Abstract
Background
A large proportion of patients with cryptogenic stroke or transitory ischemic attack (TIA) have underlying subclinical atrial fibrillation (SCAF) detected on follow up. It is not clear whether SCAF is the underlying primary entity in the pathogenesis of stroke in these patients, or merely a marker of atrial myopathy associated with left atrial remodeling, fibrosis and inflammation.
Purpose
As a hypothesis generating study, we investigated a panel of selected biomarkers involved in fibrosis, inflammation, and thrombosis: growth differentiation factor 15 (GDF-15), transforming growth factor b (TGFb), galectin-3, soluble suppressor of tumorgenicity2 (sST2), von Willebrand factor (vWF), Tissue metalloprotease1 (TIMP1), Matrix metalloprotease9 (MMP9), Emmprin, Interleukin6 (IL6), C-reactive protein (CRP), Tissue factor (TF), Plasminogen activator inhibitor (PAI1), and their relation to the occurrence of SCAF during follow-up in patients after cryptogenic stroke or TIA. We hypothized that biomarker levels were increased in patients with subclinical AF.
Methods
236 patients, median age 71 years (range 21–94) of which 38% were women, with their first cryptogenic stroke or TIA were included 2–4 days after the index event and followed with an Implantable Cardiac Rhythm Monitor for >1 year. Echocardiography and blood sampling were performed at inclusion. ELISA methods were used.
Results
SCAF occurred in 84 patients (36%). Only GDF-15 was significantly increased in AF- vs no-AF patients: 1010 pg/mL (inter quartile range: 814–1416) vs 860 pg/mL (inter quartile range: 622–1197) (p=0.018), and correlated with the number of premature atrial contractions (PAC)/24h (by Holter ECG during index hospitalization) (rs=0.314, p<0.001) and AF-burden during follow-up (rs=0.149, p=0.022). Furthermore, there was a significant trend across quartiles of GDF-15 for having AF, and patients in the three highest quartiles (Q2–4) compared with Q1 had an odd ratio of having AF of 2.16 (95% CI 1.10–4.25), adjusted for sex and body mass index. The significance, however, was lost when adjusting for age, which correlated significantly to GDF-15 (rs=0.283; p<0.001). ROC curve analyses showed an AUC of 0.593 (0.52–0.68) for GDF-15 compared to 0.617 (0.54–0.69) for age. GDF-15 was also associated with co-morbidities such as hypertension (p<0.001), diabetes (p<0.001), and vascular disease (p<0.001).
Conclusion
In patients with a cryptogenic stroke or TIA experiencing SCAF during follow up, only levels of GDF-15 were elevated and correlated with PAC/24h and AF-burden. However, GDF-15 was highly related to age and co-morbidities and did not add significantly to the prediction of AF in a multivariate analysis.
Funding Acknowledgement
Type of funding source: Foundation. Main funding source(s): Stiftelsen Dam, Norwegian Atrial Fibrillation Research Network