Does the Tendency to Thrombosis Increase in Autoimmune Thyroid Dysfunction? Rotational Thromboelastometric Analysis

The tendency for thrombosis is a well-known issue in hyperthyroidism, and the tendency for both bleeding and thrombosis has been shown in hypothyroidism. In the study, we aimed to establish the tendency to thrombosis by evaluating the coagulation system in thyroid diseases using thromboelastography (TEG). Thirty patients with new-onset autoimmune-based hypothyroidism, 30 with hyperthyroidism and 30 healthy volunteer subjects were enrolled in this study. Complete blood count, prothrombin time, activated partial thromboplastin time, brinogen levels and rotational thromboelastography (ROTEM) were studied. In the ROTEM/TEG analysis, coagulation time (CT), coagulation formation time (CFT) and maximum clot rmness (MCF) were measured. Short CT and/or CFT and/or long MCF were presumed as the tendency to hypercoagulability. The mean age of subjects with hyperthyroidism and hypothyroidism was 46±15.9 years and 44.8±13 years, respectively. The thrombocyte count in the hypothyroid group and both thrombocyte counts and brinogen levels in the hyperthyroid group was signicantly higher than in the healthy group (p=0.004, p=0.001, p<0.001, respectively). According to ROTEM/TEG results, higher mean MCF gures were noted in both the hypo- and hyperthyroid groups compared with the healthy group (p=0.014 (hypothyroid vs. healthy groups), p=0.020 (hyperthyroid vs. healthy groups)). We showed a tendency for thrombosis using ROTEM in patients with both hyper- and hypothyroidism without thrombosis. The correlation between ROTEM/TEG parameters and the main components of thrombosis formation (thrombocyte count and brinogen level) indicate that the main factor might be inammation (autoimmunity), apart from thyroid hormone levels. thromboplastin INR; international normalized SD; Standard


Introduction
Thyroid dysfunction, either as hypothyroidism or hyperthyroidism, may lead to hemostatic abnormalities varying from subtle subclinical changes in coagulation tests to clinically overt bleeding or thrombosis. It is presumed that the thrombotic tendency due to coagulation factors and hypo brinolysis is increased in hyperthyroidism, whereas hypothyroidism predilects to bleeding [1,2]. Many studies showed that increased free thyroxine (fT4) was a risk factor for venous thrombosis, and contrastingly, low fT4 was protective against it [3]. However, in some published papers, it has been revealed that there is a hypercoagulable state due to the increase in some coagulation and brinolysis activation markers in hypothyroidism [4,5].
Moreover, even in patients with subclinical hypothyroidism, it has been shown that the risk of thrombosis is increased in contrast to the bleeding tendency [6,7]. Thromboelastography (TEG) is a technique used to evaluate the tendency for bleeding or coagulation, which cannot be detected by routine laboratory tests. Rotational thromboelastometry (ROTEM), as a modi ed version of TEG, has emerged as a technique for in vitro evaluation of clot rmness, clot formation, and brinolysis in complete blood samples. The viscoelastic and mechanical features of clots are measured using ROTEM by collecting information about various steps of coagulation. Thus, the coagulation process can be evaluated globally and information regarding in vivo hemostasis can be obtained by laboratory means [8]. In the literature, there are studies using ROTEM/TEG to detect coagulation and bleeding disorders in trauma, postoperative bleeding, malignancy, ischemic heart disease, liver transplant, cardiac surgeries, and even in ammatory rheumatologic diseases [8][9][10]. The use of ROTEM/TEG to evaluate coagulation disorders in thyroid diseases is not common and has been used only in patients with thyroid cancer [11]. In the study, we aimed to evaluate the hemostatic system globally using ROTEM/TEG, which was never been used before to speci cally investigate thrombotic tendency, and to establish the effects of possible disease-related factors on ROTEM/TEG parameters in thyroid diseases.

Subjects
Thirty patients with new-onset autoimmune-based hypothyroidism and 30 with hyperthyroidism before the commencement of therapy were enrolled in the study, along with 30 healthy volunteer subjects without thyroid disease. The normal range for thyroid function assays in our institution is thyroid-stimulating hormone (TSH) 0.35-4.94 µIU/mL, free thyroxine (fT4) 0.81-1.42 ng/mL, and free tri-iodothyronine (fT3) 1.71-3.71 pg/mL Hyperthyroidism was de ned as fT4 > 1.42 ng/mL and TSH < 0.35 µIU/mL and the overt hypothyroidism was de ned as TSH > 4.94 mIU/L and fT4 < 0.81ng/dL. Subjects with diabetes mellitus (DM); hypertension (HT); chronic renal failure (CRF); rheumatologic, hematologic, cardiopulmonary, or cerebrovascular disease; hematologic or solid organ malignancies; a history of venous or arterial thromboemboli; and current use of acetylsalicylic acid (ASA), non-steroid anti-in ammatory drugs (NSAIDs) or anti-coagulant/anti-aggregant drugs were excluded. In addition, women with non-thrombotic clinical signs such as thrombocytopenia, hemolytic anemia or livedo reticularis or with a history of maternal morbidity suggestive of clinical signs of antiphospholipid antibody syndrome were not included in the study. Age-and sex-matched healthy subjects enrolled; none had any chronic disease or current drug use. TEG analyses were performed in all patients with thyroid dysfunction before medical therapy and healthy volunteer subjects at the time of enrollment to the study. Informed consent was obtained from all individual participants included in the study.
TSH, fT4,fT3, thyroid antibodies [anti-thyroid peroxidase-anti-TPO (0-34 IU/mL), anti-thyroglobulin-anti-TG (0-4,11 IU/mL), thyrotropin receptor antibodies-TRAb (>1 positive)] were measured with an automated direct chemiluminescent immunoassay. Complete blood count (CBC) (Becton-Dickinson anticoagulated tubes), prothrombin time (PT) (performed using Noplastine CI Plus (Diagnostica Stago, Asnieres, France), normal range: 9.4-15 seconds), activated partial thromboplastin time (PTT) (performed using rabbit brain cephalin with kaolin as the activator (Diagnostica Stago, Asnieres, France), normal range: 26-36 seconds), brinogen (STA®Fibrinogen kit was used for the quantitative determination of brinogen levels using the clotting method of Clauss (Diagnostica Stago, Asnieres, France), normal range: 200-400 mg/dL), and ROTEM to detect the tendency to thrombosis, were simultaneously studied. Blood samples were collected using a 19-gauge needle under minimum stasis. Blood samples for ROTEM and coagulation pro le were drawn into 4.5-mL vacutainers (Becton-Dickinson) containing 3.2% trisodium citrate with a citrate/blood ratio of 1:9. With the same procedure, more blood was drawn for a CBC into Becton-Dickinson anticoagulated tubes. The main principle of ROTEM study involves a motionless cuvette and a pin which is inserted into a certain amount of whole blood lying in the cuvette. The brin-thrombocyte bonds within this clot promote movement of the needle back and forth at a certain angle and nally convert the strength power to electrical signals with the help of an electrotransducer which are then transferred to a graphic document [8].
ROTEM analysis was performed using ROTEM device Coagulation Analyzer model Gamma 2500 (Tem International, Munich, Germany). The 300 microliter sodium citrated whole blood samples of the subjects were recalci ed using 20 µl 0.2 mol/L CaCl (starTEM®; Pentapharm, Munich, Germany). Different activators (20 µl partial tromboplastin-phospholipid and ellagic acid containing rabbit brain extract as contact activator was used for intrinsic pathway activation, 20 µl tissue factor containing rabbit brain extract was used for extrinsic pathway activation) were used for activation of intrinsic (I) and extrinsic (E) pathways were used. Therefore coagulation time (CT) (time from baseline amplitude until 2 mm riseseconds), clot formation time (CFT) (time interval from 2 mm rise until 20 mm rise -seconds) and maximum clot rmness (MCF -mm) were measured using tromboelastography. Shortened CT and/or CFT and/or prolonged MCF was considered as tendency for hypercoagulability [9,10,12]. These parameters were recorded seperately for both I and E pathways.

Statistical Analysis
The statistical analysis was done using SPSS version 20.0. Continuous variables were given as mean ± standard deviation, while categorical variables as frequency and percentage. Distribution of normality was evaluated by Shapiro Wilk's test. For the comparison of normally distributed groups, the independent t-test was used if the group number was two, and one-way analysis of variance (ANOVA) was used if the group number was three. Tukey's multiple comparison tests were used for multiple comparisons of normally distributed groups. For the comparison of non-normally distributed groups, the Mann-Whitney U test was used for comparisons if the group number was two, and the Kruskal-Wallis H test was used if the group number was three. In correlation analysis, Pearson correlation analysis was used for normally distributed parameters and Spearman correlation coe cients for non-normally distributed parameters. A p value <0,05 within 95% con dence interval was accepted as signi cant.

Results
The mean age of the 30 patients with hypothyroidism and 30 with hyperthyroidism was 46±15.9 years and 44.8±13 years, respectively. All patients' sera were positive for at least one of the thyroid autoantibodies (anti-TPO, anti-TG, TRAb). When the patients with thyroid dysfunction were compared with the healthy subjects, those with hypothyroidism had a higher number of thrombocytes and those with hyperthyroidism had both higher thrombocytes and brinogen levels (p = 0.004, p = 0.001, p<0.001, respectively). Laboratory and demographic datas results are given in Table 1.

Correlation analyses
In the correlation analyses, ROTEM/TEG parameters and age, TSH, free T4 and T3, and brinogen levels and thrombocyte count were evaluated in the hypothyroid, hyperthyroid, and healthy control groups, separately. The correlation analysis results that found to be statistically signi cant are given in Table 3 and shown in Figure 1 and 2.

Discussion
This study is the rst in the literature to evaluate thrombotic tendency using the ROTEM/TEG laboratory method in thyroid diseases (hypo-and hyperthyroidism). We found signi cant results promoting thrombosis in ROTEM/TEG parameters, both in hypo-and hyperthyroidism.
Until recently, the majority of both experimental and clinical studies showed that hyperthyroidism was associated with an increased risk of thrombosis, and hypothyroidism with an increased bleeding tendency.
However, the most recent studies showed a predilection to thrombosis in patients with hypothyroidism, similar to hyperthyroidism [7,[13][14][15]. Hemostasis is a complicated process involving thrombocytes, coagulation factors, and interaction with the endothelium. Thrombus formation and resolution depend on thrombocyte function, coagulability of blood, and vascular endothelium. Thrombosis formation in hyperthyroidism is associated with various coagulation factors, and a number of factors such as HT, lipid disorders, and increased body weight may operate in thrombosis formation in hypothyroidism (both overt and subclinical hypothyroidism). Furthermore, it was shown that by also accelerating the atherosclerotic process, these stated factors might ultimately cause an increased risk of cardiovascular disorders, resulting in thromboemboli [15].
The effects of thyroid hormones and the relation with thrombocytes, coagulation factors, and brinogen levels have both been extensively assessed in studies regarding coagulation defects in thyroid diseases [3,16,17]. In these papers, decreased brinolytic activity and coagulation were detected in hyperthyroidism, and increased or decreased brinolytic activity were suggested in hyperthyroidism [18][19][20]. The classic hemostasis-coagulation tests used until recently including brinogen and factor levels do not re ect the whole, but certain components of the coagulation process [15]. ROTEM/TEG allows rapid in vitro dynamic and global evaluation of hemostasis. Additionally, it represents a rapid, portable, and veri ed method for physicians and may obviate thrombotic tendency undetectable in routine coagulation tests in various settings [8,21]. In this study, we simply measured brinogen levels and thrombocyte counts probably responsible for the pathogenesis of thrombosis in thyroid diseases, and made a global assessment by incorporating many factors regarding hemostasis using ROTEM/TEG for the rst time in the literature.
One of the most important components of the coagulation system is thrombocytes. Various thrombocyte changes were observed in overt hyperthyroidism. Homoncik et al. found increased thrombocyte function in patients with hyperthyroidism compared with euthyroid control subjects [22,23]. In our study, the thrombocyte count increased in patients with hypo-and hyperthyroid groups in comparison with the healthy control group. MCF, as a ROTEM parameter, which is affected by brinogen concentration, levels of factor XIII, thrombocyte count, thrombocyte function, and thrombin concentration. A signi cant correlation between MCF and thrombocyte count has been shown in many diseases [10,12,24]. It can be claimed that thrombocytes constitute an important component of thrombotic tendency shown by ROTEM/TEG in patient groups with thyroid dysfunction because the thrombocyte count was signi cantly correlated with both intrinsic and extrinsic ROTEM/TEG tests in both thyroid disease groups of our study.
Another important component of the coagulation system is brinogen, which contributes to thrombosis risk in various ways. Although the underlying mechanisms are obscure, it is assumed that the situation is probably due to in ammation and hypercoagulability. High brinogen levels promote a hypercoagulable state by increasing thrombocyte aggregation and thrombocyte reactivity. Additionally, brin structure is affected by brinogen levels. In vitro brin clots in the presence of high brinogen concentrations are more dense, composed of thicker brin bers, and are more resistant to brinolysis than clots formed in low brinogen concentrations [25]. Decreased brinolytic activity was also detected in hyperthyroidism due to increased levels of factor VIII, IX, von Willebrand factor (VWF), brinogen, and plasminogen activator inhibitor-1 (PAI-1) [15]. This nding was observed both in overt and subclinical hyperthyroidism due to endogenous and exogenous causes. The issue is more complicated in hypothyroidism. Hypercoagulation was observed in some studies involving patients with overt hypothyroidism, whereas hypercoagulation was present in the others. In the studies where hypercoagulation was evident, several causes including hypo brinolysis, increased brinogen and PAI-1 levels, and decreased factor VIII and X activity were blamed as possible causes of hypercoagulability in hypothyroidism [13,16,17]. In accordance with other studies, we observed higher brinogen levels in the hyperthyroid group compared with the hypothyroid and healthy control groups. Although not statistically signi cant, the hypothyroid group had higher brinogen levels, similar to the hyperthyroid group, when compared with the healthy control group. We detected a thrombotic tendency by interrelation between brinogen and ROTEM parameters in both the hypo-and hyperthyroid groups. Despite con icting studies regarding the tendency for hypercoagulability in hypothyroidism, we showed thrombotic tendency using ROTEM/TEG for the rst time in hypothyroidism in agreement with the studies by Akıncı et al. and Cantürk et al. The positive correlation between brinogen levels and MCF as a ROTEM/TEG parameter in both groups supported thrombotic tendency and demonstrated that brinogen similar to thrombocytes was another important component effective on MCF [7,13].
In addition to the changes in coagulation factors, abnormalities in thyroid hormones were also suggested as possible causes of coagulation abnormalities in thyroid diseases [15,17]. The relation between thyroid hormones and normal hemostasis is complicated, and how thyroid dysfunction affects hemostasis has not been de nitively elucidated. Furthermore, it is claimed that underlying in ammation or autoimmunity may have effects on thrombosis [2,26]. Thyroid hormones may affect many physiologic and pathologic processes including blood coagulation and atherogenesis [15]. The role of the immune system in the hemostatic system was emphasized, and it was suggested that an activated immune system had the main effect on coagulation system because a similar in uence was detected on hemostatic parameters in both subclinical and overt thyrotoxicosis [2]. Another study in 2010 supported the claim by showing the simultaneous presence of autoimmune-based antiphospholipid syndrome leading to thrombosis and thyroid autoimmunity [26]. In our study, no correlation was detected between thyroid hormones and ROTEM/TEG parameters. Furthermore, the thrombotic tendency in ROTEM/TEG parameters in both de ciency (hypothyroidism) and excess (hyperthyroidism) of thyroid hormones indicated that the level of thyroid hormones was not associated with this setting. Underlying autoimmunity or in ammation (in ammation-related thrombocyte increase) must be kept in mind as the possible main factor in the pathophysiology of thrombotic tendency evident by laboratory tests because autoimmune-based thyroid diseases were present in both groups.
TEG studies in thyroid diseases are scarce in the literature, and to the best of our knowledge, only one study enrolled 62 patients with nodular goiter, 53 with papillary thyroid cancer, and 61 healthy subjects. In this study, TEG parameters [K (clot formation time), MA ( nal clot rmness), and angle (alfa) and R] were compatible with hypercoagulability in patients with papillary cancer before undergoing surgery. However, information regarding the hormonal status of the patients, i.e. eu-/hyper-/hypothyroidism, was missing. Thrombocyte count was higher in the papillary cancer group compared with the healthy group (statistically non-signi cant); thrombocyte count was correlated with some of TEG parameters [11]. Similarly, a positive correlation was present between thrombocyte count and ROTEM/TEG parameters in our study indicating thrombotic tendency. Nevertheless, our study differs from that study by being the rst TEG/ROTEM study to include patients with hypo-and hyperthyroidism due to thyroid diseases other than thyroid malignancies.
Our study has some limitations. The rst is the cross-sectional design. Serial analyses and long-term follow-ups will shed light on the possible pathophysiologic causes of coagulation defects in thyroid diseases. Secondly, because we did not follow up with the patients, the course of those with impaired ROTEM results remained unknown. Thirdly, the lipid pro le as a possible risk factor for thrombosis was not studied in patients with hypothyroidism. Fourth, anti-phospholipid antibodies, which may be associated with thrombosis in autoimmune thyroid diseases, should not be considered as a laboratory test in asymptomatic patients. In addition, the results regarding the presence of asymptomatic antiphospholipid antibodies in autoimmune thyroid diseases are contradictory. Although this relationship is mentioned especially in Graves' patients, there are also publications showing that the frequency of asymptomatic antiphospholipid antibodies in autoimmune thyroid patients is not different from healthy individuals [27].
Lastly, we failed to include patients with thyroid dysfunction due to non-autoimmune causes as a control group. Despite all these drawbacks, thrombotic tendency according to ROTEM results in both hypo-and hyperthyroidism will contribute to the current literature.
As a result, our study is the rst in the literature in which TEG/ROTEM was used in thyroid dysfunction and we found thrombotic tendency independent from thyroid hormones in patients with both hypo-and hyperthyroidism in the absence of clinical evidence of thrombosis. The correlation between ROTEM parameters and indirect measures of in ammation, and thrombocyte counts and brinogen levels as the main constituents of thrombosis during hemostasis may indicate that the major factor is in ammation (autoimmunity), rather thyroid hormones. Assessing the hemostasis pro le of patients with both hypo-and hyperthyroidism may help to predict the future risk of clinical thromboembolism and thus help with patient management. Use of tests such as ROTEM/TEG in thyroid diseases may help physicians for globally assessing tendency for thrombosis and bleeding.
Financial disclosure None Declarations Financial disclosure: None Ethical approval: Ethics committee was approved the study (Eskişehir Osmangazi University Ethical Board of Clinical Researches in 9th May 2019/no 07).
Con icts of Interest: There is no con ct of interests about the publication of this paper.
Informed consent: The need for written informed consent was waived considering the retrospective nature of the study.
Consent for publication: This study only collected the patient's gender, age, and laboratory data and did not involve the patient's personal privacy.

Author Contributions
GBC: conception and design, organization of the study, data processing, manuscript drafting, and writing the nal version.
DUC: data collection and analyses, manuscript drafting, critical revision of the work and nal version of the manuscript.
BT: data collection and analyses, manuscript drafting and revision.
HUT: data collection and analyses, manuscript drafting and revision.  Figure 1