Despite the fact that one out of ten men in the U.S. has prostate cancer (PCa) during their lifetime (there are approximately 165,000 incident cases and 35,000 deaths from the disease annually) relatively little sound knowledge is available on the basic epidemiology of the condition. Work to date is beset by methodologic problems, for example: 1) unrepresentative convenience samples (often of patients); 2) imprecise definition and determination of prostate cancer; 3) variability in field methods and laboratory techniques; 4) single discipline studies which overlook multifactorial contributions to PCa; 5) failure to follow well characterized subjects over time, starting at earlier ages (say, 40 years); and 6) small sample sizes which provide inadequate statistical power to adequately control important covariates. This proposed retrospective cohort study will utilize data from the Massachusetts Male Aging Study (MMAS), the largest representative male endocrine data base presently available, to identify predictors of PCa. Men originally interviewed in 1987-89 (T/1) will be recontacted in 1996-97 (T/2). Two cohorts will be distinguished: (a) those with prostate cancer (n equals 50); (b) those without prostate cancer (n equals 1650). Next- of-kin of decreased subjects will be interviewed, death certificates obtained, and the Massachusetts Cancer Registry checked. Already diagnosed non-fatal cases will be interviewed and permission sought to review medical records. A third group identified through interview and PSA as at-risk to PCa will be invited to visit the MGH, where a standard urologic workup will be conducted. These methods should approach 100 percent ascertainment of PCa cases. Statistical power calculations indicate that sufficient numbers will be available to answer the major questions of interest. Using longitudinal data (8 years of follow-up) from a large representative sample, covering a broad age span (40-70 years at T/1), this study will provide basic epidemiologic information concerning PCa: the role of 17 different hormones, life styles (dietary intake, physical activity, sexual activity, cigarette and alcohol usage), fat topology, medications, demographics, occupation, medical history and surgeries, family history, among other variables. Multivariate techniques and sufficient numbers will permit control for collinearity and confounding. Apart from significant epidemiologic contributions this work will produce the first quantifiable risk assessment instrument for PCa based on sound epidemiologic data. Such an instrument will permit earlier identification of men likely to eventually develop PCa, which is a stated purpose of the RFA.
Showing the most recent 10 out of 37 publications
