Prostate Cancer Death, Progression of Prostate Cancer, and Erectile Dysfunction After Radical Prostatectomy Among Patients with Statin and Metformin Use

Prostate cancer (PCa) causes a high disease burden and a large proportion of cancer deaths worldwide. Antidiabetic drugs and statins are commonly used medications. This has raised interest in their potential impact on PCa. Among antidiabetic drugs, metformin has been the particular subject of interest.

Statins and metformin have been proposed to affect the carcinogenesis and progression of PCa with plausible biological mechanisms. However, the epidemiological evidence provided mainly by observational studies is contradictory. Current knowledge supports that statins might affect PCa progression and prognosis favorably. Concordantly, metformin has been associated with improved overall survival, while the benefits in terms of PCa progression and PCa-specific survival have remained marginal. Due to the discrepancy, further high-quality studies, especially randomized studies, have been demanded.

Radical prostatectomy (RP) is an essential treatment modality in the treatment of intermediate- and high-risk localized prostate cancer. However, erectile dysfunction (ED) is a remarkable treatment-related side effect of RP. Statins have been suggested to have beneficial effects on erectile function by reducing endothelial dysfunction, but evidence among prostatectomy patients is inconclusive.

The aim of this thesis was to estimate the effect of statins and antidiabetic drugs on the recurrence of PCa and the prognosis of PCa patients among men who underwent RP. In addition, the thesis sought to assess whether antidiabetic drugs have an effect on the tumor characteristics of PCa and whether the use of statins affects postprostatectomy ED.

The study was conducted in two study populations. In cohort I, the study population included 1314 men who underwent RP for PCa at Tampere University Hospital (TaUH) between 1995 and 2009. In cohorts II-IV, the study population included 14,242 men with PCa operatively treated in Finland between 1995 and 2013. In cohort IV, 129 men were excluded due to treatment for ED initiated prior to surgery. Patients were identified from the prostate cancer database of TaUH and Care Registry of the Finnish Institute for Health and Welfare. Comprehensive information on purchases of cholesterol-lowering drugs, antidiabetic drugs, drugs for androgen deprivation therapy (ADT), drugs for ED treatment, and drugs for the treatment of clinically important comorbidities was gathered for each patient from the National Social Insurance Institution's prescription database. The medication data were collected from 1995 to 2009 in cohort I and from 1995 to 2014 in cohorts II-IV. Standardized doses were calculated for cholesterol-lowering drugs and antidiabetic drugs.

In cohort I, tumor characteristics, follow-up information, and information on PCa recurrence were obtained from the TaUH prostate cancer database. The risk of death was assessed based on the reported cause of death. In cohorts II-IV, PCa progression and post-prostatectomy ED were analyzed indirectly based on the initiation of medical treatment. Hazard ratios (HR) and 95% confidence intervals (CI) for the risk of tumor characteristics, PCa progression, ED treatments, and PCa/all-cause death by prediagnostic and postdiagnostic drug use were calculated with the Cox proportional hazard regression model. The ability to evaluate the effect of the cumulative amount, duration, and intensity of drug use on the outcomes enabled an estimation of dose-dependence. Analyses were carried out with adjustment for age and a multivariable adjustment model that included possible confounding factors. Postdiagnostic drug use was analyzed as a time-dependent variable in order to control for immortal time bias. Advanced analysis techniques were used, such as lag-time analysis, subgroup analysis, restricted analysis, and competing risk analysis.

Antidiabetic drug use was associated with a higher risk of high-grade PCa when compared to non-use, and glycemic control did not modify the risk. Due to low number of cases, our analyses of difference between antidiabetic drug users and non- users in other tumor characteristics (stage, tissue tumor markers, and PSA) and PCa prognosis contained uncertainty. In cohort II, a large nationwide study population and longer follow-up time enabled the more accurate evaluation, and antidiabetic drug use overall was related to a worse PCa prognosis. However, metformin use before and after a PCa diagnosis was associated with a decreased risk of PCa death and ADT initiation. The risk of PCa death in metformin users remained reduced after analysis for competing risks, and the effect showed long-lasting properties in the lag time analysis. However, no dose-dependence was observed. In contrast, the use of insulin analogues and drugs increasing insulin secretion were associated with a higher risk of PCa death and ADT initiation.

Men with statin use had decreased PCa-specific mortality compared to non-users. Both prediagnostic and postdiagnostic statin use was associated with a reduced risk of PCa death and ADT initiation. Furthermore, the beneficial effect of statin use on PCa survival was long-lasting. Furthermore, the effect of postdiagnostic statin use on the risk of ADT initiation was dose-dependent.

In general, no unequivocal association between statin use and the initiation of ED treatments after RP was observed. The prediagnostic use of statins was generally not associated with ED treatments. However, in a subgroup of patients with no known cardiovascular comorbidities, the risk of ED treatment initiation was slightly higher in statin users. Men with postdiagnostic statin use had a slightly increased risk of initiating ED treatment, although a longer duration of statin use decreased the risk.

In conclusion, high-grade PCa is more frequently detected among diabetic men with no difference in PSA levels. Statins and metformin might improve the PCa-specific prognosis among patients treated with RP, and the effect of metformin on the course of PCa may differ favorably from other antidiabetic drugs. These results are hypothesis-generating, and randomized trials are still needed to evaluate the effects of statins and metformin on the PCa prognosis. Statin use slightly increases the risk of ED treatment initiation after RP. Vascular properties and endothelial function may have an effect on erectile function after RP, but more well-designed studies on the effect of statins on erectile function after RP are needed.