Causal implications of common infections and platelet function on cardiovascular disease

Student thesis: Doctoral ThesisDoctor of Philosophy (PhD)

Abstract

The modifiable risk factors, common infections and platelet function, have been shown in conventional observational studies to be associated with cardiovascular disease (CVD) incidence and progression outcomes. However, inferring causality from these observed associations are yet to be fully assessed. In this doctoral thesis, I applied epidemiological approaches to strengthen causal inferences and highlight issues related to conducting disease progression studies.

In Chapter Four, four novel genome-wide genetic loci associated with antibodies against common infections were identified, and eight human leukocyte antigen (HLA) alleles associated with Epstein-Barr virus (EBV) showed strong evidence of replication in the independent cohort, UK Biobank (UKB). These findings suggest that there is a host genetic contribution related to infection susceptibility.

In Chapter Five, findings showed that collider bias can arise in conventional observational studies and genetic studies (e.g., genome-wide association studies (GWAS), and Mendelian randomization (MR) studies) when a variable is selected upon based on factors such as disease status (e.g., myocardial infarction (MI) incidence) or medication use. Resultantly, proper consideration to account for this issue is important as otherwise findings may be misleading.

In Chapter Six, three observational associations between antibodies against common infections and thrombotic CVD outcomes were identified, although MR analyses showed little evidence for causal effects. However, the MR analyses had insufficient statistical power as it was limited by the number of available instrumental variables (IVs) for each exposure of interest.

In Chapter Seven, there was evidence of a causal effect between antibodies against EBV and platelet count (PLT). Although findings require replication as there was only one available IV associated with antibodies against EBV and consequently MR sensitivity analyses were unable to be conducted.

In Chapter Eight, findings showed evidence of a causal effect between PLT and all thrombotic CVD outcomes of interest. This suggests that PLT, as a proxy for platelet function, has a causal role in thrombotic events.
Date of Award21 Mar 2023
Original languageEnglish
Awarding Institution
  • University of Bristol
SupervisorLavinia Paternoster (Supervisor), Rebecca Richmond (Supervisor), George Davey Smith (Supervisor) & Alastair W Poole (Supervisor)

Cite this

'