There continues to be a paucity of data characterizing elderly breast cancer patients. With the increasingly older world population, oncologists are faced with an imposing challenge due to the growing cancer burden and the specific health care needs of older cancer patients [10
]. With the largest cohort of HER2-positive elderly breast cancer patients to date, the registHER study allows the unique and important opportunity to examine the natural history of disease and treatment patterns in these patients. Elderly patients had higher rates of underlying CVD and were less likely to be treated with cytotoxic therapies compared with their younger counterparts. While there was an increased incidence of CVD events in the elderly during follow-up, there was evidence of an association with comorbidities, including hypertension with complications and any CVD. Among treated patients with disease progression, the rate of CNS metastasis decreased with increasing age. This observation of a decreased incidence of CNS metastases in patients with HER2-positive MBC with increasing age has not been appreciated previously. Regardless of age, among patients receiving first-line treatment with trastuzumab, PFS was higher compared with those not treated with trastuzumab. OS was higher for patients <65 years and similar for patients ≥65 years of age.
Elderly patients (≥75 years) with HER2-positive MBC in registHER had higher rates of underlying CVD than their younger counterparts. The increased incidence of CVD events in elderly patients during 27 months of follow-up was clinically modest (the incidences of LVD and CHF were each below 5 %). However, a subanalysis noting a possible association with underlying comorbidities suggested a possible subset of elderly patients (those hypertension with complications and any underlying CVD) at an increased risk of CVD events, although these results are based on small numbers. Our findings compare favorably with a recent population-based study assessing the risk of cardiotoxicities in association with trastuzumab with/without anthracyclines in 47,806 women aged ≥65 years with breast cancer diagnosed between 1998 and 2005 [21
]. The cumulative incidence of CHF at Year 1 was 5.5 % for patients receiving anthracycline and trastuzumab, and 7.8 % for patients receiving trastuzumab without anthracycline. Consistent with our findings, risk factors for trastuzumab cardiotoxicity in another study included age >50, hypertension, and baseline cardiac dysfunction [22
]. In a recent retrospective study of 45 patients, trastuzumab-treated elderly breast cancer patients (aged ≥ 70 years) with a history of cardiac disease and/or diabetes had an increased incidence of cardiotoxicity [23
]: the overall incidence of cardiac events was 26.7 % (n
= 12 patients), in which 8 patients (17.8 %) developed asymptomatic left ventricular ejection fraction decline and 4 patients (8.9 %) developed symptomatic CHF. This incidence is higher than that reported in the recently published Cochrane meta-analysis of trastuzumab clinical trials [24
], although this report encompassed patients of all the age groups.
It is important to note that our follow-up period of 27 months likely captured all toxicity events in the registHER cohort. In a real-world, multicenter study of 499 consecutive patients with HER2-positive early breast cancer, trastuzumab cardiotoxicity most often occurred in the first 3 months of therapy (41 % of cases), with the greatest prevalence in older patients with higher creatinine levels and in patients pretreated with doxorubicin and radiotherapy [25
]. These results highlight the importance of obtaining a full medical history from patients before initiating anti-HER2 therapies to identify potential risk factors for cardiac dysfunction.
In our study, regardless of age, PFS was significantly higher in patients receiving first-line trastuzumab therapy compared with patients not treated with trastuzumab; for OS, significant improvement was observed for patients <65 years. A nonsignificant improvement was observed for trastuzumab-treated patients ≥65 years, which may be due to the small number of events in this age group. It is also possible that competing mortality from other causes in the ≥65-year-old age group diluted the mortality benefit from trastuzumab. These findings support those from a recent study by Griffiths et al. [26
] who used the national SEER database to describe a large cohort (N
= 610) of older women (mean age 74 years) with HER2-positive MBC treated with first-line or delayed trastuzumab treatment. Their findings showed that OS in older women with HER2-positive MBC treated with trastuzumab in combination with chemotherapy were similar to outcomes reported for younger patients.
The observation of decreasing CNS metastasis with increasing age may be due to underlying pathophysiology or treatment effects (or both). Other evidence suggests that older patients are less likely to develop CNS metastases. In another study of the registHER population which included 1,012 patients who had confirmed HER2-positive tumors, Brufsky et al. [18
] found that those with CNS metastases were younger (≤65 years) and also more likely to have HR-negative disease [18
]. Similarly, in a separate study in 2,685 primary breast cancer patients, those at increased risk for CNS metastasis were younger and ER− or PR−; importantly, adjuvant systemic therapies in this study were not associated with an increase in CNS metastasis risk [27
]. Indeed, as trastuzumab and most chemotherapeutic agents do not readily cross an intact blood brain barrier [18
], and as elderly patients in registHER were less likely to receive cytotoxics, the observation of decreased CNS metastasis in the elderly is potentially due to a combination of both disease pathophysiology and treatment effects. These findings warrant further research.
An inherent limitation of this study is possible “confounding by indication,” due to the nonrandomized, observational nature of the registHER study. In observational studies, bias from confounding by indication—also referred to as “treatment selection bias”—may result because selection of treatments is not random and is determined by patient and physician characteristics; the observed effect can, therefore, be influenced by factors other than the treatment. There is also the potential for residual confounding by certain clinical factors that may not have been collected or sufficiently captured. Because patients may have had a diagnosis of MBC up to 9 months prior to enrollment and patients with longer survival may be more likely to enroll, OS estimates from time of MBC diagnosis in registHER may be slightly higher than expected in a general MBC patient population. Finally, limited information was collected for cause of death in registHER (options included only “cancer” and “other”) and cause of death was missing for >5 % of deaths, which precluded the calculation of breast cancer-specific mortality rates in this study. Our findings should also be interpreted with caution, due to the small number of events in elderly patients in our study.