Background: Habitual intake of bevarages with Lactobacillus casei strain Shirota (BLS), widely taken as fermented milk products in Japan, was shown to have an association with a reduced risk of bladder cancer in a case-control study (Ohashi, 2001). Although associations between intake of isoflavones and BC risk have been inconsistent in western epidemiological studies, a large Japanese cohort study (Yamamoto,2003) showed an association of higher intake with a reduced risk of BC.Subjects and Methods: Female primary BC pts with stage 0-1, age 40-55 were selected as cases within one year from diagnosis; matched (1:2) controls by age and residential area were randomly selected from residence registry and invited to the study. Cases and controls were interviewed by trained interviewers, blinded to identification of case or control, and answered validated food frequency questionnaires on BLS and isoflavones (in major soy products). The necessary number of cases was calculated as 355 based on hypothesized odds ratio (OR) 0.55, two-sided typeI error 0.05, typeII error 0.20 and prevalence of BLS habitual intake 0.15). Statistical analysis plan for confirmatory analyses was fixed prior to data-base lock; OR and 95% confidence intervals (CI) were calculated using conditional logistic regression taking account of age-residence matching and only age matching with adjustment of possible confounding factors including age; p-values are two-sided. This study was approved by the IRBs of the Public Health Research Foundation, the sponsor of the study, and participating hospitals.Results: Total of 368 pts were invited to the study from 14 hospitals and 321 pts accepted the invitation; 8166 invitation letters were sent to candidates of controls and 884 women accepted. Excluding ineligible pts and mismatched/unapproachable controls, data consisting of 306 cases and 662 controls were analyzed. Habitual intake was averaged over 3 life-time periods (age 11-12, age 20 and 10-15 years ago). The prevalence of BLS intake with 4 or more times per week was 11.1% in cases and 16.2% in controls; OR was 0.66 (CI 0.43-1.01; p=0.054), 0.65 (0.42-1.00; p=0.054), before and after adjustment, respectively. As regards isoflavones, OR of Q4 (highest quartile; >83.1 mg/day), Q3 (56.7-83.1) and Q2 (39.3-56.7) to Q1 (<39.3) were 0.59, 0.67 and 0.82 (trend p=0.006), 0.54, 0.61 and 0.78 (trend p=0.002), respectively, before and after adjustment. Low educational level (under university level; OR=1.7(BLS), 1.6(isoflavones)), experience of benign tumor (OR=3.0,3.1) and family history (OR=2.2,2.2) were statistically significant risk factors in adjusted analyses. No interaction was observed between intake of BLS and isoflavones.Discussion: Associations of habitual intake of BLS and isoflavones with a reduced risk of BC were confirmed for Japanese middle aged women, probably due to high background prevalence of BLS and soy bean products in Japanese. Statistical confirmation of established risk factors (benign tumor and family history) may validate the quality of this study. The observed low BC risk in women with high education level may be due to their high acceptance rate based on appreciation of meaning of the study.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 1050.

2009