Statin Use and Prostate Cancer Incidence in Manitoba, Canada: A Population-Based Nested Case–Control Study

Little is known about modifiable risk factors for prostate cancer, which is the most frequently diagnosed cancer in men and a leading cause of cancer-related deaths. A link between statin use and prostate cancer risk has been proposed, but the epidemiologic evidence is inconclusive (although some evidence suggests a small, inverse association between statin use and advanced prostate cancer; ref. 1). Study results could differ, because most studies did not adjust for confounding by prostate cancer screening (2). Because statin users are more likely to be screened (possibly due to more frequent physician visits, which can increase early detection and confound results for both total and advanced prostate cancer), screening should be adjusted for (3, 4). Most studies did not examine variability due to differences in lipophilicity and potency of specific statins (1, 5). Only a few studies examined the associations for period, duration, or dose of use. We examined the association between statin use and prostate cancer incidence in the Canadian province of Manitoba.

We linked the Manitoba Cancer Registry (MCR) to the provincial prescription drugs database (DPIN) and other Manitoba Health (MH) databases (covering virtually all residents) to conduct a nested case–control study using the same methods we employed in previous studies (6, 7).

Briefly, we identified all men 40 years or older who were registered with MH during 2000 to 2014 with ≥5 years of insurance coverage (to ensure exposure data has sufficient length). Using a unique personal identifier, we linked to the MCR to identify all prostate cancer diagnoses [International Classification of Diseases for Oncology, Third Edition (ICD-O-3) code C61.9], and used incidence density sampling to match each case (on birth date, duration of coverage, and region of residence) to up to five cancer-free controls. Using the DPIN, we measured all statin use before the index date (the case's diagnosis date). We used conditional logistic regression (to account for matching) to calculate odds ratios (OR) of the association between prostate cancer risk and statin use (both as a class and for each individual statin) while adjusting for confounders, including a previously described screening score (ref. 7; data on routine bloodwork, including PSA testing, were not available). We also assessed the association by period (years before the index date), duration, and dose of statin use.

We matched 9,384 prostate cancer cases to 46,749 cancer-free controls (Table 1). Ever use of any statin was similar between cases (24.5%) and controls (25.2%). As expected, cases were screened more often (78%) than controls (35%). Otherwise, the two groups were comparable in terms of medication use, history of chronic diseases, and patterns of healthcare utilization.

Ever use of any statin was not associated with prostate cancer risk: OR = 0.96 (95% confidence interval, 0.90–1.03; Table 2). Except for pravastatin, 0.82 (0.71–0.96), ever use of individual statins was not associated with prostate cancer risk. There was also no period-, duration-, or dose-response for use of statins (Table 2) except for use of pravastatin [0.60 (0.40–0.89) among those who used it 11–20 years before the index date]. We found similar results when we limited our analysis to clinically significant prostate cancer [Supplementary Table S1; 1,636 cases; lymph nodes positive for metastatic prostate cancer or clinical stage American Joint Committee on Cancer (AJCC) 3–4; prostate cancer–specific grading is not included in the MCR].

We found no evidence of an association between statin use and prostate cancer risk. Although a 2012 meta-analysis reported a relative risk of 0.93 (0.87–0.99; ref. 8), more recent reviews highlighted significant heterogeneity between published studies, many of which relied on questionnaires for exposure data and lacked adjustment for screening (1, 2). We used routinely collected point-of-sale prescription information to avoid recall bias and attempted to adjust for screening.

The statistically significant association between pravastatin use and prostate cancer risk is not supported by other studies (5). Because we performed multiple comparisons for different statins and lacked a priori biological rationale to single out pravastatin, we cannot rule out a chance finding, especially given the absence of a dose or duration response.

A major strength of this study is the availability of high-quality, population-based health administrative databases in Manitoba. The completeness and accuracy of the MCR and MH databases are well established. Outcome misclassification is unlikely, as reporting of cases is mandated by provincial law. Although we lacked prescription information prior to 1995 (the start of DPIN), statins were infrequently prescribed before then. We lacked information on lipid levels, so we could not account for the severity of dyslipidemia (an indication for statin use), and used an indicator to adjust for screening, which may have caused residual confounding.

In conclusion, we found no evidence of an association between statin use and prostate cancer risk. In the absence of a dose or duration response, the association between pravastatin and prostate cancer risk may be due to chance.

S.M. Mahmud reports receiving commercial research grants from Merck, Sanofi, and Roche and is a consultant/advisory board member for Sanofi. No potential conflicts of interest were disclosed by the other authors.

The results and conclusions are those of the authors and no official endorsement by the Manitoba Centre for Health Policy, Manitoba Health, or other data providers is intended or should be inferred. The opinions presented in the article do not necessarily reflect those of the funders.

Conception and design: R. Bisewski, S.M. Mahmud

Development of methodology: C.H. Righolt, R. Bisewski, S.M. Mahmud

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C.H. Righolt, S.M. Mahmud

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C.H. Righolt, R. Bisewski, S.M. Mahmud

Writing, review, and/or revision of the manuscript: C.H. Righolt, R. Bisewski, S.M. Mahmud

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): C.H. Righolt, R. Bisewski, S.M. Mahmud

Study supervision: C.H. Righolt, S.M. Mahmud

The authors acknowledge the Manitoba Centre for Health Policy for use of data contained in the Manitoba Population Research Data Repository under project # 2015-021 [HIPC # 2011/2012-10, REB # HS13215(H2011:258), and RRIC # 2017-004]. Data used in this study are from the Manitoba Population Research Data Repository housed at the Manitoba Centre for Health Policy, University of Manitoba, and were derived from data provided by Manitoba Health and Cancer Care Manitoba. This work was funded by the Manitoba Health Research Council. S.M. Mahmud was partially funded by the Canada Research Chairs Program.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.