The Role of the Mediterranean Diet in Breast Cancer Survivorship

Abstract

Female breast cancer (BC) is the most frequently diagnosed cancer. Despite the improved long-term survival of BC patients, the demands of high-quality healthcare to improve BC survivorship are unmet. Emerging studies suggested that the Mediterranean diet (MD) is associated with health-related benefits in the general population; however, evidence is limited in BC survivors. The overall aim of this thesis was to explore the role of the MD in BC survivorship. Three main studies were conducted.
The systematic review synthesised the evidence from randomised controlled trials and observational studies suggested low-to-very-low certainty of an association between increased adherence to the MD and reduced mortality in BC survivors, and limited evidence of associations with outcomes reflecting the wellbeing of BC survivors.
The secondary data analysis used the UK Biobank to examine the associations between adherence to the MD and BC survivorship. BC survivors were identified by cancer registry data and the associations of MD with outcomes were tested using linear/logistic regression and survival analysis as appropriate for the outcome. The main findings showed that increased MD adherence was associated with improved quality of life. The associations with reduced risk of all-cause mortality and non-BC mortality were weakened when adjusting for potential confounders.
The Mendelian randomisation (MR) was conducted to further explore the causality, which used 16 genetic variants identified through conducting a genome wide association study and assessing for suitability as instrumental variables. The result of one-sample and two-sample MR suggested benefit of adherence to the MD with reduced risk of developing BC, but no causal relationships were found with disease progression outcomes.
Overall, although this thesis highlights the potential benefits of MD in BC survivorship, causal estimates were imprecise. Future research using improved genetic instruments and datasets with more BC cases are required to provide further robust evidence to inform interventions.

Date of Award	7 Jan 2026
Original language	English
Awarding Institution	University of Bristol
Supervisor	Sam D Leary (Supervisor), Angeliki Papadaki (Supervisor) & Tom Dudding (Supervisor)