https://orcid.org/0000-0002-2021-1660
Abstract
Statins are widely acknowledged for their application in patients with hypercholesterolemia to reduce cardiovascular morbidity and mortality. More recently, their potential to exert pleiotropic effects, particularly in impeding the proliferation of neoplastic cells, has attracted considerable attention. Prior studies have demonstrated that statins may mitigate cancer progression and micrometastasis. However, the benefits of statins in breast cancer have been inconclusive.
The aim of this meta-analysis was to evaluate the impact of statin use following a breast cancer diagnosis on breast cancer recurrence and mortality.
We performed a systematic literature search using PubMed, Embase, and Scopus for relevant articles from inception until 30th May 2023. Hazard ratios (HR) were pooled using a random-effect model. The primary outcome of interest was the risk of breast cancer recurrence. The secondary outcomes included breast cancer-specific mortality and all-cause mortality.
A total of 15 studies with 156 448 patients were included in the final analysis. The mean age of patients between statin users and non-users was 64.59 and 59.15 years, respectively. Statin use was associated with a reduction in the recurrence of breast cancer [HR 0.76, 95% confidence interval (CI): 0.67–0.87] compared with non-statin users. This trend was similar among lipophilic statin users (HR 0.73, 95% CI: 0.63–0.85) but not for hydrophilic statin users (HR 1.17, 95% CI: 0.82–1.68). Furthermore, statin users exhibited a lower risk of breast cancer mortality (HR 0.80, 95% CI: 0.66–0.96) but all-cause mortality (HR 0.82, 95% CI: 0.66–1.02) was comparable among both groups of patients. Conversely, lipophilic statins demonstrated a reduction in both all-cause mortality (HR 0.84, 95% CI: 0.75–0.93) and breast cancer mortality (HR 0.85, 95% CI: 0.74–0.99) compared to non-statin users.
Among patients with breast cancer, statin use post-diagnosis decreases the risk of breast cancer recurrence and breast cancer mortality. Furthermore, lipophilic statins exhibit an additional advantage of reduction in all-cause mortality.
PROSPERO registration: CRD42022362011.
Central illustration showing the association of statin use post-diagnosis among breast cancer patients on clinical outcomes. (Original figure created by VA & VJ using BioRender.com)
Introduction
Breast cancer stands as the most prevalent malignancy diagnosed among women on a global scale while ranking as the second leading cause of cancer-related mortality in the female population.1,2 Projections indicate a distressing surge in breast cancer incidence by the year 2040, with more than 3 million new cases annually and more than 1 million deaths per year.1 Apart from the substantial incidence rates, breast cancer recurrence poses an additional challenge. The annual hazards of recurrence have been reported to reach as high as 10%, with the pinnacle hazard occurring within the 1- to 2-year interval (13–15%).3,4
Cardiovascular diseases (CVD) represent another prominent contributor to global mortality, responsible for a significant proportion of deaths in females (47%).5 Considering the substantial prevalence of breast cancer and CVD, it is not uncommon for these two conditions to coexist among individuals.6 Moreover, the treatment modalities employed in breast cancer management have been associated with the potential to induce cardiovascular dysfunction, further exacerbating the intricate relationship between these diseases.6
Statins are an essential component of CVD treatment. Current European Society of Cardiology (ESC) guidelines recommend high-intensity statin as a class 1A pharmacological drug to achieve lowered low-density lipoprotein cholesterol for patients under the age of 70 and thereby reduce cardiovascular disease morbidity and mortality.7 The efficacy of statins in improving cardiovascular outcomes in cancer patients experiencing chemotherapy-induced cardiotoxicity is well established.8,9 However, there exists a notable scarcity of evidence consolidating the role of statins in enhancing cancer-related outcomes, such as recurrence and overall survival.
Multiple studies have yielded conflicting results regarding the association between statin use and breast cancer outcomes. While some studies suggest a significant reduction in mortality among statin users,10–13 others have found no such association compared to non-users.14–16 To address these inconsistencies, we sought to perform a meta-analysis to assess the association between post-diagnostic statin use and breast cancer recurrence and mortality among women.
Methods
Materials and methods
This meta-analysis was conducted and reported following the PRISMA (Preferred reporting items for systematic review and Meta-analysis) and MOOSE guidelines and performed according to established methods, as described previously.17–19 The pre-specified study protocol has been registered in the PROSPERO (CRD42022362011).
Search strategy
We conducted a systematic literature search in PubMed, Embase, and Scopus using predefined MESH terms by using “AND” and “OR.” The following search terms were used: (breast cancer[MeSH Terms]) AND (Statin[MeSH Terms]) OR (Statin[Other Term]) OR (CS-514[Other Term]) OR (simvastatin[Other Term]) OR (atorvastatin[Other Term]) OR (fluvastatin[Other Term]) OR (lovastatin[Other Term]) OR (rosuvastatin[Other Term]) AND (mortality[Other Term]) AND (recurrence[Other Term]) AND (outcomes[Other Term]). The search was performed from inception up until 30th May 2023 without any restrictions on the language of the studies. Search strategies are available in Supplementary Table S1.
All the studies were carefully screened and exported to the Mendeley reference manager used to handle searched citations. A manual check was carried through to crosscheck for any remaining duplicates. Two reviewers (V.J. and V.A.) reviewed the papers based on the title and abstract. Discrepancies regarding the inclusion of studies were arbitrated by another author (A.J.).
Eligibility criteria
We included studies with adult patients ≥18 years of age diagnosed with breast cancer. All prospective and retrospective cohort studies were sought to be eligible for inclusion in the study. It was decided to include studies with two arms to compare patients of breast cancer with statin users and non-statin users post-diagnostic. Only studies that exclusively enrolled patients who used statins after their primary breast cancer diagnosis were considered. Studies that did not differentiate pre-diagnostic statin use from post-diagnostic use or studies that included patients who were already using statins at the time of primary breast cancer diagnosis were excluded. Studies performed on animals, or reviews, case reports, case series, patients <18 years, studies with a single arm, and studies without outcomes of interest were also excluded from the review.
Clinical outcomes
The primary outcome of this meta-analysis was the risk of recurrence of breast cancer. Outcomes reported as second primary event or contralateral breast cancer event were not considered as recurrence. The secondary outcomes of interest were breast cancer-specific mortality and all-cause mortality.
Data extraction and quality assessment
Two authors (V.A. and V.J.) extracted the following data: study type, author, study location, study follow-up duration, patient characteristics (number, age, gender, menopausal status, and comorbidities), and primary and secondary outcomes. We used the reported estimates when reported in the form of hazard ratios (HRs). If different estimates were available, we opted for HR with the most adjusted effect measure or propensity score-matched data where available. Two investigators (A.I. and V.J.) independently appraised the potential risk of bias using the Newcastle-Ottawa (NOS) scale for observational studies.20 We then classified studies as low, moderate, or high quality based on the scores after evaluation.
Statistical analysis
Statistical analysis was performed by pooling the HR using the random effect model, with a test for overall effect reported as Z-value and 95% CIs. Statistical significance was met if 95% CI does not cross numeric “1” and P < 0.05. The heterogeneity among studies was assessed by Higgins's I-squared (I2) statistical model with I2 values. As a guide, I2 < 25% indicated low, 25–50% moderate, and >50% high heterogeneity.21 Further sensitivity analyses were performed using a leave-one-out method to check the robustness of the results. To explore the causes of heterogeneity of primary and secondary outcomes, subgroup analyses were implemented based on study-level data, including follow-up period, study design, sample size, and statin types. Assessment of publication bias was via visualization of the funnel plot.22 The study of Brewer et al., 2013, Sim et al., 2022, and Insau et al., 2022 reported the hazards for all-cause survival and breast cancer survival. These HR were extracted from these above-mentioned studies and were converted to HR of all-cause mortality and HR of breast cancer mortality, respectively, since the final outcome of interest was in terms of mortality rather than survival. This conversion was done following prespecified methods.23 The HR of survival was first converted to the probability of survival by dividing the HR by 1 + HR (P = HR/1 + HR). After that, the probability of survival is subtracted from 1 to calculate the probability of death (since both probabilities are complementary to each other). Finally, the probability of death is reconverted to HR of death by dividing the probability by 1-probability (HR = P/1−P). All statistical analyses were performed using Review Manager version 5.4 and R version 4.1.1.
Results
Study selection
The preliminary database search using the pre-specified keywords yielded 936 articles, of which 158 studies were excluded after the removal of duplicates. A total of 725 studies were further excluded from the initial post-title and abstract screening based on the inclusion and exclusion criteria and comparison arm (statin users vs. non-statin users). The full-text review was conducted for the remaining 53 articles identified during the search period, of which 38 studies were excluded: pre-diagnostic statin use, no desired outcomes, single arm, or abstract without any desired outcome. Hence, a total of 15 studies10–16,24–31 met the eligibility criteria and were included in the meta-analysis. The PRISMA flow diagram is depicted in Supplementary Figure S1.
Baseline patient demographics
A total of 15 studies with a total of 156 448 patients were included in our analysis. The mean age among the available data between statin users and non-users was 64.59 vs. 59.15 years. The most common comorbidity among patients with statin and non-statin was diabetes mellitus (DM) (12% vs. 2.5%). Among the studies that reported the menopausal status, 84.8% of women in the statin group and 74.2% of women in the non-statin group were post-menopausal. The median follow-up duration was 4.86 years. The study characteristics, demographics, and comorbidities are presented in Table 1. Most included studies showed a low risk of bias on NOS (score was > six-ninth) apart from one study, which had a score of three-ninth as it was a conference paper and there was insufficient information about its methodology31 (Supplementary Table S2).
Baseline demographics, comorbidities, and study characteristics of studies included in the meta-analysis
| Author | Country | Study design | Cohort | Sample size | Follow-up, years Median | Age, years Mean ± SD | Post-menopausal | Diabetes mellitus | Myocardial infarction |
|---|---|---|---|---|---|---|---|---|---|
| Kwan et al. 2008 | US | Prospective | Non-statin | 1408 | 2 (Range 0.92–3.24) | 57.2 ± 11.1 | 69.3% | - | - |
| Statin | 367 | 62.2 ± 9.4 | 87.1% | - | - | ||||
| Ahern et al. 2011 | Denmark | Prospective | Non-statin | 15 487 | 6.8 | 56.95* | 67.2% | 1.78% | 0.7% |
| Statin | 3282 | 60.99* | 84.3% | 7.43% | 3.5% | ||||
| Ceacareanu et al. 2011 | US | Retrospective | Total cohort | 294 | 2.5 | - | - | 33.3% | - |
| Chae et al. 2011 | US | Retrospective | Non-statin | 547 | 4.6 | 57.68 ± 14.6 | 64.0% | 12.0% | - |
| Statin | 156 | 64.12 ± 11.5 | 87.8% | 29.5% | - | ||||
| Brewer et al. 2013 | US | Retrospective | Non-statin | 650 | 2.9 | 48.9* | 48.2% | 7.5% | - |
| Statin | 73 | 58.85* | 84.9% | 32.9% | - | ||||
| Nickels et al. 2013 | Germany | Prospective | Non-statin | 2737 | 5.4 | 62.26* | 90.5% | 7.3% | - |
| Statin | 287 | 65.59* | 96.5% | 21.3% | - | ||||
| Boudreau et al. 2014 | US | Retrospective | Non-statin | 3006 | 6.3 (IQR 3.7–9.7) | - | 72.8% | - | - |
| Statin | 1210 | - | - | - | |||||
| Murtola et al. 2014 | Finland | Prospective | Non-statin | 26 941 | 3.17 (IQR 0.08–8.50) | 58** | 59.3% | - | - |
| Statin | 4151 | 3.83 (IQR 0.08–8.67) | 64** | 81.5% | - | - | |||
| Cardwell et al. 2015 | UK | Retrospective | Non-statin | 13 598 | 7.5 (IQR 4–11) | 60.9* | - | 2.0% | 1.0% |
| Statin | 4286 | 67.3* | - | 18.0% | 5.0% | ||||
| McMenamin et al. 2016 | Scotland | Retrospective | Non-statin | 10 907 | 3.5 (IQR 2–5) | 60.64* | - | 1.1% | 0.5% |
| Statin | 4233 | 70.63* | - | 11.6% | 6.8% | ||||
| Borgquist et al. 2019 | Sweden | Retrospective | Non-statin | 20 503 | 5.13 | - | - | - | - |
| Statin | 4358 | - | - | - | - | ||||
| Harborg et al. 2020 | Denmark | Prospective | Non-statin | 13 046 | 4.5 | 65.34* | 100% | - | - |
| Statin | 1727 | 65.33* | 100% | - | - | ||||
| Inasu et al. 2022 | Sweden | Prospective | Non-statin | 91 | 8.6 | 71.8* | - | - | - |
| Statin | 269 | 70.5* | - | - | - | ||||
| Sim et al. 2022 | Singapore | Retrospective | Non-statin | 6505 | 8.74 (IQR 8.59–8.88) | 51.9* | 38.4% | 6.0% | - |
| Statin | 1353 | 59.7* | 66.4% | 13.4% | - | ||||
| Scott et al. 2023 | New Zealand | Retrospective | Non-statin | 10 916 | 4.51 (Range 0.01–10.99) | 57.1* | - | 1.0% | - |
| Statin | 4060 | 65.3* | - | 10.0% | - |
| Author | Country | Study design | Cohort | Sample size | Follow-up, years Median | Age, years Mean ± SD | Post-menopausal | Diabetes mellitus | Myocardial infarction |
|---|---|---|---|---|---|---|---|---|---|
| Kwan et al. 2008 | US | Prospective | Non-statin | 1408 | 2 (Range 0.92–3.24) | 57.2 ± 11.1 | 69.3% | - | - |
| Statin | 367 | 62.2 ± 9.4 | 87.1% | - | - | ||||
| Ahern et al. 2011 | Denmark | Prospective | Non-statin | 15 487 | 6.8 | 56.95* | 67.2% | 1.78% | 0.7% |
| Statin | 3282 | 60.99* | 84.3% | 7.43% | 3.5% | ||||
| Ceacareanu et al. 2011 | US | Retrospective | Total cohort | 294 | 2.5 | - | - | 33.3% | - |
| Chae et al. 2011 | US | Retrospective | Non-statin | 547 | 4.6 | 57.68 ± 14.6 | 64.0% | 12.0% | - |
| Statin | 156 | 64.12 ± 11.5 | 87.8% | 29.5% | - | ||||
| Brewer et al. 2013 | US | Retrospective | Non-statin | 650 | 2.9 | 48.9* | 48.2% | 7.5% | - |
| Statin | 73 | 58.85* | 84.9% | 32.9% | - | ||||
| Nickels et al. 2013 | Germany | Prospective | Non-statin | 2737 | 5.4 | 62.26* | 90.5% | 7.3% | - |
| Statin | 287 | 65.59* | 96.5% | 21.3% | - | ||||
| Boudreau et al. 2014 | US | Retrospective | Non-statin | 3006 | 6.3 (IQR 3.7–9.7) | - | 72.8% | - | - |
| Statin | 1210 | - | - | - | |||||
| Murtola et al. 2014 | Finland | Prospective | Non-statin | 26 941 | 3.17 (IQR 0.08–8.50) | 58** | 59.3% | - | - |
| Statin | 4151 | 3.83 (IQR 0.08–8.67) | 64** | 81.5% | - | - | |||
| Cardwell et al. 2015 | UK | Retrospective | Non-statin | 13 598 | 7.5 (IQR 4–11) | 60.9* | - | 2.0% | 1.0% |
| Statin | 4286 | 67.3* | - | 18.0% | 5.0% | ||||
| McMenamin et al. 2016 | Scotland | Retrospective | Non-statin | 10 907 | 3.5 (IQR 2–5) | 60.64* | - | 1.1% | 0.5% |
| Statin | 4233 | 70.63* | - | 11.6% | 6.8% | ||||
| Borgquist et al. 2019 | Sweden | Retrospective | Non-statin | 20 503 | 5.13 | - | - | - | - |
| Statin | 4358 | - | - | - | - | ||||
| Harborg et al. 2020 | Denmark | Prospective | Non-statin | 13 046 | 4.5 | 65.34* | 100% | - | - |
| Statin | 1727 | 65.33* | 100% | - | - | ||||
| Inasu et al. 2022 | Sweden | Prospective | Non-statin | 91 | 8.6 | 71.8* | - | - | - |
| Statin | 269 | 70.5* | - | - | - | ||||
| Sim et al. 2022 | Singapore | Retrospective | Non-statin | 6505 | 8.74 (IQR 8.59–8.88) | 51.9* | 38.4% | 6.0% | - |
| Statin | 1353 | 59.7* | 66.4% | 13.4% | - | ||||
| Scott et al. 2023 | New Zealand | Retrospective | Non-statin | 10 916 | 4.51 (Range 0.01–10.99) | 57.1* | - | 1.0% | - |
| Statin | 4060 | 65.3* | - | 10.0% | - |
*Data was stratified into age groups. A weighted average was taken.
**Median
Baseline demographics, comorbidities, and study characteristics of studies included in the meta-analysis
| Author | Country | Study design | Cohort | Sample size | Follow-up, years Median | Age, years Mean ± SD | Post-menopausal | Diabetes mellitus | Myocardial infarction |
|---|---|---|---|---|---|---|---|---|---|
| Kwan et al. 2008 | US | Prospective | Non-statin | 1408 | 2 (Range 0.92–3.24) | 57.2 ± 11.1 | 69.3% | - | - |
| Statin | 367 | 62.2 ± 9.4 | 87.1% | - | - | ||||
| Ahern et al. 2011 | Denmark | Prospective | Non-statin | 15 487 | 6.8 | 56.95* | 67.2% | 1.78% | 0.7% |
| Statin | 3282 | 60.99* | 84.3% | 7.43% | 3.5% | ||||
| Ceacareanu et al. 2011 | US | Retrospective | Total cohort | 294 | 2.5 | - | - | 33.3% | - |
| Chae et al. 2011 | US | Retrospective | Non-statin | 547 | 4.6 | 57.68 ± 14.6 | 64.0% | 12.0% | - |
| Statin | 156 | 64.12 ± 11.5 | 87.8% | 29.5% | - | ||||
| Brewer et al. 2013 | US | Retrospective | Non-statin | 650 | 2.9 | 48.9* | 48.2% | 7.5% | - |
| Statin | 73 | 58.85* | 84.9% | 32.9% | - | ||||
| Nickels et al. 2013 | Germany | Prospective | Non-statin | 2737 | 5.4 | 62.26* | 90.5% | 7.3% | - |
| Statin | 287 | 65.59* | 96.5% | 21.3% | - | ||||
| Boudreau et al. 2014 | US | Retrospective | Non-statin | 3006 | 6.3 (IQR 3.7–9.7) | - | 72.8% | - | - |
| Statin | 1210 | - | - | - | |||||
| Murtola et al. 2014 | Finland | Prospective | Non-statin | 26 941 | 3.17 (IQR 0.08–8.50) | 58** | 59.3% | - | - |
| Statin | 4151 | 3.83 (IQR 0.08–8.67) | 64** | 81.5% | - | - | |||
| Cardwell et al. 2015 | UK | Retrospective | Non-statin | 13 598 | 7.5 (IQR 4–11) | 60.9* | - | 2.0% | 1.0% |
| Statin | 4286 | 67.3* | - | 18.0% | 5.0% | ||||
| McMenamin et al. 2016 | Scotland | Retrospective | Non-statin | 10 907 | 3.5 (IQR 2–5) | 60.64* | - | 1.1% | 0.5% |
| Statin | 4233 | 70.63* | - | 11.6% | 6.8% | ||||
| Borgquist et al. 2019 | Sweden | Retrospective | Non-statin | 20 503 | 5.13 | - | - | - | - |
| Statin | 4358 | - | - | - | - | ||||
| Harborg et al. 2020 | Denmark | Prospective | Non-statin | 13 046 | 4.5 | 65.34* | 100% | - | - |
| Statin | 1727 | 65.33* | 100% | - | - | ||||
| Inasu et al. 2022 | Sweden | Prospective | Non-statin | 91 | 8.6 | 71.8* | - | - | - |
| Statin | 269 | 70.5* | - | - | - | ||||
| Sim et al. 2022 | Singapore | Retrospective | Non-statin | 6505 | 8.74 (IQR 8.59–8.88) | 51.9* | 38.4% | 6.0% | - |
| Statin | 1353 | 59.7* | 66.4% | 13.4% | - | ||||
| Scott et al. 2023 | New Zealand | Retrospective | Non-statin | 10 916 | 4.51 (Range 0.01–10.99) | 57.1* | - | 1.0% | - |
| Statin | 4060 | 65.3* | - | 10.0% | - |
| Author | Country | Study design | Cohort | Sample size | Follow-up, years Median | Age, years Mean ± SD | Post-menopausal | Diabetes mellitus | Myocardial infarction |
|---|---|---|---|---|---|---|---|---|---|
| Kwan et al. 2008 | US | Prospective | Non-statin | 1408 | 2 (Range 0.92–3.24) | 57.2 ± 11.1 | 69.3% | - | - |
| Statin | 367 | 62.2 ± 9.4 | 87.1% | - | - | ||||
| Ahern et al. 2011 | Denmark | Prospective | Non-statin | 15 487 | 6.8 | 56.95* | 67.2% | 1.78% | 0.7% |
| Statin | 3282 | 60.99* | 84.3% | 7.43% | 3.5% | ||||
| Ceacareanu et al. 2011 | US | Retrospective | Total cohort | 294 | 2.5 | - | - | 33.3% | - |
| Chae et al. 2011 | US | Retrospective | Non-statin | 547 | 4.6 | 57.68 ± 14.6 | 64.0% | 12.0% | - |
| Statin | 156 | 64.12 ± 11.5 | 87.8% | 29.5% | - | ||||
| Brewer et al. 2013 | US | Retrospective | Non-statin | 650 | 2.9 | 48.9* | 48.2% | 7.5% | - |
| Statin | 73 | 58.85* | 84.9% | 32.9% | - | ||||
| Nickels et al. 2013 | Germany | Prospective | Non-statin | 2737 | 5.4 | 62.26* | 90.5% | 7.3% | - |
| Statin | 287 | 65.59* | 96.5% | 21.3% | - | ||||
| Boudreau et al. 2014 | US | Retrospective | Non-statin | 3006 | 6.3 (IQR 3.7–9.7) | - | 72.8% | - | - |
| Statin | 1210 | - | - | - | |||||
| Murtola et al. 2014 | Finland | Prospective | Non-statin | 26 941 | 3.17 (IQR 0.08–8.50) | 58** | 59.3% | - | - |
| Statin | 4151 | 3.83 (IQR 0.08–8.67) | 64** | 81.5% | - | - | |||
| Cardwell et al. 2015 | UK | Retrospective | Non-statin | 13 598 | 7.5 (IQR 4–11) | 60.9* | - | 2.0% | 1.0% |
| Statin | 4286 | 67.3* | - | 18.0% | 5.0% | ||||
| McMenamin et al. 2016 | Scotland | Retrospective | Non-statin | 10 907 | 3.5 (IQR 2–5) | 60.64* | - | 1.1% | 0.5% |
| Statin | 4233 | 70.63* | - | 11.6% | 6.8% | ||||
| Borgquist et al. 2019 | Sweden | Retrospective | Non-statin | 20 503 | 5.13 | - | - | - | - |
| Statin | 4358 | - | - | - | - | ||||
| Harborg et al. 2020 | Denmark | Prospective | Non-statin | 13 046 | 4.5 | 65.34* | 100% | - | - |
| Statin | 1727 | 65.33* | 100% | - | - | ||||
| Inasu et al. 2022 | Sweden | Prospective | Non-statin | 91 | 8.6 | 71.8* | - | - | - |
| Statin | 269 | 70.5* | - | - | - | ||||
| Sim et al. 2022 | Singapore | Retrospective | Non-statin | 6505 | 8.74 (IQR 8.59–8.88) | 51.9* | 38.4% | 6.0% | - |
| Statin | 1353 | 59.7* | 66.4% | 13.4% | - | ||||
| Scott et al. 2023 | New Zealand | Retrospective | Non-statin | 10 916 | 4.51 (Range 0.01–10.99) | 57.1* | - | 1.0% | - |
| Statin | 4060 | 65.3* | - | 10.0% | - |
*Data was stratified into age groups. A weighted average was taken.
**Median
Meta-analysis of outcomes
The pooled analysis of primary outcomes showed that the risk of recurrence of breast cancer was reduced among patients with statin users compared to non-statin users (HR 0.76, 95% CI: 0.67–0.87, I2 = 53%) (Figure 1A). All-cause mortality (HR 0.82, 95% CI: 0.66–1.02, I2 = 93%) was comparable among both groups of patients; however, statin users exhibited a lower risk of breast cancer mortality (HR 0.80, 95% CI: 0.66–0.96, I2 = 83%) (Figure 1B and C).
Forest plots of outcomes including (A) recurrence of breast cancer, (B) all-cause mortality, and (C) breast cancer mortality.
Sensitivity analysis
Results of the risk of recurrence of breast cancer and all-cause mortality remained largely unaltered in terms of magnitude and direction after leave-one-out analysis (Supplementary Figure S2–4). For all-cause mortality, we further analysed the results after excluding outliers. Sensitivity analysis following the removal of the study by Insau et al. and Murtola et al. showed that there was a reduction in all-cause mortality among statin users (HR 0.88, 95% CI: 0.84–0.93) with mild heterogeneity across studies (I2 = 26%) (Supplementary Figure S5). For breast cancer mortality, upon removal of the study by Murtola et al., the heterogeneity dropped substantially from 83% to 3%; however, the hazards increased by 5% (HR 0.85, 95% CI: 0.78–0.92) (Supplementary Figure S6).
Subgroup analyses
Subgroup analysis was conducted on primary and secondary outcomes based on study design, sample size, and follow-up period. The risk of recurrence of breast cancer remained lower among patients with statin users compared to those with non-statin users without subgroup differences (P ≥ 0.05) after stratifying by study design, sample size, and follow-up period (Supplementary Figure S7A–C). However, it was important to note that subgroup analysis based on study design showed that the heterogeneity was minimal (I2 = 0%) among prospective studies compared to higher heterogeneity (I2 = 75%) among retrospective studies, suggesting that the heterogeneity in primary analysis likely derived from these retrospective studies (Supplementary Figure S7A). Similarly, heterogeneity among subgroups with larger sample sizes (I2 = 8%) was substantially smaller compared to that of subgroups with smaller sample sizes (I2 = 64%), explaining that the sample size of studies also played a role in high heterogeneity observed in the primary analysis (Supplementary Figure S7B).
Likewise, there was no subgroup difference detected on all-cause mortality despite stratification based on study design, sample size, and follow-up period (P > 0.05) (Supplementary Figure S8A–C). Segregating studies by sample size showed that the heterogeneity was minimal among subgroups with higher sample size (I2 = 0%) compared to that with lower sample size (I2 = 61%), redemonstrating the effect of sample size on the overall heterogeneity observed on all-cause mortality (Supplementary Figure S8B). It was interesting to observe that among statin users, studies with larger sample sizes (>10 000 patients) showed a reduction of 13% in all-cause mortality (HR, 0.87, 95% CI: 0.83–0.92) compared with the non-statin group of patients. However, no difference was found in studies with smaller sample sizes (HR, 0.96, 95% CI 0.82–1.13) (Supplementary Figure S8B). The follow-up period also appeared to contribute to the overall heterogeneity, given that the heterogeneity among the studies with a longer follow-up period (I2 = 0%) was substantially lower than that of studies with a shorter follow-up period (I2 = 54%) (Supplementary Figure S8C).
Lastly, because there were only a limited number of studies available for each treatment group, we conducted a subgroup analysis solely based on the sample size and follow-up for breast cancer mortality (Supplementary Figure S9A-B). Overall, there was no subgroup difference detected (p>0.05) for both sample size and follow-up period.
Meta-analysis based on the type of statins used.
We additionally conducted an analysis based on the type of statins used (hydrophilic vs. lipophilic). The use of lipophilic statin (including 3 studies) was associated with a lower risk of breast cancer recurrence (HR 0.73, 95% CI: 0.63–0.85), while there was no difference among the use of hydrophilic statin and non-statin users (HR 1.17, 95%CI: 0.82–1.68) (including 2 studies) (Figure 2). Similarly, on all-cause mortality, the use of lipophilic statin (including 3 studies) was associated with a lower risk of all-cause mortality (HR 0.84, 95% CI: 0.75–0.93), while there was no difference between patients with hydrophilic statin users and non-statin users (HR 0.95, 95% CI: 0.84–1.08) (including 3 studies) (Figure 3). Lastly, on breast cancer mortality, there was a trend of reduction in breast cancer mortality among lipophilic statin users compared to non-statin users (HR 0.85, 95% CI: 0.74–0.99) (including 3 studies) (Figure 4). Similarly, there was no difference between hydrophilic statin users and non-statin users (HR 0.99, 95% CI: 0.84–1.18) (including 3 studies) (Figure 4).
Forest plots of recurrence of breast cancer based on the type of statins used.
Forest plots of all-cause mortality based on the type of statins used.
Forest plots of breast cancer mortality based on the type of statins used.
Publication bias
Visualization of funnel plots was implemented to assess the publication bias of included studies. Overall, there appeared to be minimal or no funnel plot asymmetry that could be suggestive of publication bias for all the outcomes, including recurrence of breast cancer, all-cause mortality, and breast cancer mortality (Supplementary Figure S10–12).
Discussion
This meta-analysis assessed the effects of using statins (following diagnosis) on the risk of breast cancer recurrence, breast cancer-specific mortality, and all-cause mortality. The findings of the meta-analysis indicated that with statin use, there was a 24% reduction in breast cancer recurrence and a 20% reduction in mortality due to breast cancer. Additionally, the analysis showed that lipophilic statin use was associated with a 27% reduced risk of breast cancer recurrence, a 16% reduction in all-cause mortality, and a 15% reduction in breast cancer-specific mortality; however, no differences in these outcomes were found with the use of hydrophilic statins.
Several studies have implicated the role of statins in patients with breast cancer, which could explain the strong association observed between statins and recurrence-free survival for breast cancer patients. A study by Wang et al. hypothesized that statins’ therapeutic role in breast cancer patients is due to their anti-proliferative effect by inducing G0/G1 or G2/M arrest through a reduction of CDK4/6 and Cyclin D1.32 Additionally, statins have shown the ability to suppress modification of farnesyl pyrophosphate (FFP) and geranylgeranyl pyrophosphate (GGPP) and activation of Ras and Rho GTPases, owing to another anti-proliferative effect through cell-cycle arrest in breast cancer cells.33 Statins have been shown to alter gene transcription through the inhibition of DNA methyltransferases, thereby reducing breast cancer proliferation.34 Levels of matrix metalloproteinase (MMP)-9, usually expressed in endothelial cells, have been reduced following lipophilic statin therapy in breast cancer patients, impacting the ability of endothelial cells to initiate angiogenesis and therefore reducing their invasive capability.35 Furthermore, another study showed that treatment with a lipophilic statin in patients with breast cancer has the potential to downregulate CD44, a cancer stem cell marker, therefore reducing invasion and metastasis.36 Another mechanism where statins can reduce invasion is through decreasing transferrin receptor expression, leading to a concomitant depletion of iron in breast cancer cells.37 Combining the findings of our meta-analysis with the biological effects of statins, it appears that statins, particularly lipophilic statins, may have a crucial role to play in managing breast cancer patients.
A thorough understanding of the included studies is crucial in interpreting the findings of this meta-analysis. It is important to note that the studies included in this analysis specifically focused on breast cancer patients who started using statins after their cancer diagnosis. Consequently, the allocation of patients to the exposed and unexposed groups did not occur simultaneously. This sequential allocation approach introduces the possibility of immortal time bias. The consequence of this is the artificial inflation of observed survival or event-free time among the statin users relative to the non-statin users. Non-compliance to statin therapy represents another significant factor that warrants careful consideration. The majority of the studies included in this analysis relied upon data obtained from national registries or databases.10,11,13,16,24–28 In these studies, the ascertainment of statin use relied primarily on the prescription records provided by physicians, which, although employed as a marker of statin use, may not consistently capture the true adherence to the prescribed statin regimen. Although these studies made efforts to include only patients whose statin prescriptions were dispensed by pharmacists, acknowledging the restricted availability of statins as over-the-counter drugs in most countries, it is crucial to emphasize that the mere act of dispensing medication does not automatically ensure its subsequent consumption. Non-compliance, encompassing both the failure to initiate treatment and premature discontinuation, can potentially introduce an element of bias leading to an underestimation of the actual impact of statin therapy on breast cancer outcomes.
Another crucial aspect that differentiated the studies analyzed pertains to their approach toward covariate adjustment. While all the studies presented adjusted estimates for the hazard ratio, only some of them did not adjust for cancer stage, radiotherapy, chemotherapy, hormone therapy, concomitant medications, etc.11,16,26,27,31 Failing to control confounding factors can result in an over- or underestimation of the true relationship between statin use and cancer prognosis. Furthermore, the included studies were unable to adequately examine confounding by indication with comparison to non-statin antilipemic. Consequently, it remains possible that the observed association between statin use and breast cancer outcomes could be a surrogate for the effects of an associated condition, such as hyperlipidemia. Besides, there were differences in the characteristics of the sample population as well. The countries and, thus, the races in which the included studies were conducted differed (Table 1), which may have important implications for the generalizability and interpretation of our results. The incidence, prevalence, and distribution of breast cancer vary widely across different geographic regions and ethnic groups.2,6 Moreover, differences in lifestyle factors, environmental exposures, and healthcare access across different countries and races may also contribute to disparities in the observed cancer outcomes. Therefore, it is essential to interpret our results within the context of these variations and to consider the potential influence of these factors on the overall conclusions drawn from this meta-analysis.
Nevertheless, this meta-analysis provides compelling evidence regarding the association between post-diagnostic statin use and cancer recurrence and mortality in breast cancer patients. Our findings not only shed light on the reduced risk of breast cancer recurrence with statin use but also have significant implications for cardiovascular health. It has been well-established that cancer patients undergoing chemotherapy and radiation therapy are prone to cardiac dysfunction, leading to associated morbidity and mortality.38,39 Breast cancer treatments directly contribute to cardiovascular complications, including congestive heart failure and cardiomyopathies.6,40 Previous studies have demonstrated that statins when used as a prophylactic measure in patients receiving chemotherapy, can improve cardiovascular outcomes in breast cancer patients.8,41,42 By further demonstrating the potential benefits of statin use in reducing breast cancer recurrence, particularly among lipophilic statin users, our results suggest that statins may provide an additional advantage by improving cancer outcomes simultaneously.
Limitations
The major limitation of this study was that data was largely derived from observational studies, whereby confounding bias could not be reliably ruled out. Given that, adjusted effect sizes, where available, were selectively extracted and analyzed to ensure the robustness of results. Further sensitivity and subgroup analyses were conducted to explore the causes of high heterogeneity. Importantly, we were able to isolate studies that could be potential outliers and identify factors influencing the homogeneity of results, namely the study design, follow-up period, and/or the sample size of included studies. Unfortunately, due to limited data availability, specific subgroup analyses based on factors such as menopausal status and histopathological features of the cancers, including hormone receptors and HER2/neu status, could not be performed. The absence of these subgroup analyses limits our ability to make definitive conclusions about the impact of statins in these specific patient populations. Furthermore, meta-regression analysis could not be performed due to the inadequate availability of data on comorbidities and other baseline characteristics.
Conclusion
In conclusion, the outcomes of this meta-analysis indicate that the utilization of statins following a breast cancer diagnosis is linked to a diminished likelihood of breast cancer recurrence and breast cancer mortality. Furthermore, lipophilic statins exhibit an additional advantage of the reduction in all-cause mortality. These findings highlight the potential role of statins, especially lipophilic statins, in the management of breast cancer patients as an adjunct therapy. Future research, including prospective studies and randomized controlled trials, is warranted to further elucidate the underlying mechanisms and confirm the clinical implications of statin use in breast cancer management, especially in specific subgroups such as postmenopausal women and those with more severe breast cancer subtypes.
Acknowledgement
None
Source of funding
None
Data availability statement
The data underlying this article are available in the article and its online supplementary material.
Ethical approval
Since this is a review article of previously published studies, ethical approval is not required.
Conflict of interest: None declared by authors.
Author's Contribution
V.J. contributed to the conception or design of the work. S.A., VA, and M.S. contributed to the acquisition, analysis, or interpretation of data for the work. V.J., M.S., S.A., V.A., M.H., TK., A.J., and V.A., drafted the manuscript. V.J., V.A., and A.J. critically revised the manuscript. All gave final approval and agreed to be accountable for all aspects of work, ensuring integrity and accuracy.
Disclosure Statement
Abstract of this study was accepted for in-person presentation.
At the ‘AHA23’ conference.
Funding Statement
This research received no external funding.
