Causal relevance of clonal haematopoiesis with cardiac disease and adverse remodelling: a Mendelian randomisation study.

BACKGROUND: Many observational studies highlight clonal haematopoiesis (CH) as a novel determinant of cardiovascular disease (CVD). However, disentangling cause and effect from important confounders, such as age and smoking, is challenging. OBJECTIVES: Mendelian randomisation (MR) was used to assess the causal relationships of CH with (1) major CVD outcomes associated with adverse remodelling, and (2) cardiovascular magnetic resonance (CMR) phenotypes which have not been examined previously. METHODS: Uncorrelated (r2<0.001), genome-wide significant (p<5×10-6) single nucleotide polymorphisms were extracted from Genome-Wide Association Study summary statistics for CH (any subtype), gene-specific CH subtypes (DNMT3A and TET2), and CH clonal size subtypes (small clone and large clone). Mendelian Randomisation using a Robust Adjusted Profile Score (MR-RAPS) was used for analyses on outcomes of atrial fibrillation (AF), heart failure and 13 CMR phenotypes. Multiple comparisons in the discovery analyses were accounted for by Benjamini-Hochberg correction. RESULTS: Both DNMT3A-CH and small-clone-CH were associated with increased AF risk. Overall-CH was associated with larger left ventricular end-diastolic volume. DNMT3A-CH was associated with larger right atrial size, and left and right ventricular end-diastolic volumes. TET2-CH was associated with higher myocardial native T1 time. Small-clone-CH was associated with larger left atrial size and lower aortic distensibility. CONCLUSIONS: Common forms of CH are associated with higher AF risk and adverse remodelling patterns comprising larger atrial and ventricular sizes, myocardial fibrosis, and reduced aortic compliance. Using MR methods, this study triangulates previous observational studies and provides new evidence to support likely causal links between CH and CVD. This study, for the first time, describes associations of CH with adverse CMR phenotypes suggesting early remodelling patterns; these changes may indicate a window of opportunity for intervention such as by risk stratification and early preventative strategies to improve patient outcomes; however, further examination of the utility of such interventions is warranted.