Effects of breast cancer hormone treatment on lipids and subsequent cardiovascular diseases

[摘要] Selective estrogen receptor modulators (SERMs), the earliest breast cancer (BC) hormone treatment, improve survival of hormone positive (HR+) BC patients. Aromatase inhibitors (AIs) further prolong disease free survival (DFS) for postmenopausal HR+ BC patients, but their benefits in overall survival (OS) are inconclusive. An important factor that impacts OS but not DFS is non-BC death, and cardiovascular disease (CVD) is one of the major contributors. In addition to its high prevalence in the aging population, CVD may be associated with BC cancer treatment. BC hormone treatment influences serum estrogen concentration, which is related to the occurrence of CVD; therefore, it is critical to investigate the impact of BC hormone treatment on CVD, especially in the era when prolonged hormone treatment is increasingly popular in BC care.This dissertation consists of 3 projects: 1) enhancement of an existing Bayesian Network Meta-analysis (NMA) method by allowing more reporting formats, so more studies may be included and expanding an existing consistency check of direct and indirect evidence for a NMA to include longitudinal data; 2) a Bayesian NMA exploring the effects of BC hormone drugs on changes of lipid profiles during hormone treatment; and 3) a study examining the effects of hormone treatment on dyslipidemia/coronary heart disease (CHD) throughout the whole course of BC treatment using electronic medical records. The result from the second study showed that patients on SERMs had better lipid profiles and tended to reduce the risk of dyslipidemia/CHD. The third study revealed that the beneficial effects disappeared after discontinuing SERMs. NMA demonstrated that individual AIs impact changes of lipid profiles differently. However, the third study combining AIs as a whole did not reveal the significant impact on risk of dyslipidemia/CHD, which may be caused by mixed effects of individual AIs and insufficient statistical power due to few outcome events. The public health significance of this dissertation is to provide methods promoting applicability of NMA for longitudinal data, and to show the effects of BC hormone treatment on lipids and CHD. BC patients with pre-existing risk factors of CVD should be monitored more frequently when they are on hormone treatment.

[发布机构] the University of Pittsburgh