Trastuzumab has improved the clinical course of HER2-positive breast cancer in the metastatic, adjuvant, and neoadjuvant settings.^1–6 The joint analysis of the National Surgical Adjuvant Breast and Bowel (NSABP) B-31 and North Central Cancer Treatment Group (NCCTG) N9831 trials demonstrated that the addition of 1 year of trastuzumab to sequential anthracycline-taxane chemotherapy reduced the risk of recurrence by approximately half (hazard ratio [HR], 0.48; P < .0001) and the risk of death by one third (HR, 0.67; P = .015).⁵ A similar benefit was seen in the Herceptin Adjuvant (HERA) trial, for patients randomly assigned to 1 year of trastuzumab (or not) given only after the completion of adjuvant chemotherapy.⁴ In all three of these initial reports, and as predicted by experience in the earlier metastatic disease trials, trastuzumab’s benefits are associated with the potential for increased cardiotoxicity.¹ With broad adoption of adjuvant trastuzumab, it is appropriate to ask several related questions on this issue. How frequent and how severe is the toxicity and what modulates the risk?

Because of the cardiac safety signal in the earlier metastatic disease trials,⁷ the risk of serious and life-threatening cardiotoxicity with the combination of anthracyclines and trastuzumab was a recognized concern when these phase III adjuvant studies of trastuzumab commenced. To mitigate this risk, strict cardiac criteria that excluded patients predicted to have greater likelihood of cardiac toxicity were used. In addition, rigorous, if arbitrary, cardiac monitoring was incorporated to allow early detection of any signals of increased cardiac risk. Not knowing the magnitude of the benefit for trastuzumab, early stopping rules allowed an increase in congestive heart failure (CHF) and cardiac deaths of up to 4% in some studies, a limit that was nearly breached.^8,9

Collectively, the available studies demonstrated that the risk of serious cardiotoxicity with adjuvant trastuzumab was lower than the preset threshold for early stopping and that trastuzumab-related cardiotoxicity, in contrast to anthracycline-induced cardiotoxicity, may occur as an early and reversible event. Perhaps because of increased awareness and scrutiny, the most frequent events were asymptomatic declines in left ventricular ejection fraction (LVEF) for which the long-term clinical significance is largely unknown.

In this issue of the Journal of Clinical Oncology, updated cardiac safety data are presented from patients treated with anthracyclines and trastuzumab in three of the previously reported adjuvant trastuzumab trials. First, Russell et al¹⁰ retrospectively reviewed patients with CHF in the NSABP B-31 and NCCTG N9831 studies. Patients with possible cardiac events were identified from the trial database and reviewed by a committee of six oncologists and cardiologists. These reviewers were independent of the original study and were blinded to trastuzumab treatment and the investigators’ original cardiac diagnosis. CHF was rigorously defined by symptoms, physical signs, and objective findings and included an LVEF drop of 10% or a drop to an absolute LVEF of < 50% by locally obtained multigated radionuclide angiography or echocardiogram. This independent review confirmed earlier risk estimates by showing that the addition of trastuzumab to anthracycline-based chemotherapy increased the incidence of CHF nearly four-fold, from 0.45% to 2.0%, at a median follow-up of 1.9 and 2.0 years, respectively. Among those assigned to trastuzumab, 133 (7.4%) patients with possible cardiac events were reviewed, yielding 36 that were deemed to have definitely suffered a cardiac event. Again, consistent with earlier reports,¹¹ Russell et al note that a majority of these patients recovered completely (20 [55.5%]) or partially (11 [30.6%]). In addition, the independent review highlighted three cardiac deaths (< 0.2%) among 1,799 patients randomly assigned to trastuzumab. Two of these occurred before the initiation of trastuzumab. In an attempt to develop a risk stratification model for these adverse events, Russell et al identified age older than 50 years (2.84-fold increase v age < 50 years) and low LVEF post anthracycline treatment (P = .0106) as possible risk factors for cardiac events. Patients with baseline hypertension had a 1.89-fold increased risk of a cardiac event, although this result was not statistically significant.

In another article in this issue of JCO, Procter et al¹² report a similar analysis from the HERA study. As expected, the most common cardiac event noted was a “significant” LVEF drop, defined as an absolute decline of at least 10 percentage points from baseline to below 50%. This occurred in 164 (9.8%) trastuzumab-treated patients compared with 49 (2.9%) controls. The clinical significance of these LVEF declines is unclear because they may have been spurious, or at least quickly reversible, since fewer than half these patients overall (73 or 4.3% overall) had a confirmed cardiac event (CHF or second confirmed LVEF decline). Indeed, these observed changes in LVEF rarely translated into symptomatic CHF, which occurred in 32 (1.9%) trastuzumab-treated patients versus two (0.1%) controls (difference, 1.8%; 95% CI, 1.1% to 2.5%).

In both of these articles, the increased rate of clinical CHF (1.9% and 2%) at a median follow-up of 3.6 and 2 years should be seen in the context of the absolute improvements in disease-free survival at 3 years (6.3% and 11.7%, respectively) attributable to the addition of adjuvant trastuzumab to anthracycline-based chemotherapy.^5,13 Both articles suggest that these events are usually reversible. For example, in the article by Procter et al, 57 (78%) of 73 were deemed to have a favorable outcome. Reversibility is also suggested by the finding that they attributed no cardiac deaths to trastuzumab.¹³

Each of these two independent reviews of large prospective clinical trials offer evidence that long-term cardiotoxicity with trastuzumab following anthracycline-based chemotherapy for HER2-positive early breast cancer is relatively infrequent and generally associated with a favorable outcome. If this is reassuring, there remain several cautions and considerations. In these two analyses, cases were selected for review on the basis of previously documented cardiotoxicity, and the independent committees did not review original data from all enrolled patients. Although unlikely, it remains possible that other cardiac events may not have been reported. In addition, the independent review committees assessed reversibility on the basis of the weight of available evidence, but a prospective assessment of reversibility would have been preferable, as would incorporation of a standardized management approach for suspected CHF and possibly a quality-of-life measure to determine the true clinical significance of these events. Although the relatively constant rate of cardiac events in NSABP B-31 in years 1 to 3 suggests that longer-term follow-up is unlikely to change the results of any of these studies,¹⁴ relatively small numbers of patients have reached 4-year follow-up (> 500 in HERA). Since anthracycline-mediated cardiotoxicity may occur many years after treatment, we could still see unanticipated long-term effects in the future.

The articles by Russell et al¹⁰ and Procter et al¹² add to the evidence that patients with HER2-positive early breast cancer and normal baseline LVEF can be safely treated with sequential anthracycline-taxane–based chemotherapy incorporating trastuzumab. The question many will ask is whether we should avoid the entire issue by substituting an evidence-based alternative such as docetaxel, carboplatin, and trastuzumab (TCH) from Breast Cancer International Research Group (BCIRG) 006, or even an as yet untested approach like docetaxel plus cyclophosphamide with trastuzumab (or other drugs). The former can be discussed with available phase III data; the latter has not yet been reported. To date, the significant evidence as reviewed above supports the routine use of sequential anthracyclines and trastuzumab with an overall low risk of cardiotoxicity demonstrated in more than 4,500 patients and confirmed on independent cardiac reviews of almost 3,500 of these (Table 1) . TCH is an alternative therapy, albeit one that has been tested only in a single trial where it did not appear to have inferior efficacy to a relatively standard anthracycline/taxane/trastuzumab regimen. For TCH, there are not yet long-term cardiac safety data for the approximately 1,000 treated patients.

An alternative way of framing the issue is to ask whether there is adequate evidence or reason to abandon anthracyclines followed by trastuzumab for HER2-positive breast cancer. Some in our community have already made this decision, but it is worth noting that the advantages of continuing to treat this way are that it offers a more consistently demonstrated overall survival advantage. The cardiac event rate for sequential anthracyclines and trastuzumab is higher than the same treatment without trastuzumab, but still low, and the events are largely reversible. The sequential use of generic anthracyclines and taxanes can possibly be less expensive than agents such as docetaxel; if the dose-dense approach is used, it can be safer and faster.¹⁶ In this regard, we note that in contrast to some preconceived expectations, patients with normal baseline LVEF can safely be treated with dose-dense anthracycline-based chemotherapy and trastuzumab with no evidence of an increased risk of short- or long-term cardiotoxicity to date.^16,17 Additionally, experience at the M. D. Anderson Cancer Center has demonstrated that trastuzumab can be combined with both anthracyclines and taxanes in the preoperative setting with no increase in cardiac toxicity.^18,19 This finding was confirmed in the randomized phase III Neoadjuvant Herceptin (NOAH) study, in which the addition of trastuzumab to preoperative chemotherapy doubled the complete response rate and, despite concurrent anthracycline-trastuzumab therapy, was associated with a similar low risk of cardiotoxicity.²⁰

Yet another way to approach the issue is to ask whether there are patients without overt cardiac dysfunction whom we should select for an available non-anthracycline containing adjuvant regimen with trastuzumab. The current reports add to the growing evidence that older patients with borderline LVEF function and baseline hypertension may be at increased risk of cardiotoxicity, but no study has prospectively compared treatment regimens in this subset, and the alternative (TCH) might or might not be as toxic or more toxic in other ways for this population.¹⁵ The risk-benefit calculations for these patients are challenging, and predictive biomarkers to improve risk stratification and to predict for trastuzumab-induced cardiotoxicity are needed. Unfortunately, putative cardiac biomarkers such as troponin I and C-reactive protein have been disappointing, with no validated serum marker for cardiotoxicity emerging to date.²¹

Although the overwhelming body of evidence from rigorously conducted randomized control studies supports anthracyclines as the mainstay of current adjuvant trastuzumab-based chemotherapy regimens, the potential omission of anthracyclines for appropriately selected patients has become a major research focus. Subgroup tissue analyses from prospective trials attempting to more precisely define the role of anthracyclines have had conflicting results; some studies suggest that only HER2-positive tumors benefit from anthracyclines,^22,23 and other data show that HER2-positive tumors do not benefit from the addition of anthracyclines.²⁴ Attempts to omit anthracyclines from trastuzumab-based chemotherapy have focused on possible biomarkers such as the topoisomerase II alpha (TOPO-II) gene, the putative target of the anthracyclines, which is closely related to the HER2 gene on chromosome 17.²⁵ Preliminary data from a subgroup analysis from BCIRG 006 suggested that anthracyclines may not be needed for trastuzumab-treated patients with coamplification of HER2 and TOPO-II. In contrast, recent data from the National Cancer Institute of Canada Clinical Trials Group Mammary.5 (MA5) trial, which randomly assigned patients to CMF or anthracycline-based chemotherapy, suggested that the anthracycline benefit was confined to patients with tumors with TOPO-II gene deletions as well as amplifications.²⁶ Most strikingly, in a multivariate analysis, when adjusted for TOPO-II status, HER2-positive tumors continued to derive a survival benefit from anthracyclines (HR, 0.44; P = .02).²⁶ The reasons for the extensive inconsistencies surrounding the putative predictive effect of TOPO-II are largely unclear. A possible explanation is that many studies have depended on fluorescent in situ hydridization (FISH) to determine TOPO-II status, and the width of commercial FISH probes may limit specificity. High-resolution techniques, such as comparative genomic hybridization or representational oligonucleotide microarray analysis (ROMA), which allow a more precise definition of these genes, have demonstrated that so-called amplification of HER2 and TOPO-II may be far more complex than previously thought.²⁷ In part, the expression of other genes in the same area on chromosome 17 may play an important role in the interaction of HER2 and TOPO-II as detected by FISH.²⁸ Finally, it is important to recall that TOPO-II is only one of several targets of anthracyclines in cancer. Given these uncertainties, TOPO-II status cannot be used to tailor therapy and, as recently noted in JCO, there are currently insufficient data to recommend replacing anthracyclines in general in adjuvant chemotherapy regimens.²⁹ The current studies extend this caution to the HER2-positive setting in which trastuzumab is used.

One more related issue is the definition of the optimum duration of trastuzumab therapy and the potential interaction of duration on cardiac events. In the smaller Finland Herceptin (FinHer) study, the magnitude of benefit for 9 weeks of trastuzumab was similar to that seen in larger studies incorporating 1 year of therapy (Table 1).³ In the combined N9831–B-31 study, 29.5% of patients discontinued trastuzumab before the planned 1-year duration for reasons other than recurrence, compared with only 10.2% in the HERA trial.^5,12 Despite this, there is neither a suggestion that the shorter delivered dose of trastuzumab was associated with an inferior outcome nor a change in cardiac risk. Relevant studies are ongoing, such as the French Protocol of Herceptin Adjuvant With Reduced Exposure (PHARE) study, in which patients are randomly assigned to either 6 months or 1 year of adjuvant trastuzumab, and the Synergism Or Long Duration (SOLD) study, in which patients are randomly assigned to 9 weeks or 1 year of trastuzumab in combination with sequential taxane-anthracycline chemotherapy.^30,31

The incorporation of trastuzumab into adjuvant chemotherapy regimens for HER2-positive early breast cancer has been a major clinical advance. Long-term data from HERA and N9831–B-31 suggest that the benefits are durable and that cardiotoxicity is uncommon and generally reversible. Confirmation of equivalent efficacy for a non-anthracycline–containing regimen will be important as will the development of biomarkers and predictive models for toxicity.