Atrial fibrillation (AF) is the most frequent cardiac arrhythmia, and is associated with significant health and economic burden on patients worldwide (1). However, the preventive measurement for AF is not satisfied so date due to the unclear underling mechanism (2, 3).

As we all know, AF and coronary artery disease (CAD) often co-exist based on multiple similar risk factors. The prevalence of CAD in patients with AF is estimated to range from 17 to 46.5% (4, 5). Risk factors for CAD such as elderly age and diabetes mellitus (DM) usually result in atrial remodeling, which includes structural and electrical transformation that contribute to the incidence and development of AF (6–8). Therefore, it is critical to identify and control these risk factors for the prevention and treatment of AF in individuals with CAD.

Some correlative processes including inflammation, endothelial dysfunction and oxidative stress have been demonstrated to be associated with cardiovascular disease in basic and clinical research (9–11). Bilirubin, a tetrapyrrolic compound, can be oxidized by H2O2 during inflammatory to form several degradation products that may contribute to the pathogenesis of certain diseases (12). The direct bilirubin (DBIL) is considered an inflammatory marker and has been proved to play an important role in predicting the poor prognosis of patients with acute and chronic heart failure (13, 14). Moreover, DBIL has been identified as an independent risk factor for new-onset postoperative atrial fibrillation (POAF) after cardiac surgery (15). A retrospective cohort study conducted on thyrotoxic patients receiving radioactive iodine therapy revealed that DBIL was an important risk factor for predicting AF (16). However, Weiping Sun (17) found that DBIL was not an independent predictive factors for the occurrence of paroxysmal atrial fibrillation (pAF). In addition, a case–control study reported significantly lower levels of DBIL in AF patients compared to healthy individuals (18).

Previous studies on the association between serum DBIL and AF risk have yielded inconsistent results. The epidemiological evidence regarding the association of DBIL with risk of AF among patients with CAD remains unclear. Therefore, our study aims to determine whether an elevated serum DBIL concentration is associated with a higher risk of AF in a group of CAD patients.

In the study, there were two key findings. Firstly, the results suggested a significant relationship between elevated DBIL and AF risk in patients with CAD. Secondly, the correlation of elevated DBIL with the risk of AF in CAD patients was significantly higher among those older than 65 years and with a BMI <25 and DM.

Circulating bilirubin concentrations are primarily derived from hemoglobin of aged or damaged red blood cells. Corresponding, in this present study, it was observed that patients without liver disease in DBIL T3 or T2 had significantly higher concentrations of hemoglobin in contrast to those in T1 (p = 0.001). A review shows that increasing plasma bilirubin, acting as an antioxidant and metabolic hormone, can drive metabolic adaptions that improve deleterious outcomes of weight gain and obesity, such as inflammation, T2DM, and cardiovascular diseases (19). Previous studies have proposed a link between bilirubin and AF. However, a recent study conducted on Chinese elderly individuals with CAD and DM demonstrated that DBIL was not a predictor of AF (20). Additionally, a cross-sectional study conducted in patients with paroxysmal atrial fibrillation (pAF) suggested that patients with pAF had significantly higher levels of TBIL and indirect bilirubin, but not DBIL (21). Nevertheless, another study conducted in patients with nonvalvular chronic AF without any other cardiovascular disease revealed that the levels of total, direct and indirect serum bilirubin were significantly lower among the patients with AF compared to controls (18). The reports were inconsistent and controversial. Our study showed that patients with AF had significantly higher levels of both TBIL and DBIL compared to those without AF (p = 0.003 and p < 0.001, respectively). However, there was no significant difference in liver and kidney function between the two groups. Additionally, univariate and multivariate analysis demonstrated that DBIL was significantly associated with AF in patients with CAD. These findings suggest that DBIL could be a noteworthy marker of AF. Differences in study populations may cause inconsistent results. Moreover, the AF in our study was defined as paroxysmal AF, persistent AF, and permanent AF. Further stratified analysis will be used to evaluate these findings in future studies.

The underling mechanism that may explain the positive association between DBIL and AF have been suggested in several studies, but the precious pathways remain unclear. For decades, both TC and LDL-C have been considered causal factors for atherosclerotic cardiovascular disease. Traditional cardiovascular risk factors, including age, sex, DM, hypertension, for incident AF have been described in early studies (21, 22). Based on the above information, the association between dyslipidemia and atrial fibrillation (AF) has been evaluated. Current studies have indicated that elevated TC and LDL-C levels are inversely associated with incident AF (23, 24). The mechanism underpin the interesting inverse association may be related to the stabilizing effect of cholesterol on myocardial membranes (25–27). Consistently, logistic regression analysis in our study described the role of LDL-C in reduced AF risk. It is also worth noting that DBIL are negatively correlated with TC and LDL-C (28). Similar results were obtained in our study. And analysis revealed that DBIL was inversely associated with TC, TG, LDL-C, and ApoB. Bilirubin might induce hepatic fat utilization and reduce de novo lipogenesis, thus lowering lipid levels (29). This may partly explain why atrial fibrillation occurs more frequently in these patients.

Furthermore, as an independent risk factor, DBIL-induced AF may be associated with an inflammatory response. Heme oxygenase-1 (HO-1) and bilirubin IXalpha are predominantly accumulated in the perinuclear regions of foam cells (30). Additionally, foam cells are known to initiate the inflammatory process that may be linked to the pathogenesis of AF (31). Moreover, higher bilirubin levels may reflect increased foam cell presence, which plays a pivotal role in inflammation by regulating the production of inflammatory cytokines and matrix metalloproteinases (32). It was reported that DBIL levels were negatively associated with hsCRP levels (28). Further studies are needed to evaluate the potential role of DBIL in inflammation related to AF, due to the lack of C-reactive protein (CRP), interleukin-6 (IL-6), tumor necrosis factor alpha (TNFα), neutrophil to lymphocyte ratio (NLR), and other inflammatory markers.

The role of DBIL in oxidative stress, which may be associated with AF, is also reported. Defined doses of bilirubin could be considered as mitochondria targeted medication against inflammasome-related diseases (33). However, the molecular pathway(s) connecting reactive oxygen species (ROS) and AF are unknown. The Ca2+/calmodulin-dependent protein kinase II (CaMKII) has recently been suggested to be a ROS activated proarrhythmic signal contributing to AF (34). Oxidant stress levels evaluated by urinary 8-hydroxy-2′-deoxyguanosine (8-OHdG), a biomarker of oxidative DNA damage, and urinary biopyrrin, an oxidative metabolite of bilirubin have been proven to be significantly increased in AF patients (35).

Insulin resistance (IR) has been demonstrated to be an important risk factor for cardiovascular disease. A Prior study reported that bilirubin is associated with insulin resistance. Participants with lower bilirubin had a significantly higher risk of type 2 diabetes (36). Additionally, a cohort study performed by Puerto-Carranza et al. demonstrated that bilirubin can regulate insulin secretion and glucose uptake (37). Furthermore, there was a steep increase in the risk of incident AF associated with relatively low homeostatic model of insulin resistance (HOMA-IR) (38). To date, the TyG index has been shown to be a promising indicator of IR status (39, 40). Consistently, our study found significantly lower TyG levels in the higher DBIL group. These findings may partially elucidate the underling association of DBIL with increased AF risk. Our results could also be attributed to difference in ethnicity and sample size. Further studies are needed to investigate whether this mechanism palys an important role in atrial remodeling and the development of AF.

This study also showed that male gender was inversely correlated with the risk of AF, while older age was identified as an independent risk factor. These findings are consistent with previous studies (17, 28). Previous studies have demonstrated that old age is one of the most recognized risk factors for new onset-atrial fibrillation. Old age has been shown to induce structural and electrical atrial remodeling which increase the risk of developing atrial fibrillation. The prevalence of AF increases dramatically with advancing age (41, 42). Hence, the average age of individuals with AF is significantly higher than that in those without AF. After adjustment for confounding factors, logistic regression analysis revealed that higher age was a significant risk factor for AF. Furthermore, subgroup analysis revealed that the association of higher DBIL levels with the risk of AF was stronger in CAD patients older than 65 years. Prior studies have provided evidence that higher BMI causally increased the risk of AF (43, 44). However, the positive association was not observed in our study. Further stratified analysis suggested that the higher baseline DBIL level was associated with AF risk in individuals with BMI < 25, but not in those with BMI ≥ 25. Selection bias may results in this difference. In addition, Oka et al. (45) found that low BMI and female sex could predict reduced post-ablation left atrial emptying fraction, which was associated with recurrence of AF after ablation. Another possible explanation is that low BMI is a significant predictor of non-pulmonary vein (PV) foci, which plays an important role in the recurrence of AF (46). Prospective studies should be conducted to explore these results further. Prior studies have demonstrated that DM is one of the most common risk factors for the development of AF due to atrial structural and electrical remodeling. Consistent with previous studies (47, 48), a significant association between higher DBIL and AF risk was observed in DM patients. The results of this study suggest that we should pay attention to the DBIL levels of elderly CAD patients with lower BMI and DM in order to detect AF early.

There are several limitations in our study that should be noted. Firstly, this was a cross-sectional study conducted at a single center, and there was no patient follow-up. The results may be prone to selection or information bias. Secondly, the diagnosis of AF based on 12-lead electrocardiogram (ECG) or 24-h Holter may inevitably underestimate rates of AF. Thirdly, the biomarkers of inflammatory including high-sensitivity C-reactive protein and B-type natriuretic peptide, as well as left atrial diameter which need to be considered were not presented and studied. Finally, since AF can occur both paroxysmal and non-paroxysmal, the results should be assessed separately for each category in future studies.

In conclusion, this present study found a robust nonlinear correlation between serum DBIL levels and AF in patients with CAD. This association was more significant in DM patients and those aged ≥65 years, as well as BMI <25 patients. These findings may help identify CAD patients at high risk of AF by measuring the DBIL levels. In the future, we look forward to design a multicenter study with a larger sample size and long-term follow-up to explore the relationship between DBIL and AF. In addition, DBIL may be a potential target for therapeutic intervention to reduce the risk of AF confirmed by future studies. Further studies are necessary to investigate the role of DBIL in reducing AF risk.