The effect of statins on prostate cancer recurrence and mortality after definitive therapy: a systematic review and meta-analysis

In this work, we aim to further analyze the association of statins use with biochemical recurrence (BCR) of prostate cancer (PCa) and PCa-specific mortality after definitive therapy. A systematic literature search of PubMed, MEDLINE, and EMBASE through Jul 2015 was conducted. Pooled Hazard ratio (HR) estimates with corresponding 95% confidence intervals (CIs) were calculated using random-effects model. STATA version 10 (Stata corporation, college station, TX) was employed to conduct all statistical analyses. A total of 22 and 8 studies contributed to the biochemical recurrence analysis and PCa-specific mortality, respectively. 13 trials were included for BCR-free survival analysis. The combined result showed statins users had lowered 12% BCR risk of PCa compared with non-users (HR = 0.88, 95%CI: 0.765–0.998) (p < 0.05). The association was null among the men who underwent radical prostatectomy as primary therapy (HR = 0.96, 95%CI: 0.83–1.09), while the improved outcomes had be seen among patients who received radiation therapy (HR = 0.67, 95%CI: 0.48–0.86). After excluding the patients undergoing ADT, participants did not benefit from statins use (HR = 0.94, 95%CI: 0.77–1.11). Meanwhile, long-term statins using did not alter recurrence risk. A lower risk of prostate cancer-specific mortality was observed among statins users (HR = 0.68, 95%CI: 0.56–0.80). There was a plausible trend towards increasing the BCR-free survival rate among statins users.

Scientific RepoRts | 6:29106 | DOI: 10.1038/srep29106 Methods Study selection. We performed a literature search without language restrictions using the databases of PubMed (Jan 1967-April 2015), MEDLINE (Jan 1967-April 2015), EMBASE (Jan 1990-April 2015 to include all studies that investigate the association of statins use with prostate cancer outcomes. In addition, a manual search was conducted to identify additional relevant studies. After removing duplicate publications, two reviewers (Tan & Wei) independently assessed all remaining results by checking titles and abstracts. All publications, including abstracts, were eligible for retrieval. When studies reported outcomes from similar or overlapping databases or cohorts, only data from the most recent publication were included. The studies included should analyze the BCR, PCa-specific mortality, or BCR-free survival. BCR defined as a post-treatment PSA value of 0.2 ng/ml or greater in men who underwent radical prostatectomy; nadir PSA level + 2 ng/ml (Phoenix criteria), for men treated with radiation therapy; or any PSA increase in men treated with primary androgen deprivation therapy; and not any evidence of clinical and/or radiographically detected disease. We adapted a PRISMA (preferred reporting items for systematic reviews and meta-analyses) flow-chart to depict the study selection.
Data extraction and analysis. Data from each study were independently extracted by two reviewers (Tan & Wei) using a standardized data-extraction form. Any disagreements were resolved by consensus or by consultation with a third reviewer (Yang). Adjusted multivariate Hazard ratio (HR) with corresponding 95% confidence intervals (CIs) were used to assess potential association between statins use and BCR and mortality of prostate cancer following treatment. In addition, we also tried to contact authors via e-mail to obtain further information that had not been reported in their published articles. We identified heterogeneity between studies using the standard Cochran's Q test with a significance level of α = 0.10. We also examined heterogeneity with the I 2 statistic, which quantifies inconsistency across studies to assess the impact of meta-analysis heterogeneity. An I 2 statistic of 50% or more indicates a considerable level of heterogeneity. When heterogeneity was found, we attempted to determine potential sources of heterogeneity through stratification by various factors, and inference analysis and exclusion sensitivity analyses. Publication bias was detected using both the Begg's and Egger's tests. Statistical significance was determined using the two-tailed test, where P < 0.05 was considered significant. STATA version 10 (Stata corporation, college station, TX) was employed to conduct all statistical analyses.

Results
A total of 3229 publications were identified during the initial search (see Additional file 1.), and after employing exclusion criteria, a total of 22 studies were included for PCa recurrence analysis   (Table 1), 8 trials were available for PCa-specific mortality 22,[30][31][32][33][34][35][36] (Table 2), and 13 for BCR-free survival analysis 8,9,11,14,17,20,25,29,[36][37][38][39][40] (Table 3). Among 22 studies included in meta-analysis for PCa recurrence, 14 included a population primarily treated with radical prostatectomy (RP). Meanwhile, men in seven studies were treated with radiation therapy (RT-either external beam, brachytherapy or a combination of them). And other one study in which included patients were treated with RP, RT with or without ADT (androgen deprivation therapy) or ADT only 22 . Among 8 trials included for PCa specific mortality, one included a population of diabetic men 34 . Two abstract were included in BCR-free survival analysis.

Statins and biochemical recurrence risk of PCa.
Most studies demonstrated that statins had a neutral effect on recurrence after prostate cancer diagnosis. Overall, the combined result showed statins users had lowered the risk of recurrence compared with non-users (HR = 0.88, 95%CI: 0.765-0.998, I 2 = 65.6%, p < 0.001) (Fig. 1). The association was null among the men who underwent RP as primary therapy (HR = 0.96, 95%CI: 0.83-1.09, I 2 = 57.5%, p = 0.004). However, the combined result showed an inverse association between statins use and recurrence after treatment of RT (HR = 0.67, 95%CI: 0.48-0.86, I 2 = 60.9%, p < 0.001). It is really important to exclude patients undergoing ADT, in order to reduce the correlation between metabolic syndrome secondary androgen deprivation and BCR, as reported in literature. After excluding patients undergoing ADT, 11 studies were included and the combined estimates suggested that no patient had benefit from statins use after local therapy (HR = 0.94, 95%CI: 0.77-1.11) (Fig. 2). Additional file 2 and file 3 shown, long-term statins use still cannot alter recurrence risk (HR = 0.90, 95%CI: 0.72-1.07). No publication bias was detected among 22 studies when Begg's and Egger's tests were conducted (p = 0.81, p = 0.87, respectively) (see Additional file 4).
In the stratified analysis, the pooled HR estimates suggested that reverse relationship between statins use and PCa recurrence was observed in the US (HR = 0.86, 95%CI: 0.73-0.99). However, we failed to find a similar relation in the non-US (HR = 1.01, 95%CI: 0.67-1.35). When stratified analyses were performed on the results with or without adjusted for age, BMI, or PSA, we observed that statins use had a neutral effect on PCa recurrence when adjusted for them. Post-operation and pre-operation statins use both cannot alter PCa recurrence risk. Detailed data are illustrated in Table 4.
While there was limited data of the effect of statins on local recurrence and metastasis of PCa after treatment. For low-, intermediate-and high-risk group, statins use did not affect the recurrence risk 14 . The Mondual et al. also observed that statins using was not associated with the progression to metastasis or death of PCa after treatment 16 . Statins and prostate cancer specific mortality. Eight studies evaluated exposure to statins and the recurrence of PCa. Figure 3 graphed the HRs and 95%CIs from individual studies and the pooled results. Among them, six studies showed an inverse association. The combined estimates of eight studies showed a decreased risk (HR = 0.68, 95%CI: 0.56-0.80) ( Table 5). After excluding Serruys et al. 30 , due to its large confidence interval (HR = 0.98, 95%CI: 0.14-6.92), the result of seven studies was consistent and stable (HR = 0.68, 95%CI: 0.55-0.80). There was no publication bias (Begg's test, p = 0.386; Egger's test, p = 0.063). After adjusting for age, BMI, or PSA, the same trends were observed in these subgroups (see Table 5).
Scientific RepoRts | 6:29106 | DOI: 10.1038/srep29106 Effect of statins on BCR-free survival. Table 3 showed 13 retrospective studies were included to evaluate the effect of statins on biochemical recurrence-free survival. The 5-year BCR-free survival rate of statins users ranged from 63.7 to 97.2% and that was 57-89.6% for non-users. The statins users had much higher 10-year BCR-free survival rate compared with non-users. Sensitivity analysis. Sensitivity analyses were conducted to evaluate the effect of each study on the overall estimates by sequentially excluding each study in turn. In our meta-analysis, we found that probably no study could affect the summary of risk estimates, means that the results remained stable (data not shown).

Discussion
In this meta-analysis among men following definitive treatment of localized prostate cancer, approximately 12% reduction risk of recurrence in statins users was found (p < 0.001). In cumulative meta-analysis, a reduction risk was found for the first time in 2010, when Hamilton et al. 12 joined the analysis. When analyzed by treatment modality, statins use had a neutral effect on BCR risk among men who underwent radical prostatectomy. On the contrary, there was almost 33% lower risk of recurrence among statins users who received RT as primary treatment modality. These are consistent with the results from subgroup patients who did not receive adjuvant ADT. The benefits were null when estimates controlled for age, BMI or PSA. Overall, there was 32% lower risk of prostate cancer-specific mortality among statins users. To our knowledge, this is the first meta-analysis to address the relationship between PCa-specific mortality and statins use.
Whether the improved outcomes seen among patients who underwent RT reflecting that statins directly inhibit PCa progression, act as a radiosensitizer, or are the result of either residual confounding or type I error of studies is unknown. One postulated possible explanation was statin-induced radiosensitizing effects. Prostate cancer cell death was increased in both in vitro and in vivo models when combined the statins and ionizing radiation, which might attributed to the MYC oncogene, as statins decreasing cellular MYC levels 11,41 . Statins also play a synergistic role with radiation by promoting autophagy pathway, causing prostate cancer cell death 42 . Moreover, Gutt et al. 11 proved that a larger benefit was seen in men treated with lower radiation doses. Clinically, RT combined with ADT were often used to treat PCa. In our work, seven studies analyzed men who treated by RT, but six of them included patients who received ADT. As ADT could reduce PCa recurrence risk, the effect of statins could not be confirmed there.
Intriguingly, our findings do not support the hypothesis that statins, when taken at low dose for managing hypercholesterolemia, can reduce the risk of PCa recurrence among men treated with RP as primary therapy. Although many studies found the statins had antineoplastic properties, two precious high quality meta-analyses had confirmed that statins use could not reduce the risk of prostate cancer 5
is that statins are associated with a dose-dependent reduction in the risk of biochemical recurrence. The effect of statins on lowering recurrence risk was observed only when taken at doses ≥ 1 DE (dose equivalents) 12 . Many in vitro studies that demonstrated these protective effects also used statins concentrations that were much higher than those seen with standard therapeutic use. This meta-analysis suggested that regular statins use reduced 32% risk of PCa-related mortality. Meanwhile, Marcella et al. 31 evaluated effect of statin type on prostate cancer mortality, which showed that hydrophilic and lipophilic statins both lowered the risk of PCa-specific mortality. However, there was a trend based on potency in which high-potency statins were associated with a 73% risk reduction, whereas this effect was not observed in low-potency statins.
One of 13 trials evaluating the BCR-free survival suggested that statins users had a lower 5-year BCR-free survival rate compared with non-users 17 , while some reveled reversely 9,25,37 or showed there was no difference between two groups 20 . Meanwhile, the BCR-free survival rate at 8 or 9-year were still inconsistent 8,14,40 . However, the BCR-free survival rate at 10 year among statins users was much higher compared with non-users 36,38 . These inconsistencies might be attributed to the composition of risk group of patients, as lower-risk group had higher 5-year BCR-free survival rate compared with intermediate-risk and high-risk group 9,11,14,17 . Meanwhile, treatment modality and the radiation doses among men who underwent RT also contributed to this inconsistency 12 . Overall, there is a plausible trend towards increasing the BCR-free survival rate among statins users.
Given the known that income, education and health insurance coverage influence access to appropriate early detection, statins users are more likely to get PSA testing done and treatment, leading percent of low or intermediate-risk patients much higher among statins users than that in non-users. This might cause an illusion that statins have protective effect. The possible confounding effect arising due to the indications for which statins are prescribed also needs to be emphasized. Statins users are more likely to be obese or older as compared to non-users. This could affect PCa outcomes or progression. The subgroup analyses of 6 studies which adjusted for BMI or age reveled a neutral effect on BCR. This effect was also observed in the combined estimates of four studies adjusted for both BMI and age.

Study (year) Country
Duration of followup (Mean ± SD ) All male subjects (%)  Table 3. Characteristics of 13 studies evaluating the effect of statins on Biochemical Recurrence-Free Survival. * NR, not report. # RT, radiation therapy. $ RP, radical prostatectomy.
A previous meta-analysis, including 8 studies, initially addressed the relationship between statins use and PCa recurrence and showed that there was no association between statins use and recurrence after prostate cancer diagnosis (HR = 0.91, 95%CI: 0.72-1.13), this effect was consistent with men received RP or RT 45 . After that, another meta-analysis included 13 studies in the formal meta-analysis was conducted, which showed the same relationship (HR = 0.90, 95%CI: 0.74-1.08), while a reverse association was initially observed among men accepted RT (HR = 0.68, 95%CI: 0.49-0.93) 7 . Compared with Park et al. 7 , after including 9 more new studies  that provided stronger evidence, we found a reverse association with statins among men following definitive treatment. Besides, Oh et al. 25 updated their data in 2014, which may also contribute to this inconsistency. Unexpectedly, the combined result of these 10 studies showed a neutral effect of statins use on recurrence risk (HR = 0.88, 95%CI: 0.70-1.05).
Some limits in our meta-analysis should be mentioned. First, the available literature was limited to English language, which might be subject to selective bias and confounding. Second, there was significant heterogeneity among studies that might be attributed to various percent of risk groups. In a study, the beneficial effect was limited to high-risk patients, which proved that risk groups could affect the recurrence-free survival 14 . While, the data in different risk groups was limited; thus, the future studies addressing this topic should take the risk groups into account. Another potential factor contributed to this heterogeneity was starting time of statins use, which means some patients started using statins before operation (RT or RP or both), some started using during operation, while others received statins for the first time after operation. Third, the overall result showed that statins use lowered the recurrence of PCa, while the upper confidence interval close to 1.00. Therefore the result of this part needs to be explained with caution. At last, limited prospective sequential data was available at the moment, lowering the strength of evidence in this meta-analysis.

Conclusion
In summary, our work demonstrated that statins use could not lower PCa recurrence risk among all patients underwent RP. Although the improved outcomes might be seen among statins users who underwent RT, more  future studies excluded men who received ADT pre-or post-radiation therapy should be done. In addition, statins use decreased the prostate cancer-specific mortality risk. A plausible trend towards increasing the BCR-free survival rate among statins users was also observed.