Association Between Aromatase Inhibitors and Myocardial Infarction Morbidity in Women With Breast Cancer: A Meta-Analysis of Observational Studies

Abstract

Background

The cardiovascular toxicity of aromatase inhibitors (AIs) for women with estrogen receptor-positive breast cancer is controversial. We aimed to evaluate the association between AIs and the risk of myocardial infarction (MI) in women with estrogen receptor-positive breast cancer based on real-world studies.

Method

PubMed, Embase, and Cochrane Library were searched to identify studies that estimated the association between MI risk and AIs. A random-effects model was used to evaluate the hazard ratio (HR) and 95% confidence intervals (CIs) of the predefined outcomes.

Results

A total of 134 476 patients from eight cohort studies were enrolled in our analysis. For MI incidence, no significant difference was found between the users of AIs and non-users (HR: .98, 95% CI: .83-1.17). The subgroup analysis of patients without a history of cardiovascular disease (CVD) suggested a reduced risk of MI (HR: .86, 95% CI: .77-.96). No significant difference was found for ischemic stroke (HR: .93, 95% CI: .82-1.07) and heart failure (HR: 1.24, 95% CI: .92-1.66) between the two groups.

Conclusion

Based on real-world data, AIs may be a safe treatment route for patients with estrogen receptor-positive breast cancer and those with a history of CVD. AIs caused a major decrease in MI in patients without CVD history. However, more in-depth investigations are needed to explore the association between AI use and the incidence of MI in the treatment of estrogen receptor-positive breast cancer.

Keywords

myocardial infarction aromatase inhibitors tamoxifen breast cancer meta-analysis

Introduction

Breast cancer has become the most commonly diagnosed female malignancy in various nations,¹ and it is matched by the highest mortality. As a mainstay of treatment, aromatase inhibitors (AIs, ie, anastrozole, letrozole, and exemestane) and tamoxifen are generally prescribed for women with estrogen receptor-positive breast cancer.² AIs have been proven superior to tamoxifen, with third-generation AIs displacing tamoxifen as the cornerstone endocrine treatment for estrogen receptor-positive breast cancer. Data from several studies showed that AIs significantly reduced recurrence and improved the overall survival rates of breast cancer.³ However, cardiovascular adverse events (CVAEs) have become a major cause of noncancer-related chronic morbidity and mortality due to the decline in breast cancer mortality, and the cardiovascular toxicity of therapies has been observed in recent years.⁴

Several meta-analyses based on randomized controlled trials (RCTs) have been conducted to investigate AI-related cardiovascular toxicity.^5-7 Although RCTs are acknowledged as the “gold standard” for evaluating the safety and effectiveness of drugs, the findings of clinical trials may not be translatable to situations outside clinical trial settings because the inevitable strict inclusion and exclusion criteria usually mean that the trial participants are not representative of the patient populations encountered in clinical practices.⁸ Real-world data may be acquired from electronic health and claims databases with generally observational, retrospective, or prospective data collected over a long period of time. These data can accordingly provide information regarding the long-term safety, particularly rare events, and effectiveness of agents in large heterogeneous populations; when handled correctly, they are a vital stand-alone source of evidence that complements RCTs and other studies, which together provide the basis for decision-making.⁹ To date, information on the cardiovascular toxicity of AIs toward patients with a cardiovascular disease (CVD) history based on real-world data is lacking. An increasing number of proofs from real-world studies, which were used to be ignored and are generally based on databases and with long-term follow-up, have gained attention.¹⁰ Several population-based observational studies reported that the use of AIs had no significant association with the increased risk of myocardial infarction (MI), which was inconsistent with the findings of previous meta-analyses.^11-16 Another meta-analysis showed a similar result.¹⁷ However, these studies failed to explore further the association between AIs and MI morbidity in women with breast cancer of diverse demographic characteristics, such as patients with a history of CVD. Accordingly, we conducted a meta-analysis of the association between AI treatment and the risk of MI based on real-world data.

Methods

The present meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines¹⁸ and Meta-analysis of Observational Studies in Epidemiology.¹⁹ The protocol was registered with INPLASY with the number INPLASY202270086. As the present article is a meta-analysis, ethical approval was not needed.

Literature Search

PubMed, Embase, and Cochrane Library were searched independently and systematically by two members to identify literature from inception to February 2022. The search strategy was based on the combination of Medical Subject Headings and terms, such as “aromatase inhibitors” or “fadrozole” or “aminoglutethimide” or “aromatase inhibitor*” or “aminoglutethimide” or “anastrozole” or “letrozole” or “femara” or “exemestane” or “aromasin” or “vorozole” or “rivizor” or “formestane” or “lentaron” or “fadrozole” or “afema” and “breast neoplasms” or “breast cancer*” or “breast neoplasm*” or “breast carcinoma*” or “breast tumor*” and “myocardial Infarction” or “coronary syndrome*” or “heart infarct*” or “coronary” and “occlusion*” or “myocard*” and “infarct*” or “cardiac infarct*” or “heart attack*” or “cardiovascular strok*” or “STEMI” or “NSTEMI” or “AMI.” The search was restricted to studies published in English. In addition, all referenced articles in the included studies were retrieved manually to acquire more relevant research.

Selection Criteria

Endnote was used to exclude duplicates. The titles and abstracts of the remaining publications were scanned by two authors independently. The inclusion criteria were applied to classify the articles as included, excluded, or uncertain. We assessed the full texts carefully, and agreements were reached by discussion or engagement of a third reviewer when necessary. The criteria for study selection were as follows: (1) population: female participants with estrogen receptor-positive breast cancer and aged more than 18 years old; (2) intervention: AIs prescribed as adjuvant endocrine or extended adjuvant endocrine therapy; (3) comparison: comparison made between patients of AI treatment vs non-users; (4) outcome: reported incidence of MI for women with estrogen receptor-positive breast cancer; (5) design: cohort or case-control study. Finally, studies without odds ratio, hazard ratio (HR), or risk ratio and those without available data were excluded.

Data Extraction

Two authors extracted the following data separately for each included study: first author, year of publication, data source, study type, age, sample size, treatment, and outcomes. Discrepancies were resolved by consensus or a third reviewer.

Risk of Bias Assessment

The Newcastle–Ottawa Scale (NOS) was used to evaluate the risk of bias.²⁰ The NOS was used to assess the quality of the included studies through three parameters: selection, comparability, and outcome. Maximum scores of 4, 2, and 3 were assigned to the selection, comparability, and outcome, respectively. Studies were categorized as low- (less than 5 points), moderate- (5-7 points), and high-quality (more than 7 points) research.

Grading of Recommendations Assessment, Development, and Evaluation Evaluation

The GRADE was used to determine the quality of evidence based on the following criteria: study design, risk of bias, inconsistency rating in results, rating of indirectness of evidence, and others. The evidence quality was determined as high, moderate, low, and very low.

Primary and Secondary Outcomes

The primary outcome was MI after the AI treatment in women with estrogen receptor-positive breast cancer. The secondary outcomes were ischemic stroke (IS) and heart failure (HF). The outcomes in all studies were ascertained by administrative diagnostic codes.

Statistical Analysis

This meta-analysis was conducted using Stata (Stata Version 16.0; Stata Corporation, College Station, TX, USA). The pooled HR and 95% confidence intervals (CIs) were calculated for binary outcomes, and a P value of less than .05 was deemed statistically significant. The heterogeneity was evaluated using the I-square (I²) statistic. The random-effect model was applied to the present meta-analysis, considering the probably high heterogeneity due to clinical and methodological factors. Subgroup analysis was conducted based on CVD history, propensity score-matched cohort, tamoxifen or no treatment, age of patients, and population-based studies to investigate the association between AI usage and MI morbidity in women with estrogen receptor-positive breast cancer. Considering immortal time bias, we included studies with time-dependent model to conduct a subgroup analysis. To assess the stability of the primary outcome, we performed a sensitivity analysis by deleting trials sequentially. The contour-enhanced funnel was used to assess the potential biases of the included studies, and Egger’s test was applied to further analyze the biases quantitatively.²¹ We also calculated the number needed to harm (NNH), which is the number of patients needed to be treated by AIs for an adverse event to occur in an individual who would not otherwise have suffered.

Results

Eligible Studies

We identified 369 potentially relevant articles in the initial search. A total of 36 duplicated studies were removed, and 294 articles were excluded based on their titles and abstracts. Subsequently, 39 studies were evaluated by their full text. Then, 31 studies were eliminated for the following reasons: 19 trials were not real-world studies, 1 publication was not published in English, 4 articles lacked available data, and 7 studies were conducted based on the same databases. Ultimately, eight real-world studies involving 134 476 patients were included in the final analysis.^11-16,22,23 For studies based on the same databases, we included the latest research to obtain the most updated data. Figure 1 shows the search process.

Figure 1.

Flow diagram showing search strategy and inclusion and exclusion of studies for meta-analysis.

Study and Patient Characteristics

Table 1 summarizes the characteristics of the included studies and patients. All the included studies were cohort studies. Eight trials had sample capacities ranging from 74 to 39 315. Given the flaw of the study design of one research, all participants were women who underwent cardiac angiography, which indicated that the patients enrolled may have suffered a relatively high incidence of CVD. Thus, we eliminated the study in our analysis. The mean ages in the included studies were variable, and nearly all women in the cohorts were identified as postmenopausal. Follow-up periods in the included studies ranged from 1.3 years to 5 years. Of the included eight studies, six were population-based research. Three trials conducted a subgroup analysis based on the CVD history, and three studies enrolled patients without a history of CVD. Confounding variables were balanced by a propensity score-matched cohort in six trials.

Table 1.

Patient Characteristics at Baseline in Observational Studies of AIs and Tamoxifen Included in the Study.

Study	Database	Study type	Intervention	No. of patients	Age distribution, yr	Outcome: no. of events/no. exposed (no. of events/no. unexposed)
Ligibel, 2011	HIRDSM	Cohort	AIs vs no hormonal treatment	39 315	AIs: 64.6	MI, IS
Ligibel, 2011	HIRDSM	Cohort	AIs vs no hormonal treatment	39 315	No therapy: 67.9	MI, IS
Seruga, 2014	UCCL	Cohort	AIs vs tamoxifen	74	AIs: 42-81 (69)	MI
Seruga, 2014	UCCL	Cohort	AIs vs tamoxifen	74	Tamoxifen: 49-84 (68)	MI
Abdel, 2016	The Ontario drug benefit program	Cohort	AIs vs tamoxifen	9350	AIs: 66-77 (71)	MI
Abdel, 2016	The Ontario drug benefit program	Cohort	AIs vs tamoxifen	9350	Tamoxifen: 68-80 (74)	MI
Haque, 2016	SEER	Cohort	AIs vs tamoxifen	8014	All: 51-83 (66.8)	MI, HF
Kamaraju, 2018	SEER	Cohort	AIs vs tamoxifen	5648	All: > 67	MI
Choi, 2020	NHID	Cohort	AIs vs no hormonal treatment	38 391	AIs: 50-76 (63.3)	MI, IS
Choi, 2020	NHID	Cohort	AIs vs no hormonal treatment	38 391	Tamoxifen: 50-78 (63.8)	MI, IS
Khosrow-Khavar, 2020	CPRD	Cohort	AIs vs tamoxifen	17 922	AIs: 46-90 (68.1)	MI, IS, HF
Khosrow-Khavar, 2020	CPRD	Cohort	AIs vs tamoxifen	17 922	Tamoxifen: 46-87 (67.9)	MI, IS, HF
Franchi,2021	The healthcare utilization databases of Lombardy	Cohort	AIs vs tamoxifen	33 946	All: ≥50	MI, IS, HF

AIs: aromatase inhibitors, MI: myocardial infarction, IS: ischemic stroke HF: heart failure, HIRDSM: Health Core Integrated Research Database, UCCL: University Clinical Center of Ljubljana, SEER: Surveillance, Epidemiology, and End Results, NHID: the National Health Information Database, CPRD: Clinical Practice Research Datalink.

Bias Assessment and Sensitivity Analysis

The quality of each study was consistent according to the NOS assessment (Table 2). All studies were scored 8 or higher, implying their low risk of introducing biases. A visual inspection of the contour-enhanced plot suggested that the plot was asymmetric (Figure 2). However, Egger’s test obtained a P value of .797, indicating that no publication bias was found in the present analysis. In addition, sensitivity analysis showed that the effect estimate remained unchanged, which indicated the robustness of the results.

Table 2.

NOS Score for Observational Studies Included.

Key points for cohort studies assessment		Ligibel, 2011	Seruga, 2014	Abdel, 2016	Haque, 2016	Kamaraju, 2018	Choi, 2020	Khosrow-Khavar, 2020	Franchi, 2021
Selection	Representativeness of the exposed cohort	★	★	★	★	★	★	★	★
	Representativeness of the nonexposed cohort	★	★	★	★	★	★	★	★
	Ascertainment of exposure	★	☆	★	★	★	★	★	★
	Outcome of interest was not present at the beginning of study	☆	★	☆	★	☆	★	☆	★
Comparability	Comparability of cohorts because of the design or analysis	★	★	★	★	★	★	★	★
Outcome	Assessment of outcome	★	★	★	★	★	★	★	★
	Follow-up long enough for outcomes to occur	★	★	★	★	★	★	★	★
	Adequacy of follow up of cohorts	★	★	★	★	★	★	★	★
NOS score		8	8	8	9	8	9	8	9

NOS: Newcastle-Ottawa Scale; ★: Satisfy the condition; ☆: Dissatisfy the condition.

Figure 2.

Contour-enhanced funnel plot for the association between AIs and MI.

Primary Outcome

Table 1 displays the counts of MI events in the eight studies. The pooled results of the analysis showed no significant difference between AI users and non-users (HR: .98, 95% CI: .83-1.17; Figure 3). However, moderate heterogeneity was observed across studies (I² = 57.52%). In terms of the absolute risks, 1.20% and .96% of patients in the AI groups and non-users experienced MI, respectively (difference in absolute risk = .3%, NNH = 416).

Figure 3.

The forest plot for the association between AIs and MI.

Secondary Outcome

Patients prescribed AIs were associated with a nonsignificant, reduced risk of IS (NNH: 609) compared with the non-users (HR: .93, 95% CI: .82-1.07; I² = 37%, Figure 4). A nonsignificant difference was found in HF incidence (HR: 1.24, 95% CI: .92-1.66; I² = 77%, Figure 5).

Figure 4.

The forest plot for the association between AIs and IS.

Figure 5.

The forest plot for the association between AIs and HF.

Subgroup Analysis

Subgroup analysis suggested that the patients prescribed AIs without CVD history had a significantly low MI incidence (HR: .86, 95% CI: .77-.96; I² = 0%, Figure 6). However, no significant difference was reported in the incidence of MI in patients with a history of CVD (HR: 1.17, 95% CI: .83-1.65; I² = 13%, Figure 6). In the subgroup analysis of the propensity score-matched cohort, forest plots demonstrated no significant difference between the two groups in terms of MI incidence, and the results of subgroup analysis based on elderly (>65 years) and population-based studies were similar (Figures S1–S3). In addition, the AI group showed a significantly low incidence of MI compared with no treatment (HR: .84, 95% CI: .71-.99; I² = 0%, Figure S4). However, a nonsignificant difference was observed in the comparison between AIs and tamoxifen (HR: 1.08, 95% CI: .84-1.40; I² = 66.97%, Figure S4), but when a study by Abdel et al was excluded, the I² changed (HR: .96, 95% CI: .79-1.17; I² = 40.61%, Figure S5). Subgroup analysis based on time-dependent model showed similar result with primary outcome with a low heterogeneity (HR: .94, 95% CI: .81-1.09; I² = 21.27%, Figure S6).

Figure 6.

Subgroup analysis based on CVD history for the association between AIs and HF.

GRADE Evaluation

All included studies were observational studies. The quality of outcomes was assessed by GRADE evaluation. As the I² statistics were moderate and high, the corresponding inconsistency was downgraded to “serious” and “very serious,” respectively, for the results of MI and HF. GRADE evaluations for MI and HF were very low, and the quality for IS was low. Table 3 summarizes the overall results of the GRADE evaluation.

Table 3.

Result of GRADE Evaluation.

Outcomes	Number of studies	No of participants	Relative effect (95% CI)	Quality of the evidence (GRADE)	Comments
MI	7	134 476	RR: .98	⊕○○○	Inconsistent was “serious”
MI	7	134 476	(.83 to 1.17)	very low	Inconsistent was “serious”
IS	5	119 404	RR: .93	⊕⊕○○
IS	5	119 404	(.82 to 1.07)	low
HF	3	41 698	RR: 1.24	⊕○○○	Inconsistent was “very serious”
HF	3	41 698	(.92 to 1.66)	very low	Inconsistent was “very serious”

MI: myocardial infarction; IS: ischemic stroke; HF: heart failure; RR: risk ratio.

Discussion

The present work included real-world studies to evaluate the relationship between AIs and the risk of MI. The results of the meta-analysis showed that the use of AIs did not increase the incidences of MI, IS, and HF in women with estrogen receptor-positive breast cancer compared with non-users. The subgroup analysis of MI incidence suggested that AIs were associated with a significantly reduced MI risk in patients without CVD history. However, no similar result was observed in women with a history of CVD. Notably, the changes in I² values indicated that CVD history was part of the source of heterogeneity. We assumed that such a result may be related to the notion that patients who had suffered from acute coronary syndromes are more inclined to experience recurrent CVAEs.²⁴ In addition, patients with a CVD history have usually been associated with the development of dyslipidemia. Thus, the conclusion of Holmes et al,²⁵ who suggested that harmful lipoprotein particles can elevate the risk of MI, may support our findings in patients with a history of CVD. In addition, subgroup analysis based on the elderly, propensity score-matched cohort, and population-based studies showed similar results to the primary outcome. Compared with those with no hormonal treatment, the patients prescribed AIs showed a reduced MI morbidity. However, the upper CI approached the null value. Accordingly, we should be cautious in the interpretation of the results. In contrast, in studies employing tamoxifen, no difference was found between the two groups. The low I² in subgroup analysis based on tamoxifen or no treatment and time-dependent model indicated that the high heterogeneity may partly be due to these factors.

Several meta-analyses have focused on the cardiovascular side effects of AIs on women with estrogen receptor-positive breast cancer, as cardiac toxicity emerged as a major concern in the treatment of breast carcinoma with endocrine therapy. However, all the studies enrolled were not designed for CVAEs specifically, and the outcome definitions in these meta-analyses were discordant. A study where the pooled outcome was graded 3 and 4 CVAEs suggested that the incidence of CVAEs was higher in participants receiving AIs than in those prescribed tamoxifen. The finding was statistically significant, although with a low absolute difference (approximately .50%) and a large NNH (>180 patients).⁷ Conversely, another recent analysis defined outcomes as all CVAEs and suggested that the high CVD incidence associated with AI treatment was likely due to the cardiovascular protective effect of tamoxifen.⁵

Previous studies compared the cardiovascular toxicity of AIs with that of tamoxifen or nonhormonal treatment but reached discordant conclusions. Three population-based studies suggested no significant difference in the MI morbidity between AI- and tamoxifen-treated patients with estrogen receptor-positive breast cancer during follow-up periods.^11,12,14 However, Haque et al¹⁶ reported a negative association; the CVD risk factors and covariates were well controlled in the two groups in their study. Nevertheless, two other research claimed a higher cause-specific hazard of MI in the AI group than in the tamoxifen group.^22,23 Hence, a meta-analysis based on real-world data will be helpful in settling the real-world disagreement between these treatments.

Previous studies have proposed differential mechanisms for the association between AIs and MI incidence. An experiment on male rats raised a possible mechanism in which AIs predispose male mice to the progress of atherosclerosis with a direct effect on the endothelium.²⁶ This condition may directly explain the association of AI treatment with a high MI risk in previous meta-analyses. However, the applicability of this mechanism to humans requires further verification. In addition, two studies suggested that AIs may increase the incidence of MI by inducing a rise in the level of low-density lipoprotein cholesterol (LDL-C).^6,12 AI studies, such as those on Adjuvant post-Tamoxifen Exemestane vs Nothing Applied (ATENA) or The National Cancer Institute of Canada Clinical Trials Group MA.17, employed placebo or no treatment and therefore provided accountable results on the influence of AI treatment on lipid levels. The ATENA study reported that the AI group and non-users exhibited increases in the total cholesterol and LDL-C levels from the baseline but without a significant difference between groups.²⁷ The MA. 17 study obtained consistent findings: participants who received letrozole and placebo showed elevated lipid levels compared with the baseline,²⁸ and no evident difference was observed between groups. Another widely known mechanism is that tamoxifen exerts a cardio-protective effect by altering the serum lipid levels. This mechanism may indirectly explain why the difference in the association between AI or tamoxifen treatment and MI morbidity can be attributed to the cardio-protective effect of tamoxifen rather than the potential cardiotoxicity of AIs.²⁹ All these mechanisms are associated with serum lipid alteration and atherosclerosis. However, these hypotheses cannot explain the findings of our meta-analysis, which showed no significant difference between AI users and non-users in terms of MI morbidity. We assumed that the phenomenon was related to active lipid management in clinical practice, in which participants are usually prescribed lipid-lowering drugs once dyslipidemia is detected during clinical hormonal treatment, especially among patients with a CVD history.

Some RCTs included in previous meta-analyses were small-sample studies, whose restricted statistical power may restrain the detection of differences in the incidence of MI.³⁰ Real-world studies can be used to evaluate the safety of interventions administered outside the setting of clinical trials and improve the limited statistical power for detecting possible differences in the incidence of rare CVAEs, such as MI and HF. The present meta-analysis was conducted based on real-world studies with large sample sizes. Another important strength of our meta-analysis was that the characteristics of patients included were usually more similar to those of patients in clinical practice. Thus, the target population was representative of patients in clinical practice. We also calculated the NNH and absolute risk of MI to detect differences in the incidence of rare adverse events. In addition, subgroup analysis was conducted to search for the source of heterogeneity, that is, the CVD history.

Our meta-analysis also encountered some limitations. First, this meta-analysis was a pooled analysis of published observational studies that employed different AI choices, treatment doses, treatment lengths, and follow-up periods. We did not account for adjusting confounders given the insufficiency of data provided by these studies. This condition may reduce the accuracy of the findings on the cardiovascular toxicity of AIs, although most of the included studies adjusted for the observed confounders. Another important drawback was the potential bias that might have been introduced because most patients were not randomly allocated to the treatment arms. Rather, doctors assigned them to particular treatments based on their medical conditions. However, most of the included studies used a propensity score-matched cohort to ensure the comparability between the two groups, and our subgroup analysis based on the propensity score-matched cohort showed a similar result to the primary outcome. Third, the I² statistic indicated moderate heterogeneity in the study despite the similar designs of all studies and the use of random-effect modeling. However, the subgroup analysis suggested that the CVD history of patients may be the source of heterogeneity. In addition, the sensitivity analysis suggested the robustness of the findings of the present study, and Egger’s test did not reveal significant biases in the analysis. Finally, despite the large sample size, the result of the analysis generated a CI approaching a null value, indicating that the present analysis may still have a relatively limited statistical power to detect differences due to the low number of exposed events. Thus, more experimental and epidemiological studies are required to delineate further the association between AIs and cardiovascular health in patients with estrogen receptor-positive breast cancer.

Conclusion

In summary, this meta-analysis of real-world studies indicated that AIs may be a safe treatment route for patients with estrogen receptor-positive breast cancer in clinical practice. However, more well-designed, multicenter, and large-sample studies are needed to ensure the scientific, objective, and reliable conclusions of trials in future clinical research in investigations of the cardiovascular toxicity of AI treatment in patients with estrogen receptor-positive breast cancer.

Supplemental Material

Supplemental Material - Association Between Aromatase Inhibitors and Myocardial Infarction Morbidity in Women With Breast Cancer: A Meta-Analysis of Observational Studies

Supplemental Material for Association Between Aromatase Inhibitors and Myocardial Infarction Morbidity in Women With Breast Cancer: A Meta-Analysis of Observational Studies by Jing-Chao Sun, Ze-Fan Sun, Chao-Jie He, Chang-Lin Zhai, and Gang Qian in Cancer Control

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Jiaxing Institute of Arteriosclerotic Diseases (2020-dmzdsys), Pioneer innovation team of Jiaxing Arteriosclerotic Diseases Research Institute (XFCX--DMYH), Provincial-Municipal Joint Construction of Key Medical Disciplines in Zhejiang Province (2019-ss-xxgbx), Zhejiang Provincial Basic Public Welfare Research Program of China under (LGF21H0006), Jiaxing Key Innovation Team Fund (2018-xjqxcxtd).

Abbreviations

AI: aromatase inhibitor

MI: myocardial infarction

HR: hazard ratio

CI: confidence interval

CVD: cardiovascular disease

CVAEs: cardiovascular events

NNH: number need to harm

RCT: randomized controlled trial

MeSH: Medical Subject Heading

IS: ischemic stroke

HF: heart failure

NOS: Newcastle-Ottawa Scale

I2: I-square

ATENA: Adjuvant post-Tamoxifen Exemestane vs Nothing Applied.

Availability of Data and Materials

Because this is a meta-analysis, all of data included in this study could be found in the included references.

ORCID iD

Jing-Chao Sun

Supplemental Material

Supplemental material for this article is available online.

References

Ghoncheh

Pournamdar

Salehiniya

. Incidence and mortality and epidemiology of breast cancer in the world. Asian Pac J Cancer Prev. 2016;17(S3):43-46. doi:10.7314/apjcp.2016.17.s3.43.

National Comprehensive Cancer Network (NCCN) . Clinical Practice Guidelines in Oncology. Breast Cancer Version 4. 2021; 2021. https://www.nccn.org/professionals/physician_gls/pdf/breast.pdf. Accessed 6 March 2021.

Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) . Aromatase inhibitors versus tamoxifen in early breast cancer: Patient-level meta-analysis of the randomised trials. Lancet. 2015;386(10001):1341-1352. doi:10.1016/S0140-6736(15)61074-1. Epub 2015 Jul 23.

Upshaw

Hubbard

, et al. Physical activity during and after breast cancer therapy and associations of baseline physical activity with changes in cardiac function by echocardiography. Cancer Med. 2020;9(17):6122-6131. doi:10.1002/cam4.3277. Epub 2020 Jul 9.

Khosrow-Khavar

Filion

Al-Qurashi

, et al. Cardiotoxicity of aromatase inhibitors and tamoxifen in postmenopausal women with breast cancer: A systematic review and meta-analysis of randomized controlled trials. Ann Oncol. 2017;28(3):487-496. doi:10.1093/annonc/mdw673.

Amir

Seruga

Niraula

Ocaña

. Toxicity of adjuvant endocrine therapy in postmenopausal breast cancer patients: A systematic review and meta-analysis. J Natl Cancer Inst. 2011;103(17):1299-1309. doi:10.1093/jnci/djr242. Epub 2011 Jul 9.

Cuppone

Bria

Verma

, et al. Do adjuvant aromatase inhibitors increase the cardiovascular risk in postmenopausal women with early breast cancer? Meta-analysis of randomized trials. Cancer. 2008;112(2):260-267. doi:10.1002/cncr.23171.

Blonde

Khunti

Harris

Meizinger

Skolnik

. Interpretation and impact of real-world clinical data for the practicing clinician. Adv Ther. 2018;35(11):1763-1774. doi:10.1007/s12325-018-0805-y.

Gokhale

Stürmer

Buse

. Real-world evidence: The devil is in the detail. Diabetologia. 2020;63(9):1694-1705. doi:10.1007/s00125-020-05217-1.

10.

Matthews

Stanway

Farmer

, et al. Long term adjuvant endocrine therapy and risk of cardiovascular disease in female breast cancer survivors: Systematic review. BMJ. 2018;363:k3845. doi:10.1136/bmj.k3845.

11.

Franchi

Tritto

Tarantini

Navazio

Corrao

. Adjuvant hormonotherapy and cardiovascular risk in post-menopausal women with breast cancer: A large population-based cohort study. Cancers. 2021;13(9):2254. doi:10.3390/cancers13092254.

12.

Kamaraju

Shi

Smith

Nattinger

Laud

Neuner

. Are aromatase inhibitors associated with higher myocardial infarction risk in breast cancer patients? A medicare population-based study. Clin Cardiol. 2019;42(1):93-100. doi:10.1002/clc.23114. Epub 2018 Dec 7.

13.

Choi

Kim

Park

, et al. Effects of tamoxifen and aromatase inhibitors on the risk of acute coronary syndrome in elderly breast cancer patients: An analysis of nationwide data. Breast. 2020;54:25-30. doi:10.1016/j.breast.2020.08.003. Epub 2020 Aug 19.

14.

Khosrow-Khavar

Filion

Bouganim

Suissa

Azoulay

. Aromatase inhibitors and the risk of cardiovascular outcomes in women with breast cancer: A population-based cohort study. Circulation. 2020;141(7):549-559. doi:10.1161/CIRCULATIONAHA.119.044750. Epub 2020 Feb 17.

15.

Ligibel

James O’Malley

Fisher

Suissa

Azoulay

. Risk of myocardial infarction, stroke, and fracture in a cohort of community-based breast cancer patients. Breast Cancer Res Treat. 2012;131(2):589-597. doi:10.1007/s10549-011-1754-1. Epub 2011 Sep 1.

16.

Haque

Shi

Schottinger

, et al. Cardiovascular disease after aromatase inhibitor use. JAMA Oncol. 2016;2(12):1590-1597. doi:10.1001/jamaoncol.2016.0429.

17.

Zheng

. Risk of cardiovascular disease in breast cancer patients receiving aromatase inhibitors vs. tamoxifen: A systematic review and meta-analysis. J Clin Pharm Ther. 2022;47(5):575-587. doi:10.1111/jcpt.13598. Epub 2022 Jan 4.

18.

Moher

Liberati

Tetzlaff

Altman

PRISMA Group . Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. PLoS Med. 2009;6(7):e1000097. doi:10.1371/journal.pmed.1000097. Epub 2009 Jul 21.

19.

Stroup

Berlin

Morton

, et al. Meta-analysis of observational studies in epidemiology: A proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000;283(15):2008-2012. doi:10.1001/jama.283.15.2008.

20.

The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomized studies in meta-analyses; 2019. http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp. Accessed 6 March 2021.

21.

Harbord

Harris

Sterne

JAC

. Updated tests for small-study effects in meta-analyses. STATA J. 2009;9:197-210.

22.

Abdel-Qadir

Amir

Fischer

, et al. The risk of myocardial infarction with aromatase inhibitors relative to tamoxifen in post-menopausal women with early stage breast cancer. Eur J Cancer. 2016;68:11-21. doi:10.1016/j.ejca.2016.08.022. Epub 2016 Sep 30.

23.

Seruga

Zadnik

Kuhar

Daniel

Winer

Keating

. Association of aromatase inhibitors with coronary heart disease in women with early breast cancer. Cancer Invest. 2014;32(4):99-104. doi:10.3109/07357907.2014.880452. Epub 2014 Feb 18.

24.

Mach

Baigent

Catapano

, et al. 2019 ESC/EAS Guidelines for the management of dyslipidaemias: lipid modification to reduce cardiovascular risk. Eur Heart J. 2020;41(1):111-188. doi: 10.1093/eurheartj/ehz455. Erratum in: Eur Heart J. 2020 Nov 21;41(44):4255.

25.

Holmes

Millwood

Kartsonaki

, et al. Lipids, lipoproteins, and metabolites and risk of myocardial infarction and stroke. J Am Coll Cardiol. 2018;71(6):620-632. doi:10.1016/j.jacc.2017.12.006.

26.

Nathan

Shi

Dinh

, et al. Testosterone inhibits early atherogenesis by conversion to estradiol: Critical role of aromatase. Proc Natl Acad Sci U S A. 2001;98(6):3589-3593. doi:10.1073/pnas.051003698. Epub 2001 Mar 6.

27.

Markopoulos

Dafni

Misitzis

, et al. Extended adjuvant hormonal therapy with exemestane has no detrimental effect on the lipid profile of postmenopausal breast cancer patients: Final results of the ATENA lipid substudy. Breast Cancer Res. 2009;11(3):R35. doi:10.1186/bcr2320. Epub 2009 Jun 16.

28.

Wasan

Goss

Pritchard

, et al. The influence of letrozole on serum lipid concentrations in postmenopausal women with primary breast cancer who have completed 5 years of adjuvant tamoxifen (NCIC CTG MA.17L). Ann Oncol. 2005;16(5):707-715. doi:10.1093/annonc/mdi158. Epub 2005 Apr 7.

29.

Younus

Kissner

Reich

Wallis

. Putting the cardiovascular safety of aromatase inhibitors in patients with early breast cancer into perspective: A systematic review of the literature. Drug Saf. 2011;34(12):1125-1149. doi:10.2165/11594170-000000000-00000.

30.

Tsang

Colley

Lynd

. Inadequate statistical power to detect clinically significant differences in adverse event rates in randomized controlled trials. J Clin Epidemiol. 2009;62(6):609-616. doi:10.1016/j.jclinepi.2008.08.005. Epub 2008 Nov 14.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.28 MB