Introduction

Over the last several decades, cancer therapy continues to advance, resulting on one hand with increased survival rates, but on the other hand strengthen the importance of long-term side effects from chemotherapeutic drugs [1, 2], with cardiovascular disease being a leading cause of death among patients with cancer [3]. Cancer therapy-related cardiac dysfunction (CTRCD), defined by a left ventricular ejection fraction (LVEF) reduction, is a well-documented side effect of certain therapeutic agents [4]. LVEF has been shown to be an important indicator of the outcome as a monitor of heart function [5]. The use of LVEF as a primary measurement of overall cardiac function, however, requires a substantial tissue function loss, often irreversible [6,7,8], before being clinically detectable [9,10,11]. Global longitudinal strain (GLS), a parameter of 2D speckle-tracking echocardiography (STE) has been shown to provide clinicians with information on more subtle left ventricular function changes [12] and is associated with overall CTRCD outcomes [13, 14]. GLS has been shown to be useful in the prognostication of all-cause mortality [15]; however, its ability to predict all-cause mortality in patients receiving cancer therapy has not been well documented. Routine use of GLS in patients receiving cancer therapy has not been fully adopted yet, due to limited data [16, 17]. Using GLS routinely as a measure of cardiac function, the aim of this study was to evaluate the frequency of reduced GLS among patients receiving cancer therapy, and whether it is associated with CTRCD development and all-cause mortality.

Methods

Study population

The study population is part of the Israel Cardio-Oncology Registry (ICOR)—a prospective registry enrolling all adult patients evaluated in the cardio-oncology clinic at Tel Aviv Sourasky Medical Center. All patients signed an informed consent at the first visit in the clinic and are then followed prospectively. The registry was approved by the local ethics committee and is registered in clinicaltrials.gov (Identifier: NCT02818517).

The clinic follows adult patients who are currently receiving cancer therapy. In general, the registry includes three types of populations: patients that developed cardiovascular complications during therapy; high-risk patients with baseline risk factors and as of February 2017, the clinic evaluates preventively all patients planned for anthracyclines (ANT) therapy. From October 2016 to August 2018, 419 patients receiving cancer therapy were evaluated, of which 128 patients were excluded due to not performing GLS assessment, leaving 291 patients for analysis.

Study protocol

Past medical history, cardiac risk factors, cancer type and chronic medical treatment were noted in all patients. Regarding the ongoing cancer therapy, we analyzed only therapy associated with LVEF dysfunction, according to the European Society of Cardiology position paper 2016 [4]. At least one echocardiogram, including GLS assessment, was performed for each patient in the study. Patients were divided into two groups—reduced GLS group vs. preserved GLS group. Preserved GLS was defined as ≤ − 17% adhered to the standard benchmark set by previous studies [18]. Both groups were evaluated for the parameters associated with reduced GLS; the risk of EF reduction and CTRCD development, defined as a LVEF reduction of > 10%, to a value below 53% [19] at following echocardiogram evaluation and all-cause mortality retrieved from the electronic records of the governmental population.

Echocardiography

Three standard apical views (4-chamber, 2-chamber, and apical long-axis) were recorded using a General Electric system, model Vivid S70 echocardiogram and were performed by the same vendor, technician and interpreting cardiologist in order to prevent inter-vendor variability. Left ventricle (LV) diameters were measured from the parasternal short axis, by means of a two-dimensional or a two-dimensional-guided M-mode echocardiogram of the LV, at the papillary muscle level [20]. LVEF was calculated by the biplane method.

Images were acquired using high frame rate (> 50 frames/s) [21], and thereafter stored digitally for offline analysis. GLS was measured using STE software and tracking within an approximately 5 mm wide region of interest. A mid-systolic frame was used to initialize LV boundaries which were then automatically tracked throughout the cardiac cycle. Manual corrections were performed to optimize boundary tracking as needed. Optimization of images for endocardial visualization through adjustment of gain, compress, and time-gain compensation controls were done prior to acquisition (Fig. 1).

Fig. 1
figure 1

A patient with reduced global longitudinal strain developing cancer therapy-related cardiac dysfunction

Full size image

Statistical analysis

Categorical variables were reported as frequency and percentages. Continuous variables were evaluated for normal distribution using histogram and Q–Q plot. Normally distributed continuous variables were reported as mean and standard deviation (SD) while skewed data were presented as medians and interquartile ranges (IQR). Categorical variables were compared between categories using Chi-square test or Fisher’s exact test and continuous variables were compared using independent samples t test or Mann–Whitney test. Univariable logistic regressions were used to evaluate the association between baseline GLS and CTRCD development. A propensity score was modeled from baseline characteristics specified in Tables 1 and 2. A propensity score-adjusted logistic regressions were then used. Cox regressions were used to evaluate the association between baseline GLS and all-cause mortality. Cox regressions were adjusted for propensity scores which were modeled according to the above-specified parameters. A two-tailed p < 0.05 was considered statistically significant. Analyses were performed with SPSS (IBM Corp. Released 2013. IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: IBM Corp.).

Table 1 Baseline characteristics according to global longitudinal strain
Full size table
Table 2 Caner type and chemotherapeutic parameters according to global longitudinal strain groups
Full size table

Results

Baseline parameters

Of 291 patients evaluated consecutively, 48 (16%) patients were included in the reduced GLS group, according to the first GLS evaluation, while the remaining 243 patients were included in the preserved GLS group. Patients in the reduced GLS group were older (68[60–77] vs. 61[48–69], p < 0.001) with a male predominance (52% vs. 25%, p < 0.001) (Table 1). Among this group, cardiac morbidities were observed at a significantly higher prevalence, including ischemic heart disease (31% vs. 9%, p < 0.001), systolic dysfunction (LVEF < 50%) (46% vs. 0.4%, p = 0.001), atrial fibrillation (17% vs. 6%, p = 0.016) and chronic renal failure (13% vs. 1%, p < 0.001). However, aside from hyperlipidemia (42% vs. 20%, p = 0.001), no significant differences were noted in other cardiovascular risk factors (Table 1).

Patients with reduced GLS were more likely to be treated with beta blocker (BB) (62% vs. 23%, p < 0.001), however, no significant differences were observed regarding angiotensin converting enzyme inhibitor (ACEI) or angiotensin II receptor blocker (ARB) treatment (44% vs. 28%, p = 0.26).

At first evaluated echocardiography, patients in the reduced GLS group had as expected lower median GLS (14.7[11–16] vs. 20.7[19–22], p < 0.001) as well as lower EF (50[40–58] vs. 60[60–60], p< 0.001) and higher left ventricle end diastolic diameter (LVEDD) (51[46–54] vs. 46[43–49], p < 0.001) and left ventricle end systolic diameter (LVESD) (32[27–39] vs. 25[23–28], p < 0.001) (Table 1).

Cancer type and chemotherapeutic agents

The study population included different types of cancer (Table 2). Breast cancer was the most frequent type of cancer (55%), however, the reduced GLS group was seen to have a lower prevalence of this type of malignancy (31% vs. 60%, p < 0.001). On the other hand, lung cancer and hematologic cancer were more frequent among the reduced GLS group (15% vs. 3%, p = 0.005 and 25% vs. 12%, p = 0.013, respectively) (Table 2). Trastuzumab and pertuzumab (both recombinant humanized monoclonal antibodies against HER2) therapy were significantly used more frequently among the reduced GLS group (19% vs. 8%, p = 0.029 and 13% vs. 3%, p = 0.016, respectively). Similarly, the use of proteasome inhibitor and tyrosine kinase inhibitor therapy were more frequent among the reduced GLS group (6% vs. 1%, p = 0.033 and 8% vs. 1%, p = 0.016, respectively). No significant difference was observed regarding treatment with doxorubicin (a type of ANT) (21% vs. 18%, p = 0.657; Table 2).

Outcomes according to GLS

Overall, 45 (94%) patients in the reduced GLS and 192 (79%) patients in the preserved GLS performed follow-up echocardiography exam (Table 3). Over a median follow-up of 2.9 months (IQR 1.8–5.2) (Table 3), 11 (5%) patients developed CTRCD, with a substantially higher prevalence among the reduced GLS group (16% vs. 2%, p = 0.001) (Table 3; Fig. 1). Patients with preserved GLS had a significantly lower risk for CTRCD development (OR 0.11, 95%CI 0.03–0.41, p = 0.001), with every 1-unit improvement of GLS the risk of CTRCD decreased by 16% (OR 0.84, 95%CI 0.73–0.95, p = 0.007) (Table 4). After adjustment for baseline characteristics, including cardiovascular risk factors and systolic function, preserved GLS remained significantly associated with a lower risk for CTRCD development (OR 0.11, 95%CI 0.02–0.64, p = 0.014), with every 1-unit improvement lowering the risk by 19% (OR 0.81, 95%CI 0.67–0.98, p = 0.032). Interestingly, after adjusting for BB, ACEI/ARB treatment, GLS did not remain significantly associated to CTRCD (p = 0.166) (Table 4).

Table 3 Cancer therapy-related cardiac dysfunction according to global longitudinal strain groups
Full size table
Table 4 Global longitudinal strain as a predictor for adverse outcomes
Full size table

Similarly, reduced GLS was independently associated with any EF reduction (p = 0.004), remaining significant after adjustment for cardiovascular risk factors and systolic function (p = 0.049), and again, after adjustment for BB, ACEI/ARB treatment reduced GLS did not remain significant (p = 0.227) (Table 4).

One-year all-cause mortality was higher among patients with reduced GLS (17% vs. 1%, p < 0.001) (Table 5). Patients with preserved GLS had a significantly lower risk for all-cause mortality (OR 0.31, 95%CI 0.16–0.62, p = 0.001), with every 1-unit improvement of GLS, the risk decreased by 14% (OR 0.86, 95%CI 0.8–0.93, p < 0.001) (Table 4). However, after adjustment for cardiovascular risk factors and systolic function, GLS did not remain significantly associated to all-cause mortality (p = 0.132) (Table 4).

Table 5 One-year mortality according to global longitudinal strain groups
Full size table

Discussion

In the present study, we emphasized the importance of using GLS assessment routinely among patients receiving cancer therapy, which may allow early cardiac dysfunction diagnosis.

According to the Expert Consensus for Multimodality Imaging Evaluation of the American Society of Echocardiography [19], GLS is the optimal parameter of deformation for the early detection of subclinical LV dysfunction among patients with cancer. However, due to the lack of large randomized control trials, currently there is no evidence to guide specific GLS surveillance among patients receiving cancer therapy [4]. In our study, we found that reduced GLS is frequent (16%) and is related to high prevalence of ischemic heart disease, systolic dysfunction, atrial fibrillation, chronic renal failure and hyperlipidemia. However, no significant differences were noted regarding smoking, hypertension and diabetes mellitus. Past studies have also noted the correlation between cardiovascular risk factors, especially chronic renal failure, and reduced GLS [22, 23].

Doxorubicin and trastuzumab are well studied and known to cause cardiotoxicity [19], expressed by LVEF and GLS reduction [8, 24, 25]. As expected, in our study trastuzumab treatment was significantly more frequent among the reduced GLS group, however, no differences were noticed regarding doxorubicin therapy. This finding may be explained by the fact that the majority of patients treated with doxorubicin in our registry, which has a dose-dependent risk of developing cardiotoxicity [4], were breast cancer patients who are exposed to lower therapeutic doses (240 mg/m2) and therefore a low correlation was noticed in our study. Similarly to trastuzumab, treatment with pertuzumab, proteasome inhibitor and tyrosine kinase inhibitor were more frequent among the reduced GLS group, which may imply the need for routine follow-up among that specific population, however; larger trials are needed to support this data.

A number of studies, including the PRADA [26] and OVERCOME [27] trials, implied that routine baseline use of BB, ARB and ACEI provides protection against early decline in global LV function. However, currently there is no evidence to guide specific cardio-protection treatment according to GLS surveillance [28]. Interestingly, in our trial we observed a significantly high BB use among the reduced GLS group. This discrepancy can be attributed to the fact that the BB treatment was administered not as a result of the reduced GLS, but rather due to the high prevalence of cardiac morbidities in the mentioned group.

Past studies [13, 17, 29] have shown an association between reduced GLS and LV dysfunction, however, most of the studies were small, retrospective and mainly included breast cancer patients. Our study is novel through evaluating prospectively a large population with diverse types of cancer. We demonstrated that preserved GLS is associated with a lower risk for CTRCD development and, furthermore, with any 1-unit improvement in GLS, the risk of CTRCD decreased. Importantly, this study’s increased strength, compared to past studies [30], comes from using multivariable analyses adjusted for a propensity score which was modeled from all baseline characteristics, showing that the association remained significant through the adjustment to cardiac risk factors and systolic function. However, adding BB, ACEI/ARB treatment to the model evoked its significance, which may imply that cardio-protective treatment may prevent CTRCD development.

Using univariable analysis, reduced GLS was associated with all-cause mortality, however, after multivariable analyses adjustment, the association did not remain significant. Regrettably, the specific causes of death for most of the patients were unknown since it occurred out of our hospital. The relation of GLS to all-cause mortality was implied in the past [30] and may be explained as GLS being a marker for severe disease, elevated inflammatory cytokines [31] and overall cardiac stress. This may also support that after adjustment to cardiac risk factors, reduced GLS did not emerge associated with all-cause mortality; however, this can also be explained by the small number of deaths.

Our study has several limitations. First, it was a single center study. Second, we evaluated the patients at different time points of their therapy and therefore outcome data may not account for patients who developed GLS reduction or CTRCD later. Finally, the relatively short period of difference between echocardiography assessments might have influenced the prevalence of CTRCD development and all-cause mortality.

In summary, our study shows that reduced GLS is frequent among patients receiving cancer therapy and specifically among patients with cardiac risk factors or cancer therapy such as trastuzumab, pertuzumab, proteasome inhibitor and tyrosine kinase inhibitor, which may imply the need for close GLS follow-up among that specific population. Moreover, we implied that reduced GLS is associated with CTRCD development, independently of comorbidities and LVEF. Interestingly, our data suggest that treatment with BB, ACEI/ARB may prevent CTRCD development among patients with reduced GLS.