OUP user menu

Tissue Doppler imaging and conventional echocardiography after anthracycline treatment in adults: Early and late alterations of left ventricular function during a prospective study

Sophie Tassan-Mangina, Daniela Codorean, Marie Metivier, Brigitte Costa, Chantal Himberlin, Christelle Jouannaud, Anne Marie Blaise, Jacques Elaerts, Pierre Nazeyrollas
DOI: http://dx.doi.org/10.1016/j.euje.2005.04.009 141-146 First published online: 1 March 2006


Aim of the study To evaluate the short and long-term effects of anthracycline chemotherapy in adults using conventional echocardiography and pulsed tissue Doppler imaging (TDI).

Methods and results Twenty patients were included of which 16 had a complete follow up. They underwent an echocardiography before chemotherapy, 1–3 months and 3.5±0.6 years after the treatment. We recorded pulsed TDI at the mitral annulus, the basal segments of the left ventricular (LV) lateral and posterior walls; peak velocities in systole (Sm), early (Em) and late diastole and the isovolumic relaxation time (IVRT) were measured. The cumulative dose of doxorubicin was 211±82g/m2. Early after anthracycline therapy, we observed changes in the diastolic LV function with a decrease of the mitral E peak velocity and TDI Em. At the late control, diastolic changes were more pronounced and associated with an alteration of the systolic function (LV ejection fraction and Sm). Four patients had a LV ejection fraction <50%; in these patients we observed a mitral annulus IVRT <80ms at the early control; this could be of interest to predict later impairment of the LV ejection fraction.

Conclusion We found early changes in LV diastolic function and observed that late impairment of the LV ejection fraction occurred frequently after anthracycline therapy, despite normal systolic LV function during the first months of follow-up.

  • Anthracycline
  • Tissue Doppler imaging
  • Left ventricular ejection fraction


Anthracycline chemotherapy is an effective therapy for numerous types of malignant tumours. However, the late onset of cardiac toxicity is not unusual and may be largely underestimated. In this setting, physicians usually survey the left ventricular (LV) ejection fraction using echocardiography or radionuclide angiography, mainly during treatment and the year following its completion.1–3 Tissue Doppler imaging (TDI) has been developed for more than 10 years and is now used in everyday practice. It allows measurement of diastolic and systolic velocities of the ventricular walls and of the mitral annulus. TDI appreciation of LV diastolic performances appears to be less conditioned by loading conditions than evaluation with conventional Doppler, and, thus, more reproducible. TDI may also disclose changes in regional function that do not reflect on global LV ejection fraction.4–6 Thus, TDI could be of interest to improve the evaluation of cardiac function changes after anthracycline therapy.


Study population

The aim of this study was to evaluate the early and late effects of anthracycline chemotherapy in adults, not only with conventional echocardiography but also using TDI. We included 20 patients (5 men and 15 women), 38±10 years old, who had no history of heart disease or exposure to chemotherapy, and who required anthracycline therapy for breast cancer (n=11), lymphoma (n=5), or acute leukemia (n=4). During the follow up, 3 patients died and 1 refused a further clinical assessment.


Echocardiography was performed before treatment (n=20), 1–3 months after the completion of chemotherapy (n=18), and 3.5±0.6 years afterwards (n=16) as according to the recommendations of the American Society of Echocardiography. We used a VIVID 7 GE Medical System or a Sequoia Accuson ultrasound system with TDI capabilities. LV end diastolic and end systolic volumes were obtained from the apical 4- and 2-chamber views according to Simpson's rule.7 LV ejection fraction was derived from these volumes and was considered abnormal at under 50%. Pulsed Doppler examination of the LV inflow was performed from the 4-chamber view with the sample volume placed between the mitral leaflet tips. We measured the peak early diastolic inflow velocity E, the peak late diastolic inflow velocity A, the ratio E/A, and E deceleration time. We used pulsed TDI to record the velocity profile of the mitral annulus and of the basal segment of the LV lateral wall from the apical 4-chamber view (longitudinal velocities). The TDI velocity profile of the basal segment of the posterior wall was obtained from the parasternal long-axis view (radial velocities). At each site we measured the TDI peak velocity in systole (Sm), early (Em) and late diastole (Am) and the isovolumic relaxation time (IVRT). Quantitative results are expressed as mean±SD.

Statistical analysis

Three consecutive beats in sinus rhythm were averaged. Evolution over time was analysed with ANOVA for repeated measure and, when significant, post hoc assessment of differences between early or late data and baseline, was based on non-parametric Wilcoxon 2-sample test, or Fisher's exact test if needed. Differences in magnitude of significant relative changes between conventional and TDI measurements, for systolic and diastolic function, were analysed using non-parametric Wilcoxon 2-sample test. A p-value <0.05 was considered significant.


The data from the 3 echocardiograms are summarized in Tables 1 and 2. The first echocardiogram was normal in all patients. The LV ejection fraction was 72±6% before chemotherapy, the lowest being 59%. The total cumulative dose of doxorubicin was 211±82g/m2 (100–375g/m2). Early assessment (1–3 months after the completion of chemotherapy) displayed significant changes in diastolic myocardial function, alterations in LV inflow (decrease of E peak velocity and E/A ratio) and in TDI spectrum (decrease of radial and longitudinal Em). At that time we observed no significant change in systolic myocardial function (LV end-systolic volume, LV ejection fraction and Sm). At the final evaluation (3.5±0.6 years) we observed changes in both diastolic and systolic function: changes regarding Em and Am were more pronounced and LVEF significantly decreased, as did Sm. It is noteworthy that, even if the mean LV ejection fraction of the study group was at the lower limit of normal values, 4 out of 16 patients had an LV ejection fraction lower than 50% and 6 lower than 55%; none of them, however, had clinical symptoms of congestive heart failure. Changes in LV ejection fraction were not correlated to doxorubicin cumulative dose.

View this table:
Table 1

Data obtained during conventional transthoracic echocardiography before chemotherapy (TTE1), 1–3 months after treatment (TTE2), and 3.5±0.6 years later (TTE3)

TTE1, n = 20TTE2, n = 18TTE3, n = 16ANOVAa1 vs 2b1 vs 3b
Heart rate (beat/min)82±1277±1176±13ns
End-diastolic LV volume (mm3)75±2565±1966±17ns
End-systolic LV volume (mm3)22±1121±929±100.009ns0.023
LV ejection fraction (%)72±671±856±8<0.001ns<0.001
Peak early diastolic velocity of LV inflow E (cm/s)74±1366±1067±110.0190.0180.04
Peak late diastolic velocity of LV inflow A (cm/s)51±1253±1157±11ns
E/A ratio1.50±0.371.27±0.221.20±0.210.0020.0250.002
E deceleration time (ms)174±47149±29159±36ns
Isovolumic relaxation time (ms)91±1594±1993±14ns
  • LV, left ventricular; ns, not significant.

  • a Anova for repeated measure.

  • b Wilcoxon 2-sample test if Anova p-value <0.05.

View this table:
Table 2

Tissue Doppler imaging data obtained before chemotherapy (TTE1), 1–3 months after treatment (TTE2), and 3.5±0.6 years later (TTE3)

TTE1, n = 20TTE2, n = 18TTE3, n = 16ANOVAa1 vs 2b1 vs 3b
Basal segment of the LV lateral wallPeak systolic velocity Sm (cm/s)12±211±38±2<0.001ns0.003
Isovolumic relaxation time (ms)90±2884±1376±12ns
Peak early diastolic velocity Em (cm/s)17±314±411±3<0.0010.040.001
Peak late diastolic velocity Am (cm/s)11±210±37±2<0.001ns0.002
Basal segment of the LV posterior wallPeak systolic velocity Sm (cm/s)9±28±26±1<0.002ns<0.001
Isovolumic relaxation time (ms)75±1969±1060±15ns
Peak early diastolic velocity Em (cm/s)14±312±38±2<0.0010.027<0.001
Peak late diastolic velocity Am (cm/s)8±37±25±1<0.001ns0.001
Mitral annulusPeak systolic velocity Sm (cm/s)13±313±29±2<0.001ns0.001
Isovolumic relaxation time (ms)81±1185±976±100.002ns0.006
Peak early diastolic velocity Em (cm/s)17±317±412±3<0.001ns0.001
Peak late diastolic velocity Am (cm/s)12±212±28±2<0.001ns0.001
  • LV, left ventricular; ns, not significant.

  • a Anova for repeated measure.

  • b Wilcoxon 2-sample test if Anova p-value <0.05.


Chronic cardiac toxicity induced by anthracycline chemotherapy has been known for four decades.1 It occurs more frequently within 1 year of treatment, with a peak incidence 1–3 months after chemotherapy, and is not reversible. Late-onset cardiac toxicity has also been documented, especially in children.2–3 For Steinherz the risk of long-term cardiomyopathy is higher in patients who have abnormal echocardiograms at the end of their therapy: in his study only 12% of patients with a normal echocardiogram during the year after completion of chemotherapy, have abnormal echocardiograms during the follow-up.2 However, we lack long-term studies for cancer survivors treated during adulthood; as a consequence, long-term cardiac follow-up is often neglected in adult patients. Such a follow-up may be of interest since 20% of the patients initially included in our study had a late impairment of LV ejection fraction. As life-expectancy increases with improvement in treatments for malignant tumours, we have to take into account the potential late consequences of anthracycline cardiac toxicity to guide patients' care. Our results are in the range of Steinherz's who found a 23% incidence of late cardiac damage after childhood malignancy. The incidence of congestive heart failure increases rapidly for a dose of doxorubicin greater than 550mg/m2, which has been recognised as an empiric limiting dose for doxorubicin-induced cardiotoxicity.8 However, as in our work, subclinical cardiac damage has been documented, for doses lower than this threshold, in several studies using either echocardiography or radionuclide ventriculography.9–13 Within a few months of the completion of chemotherapy, we observed significant changes in diastolic left ventricular function, identified by LV inflow analysis, whilst changes in systolic function occurred later together with even more pronounced diastolic changes. Our findings are similar to those observed in previous studies where diastolic impairment arises first.9,11,13,14 For Stoddard, it could predict systolic dysfunction.10 However, from our data, early significant changes (E, Em, E/A) did not appear to predict accurately a LV ejection fraction lower than 50% at 3 years. TDI isovolumic relaxation time, measured at mitral annulus during early assessment, was shorter in these patients (positive and negative predictive value of 100% and 91%, respectively, to predict late LV ejection fraction under 50% for patients with an IVRT under 80ms, p<0.03) (Fig. 1). We found no significant predictive value of IVRT measured from flow Doppler, or from basal segment TDI. In our study, mitral annulus IVRT appeared to outperform both standard Doppler IVRT, which was not modified by anthracycline therapy, and basal segment measurements, which had greater variance. The lack of modification of flow Doppler IVRT is not surprising, however, conventional IVRT shortening is a well known marker of severe left ventricular dysfunction, when filling pressure increases; obviously this did not occur in our patients, who had no sign of heart failure or severe left ventricular diastolic dysfunction.

Figure 1

Relationship between late alteration of left ventricular function and mitral annulus isovolumic relaxation time value early after chemotherapy.

Assessment of normal values for TDI measurements is still controversial and vary from one study to another, thus it is difficult to assess if our Doppler measurements remain in the limits of normal values. LV ejection fraction and conventional Doppler measurements also depend on loading conditions, which is a limiting factor during follow-up since haemodynamic settings may change over time in such patients.15,16 TDI assesses the velocity of myocardial structures, instead of blood flow and has proven its reliability in myocardial function evaluation and a greater, if not absolute, independence from loading conditions.4–6

Consequently, it has been used to detect abnormal cardiac function in various disease groups such as LV hypertrophy, restrictive cardiomyopathy or ischemic cardiopathy.17 The TDI velocity profile of the mitral annulus is usually considered as a marker of the global LV function, while velocity profiles obtained in myocardial segments of the left ventricle reflect regional myocardial function. Because of Doppler orientation, radial LV shortening and relaxation are best reflected by the TDI profile of the posterior LV wall, whilst the lateral wall and mitral annulus provide information on the longitudinal LV function.17 In our study, TDI confirms the early diastolic and late systolic impairment of the left ventricular function following moderate dose anthracycline therapy. The changes in LV function are global. The changes in the mitral annular velocities agree with those obtained in the posterior and lateral segments of the left ventricle.

In our small population of young patients, with normal systolic and diastolic function before treatment, we found no important advantage of TDI, since standard echocardiography also displayed significant changes. However, changes were more pronounced for most DTI measurements compared with standard Doppler or ejection fraction evaluation (Fig. 2).

Figure 2

Relative significant changes between 1st and 2nd (n=18), and 1st and 3rd (n=16) echocardiographic assessment. E, peak mitral inflow early diastolic velocity; LVEF, left ventricular ejection fraction; Sm, peak tissular Doppler systolic velocity; Em, peak tissular Doppler early diastolic velocity. *p<0.05 and **p<0.01 for comparison with LVEF; †, p<0.01 for comparison with mitral E velocity.

It is also of clinical interest that a TDI study may help to screen people with a high risk of late anthracycline cardiomyopathy: an isovolumic relaxation time on the mitral annulus below 80ms, early after the completion of chemotherapy, proved to be a good predictor of late LVEF impairment in our study. However, with regard to the value of TDI to predict late impairment of ventricular function, these results must be considered as preliminary; the small sample size of patients makes a larger study mandatory to confirm our results. Nevertheless, we did observe that changes in diastolic parameters occur early compared to changes in LV ejection fraction and that TDI evaluation of regional and global LV function is very significantly modified after anthracycline chemotherapy.

Our study combined TDI and conventional echocardiography for long-term cardiac follow-up after anthracycline therapy in adults and underlines the necessity of prolonged cardiac vigilance since the LV ejection fraction is impaired in up to 25% of our patients at 3 years despite the preservation of systolic left ventricular function in all patients during the first year of follow-up.


We thank Sean Joyce for his helpful review of the manuscript.


View Abstract