The increase in rates of clinical heart failure in pivotal trials of trastuzumab in women with metastatic disease, especially those who were receiving or who had recently received anthracyclines, prompted establishment of the Cardiac Review and Evaluation Committee or CREC. Initial CREC criteria used to describe cardiac dysfunction and subsequently employed in many adjuvant trials were (1) cardiomyopathy characterized by a global decrease in left ventricular ejection fraction (LVEF); (2) signs or symptoms of congestive heart failure (CHF); (3) decline in LVEF of at least 5 % to less than 55 % with signs or symptoms of CHF; or (4) decline in LVEF of at least 10 % to less than 55 % without signs or symptoms of CHF (Seidman et al. 2002). The definition of cardiac dysfunction in oncology trials has differed somewhat from trial to trial making cross trial comparisons somewhat difficult. For example a drop in LVEF by 10 points or more to less than 55 % was considered evidence of cardiac dysfunction in the NSABP B-31 trial but the HERA trial required an absolute drop of 10 points or more to an LVEF of less than 50 % (Tan-Chiu et al. 2005; Piccart-Gebhart et al. 2005). Research in cardiac dysfunction/heart failure in breast cancer survivors would be facilitated by agreement on standard terminology for assessment in clinical trials. Use of American Heart Association criteria for heart failure is appropriate: Stage A risk factors present but no structural damage; Stage B structural damage but no signs or symptoms; Stage C structural damage with current or prior symptoms; and Stage D structural damage with symptoms with any physical activity or at rest (Yancy et al. 2013) (see Table 14.2). Asymptomatic reduction in left ventricular ejection fraction (LVEF) by 10 or more points to <50 % is generally viewed as representative of the Stage A to Stage B transition and is associated with a significant increase in cardiac mortality (Lenihan et al. 2013). Consequently, an asymptomatic reduction in left ventricular ejection fraction by 10 or more points to <50 % has come to represent a surrogate endpoint for cardiac injury in clinical trials and in the remainder of this manuscript will be referred to as subclinical cardiac toxicity and/or cardiac dysfunction.

Although many drugs can result in cardiac damage (Yeh et al. 2004; Bovelli et al. 2010), we will focus here on the two most commonly used types of agents with the greatest potential for cardiac dysfunction, namely anthracyclines and HER-2 targeted agents.

Although, oncologists are increasingly aware of the potential adverse effects of excessive weight on breast cancer outcomes and mortality, there has not been the same emphasis on physical activity and cardiorespiratory fitness.

Physical activity is associated with lower cardiovascular, all cause and breast cancer mortality (George et al. 2011; Irwin et al. 2011; Dhaliwal et al. 2013). In the Women’s Health Initiative 9 MET hours or about 3 h of fast walking per week pre or post diagnosis of breast cancer was enough to see an ~40 % decrease in breast cancer and all-cause mortality compared with sedentary women (Irwin et al. 2011). The important message then for women is to become active even if they were not prior to their diagnosis of breast cancer.

Cardiorespiratory fitness as measured by oxygen consumption at peak exercise (VO2_max or VO2_peak) is correlated with cardiovascular and all-cause mortality (Peel et al. 2009). In a general population increasing fitness by only one metabolic equivalent (1 MET corresponds to 3.5 mL/min/kg of oxygen consumption) is estimated to reduce risk of death by 13 % (Kodama et al. 2009). Individuals with low CRF (<7.9 METs) had a 40 % increase in all-cause mortality and 47 % increase in risk for cardiovascular events compared with those with intermediate CRF (7.9–10.8 METs). Women with low CRF had a 70 % increase in all-cause mortality and 56 % increase in cardiovascular events compared to those with high CRF (≥10.9 METs) (Kodama et al. 2009). Further, fitness may be a more important predictor of mortality than Body Mass Index (BMI) until one reaches the extremes of obesity where serious metabolic abnormalities are prevalent. In a recent meta-analysis, fit overweight and obese women had similar mortality as fit normal weight women, and were generally better off than normal weight unfit women unless the obese fit woman also had a chronic disease (Barry et al. 2014). 150 min of moderate physical activity per week or that sufficient to expend at least 1,000 Kcal per week is likely to allow most women to attain at least the lower bound of the intermediate CRF category (Lee and Skerrett 2001).

Women with breast cancer appear to have lower baseline CR fitness than age-matched individuals without cancer and fitness declines during treatment (Jones et al. 2012; Peel et al. 2014). Low fitness was reported in approximately half of women who had undergone chemotherapy with anthracyclines + taxanes with or without trastuzumab ~2 years after chemotherapy completion despite a normal LVEF in all but 8 % (Jones et al. 2007). These rates were dramatically higher than age-matched controls who had not received chemotherapy (Jones et al. 2007).

Research efforts need to focus on effective interventions which will prevent reduction of physical activity and fitness during and after treatment. Courneya et al. using a supervised aerobic and strength training exercise program have demonstrated that higher volume exercise can be safely delivered during adjuvant treatment, which in turn appears correlated with improved disease free and overall survival (Courneya et al. 2013, 2014). Higher intensity exercise may also favorably modulate pro-inflammatory biomarkers associated with cardiovascular disease risk (Fairey et al. 2005).

Methods to safely increase physical activity and fitness in a more practical, home-based environment following initial in-person training need to be a research priority, particularly for older women and those with a low level of fitness at baseline. Preliminary pilot studies suggest this is possible (Burnett et al. 2013). Further, simplified methods of assessing fitness which can easily be employed in an oncologist’s office such as the 6 min step test need to be validated against more complex tools such as VO2_max or VO2_peak (Hamilton and Haennel 2000; Simonsick et al. 2006).

Prediction models in older women at highest risk for cardiac toxicity have been suggested using variables of age, type of adjuvant therapy, and prior history of hypertension, diabetes, and symptomatic heart disease (Ezaz et al. 2014). Research is needed to develop more sensitive and specific cardiac dysfunction prediction models for women considering potentially cardiotoxic therapy which would include risk biomarkers and historical variables. These prediction models would ideally have the ability to include some of the new very sensitive indicators of left ventricular function (Fallah-Rad et al. 2011), in addition to weighted factor scores for age, BMI, a measure of physical activity or cardiorespiratory fitness, hypertension, diabetes, dyslipidemia, and any prior cardiac event. Ideally the models could also be adapted to include provision for genetic polymorphisms predisposing to cardiac risk such as the common HER-2 Ile655Val allele in the HER-2 gene (Beauclair et al. 2007), and polymorphisms in NADPH oxidase, MDR 1, MDR 2, catalase, superoxide dismutase, NADPH:quinone oxidoreductase, carbonyl reductase 3, and glutathione s-transferase (Wojnowski et al. 2005; Deng and Wojnowski 2007).

Cardiac MRI is a sensitive method for detecting left ventricular remodeling and early subclinical cardiac toxicity and is considered by some to be the gold standard (Fallah-Rad et al. 2011). It is currently being used in the MANTICORE trial to measure left ventricular end diastolic volumes (see below). However, cardiac MRI is expensive and thus not optimal for serial monitoring in the clinical community setting (Bellenger et al. 2000). Longitudinal or global left ventricular strain, which can be calculated from an echocardiogram with the appropriate software, is a measure of cardiac muscle deformability. The greater the negative value the more powerful the contraction. Normal strain values vary with age, sex and the software system used from approximately −15 to −22 (Yingchoncharoen et al. 2013; Cheng et al. 2013), but a value of −19 or more negative is highly unlikely to be associated with cardiac dysfunction or clinical congestive heart failure over the ensuing 3 months of cardiotoxic treatment (Sawaya et al. 2011, 2012). Adding a serum troponin drawn before and immediately after chemotherapy to left ventricular strain is reported to improve both sensitivity and negative predictive value for left ventricular dysfunction after anthracyclines (Fallah-Rad et al. 2011; Cardinale et al. 2010). Although some studies have suggested additional serum markers such as and high sensitivity c-reactive protein (hsCRP) may aid in predicting cardiac dysfunction, others have not found that the addition of serum biomarkers in addition to troponin to substantially increase sensitivity (Onitilo et al. 2012), in 78 patients receiving both doxorubicin and trastuzumab, found the addition of hsCRP and BNP did not add additional predictive information to ultrasensitive troponin and myeloperoxidase in predicting dysfunction by CREC criteria (Ky et al. 2014). 3D echocardiography may be more sensitive than 2D to declines in left ventricular function (Khouri et al. 2014).

Studies incorporating one or more of these markers into the monitoring process for cardiac toxicity (i.e., PREDICT trial) are currently ongoing. Left ventricular strain has perhaps the greatest potential in that the software is relatively easy to add to an echocardiogram machine and is a non-invasive procedure. Unfortunately, there are variations in the software and interpretations of results such that at present there is not a clear indication as to what constitutes a definitely abnormal strain or the amount of change which should be viewed with alarm.

Trials comparing the relative efficacy of newer sensitive biomarkers of subclinical injury prior to significant decline in 2D echo left ventricular ejection fraction, including high sensitivity troponin and other serum biomarkers, left ventricular strain, and cardiac MRI are ongoing and results are awaited with interest. Translational trials incorporating prophylactic treatment with beta blockers and angiotensin converting enzyme inhibitors at first signs of subclinical dysfunction are warranted.

Beta blockers, angiotensin converting enzyme inhibitors (ACEI), and angiotensin receptor blockers (ARBs) are commonly used to treat heart failure regardless of the inciting event. They are recommended in the general cardiovascular setting for asymptomatic women with a drop in left ventricular ejection fraction below 40 % or women with symptomatic congestive heart failure regardless of the amount of decline in LVEF (Yancy et al. 2013). Patients who develop symptomatic cardiac dysfunction while on treatment with a cardiotoxic agent should be treated as would anyone not on chemotherapy (Yeh et al. 2004). For women with early breast cancer, treatment with the cardiotoxic agent should be discontinued and non-cardiotoxic therapy substituted where possible. For women with metastatic HER-2 amplified breast cancer, temporarily discontinuing HER-2 targeted agents, treatment with blockers and ACEIs and then cautious re-introduction of the HER-2 targeted agent is often successful with continued indefinite use of the beta blocker and ACE inhibitor (Vaz-Luis et al. 2014).

What about asymptomatic women with subclinical cardiac toxicity whose LVEF is below 50 % but above 40 %? From the general population experience, it is clear that individuals with an asymptomatic ejection fraction below 50 % have a ~3.5-fold increase in all- cause mortality over the ensuing 9–10 years (Yeboah et al. 2012a). Use of ACEI and beta blockers in women with LVEF <50 %, and appropriate surgical treatment of heart valve problems resulting from radiation or medical or surgical treatment of coronary artery blockage is suggested (Lenihan et al. 2013; Lenihan 2012).

An area gaining momentum is the use of prophylactic drugs to reduce the incidence of asymptomatic left ventricular dysfunction (Swain and Vici 2004; van Dalen et al. 2008), particularly if early warning biomarkers can be used to trigger the use of the agent. Prophylactic cardioprotective therapy is not a new concept. A decade ago Swain et al. published a study suggesting doxorubicin induced cardiac toxicity can be reduced ~70 % by also giving the iron chelating agent dexrazoxane (Smith et al. 2010; Swain and Vici 2004). The current recommendation is that dexrazoxane can be given prophylactically once the cumulative doxorubicin dose exceeds 300 mg/m² (Yeh et al. 2004). Dexrazoxane is not widely used in clinical practice due to concerns that it may reduce effectiveness of doxorubicin as it binds to both TOPO 2A and TOPO 2B (Sawyer 2013).

Since individuals with a low LVEF, even if asymptomatic, have an increased risk of cardiac events and mortality, prophylactic cardio-preventive treatment trials are ongoing in the adjuvant setting of HER-2 positive breast cancer. Several small trials have shown cardioprotective effects of ACEI or beta blockers, especially when given with anthracyclines (Kalay et al. 2006; Bosch et al. 2011). Adjuvant trials with endpoints of symptomatic heart failure or a 2D echo measured LVEF of <50 % would require a very large number of participants. An alternative approach is to use a more sensitive indicator of early cardiac dysfunction. In the MANTICORE trial (Fig. 14.1) women taking trastuzumab are randomized to prophylactic cardiac protective therapy with beta blockers and/or ACEI but the primary outcome is change in left ventricular end diastolic volume as measured by cardiac MRI (Pituskin et al. 2011). With this endpoint as opposed to 2D echo LVEF, the trial coordinators are predicting that 158 randomized subjects will be adequate. Another approach to assess the benefit of prophylactic therapy is to use the traditional endpoint of LVEF drop by 10 points to <50 % but with a much higher risk group as the cohort, such as women with HER-2 amplified metastatic disease. Since these women have generally been previously treated with cardiotoxic drugs and since survival is dependent on being able to continue to receive HER-2 targeted therapy after first progression, the study question is of practical importance (Extra et al. 2010). Such a trial has been proposed in the Southwest Oncology Group Survivorship Committee in women with first and second line HER-2 amplified metastatic disease on HER targeted therapy with LVEF >50 % at baseline to determine whether prophylactic beta blocker therapy with Carvedilol will reduce cardiac dysfunction (See Fig. 14.2). Other target populations for prophylactic cardio-protective therapy where the event rate is likely to be high in either the metastatic or adjuvant setting are women 70 or over, high coronary artery calcium, and hypertension, or a high score on a CVD risk prediction model (Yeboah et al. 2012b).