A retrospective analysis was conducted that included 1884 patients, 1494 males and 390 females, who were admitted to Fuwai Hospital for ISR and successfully received coronary intervention between January 2017 and December 2018. ​According to the presence or absence of myocardial infarction, recurrent ISR and target vessel revascularization during the follow-up, the patients were divided into a MACE group (n = 215) and a non-MACE group (n = 1669). The median follow-up duration was 35 months.

Each individual signed an informed consent form. The Fuwai Hospital’s ethical committee approved this study.

The inclusion criteria were as follows: (1) individuals whose ISR was verified by coronary angiography (ISR was defined as the loss of $\ge$50% of the coronary lumen in the area of a previously stented lesion, and stenosis within 5 mm of the edge of the stent by coronary angiography was also defined as ISR); and (2) successful revascularization.

The exclusion criteria as follows: (1) patients who had bare metal stents implanted; (2) patients with stent implantation in bypass vessels; and (3) patients with severe heart, liver, renal insufficiency, abnormal thyroid function, anemia, infection or tumor.

(1) Age, sex, height, weight, previous medical history (hypertension, diabetes, hyperlipidemia, stroke history, chronic renal insufficiency history) and other data were collected. (2) Operation information, including ISR location, number and intervention treatment (stent reimplantation or drug-coated balloon use) were collected. (3) Fasting venous blood was collected from all patients in the morning before operation, and tests for routine blood, liver and kidney functions, electrolytes, fasting blood glucose, blood lipids and thyroid function (chemiluminescent immunoassay, reference range: thyroid stimulating hormone (TSH): 0.56–5.91 $\mu$IU/mL, free triiodothyronine (FT3): 3.53–7.37 pmol/L, free thyroxin (FT4): 7.98–16.02 pmol/L), N-terminal pro brain natriuretic peptide (NT-proBNP), and other related tests were performed. The ventricular ejection fraction was determined by echocardiography, and the estimated glomerularfiltrationrate (eGFR) values were calculated using the Cockcroft-Gault formula. (4) The patients were followed up for MACEs, including occurrence time, and outpatient drug use.

SPSS 25.0 (IBM Corp., Armonk, IL, USA) and R version 4.4.0 (R Core Team, 2013, Vienna, Austria) were used for statistical analyses of the data. Continuous variables are expressed as the mean $\pm$ standard deviation (normal distribution) and median (nonnormal distribution). Two independent sample t tests (normal distribution) or nonparametric tests were used to compare the two groups (nonnormal distribution). For comparisons between the two groups, the chi-square test was employed to describe count data as a frequency (example) and rate (%). The correlation between other variables and MACEs was described using Spearman’s correlation analysis. Univariate and multivariate logistic regression analyses (backward LR) were used to determine independent predictors of MACEs in ISR patients. Receiver operating characteristic (ROC) curve analysis was performed using R version 4.4.0 to determine the optimal threshold as well as the sensitivity and specificity of the threshold. Based on the ideal threshold, patients were separated into two groups, and survival differences between the two groups were examined using Kaplan‒Meier survival analysis. p 0.05 indicated a statistically significant difference.

There were no differences in age, sex, body mass index (BMI), history of hypertension, hyperlipidemia, stroke, diabetes, history of chronic renal insufficiency, ISR reference vessel size, ISR events, or therapy between the two groups. The percentage of patients receiving ticagrelor as an outpatient treatment in the MACE group was considerably greater than that in the non-MACE group (78 (36.3%) vs. 451 (27%), p = 0.006). Albumin (43.42 $\pm$ 4.77 vs. 44.17 $\pm$ 4.46, p = 0.021), direct bilirubin (2.5 (2, 3.5) vs. 2.8 (2.07, 3.73), p = 0.036) and FT3 (2.85 $\pm$ 0.43 vs. 2.92 $\pm$ 0.42, p = 0.019) levels in the MACE group were significantly lower than those in the non-MACE group, and there was no significant difference in other indicators between the two groups (Table 1).

Albumin (r = –0.053, p = 0.021) and FT3 (r = –0.054, p = 0.019) were substantially adversely linked with MACEs, according to Spearman correlation analysis. There was no significant correlation between direct bilirubin and MACEs (r = –0.037, p = 0.109) (Table 2).

The correlation indicators described above were included in a univariate logistic regression analysis, which showed that both albumin and FT3 were predictors of MACEs. Multivariate logistic regression analysis showed that FT3 was an independent predictor of postoperative MACEs in ISR patients (Odds Ratio (OR) = 0.626, 95% CI [0.429–0.913], p = 0.015) (Table 3).

The ROC curve of MACEs was analyzed for FT3, and the optimal threshold for FT3 was determined using the Youden index. When FT3 2.785 pmol/L, the sensitivity was 52.6%, the specificity was 42.8%, and the area under the curve was 54.7% (p = 0.025).

Patients were divided into the FT3 $\ge$2.785 group and the FT3 2.785 group according to the ROC-derived threshold. The findings of the Kaplan‒Meier survival analysis revealed that the incidence of MACEs was considerably lower in the FT3 $\ge$2.785 group than in the FT3 2.785 group (Hazard Ratio (HR) = 0.76, 95% CI [0.88–0.994], p = 0.044) (Fig. 1).

Fig. 1.

Kaplan‒Meier survival analysis on FT3 2.785 group and $\mathrm{\ge }$2.785 group. FT3, free triiodothyronine.

Thyroxine, including tetraiodothyronine (T4) and triiodothyronine (T3), is an influential endocrine hormone that maintains the normal growth and development of the human body, the function of various organs, and the balance of calcium [12]. T4 is its main storage form and is transformed into biologically active T3 by type I and type II deiodinases [12]. In recent years, it has been shown that T3 not only functions as an endocrine hormone but is also widely involved in the regulation of lipid metabolism [13], endothelial function [14], angiogenesis [15, 16], blood pressure, and myocardial contractility [12], all of which are closely linked to the formation and development of atherosclerosis. In recent years, a number of studies have found a correlation between thyroid hormone and acute coronary syndrome (ACS), and low FT3 is an independent risk factor for coronary artery severity and MACE after PCI [17, 18, 19, 20, 21, 22]. Studies have also explored the relationship between thyroxine and ISR [23]. Canpolat et al. [23] reported that elevated plasma FT4 levels were an independent risk factor for ISR after BMS implantation, but few studies have investigated the relationship between FT3 and ISR. In this study, it was found that low FT3 was an independent risk factor for myocardial infarction, recurrent ISR, and TVR in patients with ISR after revascularization, and its mechanism may involve the following aspects.

Patients with hypothyroidism often have abnormal lipid profiles. Tian et al. [24] found that TSH activated the cAMP/PKA/CREB signaling pathway by binding to TSH receptors on the surface of the rat liver cell membrane. This directly upregulates the expression of 3-hydroxy-3-methylglutaryl-CoA reductase (HMG-CoA reductase) and promotes cholesterol synthesis. Bakker et al. [13] found that T3 can upregulate the expression of LDL-C receptor genes on the surface of the liver cell membrane, increasing the synthesis of LDL-C receptors and thus accelerating the clearance of LDL-C from the circulation. However, Bonde et al. [25] observed that proprotein convertase subtilisin/kexin type 9 (PCSK9) levels in hyperthyroid patients were approximately 22% lower than those in the normal population and that FT3 levels were significantly negatively correlated with plasma total cholesterol, very low density lipoprotein cholesterol (VLDL-C), and LDL-C. It was hypothesized that FT3 can reduce plasma LDL-C levels by upregulating the expression of the LDL-C receptor gene and lowering the concentration of PCSK9 in hepatocytes. The event group had considerably lower FT3 levels than the non-MACE group in this study, but there was no significant difference in lipid profiles between the two groups. This result could be related to the diminished effect of FT3 on lipid levels in both groups due to statin use.

T3 is also involved in the regulation of vascular endothelial cell function. When Carrillo-Sepúlveda et al. [14] studied the effect of thyroxine on the thoracic aorta of rats, it was found that T3 activated nitric oxide (NO) synthase in endothelial cells and smooth muscle cells via the phosphatidylinositol 3-kinase/protein (PI3K/Akt) signaling pathway, resulting in NO production that triggered a rapid diastolic response in the smooth muscle of the vessel. This regulatory function is weakened when T3 levels decrease.

T3 also inhibits vascular calcification. Sato et al. [26] reported that T3 affected the expression of genes associated with calcification in rat smooth muscle cells of the aorta. In in vitro studies using cultured human coronary smooth muscle cells, physiological concentrations of FT3 (15 pmol/L) increased the mRNA levels of matrix Gla protein (MGP), which is considered to be a potent inhibitor of vascular calcification in vivo. These results suggest that physiological concentrations of thyroid hormone can directly promote the expression of the MGP gene in smooth muscle cells via thyroid hormone nuclear receptors, thereby preventing vascular calcification in vivo. Ittermann et al. [27] noted that thyroid hormone reduction modifies the smooth muscle structure of the arteries and causes thickening of the vessel walls and decreased compliance, which further aggravates atherosclerosis.

Vascular endothelial cells are damaged by the implantation of coronary stents, and physiological contractions and diastolic function of the vessel walls are inhibited. When the plasma FT3 concentration is reduced, NO secretion by endothelial cells is further suppressed, and the inhibition of vascular calcification is weakened.

Oxidative stress plays a crucial role in myocardial ischemia‒reperfusion (I/R) injury. Studies [28, 29] have shown that T3 can increase the activity of superoxide dismutase (SOD) and glutathione peroxidase (GSH-px) in myocardial cells, reduce the production of reactive oxygen species (ROS), and protect against myocardial I/R injury. In the ischemia‒reperfusion injury model, T3 supplementation significantly reduces I/R-induced ROS production and mitochondrial superoxide concentrations, thus acting as a myocardial protective feature.

In addition, the proportion of patients treated with ticagrelor as outpatients was higher in the MACE group than in the non-MACE group in this study. This finding may be due to the fact that follow-up physicians switched to ticagrelor instead of clopidogrel because of MACE events occurs.

​Although there was no significant difference in MACE incidence between diabetic and non-diabetic patients in this study, a number of previous studies have found a higher rate of ISR in diabetic patients after PCI [30, 31, 32, 33, 34]. Study [30] has shown that insulin resistance and elevated cytokine levels are associated with a higher incidence of ISR after PCI even in patients with normal glucose tolerance. Elevated inflammatory markers (interleukin 6; interleukin 1; tumor necrosis factor alpha; C-reactive protein, etc.) and impaired endothelial function can be detected in pre-diabetic patients with coronary heart disease, resulting in abnormally elevated levels of this cytokine after PCI [31]. Overactive inflammatory pathways that cause clotting can also cause restenosis in coronary stents and a poorer prognosis [32, 33]. Currently, it is believed that sodium-glucose cotransporter 2 (SGLT2)-inhibitors can improve clinical outcomes in patients after PCI [34]. However, the relationship between SGLT2-inhibitors and cardiovascular outcomes in PCI patients was not widely recognized at the time this study was initiated 10 years ago and the proportion of diabetics using SGLT2-inhibitors in this study was extremely low. The effect of SGLT2-inhibitors on interventional therapy in ISR-CTO patients with diabetes mellitus was not analyzed.