Rare and Common Genetic Variation Underlying Atrial Fibrillation Risk
Authors, Journal, Affiliations, Type, DOI
- Authors: Oliver B. Vad, Laia M. Monfort, Christian Paludan-Müller, et al.; for the Geisinger MyCode Community Health Initiative and Regeneron Genetics Center Research Team
- Journal: JAMA Cardiology 2024;9(8):732–740
- Affiliations: Department of Cardiology, Copenhagen University Hospital – Rigshospitalet; Department of Biomedical Sciences, University of Copenhagen; Regeneron Genetics Center, Tarrytown, New York
- Type: Genetic association and nested case-control study (observational)
- DOI: https://doi.org/10.1001/jamacardio.2024.1528
Overview
Using whole-exome sequencing of 403,990 UK Biobank participants, this study identified rare predicted loss-of-function (pLOF) variants in 6 genes (TTN, RPL3L, PKP2, CTNNA3, KDM5B, C10orf71) significantly associated with AF, with 5 of 6 replicated in an external cohort of >160,000 individuals. The key finding is a multiplicative interaction between rare pLOF variants and a polygenic risk score (PRS): together they confer an OR of 7.08 for AF. Carriers with a top-quintile PRS face a 10-year absolute AF risk of 24% (males >60 years). Critically, pLOF variants — predominantly driven by TTN — are associated with a 3× increased risk of cardiomyopathy both before and after AF diagnosis, supporting the concept that AF may be the first clinical manifestation of an underlying cardiomyopathy in genetically predisposed individuals.
Keywords
Atrial fibrillation · Rare variants · Predicted loss-of-function · Whole-exome sequencing · Polygenic risk score · Cardiomyopathy · Heart failure · TTN · PKP2 · UK Biobank · Gene-based burden test
Key Takeaways
Study Design and Population
- UK Biobank whole-exome sequencing: 403,990 individuals (54.1% female; median age 58 years; European ancestry); median follow-up 13.3 years (end July 2022).
- Case definition: 31,124 individuals with AF (6,677 prevalent at inclusion + 24,447 incident AF during follow-up).
- Gene-based burden test: Rare pLOF variants (MAF <1%) collapsed across 17,979 genes; significance threshold P <2.77 × 10⁻⁶ (correction for ~18,000 genes). Firth logistic regression adjusted for age, sex, and 10 principal components.
- External replication cohort: 17,910 individuals with AF and 149,348 controls.
Gene-Based pLOF Associations with AF (Table 2)
Six genes reached exome-wide significance for pLOF variants and AF:
| Gene | Protein | OR (95% CI) | P value | Replicated? |
|---|---|---|---|---|
| TTN (PSI90) | Titin (constitutively expressed cardiac exons, spliced-in >90%) | 3.85 (3.25–4.55) | 1.09 × 10⁻⁵⁵ | Yes |
| TTN (N2BA-N2B) | Titin (cardiac isoforms) | 2.17 (1.93–2.43) | 2.07 × 10⁻⁴⁰ | Yes |
| TTN (all) | Titin | 1.77 (1.60–1.95) | 3.38 × 10⁻³⁰ | Yes |
| RPL3L | Ribosomal protein L3-like | 1.56 (1.38–1.77) | 1.69 × 10⁻¹² | Yes |
| PKP2 | Plakophilin-2 | 2.12 (1.60–2.82) | 2.21 × 10⁻⁷ | Yes |
| CTNNA3 | Catenin α3 | 2.79 (1.88–4.14) | 3.74 × 10⁻⁷ | Yes |
| C10orf71 | Cardiac-enriched FHL2-interacting protein (CEFIP) | 2.40 (1.69–3.39) | 7.83 × 10⁻⁷ | No |
| KDM5B | Lysine demethylase 5B | 2.70 (1.80–4.06) | 1.76 × 10⁻⁶ | Yes |
- Associated genes were predominantly expressed in cardiomyocytes; KDM5B was the exception (expressed across all cell types). C10orf71 was almost exclusively expressed in muscle tissue.
- Secondary missense variant analysis replicated one association: UBE4B (OR 1.22; P = 5.90 × 10⁻⁷) — predominantly driven by a single missense variant.
- TTN PSI90 (constitutively expressed cardiac exons) had the strongest association — restricting analysis to functionally expressed exons greatly increases signal.
Novel Gene Mechanisms
- CTNNA3 (catenin α3): Cytoskeletal protein interacting with cardiomyocyte desmosomes; putatively associated with ARVC and familial AF. Confirms a role for desmosome-associated cytoskeletal proteins in AF beyond classical ARVC genes.
- KDM5B (lysine demethylase 5B): Not previously associated with arrhythmias. Encodes a histone H3K4 demethylase thought to play a role in cardiac fibrosis — a common substrate for re-entry arrhythmias. Novel mechanism: epigenetic regulation of arrhythmia substrate.
- RPL3L (ribosomal protein L3-like): Involved in ribosomal function and muscle growth. The association was primarily driven by a single splice-donor variant previously reported in an AF GWAS, providing mechanistic annotation.
- C10orf71 (CEFIP): Located at a locus near a known AF GWAS signal; protein may locate to sarcomere Z-discs and regulate cardiomyocyte hypertrophy. Association did not replicate — uncertain causal role.
Combined PRS + pLOF Risk: Multiplicative Interaction
- PRS alone: 1-SD increase in AF PRS → OR 1.53 (95% CI 1.51–1.55); AUC 0.76 (reference model without PRS: AUC 0.74).
- PRS alone in pLOF carriers: OR per SD 1.69 (vs. 1.53 in non-carriers) — suggests rare variant carriers are also more sensitive to polygenic risk.
- Top quintile PRS + pLOF variant: OR 7.08 (95% CI 6.03–8.28) for AF — dose-response increase across all 10 PRS × carrier strata.
- Incident AF hazard ratio (top PRS quintile + pLOF): HR 4.78 (95% CI 4.06–5.63) vs. non-carriers with low PRS.
- Cumulative absolute risk by age 80: 28.6% in top PRS + pLOF carriers vs. 12.1% (mid PRS, non-carrier) and 8.1% (low PRS, non-carrier).
10-Year Absolute Risk of AF (Figure 2)
- Highest risk group (males >60 years, top PRS quintile, pLOF carrier): 24%; equivalent female group: 16%.
- Males >60 years, top PRS quintile, non-carrier: 21%.
- Modifiable risk factors compound polygenic + rare variant risk: BMI ≥30 + hypertension + top PRS: comparable absolute risk to pLOF carriers.
- Males <60 years with top PRS + pLOF: 10%; equivalent females: 5%.
Genetic Predisposition and Risk of Cardiomyopathy / Heart Failure
All analyses used cause-specific Cox regression with AF, HF, and cardiomyopathy as competing events:
Full cohort (without AF at baseline; n = 395,528):
- pLOF variants (all 5 replicated genes) → incident AF: HR 1.85 (1.69–2.02)
- pLOF variants → incident cardiomyopathy: HR 3.13 (2.24–4.36; P <0.001)
- pLOF variants → incident HF: HR 1.51 (1.26–1.82)
- When TTN excluded: CM HR drops to 1.39 (0.80–2.40, non-significant); HF HR drops to 1.01 (non-significant) — TTN drives almost all the CM/HF association.
- AF PRS per SD: associated with AF (HR 1.45) but not with cardiomyopathy (HR 0.99) or HF (HR 1.02) — common polygenic variation is arrhythmia-specific, not cardiomyopathy-specific.
In individuals with AF (n = 21,154; no prior HF/CM):
- pLOF variants → incident cardiomyopathy post-AF: HR 2.98 (1.89–4.69; P <0.001)
- When TTN excluded: HR 1.01 (non-significant) — again TTN-driven.
- AF PRS: not associated with cardiomyopathy or HF after AF.
Interpretation: For pLOF variant carriers (especially TTNtv), AF may be the first disease manifestation preceding more severe cardiomyopathy. The common genetic AF PRS does not carry this cardiomyopathy risk — supporting fundamentally different biological pathways for rare vs. common genetic AF risk.
Clinical Implications
- Rare pLOF variants with large effect sizes may justify genetic testing in specific high-risk groups (early-onset AF, family history, coexisting CM features).
- Prior data (Yoneda 2021, 2022) showed higher rare variant rates and increased mortality in younger patients — consistent with findings from this general population cohort.
- Aligns with AHA/ACC 2023 recommendation (Class IIb): genetic testing/CM surveillance may be reasonable in AF onset <45 years.
- Integration of PRS + rare pLOF variant status could enable population-level risk stratification as WES becomes cheaper and more accessible.
Limitations of the document
- European ancestry only: Results may not be generalizable to other ethnicities; population stratification corrections applied but residual confounding possible.
- Strict significance threshold (P <2.77 × 10⁻⁶) may have limited power to detect additional AF-associated genes.
- Observational design: Gene-based associations cannot be assumed causal without functional validation. Confounding by indication and residual confounding possible.
- C10orf71 did not replicate externally; causal role uncertain.
- UK Biobank healthy volunteer bias: Enrolled individuals are healthier on average than the general UK population — absolute risk estimates may be conservative.
- Age at inclusion 40–69: Not a dedicated early-onset AF cohort; the youngest patients are included but the study is not powered to characterize very early-onset AF (<30 years) specifically.
Key Concepts Mentioned
- concepts/Early-Onset-Atrial-Fibrillation — pLOF variants provide population-level corroboration of cardiomyopathy gene involvement in AF
- concepts/Genetic-Testing-in-AF — combined PRS + rare variant model; clinical case for genetic testing
- concepts/Arrhythmogenic-Cardiomyopathy — CTNNA3 and PKP2 link AF to desmosomal disease
Key Entities Mentioned
- entities/TTN — strongest and most prevalent pLOF AF association; drives all cardiomyopathy/HF signal
- entities/PKP2 — pLOF OR 2.12 for AF; replicated; corroborates GWAS locus; desmosomal pathway
- entities/Atrial-Fibrillation — primary outcome; PRS + pLOF multiplicative model
- entities/DCM — CTNNA3 and TTN overlap with cardiomyopathy gene landscape
Wiki Pages Updated
- Created: wiki/sources/Biobank-AF-JAMA-2024.md
- Updated: wiki/entities/TTN.md (population-level pLOF AF evidence; CM/HF risk driven by TTNtv)
- Updated: wiki/entities/PKP2.md (population-level pLOF AF association OR 2.12; replicated)
- Updated: wiki/concepts/Genetic-Testing-in-AF.md (PRS + pLOF combined model; OR 7.08; 10-year absolute risk data)
- Updated: wiki/concepts/Early-Onset-Atrial-Fibrillation.md (UK Biobank population-level pLOF data)
- Updated: wiki/wikiindex.md
- Updated: log.md