Major findings
Based on the longitudinal cohort of ARIC, we found a graded risk of AF in patients without known CVDs with TyG levels after a median follow-up of 24.26 years. However, the full adjusted incidence of AF showed a U-shaped trend with a TyG level, with a nadir at a TyG level of approximately 8.80–9.20. Cox analysis and exposure-effect analysis further confirmed the U-shaped association. In a further sex-specific analysis, a U-shaped association between the TyG index and incident AF still existed in females but not in males. Taken together, this study suggested a U-shaped association between the TyG index and incident AF in people without known cardiovascular disease. Sex may be a modifier in the association between the TyG index and AF incidence.
HOMA-IR was previously considered the gold standard for IR. Recently, as an expected new estimating model for IR, the TyG index has been extensively investigated and approved with the strength of convenience compared to HOMA-IR11. Guerrero-Romero et al. found high sensitivity (96.5%) and specificity (85.0%) of the TyG index for recognizing IR(51). The significant association between the TyG index and an increased risk of cardiovascular disease and mortality has been extensively explored(28, 34, 52–55), regardless of diabetes status. However, no studies have explored the association of the TyG index with atrial fibrillation in the general population.
Hence, it is not surprising that the higher TyG index is associated with an increased risk of AF in the present study. A positive association between the TyG index and AF risk has been observed in certain patients with established cardiovascular diseases. In both cohorts, including those undergoing septal myectomy and percutaneous coronary intervention due to hypertrophic obstructive cardiomyopathy and myocardial infarction, the TyG index showed moderate predictive ability for postoperative AF incidence(56, 57). Among patients who underwent septal myectomy for hypertrophic obstructive cardiomyopathy, the TyG index showed moderate predictive ability for postoperative AF (AUC = 0.723)(56). In patients who received percutaneous coronary intervention for ST-segment elevation myocardial infarction, the TyG index also showed moderate predictive value for new-onset AF (AUC = 0.758)(57). However, we did not assess the predictive ability of the addition of TyG in the existing risk scores for AF incidence, such as the Framingham Heart Study (FHS) score, the ARIC score, and the Cohorts for Heart and Aging Research in Genomic Epidemiology Atrial Fibrillation (CHARGE-AF) score, Thrombolysis in Myocardial Infarction 48 (TIMI-AF) and C2HEST(58–62), which may be further studied.
We showed that the crude rate of AF incidence increased with the TyG index. However, after full adjustment, we found a U-shaped association of the TyG index with incident AF, which suggests that both lower and higher levels of the TyG index are risk factors for AF. Similarly, a longitudinal cohort study from South Korea showed a nonlinear association between IR and AF incidents in the nondiabetic population after a median follow-up of 12.3 years, with atrial fibrillation development increasing at HOMA-IR levels between approximately 1-2.5 and then plateauing afterward(63). Several potential mechanisms may be responsible. A strong association between a high TyG index and an increased risk of developing AF may be due to IR. IR in the liver is caused by impaired insulin metabolism influencing glucose metabolism and enhancing insulin-mediated lipogenesis to cause hyperglycemia(64). Adipo-IR has been associated with β cell dysfunction, which begins in individuals with normal glucose tolerance (65). Hyperglycemia and β cell impairment are both related to cardiomyocyte metabolism and cardiac function. All the above pathological changes may contribute to the incidence of atrial fibrillation. On the other hand, we also showed that low TyG index values were associated with an increased risk of AF events, which may be explained with caution since previous studies showed a linear relationship between TyG and cardiovascular disease. Further results are needed to confirm these results.
Sex is a well-known modifier for the incidence and development of AF. A previous meta-analysis showed that a higher TyG index was associated with coronary artery disease/CVD, even after removing sex-unadjusted studies (HR = 1.59, 95% CI: 1.21–2.09)(35). We showed a different shape of dose‒response association trend between the TyG index and AF incidence in different sexes. Similar to our study, this study showed a sex difference (P = 0.045) in the relationship between a high TyG index and subclinical atherosclerosis in nondiabetic patients(66). A significantly higher prevalence of subclinical atherosclerosis in the high TyG index group than in the low TyG index group (odds ratio [OR] = 1.510; 95% CI: 1.010–2.257) was observed in nondiabetic females but not in nondiabetic males (OR = 0.827; 95% CI: 1.556–1.231)(66). Even when cardiovascular risk factors are similar, the rate of change in BMI and worsening of lipid levels is greater in women than in men(67). Differences in the location of fat storage in men and women may play a role in the development of IR and diabetes(68, 69). Females are more likely to store fat subcutaneously than males(70). Because visceral fat is strongly associated with insulin resistance, women may need to gain more weight and experience more significant worsening of associated metabolic risk factors to achieve the same level of visceral fat storage(71). This implies that sex differences in metabolic risk factors result from elevated blood glucose levels and diabetes(72). Additionally, in rodent models, endogenous estrogens may play a role in higher insulin sensitivity in women(73). In a large clinical study, menopausal hormone therapy in postmenopausal women improved insulin sensitivity through estrogen receptors in the liver, muscle, and adipose tissue(74, 75). The protective effect of estrogen is lost when women are exposed to risk factors such as hypertension and hyperlipidemia(76, 77), which may cause a sex difference.
Diabetes is another vital factor. Several previous studies have suggested that increased IR is associated with an increased risk of cardiovascular events in patients without diabetes(78, 79). In our sensitivity analysis in nondiabetic individuals, the shape of the TyG index and AF did not significantly change. Due to the limited sample size of T2DM patients in the present cohort (8.9%), the relationship between the TyG index and AF was not assessed. Further studies are needed to clarify the role of the TyG index on AF in the diabetic population.
Strengths and limitations:
Our study has several strengths. To our knowledge, we are the first to investigate the association between the TyG index and the incidence of AF in individuals without known cardiovascular disease. Second, our data were from a well-designed large cohort study that included a large sample size and diverse populations with adequate follow-up. Third, we further investigated the sex-specific association of the TyG index with AF incidence and found that it may be a modifier. However, our study also had limitations. First, our analysis is based on an observational cohort study. Several confounding factors, such as nutritional status, may have influenced our results. In our analysis, the TyG index was calculated based on the blood sample at baseline. However, considering the median follow-up of 24.26 years, a variety of TyG index values during follow-up may provide better insight into the risk of incident AF. Second, the population in our study was American, and the generality in other populations needs to be further verified.