Kidney Week

Abstract: FR-PO0519

Four eGFR-Dependent Proteomic Trajectories of Incident Myocardial Infarction Across the Kidney Function Spectrum: A UK Biobank Proteome-Wide Association Study

Session Information

• CKM: Clinical - Trials, Epidemiology, and Biomarkers
  October 23, 2026 | Location: Exhibit Hall A, Convention Center
  Abstract Time: 10:00 AM - 12:00 PM

Category: Cardiovascular-Kidney-Metabolic Health

• 602 Cardiovascular-Kidney-Metabolic Health: Clinical

Authors

• Na, Ki Ryang, Chungnam National University, Daejeon, Korea (the Republic of)

Ki Ryang Na, DrPH

• Lee, Yongwoo, Chungnam National University Hospital, Daejeon, Korea (the Republic of)

Yongwoo Lee, MD

• Jeong, Seongyeop, Chungnam National University Hospital, Daejeon, Korea (the Republic of)

Seongyeop Jeong

• Choi, Yeolim, Chungnam National University Hospital, Daejeon, Korea (the Republic of)

Yeolim Choi, MD

• In, Seungyong, Chungnam National University Hospital, Daejeon, Korea (the Republic of)

Seungyong In, MD

• Lee, Eu Jin, Chungnam National University Hospital, Daejeon, Korea (the Republic of)

Eu Jin Lee, MD

• Park, Heewon, Chungnam National University Sejong Hospital, Sejong, Korea (the Republic of)

Heewon Park, PhD

• Kim, Hae Ri, Chungnam National University Sejong Hospital, Sejong, Korea (the Republic of)

Hae Ri Kim, MD, PhD

• Choi, Dae Eun, Chungnam National University, Daejeon, Korea (the Republic of)

Dae Eun Choi, MD, PhD

Background

Cardiovascular risk in chronic kidney disease spans the full eGFR spectrum—including supranormal eGFR (hyperfiltration)—yet prior proteomic studies of incident myocardial infarction (AMI) treated kidney function dichotomously. We performed a three-tier proteome-wide association study (PWAS) to map how protein–AMI associations vary across the kidney function continuum and to identify the molecular programs at each eGFR domain.

Methods

Cox PWAS of 2,923 Olink Explore plasma proteins were run in 31,436 UK Biobank participants across three populations: overall, eGFR <60 mL/min/1.73m² (N=744), and eGFR <75 (N=3,719). Proteins reaching significance in all three were designated consensus proteins. A prespecified PWAS was performed in the hyperfiltration subgroup (eGFR ≥100, N=8,163). Grade-stratified Cox and natural cubic spline interaction models characterized eGFR-dependent HR trajectories. Pathway enrichment used clusterProfiler; incremental prediction was tested with LASSO-selected features.

Results

Over median 13.15-year follow-up (1,649 AMI events), staged screening identified 362 AMI-associated proteins (FDR<0.05) and 16 consensus proteins, segregating into four eGFR-dependent trajectory patterns: (1) progressive uremic amplification at low eGFR (10 proteins); (2) mid-range peak at G2–G3a (EDA2R, RELT, IGFBP4); (3) sustained protection across all strata (EGFR, PLTP); and (4) hyperfiltration peak unique to MMP12 (HR 2.30, 95% CI 1.33–3.97; confirmed in the prespecified G1b PWAS, HR 1.93, p<0.001). The MMP12-hyperfiltration signal persisted after excluding current smokers and using cystatin C-based eGFR. Pathway analysis revealed three programs: universal cytokine-driven immunoinflammation; ECM/fibrosis in CKD; and a vascular-injury program at hyperfiltration. Adding proteomic features improved AMI discrimination (AUC 0.761→0.805; DeLong p=0.0087).

Conclusion

A four-pattern molecular taxonomy frames AMI risk across the kidney function spectrum, with MMP12 uniquely identifying elevated risk at hyperfiltration—challenging the assumption that supranormal eGFR is cardiovascularly benign and providing a molecular framework for eGFR-stratified cardiorenal risk assessment.

Funding

• Government Support – Non-U.S.