OUP user menu

Arrhythmia occurrence with takotsubo cardiomyopathy: a literature review

Faisal F. Syed, Samuel J. Asirvatham, Johnson Francis
DOI: http://dx.doi.org/10.1093/europace/euq435 780-788 First published online: 3 December 2010


Aims Takotsubo cardiomyopathy (TC) or the apical ballooning syndrome is a reversible cardiomyopathy mimicking acute myocardial infarction (AMI). Although malignant arrhythmia is considered less likely to occur in TC than with AMI, sporadic reports of malignant arrhythmia with TC, however, have been reported. We reviewed the medical literature on TC and arrhythmias and describe in the summary the reported findings and discuss possible specific scenarios where arrhythmia may be more likely in patients with TC.

Methods and results Articles were identified on PubMed using the MeSH terms ‘Takotsubo Cardiomyopathy’ or ‘Apical Ballooning Syndrome’. Seventy-four unique case series with five or more TC patients were identified, with a cumulative total of 1876 cases. Twelve series (242 cases) were excluded because Mayo criteria were not met. Twenty-five series (816 cases, 43.5%) reported on arrhythmia and were included in the analysis.

Conclusion Areas for further prospective study include the duration and mechanism of residual risk and optimal risk stratification and modification. The current state of evidence would support beta-blockers in the acute setting to control tachyarrhythmia, although there is little evidence to support their use beyond convalescence when used for this indication alone. Those in heart failure and cardiogenic shock should be managed with established evidence-based therapies for these conditions. The use of internal defibrillators in TC requires consideration only on a case-by-case basis.

  • Arrhythmia
  • Tachycardia
  • Bradycardia
  • Takotsubo cardiomyopathy
  • Apical ballooning syndrome


Takotsubo cardiomyopathy (TC), or the apical ballooning syndrome, is a form of reversible cardiomyopathy mimicking acute myocardial infarction (AMI) in both symptomatology and electrocardiographic findings but without significant coronary artery disease on angiography.1 Excessive sympathetic discharge is considered to be a potential primary pathogenic factor.

One of the original descriptions was from Japan in 1991, when Dote et al.2 described myocardial stunning due to multivessel coronary spasms in five cases. There was basal hyperkinesis and akinesis of apical, diaphragmatic, and/or anterolateral segments on left ventriculography. The name takotsubo cardiomyopathy was given to this entity later because of the resemblance of the appearance of the left ventriculogram to the traditional Japanese octopus trap or takotsubo. The name ampulla cardiomyopathy has also been used in reference to this entity (Figure 1).

Figure 1

Fluoroscopic right anterior oblique projections demonstrating typical angiographic findings in takotsubo cardiomyopathy. (A): left ventriculogram in diastole; (B): left ventriculogram in systole; (C): Left coronary angiogram.

Little is known about the risks and outcomes associated with arrhythmia in the setting of TC. The electrocardiographic ST-T-wave changes that accompany the presentation of TC are often associated with prolongation of the QT interval (Figure 2), yet paradoxically arrhythmias are reported to be uncommon.3 When arrhythmias do occur, they may be an important determinant of a short-term outcome. Understanding the arrhythmic substrate and triggering factors, and identifying the patient at risk of life-threatening arrhythmia in TC would allow an approach to risk stratification and tailored therapy. This article aims to summarize the prevalence and outcome of arrhythmias reported in TC based on a systematic review of published case series.

Figure 2

Electrocardiogram corresponding to Figure 1. Electrocardiogram demonstrates evolving repolarization abnormality in association with QT-prolongation.


Articles were identified on PubMed using the MeSH terms ‘Apical Ballooning Syndrome’ and ‘Takotsubo Cardiomyopathy'. Case series were selected with reporting on five or more cases. Proposed Mayo Clinic criteria were used to define the diagnosis of TC (Table 1).1 Manual screening of references of selected articles was used to identify additional case series. When a patient population was reported on in separate publications, the most recent publication was selected. The following data were extracted as available: number of cases; number with each of the following arrhythmias during index hospital presentation: ventricular fibrillation (VF), sustained and non-sustained ventricular tachycardia (VT), ventricular asystole, atrioventricular (AV) block, sinus node dysfunction (SND, to include descriptions of sick sinus syndrome, sinus pauses, and sinus bradycardia), and atrial fibrillation (AF); number of sudden cardiac deaths (SCD), and death from other causes. Cumulative prevalence and hospital mortality rates were calculated. Clinical events occurring within 1 week of discharge were considered as arising within the hospital episode.

View this table:
Table 1

Proposed Mayo Clinic criteria for the diagnosis of takotsubo cardiomyopathy

Transient hypokinesis, akinesis, or dyskinesis of the left ventricular mid-segments, with or without apical involvement. Regional wall motion abnormalities extend beyond a single coronary vascular bed.
No obstructive coronary disease or acute plaque rupture (determined angiographically)
New electrocardiographic abnormalities (ST-segment elevation, T-wave inversion, or both) or modest elevation in cardiac troponin level
No pheochromocytoma or myocarditis.


Seventy-four unique case series with five or more TC patients were identified, with a cumulative total of 1876 cases (Table 2). Twelve series (242 cases) were excluded because Mayo criteria were not met. Twenty-five series (816 cases, 43.5%) reported on arrhythmia and were included in the analysis (Table 3). The studies either recruited consecutive cases in a prospective manner,411 or identified them retrospectively from a cohort of those presenting as acute coronary syndromes or myocardial infarction at their institutions.

View this table:
Table 2

Study selection

ReferenceCityYearNumber of cases
 1Tsuchihashi et al.15Sapporo, Morioka, Sendai, Omiya, Yokohama, Matsumoto, Osaka, and Kumamoto, Japan200188
 2Kurisu et al.12Hiroshima, Japan200230
 3Akashi et al.16Kawasaki, Japan20037
 4Desmet et al.13Leuven, Belgium200313
 5Matsuoka et al.41Matsusaka, Mie, Japan200310
 6Wittstein et al.4Baltimore, MD, USA200519
 7Hertting et al.28Hamburg, Germany200632
 8Bonnemier et al.5Lubeck, Germany200624
 9Patel et al.24Philadelphia, PA, USA20075
 10Fujiwara et al.6Yamagata, Japan200711
 11Santos et al.19Faro, Portugal20075
 12Bonello et al.17Marseille, France200814
 13Dib et al.18Rochester, MN, USA2008105
 14Buja et al.7Venice, Italy200828
 15Kume et al.9Kurashiki, Japan20085
 16Mitsuma et al.23Niigata, Japan200821
 17Yoshioka et al.8Hokkaido, Japan200834
 18Lopes et al.21Almada, Portugal20085
 19Traulle et al.22Saint-Quentin, France200814
 20Singh et al.26Springfield, IL, USA2009107
 21Regnante et al.14Providence, RI, USA200970
 22Previtali et al.10Pavia, Italy200918
 23Jim et al.25Hong Kong, China20098
 24Nielsen et al.27Vejle, Denmark20097
 25Sharkey et al.11Minneapolis, MN, USA2010136
Total (%)816 (43.5)
Excluded: no arrhythmia data
 1Dote et al.2Hiroshima, Japan19915
 2Kawai et al.42Tokyo, Japan20009
 3Yamasa et al.43Nagasaki, Japan200216
 4Abe et al.44Shimada, Shizuoka, Japan200317
 5Ito et al.45Gifu, Japan200310
 6Ogura et al.39Komatsushima, Japan200313
 7Inoue et al.46Kanazawa, Japan200518
 8Abdulla et al.47Sydney, Australia200635
 9Elian et al.48Tel Aviv, Israel200613
 10Athanasiadis et al.49Stuttgart, Germany200623
 11Haghi et al.50Mannheim, Germany200734
 12Mitchell et al.51Houston, TX, USA200722
 13Yoshida et al.52Nagoya, Japan200715
 14Cangella et al.53Mercogliano (Av.), Italy20076
 15Tomcsanyi et al.54Budapest, Hungary20076
 16Spedicato et al.55Udine, Italy200829
 17Burri et al.56Birmingham, AL, USA20085
 18Burgdorf et al.57Luebeck, Germany200850
 19Sganzerla et al.58Bergamo, Italy20087
 20Valbusa et al.59Genoa, Italy200822
 21Ibanez et al.60Madrid, Spain200820
 22Gerbaud et al.61Bordeaux, France200815
 23Nunez Gil et al.62Madrid, Spain20085
 24Fang et al.63Tainan, Taiwan200810
 25El Mahmoud et al.64Boulogne and Montfermeil, France200832
 26Eitel et al.65Leipzig, Germany200859
 27Eshtehardi et al.66Bern, Switzerland200941
 28Lee et al.67Singapore200910
 29Meimoun et al.68Compiegne, France200920
 30Vidi et al.69Burlington, MA, USA200934
 31Morel et al.70Strasbourg and Haguenan, France200917
 32Barker et al.71Charlottesville, VA, USA20099
 33Fitzgibbons et al.72Worcester, MA, USA200940
 34Nef et al.73Bad Nauheim, Germany200916
 35Banihashemi et al.74Baltimore, MD, USA20096
 36Lee et al.75ChunCheon, South Korea200945
 37Lee et al.75Florence, Italy200984
Total (%)818 (43.6)
Excluded: criteria not met
 1Seth et al.76Worcester, MA, USA200312
 2Park et al.77Seoul, South Korea200526
 3Derian et al.78Park Ridge, IL200712
 4Yoshimura et al.79Osaka, Japan20087
 5Winchester et al.80Charlottesville, VA, USA200831
 6Koeth et al.81Ludwigshafen, Germany200820
 7Hombach et al.82Ulm, Germany200812
 8Fazio et al.83Catania, Palermo and Rome; Italy. Praga, Czech Republic. Kawasaki, Japan200840
 9Abdel-Aty et al.84Calgary, Canada20097
 10Gaibazzi et al.85Parma, Italy20097
 11Chao et al.86Washington DC, USA200947
 12von Korn et al.87Neustadt/Weinstrasse, Germany200921
Total (%)242 (12.9)
View this table:
Table 3

Arrhythmia and inpatient mortality

ReferenceYearNumber of casesQTc (mean ± SD)/msVFSustained VTNon-sustained VTAsystoleAV blockSinus node dysfunctionAFSCD index episodeNon-SCD index episode
Tsuchihashi et al.1520018826c04961
Kurisu et al.1220023010
Akashi et al.162003713020
Desmet et al.13200313450 ± 83 (>400 in 12)111
Matsuoka et al.41200310555 ± 450000
Wittstein et al.4200519452 (range 490–592)1050
Hertting et al.28200632001010
Bonnemier et al.5200624a446 ± 4000200101
Patel et al.242007501000100
Fujiwara et al.6200711000100
Santos et al.1920075510 ± 60 (>440 in 4)0001100
Bonello et al.1720081411021
Dib et al.182008105434 ± 2521122174
Buja et al.72008280101
Yoshioka et al.82008344b404
Traulle et al.22200814000100
Lopes et al.212008500011d000
Kume et al.920085000300
Mitsuma et al.23200821000200
Singh et al.262009107470 ± 50100681
Regnante et al.142009703b01
Previtali et al.10200918460 ± 56 (>460 in 6)00000200
Jim et al.25200980001100
Nielsen et al.2720097479 ± 137001000200
Sharkey et al.11201013610112
% of total cases2.
  • Blank fields–data not reported.

  • VF, ventricular fibrillation; VT, ventricular tachycardia; AV, atrioventricular; AF, atrial fibrillation; SCD, sudden cardiac death.

  • aData from 22 patients with Holter monitoring, one case with AF, and one case with death from cardiogenic shock. Two patients with poor quality Holter recordings excluded.

  • bVF or VT not differentiated.

  • cSustained and non-sustained ventricular tachycardia not differentiated.

  • dSame case had both AV block and asystole.

Ventricular arrhythmia

There were 15 reported cases of VF, with a prevalence of 1.8% (15 of 816 cases). Data on VT were unavailable in two studies (43 patients, 7.6%).12,13 Two studies reported jointly on VF and sustained VT [present in 7 (6.7%) of 104 patients].8,14 There were 18 (2.2%) reported cases with VT, of which 4 (0.5%) were specifically reported as sustained VT, and 8 (1.1%) as non-sustained VT. One study with 88 patients (15.6% of total) did not specify whether six cases of VT were sustained or non-sustained.15 Considering these six cases as sustained VT, the combined case frequency with VF or sustained VT was 3.4% (28 of 816 cases), which was similar when series with unavailable data on VT were excluded (26 of 773 cases, 3.3%).

Documented potentially life-threatening ventricular tachyarrhythmias were the modality of presentation in nine cases (1.1%).11,13,1518 In seven (0.9%) of these, the patient presented with SCD secondary to VF (n = 6) and VT (n = 1). In two cases, VT was reported as a possible triggering event for TC.15,16 In the series from Akashi et al.16, the patient presented with recurrent syncope on a background of having had no other cardiac history and was found to have a polymorphic tachycardia, subsequently diagnosed as idiopathic VF. Takotsubo cardiomyopathy was demonstrated on echocardiography after ventricular arrhythmia was documented, 6 h after admission.16

Tsuchihashi reported eight cases (9.1% of 88 cases) with VT, of which two developed VF. In one of the cases, VT was thought to be the triggering event for TC. Buja et al.7 described a case with monomorphic sustained VT successfully treated with electrical cardioversion. Bonello et al.17 described a case with monomorphic sustained VT of right bundle branch block morphology and right-axis deviation resulting in recurrent cardiac arrest and death. Akashi et al.16 described three patients with non-sustained VT and reported therapeutic benefit from intravenous magnesium in one patient; the remainder were reported to not require therapy.

A total of five cases of ventricular asystole were reported (0.6% prevalence), often as a mode of presentation.1820 Lopes et al.21 described a case with intermittent asystole developing from complete heart block.

Atrial fibrillation

Atrial fibrillation was reported in a cumulative 38 (4.7% of 816) cases. There was variation in how studies reported AF. Five studies (seven cases) reported on AF as a new complication of TC.5,10,19,22,23 Seven studies (19 cases) did not clarify when documenting AF during the acute episode whether there was a past history of the arrhythmia.6,9,15,2427 Three studies reported on nine cases with a known past history of AF.16,18,27

Conduction tissue dysfunction

Presumed new onset SND was reported in three series, with a total of 11 patients affected (1.3% total cases, 4.5% of total in series which reported). Tsuchihashi et al.15 reported sinus bradycardia in nine cases. Dib et al.18 reported one patient with sinus pauses of up to 7.4 s at presentation, who later developed asystole requiring temporary pacemaker and inotropic support, with eventual recovery. Another patient had SND in conjunction with Mobitz type I AV block. Hertting et al.28 reported one patient with symptomatic sick sinus syndrome during the index episode requiring permanent pacemaker implantation.

Presumed new onset AV nodal dysfunction was reported in eight series, with a total 24 patients affected (2.9% of total cases, 5.6% of patients in series which reported on AV nodal dysfunction). Dib et al.18 described one case with Mobitz type I AV block noted above. This was associated with a right bundle branch block, but there was no evidence of infra-Hisian conduction disease on invasive electrophysiology testing. It persisted at the 16 months’ follow-up. Another patient in the same series had Mobitz type II AV block and a ventricular rate of 32 per min, requiring permanent pacemaker. She had an uneventful recovery of wall motion abnormalities with follow-up to 7 years. Sharkey et al.21 and Lopes et al.29 each reported a patient with transient complete heart block. Tsuchihashi et al.15 reported four patients with AV block, Yoshioka et al.8 reported two cases, Singh et al.26 reported six cases, and Jim et al.25 reported one case; none of these reports specified the degree or level of block. Wittstein et al.4 reported five patients with PR-interval prolongation.


In-hospital mortality was 2.5% (20 of 578 cases). Causes of death were sudden cardiac (n = 9, 1.1%), the underlying predisposing disorder (n = 5), cardiogenic shock (n = 3), multi-organ failure (n = 2), and heart failure (n = 1). Sudden cardiac death was associated with ventricular arrhythmias in the majority; one death was from asystole and in another the rhythm was unknown when death suddenly occurred 3 days after discharge.30 Sharkey reported ongoing excess mortality after the acute episode when compared with the general population matched for age and sex [standardized mortality rate 1.7 (95% CI 1.1–2.7); P = 0.016], predominantly from non-cardiac causes and with a much higher rate in the first year following presentation when compared with subsequent years [6.8 (4.0–11.5)] vs. 0.6 (0.3–1.4); P < 0.0001].

Of the three with VF in the Mayo Clinic series, one could not be resuscitated acutely, another patient suffered SCD 4 days after discharge, and the third was reported to be symptom-free 7 months after discharge.18 Desmet et al.13 reported one patient who presented with VF and had a similar self-limiting episode after 9 months from hospital discharge, with non-specific chest discomfort, deep negative T-waves in anterior leads, no arrhythmia, and no further symptoms for the remaining 3 months’ follow-up. In the initial report of 22 patients, Sharkey et al.29 described a patient presenting with VF treated by implantable cardioverter defibrillator. Bonello et al.17 reported a patient presenting with VF who had recurrent VF and death during the index hospital admission.

Dib et al.18 compared cases with life-threatening arrhythmia to those without and reported a longer maximal R-R interval (30.6 ± 14.5 vs. 14.5 ± 17 ms; P = 0.0004), longer QTc interval (491 ± 81 vs. 434 ± 25 ms; P = 0.0025), and longer PR interval (207 ± 96 vs. 162 ± 24 ms; P = 0.02) with life-threatening arrhythmia. When indexed to the heart rate, no difference in the PR interval was noted, indicating a shortening of PR interval in TC patients with arrhythmias at a higher sinus rate. There was also a higher prevalence of AF (66.6 vs. 9.6%; P = 0.001) and reduced beta-blocker use at presentation [2 of 6 (33.3%) vs. 25 of 31 (80.6%); P = 0.02]. Regnante identified pre-existing angiotensin enzyme-inhibitor use (0% meeting endpoint; P = 0.026) and intubation at presentation (7.1% meeting endpoint; P = 0.004) as significant univariate predictors of the combined endpoint of mortality, cardiogenic shock, and sustained VT or VF.14

On follow-up, Tsuchihashi et al.15 reported one SCD in the community in a patient with documented multivessel coronary spasm at baseline on intracoronary acetylcholine provocation, representing 10% (1 of 10 patients) with induced coronary spasm on presentation in this series. Buja et al.7 reported two patients with SCD in the community, 1 and 14 months after discharge.7 Lopes et al.21 reported one patient with SCD 3 months after discharge.


To the best of our knowledge, we reviewed all available series of TC that reported on arrhythmia and studied the total 816 published cases. Sustained VT and VF were reported in 3.4% and a mode of presentation in 1.1% of cases, requiring emergent resuscitation. They were also the major mode of death, with SCD being reported in 1.1% during the index episode and a further 0.5% suffering a SCD weeks to months later. The current published clinical data, therefore, suggests that ventricular arrhythmia is a clinically important entity in TC, albeit less frequent than expected given the sympathetic substrate and accompanying QT interval prolongation. In comparison, the MADIT-II trial and SEARCH-MI registry reported appropriate device therapy for ventricular arrhythmia in 17–21% of patients at 1 year and 27–31% at 2 years,31,32 whereas 9.4% of patients in the control arm of MADIT-II suffered an arrhythmic death.33 Atrial fibrillation, both transient and persistent, was reported in 4.7% of cases, SND in 1.3%, and AV node dysfunction in 2.9%. Both new onset AF and nodal dysfunction may be explained by catecholamine stress and secondary increase in vagal tone.

It is currently unclear whether TC is the cause or effect of arrhythmia in the light of cases described above where there was evolution after discrete episodes of tachyarrhythmia.15,17 Wittstein et al.4 provides evidence that plasma catecholamine levels in TC at presentation are raised above and beyond those resulting from MI and left ventricular failure (in comparison with Killip class III MI). In support of this are reported autonomic changes induced during VF, defibrillation threshold testing and AV node re-entry.34

Insights from magnetic resonance imaging have given an understanding of the evolution of TC at a tissue level, characteristically with an absence of scar formation.35 This makes re-entry an unlikely mechanism for arrhythmia and lends support to catecholaminergic, automatic, or fascicular tachycardias as more likely entities, with related cytosolic calcium overload.4,17,36 In this setting, beta-blockers have theoretical value, and though their use may be limited by haemodynamic limitations acutely,17 it may explain the findings by Dib et al.18 suggesting that pre-existing beta-blocker use may protect from ventricular arrhythmias. Multivessel coronary spasm has been reported as a cause of TC,2,37 and in these circumstances, there may be a transient ischaemic trigger for arrhythmia.

In a literature review of patients with TC and QT interval prolongation, Samuelov-Kinori et al.38 compared those who developed torsades with those who did not and found a higher prevalence of male sex (26.7 vs. 5.8%; P = 0.01) and longer QTc interval (679.9 ± 230.6 vs. 555.9 ± 63.8 ms, P = 0.06) in those who developed torsades. They also observed that a known risk factor for torsades was present in 80% of those who developed it, including congenital long-QT syndrome, bradycardia, hypokalaemia, recent conversion from AF to sinus rhythm, and using QT-prolonging agents. In TC, similar mechanisms relating to autonomic dysregulation of the QT interval seen after AMI may be involved in a predisposition to arrhythmia, given the similarities in the QT interval prolongation.39 This has traditionally been ascribed to an alteration in autonomic QT modulation following AMI.40 In TC, marked prolongation of the QT interval for up to 48 h after presentation was reported by Wittstein et al.4 in all 19 patients, with Matsuoka et al.41 reporting similar findings. However, Matsuoka et al.41 reported an absence of arrhythmia with TC despite prolonged QT intervals and QT dispersion present for several weeks after the index event. On the other hand, Dib's report focused on the interaction with heart rate variability and reported that the relationship between rate and QT interval is preserved in TC, unlike AMI.18 These findings and a relative depression in the indices for heart rate variability were also reported by Bonnemeier et al.5 on 24 h Holter monitoring of 22 patients on Day 3 of presentation with TC. These observations suggest that a separate mechanism may predispose to arrhythmia in TC, related to QT hysteresis and the RR-QT interval relationships, such that patients with TC with causes for significant rate variation may be at a particular risk for arrhythmia. Such variability may arise from imbalance in autonomic control, sinus or AV nodal dysfunction, or AF.

The literature generally provides limited data on the duration of cardiac monitoring used to identify transient arrhythmias, which may have prognostic value in addition to arrhythmias manifesting clinically. The optimal duration of monitoring in this condition is unclear, as is the risk of recurrent events after initial recovery. Whether those patients suffering SCD remote to the index episode had an ongoing arrhythmia substrate residual from the initial insult from TC, or possibly as an ongoing substrate which was the cause of TC in the first instance, or even a recurrence of TC itself, remains to be clarified. There is currently insufficient evidence on which to recommend internal defibrillator use, which can only be considered on a case-by-case basis at this point in time. Further work is required to evaluate the promise of beta-blockers and ACE-inhibitors in this condition, which currently have been used empirically on the grounds of heart failure.14,18

There are some intrinsic limitations as a result of study design. We excluded a number of case series to standardize the case definition. We also excluded those case series reporting on fewer than five patients to minimize on publication bias. Event-driven data may nevertheless be over-reported, whereas there may be under-reporting of non-sustained VT, asymptomatic paroxysmal AF, and the duration of arrhythmia risk, as the published experience with TC has not prospectively focused on arrhythmias associated with the condition. It was not possible to systematically pool data to assess demographic risk factors for arrhythmia from the published data.


This report describes the currently published case frequencies of the various arrhythmias that complicate TC and provides an insight into possible mechanisms. Areas for further prospective study include the duration and mechanism of residual risk and optimal risk stratification and modification. The current state of evidence would support beta-blockers in the acute setting to control tachyarrhythmia, although there is little evidence to support their use beyond convalescence when used for this indication alone. Those with heart failure and cardiogenic shock should be managed with established evidence-based therapies for these conditions. The use of internal defibrillators in TC requires consideration only on a case-by-case basis.

Conflict of interest: none declared.


View Abstract