The Prognostic Effect of Statin Use on Urologic Cancers

Abstract

Recent studies suggest that statin may benefit cancer prognosis, especially through its radiosensitization effect. But controversy exists in other studies. Hence, we performed a meta-analysis of results from 35 studies to evaluate the effect of statin use on urologic cancers.

We conducted computerized search from PubMed, Embase, and ISI Web of Knowledge through May 2015, screened the retrieved references, and collected and evaluated relevant information. We extracted and synthesized corresponding hazard ratios (HR) and confidence interval (CI) by using Review Manager 5.3 and STATA 13. This review was registered at PROSPERO with registration No. CRD42015020171.

We selected total 35 retrospective studies and conducted a meta-analysis of results from these studies. The pooled results suggested no benefit of statin use to bladder cancer and renal cell carcinoma, except overall survival [HR = 0.81, 95% CI: 0.69–0.96]. However, significant improvement of prostate cancer prognosis including overall survival [HR = 0.82, 95% CI: 0.70–0.97] and cancer-specific survival [HR = 0.70, 95% CI: 0.59–0.83] was indicated, but not including tumor progression [HR = 0.84, 95% CI: 0.62–1.14]. Statin use improved biochemical recurrence of prostate cancer in radiotherapy patients [HR = 0.68, 95% CI: 0.54–0.85] but not in radical prostatectomy patients [HR = 0.97, 95% CI: 0.82–1.15].

Current evidence suggests no benefit of statin use to bladder cancer and renal cell carcinoma, except in overall survival. While statin use benefited prostate cancer patients in overall survival, cancer-specific survival but not in tumor progression; it also improved biochemical recurrence in radiotherapy patients but not in radical patients. To verify these results, randomized controlled trials are necessary.

INTRODUCTION

Statin is a 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitor and widely used for hypercholesterolemia patients. Recent studies prompt to indicate statin as a panacea because of its effects in treating variant diseases. A previous study showed that statin use was a protective factor for cancer incidence risk.¹ To date, statin is known as a pleiotropic drug rather than cholesterol-lowering medication. A retrospective survey of Shared Equal Access Regional Cancer Hospital (SEARCH) database indicated that triglycerides and low-density lipoproteins were associated with increased risk of prostate cancer recurrence. In the contrary, high-density lipoproteins were associated with decreased risk of prostate cancer in dyslipidemia.² It is compatible with the conclusion of another study that statin use significantly reduces breast cancer mortality.³ Similar outcomes were also observed in numerous urological cancer treatment studies. However, controversy of real effect existed in variant studies.^4,5 Hence, we aimed to conduct a meta-analysis of well selected observational studies to evaluate prognostic effect of statin use in urinary cancer treatment, limited in renal cell carcinoma, bladder cancer, and prostate cancer. This study was registered in PROSPERO with registration number CRD42015020171.⁶

METHODS

Search and Screen Strategy

We performed a systematic literature search of PubMed, Embase, and ISI Web of Knowledge to retrieve urologic cancer clinical studies using statin through May 3, 2015. We used search key words including statin, renal cell carcinoma, bladder cancer, prostate cancer, survival and mortality, etc. The detailed search strategy was described in the supplement 1. The citations in the retrieved articles were also screened for any relevant studies. The initial screen was conducted by reviewing the title and abstract by 2 independent investigators (YL and DLS) to eliminate the irrelevant articles. Then, the full-text articles were reviewed according to eligibility criteria. Any clinical study comprising the evaluation of statin use on urologic cancer prognosis was eligible. In this article, we only include results of renal cell carcinoma (RCC), bladder cancer (BC), and prostate cancer (PCa). Articles that has abstract only, duplicated literature, overlapping patients or duplicated data presented in conferences; or does not study RCC, BC or PCa; or has no data available, were excluded. In this study, all data and analyses were based on the previous published studies, and thus no ethical approval and patient consent are required.

Data Extraction and Quality Assessment

Before data collection, a spreadsheet was designed for the key information. Data extraction was independently performed by 2 researchers (YL and DLS) and cross-checked. Meanwhile, any disagreement or uncertainty was resolved by group discussion. Data extracted from the articles included the name of the first author and publication year, country, cancer type, recruitment period, number of patients, age, main treatment, follow-up, prognostic outcomes, definition of outcomes, and adjusted factors. The data were extracted from the original articles. During data extraction, multivariate outcomes were prior to univariate outcomes when both were provided, while if no multivariate results were presented, univariate outcomes were used instead. If there was no exact time to event survival data, we either estimated HR and 95% CI by the methods that were provided by Tierney et al⁷ using the given survival or mortality curve or other available data, or referred previous study outcomes⁸ or contacted the corresponding author for the original data. The extracted data from studies which have potential overlapping patients were removed before meta-analysis to avoid over-analysis. The quality assessment was carried out by using Newcastle–Ottawa Scale (NOS) for cohort study that comprised 3 domains with 8 items to evaluate bias risk.⁹ Above 5 stars of total 9 stars was deemed as good quality.

Statistical Analysis

Review Manager 5.3 (The Cochrane Collaboration, Copenhagen) was used to perform quantitative synthesis. Hazard ratio (HR) and its 95% confident interval were used to evaluate the survival outcome. First, Cochran's Q test and Higgins I² statistic were calculated for heterogeneity detection.¹⁰P ≥ 0.1 and I² ≤ 50% were deemed to no significant heterogeneity, and fixed effects model was used. Otherwise, random effects model was used. The inverse variance method was used to calculate the pooled hazard ratio. Sensitivity analysis was conducted by using the method of leave-one-out to test the feasibility of the pooled results. Publication bias was detected with Egger's regression intercept test and was only performed in outcomes comprised more than 10 studies by using STATA 13 (Stata Corp LP, College Station, TX).^10,11 A 2-tailed P < 0.05 was considered statistically significant.

RESULTS

Eligible Studies and Quality Assessment

In total, 526 abstracts were retrieved by the initial search strategy. After screening, 35 studies^4,5,12–44 including a France article³¹ were included in the qualitative and quantitative synthesis. The screening diagram was shown in Figure 1. The characteristics of included studies and the Newcastle–Ottawa Scale (NOS) quality assessment were shown in Table 1  . Outcomes included overall survival, cancer-specific survival, recurrence-free survival, progression-free survival, and biochemical recurrence.

FIGURE 1.

Literature screen diagram.

TABLE 1.

Characteristics of Included Studies

TABLE 1 (Continued).

Characteristics of Included Studies

TABLE 1 (Continued).

Characteristics of Included Studies

Survival Outcomes

In renal cell carcinoma, 4 studies were included.^15–18 Among them, 3 reported overall survival, 2 reported cancer-specific survival, and 3 reported tumor progression status. As shown in Figure 2, the pooled results of statin use in overall survival, cancer-specific survival, and tumor progression of renal cell carcinoma were HR = 0.81 (95% CI: 0.69–0.96), 0.72 (0.35–1.50), and 0.91 (0.54–1.55), respectively.

FIGURE 2.

Statin use on survival outcomes of renal cell carcinoma (RCC).

Three bladder cancer studies reported oncological prognosis.^12–14 One study reported overall survival with HR = 1.14 (0.89–1.44). Two studies reported cancer-specific survival and the pooled result was HR = 1.06 (0.87–1.29). Three studies reported recurrence-free survival with pooled HR = 1.05 (0.94–1.18). Two studies reported tumor progression and the pooled HR was 0.87 (0.65–1.15). All results were calculated by applying fixed effect model and shown in Figure 3.

FIGURE 3.

Statin use on survival outcomes of bladder cancer (BC).

Among prostate cancer studies, 5 reported overall survival outcomes and the pooled HR of statin use versus nonstatin use was 0.82 (0.70–0.97). Accordingly, 6 studies presented cancer-specific survival outcomes, the pooled result was HR = 0.70 (0.59–0.83). Tumor progression was reported in 5 studies and the pooled risk was 0.84 (0.62–1.14). All the above 3 clinical outcomes were analyzed by using the randomized effect model shown in Figure 4. Additionally, biochemical recurrence of prostate cancer became important in statin anticancer research. All prostate cancer studies subgroup were stratified by major treatment method and a subgroup analysis was conducted considering its radiosensitization effect. In radical prostatectomy subgroup, 13 studies presented biochemical result and the pooled hazard ratio of statin use versus nonstatin use was 0.97 (0.82–1.15), P = 0.73. However, in radiotherapy subgroup, 7 references reported biochemical recurrence, the pooled HR was 0.68 (0.54–0.85), P = 0.0009. Figure 5 shows the forest plot of biochemical recurrence of prostate cancer.

FIGURE 4.

Statin use on survival outcomes of prostate cancer (PCa).

FIGURE 5.

Stain use on biochemical recurrence (BCR) of prostate cancer (PCa).

Publication Bias and Sensitivity Analysis

Publication bias detection was conducted by Egger's asymmetric test only in biochemical recurrence. P value of the linear regression was 0.803 for radical prostatectomy subgroup, 0.977 for radiotherapy subgroup, and 0.463 for the entire group of prostate cancer. The results show that no significant publication bias was observed and the funnel plot is shown in Figure 6. A sensitivity analysis was conducted in prostate cancer. There is no significant change observed in cancer-specific survival, tumor progression, and biochemical recurrence after removing any included study (results were omitted).

FIGURE 6.

Funnel plot of included studies concerning biochemical recurrence (BCR).

DISCUSSION

Among medical studies, there is a great controversy on the effect of statin. Antitumor and tumor promotion effect are both presented in variant studies.^45,46 However, the mechanism of antitumor or tumor promotion effect of statin has not yet been clearly elucidated. Hindler et al⁴⁷ summarized the role of statin in cancer therapy as 4 aspects: First, statin inhibits tumor cell growth by inhibiting dolichol, geranylpyrophosphate, and farnesylpyrophosphate that are regulators of cell cycle, by inhibiting Ras and Rho that mediate cell proliferation, and by stabilizing the cell cycle kinase inhibitors p21 and p27. Second, inhibition of angiogenesis: statin has pros and cons for angiogenesis. High dose statin has an antiangiogenesis effect by inhibiting capillary tube formation and reducing vascular endothelial growth factor release. However, low-dose statin has a proangiogenesis effect by stimulating protein kinase B and activating endothelial nitric oxide synthase. Third, statin induces cell apoptosis by upregulating proapoptotic proteins and reducing antiapoptotic proteins. Fourth, statin suppresses tumor metastasis by reducing the expression of endothelial leukocyte adhesion molecule E-selectin and matrix metalloproteinase, inhibiting epithelial growth factor induced tumor cell invasion. Sun et al⁴⁸ demonstrated that cholesterol increases Ca²⁺ entry via the TRPM7 channel, which promotes proliferation of prostate cells by inducing the activation of the AKT and/or the ERK pathway. Additionally, cholesterol-mediated Ca²⁺ entry induces an increase of calpain activity that represses E-cadherin expression, which could lead to migration of prostate cancer cells. Bañez et al⁴⁹ reported that statin use significantly reduced the risk of inflammatory infiltration in prostate cancer, which was proved to be associated with cancer development and prognosis.^50–53 All the above studies attend to elucidate the possible anti-cancer mechanism of statin. While in clinical studies, there is also a great controversy on the effect of statin use for cancer patients’ prognosis. Hoffmann et al⁵⁴ reported that non-muscle-invasive bladder cancer patients, who were treated with bacille Calmette–Guérin (BCG) immunotherapy and were exposed to statin use, had worsening prognosis. This study was subsequently questioned by Kamat et al.⁵⁵ They reported no significant difference in the tumor recurrence, progression, or deaths in their cohort with 156 patients treated with BCG. A large cross-sectional study reported that statin use was associated with a reduction in the probability that older men would have an abnormal screening PSA result regardless of the PSA threshold (PSA > 2.5, > 4.0, or > 6.5 ng/mL). This reduction is more pronounced with higher statin dose, longer statin duration, and higher statin potency.⁵⁶ It revealed the anti-cancer effect of statin. Additionally, statin has a radiosensitization effect for prostate cancer both in vitro and in vivo.^25,28,57 However, some studies do not support this synergism.^21,22 Our analysis shows significant effect of statin use in prostate cancer patients underwent radiotherapy but not in patients with radical prostatectomy.

In our study, we investigated 4 clinical outcomes and 1 biochemical outcome in 3 major urologic cancers. In renal cell carcinoma treatment, statin use was associated with improvement of overall survival but not in cancer-specific survival and tumor progression. The improvement of overall survival in statin use patients was not stable and probably was derived from the protection from cardiovascular-related death.⁵⁸ In bladder cancer, no significance improvement was observed in overall survival, cancer-specific survival, or tumor recurrence and progression. A possible explanation for these results is that the inherent poor prognosis of bladder cancer overcomes the anticancer effect of statin use, or because of its intracavity. The biological behavior of bladder cancer could also be a risk for drug effect because even though in chemotherapy, numerous drug resistant or insensitive bladder cancer existed.⁵⁹ In prostate cancer, significant improvements of overall survival and cancer-specific survival, not tumor progression, were observed. Meanwhile, biochemical recurrence of PSA was intensively analyzed. Referred to previous study,⁸ stratification by major treatment methods was performed. In radical prostatectomy subgroup, no difference was observed between statin use and non-use. However, statin use significantly improved biochemical recurrence in prostate cancer patients treated with radiotherapy. Sensitivity analysis described as above did not alter the results. It is compatible with the radiosensitization effect of statin use. There seemed to be a paradox that statin use did not improve biochemical recurrence in radical prostatectomy patients but improve overall survival and cancer-specific survival. A hypothesis was that these benefits derived from radiation therapy patients. But it has not been verified. Additionally, as to clinical outcomes, small study effect was an obvious risk for the pooled results. Generally, statin use seemed to benefit prostate cancer, especially prostate cancer patients underwent radiotherapy.

Additionally, all studies included or excluded were obviously biased. Statin, unlike chemotherapeutic drugs, is a gentle medication for cancer, if it has the anticancer effect. However, the accumulative effect of statin use is unclear yet. On the other hand, the definition of statin use has not been clearly elaborated. The statin category is also different from each type. Additionally, statin use was a time-dependent covariate in these survival cohorts. Stratifying statin users by records of pre or at cancer diagnosis or treatment is not appropriate. First, the duration of statin use is volatile. A man consumed statin for 5 years is different from the man with 5-month statin consumption. Second, the statin use status of included patients was also volatile during the follow-up. For example, a statin use patient could discontinue statin consumption. Mostly, those nonstatin users could consume statin after diagnosis or treatment of cancer. This would obviously confound the survival analysis. Based on thus, randomized controlled trials would help verify this benefit. Additionally, it is too early to apply statin medication to urologic cancer patients. Adverse effect, dose, and economic factors were important obstacles that must be overcome before application though significant benefit was verified by randomized controlled trials in the future.

CONCLUSION

Our meta-analysis summarized the published literature with statin use exposure and the pooled results suggested no benefit of statin use to bladder cancer and renal cell carcinoma, except in overall survival. However, significant improvement of prostate cancer prognosis including overall survival and cancer-specific survival was indicated, but not including tumor progression. Statin use improved biochemical recurrence of prostate cancer in radiotherapy patients but not in radical prostatectomy patients. Randomized controlled trials would help verify these results.