Discovery of new biomarkers for atrial fibrillation using a custom-made proteomics chip

Another paper out of the PIVUS study. Here they looked specifically at arterial fibrillation (AF) and tried to identify potential prognostic markers. AF is a condition with a rising global incidence, creating a need for better prediction of those at risk. Many risk factors are known (weight, blood pressure, smoking, diabetes, past cardiac events etc), but potential plasma protein biomarkers have been limited. In this paper, the authors used CVD I to examine 978 subjects from the PIVUS cohort (discovery phase) and 725 from the ULSAM cohort (validation phase). 85/92 proteins were detected in >70% of samples from both cohorts and used in the study. 13 proteins were related to incident AF in PIVUS, of which 5 could be validated in the ULSAM cohort NT-pro-BNP, FGF-23, FABP4, GDF-15 & IL-6. Only NT-pro-BNP and FGF-23 were also associated with incident AF after adjustment for other known risk factors. This conservative discovery/validation approach was designed to avoid false-positives, but because the AF incidence levels were relatively small within the cohorts examined (around 15%), they also ran a secondary analysis merging the PIVUS/ULSAM datasets for meta-analysis based on the individual data – designed to avoid false-negatives. The 5 proteins from the validation phase showed the highest significance in the secondary study, but TIM-1 and AM also showed significant relation to incident AF (all 7 remained sígnificant after adjustment for known risk factors). This study confirmed earlier indications that NT-pro-BNP, FGF-23 and GDF-15 are linked to AF risk, but also identified FABP4, IL-6, TIM-1 and AM as potential new biomarkers for this disease. The authors were able to make some interesting speculations as to the possible pathophysiological roles for these proteins in AF, while remain cautious as to their involvement without further validation. On additional conclusion of interest was that adding NT-pro-BNP to the currently accepted CV risk factors for AF could be shown to add significant predictive power and improve the discrimination of the model.