By Michael H. Crawford, MD, Editor

SYNOPSIS: Administering trastuzumab after a course of anthracycline therapy for breast cancer can result in cardiac toxicity. Serial echocardiograms in this study showed that a lower initial left ventricular ejection fraction before anthracycline therapy and the amount of decrease in ejection fraction after the anthracycline course are predictive of subsequent trastuzumab cardiac toxicity.

SOURCES: Goel S, et al. Decline in left ventricular ejection fraction following anthracyclines predicts trastuzumab cardiotoxicity. JACC Heart Fail 2019;7:795-804.

Ewer MS, Ewer SM. Trastuzumab cardiotoxicity after anthracycline exposure constitutes a complex and clinically important entity. JACC Heart Fail 2019;7:805-807.

Trastuzumab therapy is an important chemotherapeutic agent for certain types of breast cancer and can cause dose-dependent, reversible cardiotoxicity (TRC). Current recommendations suggest monitoring cardiac function every three months on this therapy; however, whether findings on cardiac testing predict the development of TRC is unclear.

Goel et al designed this study to determine if baseline left ventricular ejection fraction (LVEF) or changes in LVEF and serum biomarkers or germline genetic polymorphisms can predict the subsequent development of TRC in patients receiving anthracycline-based chemotherapy followed by trastuzumab for breast cancer. This was a multicenter, prospective, observational study from 17 Australian centers. Exclusion criteria were baseline EF < 50%, pregnancy, and prior chemotherapy. LVEF was determined by echo or radionuclide blood pool imaging (MUGA) at baseline, after the completion of anthracycline therapy and every three months during trastuzumab therapy. TRC was defined as the occurrence of any of the following: cardiovascular death; cardiac arrhythmias; ischemia or infarction; New York Heart Association class III or IV heart failure; an asymptomatic decline in EF by > 15% or > 10%, with an absolute value < 50%. Also, plasma troponin and NT-proBNP were measured at the same visits. Baseline DNA was obtained from blood and sequenced. Of 222 patients who met inclusion criteria, five were excluded due to missing data, leaving 217 patients.

TRC developed in 18 patients, most because of a drop in LVEF to < 50%. Baseline characteristics between those who did and did not develop TRC were quite similar. Most TRC events occurred within the first three months after trastuzumab was started. A multivariate analysis showed that TRC was associated with lower baseline EF pre-anthracycline therapy and a greater decline in EF from pre- to post-anthracycline therapy (odds ratio [OR], 3.9 and 7.9, respectively; both P = 0.0001). Troponin, NT-proBNP, and genetic polymorphisms were not associated with TRC. The authors concluded that low baseline EF and greater declines in EF on anthracycline therapy were independent predictors of TRC on subsequent trastuzumab therapy.


Trastuzumab is a monoclonal antibody and was not expected to cause cardiac toxicity. In monotherapy trials, the incidence of cardiac toxicity was 0.4%. However, when combined with anthracycline therapy, more TRC was noted. With concomitant use, it was 16%, but the histology of the myocardium was not what is observed with anthracycline.

Anthracycline can cause myocyte destruction and replacement with fibrosis, leading to permanent damage and life-threatening reductions in LV performance. Trastuzumab is much less malignant, leading some to classify anthracycline toxicity as type 1 and trastuzumab as type 2. Thus, it makes biologic sense that if a patient is recovering from anthracycline toxicity that trastuzumab might potentiate the myocardial injury. Indeed, studies of combination anthracycline plus trastuzumab chemotherapy have shown that the longer the time between anthracycline exposure and trastuzumab administration, the less the cardiotoxicity (3% at 21 days, 0.6% at 89 days). This also fits with the Goel et al study, where most trastuzumab toxicity was observed in the first three months of therapy (in this study, trastuzumab therapy directly followed anthracycline therapy).

One goal of the Goel et al study was to identify low-risk patients who would not need frequent cardiac monitoring when on trastuzumab therapy after a course of anthracycline therapy. Contrary to other smaller studies, Goel et al found no independent predictive value of biomarkers or genetic polymorphisms. They found that LVEF before anthracycline therapy and the change from pre- to post-anthracycline therapy were strongly predictive. Post-anthracycline LVEF was not tested because it certainly would be predictive. Goel et al arrived at a formula for identifying low-risk patients. The formula carried a receiver operating characteristic value of 0.87: [(3× baseline EF) - 4.3] × (EF difference from baseline to post-anthracycline therapy). If this value is > 201%, the risk of TRC is 1.2%, and these patients would be considered low risk. The value of baseline EF has been shown in other studies, but the value of the change in EF is novel and was the most predictive (OR, 7.9).

There were limitations to the study. TRC events were uncommon (8%). Echo and MUGA were used for EF calculations (53% MUGA). This should not have affected the calculation of differences since the two techniques were not used in the same patient, but could have affected the baseline EF calculations. Also, it seems axiomatic that EF values would predict EF, and 14 of 18 patients who developed TRC met only the decrease in EF to < 50% criteria.

Further, LV strain was not evaluated, which some believe is superior to EF. However, the study is of practical value in managing patients treated with both drugs because it puts the EF values we are collecting into clinical perspective.