Soluble Fas (sFas) is known to play an important role in the development of cancers of various sites. To confirm whether or not the serum sFas level can be a predictor of cancer, we conducted a nested case-control study within a large-scale population-based cohort study in Japan. Serum samples were collected from 39,242 participants (13,839 men and 25,403 women) at baseline, all of whom were followed until 1997 for mortality and until 1994 for cancer incidence. Three controls were randomly selected and matched to each cancer case for gender, age and residential area. Serum values of sFas were measured by enzyme-linked immuno-adsorbent assay, using commercially available kits. The odds ratios (ORs) and 95% confidence intervals (95% CIs) were estimated using conditional logistic models, based on 798 total cancer mortality cases and their 2,353 matched controls. The risk of total cancer mortality was increased according to sFas levels, and the OR of the highest quartile compared with that of the lowest was 1.81 (95% CI; 1.40-2.34) after adjusting for smoking and drinking status, and body mass index. This positive association remained unaltered when cases were divided into 2 groups according to the observation period. Our results suggest that serum sFas has a possibility to detect people at high risk for cancer prior to diagnosis, since it increased before cancer diagnosis in those apparently healthy people. ' 2008 Wiley-Liss, Inc.Key words: soluble Fas; cancer; cohort study; nested case-control study Fas on membrane of many cells induces cell apoptosis by binding to the Fas ligand that is expressed predominantly in activated T cells. This interaction between Fas and Fas ligand is known to play an important role in the spontaneous death of cancer cells induced by the immune system. Another form of Fas, circulating soluble Fas (sFas), antagonizes the cell-surface Fas function and may offer a survival advantage to cells. Previous studies have demonstrated that sFas levels were increased in cancer patients from many sites.1-5 Furthermore, the overexpression of sFas in serum is correlated with a cancer stage and/or prognosis. 1,6,7 Given that much of the evidence is derived from case-control studies, it is plausible that an elevation of the sFas level in serum might contribute to a tumor escaping from Fas-induced apoptosis. However, it remains unclear whether individuals with high sFas levels are at an increased risk of cancer, i.e., whether or not sFas levels can predict cancer occurrence. Only 1 paper 8 has examined the relationship between sFas levels and ovarian cancer in a nested case-control study. The authors failed to find any association between serum sFas levels and ovarian cancer occurrences, which were diagnosed at an average of 5.1 years after blood was drawn. To test the hypothesis that elevated serum sFas could be a predictive marker of cancer and to identify people at risk of cancer, we examined the relationship between serum sFas levels and the risk of cancer mortality in a nested case-control study wi...