Thyroid function test variability and cardiovascular morbidity in hyperthyroidism

The variability of thyroid function tests (TFTs) during antithyroid drug (ATD) therapy and its association with adverse health outcomes have not been previously studied. The aim of this study was to evaluate the association of TFT variability and cardiovascular morbidity during ATD therapy.


| INTRODUCTION
The clinical course of hyperthyroidism can vary significantly depending on several factors, such as the cause and severity of the disease, patient characteristics and response to the treatment. 1,2 Graves' disease (GD) is an autoimmune disease characterized by circulating antibodies against the thyrotropin receptor and the most common cause (70%-80%) of hyperthyroidism in iodine sufficient areas. 3,4 Other common causes of hyperthyroidism include toxic multinodular goitre and solitary toxic adenoma, characterized by nodular autonomy. 3,4 The three treatment options for hyperthyroidism include longterm ATD therapy, radioactive iodine (RAI) and surgery. 1 In Europe, ATD therapy is commonly preferred as the first-line treatment of GD, and if hyperthyroidism recurs, second-line treatment with either RAI or surgery is recommended. 1,5,6 The recurrence rate of hyperthyroidism after ATD therapy is high (40%-60%), increasing the risk of associated health problems, for example cardiovascular complications. [7][8][9] As thyroid hormones have a major role in the regulation of the cardiovascular system, acute cardiovascular manifestations are common during hyperthyroidism and cardiovascular morbidity remains increased for years even after restoring normal thyroid function. 2,[10][11][12][13] In the treatment of hyperthyroidism, recent evidence highlights the importance of rapid elimination of hyperthyroidism and early definitive treatment with RAI or surgery, if remission with ATD therapy is unlikely, as patients with poor biochemical control of hyperthyroidism have an increased mortality. [14][15][16][17][18] Rapid restoration of euthyroidism has been shown to improve the survival of the patients regardless of the treatment modality. 14,18 The response to ATD therapy, however, is difficult to predict and a substantial number of patients will eventually need lifelong thyroid hormone replacement regardless of the first-line treatment option chosen. 19,20 Several studies with partly conflicting results have recognized factors that predict the recurrence of hyperthyroidism after ATD therapy in GD, 8,[21][22][23][24][25][26] whereas the response to ongoing ATD therapy regarding thyroid hormone variability during the treatment and its clinical relevance have not been studied previously.
The aim of this study was to evaluate the association of the thyroid function test (TFT) variability with the risk of cardiovascular morbidity after the initiation of ATD therapy. Furthermore, the aim was to identify baseline clinical factors related to a high TFT variability during ATD therapy.

| MATERIALS AND METHODS
In this retrospective study, all consecutive patients with a newly University Hospital, as the regional clinical guidelines instruct primary F I G U R E 1 Formation of the study cohort. The patients of the study cohort and the clinical data on their treatment were gathered from the hospital patient record system of endocrinology (Endo registry), which is an electronic system used in the endocrinology clinic aimed to support clinical practice.
The data obtained from the Endo registry included clinical characteristics of the patients registered on the first endocrinology clinic visit (Table 1) and follow-up information on possible second-line treatments with RAI or thyroid surgery. Cardiovascular morbidity was evaluated based on hospital visits to Tampere University Hospital due to a cardiovascular disease (CVD) between the first visit to the endocrinology clinic and the end of the follow-up on March 31, 2019. The data on the CVD-associated inpatient and outpatient hospital visits were obtained from the data administration services of the hospital.
The outpatient visits also included remote contacts. The cardiovascular diagnoses (ICD-10 diagnosis codes I10-99) were divided into nine subgroups presented in Table 2   first secondary diagnoses of the visit was included in the abovementioned subgroups.

| Statistical analysis
IBM SPSS Statistics version 27.0 was used in the statistical analyses.
A p-value < .05 was considered statistically significant. Continuous variables presented in Tables 1 and 3 were compared with the Mann-Whitney U test and χ 2 test was used to compare categorical variables.
The visit-to-visit variability of thyroid hormone measurements was evaluated by using the coefficient of variation (CV) as a statistical measure. The CV provides information about the dispersion of the measurements around the mean and it is commonly used, for example, in the clinical practice of diabetology to assess glycaemic variability. [28][29][30] The CV is a dimensionless number, commonly presented as the percentage of the SD of the mean, and therefore it allows the comparison of measurements with different reference   (Table 1).
There was no difference in the medians or the CVs of the TSH, fT4 and fT3 measurements during the follow-up among patients with cardiovascular morbidity compared to patients without cardiovascular morbidity (Supporting Information: Table 1).
In the univariable analyses including the whole study cohort, age, male gender, hypercholesterolaemia, diabetes, hypertension and baseline TRAbs were associated with cardiovascular morbidity ( Baseline clinical characteristics of the study cohort in relation to the follow-up fT4 median and fT4-CV are represented in Table 4. Male patients and those with diabetes had a slightly higher fT4 median than female patients and those without diabetes (p = .048 and .003, respectively). Patients with a family history of thyroid disease had a slightly lower fT4 median compared with patients without a family history (p = .028). Patients with positive TRAbs or TPOAbs had a higher fT4-CV compared with those with negative TRAbs or TPOAbs (p = .002 and .024, respectively).

| DISCUSSION
In the present study, we found that after the initiation of ATD therapy patients with autoantibody-related hyperthyroidism have a higher fT4 variability compared to patients with hyperthyroidism without autoantibodies. Age, male gender and fT4 variability are independent risk factors for cardiovascular morbidity, while TRAb positivity seems to protect from it. However, among patients with positive TRAbs, the variability of fT4 is associated with cardiovascular morbidity. Thus, fT4 variability seems to be a risk factor for cardiovascular morbidity in all hyperthyroid patients.
In previous studies, poor biochemical control of hyperthyroidism, evaluated by cumulative periods of decreased TSH or increments of 10 pmol/L in serial fT4 measurements, has been linked to increased mortality. 17,18 In 2013, Boelaert et al. 18  T A B L E 3 Cardiovascular morbidity in the univariable and multivariable analyses of the hyperthyroid patients treated primarily with antithyroid drug therapy. with cardiovascular morbidity also among TRAb positive patients.
The possible confounding effect of the aetiology of hyperthyroidism may also explain, why no difference was found in the unadjusted comparison of fT4 variability between the patients with cardiovascular morbidity and those without morbidity.
In this study, the association of TFT variability with morbidity due to any CVDs and due to the subgroup of arrhythmias was analysed. Nonarrhythmia related CVDs were not evaluated separately due to the small number of events in these groups.
Arrhythmias, mainly atrial fibrillation, accounted for most of the cardiovascular morbidity registered during the follow-up.
T A B L E 4 Clinical characteristics of the hyperthyroid patients treated primarily with antithyroid drug therapy in relation to the median and the CV of all measured fT4 values during the follow-up. Hyperthyroidism is a well-known risk factor for atrial fibrillation and has also been linked to other CVDs, such as cerebrovascular diseases, valvular diseases, cardiomyopathies, heart failure and diseases of arteries and veins, but not inevitably to coronary artery disease. 2,[9][10][11][12] Hypertension may be regarded as a risk factor, rather than being a CVD itself, although hyperthyroidism may lead to secondary systolic hypertension. 33 Therefore, the multivariable analyses were repeated excluding the patients with coronary artery disease or hypertension and the results did not change.
We found that the variability of fT4 measurements was associated with cardiovascular morbidity, whereas the variability of TSH or fT3 was not. This finding may be partly explained by the different responses of the thyroid axis hormones to the initiation of ATD therapy. After the initiation of ATD therapy, TSH may remain suppressed for several months due to delayed normalization of the pituitary-thyroid axis or continued TRAb stimulation, and fT3 may remain elevated, despite normalized fT4 levels. 34,35 Perhaps the relatively short follow-up period of this study was not optimal for evaluating the variability of TSH or fT3.
Lack of comprehensive information on pre-existing cardiovascular morbidity is a major limitation of this study, as pre-existing morbidity is a major risk factor for subsequent cardiovascular morbidity and could affect the results. Information was available on cardiovascular risk factors, which were entered as covariates in the multivariable analyses. The relatively short follow-up period and the rather small study cohort may be regarded as limitations to this study. The focus of this study was, however, on the short-term cardiovascular complications after the initiation of ATD therapy, and based on a previous study, the risk for acute cardiovascular events is highest during the first 3 months after the diagnosis of hyperthyroidism. 9 The first laboratory values measured in primary health care at the time of the diagnosis of hyperthyroidism or the doses of ATD therapy were not available in this study. Presumably, the most remarkable changes in the fT4 levels took place soon after the initiation of ATD therapy. As the initial TFT values were not available, this initial change in the TFTs was not included in the calculation of the CVs. Information on CVD-associated visits in primary health care was not available and this could underestimate morbidity, as less severe CVDs may have been treated solely in primary health care. Due to the study design, conclusions regarding the causality of the findings could not be drawn.
In conclusion, this is the first study to report that among hyperthyroid patients treated primarily with long-term ATD therapy, the variability of fT4 is associated with an increased cardiovascular morbidity. This finding emphasizes the importance of effectively controlled thyroid hormone levels during ATD therapy and early definitive treatment of hyperthyroidism if the response to ATD therapy is not optimal. Further studies are needed to verify the results of this study and to gain more insight into the clinical relevance of thyroid hormone variability and its possible associations with other adverse health outcomes, like impaired quality of life, or mortality.