In present case-control study, we found significant positive association between FT4 and IS while significant negative associations between FT3 and FT3/FT4 ratio and IS. Moreover, mediation analyses suggested that GA might play a mediating role in the associations of FT3 and FT3/FT4 ratio and IS.
In the past few years, increasing consideration has been given to the relationship between the association between levels of TH within the reference range and stroke. One prospective study conducted by Chaker L et al. [17] showed that the HR (95%CI) for all-cause stroke and fatal stroke per-SD higher FT4 was 1.08 (95% CI: 0.99, 1.15) and 1.10 (95% CI: 1.04, 1.19), respectively. And one latest retrospective cohort study[18] based on data from the National Health and Nutrition Examination Survey (NHANES) 2007–2012 also found that FT4 was significantly and positively correlated with the risk of stroke (OR:1.20; 95% CI: 1.08, 1.33). Consistent with the above studies, our results showed that higher level of FT4 in the normal range is associated with increased risk of IS. In addition to FT4, we also found that there was a significant negative correlation between FT3 and IS. Current epidemiological evidence on the association between FT3 and IS is still scarce, mainly focusing on the association between FT3 level and prognosis of stroke patients[19, 20]. In the only retrospective cohort study based on data of NHANES 2007–2012[18], although a significant negative association was found between FT3 and stroke without adjusting for any confounders, the significance disappeared after adjusting the full model. Moreover, the FT3/FT4 ratio was significantly negatively correlated with IS in our current study. Although a growing body of literature have explored the relationship between the FT3/FT4 ratio and adverse health outcomes such as diabetes[21] and all-cause death[18], there are few studies on the relationship between the FT3/FT4 ratio and stroke. So far, only Lang X et al.[18] have explored the association between FT3/FT4 ratio and incident stroke risk and reported that each 1-SD increment in the FT3/FT4 ratio decreased stroke risk by 17% (HR: 0.83; 95% CI: 0.70–0.99), which was consistent with our findings. FT3/FT4 ratio is surrogate marker of dynamic activity of DIO2 [10]. Considering that DOX2 activity is associated with increased basal metabolism [22] and lower basal metabolism is associated with stroke risk factors including diabetes [21], it is reasonable that a higher FT3/FT4 ratio has a negative correlation with stroke risk. But it is worth noting that, previous cross-sectional studies found that in the population with normal TH level, the FT3/FT4 ratio was significantly positively correlated with the index of metabolic syndrome [12, 23], and the risk of pre-diabetes and diabetes[10, 24]. And the development of CVD will affect the normal negative feedback regulation of TH [4, 18] and reduce the conversion of FT4 to FT3, showing a decrease in the FT3/FT4 ratio. Therefore, considering our case-control design, more prospective studies are urgently needed to verify our findings. In addition, no significant association of TSH with IS was detected in our study. Previous studies have explored and found significant associations between subclinical thyroid dysfunction, defined as abnormal serum TSH and normal free TH, and the stroke risk [25] However, considering that the TSH levels of the subjects in our study are within the normal range, there is no significant association between TSH and IS.
Although the exact mechanism underlying the association between thyroid function and IS risk is not clear to a large extent, we initially found that GA may partially mediate the association between TH and IS risk. GA can be used to evaluate the glycemic balance over a period of approximately 2–3 weeks. On recent prospective study conducted by Mihara A et al. [26] reported a significant positive association between GA and incident stroke. And one earlier cross-sectional study based on the Japanese general population conducted by the same team also found that elevated GA levels were closely related to carotid intima-media thickness, which is an independent predictor of stroke risk [27]. In addition, Selvin E et al. also found a significant positive correlation between baseline GA and IS risk after adjusting for traditional cardiovascular disease risk factors [28]. Our findings corresponded the above studies, which showed a positive link between GA levels and ischemic stroke. Furthermore, one recent study conducted among Chinese reported that FT3 was independently negatively correlated with GA among euthyroid subjects [29]. In current study, we found that the level of GA decreased with increasing level of FT3 and FT3/FT4 ratio, while the GA levels was associated with IS significantly. GA is the product of the non-enzymatic glycosylation reaction between glucose and albumin (ALB) and the process leads to a reduction in the antioxidant activity of ALB [14]. In addition, the glycosylation of albumin will form advanced glycosylation end products, which are closely related to the risk of CVD, including stroke [30]. Our current study implies that the association between FT3 and FT3/FT4 ratio and stroke is partially mediated by GA level. High levels of FT3 or highly active DIO2 lead to glycosylation of albumin, that is, lower GA levels, which in turn reduces the risk of stroke. However, more studies are needed to further elucidate the precise mechanisms. Our study has several limitations. First, this is a case-control study, thus can be used to establish a correlation between thyroid indices and IS risk, but cannot establish causation. Second, BMI data are missing in this study and cannot be incorporated into the model for adjustment. Additionally, the multivariate model failed to adjust for factors such as educational level, physical activity, medication history, and specific conditions (such as iodine deficiency). Therefore, the possibility of residual confounders cannot be ruled out.