Biomarkers in atrial fibrillation: an overview

Authors

  • J. A. Vílchez,

    1. Department of Cardiology, Hospital Universitario Virgen de la Arrixaca, University of Murcia, Murcia, Spain
    2. Department of Clinical Analysis, Hospital Universitario Virgen de la Arrixaca, University of Murcia, Murcia, Spain
    Search for more papers by this author
  • V. Roldán,

    1. Hematology and Medical Oncology Unit, Hospital Universitario Morales Meseguer, University of Murcia, Murcia, Spain
    Search for more papers by this author
  • D. Hernández-Romero,

    1. Department of Cardiology, Hospital Universitario Virgen de la Arrixaca, University of Murcia, Murcia, Spain
    Search for more papers by this author
  • M. Valdés,

    1. Department of Cardiology, Hospital Universitario Virgen de la Arrixaca, University of Murcia, Murcia, Spain
    Search for more papers by this author
  • G. Y. H. Lip,

    1. University of Birmingham Centre for Cardiovascular Sciences, City Hospital, Birmingham, UK
    Search for more papers by this author
  • F. Marín

    Corresponding author
    1. Department of Cardiology, Hospital Universitario Virgen de la Arrixaca, University of Murcia, Murcia, Spain
    • Correspondence to:

      Francisco Marín, MD, PhD, Department of Cardiology, Hospital Universitario Virgen de la Arrixaca, Universidad de Murcia, Ctra Madrid-Cartagena s/n, Murcia, 30120, Spain

      Tel.: 0034 968398115

      Fax: 0034 968369662

      Email: fcomarino@hotmail.com

    Search for more papers by this author

  • Disclosure

    None declared in relation to this manuscript for all authors. VR has received funding for consultancy and lecturing from Bristol-Myers-Squibb, Bayer and Boehringer Ingelheim. FM has received funding for research, consultancy and lecturing from Abbott, Boston Scientifics, Bayer, Astra Zeneca, Daiichi-Sankyo, BMS/Pfizer and Boehringer Ingelheim. GYHL has served as a consultant for Bayer, Astellas, Merck, Astra-Zeneca, Sanofi-Aventis, Aryx, Portola, Biotronic, and Boehringer-Ingelheim and has been on the speaker bureau for Bayer, Boehringer-Ingelheim, and Sanofi-Aventis.

Summary

Atrial fibrillation (AF) confers a raised risk of stroke and death, and this risk of adverse events is increased by the coexistence of other cardiovascular risk factors. The pathophysiology of AF is complex, involving the role of inflammation, structural remodelling with apoptosis, inflammation or fibrosis. These changes confer a prothrombotic or hypercoagulable state in this arrhythmia. Despite being easy to use for decision-making concerning oral anticoagulant therapy in AF, clinical risk scores used for stratification have shown modest capability in predicting thromboembolic events, and biomarkers may improve our identification of ‘high risk’ patients. Biomarkers, whether measured in the peripheral blood, urine or imaging-based may improve our knowledge of the pathophysiology of AF. Importantly these biomarkers could help in the assessment of AF prognosis. The aim of this review was to summarise the published data about biomarkers studied in AF, with focus on data from randomised prospective clinical trials and large community-based cohorts. We will also review the application of these biomarkers to prognosis on the main schemes used to help stratify risk in AF.

Review criteria

  • We have reviewed in Pubmed using the following key words: Atrial fibrillation, biomarkers, NT-proBNP, Troponin, C-reactive protein, von Willebrand factor, D-dimer, adipokines, renal biomarkers. We have also reviewed out-standing revision manuscript in atrial fibrillation and biomarkers.

Message for the clinic

  • Biomarkers could give us important information about underlying mechanisms and more importantly prognostic information.

Introduction

Atrial fibrillation (AF) is the most common cardiac arrhythmia, which is associated with increased morbidity and mortality [1]. AF increases the risk of stroke and death, and this risk of adverse events is increased by the coexistence of other cardiovascular risk factors such as heart failure, hypertension, diabetes mellitus or prior thromboembolism [2]. The pathophysiology of AF is complex and multifactorial, involving ageing and a structural remodelling whereby apoptosis, inflammation and fibrosis are the hallmarks [3]. Indeed, left ventricular dysfunction and elevated ventricular filling pressures contribute to atrial remodelling, and the role of the pulmonary veins as one of the key trigger sites for the onset of AF has also been described [4-6].

The pathways underlying thrombogenesis in AF are complex. Abnormal changes are consistent with a prothrombotic or hypercoagulable state in AF [7]. More specifically, inflammation seems to play an important role in the prothrombotic state associated with AF [8]. Furthermore, in recent years there are different data that highlight the association between inflammation to AF itself and AF-related complications [8]. Indeed, multiple reports suggest the role of inflammation is not only as a marker of incident AF, but also as an underlying mechanism involved in the induction of AF [9, 10].

Risk stratification of AF: the use of clinical risk scores

Oral anticoagulation (OAC) is highly effective in reducing stroke risk and mortality rates in patients with AF, but also increases the risk of bleeding [11]. To aid decision-making for thromboprophylaxis with OAC, several risk stratification schemes have been developed using clinical characteristics, the most popular being the CHADS₂ (congestive heart failure, hypertension, age ≥ 75 years, diabetes mellitus, and prior stroke or transient ischaemic attack) [12]. Because of several limitations [13], the CHA₂DS₂-VASc [Cardiac failure or dysfunction, Hypertension, Age ≥ 75 years (Doubled), Diabetes, Stroke (Doubled)–Vascular disease, Age 65–74 years and Sex category (Women)] has been proposed to complement the CHADS₂ score [14], reflecting a risk factor-based approach to thromboprophylaxis [15, 16].

The CHA₂DS₂-VASc score seems better at identifying the truly low risk group of patients who have an annual stroke risk < 1% [12, 14, 17]. In the view of some authors, CHA₂DS₂-VASc seems to be over-inclusive, categorising a very high percentage of subjects to be OAC eligible, being exposed to an increased risk of major bleeding [18]. The alternative would be to leave some patients untreated (or only given aspirin, which is ineffective and not any safer in terms of bleeding risk) and exposed to the risk of fatal and devastating strokes. Indeed, we recently demonstrated how CHA₂DS₂-VASc score predicted adverse events beyond thromboembolic risk in AF patients taking OAC [19].

Stroke risk is also closely related to bleeding risk, and OAC therapy needs to weigh the benefit from stroke prevention against the bleeding risk. Many thromboembolic risk factors have also been identified as bleeding risk factors (e.g., advanced age or uncontrolled hypertension) [20]. The HAS-BLED [(Hypertension, Abnormal renal/liver function, Stroke, Bleeding history or predisposition, Labile International Normalised Ratio, Elderly, Drugs/alcohol concomitantly)] has recently been proposed as a practical tool to assess the individual bleeding risk of real-world AF patients [21, 22]. The HAS-BLED score may also give important prognostic information regarding death and cardiovascular events, well as bleeding risk although HAS-BLED is a much better predictor of bleeding [23].

Clinical risk scores have shown modest capability in predicting thromboembolic events, with low values for area under the curve [14, 24, 25], with C-statistics between 0.549 and 0.638 [14]. Various studies have highlighted the incorporation of biomarkers to improve the prediction power of these scores, enhancing the identification of ‘high risk’ patients who sustain thromboembolism (Table 1) [26-29].

Table 1. Examples of studies linking biomarkers to AF prognosis in several types of events
StudyYearBiomarker evaluatedHR 95% CI (multivariate analysis, p-value)Adjusted scoreRelative IDI (%), p-value
  1. NS, No significance of the mulltivariate analysis.

End-point: stroke/TIA
Hijazi et al. [26]2012Troponin I (≥ 0.040 μg/l)1.68 (0.97–2.89), p = 0.0232CHADS₂70, p = 0.0398
CHA₂DS₂-VASc59, p = 0.0492
NT-proBNP (> 1402 ng/l)2.09 (1.22–3.58), p < 0.0001CHADS₂76, p = 0.1157
CHA₂DS₂-VASc63, p = 0.2393
Roldán et al. [29]2012hsTnT (≥ 8.04 pg/ml)2.37 (1.08–5.02), p = 0.032CHADS₂
2.44 (1.13–5.26), p = 0.023CHA₂DS₂-VASc
hsIL6 (≥ 3.35 pg/ml)1.46 (0.73–2.90), p = 0.276CHADS₂
1.45 (0.72–2.89), p = 0.295CHA₂DS₂-VASc
Hernández-Romero et al. (cardiovascular events) [75]2013Adiponectin (< 4444 ng/ml) (female patients)2.96 (1.78–4.92), p < 0.00112.8, p < 0.001
Hijazi et al. [28]2013NT-proBNP (> 1250 ng/l)2.35 (1.62–3.40), p < 0.0001CHA₂DS₂-VASc47, p < 0.0001
End-point: composite of thromboembolic outcomes
Hijazi et al. [26]2012Troponin I (≥ 0.040 μg/l)2.77 (2.06–3.72), p < 0.0001CHADS₂239, p < 0.0001
CHA₂DS₂-VASc152, p < 0.0001
NT-proBNP (> 1402 ng/l)2.79 (1.96–3.98), p < 0.0001CHADS₂211, p < 0.0001
CHA₂DS₂-VASc125, p < 0.0001
Roldán et al. [29]2012hsTnT (≥ 8.04 pg/ml)1.67 (1.09–2.56), p = 0.019CHADS₂4.4, p < 0.001
1.68 (1.10–2.57), p = 0.017CHA₂DS₂-VASc3.7, p < 0.001
hsIL6 (≥ 3.35 pg/ml)1.97 (1.29–3.02), p = 0.002CHADS₂2.2, p = 0.003
1.89 (1.23–2.89), p = 0.004CHA₂DS₂-VASc2.1, p = 0.001
Roldán et al. [56] (composite of cardiovascular events)2011vWF (≥ 221 IU/dl)2.71(1.78–4.13), p < 0.001CHADS₂12, p < 0.001
CHA₂DS₂-VASc10.3, p < 0.001
D-dimerHAS-BLED
NS
Hernández-Romero et al. [75]2013Adiponectin (< 4444 ng/ml)NS
Go et al. [82]2009Lowest eGFR (< 45 ml/min/1.73 m²) (MDRD equation)1.39 (1.13–1.71)
Proteinuria (≥ 30 mg/dl)1.54 (1.29–1.85)
Hohnloser et al. (stroke/systemic embolism) [84]2012Lowest eGFR (≤ 50 ml/min) (Cockcroft–Gault equation)0.79 (0.55–1.14), p = 0.705
Lowest eGFR (≤ 50 ml/min) (CKD-EPI equation)0.61 (0.39–0.94), p = 0.406
Lowest Cystatin-eGFR (≤ 50 ml/min)0.64 (0.39–1.05), p = 0.098
Roldán et al. [83]2013Lowest eGFR (< 30 ml/min/1.73 m²) (MDRD equation)1.42 (1.11–1.83), p = 0.006
Hijazi et al. [28]2013NT-proBNP (> 1250 ng/l)2.21 (1.73–2.82), p < 0.0001CHA₂DS₂-VASc162, p < 0.0001
End-point: mortality
Hijazi et al. [26] (Results for Vascular death)2012Troponin I (≥ 0.040 μg/l)3.20 (2.20–4.65), p < 0.0001CHADS₂End-point included in composite thromboembolic outcome
CHA₂DS₂-VASc
NT-proBNP (> 1402 ng/l)5.07 (2.95–8.71), p < 0.0001CHADS₂
CHA₂DS₂-VASc
Roldán et al. [29]2012hsTnT (≥ 8.04 pg/ml)1.79 (1.13–2.83), p < 0.001CHADS₂7.2, p < 0.001
1.99 (1.25–3.20), p = 0.004CHA₂DS₂-VASc3.1, p < 0.001
hsIL6 (≥ 3.35 pg/ml)2.48 (1.60–3.85), p < 0.001CHADS₂1.1, p = 0.025
2.19 (1.36–3.52), p = 0.001CHA₂DS₂-VASc2.7, p = 0.002
Roldán et al. [56]2011vWF (≥ 221 IU/dl)2.03 (1.24–3.32), p = 0.005CHADS₂8.5, p < 0.001
CHA₂DS₂-VASc7.1, p < 0.001
D-dimerHAS-BLED
NS
Hernández-Romero et al. [75]2013Adiponectin (< 4444 ng/ml)NS
Hermida et al. [44] hsCRP (≥ 6 mg/dl) All-cause mortality2.52 (1.49–4.25), p < 0.001CHADS₂0.147, p = 0.002
hsCRP (≥ 6 mg/dl) Cardiovascular mortality2.23 (1.09–4.57), p = 0.10CHADS₂−0.034, p = 0.40
Hohnloser et al. [84]2012Lowest eGFR (≤ 50 ml/min) (Cockcroft–Gault equation)0.86 (0.70–1.05), p = 0.627
Lowest eGFR (≤ 50 ml/min) (CKD-EPI equation)0.78 (0.63–0.96), p = 0.319
Lowest Cystatin-eGFR (≤ 50 ml/min)1.00 (0.79–1.26), p = 0.706
Roldán et al. [83]2013Lowest eGFR (< 30 ml/min/1.73 m²) (MDRD equation)1.47 (1.13–1.91), p = 0.004
Hijazi et al. [28]2013NT-proBNP (> 1250 ng/l)2.25 (1.80–2.81), p < 0.0001CHA₂DS₂-VASc270, p < 0.0001
End-point: bleeding
Hijazi et al. [26]2012Troponin I (≥ 0.040 μg/l)1.89 (1.30–2.75), p < 0.0040CHADS₂
CHA₂DS₂-VASc
NT-proBNP (> 1402 ng/l)1.28 (0.87–1.88), p < 0.5272CHADS₂
CHA₂DS₂-VASc
Roldán et al. [56]2011vWF (≥ 221 IU/dl)4.47 (1.86–10.75), p = 0.001CHADS₂
CHA₂DS₂-VASc
D-dimerHAS-BLED11, p = 0.005
NS
Hohnloser et al. [84]2012Lowest eGFR (≤ 50 ml/min) (Cockcroft–Gault equation)0.50 (0.38–0.66), p = 0.030
Lowest eGFR (≤ 50 ml/min) (CKD-EPI equation)0.48 (0.37–0.64), p = 0.004
Lowest Cystatin-eGFR (≤ 50 ml/min)0.65 (0.47–0.91), p = 0.775
Roldán et al. [83]2013Lowest eGFR (< 30 ml/min/1.73 m²) (MDRD equation)1.44 (1.08–1.94), p = 0.015
Hijazi et al. [28]2013NT-proBNP (> 1250 ng/l)1.07 (0.82–1.40), p = 0.0667CHA₂DS₂-VASc

Blood-based biomarkers, could improve our knowledge of the pathophysiology of AF (Figure 1). The study of biomarkers related to the different pathways involved in AF (myocardial injury, wall stress or inflammatory markers) could be related to clinical and echocardiographic risk factors of thromboembolism.

Figure 1.

Different pathways involved in the AF pathophysiology related to various biomarkers

The aim of this review was to summarise the published data about biomarkers studied in AF, with focus on data from randomised prospective clinical trials and large community-based cohorts. We will also review the application of these biomarkers in prognosis on the main schemes used to stratify the AF risk.

Cardiac biomarkers in AF

Myocardial injury

Cardiac troponins are contractile proteins of cardiomyocytes and are released during myocardial necrosis, and thus they are known as sensitive and specific biomarkers of myocardial injury [30]. Slight elevations in troponin levels (troponin T or I) are observed in patients with stable coronary artery disease, heart failure, and also in elderly apparently healthy individuals and have been associated with worse outcomes and increased mortality independent of conventional major coronary risk factors [30-32].

The first study that reported circulating troponin I (hsTnI) levels were associated with mortality and major adverse cardiac events in AF was in a cohort of hospitalised patients [33]. In a substudy from the RE-LY trial, raised levels of TnI could frequently be detected in patients with AF and risk factors for stroke, and TnI elevation was linked to an increased risk of stroke and mortality [26]. Moreover, risk assessment for cardiovascular death, was independently improved when TnI was added to CHADS₂ and CHA₂DS₂-VASc risk scores. These results were confirmed in the study by Roldán et al. [29], in a stable and chronic anticoagulated AF cohort, whereby increased plasma hsTnT levels were associated with an adverse prognosis in AF patients, with regard to cardiovascular events and mortality. Recently, these data were also confirmed in a substudy of the ARISTOTLE trial [27, 34].

Perhaps troponin increase (hsTnI or hsTnT) is because of AF per se, or caused by coexistent cardiovascular risk factors, or troponin may simply reflect a ‘sick heart’. Thus, there is no established explanation for the association between high troponin and stroke.

Myocyte wall stress

B-type natriuretic peptide (BNP) and the stable N-terminal portion of the prohormone, pro-BNP (NT-proBNP) are peptides synthesised by myocytes, predominantly in the left ventricle, in response to elevated wall stress. High natriuretic peptide levels correlate positively with cardiac filling pressures, making them excellent markers for abnormal LV wall stress and thus they have been proposed as a marker of LV dysfunction [35]. BNP was an independent predictor of new-onset AF in ST segment elevation myocardial infarction patients [36] elevated NT-proBNP levels independently predict an increased risk of development of AF [37].

A substudy of the RE-LY trial [26], found that NT-proBNP was predictive of thromboembolic events and cardiovascular mortality, and even after adjustment by potential confounding factors, the risk of stroke or systemic embolism was doubled to fivefold higher for cardiovascular mortality, in patients with the highest quartiles of NT-proBNP. Similar results have been shown in the ARISTOTLE trial [28], which showed improved risk stratification with NT-proBNP, doubling the risk of stroke and cardiac death.

Inflammation biomarkers in AF

Abnormal changes in systemic inflammation have been related to prothrombotic indices in AF [38]. Whether inflammation is an initiator, a consequence, or merely an association of AF is debatable and the results trying to associate the inflammatory biomarkers with AF, are inconsistent [8].

C-reactive protein (CRP), is the usual biomarker linked to inflammation and is predominantly synthesised in hepatocytes as an acute-phase reactant [39]. For example, Aviles et al. [40] demonstrated that high CRP levels predicted increased risk of developing AF and this was confirmed in other studies [41, 42]. In a small study, Conway et al. [43] reported the association between CRP and a composite outcome of stroke and death in AF. The prognostic value of CRP, to all-cause mortality and a composite of ischaemic stroke, myocardial infarction or vascular death, was displayed in a larger cohort based on the Stroke Prevention in Atrial Fibrillation III trial [41]. Another recent study [42] showed how high sensitive CRP (hsCRP), a marker of low-grade inflammation, was independently associated with AF in men, but apparently not in women, reflecting an elevate and coexisting state of coronary heart disease in their cohort of men. In a substudy of ARIC cohort, Hermida et al. [44] confirmed the results on hsCRP as a predictor of mortality with significant improvement on the CHADS₂ score by addition of this biomarker.

Other inflammation markers such as, tumour necrosis factor-α (TNF-α), interleukins, monocyte chemoattractan protein-1 (MCP-1) have also been related to AF.

Tumour necrosis factor-α, a pleiotropic proinflammatory molecule, is synthesised mainly by monocytes and macrophages, and is elevated in valvular AF [45] and persistent AF [10]. Interleukin-2 (IL2), produced by activated T lymphocytes, is associated with reduced incidence of post-operative AF [46]. Another interleukin, IL8 has also shown to be elevated in permanent AF [47]. In contrast, MMCP-1, which is a human CC chemokine, has been not associated independently with AF [48, 49].

A novel biomarker related to inflammation, osteoprotegerin has been related to cardiovascular diseases or atherosclerosis, and Nyrnes et al. [42] found a significant association with AF only on the univariate analysis, and white blood cells in the upper quartile had increased risk of AF [42].

The inflammation marker best related to AF, has been interleukin-6 (IL6), a circulating cytokine produced by monocytes, macrophages, T lymphocytes and endothelial cells that can induce a prothrombotic state [50]. We have demonstrated raised levels of IL6 in AF, which suggest the presence of an inflammatory state, although this fact appears to be related to clinical variables of the patients, rather than to the presence of AF per se [51]. Importantly, IL6 concentrations had implications for prognosis [43].

Our recent prognostic study by Roldán et al. [29] was the first to show how hsIL6 levels provided prognostic information that was complementary to clinical risk scores for prediction of long-term cardiovascular events and death, suggesting that hsIL6 and hsTnT may potentially be used to refine clinical risk stratification in AF.

Biomarkers related to prothrombotic state in AF

Atrial fibrillation provides abnormal changes in flow, evidenced by stasis in the left atrium. Moreover, abnormal changes in vessel walls include progressive atrial dilatation, endocardial denudation, and oedematous or fibroelastic infiltration of the extracellular matrix. Additionally, abnormal changes in blood constituents are well described in AF, and include haemostatic and platelet activation, as well as inflammation and growth factor changes.

These changes fulfil Virchow's triad for thrombogenesis, and are consistent with a prothrombotic or hypercoagulable state in AF [7]. Of note, abnormal concentrations of prothrombotic indices (e.g., prothrombin fragments 1 and 2 and thrombin-antithrombin complexes) are more prominent in patients with stroke who have AF [52]. Patients with atrial flutter and impaired left atrial appendage function (as shown by pulsed-wave Doppler) have increased amounts of D-dimer (DD) and β-thromboglobulin [53]. Indeed, DD (a fibrin degradation product) has also been shown to predict subsequent thromboembolic events in patients with non-valvular AF, even in those already receiving treatment with warfarin [54, 55]. However, Roldán et al. [56] did not find that DD levels in an anticoagulated AF cohort were related to prognosis, in contrast to other studies with prognostic value of DD for stroke [57, 58].

D-dimer as a marker of fibrin turnover, is essentially an index of thrombogenesis, which is raised along with clinical risk factors for thromboembolism [59]. More recent trial substudies (RE-LY or ARISTOTLE) described an association between DD levels and the risk of stroke, cardiovascular death and major bleeding outcomes independent of established risk factors including the CHADS₂ variables [27]. The risk increased with higher DD levels as evidenced by a threefold increase of stroke or systemic embolism and 3.5-fold increase for cardiovascular mortality. These results were confirmed in the ARISTOTLE substudy, which showed how DD levels at baseline, regardless OAC, were related to stroke, mortality and major bleeding [27].

Biomarkers of endothelial damage/dysfunction in AF

Abnormal haemostasis and coagulation are well described in AF and further insights into the hypercoagulable state in AF are evident [7]. Plasma levels of soluble E-selectin (sE-sel), von Willebrand factor (vWF) and soluble thrombomodulin (sTM) have been used as indexes of endothelial activation, damage/dysfunction and endothelial damage, respectively. A soluble form of thrombomodulin (sTM) is a recognised marker of endothelial dysfunction and may contribute to the hypercoagulable state in AF. Plasma sTM levels are lower in patients with persistent AF [60], but the biomarker best linked to prognosis in AF has been vWF, which is an established biomarker of endothelial damage/dysfunction with increased plasma levels observed in inflammatory and atherosclerotic vascular diseases, perhaps reflecting a damaged endothelium [61]. Furthermore, plasma vWF levels have been associated with independent risk factors for stroke (heart failure, previous stroke, age and diabetes) and stroke risk stratification schemes [43, 62]. Plasma vWf levels also refined clinical CHADS₂ risk stratification scheme for stroke and vascular events among AF patients [63].

In addition, inflammatory [58, 64] and prothrombotic [57] markers have been related to prognosis in AF patients, even in anticoagulated patients. These data were confirmed by Roldán et al. [56] whereby an increased plasma vWF levels were associated with adverse prognosis in ‘real life’ AF patients, particularly cardiovascular (mainly thrombotic) events, mortality and major bleeding. The addition of vWF as a biomarker risk factor helped to refine these clinical risk stratification schemes for stroke and bleeding.

Platelets in AF

Many abnormal changes in platelets seen in AF, could simply indicate underlying vascular comorbidities [7]. For example, Choudhury et al. [65] showed higher levels of platelet microparticles and soluble P-selectin in AF patients compared with healthy controls in sinus rhythm, but no difference was seen between patients with AF and disease-matched controls in sinus rhythm, implying the effect was more related to the underlying comorbidities. Increased amounts of β-thromboglobulin, a platelet-specific protein that indicates platelet activation and is released from α-granules during platelet aggregation and subsequent thrombus formation, have been shown in patients with both valvular and non-valvular AF compared with controls in sinus rhythm [7, 66]. Increased levels of CD62P (P-selectin) expression on platelets and platelet–leucocyte conjugates, could predispose to thrombosis and vascular events [67]. Indeed, the Rotterdam study [68] showed how plasma concentrations of soluble P-selectin were predictive of adverse clinical outcomes in elderly patients with AF. Furthermore, Hayashi et al. [69] demonstrated how acute induction of AF significantly increased the expression of P-selectin on platelets and microparticles. Other recent works illustrates the possible prothrombotic behaviour of microparticles [65, 70] in AF.

Adipokines in AF

Adipokines may be related to incident AF through several pathways, involving inflammation or through AF risk factors such as obesity and heart failure [71]. Resistin has been associated with increased insulin resistance and has proinflammatory, prohypertrophic effects [72].

Adiponectin presents anti-inflammatory, atherogenic and antihypertrophic functions [73] and both of them has been associated with multiple known risk factors for AF, including inflammation, diabetes, obesity, myocardial infarction and incident heart failure [71]. Several recent studies have linked adipokines with prognosis in AF, and one reported how high concentrations of adiponectin were related to persistent AF [74]. Another study by Rienstra et al., found higher concentrations of resistin were associated with incident AF, thus supporting the role of inflammation in AF initiation, but the relation was attenuated by adjustment for CRP. Of note, they did not detect a statistically significant association between adiponectin and incident AF [71]. In contrast, we recently reported data on adiponectin as a prognostic biomarker in AF [75].

Adiponectin levels have been related to different atherosclerotic risk factors [76, 77], and AF may be indicative of advanced atherosclerosis [78]. Our previous study found how low levels of adiponectin were independently associated with adverse cardiovascular events but only in female AF patients [75] and the lack of association in men could be because of testosterone decreasing adiponectin production [79]. Our data confirmed the importance of AF as a risk marker of atherosclerotic vascular damage and adiponectin could exert a protective role against cardiovascular diseases.

Renal function biomarkers in AF

Atrial fibrillation is usual in patients with chronic kidney disease (CKD) at different stages of severity, commonly in end renal stage, and this prevalence is increased in elderly populations [80]. The prevalence of AF increases with a decrease in the estimated glomerular filtration rate (eGFR) [81], and conversely, CKD increases the risk of thromboembolism in AF independently of other risk factors [82].

Relating renal function with final outcomes, Go et al. reported an independent stroke risk increase with reduced eGFR or if proteinuria was present [82]. Roldán et al. [83] who showed a decreased eGFR > 10 ml/min/1.73 m2 in 21% of patients, with a 1/5 of followed-up patients developing severe CKD (≤ 30 ml/min/1.73 m2). This study also showed that the presence of impaired renal function was also associated consistently with the development of adverse cardiovascular events, mortality and bleeding, even after adjusting for the CHADS₂ score. In the ARISTOTLE trial cohort, there were increased rates of stroke and bleedings using warfarin regardless the eGFR calculate by the Cockcroft–Gault and Chronic Kidney Disease Epidemiology Collaboration equations (CKD-EPI) [84].

Renal impairment constitutes a major risk factor for thromboembolic and cardiovascular events in AF [85]. Severe renal dysfunction is not included in neither of the two stroke risk stratification scores, but it has been proposed that CKD or proteinuria could be included to CHA2DS2-VASc, being the little ‘c’ letter indicating ‘chronic severe renal impairment’ [85]. Unfortunately, to properly test this hypothesis, stroke risk should be validated in non-anticoagulated populations.

The recent study published by Roldán et al., based on c-statistics and the integrated discrimination improvement, shows that adding CKD to the CHADS2 and CHA2DS2-VASc stroke risk scores did not independently improve the predictive value of current clinical risk scores [86].

A new reliable considered markers of renal function, Cystatin C and β-Trace protein, have also been considered biomarkers who reflect microvascular renal dysfunction [87, 88]. Cystatin C has also been studied related to AF in the study by Hohnloser et al. [84]. They showed how high levels of Cystatin C, when added to the eGFR equation, were associated with increased HR rates of stroke or systemic embolism, mortality and major bleeding in patients taking warfarin [84].

Genetic polymorphisms in AF

Polymorphisms associated with AF have shown relatively large risk estimates. The robustness of such estimates across populations and study designs have been recently showed in an interesting meta-analysis by Smith et al. [89]. However, limited data support the use of SNPs in risk stratification of stroke.

Several polymorphisms have been studied and highlight those related to fibrinogen and platelets [90].

The FGA T331A polymorphism (rs6050), results in a change in the α-fibrinogen gene by a threonine to alanine amino acid substitution at position 331 (T331A). This polymorphism was studied by Carter et al. [91] and showed an association with an increased mortality in patients with stroke and AF compared with those in sinus rhythm, suggesting that possession of A331 might lead to an increased susceptibility for embolisation of thrombus, possibly because of defective FXIII-dependent cross-linking.

The FGB g.4577G >A polymorphism (rs1800790), is located in the promoter region of the β-fibrinogen gene. Hyperfibrinogenemia is one important risk factor for cardiovascular disease and stroke [92] and this polymorphism has been related to coronary atherosclerotic disease [93]. Bozdemir et al. [94], found that the β-fibrinogen polymorphism was associated with the development of left atrial thrombus or spontaneous echo contrast, suggesting that β-fibrinogen g.4577G> A polymorphism could be a marker for the prediction of thromboembolism risk in patients with AF.

Marín et al. [95] reported that F13A1V34L polymorphism was independently associated with the prothrombotic and inflammatory state in AF patients, as evidenced by its association with raised tissue factor and IL6 levels, but not with platelet activation. Roldán et al. [96] also tested the possible role of Factor 7 polymorphism on AF thromboembolic risk, showing a lower prevalence of the F7 g.4727_4728ins10 (rs5742910) polymorphism amongst patients without stroke compared with patients with stroke.

Of the polymorphisms associated with platelet functionality, the integrin α2 gene (ITGA2) has had interesting results. This gene is located on chromosome 5q11.2, and the silent change in the coding region at position g.67214C>T (rs1126643) has a correlation with platelet GPIa/IIa density [97]. Thus, subjects with the T allele have an increased potential of platelet adhesion and a tendency to arterial thrombosis.

Conclusions

Biomarkers could be useful in refine clinical risk stratification scores, to identify patients at high risk for stroke and thromboembolism. Biomarkers may also increase our knowledge on AF pathogenesis. Several markers reflect the pathophysiologic process for development of AF, while others may simply be suited as markers of risk for future cardiovascular events.

Another point of view, is the possibility for a multimarker strategy that will improve better overall risk stratification, as with coronary artery disease [98] or acute coronary syndrome [99].

Also, these biomarkers could serve as indices of ongoing thrombogenesis, to test antithrombotic regimens and help decision-making on dose selection.

Funding

This work was partially supported by Sociedad Española de Cardiología, RD06/0014/039, (RECAVA) from ISCIII, Beca Cajamurcia-FFIS 2010; and PI11/00566-FEDER from ISCIII.

Acknowledgements

JA Vílchez holds a research grant ‘Río Hortega’ by the Instituto de Salud Carlos III, Madrid, Spain. D Hernández-Romero holds a postdoctoral position, ‘Sara Borrel’ grant by the Instituto de Salud Carlos III, Madrid, Spain.

Ancillary