Environmental and genetic predictors of cardiovascular ageing
File(s)
Author(s)
Shah, Mit
Type
Thesis
Abstract
Background:
Cardiovascular ageing is a process characterised by progressive structural remodelling in the cardiovascular system and associated functional decline. Cardiovascular magnetic resonance (CMR) imaging is routinely performed in healthcare, and enables detailed assessment of this decline at high resolution. The hallmarks of ageing occur at cellular, tissue and organ levels, modulated by various risk factors and genetic contributors, however it is currently unknown how significant these are to cardiovascular ageing.
Aims:
To model biological, cardiovascular ageing using machine learning approaches in order to study the environmental and genetic associations with this process.
Methods:
We used a machine learning algorithm called CatBoost, and CMR features as inputs, to predict age in 39,449 UK Biobank participants. We first trained a model of healthy cardiovascular ageing using a development set of 5065 healthy participants that were free from cardiovascular, metabolic and respiratory disease. These participants were also determined to be on no regular medications, with healthy body mass index and also non-smokers. Having trained the healthy ageing model, we predicted cardiovascular age in an evaluation set comprising the remaining UK Biobank participants. This produced a new phenotype, “cardiovascular age delta”, the difference between a participant’s predicted and actual chronological age that was used in multivariable linear regression models to compute the associations with environmental risk factors and single nucleotide polymorphisms in genome wide association studies. We also computed a cardiovascular ageing polygenic risk score in independent participants of the UK Biobank and performed phenome wide association.
Results:
Cardiovascular ageing was found to be significantly associated with conventional cardiovascular risk factors, and variants in genes that are involved in maintaining sarcomeric structure and function, the immune response to stress, vascular integrity and calcium signalling.
Discussion:
Modelling cardiovascular ageing using multiple imaging traits has enabled mechanistic insights into this process with potential novel targets for anti-ageing therapy.
Cardiovascular ageing is a process characterised by progressive structural remodelling in the cardiovascular system and associated functional decline. Cardiovascular magnetic resonance (CMR) imaging is routinely performed in healthcare, and enables detailed assessment of this decline at high resolution. The hallmarks of ageing occur at cellular, tissue and organ levels, modulated by various risk factors and genetic contributors, however it is currently unknown how significant these are to cardiovascular ageing.
Aims:
To model biological, cardiovascular ageing using machine learning approaches in order to study the environmental and genetic associations with this process.
Methods:
We used a machine learning algorithm called CatBoost, and CMR features as inputs, to predict age in 39,449 UK Biobank participants. We first trained a model of healthy cardiovascular ageing using a development set of 5065 healthy participants that were free from cardiovascular, metabolic and respiratory disease. These participants were also determined to be on no regular medications, with healthy body mass index and also non-smokers. Having trained the healthy ageing model, we predicted cardiovascular age in an evaluation set comprising the remaining UK Biobank participants. This produced a new phenotype, “cardiovascular age delta”, the difference between a participant’s predicted and actual chronological age that was used in multivariable linear regression models to compute the associations with environmental risk factors and single nucleotide polymorphisms in genome wide association studies. We also computed a cardiovascular ageing polygenic risk score in independent participants of the UK Biobank and performed phenome wide association.
Results:
Cardiovascular ageing was found to be significantly associated with conventional cardiovascular risk factors, and variants in genes that are involved in maintaining sarcomeric structure and function, the immune response to stress, vascular integrity and calcium signalling.
Discussion:
Modelling cardiovascular ageing using multiple imaging traits has enabled mechanistic insights into this process with potential novel targets for anti-ageing therapy.
Version
Open Access
Date Issued
2023-11
Date Awarded
2024-02
Copyright Statement
Creative Commons Attribution Licence
License URL
Advisor
O'Regan, Declan
Sponsor
British Heart Foundation
Publisher Department
Institute of Clinical Sciences
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)