Hyperthyroidism and vascular cell adhesion molecule-1 are associated with a low ankle-brachial index

We aimed to assess the ankle-brachial index (ABI) in patients with Graves’ disease. In the cross-sectional assessments, 81 patients with drug-naïve Graves’ disease and 235 with euthyroidism were enrolled. ABI and vascular cell adhesion molecule-1 (VCAM-1) levels were assessed. In the prospective follow-up, 32 patients with Graves’ disease were assessed again after antithyroid drugs for at least 4 weeks, and 32 age- and sex-matched controls with euthyroidism were also followed up. Patients with Graves’ disease had a higher VCAM-1 level (1309 ± 292 vs. 1009 ± 168 ng/mL, P < 0.001) and a lower ABI (0.98 ± 0.11 vs. 1.06 ± 0.10, P < 0.001) than those with euthyroidism. ABI was significantly lower in patients with hyperthyroidism and a high VCAM-1 level than in those with euthyroidism and a low VCAM-1 level (regression coefficient: − 0.050, 95% confidence interval [CI] between − 0.080 and − 0.019; P = 0.001). After treatment with antithyroid drugs, the change in free thyroxine (T4) level was inversely associated with the percentage change in ABI (regression coefficient: − 0.003, 95% CI between − 0.005 and − 0.001, P = 0.001). A synergistic effect of VCAM-1 and free T4 on ABI reduction was observed. After a longitudinal follow-up, an increase in ABI was significantly correlated with a decrease in the free T4 level.


Scientific RepoRtS
| (2020) 10:17076 | https://doi.org/10.1038/s41598-020-74267-7 www.nature.com/scientificreports/ were (1) a high serum free thyroxine (T4) level (> 17.6 pg/mL), (2) a suppressed thyroid-stimulating hormone (TSH) level (< 0.4 μIU/mL), (3) Graves' disease based on clinical manifestation, and (4) a TSH receptor antibody (TRAb) level > 1 U/L; the inclusion criteria for the control group were (1) nodular goiter, (2) a normal serum free T4 level, and (3) a normal serum TSH level. The exclusion criteria for both the case group and control group were as follows: (1) a history of pharmacological treatment for thyroid diseases; (2) a history of diabetes mellitus; (3) a history of coronary artery disease; (4) a history of end-stage kidney disease; (5) a history of severe systemic disease, such as malignancies or psychiatric disorders; (6) current acute or chronic infectious diseases; (7) current thyrotoxic crisis; (8) current autoimmune diseases other than Graves' disease; and (9) current pregnancy. Based on the assessment at baseline, we also excluded patients who had (1) a fasting plasma glucose level ≥ 126 mg/dL, (2) thyrotoxicosis other than Graves' disease, (3) an incomplete four-limb assessment of the ABI due to a known history of lower-extremity surgery or hemodialysis treatment, and (4) an ABI value > 1.40. The study complied with the Declaration of Helsinki and was approved by the Institutional Review Board of Taichung Veterans General Hospital. Written consent was obtained from each patient before the study procedures were performed (Clinical Trial Registration Number: NCT02886949, ClinicalTrials.gov).
procedure. This study included two parts of analytic assessments. For the case-control study at baseline, subjects underwent ABI assessments after an interview for the collection of medical information and anthropometric measurements. Blood samples were obtained for laboratory assessments after a fasting period ≥ 6 h. Within a week after laboratory assessments, patients attended another visit and were categorized into Graves' disease and euthyroidism groups. For the prospective observational study, 32 nonsmokers who started to use thionamide antithyroid drugs prescribed by the same research physician were consecutively invited in the case group for a follow-up assessment. We also followed up 32 age-and sex-matched nonsmokers in the control group. All of these invited subjects underwent anthropometric measurements, ABI assessments and blood sample collection after at least a 4-week period of follow-up.
Laboratory measurements. Plasma glucose levels were determined using the oxidase-peroxidase method (Wako Diagnostics, Tokyo, Japan). Serum creatinine and lipid levels were determined using commercial kits (Beckman Coulter, Fullerton, USA). Serum TSH (normal range 0.4-4 μU/mL) and free T4 (normal range 7.9-17.6 pg/mL) levels were determined using chemiluminescent immunoassays (Siemens Healthcare Diagnostics Inc., Llanberis, United Kingdom). TRAb levels were determined using the chemiluminescent method (normal level < 1 U/L; Roche Diagnostics, Mannheim, Germany). Serum C-reactive protein (CRP) was determined using an ELISA kit (R&D Systems, Minneapolis, MN, USA). The serum VCAM-1 concentration was determined by ELISA (R&D Systems, Minneapolis, USA). The estimated glomerular filtration rate (eGFR) was calculated according to the Modification of Diet in Renal Disease equation 25  ABI assessment was performed in the supine position using the validated device (VP-1000 Plus; Omron Healthcare Co. Ltd., Kyoto, Japan) after patients rested for at least 5 min. The lower systolic blood pressure of the two ankles was recorded as the ankle pressure. ABI values were calculated by dividing the recorded systolic pressure of the ankle by the higher systolic pressure of both arms 17 . The brachial-ankle pulse wave velocity (baPWV) was automatically determined using the ABI device. The reproducibility of the ABI was examined by repeated assessments in 20 subjects. A highly positive correlation of ABI values (correlation coefficient [r] = 0.937, P < 0.001) was observed between the results of the first and second measurements. The 95% confidence interval (CI) was 0.004 ± 0.050 for the bias of the ABI between repeated measurements based on Bland-Altman plots.
Statistical analysis. We present the mean ± the standard deviation (SD) for continuous variables and numbers with percentages (%) for categorical data. The clinical variables were tested for statistically significant differences using Student's t tests for continuous variables between two groups and using χ 2 tests for categorical variables. The change in each continuous variable within a group at baseline and after follow-up was analyzed by paired t tests. The association between ABI values and free T4 levels was determined using Pearson's correlation test. A two-tailed P value < 0.05 was considered statistically significant. Statistical analysis was performed using SPSS 22.0 (IBM, Armonk, NY, USA).

Results
case-control assessment. A total of 326 subjects were enrolled in this study. There were 81 subjects with Graves' disease in the case group and 235 subjects with euthyroidism in the control group (Fig. 1). The median TRAb level was 30.6 U/L (interquartile range between 5.8 and 57.1 U/L) in the Graves' disease group. The characteristics of these two groups are shown in Table 1. Patients with Graves' disease had lower TSH and higher free T4 levels (both P values < 0.001) because of different inclusion criteria from those with euthyroidism. Furthermore, the subjects with Graves' disease had a younger age (P < 0.001); a higher proportion of current smokers (P < 0.001); a lower body mass index (BMI) (P < 0.001); a higher brachial systolic blood pressure (P = 0.028); a higher heart rate (P < 0.001); a higher glucose level (P = 0.007); lower serum levels of total cholesterol (P < 0.001), high-density lipoprotein (HDL) cholesterol (P < 0.001), and triglycerides (P = 0.008); and a higher eGFR (P < 0.001) than those with euthyroidism. Notably, subjects with Graves' disease had a lower systolic blood pressure at the ankle (125 ± 21 vs. 131 ± 22 mmHg, P = 0.018) and a lower ABI value (0.98 ± 0.11 vs. 1.06 ± 0.10, P < 0.001). Subjects with Graves' disease also had a higher serum VCAM-1 level (1309 ± 292 vs. 1009 ± 168 ng/mL, P < 0.001). www.nature.com/scientificreports/ We categorized the potential factors to assess the ABI value in this study ( Table 2). The subjects with an older age, male sex, a higher BMI, a higher diastolic brachial blood pressure, a lower heart rate, a higher total cholesterol level, higher triglycerides, a lower eGFR, a higher baPWV, a lower serum VCAM-1 level, and higher CRP level had a higher ABI value than the other corresponding groups. Furthermore, ABI values were significantly correlated with free T4 levels (correlation coefficient [r] = − 0.335, P < 0.001; Fig. 2), but were not significantly correlated with CRP levels (P = 0.813).
Because of the pathophysiological relationship between VCAM-1 and hyperthyroidism, we grouped all subjects into four groups: low VCAM-1 without hyperthyroidism (n = 142), high VCAM-1 without hyperthyroidism (n = 93), low VCAM-1 with hyperthyroidism (n = 15), and high VCAM-1 with hyperthyroidism (n = 66) according to the hyperthyroid status and the serum VCAM-1 median of 1055 ng/mL. As shown in Fig. 3, there was a significant reduction in the ABI value in the patients with hyperthyroidism (P value for trend < 0.001). In addition to age and sex, we selected the confounding factors as those variables which significantly influenced both hyperthyroidism status (shown in Table 1) and the ABI value (shown in Table 2) for the multivariate regression analysis. A high VCAM-1 level with hyperthyroidism was an independent risk factor for a low ABI value after adjustments for age, BMI, heart rate, total cholesterol, triglycerides, and eGFR (regression coefficient = − 0.050, 95% CI between − 0.080 and − 0.019; P = 0.001, Table 3).
Longitudinal assessment of the subjects who had undergone antithyroid treatment. In the case group, all 32 recruited subjects completed assessments after treatment with antithyroid drugs for least 4 weeks. The median follow-up duration was 77 days (interquartile range between 57 and 123 days). The baseline characteristics of cases who participated in the follow-up were not significantly different from those of cases who did not participate in the follow-up. In the 32 subjects in the control group, the median follow-up duration was 87 days (interquartile range between 35 and 121 days). Despite being matched for age and sex, the subjects in the control group had a lower heart rate (P < 0.001), a higher total cholesterol level (P < 0.001), a higher HDL cholesterol level (P = 0.024), a lower eGFR (P < 0.001), a higher ABI value (P < 0.001), and a lower serum VCAM-1 level (P < 0.001) at baseline than the 32 subjects in the case group (Table 4). www.nature.com/scientificreports/ After treatment with antithyroid drugs, free T4 significantly decreased (− 22.189 pg/mL, 95% CI between − 27.347 and − 17.031 pg/mL), but TSH did not significantly increase (2.600 μIU/mL, 95% CI between − 1.713 and 6.912 μIU/mL). Significant increases in BMI, total cholesterol level, and HDL cholesterol level were observed, and significant decreases in systolic and diastolic brachial blood pressure and heart rate were also observed during the follow-up period. A significant increase in ABI values and a decrease in VCAM-1 levels were also observed ( Table 5). The increase in ABI values was significantly correlated with a decrease in free T4 levels (r = − 0.406, P = 0.021). In the control group, only the eGFR significantly increased and the VCAM-1 level significantly decreased during the observation follow-up period. The association between changes in free T4 levels and ABI values was not significant (r = 0.074, P = 0.689) in the control group.
Using multivariate linear regression, treatment with antithyroid drugs was an independent factor associated with an increased percentage change in the ABI value (regression coefficient = − 0.003, 95% CI between − 0.005 and − 0.001, P = 0.001) during the follow-up of the study after adjustment for age, sex, change in BMI, and change in heart rate ( Table 6).

Discussion
Our main findings are that a lower ABI was observed in patients with hyperthyroidism and high circulating VCAM-1 levels. After treatment with antithyroid drugs, ABI values were significantly increased, and the increase in the ABI values was correlated with a reduction in the circulating free T4 level. A lower ABI was associated with a higher mortality risk in subjects with an ABI value < 1.1 21,26 , and an ABI value < 1.0 should be considered to indicate a borderline abnormal ABI, as there was an association with increased CV risk even when the ABI value was > 0.9 27 . In the present study, Graves' disease was related to a significant decrease in ankle systolic blood pressure, resulting in a low ABI, as a predictor of CV disease.
Patients with Graves' disease had a higher mean VCAM-1 level than those with normal thyroid function. It has been reported that a high circulating VCAM-1 level was detected in patients with Graves' disease, and the level was positively correlated with the free T4 levels 28 . The increased VCAM-1 level resulted from enhanced expression in the endothelium due to lymphocyte infiltration in the thyroid gland 12,29,30 . However, VCAM-1 released in the blood might induce chronic inflammation and endothelial dysfunction in peripheral arteries 13 . De Ciuceis et al. 31 reported an increased level of VCAM-1 and a decreased number of endothelial progenitor cells in peripheral blood during hyperthyroid status in patients with Graves' disease. Endothelial dysfunction was also associated with an increase in asymmetric dimethylarginine in patients with Graves' disease 32,33 .
In the present study, only the VCAM-1 level, but not the CRP level, was significantly increased in patients with Graves' disease. In line with our finding, Burggraaf et al. 34 reported that the CRP level was not significantly Table 1. Clinical characteristics of the subjects with Graves' disease and euthyroidism. ABI ankle-brachial index, baPWV brachial-ankle pulse wave velocity, BMI body mass index, BP blood pressure, eGFR estimated glomerular filtration rate, Free T4 free thyroxine, HDL high-density lipoprotein, VCAM-1 vascular cell adhesion molecule-1, TSH thyroid-stimulating hormone. www.nature.com/scientificreports/ associated with thyroid function. Furthermore, the VCAM-1 level is reported to be strongly associated with PAD 24 . In the present study, high circulating VCAM-1 levels had a synergistic effect with hyperthyroidism on low ABI values. Therefore, VCAM-1 might be a mediator in the relationship between hyperthyroidism and PAD. Antithyroid drugs have been the most popular therapy used to treat hyperthyroid status 35,36 . In the present study, the ABI values increased after treatment, with reductions in both free T4 levels and VCAM-1 levels. In line with our data, antithyroid drugs have been reported to reduce circulating VCAM-1 levels 28 , and improve Table 2. ABI values according to associated risk factors. ABI ankle-brachial index, baPWV brachial-ankle pulse wave velocity, BMI body mass index, BP blood pressure, eGFR estimated glomerular filtration rate, HDL high-density lipoprotein, VCAM-1 vascular cell adhesion molecule-1, TSH thyroid-stimulating hormone. *Cutoff value based on the median. www.nature.com/scientificreports/ endothelial function 14 . Furthermore, an increased number of endothelial progenitor cells and decreased biomarkers of platelet activity were also reported after hyperthyroidism treatment 31,37 .
To the best of our knowledge, the present study is the first report of the relationship between the ABI and Graves' disease. However, there are some limitations in the present study. First, we did not examine the mechanism or causal relationship between hyperthyroidism and PAD. Second, Graves' disease was newly diagnosed in all patients with hyperthyroidism. However, we did not know the definitive time when hyperthyroidism occurred before diagnosis due to the diversity of symptoms. Third, our findings may not be applicable to patients with diabetes who were excluded due to a potential noncompressible artery at the ankle 38,39 . Fourth, we assessed only a reversible ABI reduction after treatment with antithyroid drugs but not long-term outcomes, such as  www.nature.com/scientificreports/  www.nature.com/scientificreports/ lower-extremity amputation. Finally, we did not compare the effects of different Graves' disease treatments on the ABI. It was reported that additional radioiodine therapy was associated with low long-term mortality 11 . Further large-scale studies with controlled treatment are needed to investigate the relationship between hyperthyroidism and PAD.
In conclusion, a higher free T4 level is associated with a lower ABI value in patients with untreated Graves' disease. A high circulating VCAM-1 level had a superimposing effect on the decrease in the ABI value during hyperthyroid status. After treatment with antithyroid drugs, an increase in ABI values accompanied by reductions in free T4 and VCAM-1 levels was observed.
Received: 26 March 2020; Accepted: 28 September 2020 Table 5. Differences from baseline values after follow-up. *Denotes P values for differences between case and control groups. ABI ankle-brachial index, baPWV brachial-ankle pulse wave velocity, BMI body mass index, BP blood pressure, eGFR estimated glomerular filtration rate, Free T4 free thyroxine, HDL high-density lipoprotein, TSH thyroid-stimulating hormone, VCAM-1 vascular cell adhesion molecule-1.  Table 6. The effect of changes in free thyroxine levels on the percentage change in the ankle-brachial index. B regression coefficient, CI confidence interval, Model 1 adjusted for age and sex, Model 2 adjusted for age, sex, change in body mass index, and change in heart rate.