Blood Cell Mutations Linked to 27–52% Higher Heart Failure Risk in 417K Adults
UK Biobank study finds clonal hematopoiesis—age-related blood stem cell mutations—significantly raises incident heart failure risk, especially non-DNMT3A subtypes.
Summary
A large UK Biobank study of 417,616 adults found that clonal hematopoiesis of indeterminate potential (CHIP)—age-related mutations in blood stem cells—raises heart failure risk by 27% overall. Non-DNMT3A CHIP variants (TET2, ASXL1, JAK2, spliceosome genes) drove the strongest associations, with up to 4-fold elevated risk for JAK2 CHIP. DNMT3A CHIP showed a more modest 15% increased risk, significant only for the R882 hotspot variant. Importantly, known CHIP-related comorbidities like coronary artery disease, atrial fibrillation, diabetes, and chronic kidney disease explained only about 28% of the non-DNMT3A CHIP-heart failure link, suggesting direct cardiac mechanisms are at play.
Detailed Summary
Heart failure affects millions globally and its causes remain incompletely understood. Clonal hematopoiesis of indeterminate potential (CHIP)—where aging-related mutations in blood stem cells cause a single clone to expand and affect 10–20% of people over age 70—has emerged as a potential cardiovascular risk factor. This study is among the largest and most gene-specific investigations of CHIP and incident heart failure to date.
Researchers analyzed whole-exome sequencing data from 417,616 UK Biobank participants (mean age 56 years, 56% female) free of heart failure, hematologic malignancy, and key comorbidities at baseline, following them for a median of 11.1 years. CHIP was detected in 3.3% of participants, most commonly DNMT3A (63.5% of carriers), followed by TET2 (14.4%), ASXL1 (10.3%), DNA damage repair genes (4.3%), spliceosome genes (3.0%), and JAK2 (0.8%). The primary outcome was incident heart failure, identified via ICD-10 codes.
Overall CHIP was associated with a 27% higher risk of incident heart failure after multivariable adjustment. The non-DNMT3A composite showed a 52% increased risk (aHR 1.52), with striking individual gene-level associations: JAK2 CHIP conferred a 4-fold risk increase (aHR 4.01), spliceosome CHIP a 2.45-fold increase, ASXL1 a 1.51-fold increase, and TET2 a 1.34-fold increase. DNMT3A CHIP was only nominally significant overall (aHR 1.15), and only the R882 hotspot subtype reached significance (aHR 1.41). Higher variant allele frequency showed a linear relationship with heart failure risk for non-DNMT3A but not DNMT3A CHIP, and a sex interaction was observed for DNMT3A CHIP, with stronger risk in females.
Critically, mediation analyses using propensity-score-matched groups found that the four major CHIP-associated comorbidities—coronary artery disease, atrial fibrillation, type 2 diabetes, and chronic kidney disease—collectively explained only 28.2% of the non-DNMT3A CHIP–heart failure association. This implies that roughly 72% of the excess risk operates through independent, likely direct cardiac or inflammatory pathways, positioning CHIP as a potentially targetable upstream driver of heart failure rather than merely a comorbidity marker.
Sensitivity analyses including complete-case analysis, propensity score matching, Fine-Gray competing-risk models, and stricter heart failure definitions all yielded consistent results. These findings reinforce CHIP as an independent heart failure risk factor and suggest that therapeutic strategies targeting CHIP-associated inflammation or specific mutant clones could represent a novel prevention frontier.
Key Findings
- Any CHIP associated with 27% higher incident heart failure risk (aHR 1.27) over 11 years in 417K adults.
- JAK2 CHIP conferred the highest risk (aHR 4.01); spliceosome CHIP was also strongly associated (aHR 2.45).
- Non-DNMT3A CHIP was significantly more dangerous than DNMT3A CHIP for heart failure (P for heterogeneity = .004).
- Known CHIP comorbidities (CAD, AF, T2D, CKD) explained only 28% of the non-DNMT3A CHIP–heart failure link.
- DNMT3A R882 hotspot variant, but not non-R882 DNMT3A CHIP, was independently associated with heart failure.
Methodology
Prospective cohort study using UK Biobank whole-exome sequencing data from 417,616 participants followed for median 11.1 years. Cox proportional hazards regression adjusted for age, sex, race, and multiple cardiovascular risk factors; mediation analysis used 4:1 propensity-score-matched groups and the difference method to quantify mediating comorbidity effects.
Study Limitations
UK Biobank is predominantly White British, limiting generalizability to diverse populations. CHIP was assessed only at baseline, so clonal dynamics over time were not captured. Heart failure subtypes (HFpEF vs HFrEF) could not be fully distinguished, and residual confounding from unmeasured factors cannot be excluded.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe:
