Proteomic signatures of atrial fibrillation in hospitalized patients with heart failure
Z Nezami, M A Ohlsson, P G Platonov, A Jujic, M MagnussonAbstract
Introduction
Atrial fibrillation (AF) is common among patients hospitalized with heart failure (HF) and is associated with adverse cardiovascular outcomes. However, the biological pathways underlying AF in the context of HF remain incompletely characterized. Targeted proteomic profiling may help identify circulating biomarker patterns associated with AF beyond established clinical risk factors.
Aims
To compare the circulating expression of 92 proteins between hospitalized HF patients with and without AF, and to identify proteins independently associated with AF using a two-stage analytical approach accounting for multiple testing.
Methods
The study population consisted of 324 patients hospitalized for new-onset or worsening HF. Proteomic measurements and complete AF status were available in 315 participants, who formed the analysis cohort. Atrial fibrillation was validated through patient ECG-review. Protein expression was analyzed using protein-wise linear regression. In stage 1, unadjusted models were applied to all 92 proteins, and false discovery rate (FDR) correction was used to identify proteins associated with AF. Proteins meeting FDR < 0.05 in stage 1 were taken forward to stage 2, where multivariable linear regression models adjusted for age, sex, systolic blood pressure, and body mass index were fitted. FDR correction was reapplied within the screened protein set.
Results
Among the 315 participants included in the analysis, 211 had AF and 104 did not. Mean age was 74.7±11.4 years, and 69.5% were men. In stage 1 unadjusted screening, 22 of 92 proteins were associated with AF after FDR correction. In stage 2 multivariable analyses, 13 proteins remained independently associated with AF after FDR correction. TRAP and LDLreceptor were lower in patients with AF, whereas Notch3, IGFBP7, adiponectin (APN), IGFBP1, TIMP4, SPON1, CDH5, CNTN1, MMP2, FABP4, and CTSD were higher. The strongest association was observed for TRAP (β = −0.387, p = 3.33 × 10⁻¹⁰, FDR = 7.32 × 10⁻⁹), followed by Notch3 (β = 0.296, p = 1.54 × 10⁻⁵, FDR = 1.35 × 10⁻⁴) and IGFBP7 (β = 0.404, p = 1.84 × 10⁻⁵, FDR = 1.35 × 10⁻⁴). The overall distribution of effect sizes and significance across the full protein panel is shown in Figure 1, while Figure 2 highlights the standardized between-group differences for the 13 independently associated proteins.
Conclusion
In hospitalized patients with heart failure, atrial fibrillation was associated with a distinct circulating proteomic profile. Using a two-stage FDR-controlled approach, 13 proteins remained independently associated with AF after adjustment for age, sex, body mass index, and systolic blood pressure. These findings suggest AF-related alterations in circulating proteins linked to vascular biology, extracellular matrix remodelling, and metabolic regulation, and motivate further studies to explore prognostic relevance and underlying mechanisms.Figure 1.Volcano plot of proteomicsFor image description, please refer to the figure legend and surrounding text.Figure 2.Heatmap of proteinsFor image description, please refer to the figure legend and surrounding text.