Incidence of esophageal and gastric carcinomas among American Asians/Pacific Islanders, whites, and blacks

Subsite and histology differences

Authors

  • Xiaocheng Wu M.D., M.P.H.,

    Corresponding author
    1. Epidemiology Program, School of Public Health/Louisiana Tumor Registry, Louisiana State University Health Sciences Center, New Orleans, Louisiana
    • Epidemiology Program, School of Public Health/Louisiana Tumor Registry, Louisiana State University Health Sciences Center, New Orleans, LA
    Search for more papers by this author
    • Fax: (504) 568-2493

  • Vivien W. Chen Ph.D.,

    1. Epidemiology Program, School of Public Health/Louisiana Tumor Registry, Louisiana State University Health Sciences Center, New Orleans, Louisiana
    Search for more papers by this author
  • Bernado Ruiz M.D., Ph.D.,

    1. Department of Clinical Pathology, School of Medicine, Louisiana State University Health Sciences Center, New Orleans, Louisiana
    Search for more papers by this author
  • Patricia Andrews M.P.H.,

    1. Epidemiology Program, School of Public Health/Louisiana Tumor Registry, Louisiana State University Health Sciences Center, New Orleans, Louisiana
    Search for more papers by this author
  • L. Joseph Su Ph.D.,

    1. Epidemiology Program, School of Public Health/Louisiana Tumor Registry, Louisiana State University Health Sciences Center, New Orleans, Louisiana
    Search for more papers by this author
  • Pelayo Correa M.D.

    1. Department of Clinical Pathology, School of Medicine, Louisiana State University Health Sciences Center, New Orleans, Louisiana
    Search for more papers by this author

  • The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute, National Institutes of Health, Department of Health and Human Services.

Abstract

BACKGROUND

The authors examined subsite-specific and histology-specific esophageal and gastric carcinoma incidence patterns among the Asians/Pacific Islander (API) population in the United States and compared them with those among whites and blacks.

METHODS

Data on newly diagnosed esophageal and gastric carcinomas during 1996–2000 were obtained from 24 population-based central cancer registries, representing approximately 80% of the API population in the United States. Age-adjusted rates, using the 2000 United States standard population, and age-specific rates were computed by anatomic subsite, histology, race, and gender. The difference in the age-adjusted rates between APIs and other races were examined using the two-tailed z statistic.

RESULTS

Greater than 75% of esophageal carcinomas among APIs, both males and females, were squamous cell carcinoma. Adenocarcinoma accounted for < 20% of all esophageal carcinomas. This pattern was similar to that among blacks but was completely opposite to that among whites. The rate of esophageal squamous cell carcinoma was 81% higher among API males compared with white males, but it was 64% less compared with black males. The rates of esophageal adenocarcinoma were significantly lower among APIs than among both whites and blacks both males and females. The majority of gastric carcinomas among APIs were noncardia adenocarcinoma, whereas cardia adenocarcinoma accounted for only 11% of gastric carcinomas among API males and 6% of gastric carcinomas among API females. The age-adjusted incidence rate of cardia adenocarcinoma was 23% lower among API males compared with white males, but it was 26% higher compared with black males. In contrast, the rates of noncardia adenocarcinoma among APIs were approximately 3.7 times the rate among whites for both males and females and 33% higher than the rate among blacks.

CONCLUSIONS

Subsite-specific and histology-specific incidence patterns of esophagogastric carcinoma among APIs differ from those among whites and blacks. The reasons for significantly higher rates of noncardia adenocarcinoma among APIs compared with whites and blacks need further investigation. Cancer 2006. © 2005 American Cancer Society.

Data from the Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute have shown a lower incidence rate of esophageal carcinoma and a higher incidence rate of gastric carcinoma among Asians/Pacific Islander (API) population compared with both white and black populations in the Untied States. Previous studies have suggested that esophageal and gastric carcinoma risk factors and their contributions to the occurrence of these malignancies vary by anatomic subsite and histologic type.1–6

To gain a better understanding of differences in the incidence of gastric and esophageal carcinomas between APIs and other races, data on subsite-specific and histology-specific incidence are needed. Because of the small numbers of esophageal and gastric carcinomas in the API population, studies addressing their subsite-specific and histology-specific incidence patterns for these malignancies are scarce in the United States. In one recent study, multiethnic variations of esophageal and gastric cardia carcinomas were examined, and it was found that incidence patterns were different between APIs and other races.7 That study used the SEER data, which cover 10–14% of the United States population. Because there are striking geographic variations in the incidence rates of esophageal and gastric carcinomas within the United States,8 it is important to determine whether the patterns observed in the SEER areas are consistent with patterns in much larger geographic areas. In addition, the earlier study did not examine gastric noncardia cancer incidence. The objective of the current study was to examine subsite-specific and histology-specific esophageal and gastric carcinoma incidence patterns among APIs and to compare them with the patterns among whites and blacks using the large, aggregated cancer incidence data from the North American Association of Central Cancer Registries (NAACCR). Because the rates for both esophageal and gastric carcinomas are much higher among males than among females, and the relative risks for males versus females vary substantially by geographic area within the United States,8 we stratified our analysis by race and gender in this study.

MATERIALS AND METHODS

Study Data

Cancer incidence data for 1996–2000 were from 24 population-based cancer registries that participate in the SEER Program and the Centers for Disease Control and Prevention's National Program of Cancer Registries (NPCR). The registries are Arizona, Atlanta (Georgia), California, Colorado, District of Columbia, Connecticut, Florida, Hawaii, Idaho, Illinois, Iowa, Kentucky, Louisiana, Michigan, Montana, Nebraska, New Jersey, Oregon, Rhode Island, Utah, Washington, West Virginia, Wisconsin, and Wyoming. These registries met the NAACCR criteria for high-quality incidence data and consented to the use of their data for this study.9 These states and metropolitan areas encompassed approximately 47% of the United States population, including approximately 47% of whites, 40% of blacks, and 80% of APIs.

Information on race in the registries' data was derived from medical records, which were coded according to standard race codes and then grouped into standard race recode categories in conformity with federal standards for the years that the study data represent.10, 11 API was a standard category for the classification of federal data on race (OMB) prior to the 2000 Census. Because population data for the API subgroups were not available, they had to be combined into a single, standardized API category in the current study. The API category includes Chinese, Japanese, Filipinos, Hawaiians, Koreans, Asian Indians-Pakistanis, Vietnamese, Laotians, Hmong, Kampucheans, Thai, Micronesians, Chamorrans, Guamanians, Polynesians, Tahitians, Samoans, Tongans, Melanesians, Fiji Islanders, New Guineans, and other Asians/Pacific Islanders and Orientals/Pacific Islanders. Most APIs in this study were Japanese (25.4% of all APIs), Chinese (20.8%), Korean (15.6%), Filipino (10.8%), and Vietnamese (7.4%). These 5 subgroups accounted for approximately 80% of all the esophageal and gastric carcinomas among APIs. Thai, Pacific Islander, Polynesian, Guamanian, Fiji Islander, Melanesian, and Tahitian subgroups had < 16 carcinomas each. We had no esophageal and gastric carcinomas for Chamorrans, Micronesians, or New Guineans (data not shown). For the current study, we analyzed only data for APIs, whites, and blacks. American Indians and Alaska Natives were not included because of large problems with misclassification.12 Population estimates for 1996–2000 were obtained from the SEER Program based on United States Bureau of the Census population estimates for those years.13

This study included incident invasive primary esophageal and gastric carcinomas that were coded according to International Classification of Diseases for Oncology, Second Edition topography codes C15.0–C15.9 and C16.0–C16.9.14 Malignant lymphomas from these sites were excluded. Approximately 4.8% of the tumors were excluded, because they had unknown or no microscopic confirmation. The majority of excluded tumors had no specific histology or subsite information. Esophagogastric carcinomas were grouped into 4 major anatomic subsites: esophagus (C15.0–C15.9), gastric cardia (C16.0), gastric noncardia (C16.1–C16.6), and gastric unspecified, which comprised overlapping or unspecified gastric subsites (C16.8–C16.9). Histologic categories for esophagogastric carcinomas were squamous cell carcinoma (M8050–M8082), adenocarcinomas (M8140–M8573), and other histologic types.

Statistical Analysis

Relative frequencies and average, annual, age-specific and age-adjusted incidence rates per 100,000 were computed by anatomic subsite (esophagus, gastric cardia, gastric noncardia), histologic type (squamous cell carcinoma, adenocarcinoma, and other histologic type), race (whites, blacks, and APIs), and gender (male or female). The 2000 United States standard population was used to adjust for age. Subsite-specific and histology-specific incidence rates among APIs were compared with the rates among whites and blacks using the two-tailed z statistic.15 Because of the small number of APIs by age group, age-specific incidence rates were calculated using 10-year age intervals. The rates are presented using uniform semilogarithmic scales, so that age-specific rates can be compared easily by race, gender, and age group. Counts and rates were suppressed in individual cells that contained < 16 tumor counts. The rates, standard errors, and 95% confidence intervals were generated using SEER*Stat software (version 5.3.1).16

RESULTS

The total number of invasive esophageal and gastric carcinomas diagnosed in 1996–2000 was 77,692. Thirty-eight percent of these carcinomas were esophageal carcinoma, and 62% were gastric carcinoma. For esophageal carcinoma, 85% of the patients were white, 13% were black, and 2% were API. For gastric carcinoma, 80% of the patients were white, 12% were black, and 8% were API.

Overall, the age-adjusted (2000 United States) incidence rates of esophageal carcinoma for both males and females were lower among APIs (P < 0.05) than among whites and blacks. The rates of gastric carcinoma, by contrast, were higher (P < 0.05) among APIs than among whites and blacks. The rates for both esophageal and gastric carcinomas were higher among males (P < 0.05) than among females regardless of race (Tables 1, 2).

Table 1. Percentage Distributions, Age-Adjusted (2000 United States) Incidence Rates and Rate Ratios of Esophagogastric Carcinomas by Race, Anatomic Subsites, and Histology in Selected Areasa in the United States, 1996-2000, Males
Anatomic subsite/histologyNo. (%)Age-adjusted ratesbRate ratios (95% CI)
APIsWhitesBlacksAPIsWhitesBlacksAPIs/whitesAPIs/blacks
  • 95%CI: 95% confidence interval; API: Asian/Pacific Islander.

  • a

    Data source: NAACCR 1996–2000 analytic data file, including data from 24 SEER and NPCR registries. They are Arizona, Atlanta (Georgia), California, Colorado, District of Columbia, Connecticut, Florida, Hawaii, Idaho, Illinois, Iowa, Kentucky, Louisiana, Michigan, Montana, Nebraska, New Jersey, Oregon, Rhode Island, Utah, Washington, West Virginia, Wisconsin, and Wyoming.

  • b

    Rates are per 100,000.

  • c

    International Classification of Diseases for Oncology topography codes are C15.0–C15.9.

  • d

    International Classification of Diseases for Oncology topography codes are C16.0–C16.9.

Esophagusc555 (100.0)19,077 (100.0)2539 (100.0)4.97.712.60.64 (0.60–0.69)0.39 (0.36–0.42)
 Squamous cell429 (77.3)5284 (27.7)2127 (83.8)3.82.110.51.81 (1.58–2.08)0.36 (0.33–0.39)
 Adenocarcinoma100 (18.0)12,542 (65.7)238 (9.4)0.95.01.20.18 (0.17–0.20)0.74 (0.59–0.93)
 Other histologic types26 (4.7)1251 (6.6)174 (6.8)0.20.50.90.45 (0.33–0.60)0.26 (0.18–0.37)
Stomachd2337 (100.0)24,268 (100.0)3428 (100.0)21.310.018.72.13 (2.00–2.27)1.14 (1.08–1.21)
 Adenocarcinoma2195 (93.9)22,309 (91.9)3019 (88.1)20.09.216.62.17 (2.04–2.31)1.21 (1.14–1.29)
 Cardia258 (11.0)9042 (37.2)355 (10.4)2.43.61.90.67 (0.59–0.76)1.26 (1.05–1.50)
 Noncardia1392 (59.6)8286 (34.1)1740 (50.8)12.83.59.63.67 (3.28–4.08)1.33 (1.23–1.44)
 Other unspecified subsites545 (23.3)4981 (20.5)924 (27.0)4.82.15.12.29 (2.03–2.58)0.95 (0.85–1.07)
 Other histologic types125 (6.1)1699 (8.1)352 (11.9)1.10.71.81.61 (0.27–2.04)0.62 (0.52–0.76)
Table 2. Percent Distributions, Age-Adjusted (2000 United States) Incidence Rates and Rate Ratios of Esophagogastric Carcinomas by Race, Anatomic Subsite, and Histology in Selected Areasa in the United States 1996-2000, Females
Anatomic subsite/histologyPercentage distributionsAge-adjusted ratesbRate ratios (95% CI)
APIsWhitesBlacksAPIsWhitesBlacksAPIs/whitesAPIs/blacks
  • 95% CI: 95% confidence interval; API: Asian/Pacific Islander; — suppressed because there were < 16 patients in the category.

  • a

    Data source: NAACCR 1996–2000 analytic data file, including data from 24 SEER and NPCR registries. They are Arizona, Atlanta (Georgia), California, Colorado, District of Columbia, Connecticut, Florida, Hawaii, Idaho, Illinois, Iowa, Kentucky, Louisiana, Michigan, Montana, Nebraska, New Jersey, Oregon, Rhode Island, Utah, Washington, West Virginia, Wisconsin, and Wyoming.

  • b

    Rates are per 100,000.

  • c

    International Classification of Diseases for Oncology topography codes are C15.0–C15.9.

  • e

    Count and rate are suppressed due to fewer than 16 carcinomas.

Esophagusc165 (100.0)6161 (100.0)1051 (100.0)1.21.93.90.62 (0.55–0.70)0.31 (0.27–0.35)
 Squamous cell124 (75.2)3446 (55.9)896 (85.3)0.91.13.30.83 (0.70–0.99)0.27 (0.23–0.31)
 Adenocarcinoma30 (18.2)2223 (36.1)83 (7.9)0.20.70.30.30 (0.24–0.37)0.67 (0.45–1.00)
 Other histologic typesc492 (20.5)72 (27.0)c0.20.3ee
Stomachd1679 (100.0)13,991 (100.0)2441 (100.0)12.04.39.22.79 (2.58–3.02)1.30 (1.22–1.39)
 Adenocarcinoma1578 (94.0)12,749 (91.1)2143 (87.8)11.33.98.12.89 (2.66–3.14)1.39 (1.30–1.49)
 Cardia102 (6.1)2318 (16.6)159 (6.5)0.70.70.61.03 (0.85–1.26)1.24 (0.95–1.62)
 Noncardia1070 (63.7)6573 (47.0)1334 (54.6)7.72.05.13.87 (3.45–4.34)1.52 (1.39–1.66)
 Other unspecified subsites406 (24.2)3858 (27.6)650 (26.6)2.81.22.42.37 (2.03–2.76)1.15 (1.01–1.31)
 Other histologic types92 (6.0)1140 (8.9)277 (12.2)0.60.41.01.79 (0.35–2.37)0.62 (0.50–0.77)

Age-Adjusted Subsite-Specific and Histology-Specific Incidence Rates

Esophageal squamous cell carcinoma and adenocarcinoma

Among API males (Table 1), > 77% of esophageal carcinomas were squamous cell carcinoma, and adenocarcinoma accounted for < 20% of all esophageal carcinomas. This pattern of predominant squamous cell carcinoma also was observed among black males, but a reverse pattern was seen among white males, for whom squamous cell carcinoma accounted for < 28% of all esophageal carcinomas. The incidence rate of esophageal squamous cell carcinoma among API males was 81% higher than that among white males but was 64% lower than among black males. The incidence rate of esophageal adenocarcinoma among API males, however, was lower than among both white males and black males.

Among API females, squamous cell carcinoma is the most predominant histology type, accounting for 75% of all carcinomas, and adenocarcinoma accounted for < 18% (Table 2). The incidence rates of adenocarcinoma and squamous cell carcinoma among API females were lower than among both white females and black females.

Gastric cardia adenocarcinoma

Approximately 90% of gastric carcinomas in all race and gender groups were adenocarcinoma (Tables 1, 2). Cardia adenocarcinoma accounted for 11% of gastric carcinomas among API males; this percentage was similar to that observed among black males but was much lower than that observed among white males (about 37%). The age-adjusted incidence rate of cardia adenocarcinoma was 33% lower among API males than among white males and 26% higher than among black males.

The percentages of cardia adenocarcinoma were substantially lower among females than among males for all racial groups. The percentages of cardia adenocarcinoma were about the same among API females (6.1%) and black females (6.5%) but lower than among white females (17%). Age-adjusted incidence rates of cardia adenocarcinoma did not differ between API females, white females, and black females.

Gastric noncardia adenocarcinoma

Noncardia adenocarcinoma accounted for approximately 60% of gastric carcinomas among API males, which was higher than among black males (51%) and substantially higher than among white males (34%) (Tables 1, 2). The rate of noncardia adenocarcinoma among API males was 3.7 times that among white males and 33% higher than that among black males.

Two-thirds of gastric adenocarcinoma were noncardia adenocarcinoma among API females; the percentage was higher than among both white females (47%) and black females (55%). The incidence rates of noncardia adenocarcinoma among API females, however, were almost 4 times that among white females and 52% higher than among black females.

Gastric adenocarcinoma in other nonspecific subsites

Overall, gastric adenocarcinomas in unknown subsites, including overlapping lesions, accounted for 26% of gastric adenocarcinomas (Tables 1, 2). Approximately 24% of gastric adenocarcinomas were located in unspecified anatomic subsites among the API population. This percentage was slightly lower compared with that for white females, black males, and black females and was slightly higher compared with that for white males (21%).

Age-Specific, Subsite-Specific, and Histology-Specific Incidence Rates

The age-specific incidence rates of esophageal squamous cell carcinoma were lower among API males than among black males and higher than among white males in all age groups, although the rate differences between API males and black males narrowed with advancing age (Fig. 1). The age-specific curves of esophageal adenocarcinoma for males and esophageal squamous cell carcinoma for females showed that APIs had the lowest rates among the 3 race groups for the group ages 60–69 years and ages 70–79 years, but small numbers precluded comparisons within other age groups. Data also were too scarce to analyze esophageal adenocarcinoma among API females.

Figure 1.

These charts illustrate age-specific incidence rates of esophageal squamous cell carcinoma and esophageal adenocarcinoma by race and gender in selected areas in the United States, 1996–2000 (data source: NAACCR 1996–2000 analytic data file, including data from 24 SEER and NPCR registries. They are Arizona, Atlanta [Georgia], California, Colorado, District of Columbia, Connecticut, Florida, Hawaii, Idaho, Illinois, Iowa, Kentucky, Louisiana, Michigan, Montana, Nebraska, New Jersey, Oregon, Rhode Island, Utah, Washington, West Virginia, Wisconsin, and Wyoming).

The incidence rate of cardia adenocarcinoma among API males was about the same as the rates among black males and was lower than among white males for most age groups (Fig. 2). With increasing age, the rate among API males tended to approximate the rates among whites and blacks. Among individuals age 80 years and older, the rates of cardia gastric carcinoma were higher among APIs than among whites and blacks. The incidence rates of noncardia adenocarcinoma among API males were higher at all ages than the rates among white males and black males. Similar age-specific incidence patterns by race were observed among females (Fig. 2).

Figure 2.

These charts illustrate age-specific incidence rates of cardia and noncardia gastric adenocarcinoma by race and gender in selected areas in the United States, 1996–2000 (data source: NAACCR 1996–2000 analytic data file, including data from 24 SEER and NPCR registries. They are Arizona, Atlanta [Georgia], California, Colorado, District of Columbia, Connecticut, Florida, Hawaii, Idaho, Illinois, Iowa, Kentucky, Louisiana, Michigan, Montana, Nebraska, New Jersey, Oregon, Rhode Island, Utah, Washington, West Virginia, Wisconsin, and Wyoming).

DISCUSSION

In the current study, we found that, for esophageal squamous cell carcinoma among males, blacks had the highest incidence rate, whites had the lowest rate, and the rate for APIs was intermediate. Among females, APIs had the lowest rate. Although differences in rates were found between our study areas and SEER areas, the observed patterns were consistent with a published report in which 1992–1998 SEER data were examined.7 Several factors have been associated with an increased risk of esophageal squamous cell carcinoma, such as a positive family history, tobacco and alcohol abuse, low intake of vegetables and fruits, and high intake of N-nitroso compounds.6, 17, 18 The risk of esophageal squamous cell carcinoma attributable to these risk factors varies world wide by geographic area. For instance, an association of family history with increased risk of esophageal squamous cell carcinoma was found in several case–control and cohort studies in China, Iran, and Japan.19–22 However, this familial aggregation is not observed consistently in low-risk populations, in which most studies have shown negative results.23–25 Tobacco use and alcohol use are major contributors to esophageal squamous cell carcinoma, especially in the developed countries, where incidence rates of this malignancy normally are low; whereas the roles of tobacco use and alcohol use in the etiology of esophageal carcinoma in high-incidence areas, such as the Asian esophageal carcinoma belt, are less substantial.26 Compared with whites and blacks, the prevalence of tobacco use or alcohol use is lower among APIs. The risk of esophageal squamous cell carcinoma that is attributable to tobacco use and alcohol use for APIs probably is less than for whites and blacks. Although APIs may consume more plant-based foods than whites and blacks,27, 28 some traditional dietary commodities, such as pickled vegetable, smoked or dried fish, and salted food, may increase API exposure to N-nitroso compounds, which are potent carcinogens for esophageal carcinoma.29–33 High inherited susceptibility and possible high exposure to N-nitroso compounds due to traditional dietary habits may play important roles in the higher rate of esophageal carcinoma among API males than among white males, whereas tobacco use and alcohol use may be major contributors to the higher incidence rate of this disease among black males than among API males. Smoking prevalence among Asian women in the United States is considerably lower than among other races.34 The difference in smoking prevalence between Asian-American women and white women is much greater than the difference between Asian-American men and white men. It may explain in part the lower rate of esophageal squamous cell carcinoma among API women than among both white women and black women, offsetting the effect of family history and dietary habits.

The incidence rates of esophageal adenocarcinoma among APIs in this study were lower than the rates among whites and blacks. The same pattern was observed in the SEER data, although the rates differed slightly between our study areas and the SEER areas for all racial and gender groups.7 Both chronic gastroesophageal reflux disease and obesity have been associated strongly with the risk of esophageal adenocarcinoma.35–38 Although their association is not as strong as that for esophageal squamous cell carcinoma, alcohol consumption and cigarette smoking are related to an increased risk of esophageal adenocarcinoma.4–6, 39 Previous studies have indicated that Asian Americans may have a lower rate of gastroesophageal reflux disease than whites and blacks,40–42 and the prevalence of obesity is lower in the six largest Asian-American groups (Chinese, Filipino, Asian Indian, Japanese, Korean, and Vietnamese) compared with their white and black counterparts.43 Low prevalences of gastroesophageal reflux disease, obesity, and tobacco and alcohol use among APIs may be the primary contributors to the lower rates of esophageal adenocarcinoma among APIs than among whites and blacks. In the current study, we also found that incidence rates of esophageal adenocarcinoma were lower among blacks than among whites. A similar prevalence of gastroesophageal reflux disease among whites and blacks and a higher prevalence of obesity among blacks than among whites in the United States cannot explain the observed the black-white difference,34, 44 which probably implies that other factors, such as tobacco and alcohol use, diet, and other host and genetic factors, also may play a critical role in the occurrence of this disease.45

The incidence rate of gastric cardia adenocarcinoma among API males in the current study was lower than among white males but higher than among black males. For females, the rates were about the same across racial groups. This pattern was similar to that observed in the SEER areas.7 The anatomic proximity and the similarities in incidence trends and risk factor profiles for gastric cardia and esophageal adenocarcinoma have resulted in speculation that they may be the same disease entity.46, 47 Findings from more recent studies, however, disagree with that speculation.7, 48 El-Serag et al., analyzing SEER data, reported a flat time trend for cardia carcinoma and a rising incidence for esophageal adenocarcinoma.2 Although, in the current study, we observed that the incidence of esophageal adenocarcinoma among API males was lower than among black males, the incidence of gastric cardia adenocarcinoma for the two racial groups was reversed, and a conclusion about the observed incidence patterns in this study could not be reached because of the potential misclassification of esophageal adenocarcinoma as cardia adenocarcinoma.

We also found that the incidence rates of gastric noncardia adenocarcinoma among API males and females were substantially higher than the rates among either whites or blacks. The factors that have been identified as significant contributors to an increased risk for this disease include Helicobacter pylori infection, tobacco smoking, high salt intake, low intake of vegetables and fruits, high intake of pickled food, high intake of nitrates and nitrate-related compounds, and family history.45, 49–55 It has been found that, when individuals migrate between countries or areas with different risks of gastric carcinoma, their risk patterns usually are retained or are changed only slightly.56 When incidence patterns do change in response to the new environment, the transition seems to be slower for gastric carcinoma than for other malignancies, such as colorectal carcinoma.57 This phenomenon suggests that early-life exposure is important for the risk of gastric carcinoma. In general, the prevalence of H. pylori infection is higher in developing countries than in developed countries.58, 59 Because new infections in adults are rare, because infections tend to stay active unless they are treated with antibiotics,60 and because approximately 67% of APIs in the United States are first-generation immigrants,61 the widespread infections during childhood among API immigrants may play an important role in their elevated rates of gastric noncardia carcinoma. The traditional Asian dietary commodities with higher levels of volatile N-nitrosamine than most Western-style foods,62 as discussed above, also may contribute to differences in the incidence of gastric noncardia carcinoma between APIs and other races. Because Asian countries traditionally have a higher gastric carcinoma incidence rate than the United States, the likelihood of having a positive family history for gastric carcinoma may be higher among APIs than among whites and blacks. Typically, a 50–130% elevated risk of gastric carcinoma has been observed among individuals who have a positive family history of this disease.56 Inherited genes also may play an important role in the elevated rates of gastric noncardia adenocarcinoma among APIs than among other races.

Our data also showed that 26% of gastric adenocarcinomas were located in unspecified anatomic subsites. The true reasons for the high percentage of patients with unknown subsites were not clear. Because the percentages of patients with unknown subsites were comparable by race and gender, it is unlikely that the more precise site identification for these patients would reverse the observed API-to-white or API-to-black rate ratios for gastric cardia and noncardia carcinomas.

Age-specific incidence patterns by race were consistent with the age-adjusted rates. Although age-specific incidence rates for some esophageal and gastric subsites were suppressed from this report due to small numbers of patients, the rates rose more sharply for esophageal squamous cell carcinoma and for gastric cardia and noncardia adenocarcinoma among older APIs than among older whites and blacks. The reason for the observed pattern is not clear.

There were three noteworthy limitations of this study. First, because population data for subgroups of APIs were not available, the subgroups were combined into a single API category that encompassed a broadly inclusive and diverse group. Incidence rates of esophageal and gastric carcinoma may vary strikingly among subgroups of the API category,26, 63, 64 and combining them prevented us from identifying high-risk subgroups. Second, this study was based on existing population-based cancer registry data, which were collected from hospital medical records and nonhospital reporting sources, and we were not able to evaluate misclassifications of anatomic subsite or histologic type. Third, because the majority of gastric adenocarcinomas were adenocarcinoma not otherwise specified, racial differences by histologic subtype could not be assessed.

Esophagogastric carcinoma is a complex disease. Although many factors have been associated with an increased risk of this disease, it remains unclear how different risk factors contribute to the esophageal and gastric carcinoma incidence by race and gender. The racial diversity and admixtures and the high percentage of immigrants to the United States provide us with a unique opportunity to study the etiologies of esophageal and gastric carcinomas. More research is needed to determine the attribution of each risk factor for these diseases by race and gender.

Ancillary