The role of plasma cell-free DNA as predictor of clinical outcome in heart failure – Results from the MyoVasc study
Date issued
Authors
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
License
Abstract
Heart failure (HF) represents a major cause of mortality with a prevalence of 1-2% in the adult population in developed countries. As populations age, the role of HF is expected to grow. This comes with a steep increase in healthcare costs, placing a substantial burden on society. Identifying HF risk patients earlier, possibly even before symptoms manifest, becomes crucial to initiate interventions promptly, such as lifestyle adjustments or medication. In preventive medicine, the identification of suitable biomarkers plays a key role, as they allow an objective and early disease detection.
Cell-free DNA (cfDNA) is a widely used diagnostic biomarker in clinical fields like oncology or transplantation medicine. In clinical cardiology, however, cfDNA analytics does not yet play a major role. Only a handful of rather small-scale studies have so far investigated the potential of cfDNA diagnostics in HF patients, indicating that cfDNA could be an independent risk factor for cardiovascular disease and overall mortality. The aim of the present study was therefore to evaluate the potential of cfDNA in HF diagnostics in a large cohort of subjects and to compare its predictive power to the currently most often used biomarker, NT-proBNP.
To achieve this, a reliable, reproduceable, and quick high throughput cfDNA quantification method needed to be implemented. The existing manual, time- and labour-consuming qPCR assay was automated by testing and establishing an INTEGRA pipetting robot and tuning its workflow to the special needs of high-viscosity plasma samples. The assay was adjusted to reliably produce the same test results as with the already published qPCR assay established by Neuberger et al. (183). This way a consistent measuring of the study samples was ensured. cfDNA levels were then quantified in 3109 EDTA plasma samples from the prospective MyoVasc study (NCT04064450). Two qPCR assays of different amplicon lengths (cfDNA90 bp/ cfDNA222 bp), both targeting a repetitive LINE1 element, were used for cfDNA quantification and to calculate the cfDNA integrity index, which indicates the fragmentation level of the cfDNA. Competing risk models were applied to investigate the associations of cfDNA with worsening of HF, and Cox proportional hazard regression analyses were used to assess the endpoints of cardiac death and all-cause death. C-statistics were calculated and
compared for each model. The participants were classified as 0 (healthy) or HF stages A (at risk for HF) to D (advanced HF) according to the current Universal Definition of Heart Failure. Analyses were adjusted for age, sex, cardiovascular risk factors (CVRFs) and medication (models 1-3) and additionally for NT-proBNP (model 4). Outcome data were presented as cumulative incidence plots for cfDNA90bp and 222bp levels and for the integrity index.
The cohort included 3109 study participants with an age between 34 to 85 years and 35.7% females. cfDNA concentration was lowest in stage 0/A subjects (n=534) with 10.99 (8.70/13.93) ng/ml (median (Q1/Q3)). Stage B (pre-HF) (n=923) or stage C/D subjects (n=1652) showed elevated cfDNA90bp concentrations with 13.37 (10.35/18.11) or 17.11 (12.56/22.80) ng/ml, respectively. Cox proportional hazard regression analyses indicated that the concentration of cfDNA90bp is a relevant prognostic marker for all-cause death, adjusted for age, sex, CVRFs and medication (HR = 1.312 [1.205-1.430], p < 0.0001). After additional adjustment for NT-proBNP, the effect estimates were lower, but still statistically significant (HR = 1.173 [1.073-1.282], p = 0.00046). Regarding the endpoints worsening of HF and cardiac death, the effect estimates were no longer significant after adjustment for NT-proBNP. A C-
index comparison showed the same tendency, with a significant added value of testing cfDNA additionally to NT-proBNP only when looking at all-cause death (C = 0.807 vs. C = 0.805; p = 0.050). However, cumulative incidence plots for dichotomised values of NT-proBNP and cfDNA showed the highest incidence rates for all three outcomes in patients with elevations in both biomarkers, significantly higher than in patients with elevations of NT-pro BNP alone.
The present results indicate that cfDNA is a risk factor, which independently of NT-proBNP contributes to the prediction of overall mortality (all-cause death) in the study cohort. cfDNA also appears to possess additional information value to NT-proBNP for predicting worsening of HF and cardiac death.