Atrial fibrillation (AF) is the most prevalent cardiac arrhythmia with an age-related increase in incidence (1). It is strongly associated with stroke, heart failure morbidity (2–4), and increased mortality (5, 6). Early identification of high AF risk population is crucial for avoiding the adverse consequences related to AF. There are several factors have been identified to be related to the etiology of AF, including genetic factors, environmental factors, and other complications (7). However, more than one-third of the potential risk factor contributing to AF is still unexplained. Therefore, further investigations for additional AF risk factors including smoking, cardio-metabolic factors, and several psychological factors (8–10) need to be conducted.

In recent years, the association between depression and AF has been confirmed in several basic and epidemiological studies (11, 12). Theoretically, depression is closely related to the dysregulation of the hypothalamus-pituitary-adrenal (HPA) axis and inflammation. The hyperactivation of HPA axis could induce the consistent release of cortisol, which is also a marker of cortisol resistance (13). Cortisol resistance stimulates immune activation, then the expression level of some proinflammatory cytokines, including IL-2, IL-6, IL-12, and TNF-a will be increased. These cytokines act on the brain, developing some symptoms of depression in susceptible populations (14, 15), and are capable of producing systemic inflammation. The influenced HPA axis determines that depression is often accompanied by hypertension, metabolic syndrome, and obesity (16–18), which aggravate oxidative stress and inflammation in the body. Systemic inflammation increases the risk of AF by changing the electrophysiology (i.e., affecting calcium flowing), conduction and structural substrates of the atrial (19, 20). Moreover, depression may alter the sympathetic and parasympathetic balance to induce the decreased arrhythmic threshold (1), which also influences the atrial conductivity and structural integrity (14). Smoking as an accepted risk factor of AF, is more common in people with depression, since high negative affect and low positive affect in depression might raise the patients' dependence on nicotine (21). Nicotine has been reported with the function of promoting atrial structural remodeling and interstitial fibrosis (22). Therefore, depression is a potential factor for inducing new-onset AF.

As for antidepressants which are divided into three categories including selective serotonin reuptake inhibitor (SSRI), tricyclic antidepressants and monoamine oxidase inhibitor, their cardiotoxicity has also been reported in previous studies (23, 24). Tricyclic antidepressant mainly affects intraventricular conduction, which is characterized by prolonged PR, QRS and QT intervals on the electrocardiogram (ECG) (25). SSRI tends to increase serum serotonin, which then induces the elevation of intracellular calcium level (26, 27). As a result, the amplitude of the pacemaker is increased and potentially influences the heart rhythm (27). However, on the other hand, the use of antidepressants is capable of ameliorating the imbalance conditions of proinflammatory cytokines in depression (28–30), which may reduce the risk of depression-induced AF to a certain extent.

Considering that whether depression and the use of antidepressants could increase the risk of AF remains a controversial issue in previous studies, herein, we performed a meta-analysis including all of the existing studies to detect the association between AF risk and depression.

This meta-analysis was based on the preferred reporting items for systematic review and meta-analysis (PRISMA) 2020 guidelines. Ethical approval was not provided since all data included in this study was from the published studies. The data, techniques, and materials that support the findings of this study will be available from the corresponding author according to reasonable requests.

As for bias risk assessment, the corresponding funnel plots for depression and antidepressant related studies were included in Supplementary Figures 4, 5. We also adopted Egger's and Begg's test to detect the presence of publication biases, which were presented in Supplementary Figures 6, 7. The results showed that the P-values of the two groups were ≥ 0.1, indicating that the bias risk of included studies in our meta-analysis was relatively lower.

Based on previous studies, we conducted this meta-analysis to evaluate the association between depression, antidepressants and the risk of AF. After pooling the data from 9 included observational studies, the primary outcomes of our investigation indicate that both depression and the use of antidepressants are capable of increasing the risk of incident AF.

Previous studies suggested that the occurrence of depression or depressive symptoms is closely related to some immune signaling, especially proinflammatory cytokines IL-2, IL-6, IL-12 and TNF-a (14, 40, 41). These cytokines are capable of inducing systemic inflammation, which is a potential risk factor of AF since the inflammatory cell infiltration has been observed in the atrial of AF patients (42, 43). According to the studies investigating the relationship between depression and AF, these two symptoms are in a comorbid state (44, 45). However, earlier meta-analysis pointed out that depression is related to the increased risks of sudden cardiac death, ventricular tachycardia/ventricular fibrillation, and AF recurrence (19), but the association between depression and incident AF was not considered to exist (10, 19). After reanalysis, we considered that some non-negligible limitations might affect their accuracy. As for the study from Shi et al., only two studies specific for depression and AF were included. In the study from Whang et al., the selection of the included population, such as only women or mainly white people may lead to the deviation of outcomes. In addition, people identified as depression through questionnaires were also included in this meta-analysis, which may induce a bias in the diagnosis of depression. Fu et al. only included 5 studies in their study, the majority of the study participants were from the US or Europe, the ethnic interference of study outcome also cannot be fully eliminated.

In addition, whether the use of antidepressants could impact AF risk was also a controversial issue in the previous studies (38, 46). Theoretically, tricyclic antidepressants affect cardiac conduction and cardiotoxicity (25). As for the use of SSRI antidepressants, patients' serum serotonin levels are elevated during medication. Serotonin promotes calcium overload, which may trigger focal atrial extrasystoles and increase the risk of AF (27, 47). Both of these two types of drugs are associated with prolonged QTc and increased risk of arrhythmias (48), but studies focusing on the risk of incident AF increased by monoamine oxidase inhibitors are relatively fewer. However, studies have suggested that treating with antidepressants may alter the imbalance conditions of inflammatory cytokines in depression (28–30). Our findings provide strong evidence for the view that antidepressants increase the risk of incident AF, which indicates that the prevention of AF in patients with depression deserves further research in the future.

After incorporating all of the latest relevant studies, the outcome of our analysis confirms the theoretical link between depression and incident AF. However, the substantial heterogeneity of our investigation still exists. After removing the study from Kim et al. (37), the outcome was not influenced and the heterogeneity is acceptable. The sources of heterogeneity are speculated as follows: (1) The race difference between this study and other included studies was non-negligible. Kim et al. obtained data from the Korean National Health Insurance Service database, whose study population was Asian. Except for this study, the investigation participants of other included studies in our meta-analysis were from the United States or Europe. The risk of incident AF has shown differences among different ethnic groups (49, 50). (2) This study was based on insurance claim data of ICD-10 codes for depression and AF, rather than incident AF diagnosis during follow-up. The different outcome definitions between this study and the rest studies may also lead to the existence of heterogeneity.

Previous investigations have pointed out the molecular mechanisms by which depression and the use of antidepressants might increase the risk of incident AF. These mechanisms laid a theoretical foundation for our research. After pooling all of the data from existing investigations, the results of our study quantitatively confirm that patients with depression and antidepressant users have an increased risk of new-onset AF by 15 and 16%, respectively. These data suggest that the cardiovascular health of patients with depression deserves special attention, and it is necessary to strengthen the cooperation between cardiologists and psychiatrists in the process of depression treatment. For future study, the effects of different types of antidepressants on incident AF deserves further exploration, which is helpful to formulate a more reasonable management plan for patients with depression.

Although our study has included as much data as possible and tried to avoid the influence of confounding factors, several potential limitations still exist. First of all, the substantial heterogeneity is relatively high in our study. This may be induced by analysis strategies and participant features. However, due to the limited data, subgroup analysis based on these factors cannot be carried out. Secondly, the evaluation criteria for depressive symptoms and incident AF were inconsistent, which might induce the existence of small deviations in the diagnosis of depression. Thirdly, most of the data were obtained from observational cohort studies. Although most of them were adjusted for multivariable confounding factors, the corresponding information of each included literature was not completely consistent, and the residual confounding factors cannot be completely excluded. Future studies can use the method of propensity score matching to make the baseline data of participants more comparable. Finally, the number of existing studies in the antidepressant group was relatively limited, which does not support the subgroup analysis of different types of antidepressants. Future studies can assess the effects of different types of antidepressants on incident AF after incorporating more eligible data.

Based on existing investigations, both depression and the use of antidepressants are related to the increased risk of incident AF. Further study is needed to conduct more subgroup analysis and confirm our findings.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.

This study was supported by the National Natural Science Foundation of China [No. 31960146] and Special fund for postgraduate innovation in Jiangxi Province (YC2012-B011).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.