Cardiovascular Complications of Cancer Therapy: Best Practices in Diagnosis, Prevention, and Management—Part 1

Anthracyclines

In a retrospective review of three trials, the incidence of doxorubicin-related HF was found to be 5% at a cumulative dose of 400 mg/m², 16% at a dose of 500 mg/m² and 26% at a dose of 550 mg/m² (15). However, subclinical events occurred in about 30% of the patients, even at the doses of 180–240 mg/m², about 13 years after the treatments (16). Interestingly, histopathologic changes, such as myofibrillar loss and vacuolization, can be seen in endomyocardial biopsy specimens from patients who have received as little as 240 mg/m² of doxorubicin (17). These findings suggest that there is no safe dose of anthracycline. Even doses as low as 100 mg/m² have been associated with reduced cardiac function (16,18). Nonetheless, some patients had no significant cardiac complications despite receiving doses as high as 1000 mg/m² (19). Individual susceptibility is most likely due to genetic variants in genes that regulate anthracycline cardiotoxicity (20). Other risk factors for anthracycline toxicity include cumulative dose, intravenous bolus administration, higher single doses, history of prior irradiation, use of concomitant agents known to have cardiotoxicity, female gender, underlying CV disease, age (young and elderly), delayed diagnosis, and increase in cardiac biomarkers such as troponins during and after administration (9,21–23).

Doxorubicin poisons topoisomerase 2 to cause DNA double strand break and demise of cancer cells. In the cardiomyocytes, doxorubicin targets topoisomerase 2β to induce DNA double-strand breaks and doxorubicin-bound topoisomerase 2β binds promoters of anti-oxidative and electron-transport genes to reduce their transcripts and protein expression (24,25) (Figure 1). Therefore, doxorubicin-treated cells have a marked increase in reactive oxygen species (ROS) and are defective in mitochondria biogenesis. Thus, topoisomerase 2β accounts for the three hallmarks of anthracycline-induced cardiotoxicity: myocyte death, ROS generation, and mitochondriopathy. Reduced topoisomerase 2β expression has been linked to a coding variant in the retinoic receptor γ gene, which predicts susceptibility to anthracycline-induced cardiotoxicity in childhood cancer (26).

Alkylating Agents

Alkylating agents add an alkyl group to the DNA of rapidly dividing cells and in the case of bifunctional alkylating agents, cross-linking the two DNA strands, thereby inhibiting DNA replication and cell proliferation (27). Symptoms may include arrhythmias and conduction disorders, and fulminant HF (28,29). Alkylating agents such as cyclophosphamide induce electrocardiogram (ECG) alterations in the form of low QRS voltage and non-specific T-wave or ST segment abnormalities (28,30,31). Acute symptoms usually occur within 1–2 weeks, last for a few days, and in some patients resolve without any late consequences (28,29,31). The incidence of high-dose cyclophosphamide-induced HF has been reported in as high as 28 % (32). Autopsy of patients with cyclophosphamide-induced cardiotoxicity shows hemorrhagic myocardial necrosis with interstitial edema and fibrin deposition. Complication is higher in elderly patients and in those exposed to anthracycline or mediastinal irradiation (1,29,31).

Ifosfamide is an alkylating nitrogen mustard used for the treatment of lymphomas, sarcomas, and testicular, breast, and lung carcinoma (33). There is a dose-dependent increase in HF associated with ifosfamide administration (33,34). Autopsy studies demonstrated increased heart weight, small pericardial effusions, subendocardial hemorrhage and petechial lesions in the epicardium (29).

Targeted therapies against HER-2 Pathway

Trastuzumab is a monoclonal antibody against the human epidermal growth factor receptor tyrosine kinase (HER2 or ErbB2), which regulates cell growth and repair (35). Overexpression of HER2 occurs in approximately 25% of breast cancers and confers increased proliferative and metastatic potential. HER-2 is expressed in cardiomyocytes and required for survival of cardiomyocytes (36). Mice deficient in HER-2 develop dilated cardiomyopathy and are more sensitive to doxorubicin (36).

Trastuzumab administration in HER2-positive breast cancers led to significant reductions in recurrence rate and overall mortality. A pivotal study in the metastatic setting demonstrated a 33% reduction in mortality at 1 year and an increase in median survival by 5 months (37). HF occurred in 8% of patients who received an anthracycline with cyclophosphamide; however, the incidence of HF increased to 27% with the addition of trastuzumab (37). Subsequently, several large trials confirmed efficacy of trastuzumab in increasing disease-free survival from cancer, but also established trastuzumab’s association with HF (38,39). In these trials, 1.7–4.1% of trastuzumab-treated patients developed HF when anthracycline was not part of the therapeutic regimen (38,39). Trastuzumab-related cardiotoxicity includes various degrees of LV systolic dysfunction, occasionally leading to HF (40). Symptoms are usually mild or moderate and improve following medical management and termination of drug administration (41,42). Improvement is usually seen in about four to six weeks after trastuzumab withdrawal (43). After symptomatic improvement, re-institution of trastuzumab treatment is usually possible (41–43).

Other targeted therapies against HER-2 are relatively less cardiotoxic (44–46). Perez et al (46) reviewed cardiac safety data in 44 clinical trials that used lapatinib. Of these study patients, only 1.6% experienced a cardiac event and of those only 0.2% was symptomatic. The rate of cardiac events in the lapatinib group was similar to those who did not receive lapatinib (44,46). Similarly, there was no significant increase in left ventricular dysfunction with the addition of pertuzumab to trastuzumab in the NeoSphere study (47,48). In the CLEOPATRA trial, the incidence of cardiac adverse events was 14.5% in the pertuzumab plus trastuzumab plus docetaxel arm compared to 16.4% in the control-placebo plus trastuzumab plus docetaxel arm (49,50). However, in a head-to-head comparison between lapatinib and trastuzumab, patients treated with lapatinib have shorter disease-free survival, but more non-cardiac toxicity, such as rash and diarrhea (51). Thus, one must consider both efficacy and toxicity in choosing drugs.

Vascular Endothelial Growth Factor Signaling Pathway Inhibitors

VEGF signaling pathway (VSP) inhibitors include antibody such as bevacizumab that binds to VEGF and small molecule tyrosine kinase inhibitors such as sunitinib and sorafenib, which inhibit the downstream kinase involved in VEGF receptor signaling (52). Cardiovascular side effects of VSP inhibitors include HTN, cardiomyopathy, conduction abnormalities, acute coronary syndromes, and arterial thromboses. Several VSP inhibitors also block receptors that are involved in compensatory response to stress in the cardiomyocytes. When the heart is unable to compensate for HTN-induced by VSP inhibitors, it could lead to HF (53). Therefore, maintaining good blood pressure control during VSP inhibitor therapy can prevent HF (54).

Bevacizumab is the first- or second-line chemotherapy for advanced solid tumors (55–61). Its use has been associated with HTN, thromboembolism, and cardiomyopathy (62). Approximately 2%–4% of patients treated with bevacizumab will develop HF (29). Predisposing factors include previous therapy with cardiotoxic chemotherapy drugs, such as anthracyclines (63), capecitabine (64), as well as radiation to the mediastinum (29). In a meta-analysis of VSP inhibitors and its role in cardiovascular complications, the odds ratio for cardiac dysfunction caused by VSP inhibitors was 1.35(65). . More than half (40 out of 77) of the clinical trials included in the analysis utilized bevacizumab as the drug of treatment (65). Another meta-analysis showed that bevacizumab resulted in a significant decrease in EF with relative risk index of 3.4 (66).

In a large trial of sunitinib for the treatment of metastatic renal cell carcinoma, 13% developed LV dysfunction and 3% had severe HF (67). Severe HF was more common in a single-center study of 48 patients treated for metastatic renal cell carcinoma or gastrointestinal stromal tumor (GIST) (68). In a meta-analysis of 6935 patients treated with sunitinib, the incidence of HF was 4.1 % with a relative risk of 1.81 compared to placebo (69). In a retrospective analysis of a phase I/II trial of sunitinib for the treatment of GIST, there was a 2.5 fold increase in cardiotoxicity with higher doses (70). Sorafenib has been associated with a 4 % risk of mild LV dysfunction and 1 % risk of clinical HF (71). HF due to VSP inhibitors is reversible with cessation of therapy (72).

Proteasome Inhibitors

Proteasome is a protein complex present in all cells that degrades other proteins. Inhibition of proteasome blocks cell proliferation and induces apoptosis in tumor cells, especially in multiple myeloma. Bortezomib, a reversible proteasome inhibitor, is not known to cause HF (73–75). However, 7% patients treated with carfilzomib developed new onset or worsening of preexisting HF or myocardial ischemia (76). In a phase II trial of 266 patients treated with carfilzomib monotherapy for relapsed myeloma, 10 experienced heart failure (3.8%), four (1.5%) had a cardiac arrest, and two (0.8%) had a MI during the study (77). The risk did not appear to be cumulative, at least through 12 cycles of therapy. Cardiotoxicity appears to be largely reversible with cessation of therapy and the initiation of HF treatment (78).

Baseline Risk Assessment

Patients undergoing chemotherapy should have careful clinical evaluation and assessment of CV risk factors, such as coronary artery disease, diabetes, and hypertension (Table 2) (2–6). These risk factors should be managed according to the ACC/AHA guidelines (79). This is especially important if cancer therapy is known to cause HF. An aggressive HTN management is advised for patients treated with VSP inhibitors. Physical exercise has been shown to reduce cardiotoxicity in mouse models (80).

EF assessment is mandatory to establish a baseline cardiac function before cardiotoxic cancer treatment (3). Echocardiography is the preferred modality for assessment of cardiac structure and function. MUGA scan has less inter-observer variability; however, radiation exposure limits its utility as a cardiac monitor. Magnetic resonance imaging (MRI) can be used to obtain a precise EF; however, the use of MRI is limited by its cost. Cardiac biomarkers, such as the troponins and brain natriuretic peptides (BNP), can be used to monitor for the development of cardiac dysfunction (3,5,6). However, it is not known whether routine monitoring of biomarkers is useful in changing clinical outcomes. A number of composite scores have been designed to risk-stratify cancer patients (81–83); however, they have not been validated in prospective studies.

Echocardiogram

Echocardiogram is the most important tool for serial evaluation of the heart during cancer therapy. EF should be determined using biplane method of discs according to American Society of Echocardiography guideline. If the endocardial border is not distinct, ultrasonic contrast should aid in endocardial border definition and subsequent volume calculations. However, temporal EF variability may be up to 10% with the confidence interval of 95% in 2D EF readings (84). 3D echocardiography has a temporal variability of 6% and is considered the most accurate EF measurement by echocardiography. Since CIMP is defined as a drop of EF of 10% or more or 5% or more in presence of heart failure symptoms, an accurate measurement of EF is paramount.

Myocardial Strain

Tissue Doppler imaging (TDI) and speckle-tracking strain imaging have emerged as two quantitative techniques for estimating global and regional myocardial mechanical function and have the potential to detect early signs of LV dysfunction (85). However, TDI is both user- and angle-dependent and is unable to differentiate translational motion or tethering effects from myocardial contractility. Speckle-tracking echocardiography (STE) is an angle-independent technique that uses an image-processing algorithm for analyzing motion of “speckles” or “fingerprints” within a 2D echo image, and it has replaced TDI strain as the preferred method for quantitative assessment of cardiac deformation (86,87).

Several studies have evaluated the utility of strain imaging for the detection of chemotherapy-associated cardiotoxicity. Fallah-Rad et al (88) evaluated 42 patients with breast cancer overexpressing HER-2 receiving trastuzumab in the adjuvant setting after anthracycline therapy. Within 3 months, peak global longitudinal and radial strain detected preclinical changes in LV systolic function before a decrease in EF was observed several months later. A recent prospective multicenter study by Sawaya et al demonstrated that global longitudinal strain (GLS) <19% was predictive of subsequent cardiotoxicity and was present in all patients who later developed symptoms of HF (89). Negishi et al similarly showed that a ≥11% relative reduction in GLS was predictive of subsequent trastuzumab-associated cardiotoxicity (90).

Abnormalities in strain parameters can also be seen several years after a cardiotoxic exposure. In 75 asymptomatic breast cancer survivors who received anthracycline with or without adjuvant trastuzumab, GLS was significantly decreased in the chemotherapy group 6 years after therapy compared with control subjects (91). In another meta-analysis, GLS consistently detected early myocardial changes during therapy (92). A 10% to 15% early reduction in GLS during therapy appears to be the most useful parameter for the prediction of cardiotoxicity. In late cancer survivors, global radial and circumferential strain are consistently abnormal, even with normal EF, but their clinical value in predicting subsequent ventricular dysfunction or HF has not been explored (92).

Biomarkers

The use of EF in the diagnosis of CIMP has important limitations. First, the measurement of EF is subject to technique-related variability, which can be higher than the thresholds used to define cardiotoxicity (14,84). Second, the reduction in EF is often a late phenomenon (11,88,93,94). Hence, there is a growing interest in identifying markers of early myocardial damages to predict the development of HF. Biomarkers are an economic and effective way of detecting myocardial dysfunction in apparently asymptomatic patients. The assessment of troponin and brain natriuretic peptide have been shown to be of incremental utility in identifying patients at increased risk for adverse outcomes (95–97).

Selecting a non-anthracycline regimen

A randomized study of 3222 women with HER-positive early breast cancer found that a non-anthracycline containing regimen has equal efficacy and less cardiotoxicity (122). With a non-anthracycline regimen containing docetaxel, carboplatin, and trastuzumab (TCH), patients have a 5 year disease-free survival rate of 81% compared to 84% in a anthracycline-containing regimen (ACT or ACT-H). Importantly, the TCH regimen has much lower cardiotoxicity than ACT or ACT-H. Thus, TCH was proposed as an alternative non-anthracycline containing regimen for HER-positive early breast cancers (Table 2).

Substituting doxorubicin with less-cardiotoxic anthracyclines

Over the past five decades, more than 2000 modified anthracycline chemicals have been tested in attempt to reduce cardiotoxicity while retaining tumoricidal efficacy. While several anthracycline derivatives have been evaluated in clinical trials, only epirubicin (123) and idarubicin (124) received approval for clinical use. However, a critical analysis by Cochrane database identified no difference in cardiotoxicity between epirubicin and doxorubicin at equipotent doses (125).

Continuous Infusion

Replacing bolus administration with slow infusions does not significantly affect anthracycline area under the curve but diminishes anthracycline Cmax and anthracycline accumulation in the heart (126). A Cochrane review (127) showed a significantly lower rate of clinical HF with an infusion duration of 6 hours or longer as compared to a shorter infusion duration in the adults. In the pediatric populations, the results of infusion of anthracycline have been disappointing. A randomized trial in children with high-risk acute lymphocytic leukemia found that continuous infusion offered no additional cardiac protection over bolus administration in a median follow-up of 8 years post-diagnosis (128). A follow-up at 10 years also revealed no incremental therapeutic efficacy for infusion (129). Thus, continuous infusion cannot be recommended in the pediatric population (130,131).

PEGylated Liposomal doxorubicin

PEGylated liposomal doxorubicin comprises an aqueous core of doxorubicin hydrochloride encapsulated in liposomes with a protective hydrophilic outer coating of surface-bound methoxypolyethylene glycol (132,133). Delivery of doxorubicin in a PEGylated liposomal form decreases the circulating concentrations of free doxorubicin and results in selective uptake of the agent in tumor cells. In randomized trials, PEGylated liposomal doxorubicin was as effective as doxorubicin or other traditional combination chemotherapies (134–136) (137–139). Thus, PEGylated liposomal doxorubicin is a useful option in the treatment of various malignancies (140). However, the cost associated with administering this drug has prevented its widespread adoption (141).

Dexrazoxane

Dexrazoxane was originally developed as an anticancer agent (142). Using fibroblasts from topoisomerase2β knockout mice, Lyu et al showed that dexrazoxane is protective against anthracycline-induced toxicity in topoisomerase 2β-dependent manner, linking dexrazoxane to the topisomerase2β theory of anthracycline-cardiotoxicity (25,143) (Figure 1).

The protective effect of dexrazoxane against anthracycline-induced cardiotoxicity has been demonstrated in numerous clinical trials and in adults and children (19,104,144–146). Dexrazoxane was approved in Europe and the United States for cardioprotection in patients treated with anthracyclines (Cardioxane and Zinecard) with several generic preparations available (Procard and Cardynax). In addition, dexrazoxane has been also approved for treatment of accidental extravasation of anthracyclines (Savene).

Unfortunately, one phase III trial suggested that dexrazoxane may lower the efficacy of anthracycline in treating breast cancer (19). In this trial, a significant difference in objective response was reported (47% vs. 61%, respectively, p = 0.019). While high response in the placebo group was quite unusual, other endpoints (including survival or time to progression) were not affected by dexrazoxane in this study (19,147). Careful meta-analyses of all available randomized clinical trials found no evidence that dexrazoxane lower doxorubicin’s anti-cancer effect (148,149). However, FDA approves dexrazoxane only in patients who have received more than 300 mg/m² for metastatic breast cancer and who may benefit from continued doxorubicin treatment (150).

Another controversy about dexrazoxane pertains to an increased risk of second malignancies in the pediatric cancer survivors. This was observed in survivors of Hodgkin lymphoma who had received dexrazoxane in combination with doxorubicin and etoposide. It was postulated that combining these drugs could exceed a threshold above which topoisomerase inhibitors caused genetic instability in normal tissues (151). This report led the European Medicine Agency to disapprove the use of dexrazoxane in children. However, two studies of survivors of childhood acute lymphoblastic leukemia did not detect an increased risk of second malignancies from dexrazoxane (152,153). Thus, risk/benefit analysis supports a wider clinical usage of dexrazoxane with the possible exception in patients receiving etoposide or etoposide-anthracycline combinations.

Antimetabolites

Chest pain is the most common symptom associated with 5-fluoracil (5-FU) administration. Less common side effects include myocardial infarction (MI), arrhythmias, HF, cardiogenic shock, and sudden death (159,160). High doses 5-FU (>800 mg/m²) and continuous 5-FU infusions are recognized risk factors of cardiotoxicity (7.6%) as compared to bolus injections (2%) (159,161,162). Other risks factors include preexisting cardiovascular disease, prior mediastinal radiation, and concurrent use of other chemotherapeutic agents (160,161,163–169). Cardiac events are typically short-lasting (up to 48 hours) and tend to manifest within 2 to 5 days after initiation of 5-FU (159). Ischemic alterations of the ECG have been reported in 68% of patients and associate with biomarkers elevations in 43% of the cases (162). The overall mortality ranges from 2.2% to 13% (161,165,170).

Selective activation of capecitabine (oral prodrug of 5-FU), which occurs preferentially in cancer cells, explains its lower cardiotoxicity compared to 5-FU. The incidence of capecitabine-induced cardiotoxicity ranged between 3 and 9% (160,171–173). One prospective study enrolling 644 patients with no prior history of coronary artery disease reported the occurrence of symptomatic ischemic changes of the ECG in 5.2% of patients receiving capecitabine (166). ECG changes, including ST-segment elevation, were seldom associated with elevation of serum cardiac markers (162,164,174,175).

Anti-microtubule agents

Myocardial ischemia and infarction have been reported in approximately 3% of the patients receiving paclitaxel. In 198 patients treated with paclitaxel for ovarian cancer, 0.5% experienced a MI (176). In the Cancer Therapy Evaluation Program’s Adverse Drug Reaction database, which followed more than 3400 patients, the overall incidence of grade 4 and 5 cardiac events was only 0.29% (176). These events occurred within two weeks from paclitaxel administration (176) and were observed in patients with known cardiac risk factors including HTN and CAD.

Cardiac ischemia has also been reported after administration of docetaxel. In a clinical trial with inoperable, locally advanced squamous cell carcinoma of the head and neck, 355 patients were randomized to receive a standard regimen of cisplatin and 5-FU or the same regimen plus docetaxel. Myocardial ischemia occurred in 1.7% of patients from the docetaxel arm and in 0.6% of patients from the control arm (177). Although not reported in the original study, cardiac ischemia was listed as a side effect of docetaxel the package insert (178).

Antibody-based VEGF inhibitor

In a pooled analysis of 1,745 cancer patients from 5 randomized controlled trials, the overall incidence of arterial thrombotic events (ATEs) was significantly higher in the bevacizumab group (3.8%) as compared to the control group (179). In the Bevacizumab Regimens’ Investigation of Treatment Effects (BRiTE) study, 1,953 patients receiving bevacizumab at 248 U.S. sites were followed for over 20 months (180). The overall incidence of ATEs was 2%. Of these events, >57% were cerebrovascular cardiovascular accidents (CVAs) and transient ischemic attacks (TIAs), while 30% were MI (50% lethal) (180). Independent risk factors for ATEs included age ≥75 years, poor performance status (ECOG performance status score ≥1), preexisting HTN, anticoagulation therapy, and arterial disease (180).

Small molecular tyrosine kinase inhibitors

In a double-blind, international phase III trial enrolling 569 patients with un-resectable, locally advanced, or metastatic pancreatic cancer, CVAs and MI occurred in 4.6% patients randomized to receive gemcitabine plus erlotinib compared with 1.2% patients assigned to gemcitabine plus placebo (181). Although the incidence of CVA/MI was not published in the original study by Moore et al., it was reported by the FDA in the erlotinib package insert (182).

Approximately 3% patients in clinical trials have experienced myocardial ischemia with sorafenib. In an unpublished clinical trial, MI/ischemia occurred among 2.7% of hepatocellular cancer patients treated with sorafenib compared with 1.3% of patients in the placebo group (183). Similarly, sorafenib was associated with a higher incidence of MI/ischemia compared with placebo in patients treated for renal cell carcinoma (3% vs. <1%) (184). The incidence of cardiac ischemia/MI in patients with different types of cancer receiving sorafenib ranges from 2 to 3% (183).

Proteasome inhibitors

The safety profile of carfilzomib was analyzed by Siegel et al. in 526 patients from 4 phase II clinical trials, based on which the use of carfilzomib was approved by the U.S. FDA (185). In this analysis, the overall incidence of myocardial ischemia was reported to be 3.4%. Significantly higher occurrence of ischemic events was reported in two subsequent phase III, randomized, multicenter trials, the ASPIRE and ENDEAVOR studies. In the ASPIRE study, the rate of ischemic heart disease was 5.9% versus 4.6% in the control group (186). Likewise, in the ENDEAVOR study, ischemic events occurred at a rate of 3% as compared to 2% in the control group (187). Hence, high-risk patients should be considered for an ischemic work-up prior to starting carfilzomib treatment (187).