SEARCH

SEARCH BY CITATION

Keywords:

  • hiatal hernia;
  • reflux symptoms;
  • body size;
  • esophageal and gastric cardia adenocarcinoma

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

BACKGROUND

Since the 1970s, incidence rates of esophageal and gastric cardia adenocarcinoma have risen substantially. Reasons for the increasing trends are not well understood.

METHODS

A population-based, case–control study that included esophageal adenocarcinomas (n = 222), gastric cardia adenocarcinomas (n = 277), distal gastric adenocarcinomas (n = 443), and 1356 controls was conducted in Los Angeles County. Unconditional logistic regression was used to calculate odds ratios (ORs) and corresponding 95% confidence intervals (CIs) for the 3 tumor types.

RESULTS

After adjustment for demographic factors, smoking, and body size, both hiatal hernia and reflux symptoms emerged as significant independent risk factors. Risk of esophageal adenocarcinoma was increased 3-fold (adjusted OR, 3.61; 95% CI, 2.49–5.25) among those who had reflux symptoms but did not have hiatal hernia, 6-fold (adjusted OR, 5.85; 95% CI, 3.18–10.75) among those who had hiatal hernia but did not have reflux symptoms, and 8-fold (adjusted OR, 8.11; 95% CI, 4.75–13.87) among those who had both reflux symptoms and hiatal hernia. A similar risk pattern was found in relation to history of hiatal hernia and other reflux conditions. A more modest but still significant risk pattern was observed for gastric cardia adenocarcinoma. Among control subjects, there was a significant and positive association between increasing body mass index and history of hiatal hernia and/or reflux symptoms.

CONCLUSIONS

Hiatal hernia, in combination with other reflux conditions and symptoms, was associated strongly with the risk of esophageal adenocarcinoma. These associations were more modest for gastric cardia adenocarcinomas. A significant and positive association between body size and history of hiatal hernia/reflux symptoms also was observed. Cancer 2003;98:940–8. © 2003 American Cancer Society.

DOI 10.1002/cncr.11568

Reasons for the increasing incidence trends of esophageal and gastric cardia adenocarcinoma in westernized countries are not well understood. The most consistently found risk factors for esophageal adenocarcinoma are high body mass index (BMI),1–5 symptoms of gastroesophageal reflux disease (GERD),6–8 and conditions including hiatal hernia and esophagitis.7, 9 Although a high BMI10 and history of GERD, hiatal hernia, and esophagitis9 were also significant risk factors for junctional adenocarcinoma (i.e., adenocarcinoma of the esophagus and gastric cardia combined), the influence of these factors on risk of gastric cardia adenocarcinomas alone is less consistent. A high BMI was associated significantly and positively with risk of gastric cardia adenocarcinoma in some,2, 3, 11 but not in other,1, 5 studies.Reflux symptoms were a significant risk factor for gastric cardia adenocarcinoma in a Swedish study,6 but neither reflux symptoms nor reflux conditions (hiatal hernia, esophagitis) were associated with the risk of gastric cardia adenocarcinoma in a U.S. multicenter study.7 In addition, the relation between BMI and GERD is controversial.12–14 Although a positive association between body size and reflux diseases (hiatal hernia, esophagitis) has been reported in some studies,15–20 no relation between BMI and symptoms of GERD was reported in a Swedish population-based study.14 In short-term weight reduction studies, improvement in reflux symptoms has been reported in some,21 but not in other,22, 23 studies.

During the 1990s, we conducted a large population-based case–control study of esophageal and gastric adenocarcinomas in Los Angeles County.3 We asked extensive questions regarding GERD symptoms and conditions of the esophagus and stomach with onset before the diagnosis of malignant disease. In the current article, we report our findings on the role of GERD symptoms and esophageal and gastrointestinal conditions and the interrelation between BMI and history of previous esophageal conditions and symptoms in the development of esophageal and gastric cardia adenocarcinomas.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

As described previously,3 patients included as cases in the current case–control study were men and women (age range, 30–74 years) with histologically confirmed esophageal, gastric cardia, or distal gastric adenocarcinoma and no history of malignancies at these sites. We included patients with esophageal adenocarcinoma (International Classification of Diseases for Oncology [ICOD] C15.0–C15.9) and gastric cardia adenocarcinomas (ICOD C16.0) who were diagnosed between 1992 and 1997 and patients with distal gastric carcinomas (i.e., including those diagnosed in the antrum/pylorus [ICOD C16.3, C16.4], fundus/body [ICOD C16.1, C16.2], lesser and greater curvature [ICOD C16.5, C16.6], mixed subsite [ICOD C16.8], and subsite not otherwise specified [ICOD C16.9]) who were diagnosed between 1992 and 1994. All incident malignancies were identified by the Los Angeles County Cancer Surveillance Program, a population-based tumor registry.

We conducted in-person interviews with patients, or with their next of kin (NOK) when patients were unable to be interviewed due to death or illness. Of the 1716 eligible patients identified, interviews were completed with 947 case patients (i.e., 55% [947 of 1716] of those identified and 77% [947 of 1716−310−174] of those approached). We were unable to obtain interviews for 769 patients: 310 had died or were too ill to be interviewed and did not have a NOK available for interview; physicians denied permission to contact 174 patients; 146 patients could not be located; and 139 patients did not wish to participate. We had incomplete information for 5 patients and include 942 patients in the analyses presented. Age, gender, and race distributions did not differ between case patients we interviewed and those we did not interview.

Control subjects were matched individually to interviewed case patients on gender, race, and date of birth (± 5 years). Whenever possible, we sought two control subjects for cases diagnosed with esophageal or gastric cardia adenocarcinoma to increase the statistical power of the study. Control subjects must not have had a diagnosis of stomach or esophageal carcinoma. A neighborhood control subject was sought by use of a systematic algorithm based on the address of the case patient. If the first eligible matched control subject refused to participate, the second eligible one in the sequence was invited, and so on. Of the 947 case patients interviewed, 528 had 1 participating control subject, 382 had 2 or more participating control subjects, and 37 had no eligible control subject identified. For more than half of the case patients (n = 521), the first eligible matched control subject was interviewed. The second or a later matched control was interviewed for the remaining cases.

The study was approved by the institutional review board of the Keck School of Medicine of the University of Southern California. Written informed consent was obtained from each study participant before interview.

Interview Method

Cases and their matching controls were interviewed by the same interviewer in almost all instances. NOK interviews accounted for 271 of the 942 interviews with case patients (66 for esophageal adenocarcinoma, 85 for gastric cardia, and 120 for distal gastric carcinoma patients). Although it was not feasible to blind the interviewers to case (or NOK) or control status, interviewers (and study participants) were not aware of the study hypotheses. A reference date was defined as 1 year before the date of diagnosis of the case patient. This same reference date was used for each patient's matched control subject.

We used a structured questionnaire that we developed specifically for the current study. It included questions on lifetime smoking habits, lifetime use of all types of alcoholic beverages, weight at 20 and 40 years and on the reference date (referred to as current weight), and height. We asked respondents if, before their reference date, they had been diagnosed by a physician with any of a series of conditions of the esophagus and gastrointestinal tract (gastric ulcer, duodenal ulcer, gastritis, hiatal hernia, esophagitis, Barrett esophagus, gastroesophageal [GE] reflux, conditions related to excess acid or gastric hyperacidity [referred to as excess acid], polyps in the colon or large intestine, ulcerative colitis, Crohn disease) or with other conditions not known to be related to risk of esophageal or stomach malignancy (high blood pressure, heart disease, diabetes, rheumatoid arthritis, hepatitis). If the response was yes to any of the conditions, subjects were then asked their 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 the symptom occurred and the age that the subject first experienced the symptom on a regular basis were recorded.

Statistical Analysis

Conditional and unconditional logistic regression methods were used to calculate odds ratios (ORs), as estimates of the relative risk, and corresponding 95% confidence intervals (CIs) for each tumor type of interest (esophageal, gastric cardia, and distal gastric adenocarcinomas) in relation to exposures of interest.3 We report only the results obtained using unconditional logistic regression as the results obtained using the two methods yielded similar results. All unconditional models included age (≤ 39, 40–49, 50–59, 60–69, or ≥ 70), gender (male or female), race/ethnicity (white, African American, Latino-American, or Asian American), birthplace (in the United States or outside the United States), education (less than high school, high school, some college, or college graduate or more), smoking (never smoked, formerly smoked, or currently smoked), and current BMI (weight in kilograms divided by the square of height in meters; categorized in quartiles based on the gender-specific distributions among control subjects) as covariates. We also repeated our analyses restricted to case patients who were self-respondents (n = 671). We show results separately for the three tumor types.

In our analysis regarding previous physician-diagnosed conditions of the esophagus and stomach, we excluded conditions that were first diagnosed within 2 years of the reference date. Because the 5 conditions of the esophagus (i.e., hiatal hernia, esophagitis, Barrett esophagus, excess acid, and GE reflux) are interrelated, we created a composite index of reflux conditions to include diagnoses of any of these 5 conditions 2 or more years before the subject's reference date. We recognize that Barrett esophagus is generally considered to be a premalignant lesion for esophageal adenocarcinoma24 but decided that it is an important component of the composite index. Age at diagnosis of the reflux condition was based on the youngest age at first diagnosis of any of these five conditions. Self-reported symptoms that first started within 2 years of the reference date were not included as a positive report for that symptom. A reflux symptom index was constructed to incorporate both the frequency and the duration of self-reported sour stomach/regurgitation and heartburn. Our scoring scheme is as follows: no points were given to symptoms that occurred less than weekly or had started within 2 years of the subject's reference date; 1 point was assigned to symptoms that initially occurred 3–15 years before the reference date and on a weekly basis; 1.5 points were assigned to symptoms that initially occurred 16+ years before the reference date and on a weekly basis; 2.0 points were assigned to symptoms that initially occurred 3–15 years before the reference date and on a daily basis; and 2.5 points were assigned to symptoms that initially occurred 16+ years before the reference date and on a daily basis. Therefore, a reflux symptom index of 5 (highest) was assigned to subjects who reported daily symptoms of both sour stomach/regurgitation and heartburn (2 symptoms) that initially occurred 16+ years before the reference date (Table 1). To investigate whether a high BMI increases the risk of hiatal hernia or reflux symptoms in a population of individuals who do not have malignant disease, control subjects who were diagnosed with hiatal hernia or had reflux symptoms (i.e., a reflux symptom index of 1 or higher) were compared with control patients who were not diagnosed with hiatal hernia and had no reflux symptoms (i.e., a reflux symptom index of 0). Tests for trend were computed by fitting an unconditional logistic regression model to ordinal values representing levels of exposure. All reported trend test significance levels (P values) are two-sided.

Table 1. Reflux Symptom Scores for Self-Reported Sour Stomach (Acid Indigestion/Regurgitation) and Heartburn
ScoreDescription
0Symptoms less than weekly, first occurred within 2 yrs of reference date
1Symptoms at least weekly but not daily, first occurred 3–15 yrs before reference date
1.5Symptoms at least weekly but not daily, first occurred ≥ 16 yrs before reference date
2.0Symptoms daily, first occurred 3–15 yrs before reference date
2.5Symptoms daily, first occurred ≥ 16 yrs before reference date

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The current analysis included 222 patients with esophageal adenocarcinoma, 277 with gastric cardia, and 443 with distal gastric carcinoma; these patients were compared with 1356 control subjects. Males accounted for 91% of esophageal adenocarcinoma cases, 83% of gastric cardia cases, and 59% of distal gastric carcinoma cases. Whites accounted for 78% of esophageal adenocarcinoma cases, 76% of gastric cardia cases, and 30% of distal gastric carcinoma cases. After adjustment for age, gender, race/ethnicity, birthplace, and education, current cigarette smoking was a significant risk factor for all tumor types. Alcohol use was not a risk factor for any of the tumor types. Risks of esophageal and gastric cardia adenocarcinomas also increased statistically significantly in a dose-dependent manner with increasing BMI as measured at ages 20 and 40 years and as measured recently. Details regarding these demographic and lifestyle characteristics were reported previously.3

Table 2 presents the risks of the three tumor types in relation to previous diseases of the esophagus and stomach that were diagnosed by a physician. All five conditions related to the esophagus (Barrett esophagus, hiatal hernia, GE reflux, esophagitis, and excess acid) emerged as risk factors for esophageal adenocarcinoma. Results were statistically significant for all but esophagitis. All five conditions of the esophagus also were more common among patients with gastric cardia adenocarcinoma than among controls. The associations with hiatal hernia and Barrett esophagus were statistically significant. In contrast, a history of these five conditions was unrelated to distal gastric carcinoma risk. A history of any of these 5 conditions of the esophagus (referred to as any reflux conditions) was associated with a significant 3-fold increased risk of esophageal adenocarcinoma (adjusted OR, 3.35; 95% CI, 2.35–4.78) and a 2-fold increased risk of gastric cardia adenocarcinoma (adjusted OR, 1.90; 95% CI, 1.34–2.69) (Table 2). A history of gastric ulcer was more common among patients with these three tumor types than among controls. This difference was statistically significant among patients with distal gastric carcinoma. A history of duodenal ulcer, conditions related to the large intestine (ulcerative colitis, colorectal polyps), and other conditions (hypertension, heart disease, diabetes, rheumatoid arthritis, hepatitis) did not differ significantly between patients with these tumor types and controls (Table 2).

Table 2. Adjusted Odds Ratiosa and Corresponding 95% Confidence Intervals (in Parentheses) for Esophageal, Gastric Cardia, and Distal Gastric Adenocarcinomas in Relation to History of Select Medical Conditions Diagnosed by a Physician
Conditionb beforeOdds ratio (95% confidence interval)Prevalence of condition in 1355 controls (%)
Esophageal adenocarcinoma (n = 222)Gastric cardia adenocarcinoma (n = 277)Distal gastric adenocarcinoma (n = 443)
  • NA; not available.

  • a

    The multivariate model included categoric variables for age (≤ 39 years, 40–49 years, 50–59 years, 60–69 years, 70+ years), gender, race (white, African American, Latino American, Asian American), birthplace (US-born, non-US-born), education (less than high school, high school, some college, college graduate or more), smoking (never-former, current), and body mass index (in quartiles). Quartile cut points for body mass index (kg/m2) are ≤ 23, > 23–25, > 25, ≤ 28, > 28 in males, and ≤ 22, > 22–25, >25–≤ 28.25, > 28.25 in females.

  • b

    The reference category for all estimates in this table is no history of the condition of interest. Reference date was established for each participant as 1 year before the date of diagnosis of a case patient. This same date was used for each control subject who was recruited to match a case patient.

  • c

    Years before reference date.

  • d

    All diagnoses of gastroesophageal reflux were within 3–15 years of the reference date.

  • e Included any diagnosis of hiatal hernia, esophagitis, Barrett esophagus, gastroesophageal reflux, or excess acid by a physician.

  • f

    Not available because of no positive history in case group.

Gastric ulcer (yrs)c1.87 (0.99–3.55)1.72 (0.83–3.16)2.16 (1.30–3.58)3.5
 3–151.74 (0.59–5.12)2.82 (1.23–6.48)3.11 (1.51–6.41)1.4
 16+1.95 (0.89–4.25)1.05 (0.42–2.62)1.54 (0.76–3.13)2.1
Duodenal ulcer0.91 (0.46–1.81)1.52 (0.89–2.62)1.53 (0.88–2.67)4.9
Gastritis0.79 (0.40–1.55)1.22 (0.71–2.08)1.04 (0.64–1.68)6.3
Hiatal hernia (yrs)c4.85 (3.21–7.33)2.26 (1.47–3.45)1.25 (0.73–2.12)6.7
 3–154.97 (2.90–8.53)1.50 (0.79–2.83)1.14 (0.56–2.32)3.8
 16+4.72 (2.71–8.25)3.15 (1.83–5.42)1.39 (0.65–2.96)2.9
Esophagitis1.84 (0.67–5.07)1.75 (0.66–4.64)1.92 (0.66–5.56)1.2
Barrett esophagus101.2 (12.6–813)18.3 (1.96–171)1.58 (0.08–29.6)0.07
Gastroesophageal refluxd2.74 (1.42–5.30)1.26 (0.59–2.70)0.16 (0.02–1.17)2.6
Excess acid (yrs)c3.08 (1.91–4.98)1.61 (0.96–2.68)1.55 (0.94–2.56)5.3
 3–152.28 (1.13–4.61)1.08 (0.49–2.37)1.37 (0.70–2.69)2.9
 16+3.99 (2.13–7.48)2.24 (1.15–4.36)1.79 (0.87–3.68)2.4
Any reflux (yrs)c3.35 (2.35–4.78)1.90 (1.34–2.69)1.36 (0.93–1.98)12.4
 3–153.18 (2.00–5.07)1.34 (0.81–2.22)1.39 (0.86–2.25)6.8
 16+3.53 (2.22–5.61)2.55 (1.64–3.96)1.32 (0.76–2.27)5.6
Ulcerative colitis0.39 (0.05–3.17)NAf2.15 (0.74–6.25)1.0
Colorectal polyps0.84 (0.46–1.55)0.68 (0.37–1.25)0.64 (0.34–1.20)6.2
Hypertension (yrs)c1.36 (0.98–1.90)1.32 (0.98–1.78)1.00 (0.76–1.33)28.8
 3–151.50 (1.02–2.22)1.31 (0.92–1.88)1.07 (0.77–1.49)16.4
 16+1.18 (0.75–1.87)1.33 (0.89–1.98)0.90 (0.60–1.34)12.4
Heart disease1.25 (0.86–1.82)1.13 (0.79–1.61)1.11 (0.78–1.57)17.1
Diabetes1.46 (0.88–2.40)0.85 (0.50–1.47)1.41 (0.93–2.13)7.3
Rheumatoid arthritis1.08 (0.51–2.40)0.61 (0.28–1.31)0.80 (0.45–1.44)4.5
Hepatitis1.34 (0.70–2.58)0.77 (0.37–1.60)0.31 (0.11–0.89)4.1

To determine whether previous reflux conditions and gastric ulcer may be early manifestations of malignant disease, we investigated the time interval between the diagnosis of these previous conditions and the reference date established for each participant. The adjusted ORs for esophageal adenocarcinoma associated with initial diagnosis of any reflux conditions 3–15 years and 16+ years before the reference date were 3.18 (95% CI, 2.00–5.07) and 3.53 (95% CI, 2.22–5.61), respectively. The corresponding risks for gastric cardia adenocarcinoma were 1.34 (95% CI, 0.81–2.22) and 2.55 (95% CI, 1.64–3.96), respectively. Results were strengthened somewhat by restriction of the analysis to self-respondent cases. For esophageal adenocarcinoma, the adjusted ORs associated with an initial diagnosis of any reflux conditions 3–15 years and 16+ years before the reference date were 3.49 (95% CI, 2.07–5.89) and 3.84 (95% CI, 2.30–6.40), respectively. The corresponding risks for gastric cardia adenocarcinoma were 1.32 (95% CI, 0.73–2.38) and 3.20 (95% CI, 1.98–5.16), respectively. Gastric ulcer that was diagnosed 3–15 years before the reference date was associated with a significantly elevated risk of distal (adjusted OR, 3.11; 95% CI, 1.51–6.41) and gastric cardia adenocarcinoma (adjusted OR, 2.82; 95% CI, 1.23–6.48). Gastric ulcer that was diagnosed at least 16 years earlier was not related significantly to risk of distal gastric adenocarcinoma (adjusted OR, 1.54; 95% CI, 0.76–3.13) or gastric cardia adenocarcinoma (adjusted OR, 1.05; 95% CI, 0.42–2.62) (Table 2).

Compared with controls, patients with esophageal or gastric cardia adenocarcinomas were significantly more likely to have experienced weekly or daily symptoms of sour stomach/regurgitation and heartburn more than 2 years before the reference date (Table 3) There was a significant trend of increasing risk of esophageal and gastric cardia adenocarcinoma with increasing frequency and duration of these symptoms. When we used the combined reflux symptom index (see Statistical Analysis, Table 1), individuals with the highest scores (4.5–5) showed a nearly 7-fold (adjusted OR, 6.78; 95% CI, 3.19–14.39) increased risk of esophageal adenocarcinoma and a 3-fold (adjusted OR, 3.18; 95% CI, 1.42–7.09) increased risk of gastric cardia adenocarcinoma compared with patients with no symptoms. Risk of distal gastric carcinoma was not influenced by previous reflux symptoms. Similar patterns of associations were observed when we restricted the analyses to self-respondent cases. Among patients with the highest reflux scores, risk of distal gastric carcinoma remained close to unity, whereas the risk of esophageal and gastric cardia adenocarcinoma increased more than 5-fold (adjusted OR, 5.42; 95% CI, 2.27–12.95) and 3-fold (adjusted OR, 3.15; 95% CI, 1.29–7.71), respectively (data not shown). Self-reports of gas pain also were significantly more common among patients with all three tumor types compared with controls, but there was no difference in reports of difficulty swallowing between cases and controls (data not shown).

Table 3. Adjusted Odds Ratiosa and Corresponding 95% Confidence Intervals (in Parentheses) for Esophageal, Gastric Cardia, and Distal Gastric Adenocarcinomas in Relation to Self-Reported Symptoms that Occurred at Least 3 Years before the Reference Date
SymptomOdds ratio (95% confidence interval)Prevalence of symptom/characteristic in 1355 controls (%)
Esophageal adenocarcinoma (n = 222)Gastric cardia adenocarcinoma (n = 277)Distal gastric adenocarcinoma (n = 443)
  • a

    See Table 2 footnote.

  • b Symptoms that first started within 2 years of reference date were included in the baseline category of “never or monthly.”

  • c

    Score is based on symptoms of heartburn and sour stomach/regurgitation (see Statistical Analysis for scoring system).

Sour stomach/regurgitation    
 Never or monthly1.01.01.085.9
 Weekly3.53 (2.33–5.35)1.03 (0.63–1.68)1.00 (0.64–1.57)8.5
 Daily3.96 (2.50–6.27)2.58 (1.65–4.04)1.14 (0.67–1.94)5.6
 Duration of weekly or daily symptoms (yrs)    
  3–152.96 (1.81–4.80)1.74 (1.09–2.80)1.16 (0.74–1.80)6.9
  16+4.30 (2.88–6.42)1.54 (0.98–2.41)0.92 (0.53–1.58)7.2
Heartburn    
 Never or monthly1.01.01.088.9
 Weekly3.16 (2.00–5.01)2.63 (1.72–4.01)1.37 (0.86–2.17)6.7
 Daily4.06 (2.50–6.61)2.25 (1.35–3.75)0.97 (0.50–1.88)4.4
 Duration of weekly or daily symptoms (yrs)    
  3–152.11 (1.20–3.69)2.15 (1.31–3.50)1.14 (0.70–1.88)5.7
  16+4.89 (3.18–7.53)2.77 (1.78–4.31)1.34 (0.75–2.39)5.4
Reflux symptom indexc    
 01.01.01.081.4
 1–22.85 (1.87–4.33)1.74 (1.17–2.59)1.00 (0.69–1.47)11.1
 2.5–44.40 (2.82–6.88)2.37 (1.50–3.73)1.17 (0.68–2.01)6.1
 4.5–56.78 (3.19–14.39)3.18 (1.42–7.09)0.81 (0.21–3.19)1.3

Because the magnitude of the effect of any physician-diagnosed reflux condition (this index included hiatal hernia, esophagitis, excess acid, GE reflux, Barrett esophagus) on risk was similar to that of hiatal hernia, we further investigated the risk patterns in relation to hiatal hernia and other reflux conditions (Table 4). Specifically, patients were classified into one of four groups: ‘no history of any reflux conditions’ (baseline category), ‘no hiatal hernia but history of other reflux conditions’, ‘history of hiatal hernia but no history of other reflux conditions’, or ‘history of hiatal hernia and other reflux conditions’. Risk of esophageal adenocarcinoma was not increased significantly among those who had other reflux conditions but did not have hiatal hernia (adjusted OR, 1.67; 95% CI, 0.92–3.03). The risk was increased 3-fold (adjusted OR, 3.22; 95% CI, 1.88–5.51) among those who had hiatal hernia but did not have other reflux conditions and 9-fold (adjusted OR, 9.35; 95% CI, 5.20–16.83) among those who had both hiatal hernia and other reflux conditions. A more modest risk pattern for gastric cardia adenocarcinomas was also statistically significant, whereas we found no evidence of increased risk for distal gastric carcinoma.

Table 4. Adjusted Odds Ratiosa and Corresponding 95% Confidence Intervals (in Parentheses) for Esophageal, Gastric Cardia, and Distal Gastric Adenocarcinomas in Relation to Physician-Diagnosed Hiatal Hernia, Other Reflux Conditions, and the Reflux Symptom Score
  Odds ratio (95% confidence interval)Prevalence of condition in controls (%)
Esophageal adenocarcinomaGastric cardia adenocarcinomaDistal gastric adenocarcinoma
  • a

    See Table 2 footnote.

  • b The other reflux conditions included esophagitis, excess acid, gastroesophageal reflux, and Barrett esophagus.

  • c Score is based on symptoms of heartburn and sour stomach/regurgitation (see Statistical Analysis for scoring system).

  • d

    The reference category is no history of hernia and a reflux symptom score of 0 (see footnote c in Table 3). The first diagnosis of hiatal hernia within 2 years of the reference date was counted as no hiatal hernia.

Hiatal herniaOther reflux conditions    
 NoNo1.001.001.0087.9
 YesNo1.67 (0.92–3.03)1.34 (0.78–2.29)1.46 (0.89–2.39)5.5
 NoYes3.22 (1.88–5.51)2.21 (1.32–3.69)1.31 (0.70–2.45)4.4
 YesYes9.35 (5.20–16.83)2.56 (1.29–5.11)1.21 (0.47–3.14)2.2
Hiatal herniaReflux symptom score    
 0No1.00d1.00d1.00d78.3
 1+No3.61 (2.49–5.25)2.05 (1.46–2.87)1.13 (0.80–1.58)15.1
 0Yes5.85 (3.18–10.75)2.48 (1.34–4.58)1.99 (1.03–3.82)3.2
 1+Yes8.11 (4.75–13.87)2.93 (1.68–5.13)0.67 (0.27–1.65)3.5

Risk patterns for the three tumor types in relation to hiatal hernia and self-reported reflux symptoms are also displayed in Table 4. Subjects were classified into one of four groups: ‘no history of hiatal hernia and a reflux symptom score of 0’ (baseline category), ‘no history of hiatal hernia and a reflux symptom score of 1 or higher’, ‘history of hiatal hernia and a reflux symptom score of 0’, and ‘history of hiatal hernia and a reflux symptom score of 1 or higher’. Both hiatal hernia and reflux symptoms were significant independent risk factors for esophageal adenocarcinoma. The risk of esophageal adenocarcinoma was increased 3-fold (adjusted OR, 3.61; 95% CI, 2.49–5.25) among those who had reflux symptoms but did not have hiatal hernia, 6-fold (adjusted OR, 5.85; 95% CI, 3.18–10.75) among those who had hiatal hernia but did not have reflux symptoms, and 8-fold (adjusted OR, 8.11; 95% CI, 4.75–13.87) among those who had both reflux symptoms and hiatal hernia. Similarly, a more modest risk pattern was observed for gastric cardia adenocarcinoma that was also statistically significant, whereas there was no evidence of increased risk for distal gastric carcinoma (Table 4).

Table 5 shows the relation between current BMI and history of hiatal hernia/reflux symptom scores among control subjects. The risk of hiatal hernia/reflux symptoms increased significantly with increasing current BMI in all control subjects combined (P trend = 0.05). Similar patterns of positive association were observed when we considered BMI by history of hiatal hernia (P trend = 0.23) and by reflux symptoms (P trend = 0.08). Similarly, a positive association between hiatal hernia/reflux symptoms and BMI at age 40 was found. Compared with those in the lowest quartile of BMI at age 40, those in the highest quartile showed an increased risk of hiatal hernia/reflux symptoms (adjusted OR, 1.68; 95% CI, 1.09–2.58; data not shown). The risk of hiatal hernia/reflux symptoms in relation to smoking history also was investigated (Table 5). Relative to those who had never smoked, former smokers showed a significant increased risk of hiatal hernia (adjusted OR, 2.07; 95% CI, 1.30–3.30), but there was no increased risk among current smokers. There was no significant increased risk of reflux symptoms among former or current smokers (Table 5).

Table 5. Associations of Body Mass Index and Smoking Status with History of Hiatal Hernia and Reflux Symptom Score in all Control Subjects
CharacteristicAdjusted OR (95% CI)
Hiatal herniaa (no/yes)Reflux symptom scoreb (no/yes)Hiatal hernia and reflux symptom scorec (no/yes)
  • OR: odds ratio; CI: confidence interval; BMI: body mass index; Q: quartile.

  • a

    Analysis compared control subjects with hiatal hernia and control subjects without hiatal hernia.

  • b

    Analysis compared control subjects with reflux symptoms and control subjects without reflux symptoms. A reflux score of 1+ was considered to be yes for reflux symptoms. The score is based on symptoms of heartburn and sour stomach/regurgitation (see Statistical Analysis for scoring system).

  • c

    Analysis compared control subjects with hiatal hernia and/or reflux symptoms and control subjects with neither hiatal hernia nor reflux symptoms.

  • d The multivariate model included categoric variables for age (≤ 39 years, 40–49 years, 50–59 years, 60–69 years, 70+ years), gender, race (white, African American, Latino-American Asian American), birthplace (US-born, non-US-born), and education (less than high school, high school, some college, college graduate or more).

Current BMI quartile   
 Q11.001.001.00
 Q21.18 (0.66–2.11)1.09 (0.73–1.63)1.08 (0.74–1.57)
 Q31.34 (0.75–2.39)1.33 (0.89–1.98)1.23 (0.84–1.79)
 Q41.40 (0.78–2.52)1.37 (0.91–2.04)1.43 (0.78–2.52)
 P trend0.230.080.05
Smoking status   
 Never1.01.01.00
 Former2.07 (1.30–3.30)1.07 (0.78–1.46)1.28 (0.95–1.72)
 Current0.64 (0.27–1.50)0.84 (0.54–1.29)0.83 (0.54–1.26)

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The current study has several limitations. First, the overall response rate was modest. Fifty-five percent of all eligible patients were interviewed and included in the current analysis. The patients we interviewed tended to have an earlier stage of disease at diagnosis.3 However, a history of hiatal hernia, any reflux conditions, and reflux symptoms did not differ significantly by stage at diagnosis for the three tumor types under study among the interviewed patients. Among controls, history of reflux conditions and symptoms did not differ significantly between those who were the first matched controls versus those who were not first matched controls. Second, NOK interviews accounted for 271 of the 942 interviews with case patients. The observed patterns of risks in relation to physician-diagnosed reflux condition and reflux symptoms were similar when we restricted the analyses to self-respondents. This finding suggests that the overall results of the current study cannot be due to bias resulting from the use of surrogate respondents. Third, results related to information on previous conditions of the esophagus and stomach were based on self-report, although we considered only physician-diagnosed conditions and asked extensive questions regarding each condition (e.g., age of first diagnosis, treatment). The results of the current study showed that the risk of esophageal adenocarcinoma was influenced strongly by reflux symptoms and conditions, particularly hiatal hernia, whereas the associations were more modest for gastric cardia adenocarcinoma and no associations were found for distal gastric carcinoma. Quantitative and qualitative differences in the associations by tumor type strengthen the validity of the overall findings, because it is less likely that our results are due to recall bias. In addition, previous conditions of the large intestine were not reported more frequently by cases, also arguing that the findings are credible and cannot be explained by selective recall bias among case patients. We are mindful that only some individuals seek medical advice from a physician in relation to reflux symptoms and that symptom severity, access to health care, social class, and other factors may influence whether reflux symptoms become diagnosed. Similar risk patterns in relation to self-reported reflux symptoms and diagnosed reflux conditions also strengthen the overall results. Although multiple comparisons were made in the current analysis, which may lead to chance findings, we believe our analyses have been guided by previous knowledge and sound biologic rationale.

A 7-fold increased risk of esophageal adenocarcinoma was associated with reflux symptoms that occurred daily and began at least 16 years before the reference date. These results are compatible with those reported in previous population-based, case–control studies conducted in the United States7 and Sweden.6 In the current study and in the Swedish study,6 the risk of gastric cardia adenocarcinoma was associated positively with reflux symptoms, but this finding was not supported by the U.S. multicenter study.7 Previous GE conditions also emerged as significant risk factors. All five conditions (Barrett esophagus, hiatal hernia, GE reflux, esophagitis, and excess acid) were reported more frequently by patients with esophageal and gastric cardia adenocarcinoma than by controls. Hiatal hernia and esophagitis also were significant risk factors for esophageal adenocarcinoma in the U.S. multicenter study, but these conditions were not related to risk of gastric cardia adenocarcinoma.7 The results of the current study are similar to those reported in a medical record–based case–control study of gastric cardia (62% of cases) and esophageal adenocarcinomas (38% of cases) combined.9

There is increasing recognition that hiatal hernia may be a more important cause of reflux than was previously believed.26–29 Our results further suggest that hiatal hernia alone, other reflux diseases and symptoms, and the combination of hiatal hernia with other reflux diseases or symptoms are particularly strong predictors of esophageal adenocarcinoma risk (Table 4). Reasons for our particularly strong associations with hiatal hernia are not apparent. However, some of the differences in results between studies may be due to misclassification of specific conditions and/or symptoms of the esophageal and gastrointestinal tracts as this and most other epidemiologic studies rely on self-reported histories. We note that in the current study, 12.1% of control patients (165 of 1352) reported a positive history of any physician-diagnosed reflux conditions (hiatal hernia, esophagitis, GE reflux, excess acid, Barrett esophagus). Approximately 56% of these patients (93 of 165) had hiatal hernia alone or in combination with another reflux disease, whereas 44% (75 of 165) had a reflux condition other than hiatal hernia. In addition, only 45% of control subjects with hiatal hernia (47 of 90) also reported weekly reflux symptoms, whereas only 16% of those with reflux symptoms reported hiatal hernia (47 of 251). These results among control subjects emphasize the importance of including questions on both conditions and symptoms of the esophageal and gastrointestinal tracts, because many individuals with other reflux diseases or symptoms will have no demonstrable hiatal hernia whereas others with a hiatal hernia will have no reflux symptoms or other reflux conditions.

Hiatal hernia and reflux conditions both are associated with an increased risk of Barrett esophagus, a very strong risk factor for esophageal adenocarcinoma in the current and previous studies.24, 25 However, the underlying causes of hiatal hernia and reflux conditions are not well understood. One favored hypothesis is that increasing BMI increases the risk of reflux diseases.17 Higher abdominal pressure, obesity, and other aspects of sedentary lifestyles also have been suggested.30 Among control subjects in the current study, a risk of hiatal hernia/reflux symptoms was increased significantly with increasing current BMI as well as BMI at age 40. Therefore, our results are compatible with the observations of cross-sectional,15–17 case–control,20 and follow-up studies18 that have found a significant and positive association between body size and a history of esophagitis, hiatal hernia, and symptoms of GERD. Reasons for the null findings regarding the potential correlation between body size and GERD in the Swedish study by Lagergren et al. are not apparent.14 However, their analysis compared individuals with GERD (including recurrent heartburn or regurgitation) with those without GERD so that some individuals with hiatal hernia may have been misclassified. Differences in the definition of reflux diseases, their symptoms, and high BMI in the various studies may have contributed, in part, to the inconsistent results.

In summary, in the current large, population-based case–control study in Los Angeles County, BMI was associated with the risk of hiatal hernia/reflux symptoms. Hiatal hernia, in combination with other reflux conditions and symptoms, was a predictor of risk of both gastric cardia and esophageal adenocarcinomas. The latter associations were particularly strong.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The authors thank all the study participants for their contributions. The authors also thank Annie Fung, Isaura Rivera, Timothy Stirton, and June Yashiki for their help with data collection.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES