Inflammatory proteomics profiling for prediction of incident atrial fibrillation

Abstract

Objective: Atrial fibrillation (AF) has emerged as a common condition in older adults. Cardiovascular risk factors only explain about 50% of AF cases. Inflammatory biomarkers may help close this gap as inflammation can alter atrial electrophysiology and structure. This study aimed to determine a cytokine biomarker profile for this condition in the community using a proteomics approach.

Methods: This study uses cytokine proteomics in participants of the Finnish population-based FINRISK cohort studies 1997/2002. Risk models for 46 cytokines were developed to predict incident AF using Cox regressions. Furthermore, the association of participants' C reactive protein (CRP) and N-terminal pro B-type natriuretic peptide (NT-proBNP) concentrations with incident AF was examined.

Results: In 10 744 participants (mean age of 50.9 years, 51.3% women), 1246 cases of incident AF were observed (40.5% women). The main analyses, adjusted for participants' sex and age, suggested that higher concentrations of macrophage inflammatory protein-1β (HR=1.11; 95% CI 1.04, 1.17), hepatocyte growth factor (HR=1.12; 95% CI 1.05, 1.19), CRP (HR=1.17; 95% CI 1.10, 1.24) and NT-proBNP (HR=1.58; 95% CI 1.45, 1.71) were associated with increased risk of incident AF. In further clinical variable-adjusted models, only NT-proBNP remained statistically significant.

Conclusion: Our study confirmed NT-proBNP as a strong predictor for AF. Observed associations of circulating inflammatory cytokines were primarily explained by clinical risk factors and did not improve risk prediction. The potential mechanistic role of inflammatory cytokines measured in a proteomics approach remains to be further elucidated.

Keywords: atrial fibrillation; biomarkers; epidemiology; inflammation.

Figure 1

Study flow for each FINRISK cohort.

Figure 2

Correlogram based on Pearson correlations of the measurements investigated (multiple imputed data). Measurements are ordered according to hierarchical clustering of their values; measurements with values in both cohorts are displayed as correlation meta-analyses, while those within a single cohort are shown as the value in that particular cohort. Abbreviations for cytokines are provided in the online supplemental materials.

Figure 3

Forest plots of hazard ratios (with 95%CIs) for every cytokine measurement evaluated in each cohort and the overall meta-analysis; the left figure corresponds to the simpler, age-adjusted and sex-adjusted model, while the right figure corresponds to the multiple risk factor-adjusted model. Note: Multivariable adjustment comprised baseline age, sex, body mass index, smoking, total cholesterol, systolic blood pressure, antihypertensive medication, incident myocardial infarction and incident heart failure. Estimates with 95% CIs for each measurement were ordered from lower to higher; *Statistically significant estimates ‘cohort-wise’ or overall (meta-analysis). Abbreviations for cytokines are provided in the online supplemental materials.

Figure 4

As cardiovascular risk factors only explain about 50% of AF cases, in this study risk models for 46 circulating cytokines, C reactive protein (CRP) and N-terminal pro B-type natriuretic peptide (NT-proBNP) were calculated to predict incident AF in 10 744 participants of the population-based FINRISK cohort studies 1997/2002. Incident AF was observed in 1246 persons during a median follow-up of 20.8 years. The main analyses, adjusted for participants’ age and sex, suggested that higher concentrations of macrophage inflammatory protein-1β, hepatocyte growth factor, CRP and NT-proBNP were associated with increased risk of incident AF. In further clinical variable-adjusted models, only NT-proBNP remained statistically significant. AF, Atrial fibrillation.