Aromatase inhibitors and risk of cardiovascular events in breast cancer patients: a systematic review and meta-analysis

Background Cardiovascular events (CVEs) was considered as one of the primary cause to reduce the quality of life in breast cancer patients with aromatase inhibitors (AIs) treatment, which has not been sufficiently addressed. The aim of this study was to assess the correlation between risk of CVEs and AIs in patients with breast cancer. Methods Included studies were obtained from the databases of Embase, Pubmed, Cochrane Library, Clinical Trials.gov, and reference lists. The main outcome measures were overall incidence, odds ratios (ORs), and 95% confidence intervals (CIs). Furthermore, the association and the risk differences among different tumor types, AIs,ages,or treatment regimens were conducted. Fixed-effect or random-effect models were applied in the statistical analyses according to the heterogeneity. Our analysis was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. Results Seventeen studies, which included 44,411 subjects, were included in our analyses. The overall incidence of CVEs in AIs group was 13.02% (95% CI: 8.15–20.17%) and almost all of the high-grade CVEs occurred in patients treated with AIs. The pooled ORs of CVEs was 0.9940 (95% CI: 0.8545–1.1562). Under sub-group analysis, the incidence of CVEs related to exemestane was higher than that of controls (OR = 1.1564, 95% CI: 1.0656–1.2549), but no statistical differences in risk of CVEs were found in other sub-group analysis. No evidence of publication bias was found for incidence of CVEs in our meta-analysis by a funnel plot. Conclusions These results suggest that patients with breast cancer treated with AIs do not have a significant risk of developing CVEs in comparison with the controls, and exemestane might not be considered as the alternative AI to the breast cancer patients from the perspective of CVEs. Further studies are recommended to investigate this association and the risk differences among different tumor types, AIs or treatment regimens.

AIs or treatment regimens.

Background
Breast cancer is an increasing public health problem throughout the world, which is one of the most common malignancies and causes for tumor-related deaths among women [1][2][3].Fortunately, The 5year survival rate for patients with breast cancer has elevated from an average of 53% in 2007 to 85% in 2012 and the number is still rising in recent year [4,5]. However, this positive trend in improved cancer-related mortality is weakened by an emerging increase in cardiovascular (CV) morbidity and mortality in these patients. It is estimated that over $800 million will be spent annually in the US on providing CV care for these women with breast cancer [2,6]. Several recently published research studies have studied possible etiologies of these events, and it is suggested that the increased morbidity of CV disease and events has a temporal relationship with the administration of chemotherapy for cancer [7][8][9][10][11]. Nevertheless, the causes for this increase in cardiovascular related events has not been clearly demonstrated.
Anastrozole, letrozole, and exemestane are the three proven aromatase inhibitors (AIs), and several large randomized control trials (RCTs) have demonstrated the advantage of the AIs compared with tamoxifen [12][13][14].AIs have systematically been clarified to increase the incidence of genitourinary and musculoskeletal discomfort compared to the control group [15,16], and of symptoms related to fractures [17]. However, the potential impact on cardiovascular system has not been sufficiently elucidated, and patients with cardiovascular disease risk is difficult to identify. Cardiovascular events(CVEs) related to breast cancer include hypertension, ischemic cardiovascular disease, venous thrombosis, hypercholesterolaemia, arrhythmia, cardiac failure, peripheral arterial disease, embolism, myocardial infarction, atrial fibrillation [18][19][20][21].Data regarding the CVEs in AIs treated patients was reported in several RCTs, several of these reporting increased risks with AIs[7, 22-24]. However, recent study suggested that AIs may not increase the risk of the most fatal cardiovascular events [9].
In consideration of CVEs could greatly reduce the quality of life in breast cancer patients. Thus it is of great importance to fully understand the incidence of CVEs related to AIs treatment. Therefore, we conducted a systematic review and meta-analysis to evaluate the association of AI therapies with CVEs.

Search strategy
We searched PubMed (from 1967), Embase (from 1974), and the Cochrane Library electronic databases up to the end of December 2017. Keywords included in the search were 'Aromatase inhibitors', 'Breast cancer', 'Randomized controlled trials', 'Clinical trials' and 'Controlled clinical trials'. We also searched the website of clinical trial registration (ClinicalTrials.gov) to obtain information on the registered clinical trials. The search was restricted to clinical trials and articles published in English. This study is an meta-analysis and not involves subjects, ethical approval was not required.

Study selection and quality assessment
Two investigators HY and ZJH assessed the eligibility of the trials by the search independently, and trials were retrieved for further consideration if judged pertinent by one or two investigators. Any discrepancies were identified and resolved by consensus. Clinical trials that met the following criteria were included:

1.
Patient is diagnosed with breast cancer.

2.
Prospective phase II or III RCTs, including subjects assigned to treatment with AIs.

3.
Available data regarding events of CVEs.
The quality of all included trails was assessed using Jadad scale, and scores ranged from 0 to 5 with a high score indicating a high quality study[25]. Statement could see the Table S1. All data analyses were performed by using R software, version 3.2.3 (The R foundation for statistical computing, http://www.r-project.org).

Overall incidence of CVEs in AIs group
A total of 59503 subjects from the seventeen studies were available for incidence of CVEs analysis [37,38].CVEs were reported in all studies, and it ranged from 1.1% to 60.6% in AIs group. The highest incidence of CVEs was from a phase III trials of North America, which all subjects were confirmed early-stage breast cancer [45]. Based on data from each study, the calculated overall incidence of CVEs was 13.02 (95% CI: 8.15%-20.17%, Figure 2) according to the random effects model.
High-grade CVEs were reported in 10 out of 17 studies, and it ranged from 0.34% to 24.42%. The highest incidence of high-grade CVEs was from a phase III trials of Europe, which all subjects were confirmed early-stage breast cancer [40]. Based on data from each study, the calculated overall incidence of high-grade CVEs was 3.75 (95% CI: 1.66%-8.24%, Figure S3) according to the random effects model.

Odds ratios of CVEs
To evaluate the specific contribution of AIs to the development of CVEs in subjects excluding the influence of many confounding factors such as the history of course of disease, we performed the OR

Sub-group analysis
The incidence of CVEs might be different among different tumor stages, AIs or treatment regimens.
Thus, sub-group analysis was conducted according to under-lying malignancies, AIs and follow-up periods, although there is no significantly difference on the overall incidence of CVEs between patients with AIs treatment and that of non-AIs patients. There was no significant variation in the incidence of CVEs between different tumor stages ( Figure S1), even if the patients with advancedstage breast cancer has often been thought of as the possible high risk factor for CVEs. Then, the incidence differences among follow-up periods were investigated and there was no significant variation of the incidence of CVEs between long time and short time follow-up periods in patients received AIs ( Figure S2). Additionally, we found that the incidence of CVEs related to anastrozole and letrozole was slightly higher than controls, though statistically it makes no difference. However, it is worth mentioning that the incidence of CVEs related to exemestane was higher than that of controls (OR=1.1564, 95% CI: 1.0656-1.2549, Figure 4), which suggested that exemestane might not be considered as the alternative AI to the breast cancer patients from the perspective of CVEs.

Publication bias
No evidence of publication bias was found for the OR of CVEs in our meta-analysis by a funnel plot ( Figure 5) and contour-enhanced funnel plots ( Figure S4)[48].

Discussion
It is found that CV disease and breast cancer have several overlapping risk factors, such as obesity

Ethics approval and consent to participate
Not applicable.

Consent to publish
Not applicable.

Availability of data and materials
All data and materials used in this research are freely available in electronic databases (PubMed, EMBASE, Cochrane database, www.ClinicalTrials.gov) and references have been provided. The datasets used and analysed during the current study are available from the corresponding author on reasonable request.
All data generated or analysed during this study are included in this published article and its supplementary information files.

Competing interests
The authors declare that they have no competing interests.

Funding
This research was supported by the National Natural Science Foundation of China  Table 1 Baseline characteristics of trials included in the meta-analysis.   Figure 1 Flow chart demonstrating process of study selection.

Figure 2
Forest plot for meta-analysis of incidence of CVEs with patients assigned AIs.

Figure 3
Odds ratios of AIs-associated CVEs.

Figure 4
Sub-group analysis of the incidence of CVEs related to different AIs.    Table S1.doc