This contemporary population-based study in 369 consecutive female patients receiving trastuzumab for early stage HER-2 positive BC between January 2005 and December 2015 showed that 106 (29%) patients experienced treatment interruption of ≥ 2 weeks, among whom 42 (11%) experienced permanent discontinuation of trastuzumab therapy, largely due to cardiac toxicity.
Evidence from randomized trials with adjuvant trastuzumab in combination with an anthracycline –taxane-based regimen have shown a 12–15% asymptomatic decline in LVEF with very few patients reporting permanent discontinuation (10, 11, 28–30). For instance, secondary analyses of cardiac events in NSABP B31 showed a meager 4% discontinuation of adjuvant trastuzumab (11). Careful selection of patients with limited to no medical comorbidities in clinical trials is imperative given the concern for long-term toxicity in a curative intent patient population. However, clinicians routinely have to make complex medical decisions weighing the potential benefit of HER2-directed therapy against risks of pre-existing comorbidities in everyday practice. In addition, minority and elderly patents are often underrepresented in therapeutic studies. Our study highlights this ‘real world’ disparity with a significantly higher proportion of patients with treatment interruption or permanent cessation of trastuzumab therapy for early HER2-positive BC, when only half received an anthracycline-based regimen. These results are consistent with other large retrospective series showing higher overall discontinuation of adjuvant trastuzumab in clinical practice, with Wang et al reporting up to 41% of patients discontinuing planned therapy, most events as a result of TIC (31, 32). However, a majority of patients reported by Wang et al were elderly (mean age 71.6 years), emphasizing the need for further investigation of risk factors and risk mitigation strategies in this patient population (33).
Approximately 18% of all patients in the study experienced TIC leading to treatment delays. Consistent with previous studies, TIC was responsible for the majority of treatment interruption (62%) with adjuvant trastuzumab (31, 34, 35). Shorter (< 6 months) planned duration of adjuvant trastuzumab has been studied in an attempt to reduce risk of CV events while maintaining efficacy of targeted therapy (23, 24, 36). While six months of adjuvant targeted therapy is associated with favorable CV risk, two major phase III randomized adjuvant trials reported conflicting results on non-inferiority in invasive DFS with shorter duration to standard one year of therapy (23, 24). Secondary analyses of US PHARE study revealed that metastases-free survival, a key endpoint, was significantly worse with six months vs. one year of therapy. However, long term outcomes, especially in patients experiencing treatment interruptions or discontinuation, are lacking from large prospective studies given the small number of events. Our analyses suggest any interruption in trastuzumab treatment as a prognostic factor with up to three times worse DFS and OS compared to patients who finished planned one year of therapy or had < 14 days of interruption in treatment. Multivariable analyses confirmed this effect after adjusting for major confounders, including drug-induced cardiotoxicity, indicating that worse outcomes are likely related to cancer recurrence in these patients. Similar concerns have been reported with early trastuzumab interruption in other studies as well (31–33). Gong et al reported higher risk of disease recurrence in patients stopping adjuvant trastuzumab and emphasized that early discontinuation is a stronger prognostic factor for worse survival than cardiotoxicity in this patient population. As in our study, they also reported a high proportion of patients completing therapy with over 85% patients receiving ≥ 16 doses (32). These results taken together suggest a strong need for early assessment and optimization of CV status of at-risk patients to minimize any interruption in targeted therapy.
Cardioprotective therapy, including BB and ACEi, have a well-documented safety profile and are commonly used in practice to improve cardiac outcomes in patients with preexisting heart disease. Current evidence suggests that they may be crucial in preventing treatment interruption of adjuvant trastuzumab in HER2-positive BC patients (37, 38). In a randomized, double-blinded placebo-controlled trial, Guglin et al showed that cardiotoxicity-free survival was higher for both lisinopril (HR: 0.53;p = 0.015), and carvedilol (HR: 0.49; p = 0.009) compared to placebo with fewer treatment interruptions in the intervention group (37). While we report worse OS in patients on cardioprotective therapy, this likely results from a higher proportion of patients receiving BB and/or ACEi therapy for other concurrent illnesses (72%) that may adversely impact survival.
The use of BB and ACEi prior to initiating anti-HER2 therapy needs to be carefully weighed against unintended consequences, such as rising health care costs, over-diagnosis or over-treatment associated with the increased use of cardiac surveillance and prophylactic therapy (39). Close collaboration between oncologists and cardiologists is needed to guide personalized therapy for vulnerable patients (40–43). Our institutional guidelines encourage referral to cardio-oncology for at–risk patients prior to initiating trastuzumab therapy and early intervention in patients who experience LVEF decline while on targeted therapy. A majority of patients who experienced treatment interruption (73%) in this study were referred and received timely evaluation by cardio-oncology.
Another strategy to reduce cardiac risk has been to employ non-anthracycline regimens in (neo) adjuvant therapy (44, 45). Our analyses did not show a significant difference in the use of anthracycline in patients who experienced any treatment interruption compared to controls. This is likely due to careful selection of patients deemed fit to receive adjuvant anthracycline. The advent of adjuvant TDM1 for patients with high risk residual disease and dual anti-HER2 therapies, such as combination with pertuzumab, will further impact the use of adjuvant anthracycline in these patients (46, 47).
Our study is not without limitations. While we have a detailed account of treatment data, reason for trastuzumab interruption/discontinuation, known pre-existing cardiac comorbidities, rechallenge rates and use of cardioprotective therapy, the study is limited by its retrospective nature, including possible selection bias that may impact outcome assessment. Dose titration was at the discretion of the treating clinicians. Additionally, it is a single-institution study with a limited number of events. Given the retrospective nature of chart review, it was challenging to identify time of initiation (before or after onset of TIC) and treatment indication (TIC vs. concurrent cardiac illness) for the use of BB and/or ACEi therapy in patients. This made it difficult to interpret the association of the use of cardioprotective therapy on the incidence of TIC and treatment interruption of trastuzumab. Lastly, our observational study sought to evaluate the association of treatment interruption in adjuvant trastuzumab with long-term cancer outcomes so as to augment patient care and survivorship, but cannot confirm causality.