Association Between Antihypertensive Medication Use and Breast Cancer: A Systematic Review and Meta-Analysis

Background: The prevalence rate of hypertension and breast cancer increases with advancing age. Renin-angiotensin system inhibitors (RASIs), β-blockers (BBs), calcium channel blockers (CCBs), and diuretics are widely used to treat patients with hypertension. Although, the association between the use of antihypertensive medication and breast cancer has been highly debated, recent evidence supporting this association remains controversial. Objective: To evaluate the association between the use of antihypertensive medication and the risk of breast cancer and its prognosis. Methods: This study was conducted using data from the PubMed, Embase, and Cochrane Library databases retrieved for the period from January 2000 to April 2021. Articles and their references were checked and summary effects were calculated using random- and fixed-effects models. Heterogeneity test and sensitivity analysis were also performed. Results: This meta-analysis included 57 articles, which were all related to breast cancer risk or prognosis. Assessment of breast cancer risk using the pooled data showed that the use of BBs or CCBs or diuretics was associated with increased cancer risk [BB: relative risk (RR) = 1.20, 95% confidence interval (CI) = 1.09–1.32; CCBs: RR = 1.06, 95% CI 1.03–1.08; diuretics: RR = 1.06, 95% CI 1.01–1.11]. Long-term use of diuretic increased the risk of breast cancer (RR = 1.10, 95% CI 1.01–1.20), whereas long-term RASIs treatment reduced the risk (RR = 0.78, 95% CI 0.68–0.91). In addition, we found that diuretic users may be related to elevated breast cancer-specific mortality [hazard ratio (HR) = 1.18, 95% CI 1.04–1.33], whereas using other antihypertensive medications was not associated with this prognosis in patients with breast cancer. Conclusion: Using CCBs, BBs, and diuretics increased the risk of breast cancer. In addition, diuretics may elevate the risk of breast cancer-specific mortality. The long-term use of RASIs was associated with a significantly lower breast cancer risk, compared with non-users. Thus, this analysis provides evidence to support the benefits of the routine use of RASIs in patients with hypertension, which has important public health implications.


INTRODUCTION
In 2017, hypertension was identified as a major risk factor for cardiovascular disease globally, accounting for 10.4 (9.39-11.5) million deaths and 218  million disability-adjusted lifeyears (Stanaway et al., 2018). Fortunately, antihypertensive therapy significantly reduces the risk of cardiovascular disease and death in various populations. A meta-analysis showed that reducing the systolic blood pressure by 10 mm Hg would reduce the risk of major cardiovascular disease, coronary heart disease, stroke, and heart failure by 20, 17, 27, and 28%, respectively, whereas all-cause mortality was reduced by 13% (Ettehad et al., 2016). As an effective measure to control blood pressure, antihypertensive medications are commonly prescribed worldwide, and many patients take these drugs for a long period.
Several studies have found that breast cancer is associated with a variety of antihypertensive medications, including angiotensinconverting enzyme inhibitors/angiotensin receptor blockers (ACEIs/ARBs), β-blockers (BBs), calcium channel blockers (CCBs), and diuretics. However, these results remain controversial. Raebel et al. (2017) concluded that long-term use of CCBs was not related to breast cancer risk and long-term use of ACEIs might protect against breast cancer. The Chang et al. (2016) conducted a nested case-control study of 794,533 women and did not find any relationship between antihypertensive medication use and breast cancer risk. In contrast, other studies found that longterm treatment with CCBs was associated with a higher risk of breast cancer; however, no association was observed with the use of diuretics, BB, and renin-angiotensin system inhibitors (RASIs) (Li et al., 2013;Gomez-Acebo et al., 2016). A population-based casecontrol study demonstrated that treatment with ACEIs for more than 5 years increased breast cancer risk by 14% [odds ratio (OR) 1.14, 95% confidence interval (CI) 1.06-1.22] (Hallas et al., 2012).
A recent systematic review of observational studies concluded that BB reduced the risk of breast cancer recurrence (Caparica et al., 2021), whereas ACEIs and CCBs were not associated with breast cancer development (Zhao et al., 2017a;Naghibzadeh et al., 2020). In addition, the result of a meta-analysis suggested an association between CCBs use and breast cancer risk (Wright et al., 2017). However, uncertainty still exists regarding the effect of long-term use of antihypertensive medications and a potential association between different populations, breast cancer subtypes, or both.
This question remains controversial and, therefore, in this updated meta-analysis, we aimed to assess the association between the use of various classes of antihypertensive medication and breast cancer risk, prognosis [breast cancerspecific mortality, recurrence, overall survival (OS), and disease specific survival (DSS)].

Data Sources and Search Strategy
Supplementary Table S1 shows the strategy we used to conduct a comprehensive literature search of the PubMed, Embase, and Cochrane Library electronic databases without any restriction regarding geographical parameters and publication type or language. We reviewed research articles published over a span of nearly 20 years rather than those published in the 1990s, because we considered that the technique and treatment strategies that affect patient prognosis and survival have improved with time. The search was based on the framework of adult populations exposed to antihypertensive medication compared with non-users and the articles were evaluated for diagnosis or progress of breast cancer. Based on a combination of MeSH terms, keywords, and substance names, we conducted the following string search: "antihypertensive medication," "calcium channel blockers," "beta blockers," "angiotensin-converting enzyme inhibitors," "angiotensin receptor blockers," "renin-angiotensin system inhibitors," or "diuretics" combined with "breast cancer" or "breast carcinoma." In addition, the reference lists of other reviews or meta-analyses were manually searched to identify additional related articles. This meta-analysis is not registered in the PROSPERO database, but the process is based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Page et al., 2021). The PRISMA 2020 checklist was shown in Supplementary Table S2.

Selection Criteria
Studies fulfilling the following inclusion criteria were included: 1) the exposed group included individuals who had received antihypertensive medication; 2) the study had a comparative design: antihypertensive medication users vs. non-users; 3) the outcome was breast cancer risk, breast cancer-specific mortality, recurrence, or survival (DSS or OS); 4) the study reported the relative risk (RR), OR, or hazard ratio (HR) with corresponding 95% CIs or sufficient data to calculate these parameters. When the results of multiple studies were from the same database, the study including the largest number of participants was included. The following types of studies were excluded from this meta-analysis: meta-analyses, systematic or narrative reviews, case studies, experimental laboratory articles, conference abstracts, commentaries, randomized controlled trials, repeated publications, or if the reference group was administered another class of antihypertensive medication.

Data Extraction and Quality Assessment
Two of the authors (Yuxiu Xie and Men Wang) extracted data from the included studies and any disagreements between them were resolved by discussions with between the two other relevant co-authors. We extracted the following information from the selected articles: first author and year of publication, source of data, country, study design, duration of follow-up, study outcome, sample size of participants, and covariates used for the adjustment of confounders. The risk estimate adjusted for the largest number of confounding factors was extracted.
The nine-star Newcastle-Ottawa Scale (NOS) was used to assess the methodological quality of cohort studies and casecontrol studies (Stang, 2010;Ma et al., 2020). A score of ≥7 was considered to indicate high quality.

Statistical Analysis
We conducted this meta-analysis to assess the association between the use of antihypertensive medication and breast cancer risk and prognosis. The summary risk estimates are presented as forest plots. The I 2 test was used to assess potential heterogeneity between individual studies, and an I 2 > 50% indicated significant heterogeneity (Higgins et al., 2003). A fixed-effects model was used when there was no significant heterogeneity among the studies; whereas, the random-effects models was used when there was. To explore the source of heterogeneity and evaluate the potential impact of variables, we conducted subgroup analyses based on study design, country, and duration of antihypertensive medication use in years (e.g., <5, 5-10, ≥10 years). Articles with results that showed separate risk estimates based on different classes of CCBs or a different pathological pattern of breast cancer without a summary result were treated as different studies according to the result. Funnel plots were generated and examined visually, and gauged Egger's tests were performed to assess publication bias (Egger et al., 1997). Sensitivity analyses were conducted excluding one study at a time to test the robustness of this association.

Association of Antihypertensive Medication Use with Breast Cancer Risk
Pooled data from 18 studies related to breast cancer risk showed that the use of BBs was associated with increased cancer risk (RR 1.20, 95% CI 1.09-1.32; Figure 2A and Table 2). A subgroup analysis showed a significant association with case-control (RR 1.09, 95% CI 1.05-1.12), nested case-control (RR 1.11, 95% CI 1.03-1.20) studies, and cohort (RR 1.43, 95% CI 1.07-1.91) studies. In addition, a significant association was observed in Europe (RR 1.41, 95% CI 1.16-1.70) and Asia (RR 1.12, 95% CI 1.02-1.23), but not in North America (RR 1.02, 95% CI 0.94-1.11). However, subgroup analysis of the duration of administration in years showed that the use of BBs even for ≥10 years was not associated with breast cancer risk.
A total of 24 studies reported a connection between RASIs exposure of any duration and breast cancer risk. RASIs did not increase the risk of breast cancer with a pooled RR (0.98, 95% CI 0.93-1.04), as shown in Figure 2D. However, it is worth noting that RASIs played a protective role in breast cancer when studies were restricted to those where the duration of RASI use was <5 years (RR 0.94, 95% CI 0.90-0.99) and ≥10 years (RR 0.78, 95% CI 0.68-0.91). Fifteen and seven studies were related to ACEI and ARB use, respectively, and the results suggested that use of ACEI or ARB was not related to breast cancer risk (ACEI: RR 0.99, 95% CI 0.93-1.04; ARB: RR 0.98, 95% CI 0.81-1.19, Table 2).

Association of Antihypertensive Medication Use with Breast Cancer Prognosis Breast Cancer-Specific Mortality
Twenty-nine independent studies examined the association between antihypertensive medication use and breast cancerspecific mortality, with a pooled HR of 1.04 (95% CI 0.98-1.10). When HRs for the individual antihypertensive medication (BBs, CCBs, and RASIs) were calculated in subgroup analyses, no significant results were observed ( Figure 3A). However, a significant association was observed in diuretic users (HR 1.18, 95% CI 1.04-1.33).

OS
Twenty-one studies revealed no significant link between antihypertensive medication use and breast cancer OS (HR 1.05, 95% CI 0.96-1.14, Figure 3B). Furthermore, a subgroup analysis of the classes of antihypertensive medication showed no significant association was observed.

Recurrence
In total, 27 studies reported breast cancer recurrence with the use of antihypertensive medication. However, as shown in Figure 3C, no effect was observed for any type of antihypertensive medication (HR 1.00, 95% CI 0.92-1.09).

DSS
The results of the meta-analyses on DSS are summarized in forest plots for all 10 studies; no association was found between antihypertensive medication use and breast cancer DSS (HR 0.82, 95% CI 0.63-1.08, Figure 3D).

Risk of Publication Bias
The results of the Begg's funnel plot and Egger's regression test, showed no evidence of potential publication bias in articles reporting the use of BBs (P 0.168), CCBs (P 0.151), diuretics (P 0.348), or RASIs (P 0.397) in relation to the risk of breast cancer risk (Supplementary Figure S1). In addition, a potential publication bias was observed in studies on the use of antihypertensive medication with recurrence (P 0.001), but not in other studies on prognosis (mortality: P 0.717; OS: P 0.096; DSS: P 0.125) (Supplementary Figure S2).

Sensitivity Analysis
A sensitivity analysis was performed as shown in Supplementary  Figures S3 and S4. There was no significant alteration in the pooled RRs or HRs after sequentially deleting a single study from the overall pooled analyses each time. In the sensitive analysis of diuretics users with risk of breast cancer, the pooled RR was not statistically significant after the deletion of an article (RR 1.03, 95% CI 0.98-1.08) (Biggar et al., 2013).           Abbreviations: NOS, Newcastle-Ottawa Scale; ACEI, angiotensin-converting enzyme inhibitors; ARB, angiotensin II receptor blockers; CCB, calcium-channel blockers; BB, beta-blockers; RASI, renin-angiotensin system inhibitors; BCSM, breast cancer-specific mortality; OS, overall survival; DSS, disease-specific survival; DFS, disease-free survival; NA, Not Available.

DISCUSSION
Our meta-analysis showed that the use of BBs, CCBs, or diuretics was associated with a moderate increase in the risk of breast cancer. Diuretic therapy for ≥10 years could increase breast cancer risk by 10%. Long-term use of RASIs exerted a protective effect against breast cancer, especially when used for ≥10 years. In addition, this meta-analysis indicated that diuretics are related to elevated breast cancer-specific mortality, but antihypertensive medication use did not affect recurrence, OS, or DSS. Our conclusion that BBs were associated with increased breast cancer risk was consistent with this highly influential and large studies (Numbere et al., 2015;Chang et al., 2016;Zheng et al., 2021). The prospective study data minimizes the recall bias and in the study in question, some important confounders such as alcohol status, smoking status, BMI, medication and comorbidities, were adjusted. Similarly, the positive association between diuretics and breast cancer risk mainly depended on the study that evaluated cancer risk in 2.3 million Danish women who were followed-up for 28.8 million person-years (Biggar et al., 2013). The study reported that breast cancer risk increased in the first-year of diuretic use, but the association was attributed to reverse causality, such as where symptoms of cancer, including abdominal swelling or incidental findings such as hypertension, led to diuretic exposure. To exclude care practices that affect the risk from true etiologic impact, the study further examined risk only in women who were exposed to drugs for more than one year prior to diagnosis, and found a slightly increased risk of breast cancer with increasing duration of diuretic exposure. Consistent Frontiers in Pharmacology | www.frontiersin.org with its results, we observed a duration-response association between the duration of diuretic therapy ≥10 years and breast cancer risk. Studies found that thiazide diuretics, but not other diuretics, were linked to an increasing risk of breast cancer, and the authors considered the increase in insulin resistance due to thiazide diuretics as a possible explanation (Sarafidis et al., 2007;Ni et al., 2017). A recent study showed that the use of furosemide or other diuretics before breast cancer diagnosis increased the risk of breast cancer-related death (Santala et al., 2020). Our results showed that the breast cancer mortality increased with the use of diuretics; however, we failed to analyze the effect of diuretic subtypes on breast cancer prognosis because there were few relevant studies. Thus, we could not arrive at the conclusion that different subclass of diuretics had varying associations with breast cancer. Future studies are warranted to determine the potential difference among various diuretics.
Our meta-analysis showed a positive association between CCBs use and the risk of breast cancer, which was consistent with the results of previous studies (Wright et al., 2017;Busby et al., 2018a). The elevated levels of cytosolic calcium could affect the process of apoptosis through different signaling pathways such as caspase activation, induction of endonuclease activity, or the miRNA-524-5p-BRI3-extracellular signal-regulated kinase (Erk) pathway (Iwasawa et al., 2011;Guo et al., 2014;Zhao et al., 2017b). CCBs blocks calcium entry, inhibiting this vital process, which could destroy the body's natural defense mechanism against cancer growth and promote cell survival by initiating autophagy (Sun et al., 2018). However, there is no conclusive evidence to accurately explain this observed link and the benefits of the long-term use of CCBs are controversial.
The association between RASIs and breast cancer remains controversial with some studies reporting that ARB/ACEI was not related to breast cancer risk (Teo, 2011;Datzmann et al., 2019), whereas another concluded that users of ACEIs had an increased risk (Hallas et al., 2012). However, our result demonstrated a significant (22%) reduction in breast cancer risk with ≥10 years, which is consistent with previously published observational studies (Huang et al., 2011;Li et al., 2013;Wang et al., 2013;Ni et al., 2017). In addition, in vivo and in vitro studies on RASIs have demonstrated a protective antiinflammatory, anti-proliferative, and pro-apoptotic effect on breast cancer (Herr et al., 2008;George et al., 2010). Although the effect of RASIs on breast cancer is still contradictory, RASIs should be selected more often for patients undergoing antihypertensive therapy considering its availability, feasibility, and established safety with reduced breast cancer risk.
It is important to assess the possible source of heterogeneity. First, the significance of our results might have been affected by an indication bias, such as high blood pressure (Coogan, 2013), as both hypertension and cancer are associated with obesity, alcohol, smoking, cardiometabolic abnormalities, and other comorbidities. One study reported a relationship between hypertension and cancer (Han et al., 2017). In addition, in this meta-analysis, not all included studies adjusted for the confounding factor of hypertension which would cause a certain degree of heterogeneity. Second, in observational studies of the general population, it is important to determine a dose or duration-response relationship to properly interpret the potential link between drug exposure and cancer development. However, because of the differences in drug exposure time and dose between the included studies, the results will eventually be heterogeneous. Third, most of studies were from North America or Europe but a few studies are from Asia, which produced regional heterogeneity and limited the extent to which the results can be generalized. In the subgroup analysis according to geographical area, we found reduced between-study heterogeneity and observed different results, which indicated that the association varied among different areas. Finally, because the included studies were retrieved from some population registration database, the analysis results could reduce recall bias and misclassification bias to some extent. However, population-based registries also have some shortcomings, such as the limitation of potential confounders by stored data, and the fact that a prescription was collected does not necessarily indicate that the prescribed drugs were taken. Our meta-analysis provides convincing and clear evidence of the relationship between antihypertensive drugs and the risk and prognosis of breast cancer. One strength of our study is that we conducted a stratified analysis according to the duration of antihypertensive drug use and found that using diuretics for ≥10 years could increase the risk of breast cancer, but using RASIs for ≥10 years would reduce the risk of breast cancer. This suggests that RASIs would be the best choice for patients with long-term antihypertensive drugs. Another strength of the study is that we focused on breast cancer risk and prognosis to obtain a more homogeneous group of studies. We studied the possible source of heterogeneity between studies, including study design, geographic area, and duration of antihypertensive medication use. It is noteworthy that there were also some limitations to this study. First, biases such as confounders, possible detection bias, and selection bias is inevitable in observational studies. Secondly, subgroup analyses based on age, sex, pathology of breast cancer, dose of antihypertensive medication, subclass of BB (selective BBs and non-selective BBs), CCBs (dihydropyridines and non-dihydropyridines), and diuretics (thiazide, loop, and potassium sparing diuretics) were not conducted because of the lack of relevant studies. Thirdly, our current summary evidence did not support a clear relation between long-term use of CCBs or BBs and breast cancer risk. Finally, significant heterogeneity was observed among studies focused on antihypertensive medication use and breast cancer risk and this persisted even when the data were stratified into subgroups.
The results of this meta-analysis demonstrated that BBs, CCBs, and diuretics were significantly associated with breast cancer risk. Unexpectedly, a beneficial effect of long-term use of RASI was observed against breast cancer risk. Additionally, according to the analysis of prognosis, no association was observed with any class of antihypertensive medication. Considering the limitations in this meta-analysis, further research is needed to fully clarify the association between antihypertensive medication use and breast cancer risk.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
All authors read, critically reviewed and approved the final manuscript. YX and MW conducted the database searches, screened titles, abstracts and full-texts for eligibility, PX, YD, YZ, and SY performed study quality assessments. JC and ZD planed and designed the research; DZ, NL, and NW provided methodological support/advice; PX and YD tested the feasibility of the study; YX, MW, and DZ extract data; YX performed the statistical analysis; YX and MW wrote the manuscript.