Prostate cancer survival among statin users after prostatectomy in a Finnish nationwide cohort

Background: Improved prostate cancer (PCa) survival by statin use has been reported among PCa patients managed with radiation or androgen deprivation therapy (ADT), while results are controversial for men managed surgically. We evaluate the association between cholesterol-lowering medication with initiation of ADT and disease-specific death among PCa cases who underwent radical prostatectomy in Finland between 1995 and 2013. Methods: The study cohort included 14 424 men with PCa who underwent radical prostatectomy in Finland between 1995 and 2013. Cases were identified from national hospital discharge registry. Clinical data were amended from patient files of the treating hospitals. Information on co-morbidities, additional radiation- or chemotherapy, and causes of deaths were collected from national registries. Personal-level data on medication use during 1995-2014 were gathered from national prescription database. Registry linkages were carried out using personal identification number. Lipid-lowering drugs were categorized into statins and non-statin drugs. Risk of PCa death and initiation of ADT was analyzed using Cox-regression model with adjustment for age, radiation therapy, chemotherapy, co-morbidities and other drug use. Statin use was analyzed as time-dependent variable. Delayed risk associations were evaluated in lag-time analysis. Results: Compared to non-users the risk of PCa death was significantly lower among statin users before PCa diagnosis (HR 0.70, 95%Cls 0.52-0.95). For statin use after PCa diagnosis the risk was lowered in age-adjusted analysis (HR 0.76 95%Cls 0.62-0.93) but not after multivariable-adjustment. Post-diagnostic statin use was associated with improved PCa-specific survival in 1, 3 and 5 years lag-time analyses. The risk reduction was clearest for statin use initiated 5 years earlier (HR 0.71 95%Cls 0.55-0.92). Use of statins both before and after PCa diagnosis was associated with reduced risk of ADT use (HR 0.72 95%Cls 0.65-0.80 and HR 0.73, 95%Cl 0.67-0.80, respectively). The risk of ADT decreased by increasing intensity of statin use before diagnosis. Conclusion: Statin use among surgically treated PCa patients has significant association with decreased risk of starting ADT and PCa death. The risk is lowered especially among men with statin use before PCa diagnosis and in men who used statins at high-dose. Our results are hypothesis generating due to retrospective study design.