Infection with Helicobacter pylori (H.pylori) is estimated to persist in approximately 50% of the human population. This bacterium has been commonly linked to gastrointestinal diseases such as gastritis and peptic ulcers. Current observational studies have also suggested an association with cardiovascular disease and cancer. However, there has been discordance in these findings potentially influenced by confounding. This study aims to identify single nucleotide polymorphisms (SNPs) associated with H.pylori, estimate the causal association of H.pylori with cardiovascular disease and cancer traits, and examine the direction of causality. These objectives were explored using the Avon Longitudinal Study of Parents and Children (ALSPAC) and Caerphilly Prospective Study (CaPS) cohorts. The use of genome-wide association meta-analysis was employed, and four highly suggestive SNPs possibly associated with H.pylori were identified. Two genome-wide significant SNPs identified in a previous published H.pylori genome-wide association study were used as instruments in two-sample Mendelian randomization (MR), with the four suggestive SNPs identified in the meta-analysis being excluded from analysis as they would not qualify as valid instruments, potentially violating MR assumptions. There was evidence of a causal effect of H. Pylori on LDL-cholesterol, hip circumference, breast cancer, and heart rate. However, the causal estimates suggested that H.pylori might be associated with a decrease in these traits which is in contrast to observational findings. Bidirectional MR revealed little evidence of causal effects of the outcomes on H.pylori and sensitivity analyses did not identify directional pleiotropy across instruments for each trait or heterogeneity between instruments. Overall, this study extends the scope of MR to infections and does not suffer from the limitations of observational studies, such as confounding, selection biases and reverse causation. These findings contribute to the understanding of the role of H.pylori in cardiovascular disease and cancer.
|Date of Award||20 Feb 2019|
- The University of Bristol
|Supervisor||Evangelia Stergiakouli (Supervisor) & Jie Zheng (Supervisor)|
- Mendelian randomization
- Causal inference