Article Text

Download PDFPDF

Heart failure and cardiomyopathy
Stress hormone and circulating biomarker profile of apical ballooning syndrome (Takotsubo cardiomyopathy): insights into the clinical significance of B-type natriuretic peptide and troponin levels
  1. M Madhavan,
  2. B A Borlaug,
  3. A Lerman,
  4. C S Rihal,
  5. A Prasad
  1. Division of Cardiovascular Diseases and Department of Internal Medicine, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, USA
  1. Correspondence to Dr Abhiram Prasad, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA; prasad.abhiram{at}


  • Competing interests None.

View Full Text

Statistics from

Apical ballooning syndrome (ABS), also known as stress or Takotsubo cardiomyopathy, is characterised by transient regional systolic dysfunction of the left ventricle.1,2 Wall motion abnormalities typically involve multiple coronary territories in the absence of occlusive atherosclerotic disease. ABS is increasingly recognised as an important entity in the differential diagnosis of acute coronary syndromes, yet its pathophysiology remains uncertain.2,3 In approximately two-thirds of patients, an acute emotional or physical stressful trigger can be identified.4 As a corollary, stress-related neurohumoral factors, particularly catecholamines, are speculated to play an important part in the pathogenesis.5 However, studies to date have not uniformly shown elevated plasma catecholamine levels, and the diagnostic utility of measuring catecholamines in clinical practice is unknown.5,6,7,8,9 In addition, the role of other stress hormones, such as cortisol has not been studied in this disorder.

We hypothesised that stress hormones are elevated in ABS. Thus, the primary aim of this study was to investigate the levels of plasma and urinary catecholamines and cortisol in patients presenting with ABS. In addition, given the similarity in the presenting feature of ABS and acute myocardial infarction, we compared cardiac biomarker profiles—namely, plasma B type natriuretic peptide (BNP) and high sensitivity C-reactive protein (hsCRP) between patients with ABS and patients with ST-elevation myocardial infarction (STEMI).


Study population

This case-control study included 19 consecutive ABS patients identified prospectively, using the Mayo Clinic criteria, from April 2005 to June 2007.2 A convenience sample of 10 age-matched and gender-matched patients with STEMI (seven anterior and three inferior) served as controls. The study was approved by the Mayo Clinic institutional review board and all patients consented to the use of their medical record for research purposes.

Haemodynamic assessment

Coronary angiograms were performed on the day of hospitalisation in 17 ABS and all the STEMI patients and on the second day in two ABS patients. Measurement of left ventricular end-diastolic volume (EDV), end-systolic volume (ESV) and ejection fraction (EF) by left ventriculography was performed using the monoplane area-length method.

Transthoracic echocardiography was performed in all patients on hospital day 2 using standard methods.10 Follow-up echocardiograms were available in the patients with ABS at a median interval of 40 days (interquartile range 7–44 days) after presentation. Regional left ventricular wall motion abnormality was quantified by calculating the wall motion index score. Left ventricular meridional wall stress at end systole (ESWS) and end diastole (EDWS) was calculated as previously described.11 Left ventricular filling pressures were estimated using the ratio of early transmitral filling velocity to medial annulus tissue velocity (E/E′).12

Cardiac biomarker and neurohumoral measurements

Plasma levels of free fractionated metanephrines, including metanephrine and normetanephrine, were measured in all study subjects using liquid chromatography-tandem mass spectrometry on hospital days 1 or 2.13 Fifteen patients with ABS and nine with STEMI also had evening levels of serum cortisol measured using an automated chemiluminescent immunoenzymatic assay. Serial troponin T (Elecsys; Roche Diagnostics; Indianapolis, IN, USA) and creatine kinase isoenzyme (CK-MB) levels were measured at admission and daily thereafter, with peak values recorded. Plasma hsCRP (latex particle-enhanced immunoturbidimetric assay) and BNP (immunoenzymatic sandwich assay) were measured on hospital days 1 or 2.

A 24-hour urine sample was collected in all ABS patients on hospital days 1–3 for fractionated metanephrines, and analysed using stable isotope dilution liquid chromatography-tandem mass spectrometry.13 Twenty-four hour urine fractionated free catecholamines, including unconjugated epinephrine, norepinephrine and dopamine, were also measured in 16 ABS patients on the same urine sample using high-pressure liquid chromatography.14 Liquid chromatography-tandem mass spectrometry was used to measure free cortisol levels in the 24-hour urine samples of 13 ABS patients.15 All assays were performed at the Mayo Medical Laboratories in Rochester, Minnesota. The reference range for catecholamines and metanephrines were based on those established by the laboratory.16

Statistical analysis

Continuous variables are summarised as median and interquartile range and compared between groups using the Wilcoxon signed-rank test. Categorical variables are presented as percentages. Spearman’s correlation analysis was used to assess the linearity of the relation between two variables. A two-tailed p<0.05 was considered statistically significant. The performance of BNP, troponin T and the ratio of the two in discriminating between ABS and STEMI was assessed using receiver operating characteristic (ROC) analysis. The area under the curve (AUC) and the sensitivity and specificity at the optimal cut-off point are reported. All statistical analyses were performed using JMP (Version 7, SAS Institute Inc, Cary, NC, USA).


Clinical, angiographic and haemodynamic findings

The clinical characteristics of the patients with ABS (n = 19) and STEMI (n = 10) are summarised in table 1. Risk factors, baseline medication use and presenting symptoms were similar in ABS and STEMI. A stressful emotional (n = 5) or physical (n = 10) trigger was frequently present in ABS (79%).

Table 1

Baseline clinical characteristics

Coronary angiography did not reveal spontaneous coronary spasm, embolism, thrombus or plaque rupture in any patient with ABS. Thirteen (68%) patients had normal coronary arteries or mild atherosclerosis (<50% luminal narrowing). Five (26%) patients had >50% luminal narrowing in a branch vessel (diagonal or obtuse marginal artery) along with extensive regional wall motion abnormalities that extended beyond this territory. EF measured by left ventriculography in ABS was 35 (26–39)%, and left ventricular end-diastolic pressure was elevated at 26 (20–30) mm Hg. Single, double and triple vessel disease was present in 60%, 10% and 30%, respectively, in the STEMI control group.

Stress hormones in ABS

Plasma levels of fractionated metanephrine were normal (<0.5 nmol/l) in both groups: 0.10 (0.10–0.22) in ABS and 0.16 (0.10–0.38) in controls, p = 0.29. Plasma normetanephrine levels were normal (<0.9 nmol/l) in 14 of the ABS patients 0.64 (0.43–0.97) and nine of the controls 0.53 (0.32–0.77), p = 0.44; with mild elevation in the remaining patients. Twenty-four-hour urine levels of fractionated catecholamines and metanephrines were normal in all patients with ABS (table 2). The urine normetanephrine levels were mildly elevated in two patients with ABS and normal in the remaining subjects.

Table 2

Twenty-four hour urine stress hormones levels in apical ballooning syndrome

The evening plasma cortisol levels (normal range: 2–14 μg/dl) were elevated in eight patients with ABS 16 (7.3–44.0) and three controls 13 (10.5–23.5), p = 0.95. The 24-hour urine free cortisol levels were normal in all patients with ABS (table 2).

Cardiac biomarkers in ABS

Patients with STEMI had significantly greater elevations of cardiac troponin T and CK-MB compared to those with ABS (table 1). In contrast BNP levels were significantly increased in ABS, 944 (650–2022) pg/ml relative to STEMI controls, 206 (140–669) pg/ml (p = 0.009) (fig 1). Patients with ABS were characterised by a lower troponin and higher BNP levels compared to STEMI (fig 2). In the ROC analysis, the BNP level (AUC 0.84, optimal cut-off 647 pg/ml, sensitivity 81%, specificity 75%), and the ratio of BNP to peak troponin T levels (AUC 0.95, optimal cut-off 502, sensitivity 94%, specificity 100%) differentiated ABS from STEMI (fig 2).

Figure 1

High sensitivity C-reactive protein (hsCRP) (left) and B-type natriuretic peptide (BNP) (right) levels in patients with apical ballooning syndrome (ABS) and ST-elevation myocardial infarction (STEMI).

Figure 2

Troponin T and B-type natriuretic peptide (BNP) levels in patients with apical ballooning syndrome (ABS) and ST-elevation myocardial infarction (STEMI). Patients with ABS had low level elevation in troponin T and a large rise in BNP and hence predominantly located in the upper left quadrant. Conversely, patients with STEMI had high levels of troponin T and relatively less increase in BNP, and therefore were located in the lower right quadrant of the figure. The cut-off values of troponin T and BNP were selected arbitrarily, but with the aim of visually displaying the distinct cardiac biomarker profile of the two patient populations The inset demonstrates the ROC curves for BNP levels alone and the ratio of troponin T and BNP levels.

HsCRP levels were elevated in the majority of cases and controls, with a median value of 11.0 (5.1–110.8) mg/l in the ABS cohort and 24.3 (8.1–88.6) mg/l (p = 0.78) in STEMI controls (fig 1). There was no correlation between hsCRP and troponin T or BNP levels.

Correlations between plasma BNP and haemodynamic parameters in ABS

The EF and wall motion score index measured by echocardiography at presentation in ABS are summarised in table 3. At a median follow-up of 40 days, the respective values were 63% (62–65%), (p = 0.002) and 1.00% (1.00–1.06%), (p<0.0001). All parameters of left ventricular (LV) size, function and filling pressures at the index diagnosis were similar in patients with ABS and STEMI (table 3). There was a significant correlation between BNP levels and LV end diastolic volume (EDV) (r = −0.55, p = 0.04) in patients with ABS. However, there was no significant correlation between BNP levels and other haemodynamic parameters such as EDWS, ESWS, ejection fraction, LVEDP and E/e′.

Table 3

Echocardiographic-derived left ventricular (LV) haemodynamic parameters


This study provides a comprehensive analysis of stress hormone and cardiac biomarker profiles in ABS and compares these to values obtained from patients presenting with STEMI. The major findings are that: (1) despite the presence of significant antecedent stress in most patients, the vast majority had plasma and 24-hour urine catecholamine and metanephrine levels that fell within the normal ranges; (2) plasma cortisol levels were similar to patients with STEMI, while the 24-hour urinary free cortisol levels were within normal limits; (3) BNP levels were higher in ABS despite less myonecrosis, similar LV function and similar central haemodynamics; (4) the ratio of BNP to troponin T distinguished ABS from STEMI; and (5) there was a marked elevation in the inflammatory biomarker hsCRP, in a magnitude similar to that seen in patients with STEMI.

Stress hormones in apical ballooning syndrome

Enhanced sympathetic activity has been proposed as the central aetiology of myocardial dysfunction in ABS.5 However, the findings of the present study do not support this hypothesis, and indicate that marked elevation of catecholamine level may not be a universal finding. In our cohort, plasma levels of metanephrine and normetanephrine, metabolites of epinephrine and norepinephrine, were within the normal reference range in 100% and 74% of ABS patients, respectively. Overall, the plasma fractionated metanephrine levels in patients with ABS were similar to that in those with STEMI. The 24-hour urinary excretion of epinephrine, norepinephrine, dopamine and metanephrine were also within reference range in all patients with ABS and the urinary normetanephrine level was mildly elevated in 10%.

Earlier studies have produced conflicting results regarding catecholamine levels.6,7,8,17 The most convincing evidence supporting the catecholamine hypothesis comes from the study of Wittstein and colleagues, which demonstrated elevated plasma catecholamines, metanephrines, dihydroxyphenylglycol and dihydroxyphenylacetic acid in 13 ABS patients within 1–2 days of symptom onset, with elevation persisting for 5–7 days after the initial event.5 Moreover, these levels were significantly higher than a control population of patients with acute myocardial infarction and heart failure.

There may be several potential reasons for the divergent findings between the present study and that by Wittstein and colleagues. First, the elevation in catecholamines in our cohort may have been transient and undetectable by the time the patients arrived in the hospital. The median duration from symptom onset to admission was 4 hours (interquartile range 1–8 hours). However, the absence of an elevation in the 24-hour urine catecholamine levels in virtually all patients makes this less likely. Furthermore, the study by Wittstein and colleagues reported that plasma catecholamine levels remain elevated for many days following admission suggesting that any such elevations should have been detectable in our study population.5 Second, it is conceivable that patients with ABS are a heterogeneous group with more than one mechanism to account for the systolic dysfunction. Third, the variable findings may be attributed to the complexity of measuring catecholamines, differences in assays, and differences in cut-off values for what is considered normal. The samples for the current study were analysed at a national reference laboratory, using established partition values used to diagnose phaeochromocytoma.18 These cut-offs were used because ABS resembles the cardiomyopathy associated with phaeochromocytoma and as such should be associated with a major surge in catecholamines. Fourth, the nature of the precipitating stress may influence the catecholamine response. Wittstein and colleagues reported data from patients who had ABS related to an emotional stressor, while two-thirds of the current population had a physical stressor. Moreover, most studies have also not included a control population.6,7,17,19 Finally, plasma catecholamine and metabolite levels may not correlate with elevated local myocardial adrenergic activity.20

Animal experiments suggest that the glucocorticoids may also contribute to a stress-induced cardiomyopathy. Steroid pretreatment predisposes animals to cardiac contraction band necrosis and a cardiomyopathy when exposed to a variety of stressors.21 The data regarding cortisol levels in our study are novel and have not been previously reported in ABS. We found that the evening plasma cortisol levels were elevated in 53% of ABS patients, but importantly this was no different in magnitude from elevations seen in the patients with STEMI. This suggests that glucocorticoid activation in this setting is a non-pecific response to hamodynamic compromise. The normal 24-hour urine free cortisol levels further suggest that the activation of the corticosteroid system is not sustained beyond the acute period. Overall, our findings do not lend strong support for a primary pathophysiological role of the corticosteroid system in ABS.

Cardiac biomarkers in apical ballooning syndrome

BNP levels were markedly elevated in all patients with ABS, and threefold to fourfold greater than in those with STEMI. This was in contrast to the converse finding with regard to troponin T,which was mildly elevated in ABS and markedly so in STEMI. Also, the ratio of BNP to peak troponin T could accurately differentiate between ABS and STEMI (fig 2).

Volume and pressure overload increase left ventricular wall stress and are believed to be the predominant mechanisms for the upregulation of BNP synthesis and secretion from the ventricles.22 In the present study, the frequency of congestive heart failure, and indices of systolic (for example, EF, regional wall motion index and dimensions) and diastolic function as well as wall stress were similar in ABS and STEMI, and thus did not account for the differences in BNP levels. LV end diastolic volume, a marker for myocardial stretch, correlated with BNP levels, but in contrast to patients with systolic heart failure, there was no correlation between BNP and end-diastolic wall stress in ABS.22

C-reactive protein is a marker of inflammation and an acute phase reactant. Our study indicates that the myocardial stunning that occurs in ABS is associated with a systemic inflammatory response that is comparable to that associated with acute myocardial infarction. There was no correlation between the magnitude of myonecrosis and hsCRP in patients with ABS. We speculate that the inflammatory response in ABS is not directly due to the myonecrosis, but rather an epiphenomenon, given the very minor rise in the troponin and complete recovery of contractile function. The role of inflammation in the pathogenesis of ABS merits future study.


While the current sample size is small, it is comparable to earlier studies in which catecholamine levels were found to be markedly raised.5,19 This is a retrospective analysis and is subject to the limitations of such analyses, including the inability to render conclusions regarding the mechanisms for the elevations in BNP and hsCRP among patients with ABS. The control population of ST-segment elevation myocardial infarction was not randomly selected and hence subject to selection bias. Since blood levels of the hormones were measured several hours after the onset of symptoms, and serial measurement were not performed, it is possible that a very early transient hormonal surge or delayed rise was not detected.


Extreme elevations in BNP that are out of proportion to increases in troponin should increase clinical suspicion for ABS in patients presenting with symptoms suggestive of an acute coronary syndrome. Routine measurement of catecholamines in clinical practice is unlikely to be of diagnostic value among patients suspected of having ABS. The lack of an increase in stress hormones levels in our cohort suggests that, in addition to the role of catecholamines, it is important that future studies explore alternative mechanisms for the pathophysiology of ABS. Finally, a marked systemic inflammatory response occurs in ABS, similar to that seen in acute myocardial infarction, despite the absence of significant myocardial injury.


We thank Dr William F Young Jr, Professor of Medicine, Division of Endocrinology, Mayo Clinic, Rochester, for reviewing the data and his valuable suggestions regarding the interpretation of stress hormone levels.


View Abstract


  • Competing interests None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.