Chronic Oxidative Stress as a Marker of Long-term Radiation-Induced Cardiovascular Outcomes in Breast Cancer

Abstract

While biomarkers have been proposed to identify individuals at risk for radiation-induced cardiovascular disease (RICVD), little is known about long-term associations with cardiac events. We examined associations of biomarkers of oxidative stress (myeloperoxidase, growth differentiation factor-15, 8-hydroxy-2'-deoxyguanosine [8-OH-dG], placental growth factor), cardiac injury (troponin I, cystatin-C), inflammation (interleukin-6, C-reactive protein), and myocardial fibrosis (transforming growth factor-ß) with long-term RICVD in breast cancer (BC) survivors. We conducted a nested case-control study within the Women's Health Initiative of postmenopausal women with incident BC stages I-III, who received radiation and had pre- and post-BC diagnosis serum samples. Cases (n = 55) were defined as developing incident, physician-adjudicated myocardial infarction, coronary heart disease death, other CVD death, heart failure, or stroke after BC. Cases were matched to three controls (n = 158). After adjustment, a higher 8-OH-dG ratio was significantly associated with an elevated long-term risk of RICVD, suggesting oxidative DNA damage may be a putative pathway for RICVD.

Keywords: Biomarkers; Breast cancer; Cardiovascular disease; Inflammation; Oxidative stress; Radiation.

Fig 1.. Sample flow chart.

Participants were eligible if they were diagnosed with invasive breast cancer prior to 2005 and met the following criteria: 1) had a pre- and post-breast cancer diagnosis serum sample available approximately 3 years apart and 2) had documented receipt of radiation treatment either through medical record abstraction, Medicare claims data, or self-report. Cases are defined as participants who developed a major adverse cardiac event (MACE) or heart failure (HF) after breast cancer.

Fig 2.. Boxplots for the distribution of pre- and post-breast cancer biomarker concentrations by case status.

Lower and upper box boundaries depict 25th and 75th percentiles; line inside box represents the median; whiskers extend to maximum and minimum values; “x” represents the mean. *P<0.05, **P<0.01, ***P<0.001. Abbreviations: 8-OH-dG, 8-hydroxy-2’-deoxyguanosin; c-reactive protein, CRP; GDF-15, growth differentiation factor-15; IL-6, interleukin-6; MPO, myeloperoxidase; NS, not significant at alpha 0.05; PGF, placental growth factor; TGF-B, transforming growth factor-beta; TnI, troponin-I.

Fig 3.. Adjusted associations of pre-cancer biomarkers, post-cancer biomarkers, and biomarker change ratios with MACE or HF.

Pre- and post-cancer biomarkers are log transformed to base 2. Change is modeled as the log₂ ratio of post-cancer relative to pre-cancer concentration; each unit difference in the biomarker ratio corresponds to a doubling in value compared to pre-cancer. PGF and TNI are categorized as above vs. below (reference) detection; the change value is adjusted for pre-cancer biomarker. Models are adjusted for age (5-year categories), income (< $34,999, $35,000 - $74,999, >$75,000), waist circumference (cm), smoking (pack-years), physical activity (total MET-minutes/week), cancer stage (local vs. regional), cardiac medications (yes/no).