Comparison of Risk Factors for Trastuzumab-Related Cardiotoxicity in Elderly and Non-Elderly Breast Cancer Patients

Background and Objective: Although trastuzumab-based treatment benefits patients with HER2-positive early and metastatic breast cancer improving both disease free and overall survival, cardiotoxicity has been a crucial issue. The purpose of this study is to investigate predictive risk factors for trastuzumab-related cardiotoxicity in elderly and non-elderly patients. Patients and Methods: We reviewed medical records of 260 patients treated with trastuzumab for breast cancer from January 2008 to March 2015. Cardiotoxicity was defined as a decrease of the left ventricular ejection fraction (LVEF) either>10% from the baseline or to a value <53%. The independent risk factors were determined from clinicopathological and echocardiographic parameters, using univariate and multivariate logistic models, and Receiver Operating Characteristic (ROC) curve analysis. Results: Fifty-seven elderly (≧ 65 years old) and 203 non-elderly patients (≦ 64 years old) were included. The baseline LVEF was significantly related to cardiotoxicity both in elderly and non-elderly patients. The optimal cutoff values were 67% for elderly (the area under the curve at 0.77) and 66% for non-elderly patients (the area under the curve at 0.76). The baseline LVEF value of less than 67% and 66% in elderly patients (or 5.59, 95% CI: 1.08–33.24; p=0.04) and non-elderly patients (OR 4.61, 95% CI: 1.48-17.50; p<0.01) respectively was a significant predictive factor of trastuzumab-related cardiotoxicity. Conclusions: A lower value of baseline LVEF may be the common risk factor for trastuzumab-related cardiotoxicity both in elderly and nonelderly breast cancer patients.


Introduction
Epidermal growth factor receptor 2 (HER2) is expressed in approximately 25% of all breast cancers [1].Trastuzumabbased chemotherapy significantly improves the prognostic outcomes of patients with early stage and metastatic HER2positive breast cancer [2,3].The use of trastuzumab is related to an increased risk of cardiotoxicity, which is attributed to blocking HER2 signaling in cardiac myocytes [4].Cardiotoxity ranges from 2% to 7% for trastuzumab monotherapy, 2% to 13% for trastuzumab combined with taxane, and can be as high as 27% in combination with anthracycline [5].
Previous reports of large-scaled clinical trials have suggested that older age and cardiac comorbidity are important risk factors associated with cardiotoxity among trastuzumab-treated patients [6,7].Therefore, trastuzumab-based chemotherapy is preferentially avoided for elderly patients in consideration of the risk of cardiotoxicity, while some can safely receive the treatment with trastuzumab.However, little is known regarding what factors are associated with cardiotoxicity in elderly patients.A few studies have suggested that some clinical factors such as history of cardiac disease, diabetes, hypertension, and obesity are associated with trastuzumab-related cardiotoxicity [8,9], but the potential risk factor in elderly patients warrant further investigations.
For decades, echocardiography has been a major modality for evaluation of cardiac function.Left ventricular ejection fraction (LVEF) is a crucial parameter recommended for use in clinical decision-making, diagnosis and outcome evaluation by major guidelines and trials [10,11].Previously reported, trastuzumab-related cardiotoxicity occurs more frequently in patients with a lower baseline LVEF [12,13].It is not yet clear whether the lower baseline LVEF is an independent risk factor in elderly people.
In this study, we reviewed medical records of breast cancer patients treated with trastuzumab retrospectively, and examined clinical and cardio functional parameters to find Receiver Operating Characteristic (ROC) curve analysis was applied to determine the cut off value of the baseline LVEF value for prediction of cardiotoxicity.

Results
A total of 260 patients including 57 elderly and 203 nonelderly (≦ 64 years old) patients were enrolled in this study.Table 1A and 1B show the baseline clinicopathological characteristics of those.Eight (14%) of the elderly patients and 19 (9%) of the non-elderly patients had trastuzumab-related cardiotoxicity.In elderly patients, the average LVEF value at baseline was significantly lower in those with cardiotoxicity than those without (p=0.02).In non-elderly patients, the average LVEF values at baseline (p<0.01),BMI (p=0.04), and E/e' at baseline (p=0.05), were significantly lower in those with cardiotoxicity than those without.
In the univariate analysis table 2, low LVEF values at baseline (p=0.05),some cardiovascular diseases (p=0.16),history of anthracycline exposure (p=0.25), and adjuvant usage of trastsuzumab (p=0.26) were relatively associated with cardiotoxicity in elderly patients.On the other hand, in nonelderly patients, low LVEF values at baseline (p<0.01),low BMI (p=0.02), and low E/e' at baseline (p=0.03) were identified as potential risk factors of trastuzumab-related cardiotoxicity in elderly patients by comparing with those in non-elderly ones.

Patient population and methods of treatment and observation
This was a single institutional retrospective study.Medical records of patients who were treated at the Department of Breast and Endocrine Surgery, Osaka Medical Center for Cancer and Cardiovascular Diseases from January 2008 to March 2015 were reviewed.HER2 positivity was defined as 3+ staining in immunochemistry or a positive result of fluorescence in situ hybridization (FISH) in the case of 2 + staining according to the ASCO/CAP guideline.Patients who were treated with trastuzumab for any indication (adjuvant and metastatic) were included in the study.Patients who had been given trastuzumab as an adjuvant treatment previously and were treated for recurrence during the period were also included.An elderly patient was defined as one who was 65 years old or more, according to the statement by the International Conference on Harmonization [14], on the first day of starting the treatment.Trastuzumab was administered every 3 weeks (at an 8 mg/kg loading dose followed by 6mg/kg) or weekly (at a 4mg/kg loading dose followed by 2mg/kg).All patients in this study were assessed for LVEF by 2D echocardiography scan before and every three months after starting the treatment by a team of experienced cardiologists.Cardiotoxicity during the treatment was defined as a decrease of LVEF of >10% or to a value <53% according to the definition by the American Society of Echocardiography and the European Association of Cardiovascular Imaging [15].We also examined the E/e' ratio (the ratio of early diastolic transmitral flow velocity to mitral annular velocity at the lateral wall), which is reported to be an echocardiographic parameter for predicting cardiotoxicity of trastuzumab treatment [16].Other clinicopathological data including age at treatment initiation, body mass index (BMI), estrogen receptor (ER) status, history of hypertension, diabetes mellitus, hyperlipidemia, hyperuricemia and cardiovascularrelated disease, and previous therapeutic regimen or radiation therapy were retrieved from the medical records.

Statistical analysis
The information was recorded in an Excel database and analyzed using JMP pro statistical software (Version 11.0.0,SAS Institute Inc., and Cary, NC, USA).To assess baseline characteristics, categorical variables were compared using the Fisher's exact probability test while continuous variables were analyzed using the Mann-Whitney U-test whenever appropriate.Odds ratios (or) and their 95% confidence intervals calculated from logistic regression analysis were used to evaluate the relationship of risk factors to the development of cardiotoxicity.All relevant variables being significant (p<0.05) or showing a tendency towards significance in the univariate analysis were analyzed using a multivariate logistic model.a significant associated factor.The predictive accuracy of the baseline LVEF values for trastuzumab-related cardiotoxicity was then determined by measuring the sensitivity, specificity, and area under the Receiver Operating Characteristic (ROC) curve (Figure 1).Based on the ROC curve, the optimal cutoff value of baseline LVEF as an indicator for trastuzumab-related cardiotoxicity in elderly patients was projected to be 67%, which yielded a sensitivity of 62% and a specificity of 82%, with the area under the curve a 0.77 (Figure 1A).On the other hand, the cutoff value of baseline LVEF in non-elderly patients was 66%, which had a sensitivity of 79% and a specificity of 66%, with the area under the curve at 0.76 (Figure 1B).
Table 3 shows the result of multivariate analysis including baseline LVEF and other established relatively associated factors.In elderly patients, there was an increased prediction ability of cardiotoxicity associated with a lower baseline LVEF value < 67% (OR 5.59, 95% CI: 1.08-33.24;p=0.04) after adjusting for the above possible confounders.Similarly, in nonelderly patients, a lower baseline LVEF value < 66% (OR 4.61, 95% CI: 1.48-17.50;p<0.01) was with trastuzumab-related cardiotoxicity.

Discussion
Understanding the risk factors for cardiotoxicity is very important for safer treatment with trastuzumab in breast cancer patients.In the current study, we demonstrated that the baseline LVEF level was the most significant risk factor for trastuzumab-related cardiotoxicity.Interestingly, each predictive value for cardiotoxicity of baseline LVEF in elderly and non-elderly patients was found to be approximately similar.To the best of our knowledge, there are no reports prior to this study comparing the risk factors in non-elderly with that in the elderly.This is the first report thus far demonstrating that a low value of baseline LVEF is the common risk factor for trastuzumab-related cardiotoxicity in elderly and non-elderly breast cancer patients.
Previously, Bowels et al. showed that the risk of cardiotoxicity was associated with adjuvant anthracycline and/or trastuzumab in a population-based cohort study [6].Chavez-MacGregor et al reported that older age, coronary artery disease, hypertension, and weekly trastuzumab administration increased the risk of cardiotoxicity among trastuzumab-treated patients [7].In the studies investigating risk factors in elderly patients, history of cardiac disease, diabetes, hypertension, obesity, and prior anthracycline use were associated with a higher incidence of cardiotoxicity [8,9].In our study, history of cardiac disease and prior anthracycline use were also identified as risk factors that were likely to be associated with cardiotoxicity in elderly patients.These findings are similar to those seen in previous reports [8,9].However, in multivariate analysis, we found that a lower baseline LVEF value was the most significant factor in both elderly and non-elderly patients.This result was different from those reports, because patients with low values of baseline LVEF may be excluded in such clinical trials.The baseline LVEF may be conveniently evaluated by means of echocardiography, and is considered a superior predictive marker of cardiotoxicity [17] than others.It may thus be useful to assess whether the treatment of trastuzumab-based chemotherapy is safely performed for the patient.Patients with low LVEF at baseline have to undergo careful monitoring of cardiac activity for a certain Chaves-MacGregor et al., recommends increasing the rates of cardiac monitoring for elderly patients, and adequate cardiac monitoring among trastuzumab-treated patients should be considered a marker of quality of care [18].However, no clear universal guidelines for cardiac monitoring exist for patients receiving trastuzumabbased chemotherapy.For instance, the Canadian Trastuzumab Working Group recommends that carefully selected patients receive a clinical and LVEF assessment at baseline and every three months during treatment [19].Whereas, Hart A et al., suggest that routine intensive monitoring for cardiac function is not needed after completion of trastuzumab treatment [20].There are several limitations to this study.One limitation is that serum biomarkers such as Troponin I and N-terminal pro-B type natriuretic peptide (NT pro-BNP) were not included in the examination risk factors.Troponin I is a well-established early serum biomarker of myocardial necrosis with high diagnostic and high prognostic values [13,21].NT pro-BNP levels may be used to identify the risk of developing cardiac dysfunction during trastuzumab-based treatment [22].As such, it will be necessary to examine kinds of risk factors including serum biomarkers in future.Another limitation is that the possibility of selection bias as some elderly patients with low LVEF at baseline have not been treated with trastuzumab-based chemotherapy, but we believe that this result might reflect real data in the routine clinical care for breast cancer.
In this study, we showed that a lower value of baseline LVEF is the common risk factor of trastuzumab-related cardiotoxicity both in elderly and non-elderly breast cancer patients.We conclude that if cardiac function is maintained, even elderly might be able patients to receive trastuzumabbased chemotherapy under careful observation.

Figure 1 :
Figure 1: Receiver Operating Characteristic (ROC) curve for baseline LVEF values to predict trastuzumab-related cardiotoxicity in elderly (A) and non-elderly (B) patients.(A) ROC analysis showed that the optimal cutoff value for change in baseline LVEF value was 67%.(B) ROC analysis showed that the optimal cutoff value for change in baseline LVEF value was 66%.

Table 1A :
Comparison of the factors between non-cardiotoxicity group and cardiotoxicity group in elderly patients.
* Average a Fisher's exact probability test, b Mann-Whitney U-test BMI: body mass index, ER: estrogen receptor E/e': the ratio of early diastolic transmitral flow velocity to mitral annular velocity at the lateral wall

Table 1B :
Comparison of the factors between non-cardiotoxicity group and cardiotoxicity group in non-elderly patients.

Table 2 :
Univariate logistic regression analysis to identify the risk factors for cardiotoxicity induced by trastuzumab in elderly and non-elderly patients.

Table 3 :
Multivariate logistic regression analysis to identify the risk factors for cardiotoxicity induced by trastuzumab in elderly and non-elderly patients.