Journal of Clinical Epidemiology 56 (2003) 1–9 0895-4356/03/$ – see front matter © 2003 Elsevier Science Inc. All rights reserved. PII: S0895-4356(02)00534-6 COMMENTARY Gastric cancer epidemiology and risk factors Jon R. Kelley a , John M. Duggan b,c, * a Department of Veterans’ Affairs, Commonwealth of Australia, G.P.O. Box 651, Brisbane Queensland 4001, Australia b Repatriation Medical Authority, Commonwealth of Australia, G.P.O. Box 1014, Brisbane Queensland, Australia c Discipline of Clinical Epidemiology and Biostatistics, Faculty of Medicine and Health Science, University of Newcastle, Pacific Street, Newcastle, NSW 2300, Australia Received 30 May 2001; received in revised form 26 June 2002; accepted 19 July 2002 Abstract We performed a detailed analysis of the epidemiology of gastric carcinoma, based upon a review of the literature in English. The analysis reveals many puzzling features. There has been a steady fall in the incidence of gastric carcinoma in most societies studied, but a more recent steady rise in the incidence of adenocarcinoma of the cardia and lower esophagus, largely confined to White males. Although the evidence for a major role for Helicobacter pylori (H. pylori) in the etiology of gastric corpus cancer is compelling; in Western society, it probably accounts for fewer than half the cases. The relative roles of dietary constituents such as salt and nitrites and the phenotyping of H. pylori in causation and the beneficial effects of a high fruit and vegetable diet and an affluent lifestyle, for all of which there is some evidence, are yet to be quantified. © 2003 Elsevier Science Inc. All rights reserved. Keywords: Gastric cancer; Epidemiology; H. pylori; Heredity; Diet 1. Introduction The second half of the 20th century has seen a sharp world- wide decline in both the incidence and mortality of gastric cancer. Despite this, the condition remains the world’s second leading cause of cancer mortality behind lung cancer. It has been estimated that there will have been more than 870,000 deaths from the disease in the year 2000, accounting for ap- proximately 12% of all cancer deaths [1–3]. Gastric cancer has attracted much attention from epide- miologic investigators over recent years, particularly with the emergence of H. pylori as a risk factor for the condition. This has lead to an improved understanding of the etiology and pathogenesis of gastric cancer and raised the possibility of active prevention of the disease. Differences in exposures to H. pylori and a range of other environmental factors probably account for much of the variations seen in the inci- dence of gastric cancer over time and between populations. We begin with a discussion of the pathology of gastric can- cer, then consider descriptive epidemiology, and finally re- view the evidence concerning possible etiologic factors. The evidence described in this report was identified from the En- glish language literature by searching the Medline database. 2. Pathology Approximately 90% of stomach cancers are adenocarci- nomas. Non-Hodgkin’s lymphomas and leiomyosarcomas make up most of the remaining 10%. Adenosquamous, squamous, and undifferentiated carci- nomas also occur but are rare. Other very rare malignant primary tumors of the stomach include choriocarcinomas, carcinoid tumors, rhabdomyosarcomas, and hemangioperi- cytomas. Kaposi’s sarcoma, in association with the acquired immunodeficiency syndrome, has also been reported [4]. Adenocarcinomas may be subdivided histologically into two categories: (1) a well-differentiated or intestinal type, with cohesive neoplastic cells forming gland-like structures that frequently ulcerate; and (2) a diffuse type in which cell cohe- sion is absent, resulting in infiltration and thickening of the stomach wall without the formation of a discrete mass. The in- testinal type is more common in males and older age groups. Diffuse carcinomas are relatively more common in younger age groups and have a more equal male-to-female ratio [5]. 3. Descriptive epidemiology One of the notable features of the descriptive epidemiology concerning gastric cancer is that it establishes some clear dis- tinctions between cancer localized to the gastric cardia and cancer of the rest of the stomach, as discussed below. * Corresponding author. Princeton Medical Centre, 60 Lindsay Street, Hamilton, NSW, 2303, Australia. Tel.: 61 2 49 61 2090; fax: 61 2 49 62 3443. E-mail address: duggan@hunterlink.net.au (J.M. Duggan).