Atrial fibrillation and breast cancer—Vicious twins? A systematic review and meta-analysis

Background Epidemiological studies suggest a bidirectional association between atrial fibrillation and breast cancer. This study aimed to conduct a meta-analysis to elucidate the prevalence of atrial fibrillation among breast cancer patients, and the bidirectional association between atrial fibrillation and breast cancer. Methods PubMed, the Cochrane Library, and Embase were searched to identify studies reporting the prevalence, incidence, and bidirectional association between atrial fibrillation and breast cancer. The study was registered with PROSPERO (CRD42022313251). Levels of evidence and recommendations were assessed by the Grading of Recommendations Assessment, Development and Evaluation (GRADE). Results Twenty-three studies (17 retrospective cohort studies, 5 case-control studies and 1 cross-sectional study) involving 8,537,551 participants were included. Among patients with breast cancer, the prevalence of atrial fibrillation was 3% (11 studies; 95% CI: 0.6 to 7.1%) and the incidence was 2.7% (6 studies; 95% CI: 1.1 to 4.9%). Breast cancer was associated with increased risk of atrial fibrillation (5 studies; hazard ratio [HR]: 1.43, 95% CI: 1.12 to 1.82, I2 = 98%). Atrial fibrillation was also significantly associated elevated risk of breast cancer (5 studies HR: 1.18, 95% CI: 1.14 to 1.22, I2 = 0%). Grade assessment shown low certainty of the evidence for the risk of atrial fibrillation and moderate certainty of the evidence for the risk of breast cancer. Conclusion Atrial fibrillation is not uncommon in patients with breast cancer and vice versa. There is a bidirectional association between atrial fibrillation (low certainty) and breast cancer (moderate certainty).


Introduction
Breast cancer is the most frequent cancer amongst women worldwide, accounting for about 24.5 percentage of all cancers (1). Thanks to advances in treatment, the 5-year survival rate for breast cancer patients has significantly increased to 89% (2). However, cardiovascular disease is increasingly becoming a barrier to optimal outcomes for breast cancer patients (3).
Atrial fibrillation, the most common arrhythmia, is an important cause of embolism and death from cardiovascular diseases, with a global prevalence of approximately 2-4%. Age is a strong risk factor for the incidence of atrial fibrillation. The prevalence of atrial fibrillation tends to increase with the increase in human life expectancy (4,5). One study showed that malignant neoplasms (23.1%) were the most common cause of death in patients with atrial fibrillation (6).
Recently, many epidemiological studies have shown an increased risk of new-onset atrial fibrillation in women with breast cancer (7-15), while other studies have shown an increased risk of breast cancer in women with atrial fibrillation (13,(16)(17)(18)(19)(20). It is important to note that these studies have not yielded consistent results. The incidence and prevalence of atrial fibrillation in breast cancer are unknown, and it is unclear whether there is a bidirectional association between atrial fibrillation and breast cancer. Thus, we conducted a systematic review and meta-analysis to elucidate the prevalence and incidence of atrial fibrillation among breast cancer patients and the association between atrial fibrillation and breast cancer.

Protocol registration
The study was registered in the PROSPERO (International prospective register of systematic reviews. https://www.crd.york. ac.uk/PROSPERO/ -registration number-CRD42022313251).
This meta-analysis was performed according to the PRISMA 2021 guideline for systematic review and meta-analysis (21) (Supplementary Table S1).

Search strategy
We systematically searched the PubMed, EMBASE, Cochrane Library online database up to February 25, 2022 with no language restriction. The gray literatures, published in American College of Cardiology (ACC)/American Heart Association (AHA), European Society of Cardiology (ESC) were also reviewed. The search MeSH items and keywords were as follows: ("Breast Neoplasms" [Mesh] OR "breast cancer" OR "breast cancers" OR "breast tumors" OR "breast tumors" OR "mammary Cancer" OR "mammary cancers" OR "mammary gland cancers" OR "mammary gland cancer") AND ("Atrial Fibrillation" [Mesh] OR "atrial fibrillations" OR "auricular fibrillation") OR ("Mortality" [Mesh] OR "Mortalities" OR "Case Fatality Rate" OR "Case Fatality Rates" OR "Death Rate" OR "Death Rates"). The detailed search strategy is shown in Supplementary Table S2.

Selection criteria
Eligible studies were as follows: (1) clinical trials or observational studies (cohort, case-control, or case-cohort or nest-case control) reported the prevalence or the incidence of atrial fibrillation in patients with breast cancer. (2) studies assessed the association between atrial fibrillation and breast cancer expressing with Relative Risk (RR)/HR and corresponding 95% CI or that can be calculated. For the incidence of atrial fibrillation or cancer, we only included cohort studies or the case cohort designed data.
(3) studies assessed the association between atrial fibrillation and mortality in patients with breast cancers. Two authors (X-L and X.D-Y) independently reviewed the articles for eligibility. Any discrepancies were discussed until consensus.
Accordingly, studies that met the following criteria were excluded: (1) Studies that specifically looked at the influence of breast cancer-related treatments (e.g., radiotherapy chemotherapy and adjuvant therapy) on atrial fibrillation, as anticancer treatment was linked to atrial fibrillation incident by the previous studies (22-31). (2) Reviews, meta-analyses, practice guidelines, patents, cases, replies, comments, or editorial; (3) after contacting the author, the data needed for the article still cannot be obtained.

Data extraction and quality assessment
The information was extracted from each study, including first author, publication year, country, mean age, study design, data source, follow-up, prevalence or incidence of atrial fibrillation or breast cancer, atrial fibrillation diagnosis, breast cancer diagnosis, sample size, any subgroup, effect size, and adjustments. Three author (X.D-Y and X-L Z.Q-T) extracted data. For studies that reported the prevalence of atrial fibrillation or breast cancer, the Joanna Briggs Institute (JBI) critical appraisal checklist was used to assess the study quality. The Newcastle-Ottawa quality scale (NOS) was used to quantify the quality of studies reporting association between breast cancer and atrial fibrillation. A score of JBI > 16 or NOS > 6 was regarded as acceptable quality.

Statistical analysis
We used a random-effect model to summarize the estimates and 95% confidence intervals (CIs). For incidence of atrial fibrillation or breast cancer, the exact binomial (Clopper-Pearson) method was used to calculate 95%CI (32). Estimates were standardized using the Freeman-Tukey double arcsine transformation (33). For the effect of atrial fibrillation on mortality in breast cancer patients, estimates from each study were consolidated by the method of DerSimonian and Laird (34), which designated the weight of each study based on its variance. If a study reported results separately for subgroups (e.g., age), we combined the subgroup-specific estimates using a random-effect so that each study was represented only once in the forest plot. If one study reported one more effect size, we used the most informative one. In the association outcomes, considering the difference in study design, the evidence of cohort and case-control studies was pooled separately. In cohort studies, we treat standardized incidence ratios (SIRs) as equivalent to RR.
Heterogeneity was evaluated using the Higgins I-squared (I 2 ) statistic (25,50, and 75% represent low, moderate, and high heterogeneity, respectively) (35). Publication bias was addressed by Funel plot, and Egger's and Begg's test, with the results considered to be significant when P < 0.10 (36). For the purpose of appraising the robustness and reliability of the primary study outcomes, we also carried out sensitivity analyses by omitting each study in turn. Subgroup analysis was defined as mean age (<65 vs. ≥65 years), region (America, Europe, and Asia), sample size (<5,000 vs. ≥5,000), case (<50 vs. ≥50) if enough studies were available (N > 10) considering the statistical power according to the guidelines. For the association of atrial fibrillation with breast cancer, an additional subgroup was stratified by hormone therapy. The statistical analysis was performed by the RevMan software, version 5.4.1 (The Cochrane Collaboration, Nordic Cochrane Center Copenhagen, Denmark) and Stata software, Version 16.0 (Stata Corp LP, College Station, Texas, USA). P < 0.05 double-sided was considered statistically significant. We evaluated the quality of evidence for each outcome using the GRADE method (37,38). Two authors separately assessed the quality of evidence for each outcome. We used GRADEpro GDT to provide evidence profile tables. We present the results of the outcomes as described in the outcome metric type section, footnotes are used to indicate reasons for degradation or escalation of the quality of evidence.
In the subgroup analysis, the prevalence of atrial fibrillation (3%) was highest among breast cancer patients in the Americas (5%), followed by Asia (1.7%), and Europe (0.9%) (P for subgroup difference = 0.028) ( Figure 3B). Older (age ≥ 65) patients with breast cancer (8.6%) showed a higher prevalence of atrial fibrillation than younger (age < 65) patients with breast cancer (1.9%) ( Figure 3A). Moreover, the prevalence of atrial fibrillation was higher among breast cancer patients (3.5%) in studies with larger sample sizes (≥5,000) than among breast cancer patients (2.6%) in studies with smaller sample sizes (<5,000) ( Figure 3C). The prevalence of atrial fibrillation was higher among breast cancer patients (3.7%) in studies with a     greater number of atrial fibrillation cases (≥50) than among breast cancer patients (1.9%) in studies with a smaller number of atrial fibrillation cases (<50) ( Figure 3D). However, there were no significant differences regarding age (P = 0.15), sample size (P = 0.76) or number of atrial fibrillation cases (P = 0.47) ( Table 2).

Bidirectional association between breast cancer and atrial fibrillation
A total of five articles (7, 10,11,13,14) reported the risk of newonset atrial fibrillation among breast cancer patients. As shown in Figure 4A, the risk of atrial fibrillation was significantly increased in breast cancer patients compared with non-breast cancer patients in pooled cohort studies (HR: 1.43, 95% CI: 1.12 to 1.82, I 2 = 98%).
According to the predefined criteria, subgroup analysis was not performed due to the limited number of studies.

Publication bias and sensitivity analysis
As shown in Supplementary Figure S1, funnel plots, Egger's test (both P > 0.1), and Begg's test (both P > 0.1) showed no significant publication bias in risk of atrial fibrillation, risk of breast cancer and incidence of atrial fibrillation, even though funnel plots are not recommended when the number of articles included is limited (N < 10). However, funnel plots, Egger's (p < 0.05), and Begg's tests (p < 0.05) for the prevalence of atrial fibrillation suggest publication bias. Sensitivity analysis by omitting each study indicated that our results were stable and reliable, with a range from 1.31 (95% CI: 1.01 to 1.69) to 1.56 (95% CI: 1.32 to 1.83) for the risk of atrial fibrillation and 1.16 (95% CI: 1.12 to 1.21) to 1.21 (95% CI: 1.15 to 1.29) for the risk of breast cancer (Supplementary Figure S2). For the risk of atrial fibrillation, the results remained stable when we used the fixed effects analysis model (HR: 1.28, 95% CI: 1.25 to 1.31) (Supplementary Figure S3).

Quality assessment
The GRADE tool was used to evaluate the quality of evidence of cohort studies. All the included studies were observational cohort studies, so the initial level of evidence was moderate (37). As there was greater heterogeneity in the risk of atrial fibrillation (I 2 = 98%), downgrades were given. Ultimately, the risk of atrial fibrillation was evaluated as low certainty, and the risk of breast cancer was evaluated as moderate by the GRADE tool (Supplementary Table S6).

Major findings
Our results with a sample size of 8,537,551 individuals showed that the prevalence of atrial fibrillation among breast cancer patients was 3% and that incidence was 2.7%. Patients with breast cancer had an increased risk of atrial fibrillation of 43%, and patients with atrial fibrillation had an increased risk of breast cancer of 18% in pooled cohorts, which first showed a bidirectional association between atrial fibrillation and breast cancer (Graphical Abstract).
Age is a vital risk factor for both atrial fibrillation and breast cancer, and the prevalence of both increases with aging (4,48). A national retrospective cohort study showed that breast cancer was associated with an increased risk of atrial fibrillation from 6 months to 3 years of follow-up, both in age <60 (HR: 1.80, 95% CI: 1.38-2.35) and age >60 groups (HR: 1.14, 95% CI: 1.05 to 1.25) after adjustments (10). However, a large case-control study with a sample size of 122,100 showed a reduced risk of atrial fibrillation in the age <65 group (OR: 0.70, 95% CI: 0.69 to 0.71) and 65-80 group (OR: 0.91, 95% CI: 0.90 to 0.92) but an increased risk of atrial fibrillation in the age >80 group (OR: 1.01, 95% CI: 1.00 to 1.02) (15). Our results showed that the prevalence of atrial fibrillation among elderly individuals (8.6%) was higher than that in those aged <65 years (1.9%). Studies that included atrial fibrillation in patients with breast cancer adjusted for age, so our results suggest that the increased risk of atrial fibrillation in breast cancer patients may be independent of age.
In addition, the treatment of breast cancer patients, such as surgery, chemotherapy and radiation, can also contribute to atrial fibrillation (49,50). The most commonly used chemotherapy drugs for breast cancer, such as alkylates, anthracyclines, paclitaxel and trastuzumab have been found to be cardiotoxic and can induce atrial fibrillation in breast cancer patients (51, 52). Moreover, pain medications used by patients with advanced cancer, including nonsteroidal anti-inflammatory drugs and opioids, are associated with an increased risk of atrial fibrillation (53,54). A case-control study involving patients first diagnosed with breast cancer who had not been treated showed a significantly increased risk of atrial fibrillation in these patients compared with non-breast cancer patients (12). To some extent, the increased risk of atrial fibrillation in breast cancer patients may be independent of cancer treatment.
Similarly, drug treatment for patients with atrial fibrillation may also play a role in promoting breast cancer, such as amiodarone, which may be associated with an increased risk of cancer (55). In addition, a retrospective cohort study of 2,116,029 women showed a significantly increased risk of breast cancer among patients currently on digoxin compared with a control group (RR: 1.39, 95% CI: 1.32 to 1.46) (56). It is speculated that this may be related to the estrogen-like effect of digoxin. However, the . /fcvm. . relationship between cardiac glycosides and breast cancer remains unclear. Recent studies have shown that cardiac glycosides exert antitumor effects through different mechanisms (57, 58). A recent study in the Danish Breast Cancer Cohort confirmed that among 49,312 patients enrolled from 1995 to 2008, the overall risk of breast cancer recurrence did not change significantly in the digoxin group compared with the non-digoxin group (HR: 1.13, 95% CI: 0.88 to 1.46) (59). Although antiarrhythmic drugs may contribute to atrial fibrillation due to estrogenic effects, the risk of breast cancer in patients with atrial fibrillation remains significantly higher after adjusting for hormone replacement therapy (20). Furthermore, our subgroup analysis also showed that the association between atrial fibrillation and breast cancer persists regardless of hormone therapy adjustment (Supplementary Figure S3). Two articles (1 retrospective cohort study and 1 case-control study) (11,41) reported the association between atrial fibrillation and all-cause mortality in patients with breast cancer. One retrospective cohort study of 85,423 breast cancer patients showed that new-onset atrial fibrillation was associated with increased all-cause mortality in patients with breast cancer after 1 year of follow-up (HR: 2.15, 95% CI: 1.32 to 3.48) (11). In addition, one case-control study of 2,478,598 patients with metastatic cancer showed that atrial fibrillation was associated with increased allcause mortality in patients with breast cancer (OR: 1.43, 95% CI: 1.15 to 1.78) (41). Previous study showed an increase in all-cause mortality in patients with atrial fibrillation (11,41). Guha et al. showed that this increased all-cause mortality was mainly due to increased cardiovascular mortality (11). The most common cause of cardiovascular death is heart failure, followed by embolism and stroke. It is noteworthy that the use of anticoagulants reduced all-cause mortality, either because of fewer embolic events or because anticoagulants hindered breast cancer development (60, 61).

Underlying mechanism
Our study showed a 43% increased risk of atrial fibrillation in patients with breast cancer compared to those without breast cancer. There is accumulating evidence that breast cancer is closely related to the incidence of atrial fibrillation. Several mechanisms have been proposed to explain this association. First, previous studies have confirmed that the increased risk of atrial fibrillation in cancer patients may be related to the systemic inflammatory state caused by cancer. Inflammation has been considered a vital mechanism involved in the initiation and maintenance of atrial fibrillation (62). Inflammatory markers are significantly increased in patients with breast cancer, such as C-reactive protein, tumor necrosis factor-α, interleukin-2, interleukin-6, and interleukin-8 (63). In addition, breast cancer and atrial fibrillation develop through the same inflammatory pathways, such as the NOD-like receptor protein 3 (NLRP-3) inflammasome (62,64,65). Second, some studies have also suggested that autonomic dysfunction in cancer patients may contribute to an increased risk of atrial fibrillation. Patients with breast cancer may face long-term physical pain and mental stress, which may increase sympathetic nerve activity and lead to atrial fibrillation (66,67). However, there is .
/fcvm. . reason to suspect that the medical exposure of cancer patients is far greater than that of non-cancer patients, which may lead to detection bias. However, in a recent study (7), after adjusting for medical exposure, patients with cancer still had a significantly higher risk of atrial fibrillation than those without cancer, suggesting that this detection bias does not hold.
Our study showed an 18% increased risk of breast cancer in patients with atrial fibrillation. The current study can only show the association between breast cancer and atrial fibrillation but cannot prove a causal relationship. It is speculated that the bidirectional relationship between the two diseases may be related to common risk factors, especially the procoagulant state, as thrombin is a drivers of both atrial fibrillation and breast cancer (68). In addition, apoptosis occurs in the development of atrial fibrillation (69), which may lead to the imbalance between proapoptotic and antiapoptotic factors, thus reducing the apoptosis of cancer cells and promoting the germinal development of cancer (70). However, several studies have shown a significant increase in the risk of cancer after 90 days of atrial fibrillation and a significant decrease after 90 days, although still associated with cancer, suggesting that in addition to the detection bias caused by the medical exposure of atrial fibrillation patients, many atrial fibrillation patients may already have cancer at the time of diagnosis (16,18,19). In other words, common risk factors for both breast cancer and atrial fibrillation only may be early signs of occult cancer (16,18,19). Finally, although most studies have adjusted for confounding factors, the association may still be the result of residual confounding. Therefore, the mechanism of the association between atrial fibrillation and breast cancer is very complex, and there is no conclusion yet. Further studies are needed to clarify their association.

Comparison with previous studies
Our research is somewhat pioneering. Two previous metaanalyses reported an increased risk of cancer diagnosis within 90 days in patients with newly developed atrial fibrillation, but these studies focused on all malignancies rather than specific tumor subtypes (71,72). A recent meta-analysis reported the prevalence of atrial fibrillation in breast cancer patients across different types of cancer therapy (73). We first assessed the bidirectional association of atrial fibrillation and breast cancer patients.

Clinical implications
The overall prevalence (3%) of atrial fibrillation among breast cancer patients was similar to that of the general population; however, subgroup analysis showed an increased prevalence of atrial fibrillation among breast cancer patients >65 years of age (8.6%). Given the overall low prevalence of atrial fibrillation among breast cancer patients, screening for atrial fibrillation in breast cancer patients and screening for breast cancer in patients with atrial fibrillation are not currently recommended (4), although the prevalence of atrial fibrillation among breast cancer patients is higher than that among non-breast cancer patients. However, our subgroup analysis showed that the prevalence of atrial fibrillation among patients with breast cancer >65 years of age was four times higher than that among younger breast cancer patients and twice as high as that in the general population >65 years of age (4.4%) (4). Therefore, oncologist must be vigilant about the increased risk of atrial fibrillation in elderly patients with breast cancer.

Study limitations
Our systematic review and meta-analysis had several limitations. First, all the studies included were observational and could not prove causality. However, a bidirectional relationship between breast cancer and atrial fibrillation was found in cohort rather than case-control studies. Most cohort studies were retrospective cohort studies; therefore, reverse causality may still exist, and more prospective studies are needed. Second, most of the studies we included featured female breast cancer patients, and a considerable number of studies included very few male breast cancer patients. Since we could not contact the researchers to obtain the original data, we included these individual male patients. Third, atrial fibrillation is known to contribute significantly to the incidence of stroke. However, none of the studies we included reported an outcome event for stroke; more studies are needed to clarify the outcomes, especially stroke, in patients with atrial fibrillation and breast cancer.

Conclusions
Atrial fibrillation is not uncommon in patients with breast cancer, and there is an increased risk of atrial fibrillation in breast cancer patients and vice versa. In addition, breast cancer patients with atrial fibrillation had increased all-cause mortality. GRADE assessment indicated low certainty for the risk of atrial fibrillation and moderate certainty for the risk of breast cancer. More studies are needed to determine whether screening for breast cancer in patients with atrial fibrillation is necessary.

Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Author contributions
RQY contributed to the study concept and design and revised the draft. XDY, QWH, and XL performed the search strategy and contributed to database research, acquisition of data, and statistical analyses. All the authors participated in data analysis, reviewed, and approved the final manuscript.