METHODS: Between 12/00 and 10/06 a total of 931 PCA patients underwent radical retropubic prostatectomy. In 162 of these patients PCA was initially diagnosed via standardized 12-core trans- rectal prostate biopsy in our institution. Pathohistological examination of the prostatectomy specimen included determination of predominant Gleason grade and volume of all tumor foci. The correlation of biopsy and prostatectomy Gleason scores (2– 6, 7, 8 –10) was calculated. In addition, the influence of small secondary tumor foci representing the highest Gleason grade in the prostatectomy specimen on the incidence of Gleason score upgrading was examined. RESULTS: Due to missing data 146 of the initial 162 patients were suitable for analysis. Based on the prostatectomy specimen Gleason scores 4 – 6 were assigned to 91 (62.3%) patients, Gleason score 7 to 42 (28.8%) and Gleason scores 8 –10 to 13 (8.9%). Upgrad- ing was observed in 24 (16.4%) and downgrading in 13 (8.9%) cases (kappa 0.49; 95% CI 0.35– 0.63). In 76/146 (52.0%) patients a small secondary tumor focus showed the highest Gleason grade in the prostatectomy specimen. The Gleason grade of the largest tumor focus was consistent with the overall highest Gleason grade in 48.0% (70/ 146) of cases. The median volume of the 76 small secondary tumor foci was significantly smaller than the median volume of the 70 largest tumor foci (0.21cc [range 0.01–2.68] vs. 0.83cc [range 0.01–9.67]; p0.0001). In the 76 patients with a small secondary tumor focus representing the highest Gleason grade upgrading was observed in 23 (30.3%) cases. In contrast, in the 70 patients in whom the largest tumor focus represented the highest Gleason grade upgrading was observed in only 1 (1.4%) case. In a multivariate logistic regression adjusting for PSA, total tumor volume and prostate volume existence of a small secondary tumor focus representing the highest Gleason grade in the prostatectomy specimen remained the only independent risk factor for Gleason score upgrading (p=0.0005). CONCLUSIONS: In PCA patients diagnosed via standardized 12-core prostate biopsy Gleason score upgrading after prostatectomy is significantly more frequent when small secondary tumor foci repre- sent the highest Gleason grade in the prostatectomy specimen. Source of Funding: None 995 ANALYSIS OF STATIN MEDICATION, GENETIC VARIATION AND PROSTATE CANCER OUTCOMES Robert Hamilton*, Joseph Vijai, New York, NY; David Gallagher, Dublin, Ireland; Caroline Savage, Jasmine Bhatia, New York, NY; Mia Gaudet, Atlanta, GA; Samson Fine, Howard Scher, Ana Dutra- Clarke, Jennifer Przybylo, Steven Lipkin, Robert Klein, Peter Scardino, Hans Lilja, James Eastham, Tomas Kirchhoff, Kenneth Offit, New York, NY INTRODUCTION AND OBJECTIVES: The relationship between statin use and prostate cancer risk and outcome remains controversial. Genetic variation in cholesterol and statin metabolism pathways appears to modify the effect of statins on cholesterol and their protective associa- tion with some cancers. However, no studies to date have examined this pharmacogenetic interaction in the setting of prostate cancer. METHODS: Blood was collected for DNA extraction from 782 men of Ashkenazi Jewish ancestry treated for prostate cancer at Memorial Sloan-Kettering Cancer Center. Statin use at the time of diagnosis was recorded. Associations between statin use and bio- chemical recurrence, castration-resistant metastasis and prostate can- cer-specific survival were analyzed using Cox proportional hazards models adjusted for age, year of diagnosis, PSA, Gleason grade, clinical stage, and treatment. Subsequent analyses will assess inter- actions between 40 single nucleotide polymorphisms (SNPs) impli- cated in cholesterol biosynthesis and the relationship between statin use and prostate cancer outcome. RESULTS: In total, 140 (18%) men were taking a statin at diagnosis. Statin users were diagnosed more recently (median year: 2002 vs. 1997, p0.001), at lower clinical stages (T1:57% vs. 43%, p=0.01) and with modestly lower PSA (6.5 vs. 7.3 ng/mL, p=0.09). No significant differences were noted in Gleason scores (p=0.95). Statin use was associated with a reduced risk of biochemical recurrence (HR 0.56, 95% CI 0.39–0.79, p=0.001; Figure), but this association was weaker after multivariate adjustment (HR 0.75, 95% CI 0.51–1.10, p=0.15). Statin users had a decreased risk of metastases (HR 0.42, 95% CI 0.18–0.96, p=0.04) that was not attenuated by multivariate adjustment, and a decreased risk of prostate cancer-specific death (HR 0.32, 95% CI 0.08 –1.33, p=0.12) that was strengthened after multi- variate adjustment (HR 0.23, 95% CI 0.05– 0.95, p=0.04). Analysis of these associations stratified by SNP genotypes is in progress and will be presented. All 40 SNPs have call rates 80%. CONCLUSIONS: In this cohort, statin use was associated with a reduced risk of biochemical recurrence, metastases and prostate cancer death. If significant genetic interactions are also observed and confirmed in other cohorts, this may identify potential pathways of relevance to prostate cancer treatment. Source of Funding: Sidney Kimmel Center for Prostate and Urologic Cancers; David H. Koch provided through the Prostate Cancer Foundation; Goldstein Research Initiative of the Program in Prevention, Control and Population Research at MSKCC. 996 COX-2 INHIBITION FOR PROSTATE CANCER CHEMOPREVENTION: RANDOMIZED PRE-PROSTATECTOMY TRIAL OF CELECOXIB VERSUS PLACEBO Jason Flamiatos*, Tomasz Beer, Nicole Janeba, Kristi Eilers, Wei Tian, Mark Garzotto, Portland, OR INTRODUCTION AND OBJECTIVES: Selective cyclooxygen- ase-2 (COX-2) inhibition has shown promise for chemoprevention of a number of tumor types. However, a role for this drug class has yet to be defined in prostate cancer prevention. The objective of the trial was to study the biologic effects of selective COX-2 inhibition on target tissues in men undergoing prostatectomy. METHODS: Patients with localized prostate cancer were ran- domized to receive either celecoxib (400 mg BID) or placebo for a minimum of 4 weeks prior to prostatectomy. Prostatectomy specimens were analyzed for levels of apoptosis (primary endpoint), prostaglan- dins, and androgen receptor levels. Effects on serum prostate specific antigen (PSA) and post-operative opioid use were also measured. RESULTS: A total of 28 out of an anticipated 44 patients were accrued to the trial. There was one myocardial infarction in the post- operative period in the celecoxib arm. For this reason and the emer- gence of safety concerns with the study drug the trial was halted. Assessment of prostatectomy specimens revealed the apoptotic index in tumor epithelial cells was 0.280.47% for the celecoxib arm versus 0.421.09% in the placebo arm (p=0.3763). The apoptosis index in benign epithelial cells was 0.180.25% for the celecoxib arm versus 0.12%0.18% in the placebo arm (p=0.42). PGE2 and androgen receptor levels were similar in both cancer and benign tissues in both Vol. 185, No. 4S, Supplement, Monday, May 16, 2011 THE JOURNAL OF UROLOGY e401