Statin use and breast cancer survival – a Swedish nationwide study

Breast cancer tends to occur in older women with a median age at diagnosis of 60 years in developed countries. Consequently, many breast cancer patients may be treated for both pre- and post-diagnostic comorbidities. Frequently prescribed drugs are statins, which are cholesterol-lowering HMG-CoA reductase inhibitors used for prevention of cardiovascular diseases. In Sweden, around 20% of the adult female population is currently prescribed a statin according to The National Board of Health and Welfare of Sweden (www.​socialstyrelsen.​se). Statins block the rate-limiting step in the mevalonate pathway by inhibiting HMG-CoA reductase, predominantly in hepatocytes [1]. The reduction in intracellular cholesterol levels in the liver induces up-regulation of the low-density-lipoprotein receptor resulting in lower levels of circulating cholesterol. Beyond cholesterol metabolism, the mevalonate pathway is essential for tumor promoting effects of the oncogene p53 [2]. Additionally, isoprenoid production plays a fundamental role in cell growth regulation [3]. Recent observational studies have reported an inverse association between pre-diagnostic statin use and risk of breast cancer-related death [4, 5]. These findings have, however, been replicated to a lesser degree in studies performed outside of Scandinavia [6, 7] A recent systematic review summarized relative risks of statin use with breast cancer recurrence [8]. Given the epidemiological findings from other Scandinavian populations, we hypothesized that statins may also have anticancer effects and therefore reduce cancer related mortality in a Swedish population. In this nationwide study, we investigated breast cancer related and all-cause mortality among breast cancer patients according to statin use prior to and/or following their breast cancer diagnosis.

This nationwide study included 20,559 Swedish women diagnosed with breast cancer after 40 years of age between July 1st, 2005 and through 2008, Fig. 1. The average age at diagnosis was 64 years. During follow-up until December 31st, 2012, a total of 2669 breast cancer-related deaths occurred among the 20,559 women, who were diagnosed after 1st January 2006, (Table 2). In this entire study cohort, 2518 patients were prescribed a statin before their breast cancer diagnosis. For analyses of regular use and prognosis, the cohort of women with more than two years of follow-up were used (N = 8809), and among these 1092 were regular statin users (Table 1). The lipophilic statin, simvastatin, was the most frequently prescribed statin accounting for 82% (among the 17,804 patients considered for pre-diagnostic statin use) of all recorded statin prescriptions followed by pravastatin (hydrophilic), atorvastatin (lipophilic), rosuvastatin (hydrophilic), and fluvastatin (lipophilic). The numbers for post-diagnostic use similarly showed that simvastatin was the most frequently prescribed statin (93%) followed by pravastatin, rosuvastatin, lovastatin, and atorvastatin.

When comparing use of pre- and post-diagnostic statins, the majority of pre-diagnostic statin users would expectedly also be post-diagnostic statin users (N = 2503/2739, 91.3%). In addition, a substantial number of the non-pre-diagnostic statin users would become users of statins in the post-diagnostic setting (N = 2082/17,801, 11.7%).

This large-scaled nationwide Swedish study demonstrates that the use of cholesterol-lowering statins is associated with lower risk of breast cancer related and overall deaths among women diagnosed with breast cancer, irrespective of whether statins were used pre- or post-diagnosis. These results confirm previously presented studies based on Scandinavian cohorts [4, 5, 9] although the results have been less evident in studies from England [12], Ireland [7], and Scotland [6]. The background for these disparities is not known and may reflect both biological differences between Scandinavian and British populations, but differences in the data available may as well influence the varying results given that the number of breast cancer cases included vary across studies in addition to the available co-variates included. In line with our results, a recent comprehensive meta-analysis addressing statin use and breast cancer survival demonstrated significant survival benefits among statin users, both in terms of overall survival and disease-specific survival, which was true for both pre- and post-diagnosis statin use [13]. Another systematic review on statin use and cancer mortality, including breast cancer mortality, showed consistent findings [14]. In summary, consistent observational evidence support a reduced risk of breast cancer recurrence/mortality among statin users [4, 5, 15‐18], which have paved the way for multiple ongoing clinical trials, which are investigating the role of statins in breast cancer (such as #NCT02483871, #NCT02958852, #NCT01988571). Nevertheless, treatment predictive biomarkers enabling a more targeted approach and selection of patients likely to respond to statin treatment is required [19]. Recently, a novel multigene signature comprised of genes involved in the cholesterol biosynthesis, was shown to predict statin sensitivity [20].

Our current findings are biologically plausible since statins inhibit not only cholesterol synthesis but also reduce other important downstream products, of which several are used in cell proliferation such as membrane integrity maintenance, cell signaling, protein synthesis, and cell-cycle progression [21‐23]. Interruption of these processes in potentially remnant malignant cells following primary breast cancer may therefore inhibit further cancer growth and metastasis as depicted by an improved clinical outcome. Not only can statins have a direct impact on cancer cells through inhibition of the mevalonate pathway within the cancer cells, but the reduction of circulating cholesterol levels through hepatic pathways is indeed considered important, especially for hormone-dependent cancers such as breast cancer [24]. Cholesterol serves as a fundament building block for all steroid-based hormones, and disrupted estrogen synthesis might be beneficial for estrogen receptor (ER)-positive breast cancer. The cholesterol metabolite 27-hydroxycholesterol has been demonstrated to act as an ER-ligand that is able to promote progression of estrogen-dependent breast cancer [25]. A recently conducted breast cancer trial demonstrated statins capability to effectively reduce circulating levels of 27-hydroxycholesterol [20], thus providing suggested mechanisms of the proposed systemic statin effects in obstructing breast cancer progression.

Some limitations of this study need to be acknowledged. This large-scaled nationwide study consisting of 20,559 breast cancer patients was limited to analyzing breast cancer related and overall mortalities, whereas considering endpoints including recurrences was not possible since data on breast cancer recurrences are not yet available on a national level in Sweden. In this study, results are based on a median follow-up time of 61.6 months, which allows for interpretations of statins’ effects regarding early clinical outcome of breast cancer, whereas later occurring mortality cannot currently be depicted in this study population. The study cohort was originally designed to include breast cancer cases from year 2000 to 2008 with exclusion of women with a previous cancer, which explains the exclusion period from 1958 throughout 1999. Subsequently, the cohort was restricted to breast cancer cases from 2005 and onwards to allow for valid data on statin use with prescription data being available from that time point. Thus, data on previous cancers in the period 2000–2004 was not available in the cohort of this particular study. Statin use and breast cancer survival are associated with potential confounding factors, which may bias the results of this study. In this national study, analyses were adjusted for age, tumor stage, and diabetes, however there was no access to information regarding adjuvant treatment, which preferably should have been controlled for in the multivariate analyses. This drawback is balanced by the fact that all patients included in this study received their adjuvant treatment during a very limited period of time (2006–2008) and, importantly, no major changes in adjuvant treatment were introduced during these years (i.e. adjuvant trastuzumab was introduced in Sweden after this time point). Additional data, which may have been of interest to control for, were life-style factors and co-medications, such as aspirin. However, aspirin can be dispensed over-the-counter and are not necessarily prescribed, which can lead to misclassification. Considering the associations between statin use and overweight [26] in addition to impaired breast cancer survival among overweight and obese individuals [27], it would have been preferred to control for body mass index as a confounder in the survival analyses. In parallel with the Danish study [4], statin use in this study was dominated by simvastatin, which may limit the generalizability to statins in general. In addition, the results from this study are based on data from a Swedish population and might not apply to other countries as indicated by the disparities between the previously published studies elsewhere in Europe and studies from the US. For example, adjuvant treatments for breast cancer may differ as a result of differences in healthcare systems, access to healthcare, and treatment guidelines. Furthermore, the Swedish national breast cancer screening program is well established and available for women between the ages 40 and 75 years. A recognized and highly assessed screening program is likely to be associated with early detection of breast cancer, resulting in lower stage disease at diagnosis, which will impact survival [28]. Importantly, when interpreting the results of this study, a healthy user bias might have affected the results in cases in which statin use is associated with increased health awareness. This agrees with the likelihood of attending the screening program.

A couple of strengths of this study are noteworthy. Firstly, the considerable amount of breast cancer patients included in this study represents one of the largest studies of statins and breast cancer prognosis. Secondly, Swedish infrastructure allows for access to a high quality, national pharmaceutical database, cancer registry, and cause-of-death from which data were easily merged through the civil person number while prioritizing patients’ confidentiality as the procedures were handled by the Social Health and Welfare administration. Statins are prescribed drugs as described by the Swedish Prescribed Drug Register, and there should be a minimal risk of misclassification regarding statin use as over-the-counter administration is not possible. A possible exception would be if a patient had terminated statin use prior to July 1st, 2005 when the national register was initiated, however, we consider this unlikely since statins are usually prescribed as a lifelong treatment. Thirdly, all analyses were adjusted for a diagnosis of diabetes, which limits the possible survival bias that the diabetes might imply. Diabetic patients are commonly prescribed a statin in addition to their anti-diabetic medication, since diabetics have an elevated risk for cardiovascular diseases, which further highlights the need to consider diabetes in the analyses of breast cancer mortality. Moreover, when the analysis of post-diagnostic statin use was restricted to only include new users of statin after breast cancer diagnosis, the sample size was reduced since many post-diagnostic users were also pre-diagnostic users. Despite some loss of power, statin use following breast cancer diagnosis was similarly associated with lower risk of breast cancer related deaths compared to non-users in this subset (HR = 0.94; 95% CI 0.91–0.97). Finally, sensitivity analyses considering potential lag time bias used one-year lag-time based on calculations showing that after one year of statin use the effect reaches a plateau and is thus a reasonable lag-time estimate. Reassuringly, the effect estimates of statin use regarding both breast cancer specific survival and overall survival were strengthened in the sensitivity analyses taking lag time into consideration.

In conclusion, statin users, particularly simvastatin users, had a lower risk of breast cancer related deaths compared to non-users in this nationwide cohort of Swedish women with breast cancer diagnosed after the age of 40. Considering previous evidence from functional-, clinical- and epidemiological studies, this study adds evidence to the notion that statins seem to possess beneficial effects against breast cancer progression along with its cardiovascular benefits. At this point, there is a need for confirmative results based on randomized clinical trials to test whether statin treatment can be recommended for patients with breast cancer.