In this study, we performed a two-sample MR analysis using a training dataset and a validation dataset to systematically assess the potential causal effect of MDD on the AF risk, and our main finding showed no clear evidence supporting the causal role of genetically predicted MDD on the risk of AF.
Accumulated studies have indicated that depression was significantly associated with the AF development and progress and was expected to serve as a potential risk factor for AF. A nationwide population-based study using the Korean National Health Insurance Database with a total of 6,576,582 young adults suggested that compared patients without depression, the depressive patients showed a significantly high risk of AF (adjusted hazard ratio: 1.58; 95%CI: 1.4–1.75, P < 0.05) using full-adjusted model with a median of 7.6-years follow-up[11]. Similarly, Garg et al.[8] performed a multiethnic and long-term cohort study to determine the relationship between depression (defined with a 20-item Center for Epidemiologic Studies Depression Scale score ≥ 16 or antidepressant use) and the AF incidence, in which a total of 6,644 middle-aged and older-aged adults who were free form AF at baseline were included. They also suggested that the depressive patients with the Center for Epidemiologic Studies Depression Scale ≥ 16 and antidepressant use, respectively, was significantly associated with an increasing as high as 34% and 36% AF risk. Interestingly, Feng et al.[9] suggested that compared with no depression symptoms, mild to moderate depression symptoms were significantly linked with the increased AF risk (adjusted hazard ratio: 1.50; 95%CI: 1.20–1.80, P < 0.05), while severe depression symptoms were not (adjusted hazard ratio: 0.90; 95%CI: 0.60–1.30, P > 0.05). Meanwhile, a recent meta-analysis with nine studies performed by Fu et al.[10] also showed a consistently positive result between depression and AF risk.
Whereas, a few prospective studies showed different results. A lager Women’s Health Study with over 30,000 participants conducted by Whang et al.[21] indicated that both depressive symptoms and antidepressant use failed to be related to the incident of new-onset AF with more than ten-year follow-up. Whang et al.[21] emphasized that the negative results might be responsible to the single gender (only women) and race (predominantly white population).
Moreover, the preliminary MR studies suggested that no causal association might be existed between depression and AF. Li et al.[13] performed a MR study to evaluate the bi-directional causal association between depression and cardiovascular diseases, and the depression was defined without specific phenotypes using general methods, including self-reported diagnosis combined with conventional identification methods. The results showed that showed null causal association between depression and risk of AF with IVW analysis. Meanwhile, Lu et al.[12] broadly defined the depression (including multiple depression phenotypes) to explore the causal relationship of genetic liability to depression with AF. Consistent with Li et al.[13] results, the IVW estimate indicated no association of genetically determined depression with AF (odds ratio: 1.00; 95%CI: 0.94–1.06; P = 0.95). Considering the various depression phenotypes in the GWAS database, the broadly defined depression with multiple depression phenotypes included in the MR analysis might be responsible for the negative casual results.
In this study, we aimed to explore the causality from MDD (a serious depression phenotype) on AF based on the MR study, which facilitated to reveal causality apart from bias with superior study design. We first performed MR analysis using a training data (dataset: ieu-a-1187) with multiple data processing, including removing SNPs related with AF risk factors, removing one outlier with MR-PRESSO, and tighten instrument P value threshold. The final MR analysis suggested that MDD had no causality on AF incidence. Importantly, a consistent result was presented in MR analysis using a validation data (dataset: ebi-a-GCST005903) for MDD on AF risk.
Despite of no causality between MDD with AF in our study, the possibility of the MDD effected on the development and progression of AF could not be excluded. Multiple potential mechanisms between MDD and AF have been widely explored. Inflammation has been reported to be intensified by the depression, subsequently facilitates to increase the risk of AF incident[22]. Depression has been demonstrated to be positively associated with multiple inflammatory response, including lymphocyte proliferation, increase of acute-phase reactants (e.g., C-reactive protein), and enhancement of cytokine secretion[23]. The autonomous nervous system disorder has been proved to play a key role on the electrophysiological dysfunction of cardiomyocytes, including the increase of trigger activity and the shorted effective refractory period, ultimately leading to the initiation and progression of AF[24]. Reportedly, the catecholamine levels were significantly higher in depression patients compared with control patients, indicating that depression could lead to the abnormal increase of sympathetic nervous activity and the disorder of autonomous nervous system[25, 26]. Moreover, the dysregulated hypothalamus-pituitary-adrenal axis and renin-angiotensin-aldosterone system also belonged to the important pathophysiological change in depression[27–29]. Importantly, lifestyle-related changes and behavioral mechanisms (e.g., smoking, obesity, alcohol abuse, physical inactivity, and medication nonadherence) for depression patients could serve as the predominant risk factors for AF development[30–32].
Several strengths in our study should be addressed. First, the two-sample MR method applied in our study could achieve to explore the causality between exposure and outcome as effectively as randomized controlled traits, especially in the evaluation of the causality between two diseases. Second, our findings further supported the previous studies, suggesting that screening AF patients with genetically predicted MDD might be meaningless. More researches should focus on the relationship of environmentally determined MDD and AF, or MDD and the prognosis of AF.
Meanwhile, several limitations also should be highlighted. First, previous study reported that the race might be a potential determinant in the exploration of the causality between depression and AF[21]. All GWASs data in our study derived from European population. Although our results were consistent with the previous MR studies[12, 13], more researches included other races or multiple races should be performed to further demonstrate our results. Second, the AF types were multiple, and MDD might have a causal association with a phenotype of AF. A broader study including multiple AF phenotype subgroups could be considered in the future. Finally, the exposure and outcome datasets should include non-overlapping samples as far as possible to avoid potential bias in MR analysis. Therefore, the MDD dataset and AF dataset were acquired from different online databases, respectively. Whereas, we could not avoid the potential sample overlapping between MDD dataset and AF dataset due to the unavailable raw genetic data or individual-level data.