Abstract
Background The association between cardiotoxic chemotherapy for breast cancer treatment and congestive heart failure has been well documented in clinical trials. However, clinical trial populations often contain young, otherwise healthy people, and these results may not represent true associations in community settings. We used electronic administrative data from the CRN VDW to estimate the association between chemotherapy with anthracycline and/or trastuzumab and heart failure in a cohort of women with breast cancer.
Methods We included 13,497 women diagnosed with invasive breast cancer between 1999–2007 from eight CRN health plans. Using Current Procedural Terminology codes and National Drug Codes from the VDW, we determined whether each woman was exposed to anthracyclines alone, trastuzumab alone, anthracyclines plus trastuzumab, other chemotherapy, or no chemotherapy. We used International Classification of Disease codes and an algorithm based on number of inpatient, outpatient, and emergency department diagnoses to determine whether each woman had a prevalent (before breast cancer diagnosis) or incident (after breast cancer diagnosis) heart failure diagnosis. Using Cox proportional hazards models, we modeled the time to incident heart failure (censoring at disenrollment or death) associated with time varying exposures for each treatment category using hazard ratios (HR) and 95% confidence intervals (CI) adjusted for age, health plan, stage, year of diagnosis, and comorbidities.
Results Almost half of our study population received no chemotherapy (46.2%, 6420/13497); 28.6% (3857/13497) received anthracyclines alone, 1.2% (157/13497) received trastuzumab alone, 3.5% (470/13497) received both, and 20.4% (2748/13497) received other chemotherapy. One thousand forty-one women (7.7%, 1041/13497) were diagnosed with incident heart failure. The adjusted HR associated with anthracycline use alone was 1.19 (95%CI=0.97–1.46) compared with no chemotherapy, 3.02 for trastuzumab alone (95%CI=1.86–4.89), 6.05 for anthracycline and trastuzumab combined (95%CI=4.35–8.42), and 1.36 for other chemotherapy (95%CI=1.17–1.58). These associations strengthened after excluding women with prevalent heart failure or incident heart failure diagnosed <60 days after breast cancer diagnosis.
Conclusions Cardiotoxic outcomes after chemotherapy occur in the community setting; the combination of anthracycline plus trastuzumab in particular, is associated with an elevated heart failure risk. The CRN is an ideal setting to study chemotherapy, cardiotoxicity, and other survivorship-related outcomes.
