Shorter length androgen receptor CAG repeat (AR CAGr) and GGN repeat (AR GGNr) genotypes and high dietary fat intake are potential risk factors for prostate cancer. A recent study (Zitzmann et al. Diabetologia 46:31-9;2003) has also shown a positive correlation between AR CAGr genotypes and markers of obesity such as body fat mass, leptin and insulin. We stratified the AR genotypes into two groups by base pair length: <21 and >20 for CAGr, and <16 and >15 for GGNr; the shorter repeat length was considered the risk genotype. Using logistic regression and a case-only approach, we tested for gene-environment interaction by estimating the association between genotype and dietary fat measures (percent of total calories from different types of fat) among 455 incident Caucasian and African American prostate cancer cases, adjusting for age, caloric intake, and race (where appropriate). For the entire sample, a less than multiplicative relationship between AR CAGr risk genotype and percent of calories from polyunsaturated fatty acids (PFA) intake was observed (p=0.015). After stratifying by race, this effect was only observed among the Caucasian men (p=0.020). When the cases were stratified into low and high prostate cancer aggressiveness categories, the effect was observed only in the highly aggressive group (p=0.003). In the entire sample, logistic regression results suggested a possible greater than multiplicative relationship between AR GGNr risk genotype and percent of calories from total fatty acids (TFA, p=0.094) and from saturated fatty acids (SFA, p=0.073), however neither increase was statistically significant. After stratifying by race, a greater than multiplicative relationship between AR GGNr and TFA (p=0.047) and monounsaturated fatty acids (MFA, p=0.035) was observed among African Americans; no interactions between AR GGNr and fat measures were observed among Caucasians. When stratified on aggressiveness, there was a suggestion of a greater than multiplicative relationship between AR GGNr and TFA (p=0.063), and AR GGNr and PFA (p=0.0631), and a statistically significant interaction between AR GGNr and saturated fatty acids (p=0.016) among the cases with highly aggressive disease. Our findings suggest that the AR CAGr and GGNr genotypes and dietary fat intake may be part of a complex interaction in the androgen-dependent pathway of prostate carcinogenesis.
[Proc Amer Assoc Cancer Res, Volume 46, 2005]