Clonal haematopoiesis and cardiovascular disease, defining risk, filling gaps, and framing the future Authors: Alec P. Morley, Paul Carter, Rupen Hargreaves, Dorina-Gabriela Condurache, Michael Anderson, Sean Wen, Zahra Raisi-Estabragh Published: 2026 ABSTRACT Clonal haematopoiesis (CH) refers to the clonal expansion of haematopoietic stem cells in the absence of overt haematological malignancy and is common in older individuals. Recent evidence has heralded CH as a novel determinant of cardiovascular disease. This review explores the associations of CH with ischaemic heart disease, heart failure, atrial fibrillation and cardio-oncology, highlighting biases in the literature, exploring effects of mutation and clone size, concluding with clinical implications. Introduction Haematopoiesis begins with multipotent haematopoietic stem cells (HSCs), which can self-renew and differentiate into more specialised blood cells. Somatic mutations may give proliferative advantage to HSCs, leading to their clonal expansion. Established risk factors for these mutations are older age1, history of cytotoxic exposures such as chemoradiotherapy2–4, tobacco smoking5, sex6, obesity7, telomere length8, genetic polymorphisms6,9 (such as the TCL1A locus), ancestry10 and medications11. More than 70 different gene drivers have been implicated in clonal expansion of HSCs12, conferring different fitness effects13, with, for example, splicing factor and JAK2 mutations having relatively higher fitness compared to DNMT3A mutations, which have relatively lower fitness14. However, rates of clonal expansion can vary across the lifetime15, and be influenced by germline genetics14 and environmental exposures16, and are not driven by mutation alone15. The proportion of mutant cells found in the blood, termed the variant allele fraction (VAF), can be measured, which reflects clone size at that point, with the increase over time corresponding to clonal fitness. While people with clonal expansion of HSCs have an increased relative risk of malignant transformation, this requires additional mutations, and most never develop malignancies17, unless driven by higher fitness mutations (such as within splicing factors). Clonal expansion of HSCs with demonstrable preleukemic driver mutations to a VAF > 2% in the absence of malignancy has historically been termed clonal haematopoiesis of indeterminate potential (CHIP), although more recently updated to clonal haematopoiesis (CH) given increased understanding of prognostically significant disease associations and of smaller clones not meeting the 2% VAF threshold. Another form of clonal expansion is characterised by chromosomal abnormalities and is termed mosaic chromosomal alteration (mCA). mCA encompasses three subtypes, namely autosomal mCA, mosaic loss of chrX (mLOX) and mosaic loss of chrY (mLOY). Population studies demonstrate that CH is common, especially in older adults, affecting more than 10% of people over 65-years old1. CH is linked to a heightened risk of cardiovascular6,18–37, oncological6,38, metabolic7,39,40, hepatic41, respiratory42,43 and renal44 conditions as well as a steep 40% increase in mortality risk17, which has been supported by a recent meta-analysis45. Among these, cardiovascular diseases (CVDs) are the most important contributor to excess mortality18 – with growing evidence indicating associations with increased risk of atherosclerotic disease6,18–23, heart failure6,20,23–29 and atrial fibrillation6,30–33, as shown in Fig. 1.Fig. 1An overview of the pathogenesis of CH.CH clonal haematopoiesis, VAF variant allele fraction. Figure was created in BioRender. Morley, A. (2026) https://BioRender.com/qxzcp2l. An overview of the pathogenesis of CH. CH clonal haematopoiesis, VAF variant allele fraction. Figure was created in BioRender. Morley, A. (2026) https://BioRender.com/qxzcp2l. While some researchers have heralded CH as a novel CVD determinant, others highlight the risk of confounding with major known vascular risk factors, such as age and smoking. The bidirectional relationship between CH and CVD is recognised, and reverse causation has been raised as a potential factor in augmenting apparent associations of CH with adverse cardiovascular outcomes20,46. The issue of survival bias is also notable across many studies, with the omission of individuals at the highest mortality risk in existing cohorts potentially distorting some observed relationships with age-related illnesses. Methodological strategies to disentangle cause-and-effect and mitigate key biases are important for understanding the independent causal relationships of CH. The existing VAF threshold for CH is set pragmatically at 2% (reflecting 4% of heterozygotic cells), based in part on technical limits of detection. While modern high-resolution sequencing technologies permit detection of driver mutations at very low VAF levels, the clinical utility of ultrasensitive measurements and optimal VAF threshold for CH is not yet established. In addition, the number of driver genes included in CH definitions and the panel of CH genes varies widely even across key publications6,9,12,17, and a high proportion of CH is with unknown drivers47. Emerging evidence suggests differential disease risk across CH driver mutations and clone sizes. However, many of these analyses have limited statistical power for rarer mutations and smaller clones, limiting the robustness of their conclusions. Unravelling mutation and clone size-specific relationships with CVD subtypes is essential for understanding biological mechanisms and identifying novel therapeutic targets. Furthermore, there is an urgent need for a mutually agreed consensus on defining CH, given that associations currently described in the literature are dependent upon often differing CH definitions, and lead to limited cross-comparability and result in heterogeneity in reported effect sizes20,46,48. This narrative review evaluates the latest research linking CH with specific cardiovascular conditions, varying by mutation and clone size, and highlights sources of bias and gaps in current literature. We further discuss clinical implications and the current position of CH in relation to the Wilson–Jungner screening criteria. Literature search Electronic literature databases (MEDLINE and Embase) were searched using relevant search terms to cover all published literature from 1946 to December 2025, yielding 299 relevant publications post-screening. The number of publications relating to CH has increased substantially since 2017, from 24 published in 2017 to an almost exponential increase of over 200 in the last three years. The most relevant publications are selected for discussion in this review and grouped into three major CVD categories of atherosclerotic disease, heart failure (HF) and atrial fibrillation (AF). We also discuss CH in the context of cardio-oncology, a special population of patients at heightened risk of both CH and cardiovascular complications. The CH mutations associated with CVD discussed in this review are summarised in Table 1 and the key observational studies included in the review are summarised in Table 2.Table 1Key known mutation-CH associations which are implicated in CVDsGene (Protein)RoleMutation typeProportion of CHDNMT3A (DNA methyltransferase)Epigenetic regulatorLoss-of-function, (non-R882/R882 missense)~40–60%1,2,6,9,10,12,13,15,17,20,45–47,62,63,88,101TET2 (Tet dioxygenase)Epigenetic DNA demethylationLoss-of-function~10–35%1,2,6,9,10,12,13,15,17,20,45–47,62,63,88,101ASXL1 (Polycomb regulator)Chromatin regulatorFrameshift/nonsense~3–15%1,2,6,9,10,12,13,15,17,20,45–47,62,63,88,101JAK2 (tyrosine kinase)Myeloproliferative driverV617F missense~1–5%1,2,6,9,10,12,13,17,20,45–47,62,63TP53 (tumour suppressor)DNA damage responseMissense or truncating~1–8%1,2,6,9,10,12,13,15,17,20,45–47,62,63,101PPM1D (p53 pathway phosphatase)DNA damage responseTruncating mutations in exon 5 and 6~3–8%2,6,9,10,12,15,20,45,47,88CH clonal haematopoiesis, CVD cardiovascular disease.Table 2Summary of key observational studies of CH and CVDFirst author, yearStudy populationSamplesizeAge (mean/median/range)Ethnic backgroundCH definitionDisease focusEffect size(HR/OR, 95% CI)Atherosclerotic diseaseJaiswal et al., 201718Adults from 4 case–control studies (BioImage, MDC, ATVB, PROMIS); no known haematologic malignancy8255 (4726 CAD cases and 3529 controls)60–70 years (BioImage and MDC)Predominantly White/European (BioImage/MDC/ATVB); South Asian (PROMIS)WES-defined CH; ≥1 somatic mutation in 74 myeloid malignancy-associated genes at VAF ≥ 2%; large clone CH: VAF ≥ 10%CAD and early-onset MICH was associated with CAD (HR: 1.9, 95% CI: 1.4–2.7, p < 0.001) and early-onset MI (OR: 4.0, 95% CI: 2.4–6.7, p < 0.001); VAFs ≥10% associated with coronary artery calcification (OR: 12; 2.4–64, p = 0.002) in the BioImage cohortJaiswal et al., 201718Adults from 2 case–control studies (BioImage, MDC) and 3 prospective cohorts (JHS, FUSION, FHS); no known haematologic malignancy4578(240 CH cases and 4438 controls)60–70 years (BioImage and MDC)Predominantly White/European (BioImage/MDC/FUSION/FHS); African–American (JHS)WES-defined CH; ≥1 somatic mutation in 74 myeloid malignancy-associated genes at VAF ≥ 2%; gene-specific analyses (DNMT3A, TET2, ASXL1 and JAK2);CAD and early-onset MIDNMT3A, ASXL1, JAK2 and other-CH all associated with CAD (DNMT3A: HR: 1.7, 95% CI: 1.1–2.6, p = 0.01; ASXL1: HR: 2.0 95% CI: 1.0–3.9, p = 0.05; JAK2: HR: 12, 95% CI: 3.8–38.4, p < 0.001; other: HR: 2.2, 95% CI: 1.3–3.7, p = 0.002), TET2, ASXL1, JAK2 and other mutations all associated with early onset MI (p ≤ 0.02).Kar et al., 20226Adults from UKB with WES200,45338–72 yearsPredominantly White/EuropeanWES of 43 CH genes; somatic mutations filtered against predefined CH driver variants; stratified by gene (DNMT3A, TET2, ASXL1, JAK2, SRSF2 + SF3B1) and clone size (VAF < 0.1 vs. ≥0.1)Incident CAD and MIOverall-CH and incident CAD (HR: 1.36, 95% CI: 1.22–1.53, p = 2.00 × 10−7, with correction for smoking)Overall-CH and incident MI (HR:1.32, 95% CI: 1.23–1.42, p = 1.25 × 10−12, with correction for smoking)Zekavat et al., 202319Adults from MGBB and UKB; leukaemia-free50,122(MGBB: 12,465 [657 CH cases, 11,808 controls]; UKB: 37,657 [2,194 CH cases, 35,463 controls])MGBB mean age 60.12 years in CH cases, 46.13 years in controls; UKB 60.59 years in CH cases, 56.81 years in controls86.1% White CH cases, 80% controls (MGBB); 100% White CH cases & controls (UKB)WES-defined CH; somatic mutations in haematological malignancy-related genes at VAF ≥ 2%; gene-specific analyses (DNMT3A, TET2, TP53, PPM1D, ASXL1 and splicing factors); large clones VAF > 10%Incident CADGene-specific associations with CAD including TET2-CH (HR: 2.05, 95% CI: 1.54 − 2.71, p = 6.3 × 10−7), TP53-CH (HR: 2.31, 95% CI: 1.03 − 5.16, p = 0.042), PPM1D-CH (HR: 2.63, 95% CI: 1.40 − 4.95, p = 2.7 × 10−3), ASXL1-CH (HR: 1.73, 95% CI: 1.09 − 2.76, p = 0.021),Stacey et al., 202320Adults from UKB and Icelandic deCODE cohort176,219(UKB: 130,709; deCODE: 45,510)Median58.4 years (UKB); 53.0 years (deCODE)Predominantly White/EuropeanMutational-barcode CH from WGS; ≥20 low-VAF singleton mosaic mutations (VAF 0.1–0.25) after artefact/germline filteringCAD, MI, revascularisation composite, HF; all-cause mortalityNo association with CVD in UKB (HR: 1.01, 95% CI 0.94–1.08, p = 0.88)Heimlich et al., 202421Prospective cohort of patients undergoing invasive coronary angiography1142Median 68 years (CH) vs. 59 years (non-CH)73% WhiteTargeted deep sequencing; ≥1 CH driver mutation at VAF ≥ 2%Coronary stenosis (left main, LAD), including obstructive/non-obstructive lesionsOverall-CH with left main stenosis (OR: 2.44, 95% CI: 1.40–4.27, p = 0.021)TET2-CH with obstructive LAD stenosis (OR: 2.62, 95% CI: 1.34–5.11, p = 0.039)Diez-Diez et al., 202422PESA healthy cohort3692Median45 yearsPredominantly White/EuropeanTargeted sequencing of 54 CH-driver genes; CH = myeloid CH genes; large clones VAF ≥ 2%Subclinical incident atherosclerosisCH increased risk of de novo femoral atherosclerosis at 3 years (OR:1.53, 95% CI: 1.12–2.07, p = 0.007) with larger clones (VAF ≥ 2%) showing a higher risk (OR: 2.06, 95% CI 1.22 − 3.37, p = 0.005).Zuriaga et al., 202423Adults from MGBB and UKB; leukaemia-free474,417 (MGBB: 37,181; UKB: 437, 236)MGBB: mean 61.7 ± 11.3 years; UKB: mean 56.5 ± 8.1 years88.4% White (MGBB); 83.7% White (UKB)WES-defined CH; somatic mutations in haematological malignancy-related genes at VAF ≥ 2%; gene-specific analyses (TET2, DNMT3A and ASXL1)Composite CVD outcome (CAD, MI, ischaemic stroke andall-cause mortality) and MI; prevalent (MGBB); incident (UKB)Within MGBB, associations of overall-CH, TET2-CH and ASXL1-CH with prevalent composite CVD outcome.Within UKB, associations of overall-CH, TET2-CH, DNMT3A-CH and ASXL1-CH with incident composite CVD, associations of overall-CH and ASXL1-CH with incident MI (overall-CH: HR: 1.09, 95% CI: 1.02 − 1.15, p = 7.5 × 10−3; ASXL1-CH: HR: 1.29, 95% CI: 1.10 − 1.51, p = 1.5 × 10−3)Heart FailureDorsheimer et al, 201824Chronic ischaemic HF patients taken from trials examining the effects of intracoronary administration of autologous bone marrow cells200Median age 65 [56–72] years-Deep amplicon sequencing of 56 CH/myeloid genes in bone marrow; ≥1 somatic mutation with VAF ≥ 2%; prognostic analyses focused on DNMT3A/TET2Death or HF hospitalisationHigher mortality with HF hospitalisation among those with DNMT3A or TET2 mutations (HR 2.1, 95% CI: 1.1–4.0, p = 0.02). Higher mortality associated with clonal size (VAF)Pascual-Figal et al., 202125Single-centre cohort of ambulatory patients with ischaemic and non-ischaemic HF with reduced left ventricular ejection fraction62Mean74 ± 7 years-Targeted deep sequencing (54 CH genes); ≥1 driver mutation with VAF ≥ 2%HF death or HF hospitalisationDNMT3A or TET2 mutations showed accelerated HF progression in terms of death (HR: 2.79, 95% CI 1.31 − 5.92, p = 0.008), death or HF hospitalisation (HR: 3.84, 95% CI: 1.84 − 8.04, p < 0.001) and HF-related death or HF hospitalisation (HR: 4.41, 95% CI: 2.15 − 9.03, p < 0.001)Kar et al., 20226Adults from UKB with WES200,45338–72 yearsPredominantly White/EuropeanWES of 43 CH genes; somatic mutations filtered against predefined CH driver variants; stratified by gene (DNMT3A, TET2, ASXL1, JAK2, SRSF2 + SF3B1) and clone size (VAF < 0.1 vs. ≥0.1)Incident HFOverall-CH and incident HF (HR: 1.21, 95% CI: 1.07–1.36, p = 1.38 × 10−2)Large clone CH and incident HF (HR: 1.30, 95% CI: 1.09 − 1.55, p = 1.76 × 10−2)Stacey et al., 202320Adults from UKB and Icelandic deCODE cohort176,219(UKB 130,709; deCODE 45,510)Median58.4 years (UKB); 53.0 years (deCODE)Predominantly White/EuropeanMutational-barcode CH from WGS; ≥20 low-VAF singleton mosaic mutations (VAF 0.1–0.25) after artefact/germline filteringCAD, MI, revascularisation composite, HF; all-cause mortalityNo association with CVD in UKB (HR: 1.01, 95% CI 0.94–1.08, p = 0.88)Schuermans et al., 202428Adults from JHS and WHI; no prevalent HF8090(JHS 2927; WHI 5163)JHS: median 56 years [46–65]; WHI: median 67 years [62–72]100% Black (JHS); 66% non-Hispanic White (WHI)WGS-defined CH; ≥1 somatic mutation in 74 leukaemia-associated genes at VAF > 2%. Gene-specific subtypes: DNMT3A and TET2. Large clone CH: VAF > 10%HFpEF and HFrEFTET2-CH was significantly associated with HFpEF (HR: 2.35, 95% CI: 1.34 − 4.11, p = 0.003)Sikking et al., 202429Non-ischaemic DCM patients from the Maastricht Cardiomyopathy Registry520Median 57 years [49–65]Predominantly White/EuropeanCH driver mutations with VAF ≥ 0.01% detected using the single-molecule molecular inversion probe techniqueNon-ischaemic DCM prognosisCH with a VAF cutoff of 0.36% was associated with a higher risk of cardiac death (HR: 2.33, 1.24 − 4.40). CH with a VAF cutoff of 0.06% was associated with a high risk of all-cause mortality (HR: 1.72, 95% CI 1.10 − 2.69)Zuriaga et al., 202423Adults from MGBB and UKB; leukaemia-free474,417 (MGBB: 37,181; UKB: 437, 236)MGBB: mean 61.7 ± 11.3 years; UKB: mean 56.5 ± 8.1 years88.4% White (MGBB); 83.7% White (UKB)WES-defined CH; somatic mutations in haematological malignancy-related genes at VAF ≥ 2%; gene-specific analyses (TET2, DNMT3A and ASXL1)Composite CVD outcome (CAD, MI, ischaemic stroke, andall-cause mortality) and HF; prevalent (MGBB); incident (UKB)Within MGBB, associations of overall-CH, TET2-CH and ASXL1-CH with prevalent composite CVD outcome, association of overall-CH with prevalent HF (OR: 1.08, 95% CI 1.00 − 1.16, p = 0.05)Within UKB, associations of overall-CH, TET2-CH, DNMT3A-CH and ASXL1-CH with incident composite CVD, associations of overall-CH, DNMT3A-CH and ASXL1-CH with incident HF (overall-CH: HR: 1.22, 95% CI: 1.15 − 1.30, p = 7.3 × 10−11; DNMT3A-CH; HR: 1.15, 95% CI: 1.05 − 1.24, p = 1.2 × 10−3; ASXL1-CH: HR: 1.40, 95% CI: 1.19 − 1.64, p = 4.8 × 10−5)Atrial FibrillationKar et al., 20226Adults from UKB with WES200,45338–72 yearsPredominantly White/EuropeanWES of 43 CH genes; somatic mutations filtered against predefined CH driver variants; stratified by gene (DNMT3A, TET2, ASXL1, JAK2, SRSF2 + SF3B1) and clone size (VAF < 0.1 vs. ≥0.1)Prevalent AFOverall-CH and AF (HR: 1.13, 95% CI: 1.04–1.22, p = 1.68 × 10−2, with correction for age and smoking)Ahn et al., 202430Adults with non-valvular AF and non-AF healthy controls from Seoul National University Hospital; prognosis evaluated in AF patients in UKB25,631(Korean discovery cohort: 4345; UKB: 21,286)50–79 yearsEast Asian (Koreandiscovery cohort); predominantly White (UKB)Somatic mutation in any of 24 CH-associated genes with VAF ≥ 1.5%, detected by deep-targeted sequencingAFHigher prevalence of overall-CH, DNMT3A-CH and TET2-CH in AF within discovery cohort (overall-CH: HR: 1.44, 95% CI: 1.16 − 1.77, p < 0.001; DNMT3A-CH; HR: 1.57, 95% CI: 1.21 − 2.04, p < 0.001; TET2-CH: HR: 1.72, 95% CI: 1.16 − 2.55, p = 0.008). CH in AF associated with higher incidence of composite HF, ischaemic stroke or death (HR:1.32; 1.20-1.45, p < 0.001).Schuermans et al., 202431Adults from UKB without prevalent arrhythmias/ haematologic malignancy410,702Median 58 years [50–63]Predominantly White/EuropeanWES-defined CH; VAF ≥ 2% (any CH) and VAF ≥ 10% (large clone CH); gene subtypes (DNMT3A, TET2, ASXL1, JAK2, PPM1D, TP53 and spliceosome)Incident arrhythmias: brady, supra/ventricular, AF, cardiac arrest; myocardial fibrosis in the imaging subsetAny CH: HR 1.09 to 1.16 across arrhythmia subtypes (including AF: HR: 1.11, 95% CI: 1.04–1.18, p = 0.001), cardiac arrest (HR: 1.29, 95% CI: 1.06–1.58, p = 0.012) and myocardial fibrosis (OR: 1.19, 95% CI: 1.01 − 1.41, p = 0.035), with increases in magnitude of all associations for large clone CHSaadatagah et al., 202432Adults form ARIC and UKB without haematologic malignancy/mitral valve disease199,982 (ARIC: 4131; UKB: 195,85)ARIC: mean 75.8 ± 5.2 years; UKB: mean 56.4 ± 8.1 years77% White (ARIC); 94% White (UKB)WES-defined CH; VAF ≥ 2% (any CH) and VAF ≥ 10% (large clone CH); gene subtypes (DNMT3A, TET2 and ASXL1)AFAny large clone CH: incident AF (HR: 1.12, 95% CI: 1.01 − 1.25, p = 0.036)TET2-CH: incident AF (HR: 1.18, 95% CI: 1.01 − 1.38, p = 0.043), with risk increasing for larger clone size (HR: 1.29, 95% CI: 1.05 − 1.59, p = 0.015)ASXL1-CH: incident AF (HR: 1.33, 95% CI: 1.02 − 1.73, p = 0.033), with risk increasing for larger clone size (HR: 1.45, 95% CI: 1.02 − 2.07, p = 0.040)Teng et al., 202533Patients with JAK2V617F–positive MPN439Median57 years [12–87]Predominantly East AsianJAK2V617F gene mutation; additional somatic mutation profiling by NGS (TET2, DNMT3A, ASXL1, FAT1 and EP300)AFTET2 mutation independently predicted AF (HR: 4.36, 95% CI: 1.05–18.06, p = 0.042); IL-1β elevation also predicted AF (HR: 5.48, 95% CI: 1.55–28.12, p = 0.012)AF atrial fibrillation, ARIC Atudy, ATVB Atherosclerosis, Thrombosis and Vascular Biology Italian Study, CAD coronary artery disease, CH clonal haematopoiesis, CI confidence interval, CVD cardiovascular disease, DCM dilated cardiomyopathy, FHS Framingham Heart Study, FUSION Finland–United States Investigation of NIDDM [Non–Insulin-Dependent Diabetes Mellitus] Genetics, HF heart failure, HFpEF heart failure with preserved ejection fraction, HFrEF heart failure with reduced ejection fraction, HR hazard ratio, JHS Jackson Heart Study, LAD left anterior descending artery, MDC Malmö Diet and Cancer, MGBB Mass General Brigham Biobank, MI myocardial infarction, NGS next-generation sequencing, OR odds ratio, PESA Progression of Early Subclinical Atherosclerosis, PROMIS Pakistan Risk of Myocardial Infarction Study, UKB UK Biobank, VAF variant allele fraction, WES whole-exome sequencing, WGS whole-genome sequencing, WHI Women’s Health Initiative. Table 1Key known mutation-CH associations which are implicated in CVDsGene (Protein)RoleMutation typeProportion of CHDNMT3A (DNA methyltransferase)Epigenetic regulatorLoss-of-function, (non-R882/R882 missense)~40–60%1,2,6,9,10,12,13,15,17,20,45–47,62,63,88,101TET2 (Tet dioxygenase)Epigenetic DNA demethylationLoss-of-function~10–35%1,2,6,9,10,12,13,15,17,20,45–47,62,63,88,101ASXL1 (Polycomb regulator)Chromatin regulatorFrameshift/nonsense~3–15%1,2,6,9,10,12,13,15,17,20,45–47,62,63,88,101JAK2 (tyrosine kinase)Myeloproliferative driverV617F missense~1–5%1,2,6,9,10,12,13,17,20,45–47,62,63TP53 (tumour suppressor)DNA damage responseMissense or truncating~1–8%1,2,6,9,10,12,13,15,17,20,45–47,62,63,101PPM1D (p53 pathway phosphatase)DNA damage responseTruncating mutations in exon 5 and 6~3–8%2,6,9,10,12,15,20,45,47,88CH clonal haematopoiesis, CVD cardiovascular disease. Key known mutation-CH associations which are implicated in CVDs CH clonal haematopoiesis, CVD cardiovascular disease. Table 2Summary of key observational studies of CH and CVDFirst author, yearStudy populationSamplesizeAge (mean/median/range)Ethnic backgroundCH definitionDisease focusEffect size(HR/OR, 95% CI)Atherosclerotic diseaseJaiswal et al., 201718Adults from 4 case–control studies (BioImage, MDC, ATVB, PROMIS); no known haematologic malignancy8255 (4726 CAD cases and 3529 controls)60–70 years (BioImage and MDC)Predominantly White/European (BioImage/MDC/ATVB); South Asian (PROMIS)WES-defined CH; ≥1 somatic mutation in 74 myeloid malignancy-associated genes at VAF ≥ 2%; large clone CH: VAF ≥ 10%CAD and early-onset MICH was associated with CAD (HR: 1.9, 95% CI: 1.4–2.7, p < 0.001) and early-onset MI (OR: 4.0, 95% CI: 2.4–6.7, p < 0.001); VAFs ≥10% associated with coronary artery calcification (OR: 12; 2.4–64, p = 0.002) in the BioImage cohortJaiswal et al., 201718Adults from 2 case–control studies (BioImage, MDC) and 3 prospective cohorts (JHS, FUSION, FHS); no known haematologic malignancy4578(240 CH cases and 4438 controls)60–70 years (BioImage and MDC)Predominantly White/European (BioImage/MDC/FUSION/FHS); African–American (JHS)WES-defined CH; ≥1 somatic mutation in 74 myeloid malignancy-associated genes at VAF ≥ 2%; gene-specific analyses (DNMT3A, TET2, ASXL1 and JAK2);CAD and early-onset MIDNMT3A, ASXL1, JAK2 and other-CH all associated with CAD (DNMT3A: HR: 1.7, 95% CI: 1.1–2.6, p = 0.01; ASXL1: HR: 2.0 95% CI: 1.0–3.9, p = 0.05; JAK2: HR: 12, 95% CI: 3.8–38.4, p < 0.001; other: HR: 2.2, 95% CI: 1.3–3.7, p = 0.002), TET2, ASXL1, JAK2 and other mutations all associated with early onset MI (p ≤ 0.02).Kar et al., 20226Adults from UKB with WES200,45338–72 yearsPredominantly White/EuropeanWES of 43 CH genes; somatic mutations filtered against predefined CH driver variants; stratified by gene (DNMT3A, TET2, ASXL1, JAK2, SRSF2 + SF3B1) and clone size (VAF < 0.1 vs. ≥0.1)Incident CAD and MIOverall-CH and incident CAD (HR: 1.36, 95% CI: 1.22–1.53, p = 2.00 × 10−7, with correction for smoking)Overall-CH and incident MI (HR:1.32, 95% CI: 1.23–1.42, p = 1.25 × 10−12, with correction for smoking)Zekavat et al., 202319Adults from MGBB and UKB; leukaemia-free50,122(MGBB: 12,465 [657 CH cases, 11,808 controls]; UKB: 37,657 [2,194 CH cases, 35,463 controls])MGBB mean age 60.12 years in CH cases, 46.13 years in controls; UKB 60.59 years in CH cases, 56.81 years in controls86.1% White CH cases, 80% controls (MGBB); 100% White CH cases & controls (UKB)WES-defined CH; somatic mutations in haematological malignancy-related genes at VAF ≥ 2%; gene-specific analyses (DNMT3A, TET2, TP53, PPM1D, ASXL1 and splicing factors); large clones VAF > 10%Incident CADGene-specific associations with CAD including TET2-CH (HR: 2.05, 95% CI: 1.54 − 2.71, p = 6.3 × 10−7), TP53-CH (HR: 2.31, 95% CI: 1.03 − 5.16, p = 0.042), PPM1D-CH (HR: 2.63, 95% CI: 1.40 − 4.95, p = 2.7 × 10−3), ASXL1-CH (HR: 1.73, 95% CI: 1.09 − 2.76, p = 0.021),Stacey et al., 202320Adults from UKB and Icelandic deCODE cohort176,219(UKB: 130,709; deCODE: 45,510)Median58.4 years (UKB); 53.0 years (deCODE)Predominantly White/EuropeanMutational-barcode CH from WGS; ≥20 low-VAF singleton mosaic mutations (VAF 0.1–0.25) after artefact/germline filteringCAD, MI, revascularisation composite, HF; all-cause mortalityNo association with CVD in UKB (HR: 1.01, 95% CI 0.94–1.08, p = 0.88)Heimlich et al., 202421Prospective cohort of patients undergoing invasive coronary angiography1142Median 68 years (CH) vs. 59 years (non-CH)73% WhiteTargeted deep sequencing; ≥1 CH driver mutation at VAF ≥ 2%Coronary stenosis (left main, LAD), including obstructive/non-obstructive lesionsOverall-CH with left main stenosis (OR: 2.44, 95% CI: 1.40–4.27, p = 0.021)TET2-CH with obstructive LAD stenosis (OR: 2.62, 95% CI: 1.34–5.11, p = 0.039)Diez-Diez et al., 202422PESA healthy cohort3692Median45 yearsPredominantly White/EuropeanTargeted sequencing of 54 CH-driver genes; CH = myeloid CH genes; large clones VAF ≥ 2%Subclinical incident atherosclerosisCH increased risk of de novo femoral atherosclerosis at 3 years (OR:1.53, 95% CI: 1.12–2.07, p = 0.007) with larger clones (VAF ≥ 2%) showing a higher risk (OR: 2.06, 95% CI 1.22 − 3.37, p = 0.005).Zuriaga et al., 202423Adults from MGBB and UKB; leukaemia-free474,417 (MGBB: 37,181; UKB: 437, 236)MGBB: mean 61.7 ± 11.3 years; UKB: mean 56.5 ± 8.1 years88.4% White (MGBB); 83.7% White (UKB)WES-defined CH; somatic mutations in haematological malignancy-related genes at VAF ≥ 2%; gene-specific analyses (TET2, DNMT3A and ASXL1)Composite CVD outcome (CAD, MI, ischaemic stroke andall-cause mortality) and MI; prevalent (MGBB); incident (UKB)Within MGBB, associations of overall-CH, TET2-CH and ASXL1-CH with prevalent composite CVD outcome.Within UKB, associations of overall-CH, TET2-CH, DNMT3A-CH and ASXL1-CH with incident composite CVD, associations of overall-CH and ASXL1-CH with incident MI (overall-CH: HR: 1.09, 95% CI: 1.02 − 1.15, p = 7.5 × 10−3; ASXL1-CH: HR: 1.29, 95% CI: 1.10 − 1.51, p = 1.5 × 10−3)Heart FailureDorsheimer et al, 201824Chronic ischaemic HF patients taken from trials examining the effects of intracoronary administration of autologous bone marrow cells200Median age 65 [56–72] years-Deep amplicon sequencing of 56 CH/myeloid genes in bone marrow; ≥1 somatic mutation with VAF ≥ 2%; prognostic analyses focused on DNMT3A/TET2Death or HF hospitalisationHigher mortality with HF hospitalisation among those with DNMT3A or TET2 mutations (HR 2.1, 95% CI: 1.1–4.0, p = 0.02). Higher mortality associated with clonal size (VAF)Pascual-Figal et al., 202125Single-centre cohort of ambulatory patients with ischaemic and non-ischaemic HF with reduced left ventricular ejection fraction62Mean74 ± 7 years-Targeted deep sequencing (54 CH genes); ≥1 driver mutation with VAF ≥ 2%HF death or HF hospitalisationDNMT3A or TET2 mutations showed accelerated HF progression in terms of death (HR: 2.79, 95% CI 1.31 − 5.92, p = 0.008), death or HF hospitalisation (HR: 3.84, 95% CI: 1.84 − 8.04, p < 0.001) and HF-related death or HF hospitalisation (HR: 4.41, 95% CI: 2.15 − 9.03, p < 0.001)Kar et al., 20226Adults from UKB with WES200,45338–72 yearsPredominantly White/EuropeanWES of 43 CH genes; somatic mutations filtered against predefined CH driver variants; stratified by gene (DNMT3A, TET2, ASXL1, JAK2, SRSF2 + SF3B1) and clone size (VAF < 0.1 vs. ≥0.1)Incident HFOverall-CH and incident HF (HR: 1.21, 95% CI: 1.07–1.36, p = 1.38 × 10−2)Large clone CH and incident HF (HR: 1.30, 95% CI: 1.09 − 1.55, p = 1.76 × 10−2)Stacey et al., 202320Adults from UKB and Icelandic deCODE cohort176,219(UKB 130,709; deCODE 45,510)Median58.4 years (UKB); 53.0 years (deCODE)Predominantly White/EuropeanMutational-barcode CH from WGS; ≥20 low-VAF singleton mosaic mutations (VAF 0.1–0.25) after artefact/germline filteringCAD, MI, revascularisation composite, HF; all-cause mortalityNo association with CVD in UKB (HR: 1.01, 95% CI 0.94–1.08, p = 0.88)Schuermans et al., 202428Adults from JHS and WHI; no prevalent HF8090(JHS 2927; WHI 5163)JHS: median 56 years [46–65]; WHI: median 67 years [62–72]100% Black (JHS); 66% non-Hispanic White (WHI)WGS-defined CH; ≥1 somatic mutation in 74 leukaemia-associated genes at VAF > 2%. Gene-specific subtypes: DNMT3A and TET2. Large clone CH: VAF > 10%HFpEF and HFrEFTET2-CH was significantly associated with HFpEF (HR: 2.35, 95% CI: 1.34 − 4.11, p = 0.003)Sikking et al., 202429Non-ischaemic DCM patients from the Maastricht Cardiomyopathy Registry520Median 57 years [49–65]Predominantly White/EuropeanCH driver mutations with VAF ≥ 0.01% detected using the single-molecule molecular inversion probe techniqueNon-ischaemic DCM prognosisCH with a VAF cutoff of 0.36% was associated with a higher risk of cardiac death (HR: 2.33, 1.24 − 4.40). CH with a VAF cutoff of 0.06% was associated with a high risk of all-cause mortality (HR: 1.72, 95% CI 1.10 − 2.69)Zuriaga et al., 202423Adults from MGBB and UKB; leukaemia-free474,417 (MGBB: 37,181; UKB: 437, 236)MGBB: mean 61.7 ± 11.3 years; UKB: mean 56.5 ± 8.1 years88.4% White (MGBB); 83.7% White (UKB)WES-defined CH; somatic mutations in haematological malignancy-related genes at VAF ≥ 2%; gene-specific analyses (TET2, DNMT3A and ASXL1)Composite CVD outcome (CAD, MI, ischaemic stroke, andall-cause mortality) and HF; prevalent (MGBB); incident (UKB)Within MGBB, associations of overall-CH, TET2-CH and ASXL1-CH with prevalent composite CVD outcome, association of overall-CH with prevalent HF (OR: 1.08, 95% CI 1.00 − 1.16, p = 0.05)Within UKB, associations of overall-CH, TET2-CH, DNMT3A-CH and ASXL1-CH with incident composite CVD, associations of overall-CH, DNMT3A-CH and ASXL1-CH with incident HF (overall-CH: HR: 1.22, 95% CI: 1.15 − 1.30, p = 7.3 × 10−11; DNMT3A-CH; HR: 1.15, 95% CI: 1.05 − 1.24, p = 1.2 × 10−3; ASXL1-CH: HR: 1.40, 95% CI: 1.19 − 1.64, p = 4.8 × 10−5)Atrial FibrillationKar et al., 20226Adults from UKB with WES200,45338–72 yearsPredominantly White/EuropeanWES of 43 CH genes; somatic mutations filtered against predefined CH driver variants; stratified by gene (DNMT3A, TET2, ASXL1, JAK2, SRSF2 + SF3B1) and clone size (VAF < 0.1 vs. ≥0.1)Prevalent AFOverall-CH and AF (HR: 1.13, 95% CI: 1.04–1.22, p = 1.68 × 10−2, with correction for age and smoking)Ahn et al., 202430Adults with non-valvular AF and non-AF healthy controls from Seoul National University Hospital; prognosis evaluated in AF patients in UKB25,631(Korean discovery cohort: 4345; UKB: 21,286)50–79 yearsEast Asian (Koreandiscovery cohort); predominantly White (UKB)Somatic mutation in any of 24 CH-associated genes with VAF ≥ 1.5%, detected by deep-targeted sequencingAFHigher prevalence of overall-CH, DNMT3A-CH and TET2-CH in AF within discovery cohort (overall-CH: HR: 1.44, 95% CI: 1.16 − 1.77, p < 0.001; DNMT3A-CH; HR: 1.57, 95% CI: 1.21 − 2.04, p < 0.001; TET2-CH: HR: 1.72, 95% CI: 1.16 − 2.55, p = 0.008). CH in AF associated with higher incidence of composite HF, ischaemic stroke or death (HR:1.32; 1.20-1.45, p < 0.001).Schuermans et al., 202431Adults from UKB without prevalent arrhythmias/ haematologic malignancy410,702Median 58 years [50–63]Predominantly White/EuropeanWES-defined CH; VAF ≥ 2% (any CH) and VAF ≥ 10% (large clone CH); gene subtypes (DNMT3A, TET2, ASXL1, JAK2, PPM1D, TP53 and spliceosome)Incident arrhythmias: brady, supra/ventricular, AF, cardiac arrest; myocardial fibrosis in the imaging subsetAny CH: HR 1.09 to 1.16 across arrhythmia subtypes (including AF: HR: 1.11, 95% CI: 1.04–1.18, p = 0.001), cardiac arrest (HR: 1.29, 95% CI: 1.06–1.58, p = 0.012) and myocardial fibrosis (OR: 1.19, 95% CI: 1.01 − 1.41, p = 0.035), with increases in magnitude of all associations for large clone CHSaadatagah et al., 202432Adults form ARIC and UKB without haematologic malignancy/mitral valve disease199,982 (ARIC: 4131; UKB: 195,85)ARIC: mean 75.8 ± 5.2 years; UKB: mean 56.4 ± 8.1 years77% White (ARIC); 94% White (UKB)WES-defined CH; VAF ≥ 2% (any CH) and VAF ≥ 10% (large clone CH); gene subtypes (DNMT3A, TET2 and ASXL1)AFAny large clone CH: incident AF (HR: 1.12, 95% CI: 1.01 − 1.25, p = 0.036)TET2-CH: incident AF (HR: 1.18, 95% CI: 1.01 − 1.38, p = 0.043), with risk increasing for larger clone size (HR: 1.29, 95% CI: 1.05 − 1.59, p = 0.015)ASXL1-CH: incident AF (HR: 1.33, 95% CI: 1.02 − 1.73, p = 0.033), with risk increasing for larger clone size (HR: 1.45, 95% CI: 1.02 − 2.07, p = 0.040)Teng et al., 202533Patients with JAK2V617F–positive MPN439Median57 years [12–87]Predominantly East AsianJAK2V617F gene mutation; additional somatic mutation profiling by NGS (TET2, DNMT3A, ASXL1, FAT1 and EP300)AFTET2 mutation independently predicted AF (HR: 4.36, 95% CI: 1.05–18.06, p = 0.042); IL-1β elevation also predicted AF (HR: 5.48, 95% CI: 1.55–28.12, p = 0.012)AF atrial fibrillation, ARIC Atudy, ATVB Atherosclerosis, Thrombosis and Vascular Biology Italian Study, CAD coronary artery disease, CH clonal haematopoiesis, CI confidence interval, CVD cardiovascular disease, DCM dilated cardiomyopathy, FHS Framingham Heart Study, FUSION Finland–United States Investigation of NIDDM [Non–Insulin-Dependent Diabetes Mellitus] Genetics, HF heart failure, HFpEF heart failure with preserved ejection fraction, HFrEF heart failure with reduced ejection fraction, HR hazard ratio, JHS Jackson Heart Study, LAD left anterior descending artery, MDC Malmö Diet and Cancer, MGBB Mass General Brigham Biobank, MI myocardial infarction, NGS next-generation sequencing, OR odds ratio, PESA Progression of Early Subclinical Atherosclerosis, PROMIS Pakistan Risk of Myocardial Infarction Study, UKB UK Biobank, VAF variant allele fraction, WES whole-exome sequencing, WGS whole-genome sequencing, WHI Women’s Health Initiative. Summary of key observational studies of CH and CVD 4578 (240 CH cases and 4438 controls) Overall-CH and incident CAD (HR: 1.36, 95% CI: 1.22–1.53, p = 2.00 × 10−7, with correction for smoking) Overall-CH and incident MI (HR:1.32, 95% CI: 1.23–1.42, p = 1.25 × 10−12, with correction for smoking) 50,122 (MGBB: 12,465 [657 CH cases, 11,808 controls]; UKB: 37,657 [2,194 CH cases, 35,463 controls]) 176,219 (UKB: 130,709; deCODE: 45,510) Median 58.4 years (UKB); 53.0 years (deCODE) Overall-CH with left main stenosis (OR: 2.44, 95% CI: 1.40–4.27, p = 0.021) TET2-CH with obstructive LAD stenosis (OR: 2.62, 95% CI: 1.34–5.11, p = 0.039) Median 45 years Composite CVD outcome (CAD, MI, ischaemic stroke and all-cause mortality) and MI; prevalent (MGBB); incident (UKB) Within MGBB, associations of overall-CH, TET2-CH and ASXL1-CH with prevalent composite CVD outcome. Within UKB, associations of overall-CH, TET2-CH, DNMT3A-CH and ASXL1-CH with incident composite CVD, associations of overall-CH and ASXL1-CH with incident MI (overall-CH: HR: 1.09, 95% CI: 1.02 − 1.15, p = 7.5 × 10−3; ASXL1-CH: HR: 1.29, 95% CI: 1.10 − 1.51, p = 1.5 × 10−3) Mean 74 ± 7 years Overall-CH and incident HF (HR: 1.21, 95% CI: 1.07–1.36, p = 1.38 × 10−2) Large clone CH and incident HF (HR: 1.30, 95% CI: 1.09 − 1.55, p = 1.76 × 10−2) 176,219 (UKB 130,709; deCODE 45,510) Median 58.4 years (UKB); 53.0 years (deCODE) 8090 (JHS 2927; WHI 5163) Composite CVD outcome (CAD, MI, ischaemic stroke, and all-cause mortality) and HF; prevalent (MGBB); incident (UKB) Within MGBB, associations of overall-CH, TET2-CH and ASXL1-CH with prevalent composite CVD outcome, association of overall-CH with prevalent HF (OR: 1.08, 95% CI 1.00 − 1.16, p = 0.05) Within UKB, associations of overall-CH, TET2-CH, DNMT3A-CH and ASXL1-CH with incident composite CVD, associations of overall-CH, DNMT3A-CH and ASXL1-CH with incident HF (overall-CH: HR: 1.22, 95% CI: 1.15 − 1.30, p = 7.3 × 10−11; DNMT3A-CH; HR: 1.15, 95% CI: 1.05 − 1.24, p = 1.2 × 10−3; ASXL1-CH: HR: 1.40, 95% CI: 1.19 − 1.64, p = 4.8 × 10−5) 25,631 (Korean discovery cohort: 4345; UKB: 21,286) East Asian (Korean discovery cohort); predominantly White (UKB) Any large clone CH: incident AF (HR: 1.12, 95% CI: 1.01 − 1.25, p = 0.036) TET2-CH: incident AF (HR: 1.18, 95% CI: 1.01 − 1.38, p = 0.043), with risk increasing for larger clone size (HR: 1.29, 95% CI: 1.05 − 1.59, p = 0.015) ASXL1-CH: incident AF (HR: 1.33, 95% CI: 1.02 − 1.73, p = 0.033), with risk increasing for larger clone size (HR: 1.45, 95% CI: 1.02 − 2.07, p = 0.040) Median 57 years [12–87] AF atrial fibrillation, ARIC Atudy, ATVB Atherosclerosis, Thrombosis and Vascular Biology Italian Study, CAD coronary artery disease, CH clonal haematopoiesis, CI confidence interval, CVD cardiovascular disease, DCM dilated cardiomyopathy, FHS Framingham Heart Study, FUSION Finland–United States Investigation of NIDDM [Non–Insulin-Dependent Diabetes Mellitus] Genetics, HF heart failure, HFpEF heart failure with preserved ejection fraction, HFrEF heart failure with reduced ejection fraction, HR hazard ratio, JHS Jackson Heart Study, LAD left anterior descending artery, MDC Malmö Diet and Cancer, MGBB Mass General Brigham Biobank, MI myocardial infarction, NGS next-generation sequencing, OR odds ratio, PESA Progression of Early Subclinical Atherosclerosis, PROMIS Pakistan Risk of Myocardial Infarction Study, UKB UK Biobank, VAF variant allele fraction, WES whole-exome sequencing, WGS whole-genome sequencing, WHI Women’s Health Initiative. Atherosclerotic disease CH has been repeatedly linked with increased risk of atherosclerotic CVDs, independent of traditional risk factors (such as smoking, obesity and hyperlipidaemia). In a seminal work, Jaiswal et al.18 evaluated the relationship of CH with atherosclerotic disease across multiple cohorts, presenting a prospective and a complementary retrospective analysis, which provided insights into mutation-specific risks. Firstly, using two retrospective case-control cohorts capturing 7,245 participants, they demonstrated four times greater odds of early-onset myocardial infarction (MI) in those with CH compared to non-carriers (OR: 4.0, 95% CI: 2.4–6.7, p < 0.001). Second, in a nested prospective study using data from two cohorts including 1010 patients, those with CH had a nearly 2-fold greater risk of incident MI or coronary revascularisation (HR: 1.9, 95% CI: 1.4–2.7, p < 0.001). Heimlich et al.21 investigated relationships of CH with the extent and pattern of coronary atherosclerosis in 1142 patients undergoing clinically indicated coronary angiography. Individuals with CH had a similar risk of having any coronary stenosis or obstructive stenoses overall compared with controls, but they were more likely to have prognostically adverse patterns of disease, with higher odds of obstructive disease in the left main stem (OR: 2.44, 95% CI: 1.40–4.27, p = 0.021). The cross-sectional nature of this analysis precludes temporal separation of CH onset from development of atherosclerotic disease, limiting causal inference, in particular due to reverse causation. While the researchers adjust for major demographic and vascular risk factors, residual confounding cannot be fully excluded. Furthermore, the mechanism through which CH would specifically target specific parts of the coronary anatomy is unclear. Other researchers have explored the idea of reverse causation in the associations between CH and atherosclerotic CVD. Heyde et al.49 demonstrated that atherosclerosis increases proliferation rates of both human and mouse haematopoietic stem cells and proposed that the increased CH of varying mutation subtypes associated with atherosclerotic disease could, instead, be due to atherosclerosis uniformly expanding clones across driver genes to detectable levels, thereby increasing CH detection. Indeed, experimental atherogenesis accelerated the expansion of Tet2−/− cells, although whether this occurs uniformly across other gene mutations or with heterozygotic mutations found much more commonly in CH was not studied. Carter et al.50 performed similar experiments, and although experimental atherogenesis also promoted faster expansion of Tet2−/− cells, atherogenesis had no impact on clonal expansion of heterozygotic Tet2 mutant cells, or in the context of other driver genes (Dnmt3a or Jak2). This questions the relevance of an expansion of homozygotic mutations and further demonstrates that any effect is unlikely to be uniform across all genes. This is supported by an analysis of 4184 otherwise healthy middle-aged individuals in the Progression of Early Subclinical Atherosclerosis (PESA) study22, which demonstrated a unidirectional association of CH mutations with de-novo femoral atherosclerosis over 6 years of prospective follow-up. In addition, Zuriaga et al.23 demonstrate associations of overall-CH with incident coronary artery disease (HR: 1.07, 95% CI: 1.02–1.13, p = 5.8 × 10−3) and incident MI (HR: 1.09, 95% CI: 1.02–1.15, p = 7.5 × 10−3) in an analysis of 437,236 UK Biobank (UKB) participants. Jaiswal et al.18 conducted a meta-analysis of five prospective studies, demonstrating heightened but differential risk of coronary artery disease across common driver mutations. DNMT3A, TET2 and ASXL1 mutations were linked to between 1.7- and 2.0-times greater risk, whilst JAK2V617F mutation was associated with 12.0 times greater risk. Given that DNMT3A is the most common CH driver mutation at a population level, this analysis likely represents true differences in risk rather than variations in statistical power, unlike the small cohort for JAK2-CH. However, Heimlich et al.21 report no association of DNMT3A-CH with the measured angiographic outcomes (any stenosis, obstructive stenosis, anatomically significant stenosis). Consistent with these observations, other analyses in large biobanks6,9,20 and using Mendelian Randomisation6 (MR) report no or weak associations of DNMT3A-CH with atherosclerotic outcomes. This is also somewhat corroborated in experimental studies by Rauch et al.51 and Carter et al.50, who found that bone marrow chimeric models of Dnmt3a+/− CH did not develop augmented atherosclerosis. Rauch et al.51 found that homozygotic perturbations in Dnmt3a−/− mice did accelerate atherosclerosis and posited some mechanistic insight as to why an association of DNMT3A-CH with CAD may exist51. An augmented pro-inflammatory cytokine profile (increased IL-1β, IL-6 and Cxcl chemokines) in mutant Dnmt3a−/− cells was observed, which is convergent to Tet2−/− clones and the authors propose similar processes drive coronary atherosclerosis in these two common mutations. We therefore highlight that despite some limited epidemiologic and experimental evidence of DNMT3A-CH augmenting atherosclerosis, this is largely unsupported by analysis of adequately powered human cohorts. In comparison to DNMT3A mutations, TET2 mutations have demonstrated stronger associations with atherosclerotic coronary artery disease. Mutation-specific analysis conducted by Jaiswal et al.18 demonstrated an 8-fold higher odds of early-onset MI (OR: 8.3, 95% CI: 1.2–352, p = 0.020) and a non-significantly increased risk of incident coronary artery disease (HR: 1.9, 95% CI: 1.0–3.7, p = 0.060). Both analyses are severely underpowered, including a total of 13 and 29 participants from pooled studies, respectively, with marked imprecision of the estimate for early-onset MI. A high-risk pattern of coronary artery atherosclerosis was reported by Heimlich et al.21 amongst a small cohort of 42 individuals with TET2-CH who exhibited a significantly increased risk of obstructive left anterior descending artery stenosis (OR: 2.62, 95% CI: 1.34–5.11, p = 0.039), whereas no associations were found for DNMT3A-CH. Carter et al.50 found augmented atherosclerosis in Tet2 experimental models to be specific to female mice. Thus far, sex-dependent effects of CH on CVD in humans have been understudied, but this raises the possibility that CH underlies some of the late-life rise in CVD in women, which warrants further investigation. Zuriaga et al.23 examined associations of DNMT3A-, TET2- and ASXL1-CH with prevalent and incident cardiovascular outcomes in the Mass General Brigham Biobank (MGBB, n = 37,181) and UKB (n = 437,236), respectively. Within cross-sectional analyses in the MGBB subset, TET2-CH (OR: 1.14, 95% CI: 1.00–1.31, p = 0.050) and ASXL1-CH (OR: 1.51, 95% CI: 1.23–1.86, p = 8.5 × 10−5) were significantly associated with a composite of any CVD outcome, while relationships with DNMT3A-CH were non-significant. Associations with prevalent disease-specific associations (coronary artery disease, MI, stroke and HF) were non-significant for all mutations. In prospective associations within UKB over more than 14 years of follow-up, all three mutations were significantly linked to a higher incidence of any cardiovascular events. TET2-CH showed no significant associations with any of the disease-specific incident cardiovascular outcomes, while ASXL1-CH was significantly linked to incident coronary artery disease and MI, and DNMT3A-CH was linked to a higher risk of incident HF. Augmented atherosclerosis by TET2-CH is supported by experimental work by Fuster et al.52 who examined the relationship of Tet2 mutations in murine models with atherosclerosis and identified that Tet2 mutations accelerate atherosclerosis in atherosclerosis-prone (Ldlr−/−) models. They further examined the mechanisms underlying this association and found evidence of increased NLRP3-mediated IL-1β secretion, thus indicating inflammatory cytokine overproduction may be responsible; furthermore, therapeutically intervening with NLRP3 inhibition provided a greater atheroprotective effect with Tet2-mutant Ldlr−/− mice than non-mutants. Such interventions will need to be tailored to mutational subtypes with augmented NLRP3 function; Carter et al.50 found that whilst Tet2 mutant macrophages displayed increased NLRP3 activity, the opposite was true for Dnmt3a mutations that had lower activity, whilst Jak2 was unaffected. Thus, targeting NLRP3 may have a specific benefit in TET2-CH. Consistent with the previously mentioned observations by Zuriaga et al.23, ASXL1 mutations also conferred a higher risk of coronary artery disease in the meta-analysis by Jaiswal et al.18 (HR: 2.0, 95% CI: 1.0–3.9, p = 0.050). Sato et al.53 examined the association of Asxl1 mutant macrophages in atherosclerosis-prone (Ldlr−/−) mice, implicating a disinhibition of the innate immune system and NF-κB activation (leading to NLRP3 inflammasome activation), thus contributing to similar inflammatory cytokine overproduction as seen with TET2 mutations. Myeloproliferative neoplasms driven by JAK2 mutations increase cardiovascular risk through a number of mechanisms, including heightened risk of both arterial and venous thromboses54–56. It is therefore unsurprising that Jaiswal et al.18 observed a very large association of JAK2-CH with coronary artery disease (HR: 12.0, 95% CI: 3.8–38.4, p < 0.001). However, given the small sample size in this study, the association requires further examination. Low-level JAK2 mutations with a prevalence of 3.1% have been observed within a Danish study57 of 19,958 individuals, which could suggest JAK2 is a population-level risk factor for CVD, including atherosclerosis. Fidler et al.58 demonstrated increased atherosclerotic plaque size in Jak2V617F murine models, and although they also implicated a pro-inflammatory phenotype including increased IL-1β production, mechanistically this diverged from Tet2 murine models with the AIM2 inflammasome (rather than NLRP3) driving oxidative stress and DNA damage with large necrotic cores. Furthermore, work conducted by Wang et al.59 which transplanted 100% Jak2 mutant cells into murine models, better modelling myeloid neoplasia than CH per se, identified increased neutrophil content in early atherosclerotic lesions, and that Jak2V617F erythrocytes were more susceptible to erythrophagocytosis, which drove progression of late plaques. This highlights that JAK2-CH differs in the pathophysiology of atherosclerosis, both compared to other driver genes, but also temporally, necessitating further investigation. TP53 mutations are believed to contribute to atheromatous disease through augmenting cellular proliferation rather than the inflammatory mechanisms previously discussed. Zekavat et al.19 found TP53-CH to be associated with incident atherosclerotic disease across a variety of vascular beds in UKB and MGBB, with an over 2-fold increase in coronary artery disease (HR: 2.31, 95% CI: 1.03–5.16, p = 0.042) and a 5-fold increase in peripheral artery disease (HR: 4.98, 95% CI: 1.23–20.09, p = 0.024). Through investigation of Trp53−/− murine models, they noted an increase in p53-deficient macrophage proliferation in aortic atherosclerotic lesions, which highlights the heterogeneity of mechanisms leading to atheromatous disease driven by disparate CH mutations. We highlight that biallelic knock-out of p53 may represent a rare and more extreme phenotype than is observed in human TP53-CH, although the authors report that this recapitulates the dominant negative effect of TP53 mutations in CH. A phenome-wide association study (PheWAS) in the UKB conducted by Lin et al.60 revealed that autosomal mCA, mLOX and mLOY were associated with acute MI (autosomal mCA: OR: 1.38, 95% CI: 1.26–1.51, p = 7.69 × 10−12; mLOX: OR: 1.75, 95% CI: 1.49–2.06, p = 1.64 × 10−11; mLOY: OR: 1.79, 95% CI: 1.71–1.88, p = 2.78 × 10−125). Autosomal mCA and mLOY were additionally associated with atherosclerosis (autosomal mCA: OR: 1.98, 95% CI: 1.43–2.73, p = 3.15 × 10−5; mLOY: OR: 2.37, 95% CI: 1.94–2.89, p = 1.36 × 10−17), and mLOX with chronic ischaemic heart disease (OR: 1.64, 95% CI: 1.45–1.87, p = 2.91 × 10−14). This further supported prior work by Haitjema et al.61 who examined associations of mLOY with a cohort of 610 men who underwent carotid endarterectomy from the Athero-Express Biobank; mLOY was associated secondary major cardiovascular events (HR: 2.28, 95% CI: 1.11–4.67, p = 0.02), but was not associated with atheroma size (OR: 2.15, 95% CI: 1.06–4.76, p = 0.04) after Bonferroni correction. In conclusion, various driver genes have been associated with atherosclerosis, and overlapping mechanisms have been implicated, which may form a mechanistic framework. Broadly, they can be classified into: (1) inflammatory cytokine overproduction (TET2-, JAK2-, ASXL1-, with weaker evidence for DNMT3A-CH), (2) impaired inflammatory resolution (JAK2- and TET2-CH) and (3) extensive macrophage proliferation (JAK2- and TP53-CH). These classifications are not mutually exclusive, with significant crossover between mechanisms. The associations of some CH mutational subtypes and atherosclerosis are, however, based on small numbers of participants, limiting the precision and statistical power to estimate true effects, particularly for mutations with low population prevalence. Further longitudinal studies across different cohorts are warranted to validate existing findings and provide greater confidence in causal associations of CH with atherosclerotic outcomes. Heart failure Recent findings support the role of CH as a key driver and prognostic indicator for HF of various aetiologies. In an early study, Dorsheimer et al.24 report the prevalence of CH in 18.5% of 200 patients with chronic ischaemic HF, which is higher than that reported in unselected cohorts17,62,63. Furthermore, they found that DNMT3A- and TET2-CH independently doubled the risk of death or HF hospitalisation over a median prospective follow-up period of 4.4 years, highlighting CH as a marker of prognosis in this context. This risk was also present for those with small clone fractions, but those with larger clone sizes had an incrementally poorer prognosis. Sano et al.64 investigated the role of DNMT3A and TET2 mutations in an Ang II murine model of HF, revealing associations with greater macrophage and T-cell infiltration of the myocardium and adverse cardiac remodelling (via inflammation). These initial works provided the foundation for further investigation of the role of CH within HF, whilst also indicating inflammation as a key mediator of associations with DNMT3A- and TET2-CH. Subsequent studies have revealed inconsistencies in the associations of CH with HF, highlighting potential variations in the relationship by driver mutation and HF subtype (HF with reduced or preserved ejection fraction – HFrEF and HFpEF). Zuriaga et al.23 report no significant associations between prevalent HF and overall-CH or CH related to three common driver mutations (DNMT3A, TET2 or ASXL1) in 37,181 participants from the MGBB. Conversely in their prospective analysis of 437,236 UKB participants, they report significant associations with incident HF and overall-CH (HR: 1.22, 95% CI: 1.15–1.30, p = 7.3 × 10−11), DNMT3A-CH (HR: 1.15, 95% CI: 1.05–1.24, p = 1.2 × 10−3), and ASXL1-CH (HR: 1.40, 95% CI: 1.19–1.64, p = 4.8 × 10−5). In a smaller prospective analysis of 8090 individuals from The Jackson Heart Study and Women’s Health Initiative, Scheurmans et al.28 did not find a significant association between DNMT3A-CH (or overall-CH) and incident HF (either HFpEF or HFrEF). Divergence was found for TET2-CH, which was associated with elevated risk of HFpEF (HR: 2.35, 95% CI: 1.34–4.11, p = 0.003), but not HFrEF. The concurrent increase in CRP levels found in TET2-CH raised the possibility of an inflammation-driven process, with the intersection between inflammation and metabolism increasingly recognised to be central to HFpEF pathogenesis. However, the lack of association with HFrEF raises the question of whether HF associations are driven by HFpEF or whether the findings by Scheurmans et al.28 were due to differences in the size and composition of the underlying populations studied, including a more racially diverse and female-predominant cohort. These observations may also be attributed to the biologically opposing functions of DNMT3A and TET2, with roles in gene repression and activation, respectively, and some evidence for opposing actions on inflammasome activity50. However, previous studies51 have identified a paradoxical concordance of macrophage phenotypes, which has yet to be reconciled within the context of HFpEF. Yu et al.27 also found no significant association with DNMT3A-CH and incident HF, although this contrasts with evidence from Pascual-Figal et al.25 who found significantly faster HFrEF progression with a greater risk of overall mortality, HF death or hospitalisation. Furthermore, a recent MR study65 suggested causal associations of DNMT3A-CH with adverse remodelling patterns of larger left and right ventricular end-diastolic volumes. The associations of TET2-CH and HF have been found relatively consistently throughout the literature, with increased incidence (of HFpEF)28, faster progression24, supportive pre-clinical models66 and an implication of NLRP3 activation and IL-1β production leading to adverse cardiac remodelling. Similarly, less common driver mutations, including ASXL1 and JAK227,67 have been associated with a higher incidence of HF, with an effect graded to clonal size. Such associations were significant with or without antecedent coronary artery disease, indicating the effect of CH is not solely through driving ischaemic disease – lending supportive evidence for alternative mechanisms (such as inflammation), which perhaps also explains greater consistency in reported association with HFpEF subtypes. Finally, despite associations of CH with HF being attenuated by smaller clone size, Sikking et al.29 identified a 21% prevalence of CH in a study of 520 dilated cardiomyopathy patients, with clones below 2% VAF having prognostic significance of all-cause mortality (HR: 1.72, 95% CI: 1.10–2.69, p = 0.017), with a VAF threshold of 0.06%. This finding further strengthens the impact of CH in non-ischaemic HF and highlights that very small clones are not benign and can demonstrably influence morbidity and mortality. The UKB PheWAS conducted by Lin et al.60 revealed a positive association between mLOY and HF (OR: 1.67, 95% CI: 1.51–1.85, p = 1.14 × 10−22). These findings are supported by a further prospective UKB study68, which demonstrated increased risk of HF death among individuals with mLOY (HR: 1.76, 95% CI: 1.01–3.05, p = 0.045). The mechanisms underlying such an association were explored by Sano et al.68 using murine models lacking chromosome Y, which demonstrated a fibrotic phenotype and decreased cardiac output, with effects mitigated by transforming growth factor β1–neutralising antibody administration. Most current studies are limited by sample size, cross-sectional study designs, and a lack of detailed phenotyping of CH and HF required to make reliable and precise conclusions. Future studies should strive to provide more granular definitions of HF, considering underlying inherited and acquired aetiologies, as well as to distinguish preserved and reduced ejection fraction phenotypes. Atrial fibrillation Atrial fibrillation (AF) is the most common sustained arrhythmia, associated with advancing age, chronic inflammation, obesity and atrial fibrosis. CH has been implicated to contribute to pro-inflammatory and fibrotic changes within the myocardium; therefore, it is unsurprising that associations have been described between CH and AF, albeit in a mutation and clone size-dependent manner32. The largest prospective study to date, conducted by Saadatagah et al.32, assessed incident AF risk amongst 199,982 participants across the Atherosclerosis Risk in Communities (ARIC) study (median follow-up 7.0 years) and UKB (median follow-up 12.2 years). Incident AF risk was increased with any TET2-CH (HR: 1.18, 95% CI: 1.01–1.38, p = 0.043) and large clone TET2-CH (HR: 1.29, 95% CI: 1.05–1.59, p = 0.015); the importance of small clone CH was not specifically assessed. These findings are corroborated by a retrospective case-control study within an East Asian cohort from Ahn et al.30 who reported a 65% increased odds of AF with TET2-CH (OR: 1.65, 95% CI: 1.05–2.60, p = 0.030); TET2-CH was also associated with atrial enlargement and increased ventricular filling pressures – remodelling patterns predisposing to AF. Notably, though, atrial size was not significantly different in CH across a range of driver genes, including TET2-CH in the longitudinal study by Saadatagah et al.32. Given that atrial dilatation and remodelling are key factors in the pathogenesis of AF, further studies are required to elucidate the effect of CH on atrial size and function. Experimental investigations with Tet2−/− mice conducted by Lin et al.69 examined the mechanisms underlying such associations of TET2-CH with AF, which implicated NLRP3-inflammasome activation and calcium dysregulation; attenuation of AF risk was observed with oral administration of an NLRP3 inhibitor. This supported observational findings of increased IL-6 with large clone TET2-CH (which is known to be upregulated by IL-1β produced by the NLRP3-inflammasome70) in Saadatagah et al.’s32 study and provides a convincing explanation of a pro-inflammatory state mediating the association of TET2-CH with AF. ASXL1-CH also demonstrates an association with increased AF risk; in an observational study conducted by Schuermans et al.31 of 410,702 UKB participants, any size of ASXL1-CH associated with a 1.23-fold increase of AF (HR: 1.23, 95% CI 1.05–1.45, p = 0.012), albeit solely in the minimally adjusted model. However, this finding was replicated in Saadatagah et al.’s32 analysis, which reported increased risk of AF with any ASXL1-CH (HR: 1.33, 95% CI: 1.02–1.73, p = 0.033). Individuals with large clone ASXL1-CH demonstrate larger indexed left ventricular mass (B-coefficient per 1-SD increase in log-transformed mass: 0.26, 95% CI: 0.12–0.40, p = 3.5 × 10−4) and greater high-sensitivity troponin T (B-coefficient: 0.21, 95% CI: 0.10–0.32, p = 1.8 × 10−4), indicative of both hypertrophy and cardiomyocyte injury. Although these changes in cardiac structure and function provide the environment for AF generation, atrial size (which is more directly linked to AF) was not affected. Further mechanistic studies are required to investigate the role of ASXL1-CH in the development of AF, as presently there exists no murine models or genetic inferential studies examining this. Given JAK2 mutations are relatively rare within the CH population62, there has been insufficient evidence at the population level to reliably establish increased risk of AF; however, in a study conducted by Teng et al.33 in a cohort of 439 individuals with myeloproliferative neoplasms (including those with polycythaemia vera and essential thrombocytopaenia), who carry JAK2V617F missense variants, 6.6% of patients developed AF over a median follow up of 6 years. We therefore highlight this as an area for future research, given that extrapolation of results from alternative populations cannot be reliably used to infer risk within the CH population, despite having a biologically plausible association. Despite a lack of association of DNMT3A-CH with AF risk in studies conducted by Saadatagah et al.32 and Schuermans et al.31, there is evidence of an association both observationally and with genetic inferential studies. Ahn et al.30 reported a 45% increase in odds of AF with DNMT3A-CH (OR: 1.45, 95% CI: 1.09–1.93, p = 0.012), which is further corroborated by an MR study by Kar et al.6 (OR: 1.05, 95% CI: 1.02–1.07, p = 1.99 × 10−4), providing supportive evidence of causality. The mechanism for such an association is comparatively less well understood than the association observed with TET2 mutations, however a further MR study has implicated larger right atrial size65 which may contribute to AF generation; in addition, Rauch et al.51 describe concordant macrophage phenotypes with loss of function mutations in Tet2−/– and Dnmt3a−/− mice, which suggests DNMT3A-CH and AF associations may also be driven by a pro-inflammatory phenotype, highlighting the need for future investigation in this area. Associations between autosomal mCA, mLOX and mLOY with AF and flutter have been described within the UKB PheWAS conducted by Lin et al.60 (autosomal mCA: OR: 1.31, 95% CI: 1.17 − 1.46, p = 1.56 × 10−6; mLOX: OR: 1.66, 95% CI: 1.41−1.94, p = 3.70 × 10−10; mLOY: OR: 1.61, 95% CI: 1.51−1.71, p = 1.34 × 10−51). In a recent sensitivity analysis conducted by Lim et al.71, among 415,466 UKB participants, similar associations of mLOY with AF and flutter was observed among never smokers (HR: 1.07, 95% CI: 1.01–1.14, p = 0.029) and ever smokers (HR: 1.05, 95% CI: 1.01–1.11, p = 0.030) as well as men aged >60 and ≤60 years, however these were non-significant after Bonferroni correction. MR analysis in the same study provided further highly significant evidence of a causal relationship between mLOY and AF (HR: 1.15, 95% CI: 1.13–1.18, p = 1.52 × 10−21). Cardio-oncology Cancer survivors have a heightened risk of CVD compared to the general population. A substantial proportion of this elevated cardiovascular risk has been attributed to the effects of cancer therapies on the cardiovascular system, high burden of shared risk factors and biological processes related to the cancer itself, such as systemic inflammation and prothrombotic states. Many of these factors also increase the likelihood of acquiring CH-associated driver mutations, including advancing age, tobacco smoking, chronic inflammation and the direct mutagenic effects of chemoradiation. In recent years, CH has become an area of increasing interest in cardio-oncology, as accumulating data suggest that cancer patients with CH mutations face higher risks of CVD and all-cause mortality, especially those carrying larger clones. This has prompted calls for evaluating CH as a potential biomarker to enhance cardiovascular risk stratification in cancer survivorship care72,73. Recent studies indicate that patients with solid tumours display a higher prevalence of CH than the general population, with variations by cancer types73–75. For instance, in a large sequencing study of over 12,000 adults with non-haematological cancers, 23.1% of individuals over 80 years carried CH-related mutations detectable in cell-free DNA74. The most frequently mutated genes included DNMT3A, TP53, TET2 and PPM1D, with gene patterns varying by tumour type; for example, TP53 was the most frequent CH gene in melanoma and small-cell lung cancer, while DNMT3A mutations were most common in uterine and cervical cancers74. A separate analysis similarly reported that 25.1% of patients with non-haematological malignancies had at least one clonal somatic mutation75. Cancer survivors commonly develop therapy-associated CH, a distinct form of clonal expansion characterised by mutations in DNA damage response (DDR) genes such as ATM, CHEK1, CHEK2, TP53 and PPM1D that confer a selective growth advantage2–4. This form of CH is thought to be driven by the genotoxic stress induced by cytotoxic chemotherapy and radiotherapy, both of which promote the acquisition and expansion of somatic mutations. Anticancer therapies such as platinum agents, anthracyclines and thoracic radiation can promote the expansion of haematopoietic clones carrying driver mutations. In a large study of 24,146 cancer patients, Bolton et al.2 reported that external beam radiotherapy and regimens containing platinum and topoisomerase II inhibitors were most strongly correlated with CH in specific DDR pathway genes, including TP53, PPM1D and CHEK2. Consistent findings have been observed in other settings; for example, among patients with non-Hodgkin lymphoma undergoing autologous stem-cell transplantation and receiving anthracycline-based treatment, nearly 30% were found to have CH, with PPM1D and TP53 representing the dominant mutations76. Emerging evidence suggests that CH may also contribute to cardiotoxicity risk in cancer patients. In a small prospective study of 88 solid cancer patients treated with immune checkpoint inhibitors, CH more than doubled the risk of myocarditis (HR 2.74, 95% CI: 1.44–5.22, p = 0.002) and was a greater prognostic factor than dual therapy77. In a cohort of patients with breast cancer, sarcoma and multiple myeloma, CH was an independent predictor of doxorubicin-induced cardiotoxicity, even after accounting for conventional risk factors78. Supporting data from lymphoma cohorts show a similar pattern: TET2-mutant clones were independently associated with anthracycline-related cardiotoxicity79. Consistent findings have been reported in breast cancer patients treated with HER2-targeted therapies and chest radiotherapy, among whom CH, most commonly involving DNMT3A and TET2 mutations, was detected in 38% of individuals and associated with an independent risk of therapy-related cardiomyopathy72. However, it remains uncertain whether these associations are driven by a selective expansion of clones carrying driver mutations following exposure to cytotoxic therapy80,81. An important next step in the field is understanding how genotoxic treatments shape clonal dynamics and influence cardiovascular outcomes. Both CVD and CH are common in cancer patients, and CH is increasingly recognised as a contributor to cardiovascular vulnerability in this population. With growing evidence linking CH to adverse cardiac outcomes, it is emerging as a potentially useful biomarker for identifying cancer patients at heightened cardiovascular risk. Yet important uncertainties remain, including the mutation-specific and clone-size-dependent risks, the timing of clonal expansion and the degree to which therapy-associated CH differs biologically from age-related CH. Addressing these questions will be essential for defining whether and how CH should be incorporated into personalised cardio-oncology risk stratification and survivorship care. Research gaps and directions for future work Existing literature evaluating the cardiovascular associations of CH has several key biases – reverse causation, confounding, survival bias and insufficient statistical power. We will discuss these issues and potential mitigations in the remainder of this section. Reverse causation Studies reporting prevalence and cross-sectional associations of CH in cohorts selected on those illnesses (e.g. HF) are subject to reverse causation – it is not possible to distinguish the temporal relationship of CH with the CVD of interest (or its associated risk factors), which may in itself drive development of CH. Longitudinal analyses are essential to determine causal associations of CH with CVD. Residual confounding factors There are many shared exposures that drive both CVD and CH, and insufficient adjustment for these variables will introduce confounding and may falsely augment observed relationships. In multivariate regression models with CVD as the outcome and CH as the main predictor, the usual confounding factors that are adjusted for are age, sex and smoking; however, some analyses have indicated that residual confounding by smoking may influence current estimates20. Furthermore, CH is associated with a wide variety of ageing-related traits and disorders, including CVD. It is possible that chronological age may not be representative of biological age or organ-specific age. Indeed, organ-specific ageing has been quantified using plasma protein signatures82, and it would be pertinent to assess biological age46 and organ-specific age as a confounding factor in the observed relationship between CH and CVD. Survival bias The availability of large and deeply phenotyped longitudinal cohorts of unselected populations (e.g. UKB) allows mitigation of the aforementioned biases. However, these cohorts may be hampered by survival bias—in that individuals with the most unfavourable forms of CH may not survive to enter such cohorts, particularly those with recruitment in middle or older age (such as UKB). Statistical power attributed to rare mutations, small clones and phenotype definition While research biobanks can provide large sample sizes, statistical power remains insufficient for rarer mutations. Furthermore, older technologies used at the time for sample analysis in these cohorts may miss small clones, which are increasingly recognised as of potentially prognostic importance. For example, Mutect2 is the main bioinformatics tool for CH detection, but current recommendations for this approach are the inclusion of mutations with ≥2% VAF, and for selected indels, such as ASXL1 G646Wfs*12, ≥10% VAF12. In particular, JAK2, SF3B1 and U2AF1 hotspot mutations have lower coverage (median sequencing coverage of ≤31 reads) which limits the ability to detect clones with ≥2% VAF. ‘pileup’ is a more sensitive approach for detecting low-VAF mutations, particularly those supported by just 1–2 sequencing reads83, and it is able to identify mutations with ≥2% VAF missed by Mutect284. Phenotypic definitions amongst epidemiological studies, including genetic inferential studies, have been a source of controversy, particularly with regard to differences in the methodological techniques applied to ascertain CH in large-scale datasets. For example, a variety of approaches have been applied to classify CH within large datasets (such as whole genome sequencing6, whole exome sequencing12, targeted exome sequencing9 and mutational barcode-defined20 methods), yet there still remains a notable lack of an international consensus agreement on the most appropriate method of ‘calling’ CH, which has been identified to result in disease association underestimation48. We further argue that this may, at least in part, underlie the sometimes-conflicting genetic associations of CH with CVD outcomes. In addition, cardiovascular conditions are heterogeneous in terms of their aetiology and natural history; however, across most published reports, there is limited granularity of CVD definitions, which restricts inferences about biological mechanisms linking CH with disease-specific cardiovascular conditions. Establishing causality using Mendelian randomisation Genetic inferential studies using MR can be used to reduce reverse causation and confounding biases and have a role for triangulation of evidence alongside other study types – however, methodological scrutiny of the assumptions of such analyses85,86 is warranted when drawing conclusions. In this setting, genetic variants associated with CH can be used as instrumental variables to generate causal estimates on downstream outcomes (e.g. CVDs). Validity of such genetic causal inference studies is highly sensitive to the reliability of the instrumental variables, particularly in the exposure genome-wide association studies (GWAS). It is also noteworthy that it is only feasible to assess overall-CH, DNMT3A-CH and mLOY as exposures in traditional MR (such as Inverse Variance Weighted MR). For these classes of CH, more than 20 genetic risk loci have been identified, and therefore, there is a sufficient number of genetic instruments to assess these CH classes as exposures in traditional MR. As for other classes of CH (TET2-CH, ASXL1-CH, TP53-CH, JAK2-CH, autosomal mCA and mLOX), less than 10 genetic risk loci have been identified, and therefore it is not feasible to assess these classes of CH as exposures using traditional MR techniques6,9,10,41,65. Finally, appropriate care must be taken to mitigate the effects of horizontal pleiotropy, that is, the instrumental variables affect outcomes (in this case, CVD) through pathways which are not via the exposure being examined (in this case, CH). For example, SH2B3 mutations are known to promote JAK2-CH, but also in their own right have associations with CVD, hence, without appropriate adjustment, MR analyses may overcall associations of JAK2-CH with CVD87. Elucidating the mechanism of CH contributing to CVD development Experimental models have been instrumental in furthering our understanding of pathophysiology in CH. Most studies have been based on bone marrow chimeric models involving lethal irradiation to empty the bone marrow; this then allows transplantation of a mix of wild-type and mutant bone marrow cells into the irradiated mouse in the proportion required to mimic the clonality of the condition. Irradiation is a destructive process which causes profound inflammation and can alter the physiology of the heart and great vessels. Clonal dynamics and expansion in the bone marrow are likely to be extremely different when competing within an empty bone marrow. Some studies transplant 100% mutant cells, which do not mimic clonality in CH59; as such, the action of mutant cells may be supra-pathophysiological or may miss the cell non-autologous effects or interactions with WT cells within the haematopoietic compartment or their progeny. Other studies focus on homozygous mutations, which rarely occur in CH (notably humans homozygous for DNMT3A mutations are virtually non-existent), and it is unclear how this gene-dosage effect relates to the mechanisms and consequences of heterozygotic mutations in humans. Association between novel classes of CH with CVD An emerging class of CH without any identifiable driver gene mutations or chromosomal abnormalities has recently been described20,47,88. The association between this class of ‘driverless’ CH and CVD is not well understood. In a recent UKB study, individuals with CH based on the detection of passenger mutations demonstrated no association with CVD20. Similarly, a separate and recent study found no association between passenger mutation burden or clonal expansion with ischaemic stroke, coronary artery disease, or MI89. Nevertheless, these individuals with ‘barcoded’ CH encompass driver gene-mutant CH, mCAs and ‘driverless’ CH. Future work distinguishing ‘driverless’ CH as a distinct entity may help clarify our understanding of the traits and diseases, including CVD, associated with this class of CH. Furthermore, current research has focused upon myeloid CH; however, CH driven by mutations in genes associated with lymphoid malignancies has been found to cause immune dysregulation90 and to be associated with autoimmune disease, yet associations with CVDs of lymphoid CH are currently understudied. Concluding remarks Moving forward, studies should focus on large, unselected populations with longitudinally detected disease-specific cardiovascular outcomes. Multicohort analyses are needed to allow sufficient statistical power to study rarer CH driver mutations. Modern high-resolution sequencing technologies may provide new insights into the clinical significance of small clones. While there is a role for MR analyses, these should be interpreted in conjunction with other study types and scrutinised with regard to relevant biases and instrument strengths. A broader consensus on the definition of CH would support universal cross-comparability of both observational and genetic studies. Pre-clinical models remain critical to advancing biological understanding of the cardiovascular consequences of CH and to identifying potential druggable targets. Clinical implications At present, there are no widespread CH population screening programmes or any guidelines that advocate for its testing. Despite this, there do exist clinics which demonstrate the feasibility of screening and managing the sequelae of a CH diagnosis91. These clinics could provide a cohort of patients who could directly contribute to further research efforts by including them in biorepositories92; this would be particularly useful given that genomic sequencing is not commonplace in most healthcare systems and therefore identification of individuals with CH for research is limited. With growing recognition of the consequences of CH, as well as clinical trials which will guide how to treat it on the horizon, this raises the potential of CH screening programmes. Treatment: preventing sequelae or halting progression In 1968, Wilson and Jungner proposed a set of criteria for assessment of the suitability of screening for any given condition, incorporating factors related to the condition being screened, the screening test and economic implications93, as summarised in Fig. 2. For screening to be justifiable, the availability of efficacious treatments with clear guidance on whom to treat is integral. As there are currently no treatments proven by clinical trials to halt CH or its consequences, this is an important knowledge gap. However, mounting pre-clinical evidence implicates augmented inflammation, and particularly the NLRP3 inflammasome, in the pathogenesis of CH-driven CVD, which may pave the way for therapeutic strategies. In animal models of TET2-CH, the worsening of atherosclerosis52, HF66 and AF69 were all halted by inhibition of the NLRP3 inflammasome, which produces IL-1β. Furthermore, in a post-hoc analysis of the CANTOS94 trial, patients with TET2-CH derived greater benefit from neutralisation of IL-1β with canakinumab, supporting the notion of this therapeutic approach. This has motivated the IMPACT95 trial, which is recruiting patients with clonal cytopenia of unknown significance to assess the effect of canakinumab on cardiovascular outcomes alongside haematological outcomes such as progression to myeloid cancer, mutational burden and blood counts (NCT05641831). Although promising, the cost and fatal infections associated with canakinumab may be a barrier to its implementation, and other methods of targeting inflammation may be necessary. Colchicine is thought to reduce cardiovascular events by similar mechanisms, and notably, colchicine reduced atherogenesis in animal models of TET2-CH and colchicine users in the UKB and MGBB had lower rates of MI23. Additionally, the effect of alternative NLRP3 inhibitors, including selnoflast (ISRCTN10520571) and DFV890/MAS825 (NCT06097663) are being studied in patients with CH and coronary artery disease. Lastly, IL-1β drives IL-6 production, with the latter thought to be causal in CVD. IL-6 inhibition in animal models96, and genetic perturbations in humans97,98 both reduce atherosclerotic disease. As there are IL-6 inhibitors, such as tocilizumab, already in clinical use, this is another promising option for future clinical trials. An overview of ongoing clinical trials is shown in Table 3.Fig. 2CH screening: considerations and current status within the Wilson & Jungner framework.CH clonal haematopoiesis, CV cardiovascular. Figure was created in BioRender. Morley, A. (2026) https://biorender.com/3gi3e5fTable 3Ongoing clinical trials in CH / CCUS.TrialIntervention typeDrug/therapyStatusDatesClinical trial codeCHIP-HF: colchicine in heart failure patients with CHIPSmall moleculeColchicine (microtubule inhibitor)OngoingAug 2021–(ongoing)EudraCT 2021-001508-13A study of enasidenib in people with clonal cytopenia of undetermined significanceSmall moleculeEnasidenib (IDH2 inhibitor)Active, not recruitingOct 2021–Oct 2026 (est)NCT05102370Repurposing metformin as a leukaemia-preventive drug in CCUS and LR-MDSSmall moleculeMetformin (biguanide)RecruitingDec 2021–Feb 2026 (est)NCT04741945Ivosidenib for patients with clonal cytopenia of undetermined significance and mutations in IDH1Small moleculeIvosidenib (IDH1 inhibitor)Active, not recruitingApr 2022–Jan 2030 (est)NCT05030441A study to assess the safety, processing by the body, and effects on the body following 4 weeks of selnoflast dosing in participants with coronary artery diseaseSmall moleculeSelnoflast (NLRP3 inhibitor)CompletedJul 2022–May 2023ISRCTN10520571Canakinumab for the prevention of progression to cancer in patients with clonal cytopenias of unknown significance, IMPACT studyBiologic (monoclonal antibody)Canakinumab (IL-1β inhibitor)RecruitingFeb 2023–Dec 2028 (est)NCT05641831DFV890 and MAS825 for inflammation reduction in CHD patients with CHIPSmall molecule & biologic (monoclonal antibody)DFV890 (NLRP3 inhibitor) & MAS825 (anti–IL-1β monoclonal antibody)RecruitingFeb 2024–(ongoing)NCT06097663Statins in patients with clonal cytopenia of undetermined significance (CCUS) and myelodysplastic syndromes (MDS)Small moleculeAtorvastatin/rosuvastatin (HMG-CoA-reductase inhibitors)RecruitingFeb 2024–May 2027 (est)NCT05483010Curcumin to improve inflammation and symptoms in patients with clonal cytopenia of undetermined significance, low risk myelodysplastic syndrome, and myeloproliferative neoplasmsSupplementCurcumin (Supplement; anti-inflammatory/anti-oxidant)RecruitingMar 2024 – Mar 2027 (est)NCT06063486TECTONIC: IL-1β Inhibition in TET2-CHIP and ASCVDBiologic (monoclonal antibody)Canakinumab (IL-1β inhibitor)Not yet recruitingJul 2025–Apr 2030 (est)NCT06691217ASCVD atherosclerotic cardiovascular disease, CCUS clonal cytopenia of undetermined significance, CH clonal haematopoiesis, CHD coronary heart disease, CHIP clonal haematopoiesis of indeterminate potential, LR-MDS low-risk myelodysplastic syndromes. CH screening: considerations and current status within the Wilson & Jungner framework. CH clonal haematopoiesis, CV cardiovascular. Figure was created in BioRender. Morley, A. (2026) https://biorender.com/3gi3e5f Table 3Ongoing clinical trials in CH / CCUS.TrialIntervention typeDrug/therapyStatusDatesClinical trial codeCHIP-HF: colchicine in heart failure patients with CHIPSmall moleculeColchicine (microtubule inhibitor)OngoingAug 2021–(ongoing)EudraCT 2021-001508-13A study of enasidenib in people with clonal cytopenia of undetermined significanceSmall moleculeEnasidenib (IDH2 inhibitor)Active, not recruitingOct 2021–Oct 2026 (est)NCT05102370Repurposing metformin as a leukaemia-preventive drug in CCUS and LR-MDSSmall moleculeMetformin (biguanide)RecruitingDec 2021–Feb 2026 (est)NCT04741945Ivosidenib for patients with clonal cytopenia of undetermined significance and mutations in IDH1Small moleculeIvosidenib (IDH1 inhibitor)Active, not recruitingApr 2022–Jan 2030 (est)NCT05030441A study to assess the safety, processing by the body, and effects on the body following 4 weeks of selnoflast dosing in participants with coronary artery diseaseSmall moleculeSelnoflast (NLRP3 inhibitor)CompletedJul 2022–May 2023ISRCTN10520571Canakinumab for the prevention of progression to cancer in patients with clonal cytopenias of unknown significance, IMPACT studyBiologic (monoclonal antibody)Canakinumab (IL-1β inhibitor)RecruitingFeb 2023–Dec 2028 (est)NCT05641831DFV890 and MAS825 for inflammation reduction in CHD patients with CHIPSmall molecule & biologic (monoclonal antibody)DFV890 (NLRP3 inhibitor) & MAS825 (anti–IL-1β monoclonal antibody)RecruitingFeb 2024–(ongoing)NCT06097663Statins in patients with clonal cytopenia of undetermined significance (CCUS) and myelodysplastic syndromes (MDS)Small moleculeAtorvastatin/rosuvastatin (HMG-CoA-reductase inhibitors)RecruitingFeb 2024–May 2027 (est)NCT05483010Curcumin to improve inflammation and symptoms in patients with clonal cytopenia of undetermined significance, low risk myelodysplastic syndrome, and myeloproliferative neoplasmsSupplementCurcumin (Supplement; anti-inflammatory/anti-oxidant)RecruitingMar 2024 – Mar 2027 (est)NCT06063486TECTONIC: IL-1β Inhibition in TET2-CHIP and ASCVDBiologic (monoclonal antibody)Canakinumab (IL-1β inhibitor)Not yet recruitingJul 2025–Apr 2030 (est)NCT06691217ASCVD atherosclerotic cardiovascular disease, CCUS clonal cytopenia of undetermined significance, CH clonal haematopoiesis, CHD coronary heart disease, CHIP clonal haematopoiesis of indeterminate potential, LR-MDS low-risk myelodysplastic syndromes. Ongoing clinical trials in CH / CCUS. ASCVD atherosclerotic cardiovascular disease, CCUS clonal cytopenia of undetermined significance, CH clonal haematopoiesis, CHD coronary heart disease, CHIP clonal haematopoiesis of indeterminate potential, LR-MDS low-risk myelodysplastic syndromes. Screening could also be of value if treatments are discovered which halt progression of CH to myeloid malignancies, or even expansion in VAF – with larger clone sizes a key factor in determining leukaemic99–101 and cardiovascular18 risk. Numerous clinical trials are underway in this regard. Strategies being tested include metformin (NCT04741945) and vitamin C, both of which are associated with a lower risk of CH in observational studies11,102,103, and attenuate clonal expansion in animal models11,104,105. Other trials are assessing the potential of statins (NCT05483010) or curcumin (NCT06063486) in reducing progression of clonal cytopenias or low-risk myelodysplasia through suppression of inflammation. Lastly, specific mutations are being targeted, with studies of IDH1/2 inhibitors with known efficacy in IDH1/2 mutant AML being trialled in the context of IDH1/2 mutant CH (NCT05102370 and NCT05030441). A multi-faceted approach to discovering efficacious, safe treatments to prevent both the cardiovascular and haematological consequences of CH is therefore underway. Identifying CH in the clinic CH can be identified through several genetic sequencing methods with high accuracy106, which only require a blood sample, and so is minimally invasive to the patient. Sequencing will become more accessible as costs continue to decline, and low-cost assays for CH detection have been developed107. Notably, this assay has not been validated for clinical use, which should be a priority when low-cost solutions are developed to enable benefits to patients. Perhaps more of a barrier to consider is the psychological impact of receiving a CH diagnosis, and with it the uncertain risk of developing multi-system diseases with high morbidity and mortality. This was emphasised in a study of over 500 breast cancer patients surveyed on their opinions of CH testing108. Twenty-eight percent of participants experienced at least moderate anxiety on learning about CH, with the associated diseases and lack of treatments being factors that influenced this. Despite this, the majority of patients (>75%) reported that they would opt for testing for themselves and that a CH diagnosis would prompt them to change their behaviours and seek medical attention to reduce their health risks. Overall, this study suggests that within certain populations, such as oncology, CH screening may be viewed as acceptable. Given the enrichment of CH after cancer treatments109–112, this may be a suitable high-risk population for screening in the future. However, the perspectives of patients with recent breast cancer diagnosis may not be representative of the wider population, and acceptability would need to be assessed in other high-risk populations being considered for screening. Further research will guide who this would benefit, but could include those at high risk for CVD, with premature CVD or with a family history of leukaemia or CH. Conclusion CH is emerging as an important risk factor for CVDs, and more broadly for overall mortality risk; however, such risk is heterogeneous and often mutation and clone-size-specific. Our understanding of this nuanced relationship remains in its infancy, but as we look to transition from bench-to-bedside, there are knowledge gaps and inconsistencies which must be reconciled. In order to facilitate future efforts in examining associations of CH genetically with outcomes (including CVD), we highlight the need for an international consensus on methods of defining CH with respect to driver genes, mutation types and filtering strategies. CVD associations of mutation subtypes must then be clarified in longitudinal cohorts with sufficient statistical power. There is a potential role for screening and risk-stratification at a population level, which may become a reality with ongoing clinical trials designed to combat CVD risk in CH.