Article Text

Download PDFPDF

Original research
Relationship between regional left ventricular dysfunction and cancer-therapy-related cardiac dysfunction
  1. Yoshihito Saijo1,
  2. Kenya Kusunose1,
  3. Yuichiro Okushi1,
  4. Hirotsugu Yamada2,
  5. Hiroaki Toba3,
  6. Masataka Sata1
  1. 1 Department of Cardiovascular Medicine, Tokushima University Hospital, Tokushima, Japan
  2. 2 Department of Community Medicine for Cardiology, Tokushima University Graduate School of Biomedical Sciences, Tokushima, Japan
  3. 3 Department of Thoracic, Endocrine Surgery and Oncology, Tokushima University Hospital, Tokushima, Japan
  1. Correspondence to Dr Kenya Kusunose, Cardiovascular medicine, Tokushima University Hospital, Tokushima 770-8503, Japan; kusunosek{at}tokushima-u.ac.jp

Abstract

Objective The aim of our study was to assess the association between risk of cancer-therapy-related cardiac dysfunction (CTRCD) after first follow-up and the difference in echocardiographic measures from baseline to follow-up.

Methods We retrospectively enrolled 87 consecutive patients (58±14 years, 55 women) who received anthracycline and underwent echocardiographic examinations both before (baseline) and after initial anthracycline administration (first follow-up). We measured absolute values of global longitudinal strain (GLS), apical longitudinal strain (LS), mid-LS and basal-LS at baseline and first follow-up, and per cent changes (Δ) of these parameters were calculated. Among 61 patients who underwent further echocardiographic examinations (second follow-up, third follow-up, etc), we assessed the association between regional left ventricular (LV) systolic dysfunction from baseline to follow-up and development of CTRCD, defined as LV ejection fraction (LVEF) under 53% and more absolute decrease of 10% from baseline, after first follow-up.

Results LVEF (65%±4% vs 63±4%, p=0.004), GLS (23.2%±2.6% vs 22.2±2.4%, p=0.005) and basal-LS (21.9%±2.5% vs 19.9±2.4%, p<0.001) at first follow-up significantly decreased compared with baseline. Among the 61 patients who had further follow-up echocardiographic examinations, 13% developed CTRCD. In the Cox-hazard model, worse Δbasal-LS was significantly associated with CTRCD. By Kaplan–Meier analysis, patients with Δbasal-LS decrease of more than the median value (−9.7%) had significantly worse event-free survival than those with a smaller decrease (p=0.015).

Conclusions Basal-LS significantly decreased prior to development of CTRCD, and worse basal-LS was associated with development of CTRCD in patients receiving anthracycline chemotherapy.

  • echocardiography

Statistics from Altmetric.com

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.

Introduction

In recent years, cancer therapy has advanced, but patients with cancer-therapy-related cardiac dysfunction (CTRCD) have increased. Many studies and pathological assessments revealed that anthracycline chemotherapy caused degeneration and fibrosis in the left ventricular (LV) myocardium inhomogeneously, leading to cardiac dysfunction.1–5 Patients with CTRCD have poor prognoses due to increased cardiovascular events and the limited selection of anticancer drugs that do not affect the heart.6–8

Transthoracic echocardiography is widely used for evaluating CTRCD. LV ejection fraction (LVEF) is a standard measurement, but LVEF can be misleading as a marker of LV systolic function causing several problems. Recently, global longitudinal strain (GLS), an indicator of global LV systolic function, has been used to detect subclinical myocardial impairment from chemotherapy.9–11 Decreased GLS after chemotherapy is a useful parameter for responders to cardioprotective therapy.12 However, the percentage of these responders progressively decreases as the time from anthracycline administration to heart failure treatment increases. Furthermore, no complete LV functional recovery is observed if the time to heart failure treatment is longer than 6 months.13 The evaluation of LV systolic dysfunction guided by GLS may delay the timing of appropriate therapeutic intervention in patients receiving anthracycline chemotherapy. Thus, identifying a more sensitive parameter for CTRCD than the conventional parameters is an important issue for better outcomes.

We hypothesised that regional LV impairment from anthracycline chemotherapy was detected by two-dimensional speckle tracking echocardiography (2D-STE) and is a useful indicator for development of CTRCD. The aims of this study were to evaluate a heterogeneity of LV systolic dysfunction from anthracycline chemotherapy by 2D STE and to assess the relationship between this heterogeneity and development of CTRCD.

Methods

Patient population and study design

We retrospectively enrolled 87 consecutive patients (58±14 years old, 55 women) who received anthracycline chemotherapy and underwent echocardiographic study both before and after initial anthracycline administration from February 2014 to January 2016 (figure 1). We excluded patients with previous anthracycline exposure and patients who developed CTRCD at first follow-up echocardiographic study to focus on development of CTRCD after first follow-up. We also excluded patients with significant cardiovascular diseases (LVEF <53%, asynergy of LV wall, ≥moderate valve diseases, cardiomyopathy and arhythmia) at baseline. Finally, we excluded patients with inadequate images, or without consent. Clinical risk factors (hypertension, diabetes, dyslipidaemia, smoking history) were obtained. Study approval was obtained from the Institutional Review Board of Tokushima University Hospital.

Figure 1

Patient flow diagram. CTRCD: cancer-therapy-related cardiac dysfunction; LV: left ventricular; LVEF: left ventricular ejection fraction.

Timing of the echocardiographic examinations and study endpoint

We determined the timing of echocardiographic study before anthracycline administration as Tb and the timing of the first echocardiographic study after initial anthracycline administration as T1 (figure 2). Follow-up clinical visits for the study population were done at least every 3 months in our institution, and the timings of further follow-up echocardiographic examinations were assessed by experienced physicians. We defined the period from initial anthracycline administration to T1 as follow-up period 1, and the period from initial anthracycline administration to the last follow-up echocardiographic examination as follow-up period 2. In this study, an endpoint was development of CTRCD, defined as LVEF under 53% and more absolute decrease of 10% from baseline.14

Figure 2

Time flow diagram. Tb was determined as the timing of echocardiographic study before anthracycline administration, and T1 as the timing of the first echocardiographic study after initial anthracycline administration. Follow-up period 1 is defined as the period from initial anthracycline administration to T1, follow-up period 2 as the period from initial anthracycline administration to the last follow-up echocardiographic examination.

Standard echocardiographic assessment

Experienced examiners performed standard echocardiographic examinations according to the American Society of Echocardiography guidelines15 using commercially available ultrasound examination machines (Vivid E9 and E95, GE Healthcare, Milwaukee, WI, USA; iE33 and EPIQ, Philips Healthcare, the Netherlands). We performed follow-up echocardiographic examinations with the same machine at baseline. Both LV and left atrial volumes were calculated using the biplane Simpson method and subsequently indexed by body surface area. The peak early diastolic velocity (E) and peak atrial filling velocity (A) of transmitral flow were measured. The peak early diastolic (e’) and the peak atrial systolic (a’) mitral annular tissue motion velocities were measured with a sample volume placed at the lateral side of mitral annulus.

Two-dimensional speckle tracking echocardiographic assessment

2D-STE was utilised to characterise longitudinal systolic strain. Images were acquired at ≥40 frames per second at end-expiration in the apical long, two-chamber and four-chamber views and analysed in a blinded manner, offline using a vendor-independent software package (EchoInsight, Epsilon Imaging, Ann Arbour, MI, USA). Images with <40 frames per second were excluded from analysis as inadequate images. For each of the individual apical views, tracking was visually inspected throughout systole to ensure adequate border tracking, and the endocardial contours were adjusted manually when necessary. The LV wall was divided into 18 segments, and each segmental LS was automatically measured. The mean of all segmental peak strain values was considered GLS, and each mean peak strain value of 6 segments in the apical, mid and basal layers were considered apical-longitudinal strain (LS), mid-LS and basal-LS, respectively. We evaluated these variables as absolute value. These parameters were measured at both Tb and T1, and per cent changes of these parameters were calculated ([strain value at T1 – strain value at Tb]×100/strain value at Tb), respectively (ΔGLS, Δapical-LS, Δmid-LS and Δbasal-LS).

Patient and public involvement statement

Neither patients nor the public were involved in the design, conduct, reporting or dissemination of our research.

Statistical analysis

Continuous variables were expressed as mean±SD for normal distribution, or median and IQRs for skewed distribution. The differences of continuous variables between Tb and T1 were compared by a paired samples t-test, and differences between CTRCD (+) and CTRCD (−) groups by an unpaired t-test. The χ2 test was used to compare categorical variables. The association of variables with time from T1 to development of CTRCD was evaluated by the univariable and multivariable Cox Proportional Hazard analyses with the forced enter method regardless of significance. We made three multivariate COX regression hazard models using pre-specified variables (age/cancer type/Δbasal LS, age/chemotherapy type/Δbasal LS, and age/ cumulative anthracycline dose/Δbasal LS) according to previous publications.6 16 17 An HR with a 95% CI was calculated for each variable. Additionally, to compare time until CTRCD between patients with ΔGLS and Δbasal-LS below and above the median (−2.4% and −9.7%, respectively), Kaplan–Meier analyses were constructed and compared using the log-rank test. To assess differences in basal-LS/GLS between CTRCD (+) and CTRCD (−) groups, we applied a linear mixed-effects model with unstructured covariance for random effects (intercepts and slopes on time). We used basal-LS/GLS as the outcome, CTRCD (yes/no) and time (continuous) as the covariates, and their interaction (CTRCD group ×time). A value of p<0.05 was considered significant. All statistics were calculated by the MedCalc V.16.8.4 (MedCalc Software, Mariakerke, Belgium) and SPSS V.20.0 (SPSS, Chicago, IL, USA).

Results

Baseline clinical characteristics

Baseline clinical characteristics of this study group are shown in table 1. In this study, 38 patients had breast cancer, 35 patients had malignant lymphoma and 14 patients had other diseases. All patients underwent chemotherapy with anthracycline. Most patients underwent T1 echocardiograms after completion of anthracycline chemotherapy, but 11 patients with malignant lymphoma received echocardiograms before completion of anthracycline chemotherapy (all patients completed chemotherapy within 1 month after T1). In all, 40 patients received epirubicin (264±126 mg/m2), 37 patients got doxorubicin (153±108 mg/m2) and 16 patients received other anthracycline drugs (five patients with daunorubicin, five patients with aclarubicin, three patents with idarubicin and three patients with pirarubicin). Among them, four patients received two kinds of anthracyclines, and one patient received three kinds of anthracycline. In all, 16 patients (18%) received trastuzumab.

Table 1

Baseline clinical characteristics of the study population

Reproducibility of GLS

Reproducibility was expressed as the mean percentage error (absolute difference divided by the average of the two observations). Measurement was performed in 20 randomly selected subjects by one observer and then repeated on two separate days by two observers who were unaware of the previous measurements and the study time point. The intra-observer and inter-observer variability of the basal-LS was 6.4%±3.2% and 8.8%±4.6%, respectively.

Comparisons of echocardiographic variables between Tb and T1

Table 2 shows the comparisons of the conventional echocardiographic variables between Tb and T1. A median time of follow-up period 1 was 89 days (IQR: 73–165 days). LVEF (65%±4% vs 63±4%, p=0.003), E (74±4 cm/s vs 69±18 cm/s, p=0.007) and E/A (1.1±0.9 vs 0.9±0.4, p=0.012) at T1 decreased compared with those at Tb. Figure 3 shows the comparisons of STE variables between Tb and T1. GLS and basal-LS at T1 significantly decreased compared with those at Tb (GLS; 23.2±2.6% vs 22.2±2.4%, p<0.001, basal-LS; 21.9±2.5% vs 19.9±2.4%, p<0.001). On the other hand, there were no significant changes of apical-LS and mid-LS between Tb and T1 (apical-LS; 27.3±4.4% vs 26.5±4.5%, p=0.16, mid-LS; 20.7±2.9% vs 20.0±2.7%, p=0.061). ΔGLS was −4.2%, Δapical-LS −1.3%, Δmid-LS −2.5% and Δbasal-LS −8.2%.

Table 2

Comparisons of the echocardiographic variables between Tb and T1

Figure 3

Comparison of speckle tracking echocardiographic variables between Tb and T1. GLS, global longitudinal strain; LS, longitudinal strain.

Association with development of CTRCD

All 61 patients underwent further echocardiographic examinations. In all, 26 patients who could not undergo further echocardiographic examination were excluded from analysis: five patients died from cancer progression and three patients from pneumonitis; 14 patients did not undergo follow-up echocardiography after moving to other institutions; and four patients were lost to follow-up. There were no significant differences in baseline characteristics between follow-up (+) and follow-up (−) groups as shown table 1. Among them, 8 (13%) patients achieved the endpoint during a median follow-up period 2 of 375 days (IQR: 265–512 days). In univariable analyses, basal-LS at T1 and Δbasal-LS were associated with the endpoint (basal-LS; HR=0.60, 95% CI=0.43 to 0.83, p=0.002. Δbasal-LS; HR=0.91, 95% CI=0.84 to 0.99, p=0.026, table 3). Multivariate Cox models showed worse ΔGLS had significant associations with development of CTRCD in all models (table 3). Additionally, we performed Kaplan–Meier analysis according to Δbasal-LS and ΔGLS (figure 4). There was no significant difference of the event free probability between the two groups divided by the median value (−2.4%) of ΔGLS (p=0.60). On the other hand, the patients with Δbasal-LS decrease of more than median value (−9.7%) had significantly worse event-free survival than those with a smaller decrease (p=0.015).

Table 3

Univariable and multivariable analyses of association between echocardiographic variables and cancer-therapy-related cardiac dysfunction

Figure 4

Kaplan–Meyer analysis of event-free survival of ΔGLS and Δbasal-LS. CTRCD, cancer-therapy-related cardiac dysfunction; GLS, global longitudinal strain.

Comparisons of time courses of GLS and basal-LS

As expected on the basis of parameter estimates obtained by the mixed-effects model, GLS and basal-LS at baseline were similar in CTRCD (+) and CTRCD (−) groups at the beginning of the follow-up (p=0.98 and p=0.42, figure 5). The GLS and basal-LS significantly decreased at rates of −0.6±0.3 and −0.8±0.2 %/follow-up (both p<0.001). Especially in the CTRCD (+) group, basal-LS showed a significant decrease with greater rate compared with CTRCD (−) group (p=0.04). During the follow-up period, there was no difference in the decreasing rate in GLS between CTRCD (+) and CTRCD (−) groups (p=0.33).

Figure 5

Comparisons of time courses of GLS and basal-LS between CTRCD (+) and CTRCD (−) groups. GLS, global longitudinal strain; LS, longitudinal stain.

Discussion

The main findings of the present study were (1) LVEF, GLS and basal-LS significantly decreased from anthracycline administration prior to development of CTRCD; (2) worse basal-LS was significantly associated with development of CTRCD. To the best of our knowledge, this is the first report assessing the relationship between regional LV systolic dysfunction from anthracycline chemotherapy and development of CTRCD.

CTRCD is a fatal problem for patients who received chemotherapy with anthracycline, occurring in 9% of patients and 98% of them occurred within the first year.18 Myocardial impairment from anthracycline can progress rapidly, leading to congestive heart failure or irreversible cardiac dysfunction. The percentage of responders to cardioprotective therapy progressively decreased as the time from anthracycline administration to heart failure treatment increased.13 For this reason, seeking a sensitive marker to predict development of CTRCD is important for better cardiac outcomes. The previous literature reported that ΔGLS is an indicator of decreased LVEF from chemotherapy.19 However, our study showed that there was no significant association between ΔGLS and development of CTRCD. This discrepancy may result from the smaller number of subjects and shorter follow-up period from initial chemotherapeutic administration to first follow-up echocardiographic examination compared with the previous paper. In the present study, GLS also significantly decreased from anthracycline chemotherapy, but a decreased rate of GLS was smaller than that of basal-LS. In addition, GLS may underestimate the severity of LV systolic dysfunction when LV impairment occurs only regionally because GLS is an indicator of whole LV systolic function. Thus, worse basal-LS may reflect LV impairment from anthracycline chemotherapy more sensitively compared with GLS.

The heterogeneity of LV systolic dysfunction from chemotherapy has been discussed. Lange et al reported that the reduction of LS in basal segment during chemotherapy is notable in patients with breast cancer.3 On the other hand, some studies reported that LV systolic dysfunction in apical or septal side decreased from chemotherapy.2 20 21 There are some possible reasons for this discrepancy: (1) different chemotherapeutic regimens; (2) different background characteristics including race and cardiovascular risk and (3) development of Takotsubo cardiomyopathy. But, these studies did not assess the relationship between the heterogeneity of LV impairment and development of CTRCD. Results of our study implied that the regional LV systolic dysfunction in the basal segment decreased prior to other segments, and subsequently led to global LV dysfunction. Further prospective study is needed to confirm these clinical questions. Furthermore, Coutinho et al showed that regional LV systolic function (especially in the anterior or septal wall) decreased from chemotherapy by using three-dimensional STE (3D-STE), which has high accuracy to detect LV systolic dysfunction.22 However, there are some limitations including low frame rate, reproducibility of regional strain and limited machines. Whereas they evaluated the regional change of LV systolic dysfunction for every 18 segments, we evaluated it using the average strain values for six segments (base, mid and apical) at ≥40 frames per second, leading to higher reproducibility of regional strain values. In addition, because the survey for CTRCD must be performed in every hospital, we believe that 2D-STE, which is widely available and is easily used to analyse LV function, is more suitable modality for now than 3D-STE.

The mechanism of regional LV impairment from anthracycline has not been clearly shown, but it is thought to be caused by underlying diseases with increased inflammatory cytokines and loading conditions, leading to vacuolisation of myocardial fibres and myocytolysis.1 The hypothesis of this mechanism may be explained by the law of Laplase.23 The cross-sectional radius of curvature in the basal segment is greater than the other segments, and myocardium and endocardium in the basal segment are damaged by higher wall stress than the other segments.24–26 Badke et al have shown a progressive LV basal dilatation without a noticeable LV apical response in early changes by chronic volume overload.26 Moreover, Ashikaga et al reported that the increases in myofiber and sheet shear remodelling in the basal segment are also significantly greater than the apical segment by the response of early remodelling for volume overload.27 In healthy people, these adverse changes are subtle and do not significantly affect cardiac function. However, when patients are exposed to chemotherapy with cardiotoxicity, the basal myocardium which suffers under high pressure and wall stress is more likely to be caused apoptosis and fibrosis than other segments.1 28 Consequently, this subtle impairment is expressed as worse basal-LS.

Emerging echocardiographic technologies such as STE have resulted in many studies on the detection of CTRCD, and GLS has been shown to be a useful marker of therapeutic intervention for CTRCD.12 19 29 However, the conventional evaluation of LV systolic dysfunction may lead to delayed timing of the appropriate therapeutic interventional because myocardial impairment from cardiotoxicity of anthracycline can progress rapidly, and the effect of cardioprotection therapy progressively is limited. Therefore, a more sensitive marker for CTRCD than GLS is needed for effective therapeutic intervention. The present study showed that there is regional LV systolic dysfunction prior to development of CTRCD in patients receiving anthracycline chemotherapy, and the relationship between worse basal-LS and CTRCD. Therefore, evaluating basal-LS may be more useful as a marker of therapeutic intervention timing for CTRCD compared with the conventional parameters.

Limitations

First, our retrospective cohort study investigated a small number of patients in a single centre. We could not enter some variables into the model because of the relatively small number of outcomes, which posed a potential risk of model over fit. Second, timing of follow-up echocardiographic studies was judged by each physician clinically, and patients were not examined at regular intervals. Furthermore, the patients who could not undergo further echocardiographic studies due to death or moving to other institutions were excluded from analysis of the relationship between echocardiographic parameters and development of CTRCD. Therefore, we have to consider a selection bias. A further prospective study with a larger number of patients to confirm our findings (eg, Strain Surveillance of Chemotherapy for improving Cardiovascular Outcomes trial) is needed.30 Third, patients in this study were also treated with other chemotherapy drugs such as trastuzumab. However, the number of patients treated with trastuzumab in our study was relatively small, and no patients treated with trastuzumab developed CTRCD. We theorised that the influence of trastuzumab to this investigation was small. Fourth, we used multiple vendors to acquire images in this study. However, we performed the follow-up echocardiographic examinations with the same machine as baseline in the same subject, and all images were analysed using a vendor-independent software package. We believe that the variety of strain value by using multiple vendors is small.

Conclusion

Basal-LS significantly decreased prior to development of CTRCD in patients receiving anthracycline chemotherapy, and worse basal-LS was significantly associated with development of CTRCD. Detecting regional LV systolic dysfunction may be more useful for risk stratification than conventional echocardiographic parameters. We need to perform a further prospective study to confirm our findings because the present single centre study had a small number of patients.

Key questions

What is already known on this subject?

  • Global longitudinal strain (GLS) has been shown to be a useful marker of therapeutic intervention for prevention of cancer-therapy-related cardiac dysfunction (CTRCD). However, the evaluation of left ventricular (LV) systolic dysfunction guided by GLS may delay the timing of appropriate therapeutic intervention in patients receiving anthracycline chemotherapy.

What does this study add?

  • We hypothesised that LV impairment from anthracycline chemotherapy occurs heterogeneously prior to development of CTRCD. We retrospectively assessed a relationship between this heterogeneity and development of CTRCD. This study showed that basal-longitudinal strain (LS) decreased in patients receiving anthracycline, and worse basal-LS was significantly associated with development of CTRCD.

How might this impact on clinical practice?

  • Evaluating worse basal-LS may be a more sensitive marker of therapeutic intervention for prevention of CTRCD compared with conventional parameters. Because the present single-centre study had a small number of patients, a further prospective study is needed to confirm our findings.

Acknowledgments

The authors acknowledge Kathryn Brock, BA, for editing the manuscript, and Dr. Akira Tangoku, Dr. Hirokazu Takechi and Dr. Masami Morimoto, for introducing the patients.

References

Footnotes

  • Twitter @Ken_Cardiology, @echoboy7

  • Contributors KK designed the study and wrote the initial draft of the manuscript. YS contributed to analysis and interpretation of data, and assisted in the preparation of the manuscript. All other authors have contributed to data collection and interpretation, and critically reviewed the manuscript. All authors approved the final version of the manuscript, and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

  • Funding This work was partially supported by JSPS Kakenhi Grants (Number 17K09506 to K. Kusunose, and 19H03654 to M. Sata), the Takeda Science Foundation (to K. Kusunose).

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Patient consent for publication Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement Data are available upon reasonable request.