OUP user menu

The clinical features of takotsubo cardiomyopathy

Y.J. Akashi , K. Nakazawa , M. Sakakibara , F. Miyake , H. Koike , K. Sasaka
DOI: http://dx.doi.org/10.1093/qjmed/hcg096 563-573 First published online: 1 August 2003


Background: Cardiologists have recently recognized a reversible form of heart failure of unknown origin characterized by a takotsubo-shaped hypokinesis of the left ventricle on left ventriculography.

Aim: To clarify the clinical features of this cardiomyopathy.

Design: Observational study.

Methods: Seven patients with reversible ventricular dysfunction were followed for 4.5 years. Clinical course, routine examinations, and cardiac catheterizations in each patient were documented.

Results: The cardiomyopathy developed in six elderly female and one male patients (mean age 75.3 years), all of whom had been exposed to stress. Cardiac enzymes did not significantly increase, but serum norepinephrine increased remarkably (1.19 ng/ml). Coronary angiography revealed normal coronary arteries. However, left ventriculography showed akinesis in the apical segments, together with hyperkinesis in the basal segments (a takotsubo shape). The abnormal kinesis normalized within 17.4 hospital days without any treatment in five patients, and with haemodynamic support for 3 days in the other two. Endocardial biopsies did not suggest any specific pathology. The cardiac events did not recur over a 1–4 year follow-up.

Discussion: Coronary vasospasm, myocarditis and other substantial diseases previously described were ruled out as the cause of takotsubo cardiomyopathy in our subjects. Prognosis was good without any form of treatment, provided that the patients survived the severe heart failure state. Catecholaminergic or adrenoceptor-hyperactive cardiomyopathy may be the cause of this cardiomyopathy.


Recent reports have described a left ventricular dysfunction that resembles an acute myocardial infarction, but with normal coronary arteries and a shape resembling a takotsubo (a Japanese pot for fishing for octopus) on left ventriculography.119 This dysfunction, known as ‘takotsubo cardiomyopathy’,1319 characteristically resolves within a few weeks. However, its origin and detailed clinical features remain unknown. We analysed clinical data from seven patients with takotsubo cardiomyopathy who were admitted to our institution from 1998 to 2001.



A total of 472 patients with a sudden onset of heart failure, acute myocardial infarction-like abnormal Q-wave formation, and ST-T changes on the electrocardiogram (ECG) were admitted to our institution from January 1998 to August 2001. They were diagnosed as follows: 463 patients with acute myocardial infarction, seven (1.5%) with takotsubo cardiomyopathy (the subjects of this study) and two with acute viral myocarditis. Takotsubo cardiomyopathy was defined as: (i) the occurrence of heart failure similar to that observed with acute myocardial infarction; (ii) a takotsubo-shaped hypokinesis of the left ventricle on the echocardiography and left ventriculography; (iii) normal coronary angiography findings in spite of continuous ST-segment abnormalities in the ECG; and (iv) complete normalization of left ventricular dysfunction within a few weeks.

ECG and echocardiography

Standard 12-lead ECGs (25 mm/s paper speed, with a calibration of 10 mm = 1.0 mV) were recorded at the hospital admission, and at 6 h, 12 h, 24 h, 48 h, 72 h, 1 week, 2 weeks, and 4 weeks after admission. The locations of the precordial leads were marked with an indelible marker on the patient’s chest wall.

Two-dimensional echocardiography (SSH-140A, Toshiba, Tokyo, Japan) was recorded 2–3 times over the first week of hospitalization. The left ventricular ejection fraction (EF) was calculated by the single-plane area-length method with the apical two-chamber view, and expressed as a percentage.

Cardiac catheterization

Cardiac catheterization was performed within an hour of admission in all cases. In four patients, cardiac catheterization was performed when the dysfunction was determined to have normalized according to the echocardiography. Coronary angiography and left ventriculography were performed via six French catheters. In one case, right ventriculography was also performed. The left ventricular EF was calculated by the Simpson method. After the initial catheterization, a Swan-Ganz catheter was placed at a suitable position in each patient. Coronary angiography with acetylcholine provocation was performed in four patients. In one, it was done in the initial catheterization, and in the other three, in a later catheterization. Acetylcholine chloride, the test agent, was injected within 30 s. A 50 µg dose was used for the right coronary artery, and a 100 µg dose for the left coronary artery.

Haemodynamics, endocardial biopsy and blood tests

Circulatory haemodynamics were continuously observed throughout the clinical course in the coronary care unit. Endocardial biopsy was performed from the right ventricle apex in five patients (three on the first day in hospital, one on the third day, and one on the 18th day). However, histological examinations were performed only by microscope. Blood tests were performed on admission, and 1 h, 3 h, 6 h, 12 h, 24 h, 48 h, 72 h, 1 week, 2 weeks and 4 weeks following admission. All measured data were expressed as means ± SD.


The study protocol was approved by the Committee on Human Investigation of St Marianna University School of Medicine. The nature and purpose of the study and risks involved were explained, and written informed consent for participation in the study was obtained from all the subjects prior to their enrolment.


Clinical characteristics

Table 1 shows the clinical characteristics of the patients. Six of the seven were female, and the mean age was 75.3 years. One had a family history of hypertension. Past histories included hypertension in five, hyperlipidaemia in four, previous cerebral infarction in two, and a syncopal episode of unknown origin in one. Two also had a past history of heart disease: sick sinus syndrome in one and paroxysmal atrial fibrillation in the other. None had structural heart disease in their families or past histories.

View this table:
Table 1

Patients’ characteristics, family history and past history

Clinical features on admission

Table 2 shows the clinical features on admission. Five patients were admitted with chest pain, and the other two with loss of consciousness. We were able to identify possible triggers of the dysfunction in five: a common cold in two, pneumothorax in one, exercise in one, and ventricular fibrillation (VF) induced by idiopathic VF in the last. We learned of no events that could have possibly triggered the dysfunction in the other two. Patient 7 had a physically handicapped husband, and the burden of caring for him caused her severe physical and emotional stress.

View this table:
Table 2

Clinical features on admission


All of the cases exhibited an acute myocardial infarction-like change in ECG morphology, ST-segment elevation, Q-wave formation, and restricted presence of the resultant inverted T wave only in the chronic phase. Figure 1 shows a representative case. The ST-segment elevation with an abnormal Q wave persisted for 1 week, and thereafter changed to a non-Q and inverted T wave without any change in ST segment over the next 1 month.

Figure 1.

Progressive change in the electrocardiogram in patient 3. The initial electrocardiogram on admission showed ST elevation in the V1 to V4 leads. Twelve hours after admission, the ST level in leads II, III, and aVf was elevated and remained so for 2 weeks.

Table 2 shows the change in ECG morphology in each case. The change in ECG morphology was not specific to any of the leads. ST segment depression was observed in one patient, and ST segment elevation persisted for one week or more than a week in the remaining six. Ultimately, inverted T waves were observed in the leads that exhibited the ST segment change and abnormal Q wave formation. However, the resultant abnormal Q wave did not persist in all of the cases.

Blood tests

Table 3 shows that the data from the blood tests on admission and the maximum values of the cardiac enzymes. The first blood sample was taken 21.9 h after their symptom revealed. The white blood cell count on admission increased (> 8000/µl) in four patients. Cardiac enzymes, including creatinine kinase (CK) and its MB isoenzyme, did not significantly increase (MB/total CK < 10%) during the period of hospitalization, and nor did their peak value. Peak cardiac enzyme was predicted 27.5 h after admission. These laboratory findings normalized within 1 week in all of the cases.

View this table:
Table 3

Laboratory data on admission and peak cardiac enzyme

A troponin T increase (0.71 ng/ml) over the normal range (< 0.1 ng/ml) was observed. The plasma norepinephrine (NE) concentration increased in four patients (normal 0.24–0.57 ng/ml), and in each patient the plasma NE concentration was highest in the first blood sample. Moreover, a large increase in plasma brain natriuretic peptide (BNP) concentration (mean value 677.1 pg/ml, normal range < 18.4 pg/ml) was observed. The paired serum test for viral infection was negative in all cases.

Cardiac catheterization

Cardiac catheterization was performed when the change in ECG morphology stabilized in all cases. Coronary angiography revealed normal arteries without any stenosis or obstructions. Left ventriculography showed akinesis in the apical, diaphragmatic and/or anterolateral segments, and hyperkinesis in the basal segments. Figure 2 shows the left ventriculography for all of the cases, and Table 4 shows the haemodynamic data. The left ventricular EF, left ventricular end-diastolic pressure (LVEDP), pulmonary capillary wedge pressure (PCW), cardiac output (CO) and cardiac index (CI) were 39.7 ± 7.7%, 14.7 ± 6.0 mmHg, 9.3 ± 6.1 mmHg, 2.8 ± 1.3 l/min, and 2.0 ± 0.8 min/m2, respectively. Patient 7 also underwent a right ventriculography, which showed hypokinesis from the mid portion to the apical area (Figure 2h). The Swan-Ganz data indicated class III of the Forrester subset (Table 4).

Figure 2.

Left ventriculography in patients 1–7, all showing akinesis in the apical, diaphragmatic and/or anterolateral segments, and hyperkinesis in the basal segments (a–g). Right ventriculography showed hypokinesis from the mid portion to the apical region (h).

View this table:
Table 4

Initial cardiac catheterization data and supported device

Four cases underwent a second cardiac catheterization after showing improvement in the left ventricular wall motion. Data on these patients revealed normal coronary arteries and left ventricular wall motion. Induction of coronary arterial vasospasm by acetylcholine was negative in all four tested for this reaction. Patient 7 received this test on the first hospital day, and no vasospasm was induced. The other three cases underwent this test during the second catheterization, and none of them entered a vasospastic state (Table 5).

View this table:
Table 5

Clinical course and pathophysiological findings

Clinical course

Table 5 summarizes the clinical courses. None of the patients received oral medications such as digitalis, angiotensin-converting enzyme inhibitors, angiotensin-II receptor antagonists, or β-blockers during their hospital stays. However, patients 2 and 3 required haemodynamic support for 3 days. Normalization of the left ventricular contraction was complete by 17.4 ± 5.3 hospital days (EF 63.3 ± 7.0%). Several patients experienced episodes of, non-sustained ventricular tachycardia (patients 1, 3, and 7), paroxysmal atrial fibrillation (patients 2 and 6), and a pneumothorax (patient 4) during their hospital stays. In patient 3, the only patient requiring medication, intravenous magnesium sulphate (2 mg) was effective in suppressing the frequent occurrence of non-sustained ventricular tachycardia. The pneumothorax in patient 4 occurred spontaneously, and treatment with a chest tube resulted in complete healing within one week.

Patient 7 had a clinical course distinct from the other six cases. She was admitted for examination of repeated syncopal episodes of unknown origin. Obvious structural heart disease had not been demonstrated. Six hours after admission, syncope occurred. Based on the ECG monitoring, the cause of the syncope was diagnosed as polymorphic ventricular tachycardia with a short-coupled ventricular premature beat. The tachycardia spontaneously terminated 90 s later, and the echocardiogram revealed the development of a takotsubo-shaped left ventricular dysfunction. Ultimately, this resulted in a clinical course of takotsubo cardiomyopathy which normalized by the 14th hospital day without any treatment. Her tachycardia was diagnosed as the short-coupled variant of idiopathic VF, and an implantable cardioverter defibrillator was implanted.

During a follow-up period of 1 to 4 years, patient 1 died from non-cardiac causes, 2 years after the episode of cardiomyopathy, and the other six cases had not had any cardiac events.

Endocardial biopsy

As summarized in Table 5, the pathological findings did not suggest any specific heart disease.


Basis of takotsubo cardiomyopathy

There have been a number of clinical reports with findings that mimic acute myocardial infarction with reversible left ventricular dysfunction,119 especially in Japan.1,3,4,6,8,1219 In the literature, this disease appears to account for about 1% of all acute myocardial infarctions,3,14 and the incidence in our institution was similar. The clinical features common to these reports can be summarized as follows: (i) most patients are women aged > 60 years; (ii) symptoms at onset are similar to those of acute myocardial infarction; (iii) the ventricular dysfunction is shaped like a takotsubo; (iv) the coronary arteries are not diseased, (v) the dysfunction improves rapidly within a few weeks; and (vi) data on the recurrence of this cardiomyopathy have not been documented. Its origins are unclear.

Our findings are consistent with those in previous studies: (i) this reversible left ventricular dysfunction occurred suddenly in structurally normal hearts in (mostly) elderly females; (ii) all cases experienced severe stress prior to the episodes; (iii) a high level of serum catecholamines was revealed; (iv) the cases were confirmed to have normal coronary arteries, ruling out any relation to coronary vasospasm; (v) myocarditis was ruled out; (vi) the dysfunction improved to a normal value in a mean of 17.4 hospital days without any treatment, after the patients made it past the severe left ventricular dysfunction in the acute phase; and (vii) this cardiomyopathy appeared to have a good prognosis without any treatment.

Noteworthy clinical features observed in the present study were the high concentrations of catecholamine to which the patients were exposed, the lack of coronary vasospasm or myocarditis as a precipitating factor, and the complete recovery without treatment. This type of LV dysfunction has been reported to develop concomitantly with many different types of disorders, including coronary arterial spasm,2,3 phaeochromocytoma,57 Guillain-Barré syndrome,8 emotional stress,9,22,23 endocrine disease,10,20,21 pneumothorax,17 and subarachnoid haemorrhage.24,25 However, the relationship to these diseases is unclear. Our subjects were free from any serious concomitant disorders, although we identified events before the onset of the heart failure that may have triggered this cardiomyopathy in five patients: a common cold, pneumothorax, exercise, emotional stress, and a VF attack. In particular, we believe that stressful situations may trigger this cardiomyopathy. The very high serum catecholamine concentration observed in some of our cases supports this hypothesis. Owa et al.15 reported that the recovery of the I-123-metaiodobenzyl-guanidine (123I-MIBG) myocardial uptake was latest in some tracers, and furthermore, they suggested that the origin of TC was related to a disturbance of the cardiac sympathetic innervation. Our institution was able to observe the same 123I-MIBG findings in some of the patients with TC.13,17,18 The systematic catecholamine increase may have been one of the primary results of this heart failure. On the other hands, Kurisu et al.16 commented in their literature that circulating plasma NE concentration was almost normal in patients with takotsubo cardiomyopathy. We hypothesized that excessive activation of cardiac catecholamine receptors may have been another result of this heart failure, but it may also have been the origin of the cardiomyopathy.

The most well known origin of reversible ventricular dysfunction is the myocardial stunning phenomenon in ischaemic heart disease. Dote et al.3 suggested that emotional stress might result in an altered coronary vasospasm that could potentially elicit an reversible dysfunction. However, in our subjects, by the coronary angiography performed during the ST-segment elevation did not uncover any signs of coronary vasospasm. We consider myocardial stunning due to coronary vasospasm unlikely to be the origin of this cardiomyopathy in our subjects.

Clinical course

Our initial laboratory data were similar to those on an acute myocardial infarction. In most of our cases, cardiac enzymes were slightly increased on admission, and slightly further in later examinations. While this increase in the cardiac enzymes might have reflected injury to the myocytes, the peak value was too low to indicate the development of a myocardial infarction or fulminant myocarditis in any of our subjects. The ST segment change was exhibited in various leads of the ECG, and did not suggest a specific coronary arterial obstruction. Observing such an ST-segment elevation in patients with Guillain-Barré syndrome, Iga et al.8 proposed that the change was elicited by an immune response associated with an infection. However, in our cases, viral myocarditis was not found in blood tests or paired serum examinations for the various viruses.

The catecholaminergic cardiomyopathy or high levels of norepinephrine also indicated the origin of the ST segment deviation in the ECG.58 Catecholamine-induced ST-segment deviation has also been observed in an animal model.26 An increase in the plasma norepinephrine concentration was observed on admission in four of our patients. Catecholamines can directly injure the myocardium,27,28 and brain-mediated cardiac injury can be prevented experimentally by total cardiac sympathectomy.29 The catecholamine increase in our initial findings may have been due to heart failure. However, a catecholamine increase could not be ruled out as one of the origins of the cardiomyopathy, as there was no relationship between the plasma norepinephrine value and the markers of cardiac function, and most of the subjects were under some stress prior to the event. After confirming an absence of catecholamine-producible organic disease (i.e. phaeochromocytoma or Sipple syndrome) through examinations, we suspected an abnormal response in the myocardium. Ueyama et al.30 have recently shown transient left ventricular hypocontraction via excessive activation of cardiac adrenoceptors in an animal model. We support their hypothesis for the model of takotsubo cardiomyopathy.

The BNP level was remarkably elevated in the acute phase. Generally, a high level of serum BNP is a marker of poor prognosis,31 and there are several reports in which BNP was correlated with LVEDP32 or EF.33 Although very high levels of plasma BNP were observed in our subjects, the prognosis was good. Our patients’ BNP levels in the chronic phase were not evaluated.

Pathology observed in this cardiomyopathy

We were not able to find a pathological cause for this cardiomyopathy from our microscopic examinations. Pathologically focal myocytolysis is reported as an origin of this cardiomyopathy.14 In our subjects, paired-serum examinations were negative for any kind of bacteria or virus, including Mycoplasma and Chlamydia pneumoniae. While our specimens did show a mild interstitial change, we could not confirm the presence of a diffuse or narrow range of necrosis, and thus rule out viral myocarditis. Some papers have reported coronary vasospasm as an origin,3,14 whereas Kloner et al.34 reported that a pathological change in the myocardium was not demonstrated in the stunned myocardium. As infiltration of small mononuclear cells was documented in some cases; these pathological findings suggested that this cardiomyopathy was some kind of inflammatory heart disease, but not a coronary heart disease. There is also a report describing histological myocardial damage in fatal cases without coronary heart disease.35

Study limitations

This study could not be designed as a large prospective study because of the small number of patients and rarity of the disease. Further studies involving a larger number of institutions will be needed.

We have to handle our takotsubo cardiomyopathy cases differently from Syndrome X or microvascular angina patients.3641 Most of our patients were too old to undergo exercise testing, hence it was impossible to evaluate whether Syndrome X was involved. From the onset of this cardiomyopathy, measured concentrations of norepinephrine were high. In addition to our patients’ status as elderly females, there might have been a severe expression of microvascular angina in these cases.

In the present study, a high level of plasma catecholamine concentration, or an activation of cardiac catecholamine receptors was suspected as one of the factors related to the development of this cardiomyopathy. As our study failed to elucidate why the asynergy was regional, the distribution of catecholamine receptor density in the myocardium should be examined to illuminate the mechanism of the regional dysfunction.


Takotsubo cardiomyopathy is a distinct type of heart failure. Although its origin was not clearly demonstrated, coronary vasospasm, acute myocarditis, and other substantial diseases that have been previously described were ruled out as causes. Heart failure in this cardiomyopathy completely reversed without treatment, and had a good prognosis once past the severe state. This cardiomyopathy should be noted as a cause of sudden cardiac death in individuals without obvious heart disease. Although catecholaminergic or adrenoceptor-hyperactive cardiomyopathy is a possible instigating factor, further studies are needed to confirm this.


View Abstract