Association Between Statin Use and Prognosis of Breast Cancer: A Meta-Analysis of Cohort Studies

Background: Statin, a lipid-lowering drug, has been suggested to confer anticancer efficacy. However, previous studies evaluating the association between statin use and prognosis in breast cancer showed inconsistent results. A meta-analysis was performed to evaluate the association between statin use and clinical outcome in women with breast cancer. Methods: Cohort studies comparing recurrence or disease-specific mortality in women with breast cancer with and without using of statins were identified by search of PubMed, Embase, and Cochrane's Library databases. A random-effect model, incorporating the inter-study heterogeneity, was used to combine the results. Subgroup analyses were performed to evaluate the influences of study characteristics on the outcomes Results: Seventeen cohort studies with 168,700 women with breast cancer were included. Pooled results showed that statin use was significantly associated with a lower risk of breast cancer recurrence (adjusted hazard ratio [HR] = 0.72, p < 0.001) and breast cancer mortality (HR = 0.80, p < 0.001). Subgroup analysis showed that timing of statin use, statin type, study design, sample size, or quality score did not significantly affect the outcomes. However, statin use was associated with more remarkably reduced breast cancer recurrence in studies with mean follow-up duration ≤ 5 years (HR = 0.55, p < 0.001) than that in studies of >5 years (HR = 0.83, p = 0.01). Conclusions: Statin use is associated with reduced recurrence and disease-specific mortality in women with breast cancer. These results should be validated in randomized controlled trials.


INTRODUCTION
Although advances have been achieved in the prevention and treatment of breast cancer in recent decades, the disease remains one of the most common malignancies in women (1,2). It has been reported that ∼1.4 million women are diagnosed as breast cancer each year all over the world, and breast cancer remains an important cause of mortality in women (3,4). The 3-hydroxy-3-methylglutaryl CoA (HMG-CoA) reductase inhibitors, also known as statins, are the most commonly use lipid-lowering medications which have become a cornerstone for the prevention and treatment of atherosclerotic cardiovascular diseases (5). Accumulating evidence revealed that statins have various potential pharmacological effects besides their lipid-lowering efficacy, such as anti-inflammation, anti-proliferation, and anti-invasion, pro-apoptosis, immunomodulation, which are all involved in the pathogenesis of cancer (6,7). These findings highlight the potential role of statins as anticancer agents (8). Although previous studies generally did not show that statin use is related with reduced risk of breast cancer incidence (9)(10)(11), some cohort studies showed that compared with the non-users, users of statin with breast cancer may have better clinical outcomes (12)(13)(14)(15)(16)(17). However, other cohort studies did not show that statin use in women with breast cancer was associated with improved prognosis (18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28). Although several meta-analyses have been performed to evaluate the association between statin use and prognosis in women with breast cancer (29)(30)(31)(32), only studies published before 2017 were included, and the limited number of studies prevented a comprehensive evaluation of the impacts of study characteristics on the outcomes. Therefore, we aimed to perform an updated meta-analysis regarding the association between statin use and prognosis in breast cancer, by incorporating of the recently published cohorts that were not included in previous meta-analyses (17,(24)(25)(26)(27)(28). The relative large number of available studies enables us to perform comprehensive analyses regarding the influences of study characteristics on the outcomes.

METHODS
The meta-analysis was designed and performed in accordance with the MOOSE (Meta-analysis of Observational Studies in Epidemiology) (33) and Cochrane's Handbook (34) guidelines.

Literature Search
Electronic databases of PubMed, Embase, and the Cochrane's Library were systematically searched using the combination of the following terms: (1) "statin" OR "3-hydroxy-3-methylglutaryl CoA reductase inhibitor" OR "CS-514" OR "simvastatin" OR "atorvastatin" OR "fluvastatin" OR "lovastatin" OR "rosuvastatin" OR "pravastatin" OR "pitavastatin"; (2) "breast cancer"; and (3) "survival" OR "prognosis" OR "mortality" OR "death" OR "recurrence" OR "surgery" OR "operation." The search was limited to human studies with no restriction of publication language. The reference lists of original and review articles were also manually analyzed. The final literature search was performed on February 24, 2020.

Study Selection
Studies were included if they met the following criteria: (1) published as full-length articles; (2) designed as cohort studies with the minimal follow-up duration of 1 year; (3) included women with breast cancer; (4) use of statin was identified as exposure of interest; (5) documented the incidence of breast cancer recurrence or breast cancer mortality during follow-up; and (6) reported the adjusted hazard ratios (HRs, at least adjusted for age) and their corresponding 95% confidence intervals (CIs) for the above outcomes in women with breast cancer with and without the use of statin. Reviews, editorials, preclinical studies, and non-cohort studies were excluded.

Data Extracting and Quality Evaluation
Literature search, data extraction, and study quality assessment were independently performed by two authors according to the predefined inclusion criteria. If inconsistencies occurred, discussion with the corresponding author was suggested to resolve these issues. The following data were extracted: (1) name of the first author, publication year, country, and study design; (2) characteristics, number, and mean age of women with breast cancer, definition and timing of statin use, and follow-up period; and (3) number of cases with breast cancer recurrence and breast cancer mortality, and the adjusted variables when presenting the HRs. The quality of each study was evaluated using the Newcastle-Ottawa Scale (NOS) (35). This scale ranges from 1 to 9 stars and judges the quality of each study regarding three aspects: selection of the study groups; the comparability of the groups; and the ascertainment of the outcome of interest.

Statistical Analyses
The associations between statin use and breast cancer recurrence and mortality were measured by HRs in this study. To stabilize its variance and normalized the distribution, HR data and its corresponding stand error (SE) from each study were logarithmically transformed (34). The Cochrane's Q-test was performed to evaluate the heterogeneity among the include cohort studies (34,36), and an I 2 statistic was also calculated. A significant heterogeneity was considered if I 2 > 50%. A randomeffect model was used to pool the results since this model has been  Cardwell et al. Bjarnadottir et al.
In the original manuscript from Li, the HR results for breast cancer recurrence were reported according to the time of statin use (<3, 3-5, and >5 years) (27). These results were firstly pooled with a random-effect model to generate a data of HR for women with statin use of any time compared to non-users, and then the pooled HR was included in the meta-analysis. The heterogeneity among these studies was not significant (P for Cochrane's Q-test = 0.20, I 2 = 26%). Pooled results with a random-effect model showed that statin use was associated with a significantly reduced breast cancer recurrence (adjusted HR = 0.72, 95% CI: 0.60 to 0.86, p < 0.001; Figure 2A subgroup difference = 0.14; Figure 3B). In addition, subgroup analysis also showed that difference in study design, sample size, NOS, and adjustment of menopausal status, hormonal receptor status, or comorbidities did not significantly affect the results (p for subgroup difference all >0.10; Table 3). However, statin use was associated with a more remarkably reduced breast cancer recurrence in studies with mean follow-up duration ≤5 years (HR = 0.55, p < 0.001) than that in studies with mean followup duration >5 years (HR = 0.83, p = 0.01; p for subgroup difference = 0.009; Table 3).

Association Between Statin Use and Breast Cancer Mortality
Meta-analysis of nine cohort studies (14-17, 19, 21, 22, 25, 28) showed that statin use was associated with a significantly reduced risk of breast cancer mortality (adjusted HR = 0.80, 95% CI: 0.72 to 0.90; p < 0.001) with significant heterogeneity (I 2 = 55%; Figure 4A Figure 5B). Furthermore,  subgroup analysis also showed that differences in study design, sample size, follow-up duration, NOS, adjustment of menopausal status, hormonal receptor status, or comorbidities did not significantly affect the results (p for subgroup difference all >0.10; Table 3).

Publication Bias
The funnel plots for the associations between statin use and breast cancer recurrence and mortality were shown in Figures 6A,B. The plots were symmetrical on visual inspection, suggesting low risks of publication biases. Results of Egger's regression tests also showed similar results (p = 0.328 and 0.384, respectively).

DISCUSSION
In this meta-analysis of cohort studies, we found that compared to the non-users, statin use was associated with significant lower recurrence and disease-specific mortality in women with breast cancer, even after adjustment of potential confounding factors including age, cancer stage, and anticancer treatments. Subgroup analysis showed that study characteristics such as timing of statin use, statin type, statin exposure time, study design, sample size, quality score, adjustment of menopausal status, hormonal receptor status, or comorbidities did not seem to significantly affect the association between statin use and improved prognosis in women with breast cancer. However, a more remarkably reduced breast cancer recurrence was observed in studies with shorter follow-up duration (≤5 years) compared to that in studies with longer follow-up duration (>5 years). Taken together, these findings suggest that statin use is associated with reduced recurrence and disease-specific mortality in women with breast cancer, which supports the implementation of a randomized clinical trial. Several meta-analyses have been performed to evaluate the association between statin use and prognosis in women with breast cancer (29)(30)(31)(32). Although results of these meta-analyses were generally consistent the overall results of our metaanalysis, these studies only included five to eight cohort studies, which prevented subsequent analyses for the influences of study characteristics on the outcomes. A previous meta-analysis by Liu et al. published in 2017 showed that the relationship between statins use and breast cancer was remarkable in studies with lipophilic statins and statin exposure of <4 years (32). However, only seven cohorts were included in this meta-analysis, and the authors used the mean follow-up year as a reflection of statin exposure year, which made the results less reliable (32).
Compared to previous meta-analyses, our study has the following strengths. Firstly, we included up-to-date evidence from related cohort studies, which included 17 studies with 168,700 women with breast cancer. This large number of studies enables us to perform comprehensive subgroup analyses based on the data of study level. Secondly, only studies with multivariate analyses were included. Therefore, our study results indicated that statin use was independently associated with improved prognosis in women with breast cancer. Thirdly, sensitivity analyses were used to evaluate the stability of the results, which showed that the overall meta-analysis results were not affected by either of the included study. Finally, results of subgroup analyses suggested that statin was associated with a more remarkably reduced breast cancer recurrence in studies with shorter follow-up duration (≤5 years) compared to that in studies with longer follow-up duration (>5 years). One possible explanation for this finding may be that compared to shortterm recurrence, mechanisms responsible for the long-term recurrence of breast cancer could be more complicated, and the potential protective efficacy of statins might be weakened. Moreover, it has been reported that triple negative breast cancer tends to recur in <5 years whereas hormone receptor positives have longer periods of dormancy (39). The difference in molecular subtype of breast cancer may be accounted for the subgroup results. However, we could not confirm this hypothesis because the molecular subtype of breast cancer was generally not reported in studies included in the subgroup analysis according to statin exposure time. Besides, it has to be mentioned that since the exposure time of statin in each study is not necessarily correlated with the follow-up time. Therefore, the finding of the subgroup analysis may be less clinically relevant.
The mechanisms underlying the potential association between statin use and lower breast cancer recurrence and mortality remain largely unknown at current stage. A previous cohort study including 191 Korean women with breast cancer who underwent resection showed that a higher tumor expression of HMG-CoA reductase was associated with poor diseasefree survival, which suggests that the potential benefit of statin on clinical outcomes in breast cancer may involve its pharmacological effect on HMG-CoA reductase inhibiting (40). A recent study in Swedish women with breast cancer who were on statins also showed similar finding (28). More direct evidence comes from a recent experimental study, which showed that induction of tumor expression of HMG-CoA reductase led to resistance to statin induced deaths of breast cancer cells (41), which further demonstrated that the benefits of statins in breast cancer are at least partially depending on their inhibition of HMG-CoA reductase. Besides, preclinical studies also suggest that statin may exert anticancer efficacy in breast cancer via other mechanisms. It has been suggested that inhibition of protein prenylation involved in signaling pathways of carcinogenesis and cancer progression may be halted as a downstream effect of HMGCR inhibition by statins (42). In addition, simvastatin was shown to inhibit breast tumor angiogenesis via impeding hypoxia-inducible factor-1α-induced pro-angiogenic factors (43). Moreover, atorvastatin was found to inhibit the activity of breast cancer cells via inducing autophagy (44). In addition, lovastatin could mediate MCF-7 cancer cell death by interaction with p53-survivin signaling cascade (45). Taken together, the mechanisms underlying the potential benefits of statins in breast cancer are likely to be multifactorial, and further studies are warranted to determine the key molecular pathway involved.
Our study has limitations, which should be considered when interpreting the results. Firstly, although we combined HR data after multivariate adjustment, residual factors that potentially confound the association between statin use and prognosis in breast cancer may remain existing. Secondly, definition and exposure time of statin use varied among the included studies. Although our stratified analyses did not show that timing, category, or exposure time of statin use may significantly affect the outcome, these results should be validated in randomized clinical trials. In addition, our results of subgroup analyses were based on data of study level rather than individual patient level. The findings of subgroup analyses should be validated in large-scale prospective studies. Finally, a causative relationship between statin use and improved prognosis in women with breast cancer should not be retrieved from our results. Randomized clinical trials are needed to confirm whether additional treatment with statin could improve the clinical outcomes in women with breast cancer.
In conclusion, our meta-analysis showed that statin use was associated with significant reduced recurrence and diseasespecific mortality in women with breast cancer. These findings support the implementation of a randomized clinical trial to evaluate the potential benefits of statins on clinical outcomes in women with breast cancer.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/supplementary files.