Risk of primary haematologic cancers following incident non-metastatic breast cancer: A Danish population-based cohort stud

1 Introduction

Breast cancer is the most commonly diagnosed malignancy in women [1] . Population-based mammography screening, and increasingly effective treatments have contributed to increased prevalence of long-term breast cancer survivors. Hence, breast cancer survivors represent over one-third of all female five-year cancer survivors worldwide [2] . These women are at risk of developing new primary cancers either due to shared etiologies or breast cancer-directed treatment 345678910 .

Studies investigating the risk of haematologic cancer after breast cancer document an increased incidence of myeloid and lymphoid leukaemia 11121314151617 , and non-Hodgkin’s lymphoma (NHL) [1112] . Yet, most of these studies do not include women undergoing current treatment recommendations [111314] . Guideline for breast cancer treatment has evolved over the past 30–40 years [18] . In Denmark, during the 1980 s, guideline for chemotherapy included cyclophosphamide, metotrexane and fluorouracil (CMF) administered in nine four-week cycles (600, 40 and 600 mg/m ² , respectively). Since 1990, CMF was gradually replaced by cyclophosphamide, epirubicin and 5-fluorouracil (CEF) administered as nine three-week cycles (600, 60 and 600 mg/m ² , respectively). In 2007, 5-fluorouracil was replaced by taxanes administered as three cycles of docetaxel (100/m ² ) subsequent to three cycles of epirubicin and cyclophosphamide (EC, 90 and 600 mg/m ² , respectively) administered at three-week intervals [19] . In 2008, this regime was supplemented with prophylactic granulocyte colony-stimulating factor (G-CSF) [20] . Though being used throughout follow-up, docetaxel has since 2013 been gradually replaced by nine weekly cycles of paclitaxel. Tamoxifen therapy was consolidated with aromatase inhibitors for postmenopausal women from 2004 [19] ; and the duration of tamoxifen therapy was extended to up to ten years after primary diagnosis. It is important to examine the potential influence of these changes on the risk of haematologic malignancies after breast cancer.

Recent studies including women reflecting contemporary treatments may have been prone to immortal time bias [21] as women who developed haematologic malignancies in the first six months after breast cancer diagnosis were excluded, yet follow-up began at the time of diagnosis [121517] . Furthermore, the studies included women participating in a health insurance programme or were based on the SEER cancer registries, which may not be representative of the entire population [22] . We therefore evaluated the risk of haematologic cancers after primary breast cancer diagnosis in a nationwide population-based study in Denmark, where all residents have access to free tax-funded healthcare, including cancer care [23] .