causing the gastric pH temporarily to increase to levels that are much higher than those produced by H₂-receptor antagonists (32). According to the first hypothesis, the observed increase in risk of EA among regular users of long-term OTC acid neutralizing agents rather than among H₂-receptor antagonist users in this study would be explained by the fact that OTC acid neutralizing agents provide less successful acid suppression than H₂-receptor antagonists. Alternatively, if the second hypothesis is true, the observed association suggests that OTC acid neutralizing agents create a more alkaline condition in the stomach than H₂-receptor antagonists at the time of maximum symptoms and reflux, and therefore more bile reflux may reach the distal esophagus. As we discussed above, the direct contact of alkaline duodenal bile that enters the lower esophagus can cause cell turnover and is potentially carcinogenic (33); furthermore, it may promote progression of Barrett's esophagus (29, 34).

Although a number of epidemiologic studies have investigated the possible positive association between the use of H₂-receptor antagonists and the risk of gastric adenocarcinomas (7, 8, 10-12, 14), research to determine whether the risk of EA is increased among users of H₂-receptor antagonists or other antacids has been limited. In a recent case-control study, which was nested within a large prospective study in the United Kingdom, H₂-recepor antagonist and PPI users were both at increased risk of EA; the association was limited to long-term current users of H₂-recepor antagonists in the multivariable analysis that included variables for GERD, hiatal hernia, peptic ulcer, and dyspepsia in the model (12). This study was, however, based on a small number of cases and did not consider the use of OTC acid neutralizing agents or acid suppressive drugs that were available without prescription during the study period. In Sweden, Lagergren et al. (9) reported a nearly 3-fold higher risk of EA among persons who used medications for reflux symptoms at least 5 years before study than among those who took no medication. In a U.S. study, Chow et al. (5) observed a 4-fold higher risk of adenocarcinomas of the esophagus and gastric cardia in individuals who had four or more prescriptions for H₂-receptor antagonists, but the association disappeared after adjusting for the predisposing conditions. One of the drawbacks of that study, as well as other earlier studies (6, 15), is that it did not separate EAs from adenocarcinomas of the gastric cardia. Farrow and her colleagues (7) investigated the separate associations between antacid use and adenocarcinomas of the esophagus, gastric cardia, and distal stomach. In this study, long-term use of OTC antacids was associated with increased risk of EA, but no statistically significant risk increase was observed for H₂-receptor antagonist users, which is consistent with the results we report here.

Our study represents one of two large population-based epidemiologic studies within the United States that has been designed specifically to investigate the etiologies of these tumor types. One strength of this study is the definition of the reference group. Instead of combining never users with irregular users as many other studies have done, our reference group is composed of persons who have never used the antacids of interests. In addition, rather than simply adjusting for GERD and other UGI symptoms in the analyses as most of other studies have done, we have evaluated the risk estimates for antacid use separately among participants with and without a history of physician-diagnosed UGI disorders. Furthermore, our study allows comparison of risk estimates between OTC acid neutralizing agents and acid suppressive drugs; the former group has been rarely studied separately.

Of interest is the reduced risk of DGA in individuals who irregularly used OTC acid–neutralizing agents. After carefully studying the reported OTC acid–neutralizing agents in this group used irregularly, we noticed that one of the most frequently used OTC acid–neutralizing agents was Pepto-Bismol, which is used to protect the stomach lining from acid, but also kills Helicobacter pylori (35). Our analyses comparing irregular users of Pepto-Bismol to persons with no OTC antacid use found a reduction in risk of distal gastric cancer that was similar to that observed for other OTC antacid preparations (data not shown).

Limitations of this study include the potential for recall bias, a modest response rate, and the potential for less accurate reporting of exposures by next-of-kin; these have been described elsewhere (17). Next-of-kin interviews accounted for 29% of the patient reports in this study. Our results did not seem to be affected by next-of-kin interviews and were similar when next-of-kin interviews were excluded from the analyses. Our reference group had limited numbers for some analyses because of our restricted definition for the reference group, excluding irregular users; in fact, we only had two patients with EA with a history of UGI disorders who reported no antacid use. This limited the statistical power in our stratified analyses. In addition, we have few participants who reported use of PPIs because the study was conducted before widespread use of these drugs. Further research is required to fully assess the effect of H₂-receptor antagonists and PPIs, which are now available as OTC drugs and to define the mechanisms by which the common OTC drugs used by our participants affect risk of EA.

In summary, this study observed increases in risk of EA in persons who used nonprescription acid neutralizing agents regularly. This may be due to confounding by indication with the conditions for which these drugs are used etiologically related to disease risk, or the increase in risk associated with OTC drugs may be the result of alkaline reflux resulting from the acid neutralization entering the esophagus.

No potential conflicts of interest were disclosed.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank all the study participants for their contributions and Annie Fung, Isaura Rivera, Timothy Stirton, Chiu-Chen Tseng, and June Yashiki for their help with data collection and data management.