Non-cancer risks among female breast cancer survivors: a matched cohort study in Japan

Abstract

Background: The number of breast cancer (BC) survivors has increased worldwide, but the landscape of their non-cancer disease risks remains unclear, especially among Asian women.

Methods: In the JMDC claims database, which covers company employees and their family members in Japan, women aged 18-74 years with and without an incident BC were matched in a 1:4 ratio for age and entry timing to the database between January 2005 and December 2019. The risks for six cardiovascular diseases (myocardial infarction, heart failure, atrial fibrillation/flutter, ischaemic stroke, intracranial haemorrhage, and pulmonary embolism) and six non-cardiovascular diseases (major osteoporotic fractures, other fractures, gastrointestinal bleeding, urinary tract infection, infectious pneumonia, and anxiety/depression) were compared between the groups.

Findings: Comparing 24,017 BC survivors and 96,068 matched women (mean age, 50.5 years, standard deviation, 8.7 years), the incidence rates of heart failure, atrial fibrillation/flutter, and all non-cardiovascular diseases were higher in the BC survivor group. The highest adjusted hazard ratio (HR) was noted for heart failure (3.99 [95% confidence interval 2.58-6.16]), followed by gastrointestinal bleeding (3.55 [3.10-4.06]), and anxiety/depression (3.06 [2.86-3.28]). The HRs in the first year were larger than those for 1-10 years for most outcomes, whereas the HRs for fracture outcomes were larger for 1-10 years.

Interpretation: BC survivors in Japan showed an increased risk of many non-cancer diseases compared to women without BC. Most risks increased more steeply during the first year following diagnosis, whereas the risk of fractures increased later.

Funding: Competitive research funding from Pfizer Health Research Foundation in Japan.

Keywords: Bleeding; Breast cancer; Cardiovascular disease; Depression; Fracture; Infection.

Fig. 1

Flow chart of matched cohort of women with and without breast cancer.

Fig. 2

Kaplan–Meier curves of the cumulative incidence among women with and without breast cancer for each outcome. BC = breast cancer.

Fig. 3

Adjusted hazard ratios between women with and without breast cancer for each outcome during <1 year and 1–10 years from the index month. CI = confidence interval, BMI = body mass index. Note: The adjusted hazard ratios (95% CIs) are from the last model (model 4) adjusted for hypertension, diabetes, dyslipidaemia, osteoporosis (only for the fracture outcome), BMI, smoking history, and drinking habits, with multiple imputation for BMI, smoking history, and drinking habits (Appendix pp 25–28).