Evidence-based cancer prevention recommendations for Japanese.

A comprehensive evidence-based cancer prevention recommendation for Japanese was developed. We evaluated the magnitude of the associations of lifestyle factors and infection with cancer through a systematic review of the literature, meta-analysis of published data, and pooled analysis of cohort studies in Japan. Then, we judged the strength of evidence based on the consistency of the associations between exposure and cancer and biological plausibility. Important factors were extracted and summarized as an evidence-based, current cancer prevention recommendation: 'Cancer Prevention Recommendation for Japanese'. The recommendation addresses six important domains related to exposure and cancer, including smoking, alcohol drinking, diet, physical activity, body weight and infection. The next step should focus on the development of effective behavior modification programs and their implementation and dissemination.

Coping strategies and risk of cardiovascular disease incidence and mortality: the Japan Public Health Center-based prospective Study.

AIMS: Coping strategies may be significantly associated with health outcomes. This is the first study to investigate the association between baseline coping strategies and cardiovascular disease (CVD) incidence and mortality in a general population cohort. METHODS AND RESULTS: The Japan Public Health Center-based prospective Study asked questions on coping in its third follow-up survey (2000-04). Analyses on CVD incidence and mortality included 57 017 subjects aged 50-79 without a history of CVD and who provided complete answers on approach- and avoidance-oriented coping behaviours and strategies. Cox regression models, adjusted for confounders, were used to determine hazard ratios (HRs) according to coping style. Mean follow-up time was 7.9 years for incidence and 8.0 years for mortality.The premorbid use of an approach-oriented coping strategy was inversely associated with incidence of stroke (HR = 0.85; 95% CI, 0.73-1.00) and CVD mortality (HR = 0.74; 95% CI, 0.55-0.99). Stroke subtype analyses revealed an inverse association between the approach-oriented coping strategy and incidence of ischaemic stroke (HR = 0.79; 95% CI, 0.64-0.98) and a positive association between the combined coping strategy and incidence of intra-parenchymal haemorrhage (HR = 2.03; 95% CI, 1.01-4.10). Utilizing an avoidance coping strategy was associated with increased mortality from ischaemic heart disease (IHD) only in hypertensive individuals (HR = 3.46; 95% CI, 1.07-11.18). The coping behaviours fantasizing and positive reappraisal were associated with increased risk of CVD incidence (HR = 1.24; 95% CI, 1.03-1.50) and reduced risk of IHD mortality (HR = 0.63; 95% CI, 0.40-0.99), respectively. CONCLUSION: An approach-oriented coping strategy, i.e. proactively dealing with sources of stress, may be associated with significantly reduced stroke incidence and CVD mortality in a Japanese population-based cohort.

Soy food and isoflavone intake and endometrial cancer risk: the Japan Public Health Center-based prospective study.

OBJECTIVE: Compared with western populations, the consumption of soy foods among Japanese is very high and the incidence of endometrial cancer very low. We evaluated the association of soy food and isoflavone intake with endometrial cancer risk in Japanese women. DESIGN: Prospective cohort study. SETTING: Ten public health centre areas in Japan. POPULATION: Forty nine thousand one hundred and twenty-one women of age 45-74 years who responded to a 5-year follow-up survey questionnaire. METHODS: Intakes of soy foods as well as other covariates were assessed in 1995-1998 by a self-administered food frequency questionnaire. Cox proportional hazards regression models were used to estimate hazard ratios (HR) and 95% confidence intervals (CI). MAIN OUTCOME MEASURE: Incidence of endometrial cancer. RESULTS: During an average of 12.1 years of follow up, 112 newly diagnosed endometrial cancer cases were identified. Energy-adjusted intakes of soy food and isoflavone were not associated with the risk of endometrial cancer. The multivariate-adjusted HR per 25 g/day increase in the intake of soy food was 1.02 (95% CI 0.94-1.10), and the corresponding value for isoflavone intake per 15 mg/day was 1.01 (95% CI 0.84-1.22). CONCLUSION: In this population-based prospective cohort study of Japanese women, we observed no evidence of a protective association between soy food or isoflavone intake and endometrial cancer risk.

Rice, bread, noodle and cereal intake and colorectal cancer in Japanese men and women: the Japan Public Health Center-based prospective Study (JPHC Study).

BACKGROUND: Colorectal cancer (CRC) incidence rate increased rapidly in Japan between the 1950s and 1990s. We examined the association between rice intake and CRC risk in comparison with bread, noodles and cereal among Japanese adults enrolled in the Japan Public Health Center-based prospective Study. METHODS: A total of 73,501 Japanese men and women were followed-up from 1995 to 1999 until the end of 2008 for an average of 11 years. During 801,937 person-years of follow-up, we identified 1276 incident cases of CRC. Hazard ratios (HRs) and 95% confidence intervals (95% CIs) of CRC for rice, noodle, bread and cereal intake were calculated by Cox proportional hazards model. RESULTS: Overall, no significant association was observed for the highest quartile of rice intake compared with the lowest and the risk of CRC and its subsites in men (HR, 0.77; 95% CI, 0.56-1.07) and women (HR, 1.10; 95% CI, 0.71-1.68). However, a non-significant inverse trend was observed between rice intake and rectal cancer in men. No clear patterns of association were observed in bread, noodle and cereal intake. CONCLUSION: Our findings suggest that the consumption of rice does not have a substantial impact on the risk of CRC in the Japanese population.

Association of vegetable and fruit intake with gastric cancer risk among Japanese: a pooled analysis of four cohort studies.

BACKGROUND: Prospective evidence is inconsistent regarding the association between vegetable/fruit intake and the risk of gastric cancer. METHODS: In an analysis of original data from four population-based prospective cohort studies encompassing 191 232 participants, we used Cox proportional hazards regression to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) of gastric cancer incidence according to vegetable and fruit intake and conducted a meta-analysis of HRs derived from each study. RESULTS: During 2 094 428 person-years of follow-up, 2995 gastric cancer cases were identified. After adjustment for potential confounders, we found a marginally significant decrease in gastric cancer risk in relation to total vegetable intake but not total fruit intake: the multivariate-adjusted HR (95% CI; P for trend) for the highest versus the lowest quintile of total vegetable intake was 0.89 (0.77-1.03; P for trend = 0.13) among men and 0.83 (0.67-1.03; P for trend = 0.40) among women. For distal gastric cancer, the multivariate HR for the highest quintile of total vegetable intake was 0.78 (0.63-0.97; P for trend = 0.02) among men. CONCLUSIONS: This pooled analysis of data from large prospective studies in Japan suggests that vegetable intake reduces gastric cancer risk, especially the risk of distal gastric cancer among men.

Body mass index and breast cancer risk in Japan: a pooled analysis of eight population-based cohort studies.

BACKGROUND: A positive association between body mass index (BMI) and breast cancer risk among postmenopausal women has been reported, and a weak inverse association has been suggested among premenopausal women from studies in the Western population. The effects of BMI on breast cancer have remained unclear among the Asian population, especially in premenopausal women. METHODS: We assessed the associations between BMI and breast cancer incidence by a pooled analysis from eight representative large-scale cohort studies in Japan. Cancer incidence was mainly confirmed through regional population-based cancer registries and/or through active patient notification from major local hospitals. Breast cancer was defined as code C50 according to ICD10. Pooled estimates of the hazard ratios (HRs) and 95% confidence interval (CIs) for breast cancer were calculated using random-effects models. RESULTS: Analytic subjects were 183 940 women, 1783 of whom had breast cancer during 2 194 211 person-years of follow-up. A positive association between BMI and the risk of postmenopausal breast cancer was observed (trend P<0.001). The HRs for premenopausal breast cancer were 1.05 (95% CI 0.56-1.99), 1.07 (95% CI 0.76-1.52), 0.91 (95% CI 0.64-1.30), 1.15 (95% CI 0.76-1.73), 1.45 (95% CI 0.71-2.94), and 2.25 (95% CI 1.10-4.60), respectively, in BMIs of <19, 19 to <21, 21 to <23, 25 to <27, 27 to <30, and ≥30 kg/m2. These results were not substantially altered after excluding the patients who were diagnosed with breast cancer in the first 2 years of follow-up. CONCLUSIONS: The increased risk of postmenopausal breast cancer among women with higher BMIs was confirmed in Japanese. A borderline-significant positive association between BMI and premenopausal breast cancer was observed, suggesting that body mass in Asian women might have opposite effects on breast cancer compared with Western women.

Attributable causes of cancer in Japan in 2005--systematic assessment to estimate current burden of cancer attributable to known preventable risk factors in Japan.

BACKGROUND: To contribute to evidence-based policy decision making for national cancer control, we conducted a systematic assessment to estimate the current burden of cancer attributable to known preventable risk factors in Japan in 2005. METHODS: We first estimated the population attributable fractions (PAFs) of each cancer attributable to known risk factors from relative risks derived primarily from Japanese pooled analyses and large-scale cohort studies and the prevalence of exposure in the period around 1990. Using nationwide vital statistics records and incidence estimates, we then estimated the attributable cancer incidence and mortality in 2005. RESULTS: In 2005, ≈ 55% of cancer among men was attributable to preventable risk factors in Japan. The corresponding figure was lower among women, but preventable risk factors still accounted for nearly 30% of cancer. In men, tobacco smoking had the highest PAF (30% for incidence and 35% for mortality, respectively) followed by infectious agents (23% and 23%). In women, in contrast, infectious agents had the highest PAF (18% and 19% for incidence and mortality, respectively) followed by tobacco smoking (6% and 8%). CONCLUSIONS: In Japan, tobacco smoking and infections are major causes of cancer. Further control of these factors will contribute to substantial reductions in cancer incidence and mortality in Japan.

Association between body mass index and the colorectal cancer risk in Japan: pooled analysis of population-based cohort studies in Japan.

BACKGROUND: Obesity has been recognized as important risk factors for colorectal cancer. However, limited evidence is available on colorectal cancer and body mass index (BMI) in Asian population. METHODS: We conducted a pooled analysis of eight population-based prospective cohorts studies in Japan with more than 300,000 subjects to evaluate an impact of obesity in terms of BMI on colorectal cancer risk with unified categories. We estimated summary hazard ratio (HR) by pooling of study-specific HR for BMI categories with random effect model. RESULTS: We found a significant positive association between BMI and colorectal cancer risk in male and female. Adjusted HRs for 1 kg/m(2) increase were 1.03 [95% confidence interval (CI) 1.02-1.04] for males and 1.02 (95% CI 1.00-1.03) for females. The association was stronger in colon, especially in proximal colon, relative to rectum. Males showed a stronger association than females. Population attributable fraction for colorectal cancer by BMI ≥ 25 kg/m(2) was 3.62% (95% CI 1.91-5.30) for males and 2.62% (95% CI 0.74-4.47) for females. CONCLUSIONS: We found significant association between BMI and colorectal cancer risk by pooling of data from cohort studies with considerable number of subjects among Japanese population. This information is important in cancer control planning, especially in Asian population.

Gender difference in the association of insulin and the insulin-like growth factor axis with colorectal neoplasia.

OBJECTIVE: Accumulating evidence has implicated insulin and the insulin-like growth factor (IGF) axis in colorectal carcinogenesis. Of interest, adiposity is likely to impose a greater risk on men than on women, which indicates that the association of insulin and the IGF axis with colorectal neoplasia may differ by gender. However, epidemiological evidence for this possible gender difference is limited to date. METHODS: We measured plasma concentrations of C-peptide, IGF-I and IGF-binding proteins (IGFBPs) 1 and 3 in 1520 healthy volunteer examinees who underwent total colonoscopy between February 2004 and February 2005, and cross-sectionally investigated the association of these biomarkers with colorectal adenoma by gender. An unconditional logistic regression model was used to estimate odds ratios (ORs) and their 95% confidence intervals (CIs) for colorectal adenoma after adjustment for potential confounders. RESULTS: We observed a positive association of C-peptide and IGF-I (P (trend)<0.001 and 0.02, respectively) and an inverse association of IGFBP-1 (P (trend)=0.002) with colorectal adenoma in men. Adjusted ORs of colorectal adenoma for the highest compared with the lowest quartile were also statistically significant for C-peptide (OR: 2.62, 95% CI: 1.71-4.01), IGF-I (OR: 1.63, 95% CI: 1.08-2.46) and IGFBP-1 (OR: 0.49, 95% CI: 0.32-0.75). In contrast, no measurable association was seen in women. Corresponding ORs for C-peptide, IGF-I and IGFBP-1 were 0.98 (95% CI: 0.56-1.71), 0.79 (95% CI: 0.44-1.43) and 1.05 (95% CI: 0.60-1.86), respectively. The gender difference was statistically significant for C-peptide (P (interaction)=0.03) and marginally significant for IGF-I and IGFBP-1 (P (interaction)=0.14 and 0.12, respectively). CONCLUSION: Our observations suggest that insulin and the IGF axis act differently by gender in colorectal carcinogenesis, at least in its early stage. The findings of this study further our understanding of the complexities of the gender difference in the association between adiposity and colorectal neoplasia.

Green tea consumption and gastric cancer in Japanese: a pooled analysis of six cohort studies.

BACKGROUND: Previous experimental studies have suggested many possible anti-cancer mechanisms for green tea, but epidemiological evidence for the effect of green tea consumption on gastric cancer risk is conflicting. OBJECTIVE: To examine the association between green tea consumption and gastric cancer. METHODS: We analysed original data from six cohort studies that measured green tea consumption using validated questionnaires at baseline. Hazard ratios (HRs) in the individual studies were calculated, with adjustment for a common set of variables, and combined using a random-effects model. RESULTS: During 2 285 968 person-years of follow-up for a total of 219 080 subjects, 3577 cases of gastric cancer were identified. Compared with those drinking <1 cup/day, no significant risk reduction for gastric cancer was observed with increased green tea consumption in men, even in stratified analyses by smoking status and subsite. In women, however, a significantly decreased risk was observed for those with consumption of > or =5 cups/day (multivariate-adjusted pooled HR = 0.79, 95% confidence interval (CI) = 0.65 to 0.96). This decrease was also significant for the distal subsite (HR = 0.70, 95% CI = 0.50 to 0.96). In contrast, a lack of association for proximal gastric cancer was consistently seen in both men and women. CONCLUSIONS: Green tea may decrease the risk of distal gastric cancer in women.

Dietary isoflavone and the risk of colorectal adenoma: a case-control study in Japan.

We conducted a case-control study in a Japanese population to investigate the association between dietary isoflavone intake and the risk of colorectal adenoma. Participants who underwent magnifying colonoscopy with dye spreading as part of a cancer screening programme responded to a self-administered questionnaire, which included lifestyle information and intake of 145 food items, before the colonoscopy. A total of 721 case and 697 control subjects were enrolled. Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated using unconditional logistic regression models. We found a significant inverse association between dietary isoflavone intake and the risk of colorectal adenoma in men and women combined. However, the inverse association was not linear; rather, all quartiles above the first showed a similar decrease in risk, with multivariable-adjusted ORs and 95% CIs compared with the lowest quartile of 0.77 (0.57-1.04), 0.76 (0.56-1.02) and 0.70 (0.51-0.96) in the second, third and highest quartiles, respectively (P for trend=0.03). Of interest, the observed association was more prominent in women than in men. The observed ceiling effect associated with higher isoflavone intake suggests that a lower intake of dietary isoflavone might be associated with an increased risk of colorectal adenoma in Japanese populations.

Plasma 25-hydroxy vitamin D and subsequent prostate cancer risk in a nested Case-Control study in Japan: The JPHC study.

BACKGROUND/OBJECTIVES: Although vitamin D has been experimentally reported to inhibit tumorigenesis, cell growth and prostate cancer invasion, epidemiologic data regarding prostate cancer risk are inconsistent, and some studies have suggested positive but nonsignificant associations. Further, the impact of vitamin D on prostate cancer between Western and Japanese populations may differ due to different plasma vitamin D levels. SUBJECTS/METHODS: We performed a nested case-control study within the Japan Public Health Center-based Prospective (JPHC) Study in 14,203 men (40-69 years) who answered a self-administered questionnaire at baseline (1990-1994) and gave blood samples, and were followed until 2005. We identified 201 prostate cancers which are newly diagnosed during follow-up (mean 12.8 years). We selected two matched controls for each case from the cohort. We used a conditional logistic regression model to estimate the odds ratios (ORs) and 95% confidence intervals (CIs) for prostate cancer with respect to levels of 25-hydroxy vitamin D (25(OH)D) in plasma. RESULTS: We did not observe statistically significant association between 25(OH)D level and total prostate cancer (multivariate OR=1.13 (95%CI=0.66-1.94, Ptrend=0.94) for the highest versus lowest tertile) However, 25(OH) levels were slightly positively associated with advanced cancer. The results remained substantially unchanged after stratification by intake of fish or calcium intake. CONCLUSIONS: 25(OH)D level showed no association with overall prostate cancer among Japanese men in this large cohort.

Effect of vitamin C on common cold: randomized controlled trial.

OBJECTIVE: To investigate the relationship between the common cold and vitamin C supplementation. DESIGN: A double-blind, 5-year randomized controlled trial. SETTING: A village in Akita prefecture, one of the regions in Japan with the highest mortality from gastric cancer. SUBJECTS: Participants in annual screening programs for circulatory diseases conducted under the National Health and Welfare Services Law for the Aged, and diagnosed as having atrophic gastritis. Of the 439 eligible subjects, 144 and 161 were assigned to receive 50 or 500 mg of vitamin C, respectively, after protocol amendment. During the supplementation phase, 61 dropped out, and 244 completed the trial. INTERVENTION: Daily vitamin C supplementation of 50 mg (low-dose group) or 500 mg (high-dose group). RESULTS: Total number of common colds (per 1000 person-months) was 21.3 and 17.1 for the low- and high-dose groups, respectively. After adjustment for several factors, the relative risks (95% confidence interval (CI)) of suffering from a common cold three or more times during the survey period was 0.34 (0.12-0.97) for the high-dose group. No apparent reduction was seen for the severity and duration of the common cold. CONCLUSION: A randomized, controlled 5-year trial suggests that vitamin C supplementation significantly reduces the frequency of the common cold but had no apparent effect on the duration or severity of the common cold. However, considering several limitations due to protocol amendment, the findings should be interpreted with caution.

Alcohol consumption and severity of angiographically determined coronary artery disease in Japanese men and women.

The relation of alcohol consumption to the severity of coronary atherosclerosis was examined among 323 men and 220 women who underwent coronary arteriography. Severity of coronary atherosclerosis was assessed by the number of vessels obstructed > or =75% in diameter and Gensini's severity score. Alcohol consumption was divided into 5 categories in men (never, past, 1-24, 25-49, and > or =50 ml per day) and 3 categories in women (never, past, and current). Among men, odds ratios of severe stenosis (multiple-vessel disease or Gensini's score >15) decreased substantially and significantly in all current drinking categories but without dose-response effect. There was a weak, inverse association of current alcohol consumption with one-vessel disease, but not with moderate stenosis in terms of Gensini's score (< or =15). Past drinkers showed a fairly large, but statistically nonsignificant, decrease in the odds ratios of not only severe stenosis but also of moderate stenosis. Among women, current drinkers showed a small, statistically nonsignificant decrease in the risk of severe stenosis in terms of Gensini's score. These associations with alcohol use did not change after adjustment for known coronary risk factors. The present findings add to evidence that alcohol drinking confers protection against coronary atherosclerosis.

Relation between green tea consumption and the severity of coronary atherosclerosis among Japanese men and women.

PURPOSE: To examine the relation between green tea consumption and arteriographically determined coronary atherosclerosis. METHODS: Study subjects were 512 patients (302 men and 210 women) aged 30 years or older who underwent coronary arteriography for the first time at four hospitals in Fukuoka City or one hospital in an adjacent city between September 1996 and August 1997. Lifestyle characteristics including green tea consumption were ascertained before arteriography by a questionnaire supported with interview. RESULTS: 117 men (38.7%) and 50 women (23.8%) had significant stenosis of one or more coronary arteries. Green tea consumption tended to be inversely associated with coronary atherosclerosis in men, but not in women. An evident, protective association between green tea and coronary atherosclerosis was observed in a subgroup of 262 men excluding those under dietary or drug treatment for diabetes mellitus. In this subgroup, after adjustment for traditional coronary risk factors and coffee, odds ratios of significant stenosis for consumption of 2-3 cups and 4 or more cups per day were 0.5 (95% confidence interval 0.2-1.2) and 0.4 (0.2-0.9), respectively, as compared with a consumption of one cup per day or less. CONCLUSIONS: The results indicate that green tea may be protective against coronary atherosclerosis at least in men.

Job strain, Type A behavior pattern, and the prevalence of coronary atherosclerosis in Japanese working men.

OBJECTIVE: To examine the relation of type A behavior pattern and job strain to angiographically documented coronary stenosis. METHODS: Subjects were 197 male Japanese patients with a full-time job. A questionnaire-based interview elicited psychosocial and other factors. Type A behavior pattern was measured by 12 questions, and job strain by the method of Karasek. Significant coronary stenosis was defined when a 75% or greater luminal narrowing occurred at one or more major coronary arteries or when a 50% or greater narrowing occurred at the left main artery. Logistic regression analysis was used to calculate odds ratio (OR) and 95% confidence interval (CI) with adjustment for traditional coronary risk factors and job type. RESULTS: Type A behavior pattern was related to a statistically non-significant lower prevalence of the coronary stenosis especially in the absence of job strain (adjusted OR 0.6, 95% CI 0.3-1.2). Job strain was non-significantly associated with a modestly increased prevalence of coronary stenosis (OR 1.7, 95% CI 0.6-5.2). CONCLUSION: These findings suggest that both the behavioral pattern and psychosocial work environment may be related to coronary artery stenosis.

Plasma vitamin D and risk of colorectal cancer: the Japan Public Health Center-Based Prospective Study.

We investigated the association between plasma 25(OH)D and the subsequent colorectal cancer incidence risk by a nested case-control study in The Japan Public Health Center-based Prospective Study, covering 375 newly diagnosed cases of colorectal cancer from 38 373 study subjects during a 11.5-year follow-up after blood collection. Two controls were matched per case on sex, age, study area, date of blood draw, and fasting time. In a conditional logistic regression model with matched pairs adjusted for smoking, alcohol consumption, body mass index, physical exercise, vitamin supplement use, and family history of colorectal cancer, plasma 25(OH)D was not significantly associated with colorectal cancer in men or in women. However, the lowest category of plasma 25(OH)D was associated with an elevated risk of rectal cancer in both men (odds ratio (OR), 4.6; 95% confidence interval (CI), 1.0-20) and women (OR, 2.7, 95% CI, 0.94-7.6), compared with the combined category of the other quartiles. Our results suggest that a low level of plasma 25(OH)D may increase the risk of rectal cancer.

Association of body mass index and height with risk of prostate cancer among middle-aged Japanese men.

In a population-based prospective study of 49 850 Japanese men, body mass index and height were not significantly associated with risk of prostate cancer (311 cases), although small positive effects could not be ruled out in advanced cases (91 cases).

Case-control study of nonfatal myocardial infarction in relation to selected foods in Japanese men and women.

Most studies of diet and coronary heart disease (CHD) have focused on constituents rather than on whole foods. The present study examined the relationship of selected foods to nonfatal acute myocardial infarction (AMI) in Japan, with special reference to vegetables, fruits, fish, and tofu. A total of 660 cases with their first episode of AMI aged 40-79 years living in Fukuoka City or adjacent areas and 1,277 controls matched for age, sex, and residence were surveyed on lifestyle, including dietary factors. Participation rates were 87% of cases and 52% of controls. Consumption frequencies of 19 food/beverages items and daily amounts of 4 items were ascertained by interview. The final analysis was done with 632 cases and 1,214 controls. Although consumption of vegetables showed no clear association with the risk of AMI, fruit consumption appeared to reduce the risk of AMI in both men and women. The results also suggested that fish consumption was related to a decreased risk of AMI in men, although the trend was not statistically significant. In women only, tofu consumption was inversely related to the risk of AMI; relative risks for eating tofu <2, 2-3, and 4+ times per week were 1.0, 0.8, and 0.5, respectively, after adjustment for non-dietary factors (p for trend = 0.01). Further adjustment for consumption of fruit, fish and tofu did not alter the findings generally. The findings suggest that, in women at least, tofu consumption may be protective against the risk of AMI. Further studies are needed to corroborate the relationship of consumption of fish and fruit to AMI risk in Japanese men and women.

Salt and salted food intake and subsequent risk of gastric cancer among middle-aged Japanese men and women.

Evidence on the association between salt intake and gastric cancer is sparse, especially in prospective studies. We conducted a population-based prospective study in Japan, where the majority of men has been infected with Helicobacter pylori. A total of 18 684 men and 20 381 women aged 40-59 years who reported their dietary habits and did not report any serious disease at baseline were followed from 1990 to 2001. A total of 486 cases, 358 men and 128 women, with histologically confirmed gastric cancer were documented among them. The quintile category of salt intake was dose-dependently associated with gastric cancer risk in men after adjusting for potential confounding factors (P for trend <0.001), while a trend was not clear in women (P for trend=0.48). Although stratification by study area, with varied salt intake and gastric cancer incidence, attenuated the observed clear associations with salt and salted foods, the frequency categories of highly salted foods such as salted fish roe and salted fish preserves were strongly associated with the risk in both sexes. Restriction of salt and salted food intake is a practical strategy to prevent gastric cancer in areas with high risk.