点击显示 收起
Cancer Research Center of Hawaii, University of Hawaii, Honolulu, Department of Pathology, University of Cincinnati Medical Center, Cincinnati, Ohio
Department of Medicine, New York University School of Medicine, New York
Department of Internal Medicine, Faculty of Medicine, Toho University School of Medicine, Tokyo, Japan
Background.
The objective was to investigate the association of Helicobacter pylori and serum pepsinogen (PG) levels with gastric adenocarcinoma.
Methods.
Serum obtained from 299 patients at the time of cancer diagnosis and from 336 population-based control subjects was tested for PG I, PG II, and antibodies to H. pylori and to CagA.
Results.
Subjects with low PG I levels or low PG I/II ratios were at increased risk for cardia and noncardia gastric cancer, whereas those with H. pylori or CagA seropositivity had an elevated risk for noncardia cancer only. Subjects seropositive for either H. pylori or CagA who had low PG I levels had the highest odds ratio (OR) (9.21 [95% confidence interval {CI}, 4.9517.13]) for noncardia cancer, compared with subjects with neither factor. Elevated risks were also found among subjects with only 1 factor (OR, 5.40 [95% CI, 2.6111.20] for low PG I level only; OR, 4.86 [95% CI, 5.908.13] for H. pylori or CagA seropositivity only). This pattern persisted when PG I/II ratio replaced PG I level and when CagA seropositivity alone replaced H. pylori immunoglobulin G or CagA seropositivity.
Conclusions.
The results suggest that persons with both H. pylori or CagA seropositivity and a low PG I level or PG I/II ratio are highly susceptible to development of noncardia gastric cancer.
Helicobacter pylori infection has been associated with chronic superficial gastritis [1], an early step in the pathogenesis of gastric cancer [2]. H. pylori strains may possess a functional cag island, which leads to injection of CagA protein into gastric epithelial cells, where it undergoes tyrosine phosphorylation and affects signal transduction pathways [3]. Persons carrying cagA+ H. pylori strains develop more-severe gastritis [46]. Consequently, it is important to assess CagA status when testing for H. pylori. Some individuals who carry cagA+ H. pylori strains may have H. pylori antibody levels below the cutoff value and, thus, may be falsely considered to be H. pylori seronegative [7].
Chronic atrophic gastritis, a later step in the carcinogenic process, has been linked to a low serum pepsinogen (PG) I level or to a low serum PG I/II ratio [8]. PG I and PG II are immunologically distinct aspartic proteinases of pepsin, which digests proteins [9]. PG I is synthesized by chief cells in fundic gland mucosa, and PG II is produced in similar cells and in pyloric gland cells in the gastric antrum. Destruction of fundic glands by atrophic gastritis leads to a reduction in PG I levels and in the PG I/II ratio [10]. Even though H. pylori infection does not always lead to atrophic gastritis [11] and atrophic gastritis is caused by factors other than H. pylori infection, it is important to investigate the relationship between serum PG levels, H. pylori seropositivity (especially with regard to CagA status), and gastric cancer risk. Although an H. pylori infection followed by atrophy is considered to be a common pathway to gastric cancer [12], these are not proven necessary factors.
Four case-control studies, in which serum samples were obtained at the time of diagnosis of gastric cancer [8, 1315], and 3 cohort studies [10, 16, 17] have evaluated serum PG levels, H. pylori status, and gastric cancer risk, but the results are inconsistent. Because of these equivocal findings and because only 1 previous study determined CagA status [15], we conducted a case-control study of gastric cancer in Hawaii that included measurement of levels of serum PGs and antibodies to H. pylori and to CagA. Our hypothesis was that stratification for both H. pylori/CagA and location of the cancer would permit a clearer understanding of the relationship between H. pylori infection, PG levels, and gastric cancer risk.
SUBJECTS, MATERIALS, AND METHODS
Study population.
Patients initially received diagnoses of gastric cancer between September 1993 and April 1999 at 8 major hospitals on the island of Oahu and were identified by the Hawaii Tumor Registry, a member of the Surveillance, Epidemiology, and End Results program of the National Cancer Institute [18]. There were 591 patients with tissue-confirmed gastric cancer, whose pathology reports and histologic slides were reviewed by the study pathologist (G.N.S.). Of these 591, 59 had a diagnosis of gastric lymphoma or malignant stromal tumor and were excluded from the analysis. Because we were limited to using unused preoperative serum available from participating hospitals, 299 (56%) of the remaining 532 patients with gastric adenocarcinoma had a sample available for the study; these serum samples were obtained at the time of diagnosis. The patients were 2697 years of age, were residents of Oahu, and belonged to 1 of 6 ethnic groups: white, Chinese, Filipino, Japanese, Korean, or Native Hawaiian (including part Hawaiian). Ethnicity was based on having at least 3 of 4 grandparents belonging to the same ethnicity, unless it included Hawaiian heritage.
Eligible control subjects were identified from lists of Oahu residents who had been interviewed by the Health Surveillance Program of the Hawaii Department of Health. Each year, this program identifies a 1% random sample of all households in the state, with a sampling procedure modeled after that of the National Health Survey [19]. To supplement the pool of eligible control subjects 65 years of age, Oahu residents registered with the Health Care Financing Administration (now the Centers for Medicare and Medicaid Services) also were identified; this institution is estimated to service 95% of individuals 65 years of age in the United States [20].
The population-based control subjects were frequency matched to the case patients on the basis of sex, ethnicity, and 5-year age groups. A total of 759 potential control subjects without a history of gastric cancer were identified. Of these, 36 (5%) had died or were too ill to be questioned. Among the remaining 723 potential control subjects, 220 (30%) refused the interview, and 57 (8%) were unlocatable, had moved, or had difficulty communicating in English. Interviews were completed for the remaining 446 control subjects. Of these, 336 (75%) provided blood samples, whereas 84 (19%) refused and 26 (6%) did not provide blood samples for other reasons.
Pathologic classification.
The histologic classification of hematoxylin-eosinstained slides used for this study was that of Lauren [21], which classifies gastric adenocarcinomas into the following basic types: intestinal, diffuse, and mixed/other. In all, there were 212 (71%) intestinal, 65 (22%) diffuse, and 22 (7%) mixed/other cases.
The identification of subsite was based on the gross description of cancers in resection specimens. Tumors were classified as "cardia" if the cardioesophageal junction was involved. The noncardia (distal) cases included the corpus if it arose in the oxyntic mucosa, including the mucosa of the fundus, and included the antrum if it involved the pyloric antrum, including the antrocorpus junction. There were 35 cardia and 264 noncardia cases in the study.
Serologic methods.
The blood samples from case patients and control subjects were tested for the following: serum PG I, serum PG II, and IgG antibodies to H. pylori and to the CagA protein of H. pylori. The measurement of serum PG level was performed by immunoradiometric assay [22] using PG I, II RIA BEAD kits (Dainobot). Low PG levels were defined as PG I levels 30 ng/mL or PG I/II ratios 2.0. These cut points best discriminated between case patients and control subjects in our study; other studies have used different definitions of low PG levels, based on their study populations [8, 9, 10, 1417].
Assays for serum IgG antibodies to H. pylori group antigens and to a recombinant fragment of the CagA protein were performed exactly as described in prior studies [23, 24]. All assays were performed at least in duplicate on coded specimens, by laboratory personnel blinded to the identity or status of the study subjects. A total of 227 case patients had both the PG and H. pylori/CagA tests performed. Because of limited amounts of sera, 49 had only the H. pylori/CagA tests performed, and 23 had only the PG tests performed. All 336 control subjects had the H. pylori/CagA tests performed, and 334 had the PG tests performed.
Statistical analysis.
For the statistical analysis, we used unconditional logistic regression to compute odds ratios (ORs) and 95% confidence intervals (CIs) for exposures of interest [25]. Indicator measures representing PG I or PG I/II status and H. pylori or CagA status were entered as independent variables. The models were adjusted for the variables used in the frequency match: sex, ethnicity, and age (continuous). Separate models were estimated for cardia and noncardia cancers and for intestinal and diffuse cancers, to determine whether the risk factors varied by cancer subsite or histologic type; all control subjects were used in these models.
A joint-effects analysis was performed and is defined as assessment of the risk for gastric cancer among subjects with H. pylori or CagA seropositivity who have low PG values, as well as among subjects abnormal for only 1 of these factors, compared with subjects with neither factor. Because it is not expected that both factors together equal the sum of the individual factors, on either an additive or a multiplicative scale, a test of interaction is not presented in table 3.
RESULTS
Study population.
There were 299 case patients with gastric adenocarcinoma and 336 control subjects in the study. The mean (±SD) age was 70.7 (±11.8) years for case patients and 70.6 (±12.7) years for control subjects. There were a total of 183 male and 116 female case patients. Of the control subjects, 228 were men and 108 were women. Because results were similar for men and women, they were combined in the analyses.
Serum PGs.
We compared 250 case patients with gastric adenocarcinoma and 334 control subjects according to the percentage with low serum PG I levels or low serum PG I/II ratios (table 1). The ORs for gastric cancer were significant for both a low PG I level (OR, 2.66 [95% CI, 1.813.90]) and a low PG I/II ratio (OR, 2.78 [95% CI, 1.844.21]). After separation by tumor location, there were 221 noncardia (distal) cases and only 29 cardia cases. A significant association was observed for both subsites. The cases were also separated by histologic type (intestinal or diffuse); there was a positive association of a low PG I level and a low PG I/II ratio with both types of tumor.
H. pylori or CagA status.
A similar analysis was performed for 276 case patients and 336 control subjects according to H. pylori or CagA status (table 2). More case patients than control subjects were seropositive for either H. pylori or CagA (OR, 2.82 [95% CI, 1.993.99]). Similarly, more case patients than control subjects were seropositive for CagA alone (OR, 2.59 [95% CI, 1.843.65]). Next, the cases were separated by tumor location. The positive association persisted for noncardia cases (OR, 3.41 for H. pylori or CagA seropositivity; OR, 3.16 for CagA seropositivity alone) but not for cardia cases of the stomach.
When the analysis in table 2 was limited to only subjects who were H. pylori seropositive instead of subjects who were H. pylori or CagA seropositive, the OR for the noncardia cases was 2.49 (95% CI, 1.763.54) instead of 3.41, reflecting the fact that there were 24 case patients and 14 control subjects who were H. pylori seronegative and CagA seropositive.
Because of the lack of a positive association between H. pylori seropositivity and cardia cancer found in this study and in prior studies [23, 24, 26], only noncardia cases were included in subsequent analyses. When they were separated according to histologic type (table 2), the associations with being seropositive for either H. pylori or CagA or with being seropositive for CagA alone were stronger for the diffuse noncardia cases.
Joint effect of H. pylori/CagA seropositivity and PG values.
To evaluate the joint effect of H. pylori IgG or CagA seropositivity and low PG I level on gastric cancer risk, we defined those who were H. pylori and CagA seronegative and had normal PG I levels as the referent group (table 3). Compared with the referent group, subjects who were H. pylori and CagA seronegative and had low PG I levels had an OR of 5.40 (95% CI, 2.6111.20) for noncardia gastric cancer. Subjects who were seropositive for either H. pylori or CagA and had normal PG I levels had an OR of 4.86 (95% CI, 2.908.13), whereas those who were seropositive for either H. pylori or CagA and had low PG I levels had an OR of 9.21 (95% CI, 4.9517.13). A similar pattern persisted when noncardia cases were separated into intestinal and diffuse histopathologic categories. For each type, subjects who were seropositive for either H. pylori or CagA and had low PG I levels had the highest ORs; for diffuse cancers, the OR for that group was >40 (95% CI, 9.51174.60).
When PG I/II ratios were examined instead of PG I levels, subjects who had low ratios and were seropositive for either H. pylori or CagA were found to have the highest OR (6.88 [95% CI, 3.8112.44]) for noncardia cancer. The same pattern was seen for intestinal cases but not for diffuse cases, possibly because of their small sample size.
Next, we defined subjects who were CagA seronegative and had normal PG I levels as the referent group (table 3). Compared with the referent group, subjects who were CagA seronegative and had low PG I levels had an OR of 4.76 (95% CI, 2.728.34) for noncardia cancer. Subjects who were CagA seropositive and had normal PG I levels had an OR of 4.79 (95% CI, 3.007.63), and those who were CagA seropositive and had low PG I levels had an increased OR, of 6.40 (95% CI, 3.4511.89). Again, for the intestinal and diffuse types, subjects who were CagA seropositive and had low PG I levels had the highest ORs. Finally, when PG I/II ratios were used instead of PG I levels, those who had low PG I/II ratios and were seropositive for CagA were found to have the highest OR (6.08 [95% CI, 3.2211.48]) for noncardia cancer.
DISCUSSION
In the present study, subjects with low PG I levels or low PG I/II ratios were at an increased risk for cardia and noncardia cancer of the stomach, including both histologic types, as shown in a previous study [8]. Subjects with H. pylori or CagA seropositivity had an elevated risk of intestinal and diffuse cancer of the distal (noncardia) stomach, but H. pylori seropositivity was not associated with cancer of the cardia. This is consistent with findings of earlier investigations [23, 24, 26]. When we limited the analysis shown in tables 1 and 2 to the 227 case patients who had both PG and H. pylori/CagA tests performed, the results were very similar (data not shown).
It should be noted that 44% of the identified case patients and 56% of the potential control subjects did not participate in this study. As a result, selection bias is possible, although there is no a priori reason to suspect the presence of bias in the analyzed data. In addition, there can be inaccuracies in measuring H. pylori status and PG levels at the time of diagnosis, but such inaccuracies (which would cause misclassification) generally would lead to weakening rather than strengthening of the associations found in this study.
Our results support those of 2 case-control studies that showed a positive association of gastric cancer with a low PG I level, a low PG I/II ratio, and H. pylori seropositivity [8, 13]. In another study, a low PG I level or a low PG I/II ratiobut not H. pylori seropositivitywas associated with an increased risk for gastric cancer [14]. Other investigators found that, among subjects with low PG I/II ratios, H. pylori antibody levels and CagA seropositivity had an inverse instead of a positive association with gastric cancer [15]. Among those with normal PG I/II ratios in the same study, there was no association with H. pylori and CagA seropositivity. The authors suggested that extensive fundic mucosal atrophy caused a loss of H. pylori with a consequent reduction in antibody titer. Endoscopic studies have shown that the presence of H. pylori in the stomach diminishes with the progression of gastric damage, characterized by atrophy, intestinal metaplasia, and dysplasia [27, 28].
When we also separated our study subjects into those with low PG I/II ratios and those with normal PG I/II ratios, we found that, among those with low ratios, there was no significant association between H. pylori seropositivity and noncardia gastric cancer (OR, 1.49 [95% CI, 0.603.74]) or between CagA seropositivity and noncardia gastric cancer (OR, 1.64 [95% CI, 0.743.66]). These results support the view that atrophic gastritis reduces H. pylori colonization. Concurrently, among those with normal PG I/II ratios, there was a positive association of H. pylori seropositivity (OR, 2.65 [95% CI, 1.734.08]) and CagA seropositivity (OR, 3.58 [95% CI, 2.315.56]) with noncardia gastric cancer, which is consistent with results from prior studies showing a positive association of H. pylori seropositivity with gastric cancer risk [8, 13].
In our study, there were a total of 24 (9% of 276) patients with gastric adenocarcinoma and 14 (4% of 336) control subjects who had anti-CagA antibodies in their sera, even though their H. pylori IgG antibody test results were negative. We have recently confirmed that some persons who are culture positive for H. pylori do not produce sufficiently high antibody levels to antigens present in the whole-cell ELISA to be designated as H. pylori seropositive but do meet the threshold for anti-CagA antibodies [7]. Because of this phenomenon, it is important to test for antibodies to both CagA and H. pylori in epidemiologic studies, to determine potential associations of H. pylori (especially cagA+ strains) with conditions of interest.
The relatively high prevalence of CagA seropositivity with H. pylori seronegativity (9%) in patients with gastric cancer is consistent with earlier observations that this discordance is most pronounced in groups with substantial atrophic gastritis [29]. If confirmed, these findings suggest that CagA seropositivity is retained longer after the development of atrophic gastritis than is a diagnostic H. pylori titer.
Cohort studies have the advantage of obtaining serum samples before illnesses are diagnosed. Consequently, it is more certain in cohort studies that low PG levels or H. pylori seropositivity preceded the diagnosis of gastric cancer. Furthermore, the effect of the time interval between collection of serum specimens and diagnosis can be assessed. In a cohort study in Japan, Watanabe et al. reported that atrophic gastritis, identified by both a low PG I level and a low PG I/II ratio, was a more important risk factor for gastric cancer than was H. pylori infection [16]. Their results suggested that H. pylori infection has an indirect relationship with gastric cancer, through the development of atrophic gastritis.
Researchers in Finland reported that a low PG I level had a stronger association with gastric cancer than did elevated levels of serum antiH. pylori IgG [17]; the association of low PG I levels with cancer was stronger at shorter follow-up times, whereas the association of H. pylori with cancer became stronger with longer intervals until diagnosis. In addition, they showed that low PG I levels were associated with elevated gastric cancer risk, regardless of H. pylori antibody status.
In contrast, a cohort study in California [10] reported that low PG I levels without evidence of H. pylori infection was not associated with cancer (OR, 0.8), whereas H. pylori infection in the absence of low PG I levels was independently associated with cancer (OR, 2.4). For subjects in whom both H. pylori infection and low PG I levels were present, there was a marked increase in risk for distal gastric cancer (OR, 10.0). The differing time interval between serum collection and cancer diagnosis24 years in the California study and 13 years in the Finnish studymight explain their divergent results, since extensive replacement of fundic glands by intestinalized mucosa creates an unfavorable environment for H. pylori [27, 28]. Serum from the California subjects was more likely to be obtained before the process of intestinalization was extensive.
Our results showed that the magnitude of the ORs was similar for low PG I levels or PG I/II ratios and H. pylori or CagA seropositivity. When we studied the joint effect of PG I levels and PG I/II ratios with H. pylori and CagA status among patients with noncardia cancer, there was an accentuation in risk for those who had a low PG I level or PG I/II ratio, regardless of H. pylori/CagA status, as well as for those who had a normal PG I or PG I/II ratio and who had H. pylori or CagA seropositivity. The subjects in our study with noncardia cancer had extensive replacement of their fundic mucosa by intestinalized mucosa [30], so that our results were similar to those of the Finnish cohort study [17].
Nonetheless, the highest risk for noncardia gastric cancer was observed among those with either H. pylori or CagA seropositivity and a low PG I level or a low PG I/II ratio in both cohort studies [10, 17], as well as in the present study. Although both H. pylori tests and PG levels measured at the time of diagnosis have their limitations, these results suggest that persons seropositive for H. pylori, especially for cagA+ strains, who also develop atrophic gastritis are highly susceptible to noncardia gastric cancer. Having either factor alone also led to a smaller increase in risk; this increase may have been accentuated by false-negative results in these groups, especially in the low-PG group. H. pylori can induce a continuous inflammatory process in the stomach [12], which begins as superficial gastritis and can evolve in time into the extensive atrophic gastritis that is characterized by a low serum PG I level or a low PG I/II ratio. The risk for noncardia cancer is proportionate to the surface area of the metaplasia [31]. These results also emphasize the heightened risk for the diffuse histologic type of noncardia gastric cancer that is associated with atrophy and cagA+ H. pylori strains (table 3). This strong association has not, to our knowledge, been reported elsewhere and needs to be confirmed by others.
Separation of gastric cancer into cardia and noncardia cases is important in assessing the true association of risk factors with this disease. The incidence of noncardia gastric cancer, which has been the predominant gastric cancer in the United States, is steady or declining, whereas the incidence of cardia cancer has been increasing [32, 33]. Our results suggest that cardia cancer is associated with a low PG I level or a low PG I/II ratio but that H. pylori does not appear to contribute to its development.
In conclusion, persons with both H. pylori or CagA seropositivity and a low PG I level or a low PG I/II ratio in the present study were at high risk for noncardia cancer. Since cardia cancer was associated with low PG I levels or low PG I/II ratios but not with H. pylori seropositivity, separation of gastric cancer into cardia and noncardia subsites is important in identifying risk factors associated with these disease entities.
Acknowledgments
We thank the Castle Medical Center, Kaiser-Permanente Medical Center, Kuakini Medical Center, Queen's Medical Center, Straub Clinic and Hospital, St. Francis Medical Center, St. Francis Medical CenterWest, and Wahiawa General Hospital, for their cooperation and collaboration. We also thank Maj Earle, Leslie Round, Jean Sato, and Thomas Brown, for technical help and assistance.
References
1. Blaser MJ, Atherton J. Helicobacter pylori persistence: biology and disease. J Clin Invest 2004; 113:32133. First citation in article
2. Correa P. Human gastric carcinogenesis: a multistep and multifactorial process. Cancer Res 1992; 52:673540. First citation in article
3. Peek RM, Blaser MJ. Helicobacter pylori and gastrointestinal tract adenocarcinomas. Nat Rev Cancer 2002; 2:2837. First citation in article
5. Kuipers EJ, Perez-Perez GI, Meuwissen SGM, Blaser MJ. Helicobacter pylori and atrophic gastritis: importance of the cagA status. J Natl Cancer Inst 1995; 87:177780. First citation in article
6. Beales IL, Crabtree JE, Scunes D, Covacci A, Calam J. Antibodies to CagA protein are associated with gastric atrophy in Helicobacter pylori infection. Eur J Gastroenterol Hepatol 1996; 8:6459. First citation in article
7. Romero-Gallo J, Perez-Perez GI, Novick RP, Kamath K, Norbu T, Blaser MJ. Responses of endoscopy patients in Ladakh, India, to Helicobacter pylori whole-cell and CagA antigens. Clin Diagn Lab Immunol 2002; 9:13137. First citation in article
8. Fukuda H, Saito D, Hayashi S, et al. Helicobacter pylori infection, serum pepsinogen level and gastric cancer: a case-control study in Japan. Jpn J Cancer Res 1995; 86:6471. First citation in article
9. Samloff M, Stemmermann GN, Heilbrun LK, Nomura A. Elevated serum pepsinogen I and II levels differ as risk factors for duodenal ulcer and gastric ulcer. Gastroenterology 1986; 90:5706. First citation in article
10. Parsonnet J, Samloff M, Nelson LM, Orentreich N, Vogelman JH, Friedman GD. Helicobacter pylori, pepsinogen, and risk for gastric adenocarcinoma. Cancer Epidemiol Biomarkers Prev 1993; 2:4616. First citation in article
11. Kuipers EJ, Uyterlinde AM, Pena AS, et al. Long-term sequelae of Helicobacter pylori gastritis. Lancet 1995; 345:15258. First citation in article
12. Suerbaum S, Michetti P. Helicobacter pylori infection. N Engl J Med 2002; 347:117586. First citation in article
13. Asaka M, Kato M, Kudo M, et al. Relationship between Helicobacter pylori infection, atrophic gastritis and gastric carcinoma in a Japanese population. Eur J Gastroenterol Hepatol 1995; 7(Suppl 1):S710. First citation in article
14. Kato S, Onda M, Matsukura N, et al. Genetic polymorphisms of the cancer related gene and Helicobacter pylori infection in Japanese gastric cancer patients. Cancer 1996; 77:165461. First citation in article
15. Plummer M, Vivas J, Fauchère JL, et al. Helicobacter pylori and stomach cancer: a case-control study in Venezuela. Cancer Epidemiol Biomarkers Prev 2000; 9:9615. First citation in article
16. Watanabe Y, Kurata JH, Mizuno S, et al. Helicobacter pylori infection and gastric cancer. Dig Dis Sci 1997; 42:13837. First citation in article
17. Aromaa A, Kosunen TU, Knekt P, et al. Circulating anti-Helicobacter pylori immunoglobulin A antibodies and low serum pepsinogen I level are associated with increased risk of gastric cancer. Am J Epidemiol 1996; 144:1429. First citation in article
18. Young JL Jr, Percy CL, Asire AJ, et al. Surveillance, epidemiology, end results: incidence and mortality data 19731977. National Institutes of Health publication 81-2330. Washington, DC: Government Printing Office, 1981. First citation in article
19. Oyama N, Johnson DB. Hawaii Health Surveillance Program survey methods and procedures. Research and Statistics report 54. Honolulu: Hawaii State Department of Health, Research and Statistics Office, 1996. First citation in article
20. Hatten J. Medicare's common denominator: the covered population. Health Care Financ Rev 1980; 2:5364. First citation in article
21. Lauren P. The two histological types of gastric carcinoma: diffuse and so-called intestinal type carcinoma: an attempt at a histoclinical classification. Acta Pathol Microbiol Scand 1965; 64:3149. First citation in article
22. Ichinose M, Miki K, Furihata C, et al. Radioimmunoassay of serum group I and group II pepsinogens in normal controls and patients with various disorders. Clin Chim Acta 1982; 126:18391. First citation in article
23. Nomura AMY, Stemmermann GN, Chyou PH, Kato I, Perez-Perez GI, Blaser MJ. Helicobacter pylori infection and gastric carcinoma among Japanese Americans in Hawaii. N Engl J Med 1991; 325:11326. First citation in article
24. Blaser MJ, Perez-Perez GI, Klenthous H, et al. Infection with Helicobacter pylori strains possessing cagA is associated with an increased risk of developing adenocarcinoma of the stomach. Cancer Res 1995; 55:21115. First citation in article
25. Breslow NE, Day NE. Statistical methods in cancer research. Vol 1. The analysis of case-control studies [IARC scientific pub. no. 32]. Lyon: International Agency for Research on Cancer, 1980. First citation in article
26. Parsonnet J, Friedman GD, Vandersteen DP, et al. Helicobacter pylori infection and the risk of gastric carcinoma. N Engl J Med 1991; 325:112731. First citation in article
27. Karnes WE Jr, Samloff IM, Siurala M, et al. Positive serum antibody and negative tissue staining for Helicobacter pylori in subjects with atrophic body gastritis. Gastroenterology 1991; 101:16774. First citation in article
28. Muoz N, Kato I, Peraza S, et al. Prevalence of precancerous lesions of the stomach in Venezuela. Cancer Epidemiol Biomarkers Prev 1996; 5:416. First citation in article
29. Limburg PJ, Qiao Y-L, Mark SD, et al. Helicobacter pylori seropositivity and subsite-specific gastric cancer risks in Linxian, China. J Natl Cancer Inst 2001; 93:22633. First citation in article
30. Stemmermann GN, Nomura AMY, Kolonel LN, Goodman MT, Wilkens LR. Gastric carcinoma pathology findings in a multiethnic population. Cancer 2002; 95:74450. First citation in article
31. Stemmermann GN, Fenoglio-Preiser C. Gastric carcinoma distal to the cardia: a review of the epidemiological pathology of the precursors to a preventable cancer. Pathology 2002; 34:494503. First citation in article
32. Yang PC, Davis S. Epidemiological characteristics of adenocarcinoma of the gastric cardia and distal stomach in the United States, 19731982. Int J Epidemiol 1988; 17:2937. First citation in article
33. Blot WJ, Devesa SS, Kneller RW, Fraumeni JF Jr. Rising incidence of adenocarcinoma of the esophagus and gastric cardia. JAMA 1991; 265:12879. First citation in article