Abnormal thyroid function is common in takotsubo syndrome and depends on two distinct mechanisms: results of a multicentre observational study

Introduction

‘Takotsubo syndrome’ (TTS), Takotsubo cardiomyopathy or’Stress-induced cardiomyopathy’ is a rare disorder characterized by left ventricular dysfunction with typical patterns of regional wall motion abnormalities and myocardial ischaemia in the absence of obstructive coronary arteries. Clinical presentation often mimics acute coronary syndrome (ACS) with acute chest pain and ST-segment elevations accompanied by elevated biomarkers of acute myocardial injury. Therefore, most TTS patients undergo diagnostic coronary catheterization to exclude a culprit coronary artery lesion. Recovery of left ventricular function is observed in the majority of patients, and the rate of recurrence of TTS is low [1, 2].

TTS may be accompanied by significant complications that include cardiogenic shock, left ventricular thrombus formation and development of life-threatening tachyarrhythmias [3-5].

Severe emotional or physical stress preceding the onset of TTS is the most commonly recognized trigger in approximately two thirds of patients [6-8]. Furthermore, TTS affects predominantly postmenopausal women; hence, alterations in hormonal regulation are discussed to be involved in pathogenesis [9-12].

Activation of the adrenergic system seems to play an essential role in the development of TTS [13]. Studies showed significant elevation in plasma catecholamine levels in patients with TTS when compared to those with ACS [14, 15]. Paur et al. reported high levels of epinephrine to trigger apical cardiodepression in an in vivo rat model via specific beta-adrenoceptor signalling pathways, which may have evolved as a cardioprotective strategy limiting catecholamine-induced myocardial toxicity during acute stress [16]. Recently, an alteration (desensitization and subtype switching) of beta-adrenoceptors in human myocardial tissue was demonstrated in TTS, which was interpreted as a protection against catecholamine toxicity [17].

Several case reports have suggested other triggers of TTS including hyper- and hypothyroidism [11, 18]. The most frequent thyroid alteration linked to TTS is thyrotoxicosis due to various disease entities [19-22]. Previously Miyazaki et al. visualized myocardial stunning by scintigraphy in a patient with TTS and concomitant thyrotoxicosis [23]. Interestingly, in TTS the odds ratio of hypothyroidism seems to be increased, too [24].

Although the exact pathogenesis of this relationship is not yet fully understood, it has been suggested that thyroid hormones might sensitize the myocardial tissue for the effects of catecholamines [23, 25, 26]. This would explain an increased risk of TTS in stress situations in the presence of elevated thyroid hormone concentrations.

In order to obtain more comprehensive insights into the relationship between thyroid homeostasis and TTS we analysed thyroid hormone profiles during inpatient stay in the acute phase of TTS and compared the findings to results obtained in patients with acute coronary syndrome and in normal controls.

Methods

Results

Patient characteristics

After excluding 9 subjects (36%) due to use of levothyroxine or levothyroxine and amiodarone (Fig. 1) 16 patients with TTS were eligible for inclusion. Their mean ejection fraction (EF) was 42 ± 3%, and concentrations of CK, CK-MB and troponin I were 367.9 ± 116.5 U L⁻¹, 36.8 ± 3.6 U L⁻¹ and 111.7 ± 106.9 µg L⁻¹, respectively. 7 (44%) of the eligible subjects suffered from concomitant hypertension, 4 (25%) from atrial fibrillation, 7 (44%) from thoracic pain, and 1 (6%) were smokers. One of the excluded subjects was additionally on concomitant treatment with amiodarone.

An overview of clinical characteristics of the study population is reported separately for the groups in Table 1.

Table 1. Clinical characteristics of the study population

	Normal controls from NHANES sample (n = 16)	Acute coronary syndrome (n = 16)	Takotsubo syndrome (drug-naïve, n = 16)a
Age, years	65.3 ± 2.0	68.6 ± 3.0	66.3 ± 2.5
Sex
Female	15	15	15
Male	1	1	1
Body mass index, kg m−2	29.8 ± 1.8	29.3 ± 1.7	28.0 ± 2.3
Systolic blood pressure, mmHgb	138 ± 6	123 ± 7	137 ± 11
Diastolic blood pressure, mmHgb	71 ± 3	61 ± 6	86 ± 7**
Heart rate, min-1b	72 ± 3	81 ± 3	95 ± 6*,***
Thyroid diseases
Autoimmune thyroiditis	0	0	0
Radioiodine therapy	0	0	0
Thyroidectomised	0	0	0
Nodular goitre	0	0	2
Hyperthyroidism	0	0	1
Diabetes	1 (6%)	3 (19%)	3 (19%)
COPD	1 (6%)	2 (13%)	3 (19%)
Levothyroxine usage	0	0	0
Amiodarone usage	0	0	0
Beta-blocker usage	3 (19%)	0 (0%)	6 (38%)
acetylsalicylic acid (ASA) therapy	0 (0%)	3 (19%)	5 (31%)*
Anticoagulation	0 (0%)	15 (94%)**	1 (6%)****
Length of stay in intensive care unit, days	0 ± 0	5.63 ± 0.65**	3.87 ± 1.20*

• Data are reported as n or mean ± SEM.
• ^a Subjects receiving neither levothyroxine or amiodarone.
• ^b First measurement after admission or inclusion, resp.
• * P < 0.05 compared to normal controls.
• ** P < 0.001 compared to normal controls.
• *** P < 0.05 compared to ACS.
• **** P < 0.001 compared to ACS.

Phenotype of thyroid function

In the first measurement at or after admission, four of the 16 drug-naïve subjects with TTS (25%) suffered from subclinical (n = 2) or overt primary thyrotoxicosis (in three of the four cases blood sampling took place before the application of iodinated radiocontrast agents). One additional subject each developed subclinical or overt thyrotoxicosis, respectively, during the course of treatment. In four subjects FT4 concentrations were initially increased, despite nonsuppressed TSH concentration, representing an elevated set point of thyroid homeostasis. Two subjects presented with subclinical hypothyroidism. In total, six subjects demonstrated elevated FT4 concentration in the first measurement and another fraction of six subjects (38% each) had normal thyroid function during the time course of observation.

One of the two patients with known nodular goitre presented with subclinical thyrotoxicosis. The second one demonstrated initially an elevated set point but developed primary hyperthyroidism during the stay in the hospital.

In three of the subjects with TTS (19%), an elevated set point evolved from overt thyrotoxicosis, subclinical thyrotoxicosis and euthyroidism, respectively, during the seven days after admission. From the four subjects with an elevated setpoint in the initial evaluation, one developed overt thyrotoxicosis and one subclinical hypothyroidism during follow-up.

Compared to normal controls, mean FT4 concentrations and SPINA-GT were repeatedly elevated in both ACS and TTS groups, whereas FT3 and SPINA-GD were in both groups reduced (Table 2). Immediately after admission (before the subjects received radiocontrast agents) and through the first three days the set point of thyroid homeostasis was elevated in subjects suffering from TTS, as evaluated by both TTSI and TFQI. Mean TTSI was temporarily reduced in ACS but elevated in TTS. 24 and 72 h after admission to the ICU TFQI was significantly increased in TTS (Table 2, Table S1).

Table 2. Endocrine parameters in normal controls and drug-naïve patients with acute coronary syndrome and Takotsubo syndrome. Results from day 0 are from admission, before radiocontrast agents were applied

Parameter (reference range), UoM	Normal controls from NHANES sample (n = 16)b	Acute coronary syndrome (n = 16)	Takotsubo syndrome (drug-naïve, n = 16)a
TSH (0.35–3.5), mIU L−1	1.60 ± 0.17
Day 0		N/A	1.77 ± 0.40
Day 1		0.62 ± 0.12*	1.21 ± 0.45
Day 3		1.65 ± 0.50	2.11 ± 0.85
Day 7		2.27 ± 0.56	1.21 ± 0.45
FT4 (7.6–16.1), pmol L−1	10.3 ± 0.4
Day 0		N/A	15.4 ± 0.8**
Day 1		14.0 ± 0.9**	16.6 ± 0.6**
Day 3		13.2 ± 1.0*	16.7 ± 1.2**
Day 7		14.3 ± 1.3*	15.3 ± 1.5*
FT3 (4.0–5.7), pmol L−1	4.6 ± 0.1
Day 0		N/A	4.8 ± 0.2
Day 1		3.2 ± 0.1**	3.8 ± 0.4*
Day 3		3.2 ± 0.3**	3.7 ± 0.4*
Day 7		3.9 ± 0.4	3.2 ± 1.5*
SPINA-GT (1.41–8.67), pmol s−1	2.45 ± 0.23
Day 0		N/A	6.23 ± 1.56*
Day 1		9.40 ± 2.57*	11.80 ± 3.61*
Day 3		4.25 ± 0.88	4.45 ± 1.63*
Day 7		5.85 ± 2.76	6.68 ± 3.10
SPINA-GD (20–60), nmol s−1	42.2 ± 1.8
Day 0		N/A	28.8 ± 1.1**
Day 1		22.2 ± 1.5**	21.4 ± 1.9**
Day 3		23.7 ± 2.2**	20.5 ± 1.2**
Day 7		24.2 ± 1.2**	18.7 ± 7.3**
TTSI (100–150)	101.9 ± 11.0
Day 0		N/A	165.5 ± 46.0
Day 1		52.9 ± 11.4	123.4 ± 47.5
Day 3		134.6 ± 38.8	224.7 ± 88.4
Day 7		199.8 ± 44.9 *	92.5 ± 64.2
TFQI	–0.08 ± 0.08
Day 0		N/A	0.32 ± 0.10*
Day 1		–0.03 ± 0.07	0.24 ± 0.10*
Day 3		0.16 ± 0.15	0.39 ± 0.17*
Day 7		0.28 ± 0.17	0.19 ± 0.20

• Data are reported as n or mean ± SEM.
• ^a subjects receiving neither levothyroxine or amiodarone.
• ^b Single measurement available only.
• * P < 0.05 compared to single measurement in normal controls.
• ** P < 0.001 compared to single measurement in normal controls.
• *** P < 0.05 compared to ACS.
• **** P < 0.001 compared to ACS.

Based on the TTS cohort, the PAM algorithm identified two clusters in the affine space of TSH and FT4 concentrations with marginally high and low TSH concentrations, respectively, and slightly to moderately elevated FT4 concentrations (cluster 1: TSH 3.47 mIU L⁻¹, FT4 14.4 pmol L⁻¹; cluster 2: TSH 0.31 mIU L⁻¹, FT4 16.7 pmol L⁻¹, Fig. 2a).

Accordingly, the dispersion of most parameters was higher in TTS than in the other two investigated groups (Table 2). This was reflected by the fact that the proportion of endocrine parameters outside their respective reference ranges was for the majority of time points higher than expected (Fig. 3). Especially for TSH and TTSI, both reduced and elevated results were common.

A bivariate representation (Fig. 4) demonstrates that normal, high-normal and elevated concentrations of FT4 were accompanied by either rather low or high TSH concentrations, suggesting two distinct aetiological factors of high FT4 concentrations. Although the ML algorithm had no information about diagnoses the identified clusters were partly overlapping with the diagnosis groups (Fig. 2b).

Discussion

Despite the rareness of TTS, we were able to include 16 subjects and to compare their thyroid function to age- and sex-matched normal subjects and patients with ACS.

In our study, we observed in the majority of patients affected by TTS abnormal thyroid hormone laboratory test results. In order to interpret these results in a clinical context, we performed a categorization dependent on thyroid hormone characteristics.

36% of all subjects in the screening phase received levothyroxine treatment, so that they had to be excluded from analysis. This fraction is substantial and much higher than the expected population prevalence of 5–10% [35], but similar to previous reports on TTS [9, 12].

The finding that four patients suffered from subclinical or overt thyrotoxicosis suggests a possible triggering mechanism for the development of TTS, which would be in accordance with previous studies and case reports observing an association between TTS and thyrotoxicosis [11, 18].

Up to now, the distinct mechanisms by which thyrotoxicosis is able to elicit TTS remains poorly understood. Previous reports described TTS in thyrotoxicosis of different origin including true endogenous hyperthyroidism, destructive thyroiditis and factitious thyrotoxicosis [11, 19-23].

The sympatho-adrenergic system plays a pivotal role in the pathogenesis of TTS [13]. This applies to psychosocial stressors, physical strain and autonomous production of catecholamines by pheochromocytomas and paragangliomas, as demonstrated in a recent meta-analysis [50]. Thyroid hormones are able to sensitize the heart for catecholamines by stimulating the expression of beta-adrenoceptors in cardiomyocytes, with subsequent positive inotropic and chronotropic effects [25, 35-37]. Therefore, an excess of thyroid hormones might potentiate the effects of catecholamines in the myocardial tissue with resulting increased sensibility to stress events and heart stunning [38-40]. Conversely, intracellular cAMP formed by activated beta-adrenoceptors is able to stimulate the expression of a number of genes including that of type 2 deiodinase (DIO2), which converts T4 to T3 and 3,5-T2, that is to more active thyroid hormones [41]. Although not measured here, this may be an important mechanism in the pathogenesis of stress cardiomyopathy, the more as TSH is able to stimulate cAMP formation in various tissues, too [44]. Therefore, in the combination of stress with elevated concentration of T4 a positive feedback loop between catecholaminergic signalling and activation of thyroid hormones may arise, eventually igniting the transition to TTS.

The results show that somatic and especially endocrine factors may play a more profound role for the pathogenesis of TTS than previously assumed. In addition to the well-documented association of catecholamine excess and TTS events [50] our observations challenge the Mayo criteria requiring the exclusion of pheochromocytomas or paragangliomas [51].

In four subjects (in three cases before application of radiocontrast agents, in one immediately afterwards) FT4 concentrations were temporarily increased, despite nonsuppressed TSH concentration. This pattern does not result from primary hyperthyroidism but marks the constellation of an elevated ‘set point’ of the homeostatic system. An elevated set point is characterized by an increased or at least nonsuppressed TSH concentration despite an elevated FT4 concentration [42]. Temporarily or permanently elevated set points result from increased TSH secretion due to elevated hypothalamic TRH input. They have been described in conditions associated to type 2 allostasis including obesity, acute psychosis, endurance training, adaptation to cold and post-traumatic stress disorder (PTSD) [43].

The hypothalamic–pituitary–thyroid (HPT) axis acts as an adaptive, dynamic system, which in unstrained resting conditions operates as a homeostatic regulator, aiming at constant value control and maintaining serum concentrations of thyroid hormones in the vicinity of a fixed set point [44]. However, in type 2 allostatic load resulting from an expected increase in energy demand, although the cumulative energy balance being still sufficient [45], elevations of the set point of the hypothalamus–pituitary–thyroid (HPT) axis are common. This constellation is typical in psychosocial stress situations [43]. Of note, psychosocial stress is known as the main trigger for TTS, and its effect may be mediated by altered thyroid hormone concentrations consistent with type 2 thyroid allostasis. As previously explained type 2 allostatic load may even increase the sensitivity of the heart to catecholamines as fast mediators of the stress response.

A recent publication by Templin et al. demonstrated hypoconnectivity of central brain regions associated with autonomic functions and regulation of the limbic system in patients with TTS [46]. The authors concluded that autonomic-limbic integration and particularly hypothalamic regions might play an important role in the pathophysiology of TTS. A possible linkage between type 2 thyroid allostasis and TTS could be given by an interaction between the so-called brain-heart axis and the HPT axis. More carefully planned prospective studies are necessary to test this hypothesis.

The common end point of primary hyperthyroidism and a high set point of the homeostatic system is an elevated or at least high-normal free T4 concentration. In cases with high-normal FT4 levels sensitization of myocardial tissue to catecholamines may result from hormone concentrations above the personal set point of thyroid homeostasis [47]. Recently, a similar pattern of dichotomous TSH levels but consistently high FT4 concentrations was described to be associated to malignant arrhythmia in patients with structural heart disease [48]. This is consistent with the results of large population-based studies [55, 56].

In summary, abnormal thyroid hormone concentrations seem to be the rule rather the exception in TTS. In addition to primary thyrotoxicosis, an elevated set point for serum T4 concentrations in type 2 thyroid allostasis may be a major contributor to the pathophysiology of this still understudied condition. If confirmed by additional studies, further research may be warranted to define possible types of TTS (‘endocrine-type’ or ‘stress-type’).

Surrogate markers for pathogenesis and prognosis of TTS are highly needed in order to facilitate personalized treatment strategies that are tailored to the individual requirements of affected patients. This applies especially to the selection of beta-blockers, a still controversial issue [49]. In cases of elevated or even high-normal T4 concentrations, beta-blocker therapy may be beneficial by disrupting the suggested positive feedback loop between thyroid hormone activation and up-regulated beta-adrenoceptors. Additionally, beta-blockers (at least the nonselective variants, e.g. propranolol, sotalol and pindolol) may inhibit step-up deiodination resulting in the conversion of T4 to the more active hormone T3.

As a study limitation, we want to spotlight the use of two different laboratory analysis devices in the cohort, which is a consequence of the two-centre design of this investigation. However, TSH assays are calibrated with a WHO standard, and hormone concentrations did not differ between the used assays.

An additional limitation may result from the low number of TTS patients in this two-year period. Unfortunately, low sample sizes are nearly unavoidable in the study of rare diseases. This limits the conclusion to be drawn from the data. Long-term, register-based multicentre clinical studies and animal experiments may help to re-evaluate the findings of the reported results and to provide additional and more robust insights. Additionally, more elaborate machine learning techniques and complexity measures may help to find additional patterns and trends in existing data, provided that the data base is large enough.

All subjects with TTS and ACS had received iodinated radiocontrast agents during inpatient treatment. This may impair the interpretation of the results since SPINA-GT may be increased (e.g. in toxic adenoma and multinodular goitre or Graves’s disease) or decreased (via Plummer and Wolff–Chaikoff effects) after exposure to high iodine doses. Since we observed increased FT4 concentrations in the TTS cohort even directly after admission and since differences to patients with ACS persisted during follow-up this doesn’t seem to be a major confounder, however.

To our knowledge, the present study covers the largest cohort of TTS patients in which thyroid hormones were examined and evaluated in detail up to now.

Funding

This work did not receive external funding.

Conflict of interest

The authors declare that they have no conflict interest.