Higher C-reactive protein levels have been observed in patients with persistent atrial fibrillation than in patients with paroxysmal atrial fibrillation, suggesting that Creactive protein levels may be related to the burden of atrial fibrillation [14].

Higher levels of YKL-40, a new biomarker of inflammation, have recently been documented in individuals with permanent atrial fibrillation compared to patients with persistent atrial fibrillation, further supporting the concept of an association between the chronicity of atrial fibrillation and the inflammatory burden [13].

Although there is abundant data suggesting a link between inflammation and atrial fibrillation, several studies failed to show elevated levels of C-reactive protein, interleukin-6, or interleukin-8 in atrial fibrillation patients [15-19].

A recent study was conducted in our department that included 81 subjects, 72 atrial fibrillation patients (39 with paroxysmal atrial fibrillation and 33 with persistent atrial fibrillation) and 9 control patients with Wolff-Parkinson-White syndrome. In the study, the levels of several inflammatory markers were determined in peripheral, coronary sinus, left atrial and pulmonary veins. There were no significant differences among the three groups regarding interleukin-8, soluble intercellular adhesion molecule-1 or transforming growth factor β1 levels at any of the sampling sites. Similarly, C-reactive protein levels, which were only measured in the peripheral blood, did not differ in the three groups. These results suggest that the high inflammatory marker levels reported in atrial fibrillation patients could be due to associated comorbidities rather than to atrial fibrillation per se [20].

In the same study that compared paroxysmal and persistent atrial fibrillation patients with controls without any history of atrial fibrillation [20], the peripheral levels of vascular endothelial growth factor (VEGF) were higher in both paroxysmal and persistent atrial fibrillation patients compared to controls (Figure 1). In the left atrium, VEGF levels were higher in the paroxysmal atrial fibrillation group compared to the control group; however, there was no significant difference between persistent atrial fibrillation patients and controls (Figure 1).

A number of pathological conditions induce VEGF secretion, including oxygen deficiency, inflammation, and pulsatile mechanical stretch [21]. Patients with atrial fibrillation showed low levels of several inflammatory markers, regardless of the clinical form of the arrhythmia. This suggests that the inflammatory process is low grade, if present at all, in these patients. Therefore, it seems unlikely that inflammation alone could account for the high peripheral VEGF levels in atrial fibrillation patients. Despite the absence of evidence of accelerated angiogenesis in paroxysmal atrial fibrillation patients, we found high levels of VEGF in the left atrium in this population. Histopathological studies report a high incidence of atrial ischemia in up to 17% of autopsied patients with a history of atrial fibrillation and isolated atrial infarctions in more than 20% of them [22]. A transient increase in hypoxia inducible factor α (HIF1α) gene expression, which is a known trigger for VEGF secretion, is reported in cardiac myocytes in the early onset of atrial fibrillation [23]. On the other hand, the inverse relationship is also a possibility, with ischemia favouring atrial fibrillation occurrence.

A fibrillating atrium leads to irregular blood flow and may induce pulsatile vascular stretch and impaired blood rheology with subsequent VEGF secretion in the peripheral blood [24-26]. Pulsatile mechanical stretch, recently postulated to be a potent trigger for VEGF secretion from cardiac myocytes [21,27], could lead to the high intracardiac VEGF levels observed in paroxysmal atrial fibrillation.

Nevertheless, it appears that left atrial secretion of VEGF is a transient event in the natural history of atrial fibrillation. Persistent atrial fibrillation is associated with more important fibrotic changes than the paroxysmal form (Figure 2) [28,29].

VEGF itself may be involved in these fibrotic changes, probably through the induction of angiogenesis [30-32]. The spreading of cardiac fibrosis may cause progressive

cardiac stiffening, thereby reducing the degree of pulsatile stretch and subsequently diminishing intracardiac VEGF levels, as observed in patients with persistent atrial fibrillation. Thus, newly identified biomarkers, such as VEGF, may serve as predictors of atrial fibrillation progression.