학술논문

What lies behind the causal impact of body mass index level and change on human health? : added value from complementary study design and deep metabolomic phenotyping
Document Type
Electronic Thesis or Dissertation
Author
Source
Subject
adiposity
metabolites
Mendelian randomization
ALSPAC
Language
English
Abstract
Excess adipose tissue, adiposity, is associated with many diseases and overall mortality. This thesis aimed to investigate whether metabolites play an intermediary role in these associations using a variety of resources and methods to strengthen causal inference. In a comprehensive systematic review and meta-analysis (Chapter 2), the causal effects of adiposity were observed across a broad spectrum of diseases, including endometrial cancer, which was selected for subsequent analysis. These results were supported by a narrative synthesis of over 2,000 Mendelian randomization (MR) analyses, which also highlighted evidence of an association between adiposity and many, predominantly lipid-based, metabolites. Within the Avon Longitudinal Study of Parents and Children (ALSPAC), evidence for a consistent association between body mass index (BMI), waist hip ratio (WHR), and body fat percentage (BF) with up-to 230, predominantly lipid-based, metabolites and ratios was found (Chapter 4). In these linear models, the effect of adiposity persisted after adjustment for covariables and across the lifecourse. In two independent datasets, MR, a method that mitigates limitations in observational analyses, provided further evidence for an association between BMI and WHR with up-to 230, predominantly lipid-based, metabolites and ratios (Chapter 5). These effects were consistent in sensitivity analyses. The effect of BF on metabolites was frequently opposite to the effects observed for BMI and WHR, and BF in observational analyses. Evidence from observational and MR analyses identified 54 metabolites that were consistently associated with adiposity. Two of these (triglycerides in small and very small very large low density lipoprotein) were associated with endometrial cancer in MR analysis and, using multivariable MR, there was evidence for a potential intermediary role of both metabolites on the effect of WHR and BF, but not BMI, on non-endometrioid cancer, but not endometrioid cancer (Chapter 6). Weak instruments may have biased these results however. This work highlights the broad effect of adiposity on the metabolome, identifies two metabolites that may be involved in the association between adiposity and endometrial cancer, and provides a basis for future investigations of the intermediary role of metabolites.

Online Access