Article Tools
ORIGINAL REPORTS
Rapid Communications
Article Tools
Use of Statins and the Risk of Death in Patients With Prostate Cancer
Abstract
To determine whether the use of statins after prostate cancer diagnosis is associated with a decreased risk of cancer-related mortality and all-cause mortality and to assess whether this association is modified by prediagnostic use of statins.
A cohort of 11,772 men newly diagnosed with nonmetastatic prostate cancer between April 1, 1998, and December 31, 2009, followed until October 1, 2012, was identified using a large population-based electronic database from the United Kingdom. Time-dependent Cox proportional hazards models were used to estimate adjusted hazard ratios (HRs) with 95% CIs of mortality outcomes associated with postdiagnostic use of statins, lagged by 1 year to account for latency considerations and to minimize reverse causality, and considering effect modification by prediagnostic use of statins.
During a mean follow-up time of 4.4 years (standard deviation, 2.9 years), 3,499 deaths occurred, including 1,791 from prostate cancer. Postdiagnostic use of statins was associated with a decreased risk of prostate cancer mortality (HR, 0.76; 95% CI, 0.66 to 0.88) and all-cause mortality (HR, 0.86; 95% CI, 0.78 to 0.95). These decreased risks of prostate cancer mortality and all-cause mortality were more pronounced in patients who also used statins before diagnosis (HR, 0.55; 95% CI, 0.41 to 0.74; and HR, 0.66; 95% CI, 0.53 to 0.81, respectively), with weaker effects in patients who initiated the treatment only after diagnosis (HR, 0.82; 95% CI, 0.71 to 0.96; and HR, 0.91; 95% CI, 0.82 to 1.01, respectively).
There has been great interest in the antitumor effects of 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors, better known as statins. Statins have been shown to improve lipid profiles and reduce cardiovascular morbidity and mortality.1 There is now evidence that statins may have antitumor effects in a number of cancers, including prostate cancer.2
To date, a number of observational studies have investigated the association between the use of statins and different prostate cancer outcomes, but with inconsistent findings.2–11 Such discrepancies are likely a result of methodologic limitations, which included relatively small sample sizes, immortal time bias,3,5,10 and no consideration of latency time windows between the first use of statins and prostate cancer outcomes.5–10 Furthermore, none of these studies specifically assessed whether the use of statins before the prostate cancer diagnosis modified the association on the use of statins after diagnosis, because there is some evidence that prediagnostic use of statins may be associated with more favorable tumor characteristics.12–14
Thus, the primary objective of this study was to determine whether the use of statins after prostate cancer diagnosis is associated with a decreased risk of prostate cancer mortality. Secondary objectives were to determine whether such use is also associated with a decreased risk of all-cause mortality and to assess whether prediagnostic use of statins modifies the association of these mortality outcomes.
This study was conducted by linking the following four large electronic databases from the United Kingdom: the United Kingdom National Cancer Registry, the Clinical Practice Research Datalink (CPRD; previously known as the General Practice Research Database), the Hospital Episode Statistics (HES) database, and the Office for National Statistics (ONS) database. The United Kingdom National Cancer Registry contains tumor information, including site of primary growth (coded using the International Classification of Diseases, 10th Revision [ICD-10]) and tumor characteristics (such as grade, stage, and primary treatments received, although not consistently recorded). The CPRD contains data on more than 12 million individuals enrolled in more than 650 general practices.15 The HES database contains dates of hospital admissions, primary and secondary diagnoses (coded using the ICD-10 classification), and related procedures (coded using the ICD-10 classification and Office of Population Censuses and Surveys Classification of Interventions and Procedures, Fourth Version). Finally, the ONS contains the electronic death certificates of all citizens living in the United Kingdom and was used to identify the underlying cause of death (coded using the ICD-10 classification) for all patients who died during follow-up. The study protocol was approved by the Independent Scientific Advisory Committee of the CPRD and the Research Ethics Board of the Jewish General Hospital, Montreal, Quebec, Canada.
We conducted a population-based retrospective cohort study within the databases described earlier. As a first step, we used the United Kingdom National Cancer Registry to identify all patients newly diagnosed with prostate cancer (ICD-10 code: C61) between April 1, 1998, and December 31, 2009, followed until October 1, 2012. These patients were then linked to the CPRD, HES, and ONS databases. We excluded patients diagnosed with metastatic disease and those with less than 1 year of up-to-standard medical history in the CPRD before the prostate cancer diagnosis. Furthermore, all patients were required to have at least 1 year of follow-up, which was necessary for latency considerations. Thus, cohort entry was set to the year after the prostate cancer diagnosis, and all patients were observed until death (either from prostate cancer or from other causes), end of registration with the general practice, or end of study period (October 1, 2012), whichever came first.
We considered all statins on the United Kingdom market during the study period. These consisted of statins classified as hydrophilic (pravastatin and rosuvastatin) and lipophilic (atorvastatin, simvastatin, fluvastatin, and cerivastatin).
Use of statins after the prostate cancer diagnosis (ie, postdiagnostic use) was entered as a time-dependent variable in the models, allowing patients to move from a period of nonexposure to a period of exposure. Thus, patients were considered unexposed to statins up until the time of a first prescription and exposed thereafter for the remainder of follow-up. Furthermore, statin exposure was lagged by 1 year to take into account a biologically meaningful latency time window given that short duration exposures are unlikely to be associated with the mortality outcomes.
On the basis of the previous considerations, postdiagnostic statin exposure was expressed in the following three ways: ever use, cumulative duration of use, and cumulative dose. For the first approach, ever use of statins after the prostate cancer diagnosis was compared with never use up until the time of the event. For the second and third approaches, it was of interest to determine whether there was a dose-response relationship between statin use and cancer-related and all-cause mortality. Cumulative duration of use was defined, in a time-dependent fashion, as the total number of months of statin exposure, calculated by summing the durations of all prescriptions between cohort entry and the time of the event. This variable was then classified into the following four categories: less than 12 months, 12 to 24 months, 24 to 36 months, and ≥ 36 months of use. Similarly, for cumulative duration of use, patients identified as ever exposed to statins had their exposure expressed in units of defined daily dose (DDD), which is a validated measure from the WHO.16 The DDD for the 30-mg formulation of simvastatin was used as the reference, and the DDDs for the other statins were used to convert each statin to 30-mg simvastatin-equivalent doses. Cumulative dose was then calculated by summing all DDDs up until the date of the event. This variable was classified into the following four categories: less than 365, 365 to 731, 731 to 1096, and ≥ 1,096 DDDs.
Descriptive statistics were used to describe the characteristics of the entire cohort and, separately, of patients exposed and unexposed to statins before prostate cancer diagnosis. Time-dependent Cox proportional hazards models were used to estimate hazards ratios (HRs) with 95% CIs of prostate cancer mortality and all-cause mortality associated with the ever use, cumulative duration of use, and cumulative dose of statins. For cumulative duration and dose, linear trend was assessed by entering the aforementioned categories of these variables as continuous in the models. All models were adjusted for a number of potential confounders measured before cohort entry including prediagnostic statin use (listed in Table 1), as well as prostate-specific antigen (PSA) testing activity (ie, number of tests performed), which was included as a time-dependent covariate in the models. Staging information was not included as a covariate because it was missing for approximately 90% of patients. Schoenfeld residuals were examined for the time-fixed covariates, and no significant departure from the proportional hazards assumption was detected.
|
| Characteristic | Entire Cohort (N = 11,772) | Prediagnostic Statin Users (n = 3,407) | No Prediagnostic Statin Use (n = 8,365) | |||
|---|---|---|---|---|---|---|
| No. of Patients | % | No. of Patients | % | No. of Patients | % | |
| Age, years | ||||||
| Mean | 71.3 | 71.9 | 71.1 | |||
| Standard deviation | 8.8 | 7.5 | 9.3 | |||
| Race | ||||||
| White | 10,375 | 88.1 | 3,044 | 89.4 | 7,331 | 87.6 |
| Black | 134 | 1.1 | 43 | 1.3 | 91 | 1.1 |
| Other | 133 | 1.1 | 52 | 1.5 | 81 | 1.0 |
| Unknown | 1,130 | 9.6 | 268 | 7.9 | 862 | 10.3 |
| Excessive alcohol use | 872 | 7.4 | 339 | 10.0 | 533 | 6.4 |
| Smoking status | ||||||
| Never | 4,708 | 40.0 | 1,105 | 32.4 | 3,603 | 43.1 |
| Ever | 6,456 | 54.8 | 2,248 | 66.0 | 4,208 | 50.3 |
| Unknown | 608 | 5.2 | 54 | 1.6 | 554 | 6.6 |
| Body mass index, kg/m2 | ||||||
| < 18.5 | 103 | 0.9 | 21 | 0.6 | 82 | 1.0 |
| 18.5-25 | 3,672 | 31.2 | 899 | 26.4 | 2,773 | 33.2 |
| 25-30 | 5,202 | 44.2 | 1,623 | 47.6 | 3,579 | 42.8 |
| ≥ 30 | 1,996 | 17.0 | 768 | 22.5 | 1,228 | 14.7 |
| Unknown | 799 | 6.8 | 96 | 2.8 | 703 | 8.4 |
| Comorbidities | ||||||
| Chronic kidney disease | 793 | 6.8 | 448 | 13.2 | 345 | 4.1 |
| Myocardial infarction | 991 | 8.4 | 753 | 22.1 | 238 | 2.9 |
| Ischemic stroke | 485 | 4.1 | 305 | 9.0 | 180 | 2.2 |
| Transient ischemic attack | 587 | 5.0 | 312 | 9.2 | 275 | 3.3 |
| Peripheral artery disease | 1,900 | 16.1 | 999 | 29.3 | 901 | 10.8 |
| Previous cancer | 1,862 | 15.8 | 557 | 16.4 | 1,305 | 15.6 |
| Prostate-specific antigen, ng/mL | ||||||
| < 4 | 408 | 3.5 | 232 | 6.8 | 443 | 5.3 |
| 4-10 | 2,902 | 24.7 | 975 | 28.6 | 2,074 | 24.8 |
| >10 | 5,110 | 43.4 | 1,454 | 42.7 | 3,345 | 40.0 |
| Unknown | 3,352 | 28.5 | 746 | 21.9 | 2,503 | 29.9 |
| Gleason score | ||||||
| 2 | 52 | 0.4 | 9 | 0.3 | 43 | 0.5 |
| 3-6 | 2,867 | 24.4 | 843 | 24.7 | 2,006 | 24.0 |
| 7 | 2,248 | 19.1 | 709 | 20.8 | 1,527 | 18.3 |
| ≥ 8 | 1,384 | 11.8 | 472 | 13.9 | 906 | 10.8 |
| Unknown | 5,221 | 44.4 | 1,374 | 40.3 | 3,883 | 46.4 |
| Prostate cancer treatments* | ||||||
| Radical prostatectomy | 5,148 | 43.7 | 1,378 | 40.5 | 3,770 | 45.1 |
| Radiotherapy | 2,425 | 20.6 | 785 | 23.0 | 1,640 | 19.6 |
| Chemotherapy | 293 | 2.5 | 101 | 3.0 | 192 | 2.3 |
| Androgen deprivation therapy | 6,823 | 58.0 | 1,997 | 58.6 | 4,826 | 57.7 |
| Metformin | 605 | 5.1 | 435 | 12.8 | 170 | 2.0 |
| Sulfonylureas | 514 | 4.4 | 328 | 9.6 | 186 | 2.2 |
| Thiazolidinediones | 112 | 1.0 | 99 | 2.9 | 13 | 0.2 |
| Other hypoglycemic agents | 63 | 0.5 | 43 | 1.3 | 20 | 0.3 |
| Insulin | 164 | 1.4 | 121 | 3.6 | 43 | 0.5 |
| ACE inhibitors | 3,290 | 28.0 | 1,819 | 53.4 | 1,471 | 17.6 |
| Angiotensin receptor blockers | 937 | 8.0 | 556 | 16.3 | 381 | 4.6 |
| Calcium channel blocker | 3,310 | 28.1 | 1,684 | 49.4 | 1,626 | 19.4 |
| β-blockers | 3,187 | 27.1 | 1,727 | 50.7 | 1,460 | 17.5 |
| Diuretics | 4,138 | 35.2 | 1,754 | 51.5 | 2,384 | 28.5 |
| Other antihypertensives | 113 | 1.0 | 57 | 1.7 | 56 | 0.7 |
| Aspirin | 4,145 | 35.2 | 2,372 | 69.6 | 1,773 | 21.2 |
| Other NSAIDs | 5,848 | 49.7 | 1,832 | 53.8 | 4,016 | 48.0 |
| 5α-reductase | 799 | 6.8 | 287 | 8.4 | 512 | 6.1 |
| Prediagnostic use of statins | 3,407 | 28.9 | 3,407 | 100.0 | 0 | 0.0 |
Abbreviations: ACE, angiotensin-converting enzyme; NSAIDs, nonsteroidal anti-inflammatory drugs.
*Treatments received in the first year after the prostate cancer diagnosis.
In an exploratory analysis, we determined whether statin lipophilicity influenced the risk of prostate cancer outcomes, because lipophilic statins have been shown to be more potent in reducing prostate cancer cell migration and spread of metastatic prostate colonies.17 For this analysis, users of statins were further classified into one of the following three mutually exclusive categories: use of lipophilic statins only, use of both lipophilic and hydrophilic statins, and use of hydrophilic statins only. We then performed two secondary analyses. In the first, we determined whether prediagnostic use of statins acted as an effect modifier of the association between post-treatment statin use and prostate cancer and all-cause mortality. Effect modification was assessed by including interaction terms between prediagnostic and postdiagnostic statin use in the models. In the second analysis, we determined whether postdiagnostic use of statins was associated with a decreased risk of distant metastasis. Finally, for the prostate cancer mortality analysis, we conducted a sensitivity analysis to account for competing risks as a result of death from other causes using the subdistribution hazards model proposed by Fine and Gray18 adapted for time-dependent covariates.19 All analyses were two-tailed tests based on α = .05 and were performed using SAS version 9.3 (SAS Institute, Cary, NC).
A total of 11,772 patients newly diagnosed with nonmetastatic prostate cancer met the study inclusion criteria (Fig 1). The mean age at cohort entry was 71.3 years (standard deviation, 8.8 years), with a mean follow-up time of 4.4 years (standard deviation, 2.9 years). The corresponding incidence rates of prostate cancer mortality and all-cause mortality were 34.8 per 1,000 per year (95% CI, 33.2 to 36.4 per 1,000 per year) and 67.9 per 1,000 per year (95% CI, 65.7 to 70.2 per 1,000 per year), respectively.
Fig 1. Study flow chart.
Table 1 lists the characteristics of the entire cohort, as well as of the statin and nonstatin users, at baseline. As expected, statin users were more likely to have been smokers and obese (body mass index ≥ 30 kg/m2), had a higher prevalence of comorbidities, and were more likely to have used antidiabetic agents, antihypertensive drugs, and aspirin (Table 1).
Table 2 lists the results of the primary analysis. Use of statins after prostate cancer diagnosis was associated with a 24% risk reduction in prostate cancer mortality (Table 2). A dose-response relationship was observed in terms of cumulative duration of use and dose, with the HRs becoming progressively more protective with longer durations of use and higher cumulative doses (≥ 36 months of use: HR, 0.61; 95% CI, 0.49 to 0.75; and ≥ 1,096 DDDs: HR, 0.57; 95% CI, 0.46 to 0.72). In an exploratory analysis, a 23% decreased risk was observed with lipophilic statins, and a 35% decreased risk was observed with hydrophilic statins (Appendix Table A1, online only).
|
| Statin Exposure | No. of Patients (n = 1,791) | No. of Person-Years | Incidence Rate (per 100 person-years) | 95% CI (per 100 person-years) | Crude HR | Adjusted HR* | 95% CI |
|---|---|---|---|---|---|---|---|
| No use of statins after diagnosis | 1,250 | 32,283 | 3.87 | 3.66 to 4.09 | 1.00 | 1.00 | Reference |
| Use of statins after diagnosis | 541 | 19,261 | 2.81 | 2.58 to 3.05 | 0.75 | 0.76 | 0.66 to 0.88 |
| Cumulative duration of use, months | |||||||
| < 12 | 90 | 2,401 | 3.75 | 3.03 to 4.59 | 0.97 | 0.99 | 0.79 to 1.24 |
| 12-24 | 177 | 5,553 | 3.19 | 2.74 to 3.68 | 0.81 | 0.82 | 0.68 to 0.99 |
| 24-36 | 104 | 4,032 | 2.58 | 2.12 to 3.11 | 0.67 | 0.66 | 0.52 to 0.83 |
| ≥ 36 | 170 | 7,275 | 2.34 | 2.01 to 2.71 | 0.66 | 0.61 | 0.49 to 0.75 |
| P | < .001 | ||||||
| Cumulative dose, DDDs† | |||||||
| < 365 | 154 | 4,454 | 3.46 | 2.94 to 4.04 | 0.89 | 0.84 | 0.70 to 1.02 |
| 365-731 | 164 | 5,381 | 3.05 | 2.61 to 3.54 | 0.79 | 0.80 | 0.66 to 0.97 |
| 731-1,096 | 104 | 3,643 | 2.86 | 2.34 to 3.45 | 0.76 | 0.76 | 0.60 to 0.95 |
| ≥ 1,096 | 119 | 5,783 | 2.06 | 1.71 to 2.45 | 0.57 | 0.57 | 0.46 to 0.72 |
| P | < .001 | ||||||
Abbreviations: DDD, defined daily dose; HR, hazard ratio.
*Adjusted for age, year of prostate cancer diagnosis, ethnicity, excessive alcohol use, smoking status, obesity, chronic kidney disease, myocardial infarction, ischemic stroke, transient ischemic attack, peripheral artery disease, previous cancers, prostate-specific antigen level, Gleason score, metformin, sulfonylureas, thiazolidinediones, insulins, other oral antihypoglycemic agents, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, calcium channel blockers, β-blockers, diuretics, other antihypertensive drugs, aspirin, other nonsteroidal anti-inflammatory drugs, 5α-reductase inhibitors, prediagnostic statin use, prostate-specific antigen testing activity, prostatectomy, radiation therapy, chemotherapy, and androgen deprivation therapy.
†Based on an equivalent dose of 30 mg of simvastatin.
Table 3 lists the results of the secondary outcome. Use of statins after diagnosis was associated with a 14% decreased risk of all-cause mortality (Table 3), with the HRs gradually decreasing with longer cumulative durations of use and higher cumulative doses (Table 3).
|
| Statin Exposure | No. of Patients (n = 3,499) | No. of Person-Years | Incidence Rate (per 100 person-years) | 95% CI (per 100 person-years) | Crude HR | Adjusted HR* | 95% CI |
|---|---|---|---|---|---|---|---|
| No use of statins after diagnosis | 2,270 | 32,283 | 7.03 | 6.75 to 7.33 | 1.00 | 1.00 | Reference |
| Use of statins after diagnosis | 1,229 | 19,261 | 6.38 | 6.03 to 6.75 | 0.90 | 0.86 | 0.78 to 0.95 |
| Cumulative duration of use, months | |||||||
| < 12 | 195 | 2,401 | 8.12 | 7.04 to 9.32 | 1.15 | 0.93 | 0.81 to 1.08 |
| 12-24 | 334 | 5,553 | 6.02 | 5.40 to 6.69 | 0.85 | 0.88 | 0.76 to 1.03 |
| 24-36 | 253 | 4,032 | 6.28 | 5.54 to 7.09 | 0.90 | 0.88 | 0.75 to 1.03 |
| ≥ 36 | 447 | 7,275 | 6.14 | 5.59 to 6.73 | 0.86 | 0.82 | 0.69 to 0.97 |
| P | < .001 | ||||||
| Cumulative dose, DDDs† | |||||||
| < 365 | 323 | 4,454 | 7.25 | 6.49 to 8.08 | 1.03 | 0.95 | 0.83 to 1.10 |
| 365-731 | 351 | 5,381 | 6.52 | 5.87 to 7.23 | 0.92 | 0.87 | 0.74 to 1.01 |
| 731-1,096 | 233 | 3,643 | 6.40 | 5.61 to 7.26 | 0.90 | 0.85 | 0.73 to 1.00 |
| ≥ 1,096 | 322 | 5,783 | 5.57 | 4.98 to 6.20 | 0.77 | 0.85 | 0.72 to 1.00 |
| P | < .001 | ||||||
Abbreviations: DDD, defined daily dose; HR, hazard ratio.
*Adjusted for age, year of prostate cancer diagnosis, ethnicity, excessive alcohol use, smoking status, obesity, chronic kidney disease, myocardial infarction, ischemic stroke, transient ischemic attack, peripheral artery disease, previous cancers, prostate-specific antigen level, Gleason score, metformin, sulfonylureas, thiazolidinediones, insulins, other oral antihypoglycemic agents, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, calcium channel blockers, β-blockers, diuretics, other antihypertensive drugs, aspirin, other nonsteroidal anti-inflammatory drugs, 5α-reductase inhibitors, prediagnostic statin use, prostate-specific antigen testing activity, prostatectomy, radiation therapy, chemotherapy, and androgen deprivation therapy.
†Based on an equivalent dose of 30 mg of simvastatin.
Table 4 lists the results of the secondary analyses. The use of statins before prostate cancer diagnosis significantly modified the association between postdiagnostic use of statins and prostate cancer mortality (Table 4), with greater risk reductions in patients who also used statins before diagnosis (HR, 0.55; 95% CI, 0.41 to 0.74). Dose-response relationships in terms of cumulative duration of use and dose were observed but were more pronounced in patients who also used statins before diagnosis (Appendix Table A2, online only). With respect to the secondary outcome of all-cause mortality, similar patterns were observed, with the strongest risk reductions in patients who also used statins before diagnosis, whereas the HR approached statistical significance for patients who did not use statins before diagnosis (Table 4). A dose-response relationship was observed for all-cause mortality in patients who also used statins before diagnosis (Appendix Table A3, online only). After accounting for competing risks as a result of other causes of death, the decreased risk observed between postdiagnostic use of statins and prostate cancer mortality remained statistically significant (HR, 0.87; 95% CI, 0.75 to 0.99; P = .047). Finally, postdiagnostic use of statins was also associated with a decreased risk of distant metastasis (HR, 0.77; 95% CI, 0.68 to 0.88; Appendix Table A4, online only).
|
| Outcome | Statin Use Before Diagnosis | No Statin Use Before Diagnosis | P for Interaction | ||
|---|---|---|---|---|---|
| Adjusted HR* | 95% CI | Adjusted HR* | 95% CI | ||
| Prostate cancer mortality | 0.55 | 0.41 to 0.74 | 0.82 | 0.71 to 0.96 | .02 |
| All-cause mortality | 0.66 | 0.53 to 0.81 | 0.91 | 0.82 to 1.01 | .01 |
Abbreviation: HR, hazard ratio.
*Adjusted for age, year of prostate cancer diagnosis, ethnicity, excessive alcohol use, smoking status, obesity, chronic kidney disease, myocardial infarction, ischemic stroke, transient ischemic attack, peripheral artery disease, previous cancers, prostate-specific antigen level, Gleason score, metformin, sulfonylureas, thiazolidinediones, insulins, other oral antihypoglycemic agents, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, calcium channel blockers, β-blockers, diuretics, other antihypertensive drugs, aspirin, other nonsteroidal anti-inflammatory drugs, 5α-reductase inhibitors, prostate-specific antigen testing activity, prostatectomy, radiation therapy, chemotherapy, and androgen deprivation therapy.
The results of this large population-based study of patients newly diagnosed with nonmetastatic prostate cancer indicate that the postdiagnostic use of statins is associated with a 24% decrease in prostate cancer mortality, with dose-response relationships in terms of cumulative duration of use and cumulative dose. The use of statins after prostate cancer diagnosis was also associated with a decreased risk of all-cause mortality and distant metastasis. Finally, the strongest risk reductions with postdiagnostic use of statins were observed in patients who also used these drugs before diagnosis, with more modest risk reductions in patients who initiated these drugs after diagnosis.
Overall, the postdiagnostic use of statins was associated with a decreased risk in prostate cancer mortality. In a secondary analysis, however, the use of statins both before and after diagnosis was associated with a stronger risk reduction than initiation of those drugs after diagnosis. There are a number of possible explanations for the observed effect modification by prediagnostic use of statins. One possibility may be that compared with nonusers, prediagnostic statin users have more favorable tumor characteristics,12–14 which would result in improved prostate cancer outcomes. However, this was not observed in our cohort, because prediagnostic statin users were slightly more likely to have higher Gleason scores (ie, ≥ 8) compared with nonusers. Another possibility could be that prediagnostic statin users are also those who have used these drugs for the longest durations, which is concordant with our observation that longer durations of use and higher cumulative doses are associated with the strongest risk reductions. Finally, patients initiating statins before prostate cancer diagnosis may be different from patients initiating after diagnosis. Specifically, it is possible that the latter group required statins as a consequence of certain treatments, such as androgen deprivation therapy,20,21 which is known to increase lipid levels. Androgen deprivation therapy is typically prescribed to patients with advanced prostate cancer, which despite statistical adjustments, would make statins seem to have more modest effects. Nonetheless, our findings suggest that the use of statins before and/or after prostate cancer diagnosis may reduce the risk of prostate cancer mortality.
The antitumor effects by statin use in prostate carcinogenesis have been demonstrated in a number of experimental studies. These studies have shown that statins can potentially decrease the risk of prostate cancer development and progression via a variety of mechanisms through its action on lipid metabolism. Statins decrease cholesterol levels and may have antitumor effects on prostate cancer via their effects on cell proliferation, inflammation, membrane organogenesis, and steroidogenesis.22 The decreased risk of prostate cancer mortality associated with the use of statins observed in our study supports the experimental evidence suggesting a potential antitumor effect of these drugs on prostate carcinogenesis.
To date, a number of observational studies have examined the association between statins and prostate cancer outcomes, with mixed findings.2–11 Disease recurrence was the primary outcome in the majority of these studies,8,9 with two studies reporting null associations8,9 and the remaining studies reporting risk reductions ranging between 30% and 43%.5–7 With respect to mortality outcomes, one study reported that statins were associated with 41% (HR, 0.59; 95% CI, 0.37 to 0.94) and 65% (HR, 0.35; 95% CI, 0.21 to 0.58) decreased risks on all-cause mortality in patients who underwent radiotherapy and radical prostatectomy, respectively.10 However, the method of assessing statin exposure in that study introduced immortal time bias,23 which may have exaggerated the potential benefits of this therapy. With respect to prostate cancer mortality, one study reported that statins were associated with a 65% risk reduction in lethal and fatal prostate cancer,4 whereas another study observed no statistically significant reduction in prostate cancer mortality and all-cause mortality,5 although the latter study was likely underpowered to investigate these outcomes. Although most of these studies assessed statin use at the time of the prostate cancer diagnosis, none specifically quantified the association with prostate cancer outcomes with use after diagnosis. This postdiagnostic use most resembles the expected benefits of these drugs in the adjuvant setting and thus could provide critical information for the design of randomized controlled trials assessing the effects of statins in patients newly diagnosed with prostate cancer.
This population-based study has a number of strengths and some limitations. To our knowledge, this is the largest study to have investigated the association between the use of statins and prostate cancer outcomes. Furthermore, with up to 15 years of follow-up, we were able to identify a significant number of patients, and thus, the study was well powered to investigate the two outcomes of interest. This study also considered latency, which was a methodologic shortcoming of the previous studies.5–11 Exposure misclassifications from the uncertainty of whether patients complied with the treatment regimen and the lack of information on prescriptions written by specialists are possible. However, these exposure misclassifications would dilute the HRs toward the null and thus underestimate the reduced risks observed in this study. Misclassifications with our primary outcome of prostate cancer mortality are also likely to have been minimal because, unlike other cancer types, prostate cancer mortality has been shown to be generally well recorded in death certificates (κ = 0.91).24 Moreover, the overall prostate cancer mortality rate of 34.8 per 1,000 per year estimated in our study is concordant with the prostate cancer mortality rate previously reported in the United Kingdom.25 Finally, linking the United Kingdom National Cancer Registry, CPRD, HES, and ONS databases allowed us to collect and adjust for a number of potential important confounders, including ethnicity, smoking, body mass index, prostate cancer treatments, and PSA levels. Unfortunately, it was not possible to adjust for tumor stage, because this information was missing for nearly 90% of patients in the United Kingdom National Cancer Registry and there were missing data on Gleason scores and PSA levels. Although these are important predictors of our outcomes, it is unclear whether they are confounders because they may not be associated with statin exposure. Furthermore, our analyses adjusted for prostate cancer therapies received, which are closely correlated to tumor characteristics.
Another important consideration is the possible healthy-user bias previously described with statins users.26 However, it is unclear whether better lifestyle habits would fully explain the 24% risk reduction in prostate cancer mortality and the dose-response relationships observed. Furthermore, statin users seemed to be less healthy than nonstatin users (ie, more likely to have been smokers and have a higher prevalence of comorbidities), which is a characteristic more likely to be associated with worse outcomes. Furthermore, the models were adjusted for prostate cancer treatments and PSA testing activity to minimize the potential bias of compliant statin users being more likely to seek aggressive treatments. Overall, adjustment for these variables did not materially affect the point estimates, although residual confounding may be present.
In summary, the results of this study indicate that the use of statins after a prostate cancer diagnosis is associated with a decreased risk of cancer-related mortality and all-cause mortality. These decreased risks were strongest in patients who also used statins before prostate cancer diagnosis, suggesting a possible effect of longer cumulative durations of use and higher cumulative doses. Finally, although the results of this study provide evidence that the use of statins may be associated with a decreased risk of prostate cancer mortality, additional well-conducted observational studies are needed to confirm these findings before launching randomized controlled trials assessing the effects of statins in the adjuvant setting.
Processed as a Rapid Communication manuscript. See accompanying editorial on page 1
Supported by the Canadian Institutes of Health Research. L.A. is the recipient of a Chercheur-Boursier career award from the Fonds de la Recherche en Santé du Québec, and S.S. is the recipient of the James McGill Chair.
Terms in blue are defined in the glossary, found at the end of this article and online at www.jco.org.
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
Although all authors completed the disclosure declaration, the following author(s) and/or an author's immediate family member(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: None Consultant or Advisory Role: Samy Suissa, AstraZeneca (C), Boehringer Ingelheim (C), Novartis (C), Pfizer (C), Merck (C) Stock Ownership: None Honoraria: None Research Funding: None Expert Testimony: None Patents: None Other Remuneration: None
Conception and design: Oriana Yu, Serge Benayoun, Gerald Batist, Samy Suissa, Laurent Azoulay
Financial support: Laurent Azoulay
Collection and assembly of data: Maria Eberg, Samy Suissa, Laurent Azoulay
Data analysis and interpretation: Oriana Yu, Maria Eberg, Serge Benayoun, Armen Aprikian, Samy Suissa, Laurent Azoulay
Manuscript writing: All authors
Final approval of manuscript: All authors
| 1. | DA Eisenberg : Cholesterol lowering in the management of coronary artery disease: The clinical implications of recent trials Am J Med 104: 2S– 5S,1998 Crossref, Medline, Google Scholar |
| 2. | SF Nielsen, BG Nordestgaard, SE Bojesen : Statin use and reduced cancer-related mortality N Engl J Med 367: 1792– 1802,2012 Crossref, Medline, Google Scholar |
| 3. | MS Geybels, JL Wright, SK Holt , etal: Statin use in relation to prostate cancer outcomes in a population-based patient cohort study Prostate 73: 1214– 1222,2013 Crossref, Medline, Google Scholar |
| 4. | EA Platz, MF Leitzmann, K Visvanathan , etal: Statin drugs and risk of advanced prostate cancer J Natl Cancer Inst 98: 1819– 1825,2006 Crossref, Medline, Google Scholar |
| 5. | R Gutt, N Tonlaar, R Kunnavakkam , etal: Statin use and risk of prostate cancer recurrence in men treated with radiation therapy J Clin Oncol 28: 2653– 2659,2010 Link, Google Scholar |
| 6. | RJ Hamilton, LL Banez, WJ Aronson , etal: Statin medication use and the risk of biochemical recurrence after radical prostatectomy: Results from the Shared Equal Access Regional Cancer Hospital (SEARCH) Database Cancer 116: 3389– 3398,2010 Crossref, Medline, Google Scholar |
| 7. | MA Kollmeier, MS Katz, K Mak , etal: Improved biochemical outcomes with statin use in patients with high-risk localized prostate cancer treated with radiotherapy Int J Radiat Oncol Biol Phys 79: 713– 718,2011 Crossref, Medline, Google Scholar |
| 8. | DE Soto, S Daignault, HM Sandler , etal: No effect of statins on biochemical outcomes after radiotherapy for localized prostate cancer Urology 73: 158– 162,2009 Crossref, Medline, Google Scholar |
| 9. | MA Moyad, GS Merrick, WM Butler , etal: Statins, especially atorvastatin, may favorably influence clinical presentation and biochemical progression-free survival after brachytherapy for clinically localized prostate cancer Urology 66: 1150– 1154,2005 Crossref, Medline, Google Scholar |
| 10. | MS Katz, PR Carroll, JE Cowan , etal: Association of statin and nonsteroidal anti-inflammatory drug use with prostate cancer outcomes: Results from CaPSURE BJU Int 106: 627– 632,2010 Crossref, Medline, Google Scholar |
| 11. | SW Marcella, A David, PA Ohman-Strickland , etal: Statin use and fatal prostate cancer: A matched case-control study Cancer 118: 4046– 4052,2012 Crossref, Medline, Google Scholar |
| 12. | RH Breau, RJ Karnes, DJ Jacobson , etal: The association between statin use and the diagnosis of prostate cancer in a population based cohort J Urol 184: 494– 499,2010 Crossref, Medline, Google Scholar |
| 13. | S Loeb, D Kan, BT Helfand , etal: Is statin use associated with prostate cancer aggressiveness? BJU Int 105: 1222– 1225,2010 Crossref, Medline, Google Scholar |
| 14. | WR Farwell, LW D'Avolio, RE Scranton , etal: Statins and prostate cancer diagnosis and grade in a veterans population J Natl Cancer Inst 103: 885– 892,2011 Crossref, Medline, Google Scholar |
| 15. | More dimensions to data, observational data Clinical Practice Research Datalink http://www.cprd.com/ObservationalData/ Google Scholar |
| 16. | DDD: Definition and general considerations WHO Collaborating Centre for Drug Statistics Methodology World Health Organization http://www.whocc.no/ddd/definition_and_general_considera/ Google Scholar |
| 17. | M Brown, C Hart, T Tawadros , etal: The differential effects of statins on the metastatic behaviour of prostate cancer Br J Cancer 106: 1689– 1696,2012 Crossref, Medline, Google Scholar |
| 18. | JP Fine, RJ Gray : A Proportional hazards model for the subdistribution of a competing risk J Am Stat Assoc 94: 496– 509,1999 Crossref, Google Scholar |
| 19. | PK Ruan, RJ Gray : Analyses of cumulative incidence functions via non-parametric multiple imputation Stat Med 27: 5709– 5724,2008 Crossref, Medline, Google Scholar |
| 20. | H Ahmadi, S Daneshmand : Androgen deprivation therapy: Evidence-based management of side effects BJU Int 111: 543– 548,2013 Crossref, Medline, Google Scholar |
| 21. | S Shahani, M Braga-Basaria, S Basaria : Androgen deprivation therapy in prostate cancer and metabolic risk for atherosclerosis J Clin Endocrinol Metab 93: 2042– 2049,2008 Crossref, Medline, Google Scholar |
| 22. | K Pelton, MR Freeman, KR Solomon : Cholesterol and prostate cancer Curr Opin Pharmacol 12: 751– 759,2012 Crossref, Medline, Google Scholar |
| 23. | S Suissa : Immortal time bias in pharmaco-epidemiology Am J Epidemiol 167: 492– 499,2008 Crossref, Medline, Google Scholar |
| 24. | DF Penson, PC Albertsen, PS Nelson , etal: Determining cause of death in prostate cancer: Are death certificates valid? J Natl Cancer Inst 93: 1822– 1823,2001 Crossref, Medline, Google Scholar |
| 25. | Cancer incidence and mortality in the United Kingdom and constituent countries, 2002-04 Health Stat Q 35: 78– 83,2007 Medline, Google Scholar |
| 26. | CR Dormuth, AR Patrick, WH Shrank , etal: Statin adherence and risk of accidents: A cautionary tale Circulation 119: 2051– 2057,2009 Crossref, Medline, Google Scholar |
| Gleason score: | A pathologic description of prostate cancer grade based on the degree of abnormality in the glandular architecture. Gleason patterns 3, 4, and 5 denote low, intermediate, and high levels of histologic abnormality and tumor aggressiveness, respectively. The score assigns primary and secondary numbers based on the most common and second most common patterns identified. |
| PSA (prostate-specific antigen): | A protein produced by cells of the prostate gland, the blood level of PSA is used as a tumor marker for men who may be suspected of having prostate cancer. Most physicians consider 0 to 4.0 ng/mL as the normal range. Levels of 4 to 10 and 10 to 20 ng/mL are considered slightly and moderately elevated, respectively. PSA levels have to be complemented with other tests to make a firm diagnosis of prostate cancer. |
|
| Statin Use | Patients (n = 1,791) | No. of Person-Years | Incidence Rate (per 100 person-years) | 95% CI (per 100 person-years) | Crude HR | Adjusted HR* | 95% CI | |
|---|---|---|---|---|---|---|---|---|
| No. | % | |||||||
| No use of statins after diagnosis | 1,250 | 69.8 | 32,283 | 3.87 | 3.66 to 4.09 | 1.00 | 1.00 | Reference |
| Type of statin used after diagnosis | ||||||||
| Use of lipophilic statins only | 493 | 27.5 | 17,107 | 2.88 | 2.64 to 3.15 | 0.77 | 0.77 | 0.67 to 0.89 |
| Use of hydrophilic statins only | 26 | 1.5 | 1,052 | 2.47 | 1.65 to 3.57 | 0.64 | 0.65 | 0.43 to 0.98 |
| Use of both lipophilic and hydrophilic statins | 22 | 1.2 | 1,102 | 2.00 | 1.28 to 2.97 | 0.56 | 0.67 | 0.43 to 1.04 |
Abbreviation: HR, hazard ratio.
*Adjusted for age, year of diagnosis, ethnicity, excessive alcohol use, smoking status, obesity, chronic kidney disease, myocardial infarction, ischemic stroke, transient ischemic attack, peripheral artery disease, previous cancers, prostate-specific antigen level, Gleason grade, metformin, sulfonylureas, thiazolidinediones, insulins, other oral antihypoglycemic agents, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, calcium channel blockers, β-blockers, diuretics, other antihypertensive drugs, aspirin, other nonsteroidal anti-inflammatory drugs, 5α-reductase inhibitors, prediagnostic statin use, prostate-specific antigen testing activity, prostatectomy, radiation therapy, chemotherapy, and androgen deprivation therapy.
|
| Exposure to Statins | Statin Use Before Diagnosis | No Statin Use Before Diagnosis | P for Interaction | ||
|---|---|---|---|---|---|
| Adjusted HR* | 95% CI | Adjusted HR* | 95% CI | ||
| Cumulative duration of use, months | |||||
| < 12 | 0.81 | 0.55 to 1.22 | 0.95 | 0.71 to 1.28 | .11 |
| 12-24 | 0.61 | 0.44 to 0.85 | 0.90 | 0.70 to 1.15 | |
| 24-36 | 0.52 | 0.36 to 0.75 | 0.64 | 0.45 to 0.90 | |
| ≥ 36 | 0.42 | 0.30 to 0.59 | 0.73 | 0.56 to 0.94 | |
| Cumulative dose, DDDs† | |||||
| < 365 | 0.67 | 0.47 to 0.94 | 0.82 | 0.64 to 1.06 | .14 |
| 365-731 | 0.56 | 0.40 to 0.78 | 0.93 | 0.73 to 1.18 | |
| 731-1,096 | 0.55 | 0.38 to 0.79 | 0.84 | 0.61 to 1.15 | |
| ≥ 1,096 | 0.43 | 0.30 to 0.62 | 0.62 | 0.45 to 0.85 | |
Abbreviations: DDD, defined daily dose; HR, hazard ratio.
*Adjusted for age, year of diagnosis, ethnicity, excessive alcohol use, smoking status, obesity, chronic kidney disease, myocardial infarction, ischemic stroke, transient ischemic attack, peripheral artery disease, previous cancers, prostate-specific antigen level, Gleason score, metformin, sulfonylureas, thiazolidinediones, insulins, other oral antihypoglycemic agents, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, calcium channel blockers, β-blockers, diuretics, other antihypertensive drugs, aspirin, other nonsteroidal anti-inflammatory drugs, 5α-reductase inhibitors, prostate-specific antigen testing activity, prostatectomy, radiation therapy, chemotherapy, and androgen deprivation therapy.
†Based on an equivalent dose of 30 mg of simvastatin.
|
| Exposure to Statins | Statin Use Before Diagnosis | No Statin Use Before Diagnosis | P for Interaction | ||
|---|---|---|---|---|---|
| Adjusted HR* | 95% CI | Adjusted HR* | 95% CI | ||
| Cumulative duration of use, months | |||||
| < 12 | 0.98 | 0.74 to 1.31 | 1.02 | 0.84 to 1.25 | .02 |
| 12-24 | 0.66 | 0.52 to 0.84 | 0.85 | 0.71 to 1.02 | |
| 24-36 | 0.65 | 0.50 to 0.84 | 0.90 | 0.72 to 1.11 | |
| ≥ 36 | 0.56 | 0.44 to 0.71 | 0.87 | 0.74 to 1.03 | |
| Cumulative dose, DDDs† | |||||
| < 365 | 0.78 | 0.61 to 1.00 | 0.90 | 0.75 to 1.06 | .02 |
| 365-731 | 0.67 | 0.53 to 0.86 | 0.93 | 0.78 to 1.11 | |
| 731-1,096 | 0.66 | 0.51 to 0.85 | 0.87 | 0.70 to 1.09 | |
| ≥ 1,096 | 0.54 | 0.42 to 0.69 | 0.88 | 0.73 to 1.07 | |
Abbreviations: DDD, defined daily dose; HR, hazard ratio.
*Adjusted for age, year of diagnosis, ethnicity, excessive alcohol use, smoking status, obesity, chronic kidney disease, myocardial infarction, ischemic stroke, transient ischemic attack, peripheral artery disease, previous cancers, prostate-specific antigen level, Gleason score, metformin, sulfonylureas, thiazolidinediones, insulins, other oral antihypoglycemic agents, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, calcium channel blockers, β-blockers, diuretics, other antihypertensive drugs, aspirin, other nonsteroidal anti-inflammatory drugs, 5α-reductase inhibitors, prostate-specific antigen testing activity, prostatectomy, radiation therapy, chemotherapy, and androgen deprivation therapy.
†Based on an equivalent dose of 30 mg of simvastatin.
|
| Statin Use | No. of Patients (n = 2,158) | No. of Person-Years | Incidence Rate (per 100 person-years) | 95% CI (per 100 person-years) | Crude HR | Adjusted HR* | 95% CI |
|---|---|---|---|---|---|---|---|
| No use of statins after diagnosis | 1,646 | 41,582 | 3.95 | 3.77 to 4.15 | 1.00 | 1.00 | Reference |
| Use of statins after diagnosis | 512 | 18,035 | 2.84 | 2.60 to 3.09 | 0.85 | 0.77 | 0.68 to 0.88 |
Abbreviation: HR, hazard ratio.
*Adjusted for age, year of diagnosis, ethnicity, excessive alcohol use, smoking status, obesity, chronic kidney disease, myocardial infarction, ischemic stroke, transient ischemic attack, peripheral artery disease, previous cancers, prostate-specific antigen level, Gleason score, metformin, sulfonylureas, thiazolidinediones, insulins, other oral antihypoglycemic agents, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, calcium channel blockers, β-blockers, diuretics, other antihypertensive drugs, aspirin, other nonsteroidal anti-inflammatory drugs, 5α-reductase inhibitors, prostate-specific antigen testing activity, prostatectomy, radiation therapy, chemotherapy, and androgen deprivation therapy.
