OUP user menu

Global and regional myocardial function quantification by two-dimensional strain in Takotsubo cardiomyopathy

Felix Heggemann , Christel Weiss , Karsten Hamm , Jens Kaden , Tim Süselbeck , Theano Papavassiliu , Martin Borggrefe , Dariusch Haghi
DOI: http://dx.doi.org/10.1093/ejechocard/jep062 760-764 First published online: 28 May 2009


Aims This study sought to characterize global and regional systolic function in Takotsubo cardiomyopathy (TC) using two-dimensional (2D) strain imaging.

Methods and results Twelve consecutive patients (11 women, 1 man) underwent 2D echocardiography on admission and on early follow-up (34 ± 16 days). Two-dimensional images were analysed to measure longitudinal and radial strain and to calculate post-systolic shortening (PSS) and the post-systolic index (PSI). Mean age was 64 ± 14 years. Upon presentation ejection fraction, average longitudinal and radial strains were 42 ± 9%, −10.6 ± 5.5%, and 20.1 ± 17.3%, respectively. Values improved to 59 ± 8%, −17.6 ± 3.0%, and 50.2 ± 17.0%, respectively (all P < 0.001). PSS was present in 69% of segments upon presentation and in 53% of segments upon follow-up. PSI was −0.16 at baseline and improved to −0.06 upon follow-up (P < 0.001).

Conclusion Patients with TC show abnormal global and regional strain patterns during the acute phase of the disease which improve over time. However, subtle abnormalities of regional LV function seem to persist into the early follow-up period as suggested by the presence of PSS in more than half of LV segments. Long-term follow-up studies are needed to clarify whether these subtle abnormalities will further improve.

  • 2D strain
  • myocardial function
  • Takotsubo
  • cardiomyopathy
  • echocardiography


In 1991 Dote and colleagues reported on a novel syndrome consisting of transient left ventricular (LV) apical dysfunction (‘apical ballooning’), minimal elevation of myocardial enzymes in the absence of coronary artery disease, and electrocardiographic ST-T segment changes. It was named Takotsubo cardiomyopathy (TC) because of the angiographic appearance of the LV at endsystole resembling the shape of an octopus trap (Japanese tako: octopus; tsubo: pot).1 The pathogenesis of the syndrome is not well known. However, increased levels of circulating catecholamines seem to play a pivotal role.2

So far, quantitative data on regional myocardial function in this disease entity have been rare.3 Two-dimensional strain echocardiography (2D-SE) is a novel method for quantitative assessment of myocardial function. It is based on tracking the movement of stable acoustic patterns (called speckles) within the myocardium frame-by-frame throughout the cardiac cycle.47

In this study, we tested the application of this novel method for the evaluation of global and regional left ventricular (LV) dysfunction in TC.


Study population

Twelve consecutive patients with a clinical diagnosis of TC undergoing clinically indicated echocardiography were recruited for this study. Diagnosis of TC was based on the following criteria: (i) acute onset of apical and/or mid-ventricular wall-motion abnormalities of the left ventricle; (ii) rapid improvement of wall-motion abnormalities within a few days of the initial diagnosis; and (iii) exclusion of significant obstructive coronary artery disease by coronary angiography.

The study protocol was approved by the local ethical committee at our institution.


Echocardiography was performed within 48 h of diagnosis and upon early follow-up in all patients. Standard echocardiographic images were obtained using a Vivid 7 machine (GE Ultrasound, Horten, Norway) with a 2.5 MHz phased-array transducer at end-expiratory apnoea. Two-dimensional strain analysis was performed offline using custom 2D strain imaging software (EchoPac, GE Ultrasound). The endocardial borders were traced at the end-systolic frame of the 2D images. The interactive software then automatically tracked myocardial motion and divided each image into six segments. Numerical and graphical displays of deformation parameters (reflecting the average value for tracking all acoustic markers in each segment) were then generated for all six segments from each view.8 Short-axis images, just below the mitral valve level, were used to compute radial peak systolic strain. Longitudinal peak systolic strain was acquired in the apical two-chamber, three-chamber, and four-chamber views. If further shortening occurred after the end of systole, this was measured as peak strain. The difference between peak systolic strain and peak strain was calculated as the post-systolic shortening (PSS), and from this the post-systolic index (PSI) was derived from the equation: PSS/peak strain.9 Segments with systolic lengthening (rather than shortening) as indicated by positive systolic strain values in the apical views and by negative systolic strain values in the short-axis view were not used to calculate PSS and PSI. Global strain was calculated as the average longitudinal strain of the segments of two-, three-, and four-chamber views.

Statistical analysis

Continuous variables are expressed as mean ± standard deviation. Continuous variables were compared using the two-sided Mann–Whitney U-test. A P-value < 0.05 was considered statistically significant. Interobserver variability was assessed by two independent, blinded observers in five randomly selected patients. Intraobserver variability was performed by one observer re-measuring the same five patients after 3 weeks. The interobserver variability was expressed using interclass correlation coefficient (ICC).10


The clinical and echocardiographic features of patients are shown in Table 1. Eight patients had classical TC with apical ballooning and four had a variant without apical involvement. All patients were in sinus rhythm. Follow-up echocardiography was performed 34.3 ± 16.6 days after the initial evaluation (range 11–66 days). At baseline, 10 of 12 patients were on a beta-blocker. One patient was on a calcium channel-blocker. Upon follow-up, 9 of 12 patients were on a beta-blocker and none on a calcium channel-blocker.

View this table:
Table 1

Baseline clinical and echocardiographic characteristics

Age64 ± 14
Female, n (%)11 (92)
Hypertension, n (%)6 (50)
Diabetes, n (%)0 (0)
Current smoker, n (%)3 (25)
Hypercholesterolaemia, n (%)3 (25)
Presenting symptom, n (%)
 Angina7 (58)
 Dyspnoea2 (17)
 Other3 (25)
ECG upon presentation
 ST-elevation7 (58)
 T wave inversion3 (25)
 Other2 (17)
Type, n (%)
 Classical8 (67)
 Variant4 (33)
Peak troponin I (µg/L) (normal range: 0–0.4)3.30 ± 3.71
Triggering factor, n (%)
 Physical stress6 (50)
 Emotional stress1 (8)
 Other5 (42)
Biplane ejection fraction (%)
 Baseline (%)41 ± 9
 Follow-up (%)59 ± 8

At baseline, 2D strain was obtainable in 245 of 278 segments (88.1%) compared with 256 of 279 segments (91.8%) upon follow-up. Compared with baseline values, ejection fraction improved significantly (41.3% vs. 59.1%, P < 0.001) as did global strain (−10.6% vs. −17.6%, P < 0.001) and average radial strain [20.1% vs. 50.2% (P < 0.001)]. Regional inter- and intraobserver correlation coefficients at initial presentation were 0.82 and 0.98, respectively (basal 0.7 and 0.85; mid-ventricular 0.85 and 0.96; apical 0.91 and 0.99).

Table 2 shows baseline and follow-up measurements of regional strain in the different segments for the entire study population. Similar results were found for the subgroup of patients with classical TC (Table 3). An example is given in Figure 1.

Figure 1

Longitudinal strain curves from four-chamber view at baseline (A) and upon follow-up (B). Peak S, peak systolic strain; Peak G, peak strain.

View this table:
Table 2

Peak systolic regional strain values of the entire study population by segment

Strain (%)BaselineFollow-upP-value
Longitudinal; four-chamber view
Longitudinal; three-chamber view
Longitudinal; two-chamber view
  • *Not significant.

View this table:
Table 3

Peak systolic regional strain values by segment in classical Takotsubo cardiomyopathy

Strain (%)BaselineFollow-upP-value
Longitudinal; four-chamber view
Longitudinal; three-chamber view
Longitudinal; two-chamber view

In classical TC, longitudinal strain decreased from base to apex at baseline (Table 4). There was no apex-to-base gradient present upon follow-up. In variant TC, longitudinal strain was lowest at mid-LV segments (Table 5). Again, there was no apex-to-base gradient present upon follow-up.

View this table:
Table 4

Longitudinal strain values of all basal, mid-ventricular, and apical segments in classical Takotsubo cardiomyopathy

Longitudinal strainBaselineFollow-upP-value
Base (%)−15.9 ± 6.1−16.6 ± 8.10.116
Mid-LV (%)−8.4 ± 6.0*−16.5 ± 4.3<0.001
Apex (%)−1.7 ± 7.6**−16.1 ± 4.4<0.001
  • *P < 0.001 mid-LV vs. base.

  • **P < 0.001 apex vs. base.

View this table:
Table 5

Longitudinal strain values of all basal, mid-ventricular, and apical segments in variant Takotsubo cardiomyopathy

Longitudinal strainBaselineFollow-upP-value
Base (%)−13.2 ± 5.6−19.9 ± 5.80.001
Mid-LV (%)−10.4 ± 10.5*−18.4 ± 5.00.004
Apex (%)−18.8 ± 7.3**−22.3 ± 6.00.058
  • *P = 0.493 mid-LV vs. base.

  • **P = 0.014 apex vs. base.

A total of 25 segments (12%) had systolic lengthening (rather than shortening) as indicated by positive systolic strain values in the apical views and by negative systolic strain values in the short-axis view at baseline. Systolic lengthening was absent in all segments upon follow-up.

PSS was present in 69% of segments upon presentation and in 53% of segments upon follow-up. Post-systolic index was −0.16 at baseline and improved to −0.06 upon follow-up (P < 0.001).


The present study demonstrates that in TC, LV global and regional systolic function can be quantified by 2D-SE. Compared with baseline values, global strain improved in all patients upon follow-up. There was a significant apex-to-base gradient of strain at baseline, indicating the more severe involvement of the apical (and to a lesser extent) mid-ventricular segments compared with the LV base. However, despite the general perception of basal hypercontractility in TC, total longitudinal strain of the LV base was also diminished in several segments at baseline. In classical TC, this observation was predominantly because of an acute decrease of longitudinal strain in the basal anterior and anteroseptal segments. Reduced basal strain may, in part, be owing to the fact that sympathetically mediated myocardial stunning, which is believed to be the underlying pathogenic mechanism of TC,2 may also affect basal myocardium adjacent to the mid-ventricular segments. It is also conceivable that similar to what has been observed in patients with acute myocardial infarction,11 remote loading effects of dysfunctional segments, changes in LV geometry and a subsequent increase of wall stress in non-affected myocardium, and tethering of basal segments to dysfunctional mid-ventricular segments all play a contributory role to the observed decrease in basal longitudinal strain. Of note, baseline radial strain was reduced in all mid-ventricular segments; an observation that underlines the notion that TC affects myocardium beyond the territory of a single major coronary artery.

Consistent with previous reports in healthy volunteers,6 longitudinal strain upon follow-up was relatively uniform throughout the LV myocardium. Pathological PSS as indicated by high PSI and low strain values12 was present in a significant number of segments at baseline. Pathological PSS is a frequent but non-specific marker of regional myocardial dysfunction. It can be found in 78% of acutely ischaemic myocardial segments and in 79% of segments with chronic myocardial infarct scar.12 Our study demonstrates that pathological PSS occurs roughly at a similar incidence during the acute phase of TC. Of note, more than half of LV segments also showed PSS upon early follow-up. At this point in time, global parameters of LV function have almost normalized and regional LV wall-motion abnormalities are no longer present. This observation could be regarded as a possible indicator of persisting abnormalities of LV regional function into the early recovery period. Prolonged recovery of cardiac metabolism13 and long-lasting abnormalities of cardiac repolarization14 are known features of TC. However, to the best of our knowledge, this is the first study to report on long-lasting, though subtle, abnormalities of cardiac mechanical function in TC.


There are limitations to this study that must be acknowledged. Because of the small number of patients, the results should be regarded with caution. Follow-up time points were quite variable in our study ranging from 11 to 66 days. Thus, it remains unclear whether results would have been the same, if follow-up time points had been more standardized.

Some of the myocardial segments were uninterpretable owing to signal noise. This was not always caused by poor acoustic windows. Tracking problems were particularly common in the LV base, probably secondary to hyperdynamic LV contractility and annular motion at the LV base.8

A strictly basal variant of TC has been described.15 None of the patients in our study had this rare variant of the disease. It can be assumed that results will be different in this subset of patients. In fact, compared with classical TC, a reverse strain pattern can be expected in the basal variant of the disease.


Two-dimensional-SE is a novel technique that allows direct measurement of global and regional systolic function in TC. Major findings of this study are: (i) TC affects myocardium beyond the territory of a single major coronary artery; (ii) TC also alters mechanical properties of the LV base both in classical as well as variant TC; and (iii) similar to what is known from studies on myocardial metabolism and on cardiac repolarization in TC, subtle abnormalities of LV mechanical function also seem to persist into the early recovery period of the disease, although the exact time-course of these alterations could not be addressed by the present study.

Conflict of interest: none declared.


  • Both the authors contributed equally to this work.


View Abstract