Introduction

Atrial fibrillation (AF) is a main cause of morbidity and mortality due to the possibility of thromboembolic stroke [1–3]. In most cases, the point of origin of the thrombus is the left atrial appendage (LAA) [4]. For patients with known AF and a high thromboembolic risk a sufficient therapy with anticoagulants is of indisputable importance [5]. However, unresolved questions still remain [6, 7]. Is there a need for a routinely performed TEE before cardioversion, even in low risk patients? Could we dare to stop the therapy with anticoagulants after successful performed pulmonary vein isolation in patients with low risk [1]? Biomarkers which could predict an elevated thromboembolic risk would be helpful to answer these occasionally difficult questions. Increasing evidence suggests a fundamental role of inflammatory processes in atrial fibrillation (AF) [8–14]. AF itself [12] and inflammatory processes in general could be associated with markers of platelet activation. Monocyte–platelet aggregates (MPAs) are sensitive markers for both platelets and monocyte activation [15, 16] and could represent a link between inflammation and hemostasis [17]. Patients with heart failure turned out to have persistent high levels of MPAs in peripheral blood [18]. Heart failure is an established risk factor for thrombotic complications in patients with AF. An elevated content of MPAs could be further associated with worse outcome after stroke [19]. It is not known whether MPAs are increased in dependence on the extent of thrombogenicity in AF. Inflammatory circumstances or chronic inflammatory diseases can promote AF [10, 20–23]. It could be emerged that AF itself can promote inflammatory processes during atrial remodeling. AF could lead to a local infiltration of leukocytes into the atrial wall which is mediated by CD11b and CD18 integrins [24, 25]. Activated platelets are known to be able to induce activation with enhanced expression of adhesion molecules in circulating monocytes [26]. Whether AF in combination with prothrombotic conditions, like slow flow in the LAA or even an existing thrombus formation is associated with an additional increase of CD11b expression on monocytes is still unclear. In this study we examined differences in the content of MPAs and expression of CD11b on monocytes in patients with AF in dependence of their thrombogenicity.

Methods

Study population

107 patients with symptomatic and drug refractory AF underwent transesophageal echocardiography (TEE) before planned electric cardioversion or catheter ablation. In all patients clinical, laboratory and demographic data were obtained. Flow-cytometric quantification analyses were done on the day of performed TEE. Patients with a detected LA-Thrombus (n = 27) were included consecutively and then compared with patients with AF but without LA-thrombus. As the group of patients without LA thrombus would have been much larger than the group of patients with an LA thrombus we stopped the inclusion of patients without LA thrombus after 80 patients. Patients presenting an acute coronary syndrome, relevant infectious or inflammatory disorders were excluded. The patients with AF were further divided in three groups in dependency on their peak emptying blood flow velocity in the left atrial appendage. The study was performed in accordance with the Helsinki Declaration and approved by the institutional ethic committee of the Technische Universität Dresden (EK406122012). All of the participants gave their written, informed consent.

Echocardiography

All patients underwent transthoracic and transesophageal echocardiographic examinations before planned cardioversion or catheter ablation. LA diameter (from the parasternal long-axis view), evaluation of valvular heart diseases and LVEF (using biplane Simpson method) were obtained according to the American Society of Echocardiography [27]. Left atrial (LA) appendage peak emptying flow velocity (LAAEV) was determined in 99 of the 102 study participants. LAAEV below 20 cm/s was classified as a prothrombotic condition [28], blood flow >40 cm/s as non thrombogenic and LAAEV 20–40 cm/s as intermediate thrombogenic. LA-thrombus was defined as a distinct echogenic mass separate to the LA body which could be distinguished from the surrounding atrium. The presence of LA thrombus was confirmed by a second observer.

Laboratory methods

Peripheral venous blood samples were collected through non traumatic puncture from all of the study participants with minimal stasis into tubes containing sodium citrate (Sarstedt) as anticoagulants and analyzed by flow cytometry within 30 min of collection. Flow-cytometric quantification analysis of peripheral blood was performed in 102 patients. Shortly after, 50 µl blood were labeled within 10 min with CD45-FITC, CD14-APC, CD11b-PE and CD41-PE [all antibodies were purchased from Becton–Dickinson (BD), Oxford, UK]. CD11b-PE and CD41-PE were used in different tubes. As control for the determination of CD11b on monocytes tubes containing CD 45, CD14 except CD11b (fluorescence minus one) were used. Flow-cytometric measurements were performed using a BD FACSCalibur flow cytometer. Monocytes were identified by gating strategies based on CD45 expression and side scatter to select monocytes. The content of MPAs was determined by co-expression of CD41 and CD14 on monocytes. The quantification of MPAs was expressed as mean fluorescence intensity (MFI), as relative count and as absolute count. Relative counts of MPAs represent the percentage of monocytes with coexpression of the platelet marker CD41 on all monocytes. Absolute counts of monocytes (in cells per µl) were obtained by calculating the relative number of monocytes proportional to the number of the count in the standard laboratory blood test. This was performed with blood from the same puncture. Together with the ascertained percentage of monocytes, we similarly calculated the absolute number of MPAs per µl. The extent of activation on monocytes and granulocytes was determined by quantification of the cell adhesion molecule CD11b (MAC-1). The expression of CD11b was measured as mean fluorescence intensity (MFI). The values of fluorescence minus one (FMO) control were subtracted.

Statistical analysis

Statistical analysis was performed using SPSS v.18. The distribution of continuous data was examined using the Kolmogorov-Smirnov-test. Data are given as mean ± standard error of mean (SEM), unless otherwise stated. For comparison of group 1 (patients with AF, without LA-thrombus) with group 2 (AF and LA-thrombus) the data with a normal Gaussian distribution were analyzed using an unpaired Student’s t test after controlling for equality of variances with the Levene’s test. Data with a non-Gaussian distribution were analyzed using the Mann–Whitney U test. Cross-sectional comparisons among the three study populations (AF with LAAEV >40 cm/s, AF with LAAEV 20–40 cm/s and AF with LAAEV <20 cm/s) were performed with one-way ANOVA. A post hoc Tukey test was performed to assess intergroup differences. Correlation coefficients were calculated by Pearson tests. For logistic regression analysis we included predictors identified in the univariate test. A p value <0.05 was considered as statistically significant. Odds ratio and 95 % confidence interval were calculated. Receiver-operating characteristic (ROC) curve analyses were used for evaluating the optimal cutoff value, and the related sensitivity and specificity of the investigated parameters for predicting LA thrombus.

Results

Baseline characteristics

Relevant baseline clinical, laboratory and echocardiographic parameters of the study participants are summarized in Table 1. Not surprisingly, patients with detected LA thrombus showed an increased extent of already known risk factors like age, diabetes, coronary artery disease and impaired left ventricular function. In addition they had slightly elevated levels of leukocytes, D-dimer and IL-6. Other markers of systemic inflammation were only increased by trend in the LA thrombus group. The LA diameter was greater in patients with AF and detected thrombus. As expected, the CHA2DS2-VASc. score was elevated in the group with proven thrombus formation. No participant in the group with LA thrombus had a CHA2DS2-VASc.-score <1.

Table 1 Clinical and demographic characteristics of the study participants
Full size table

Monocyte–platelet aggregates

The total MPA count (147 ± 12 vs. 311 ± 29 cells/µl, p < 0.001), the relative proportion of MPAs on monocytes (34.9 ± 1.7 vs. 61.9 ± 4.7 %, p < 0.001) and the MFI of CD41 on monocytes (138 ± 13 vs. 384 ± 65, p = 0.001) were significantly increased in patients with AF and detected LA thrombus formation (Table 2). In a multivariate logistic regression analysis (Table 3) the count of MPAs remained an independent predictor of thrombus formation in patients with AF after adjustment of age, LVEF, LA diameter, diabetes, leukocyte count, D-dimer and the IL-6 level. In the ROC analysis, MPAs above 170 cells/µl predicted LA thrombus formation with a sensitivity of 93 % and a specificity of 73 % (Fig. 1). MPAs above 170 cells/µl had an odds ratio of 34.2 (95 % CI 3.6–324.7, p = 0.01) for predicting LA thrombus in patients with AF (Table 3).

Table 2 Markers of platelet and monocyte activation depending on detected LA thrombus formation
Full size table
Table 3 Multivariate logistic regression analyses for predictors of thrombus formation in patients with AF
Full size table
Fig. 1
figure 1

Receiver operating characteristic (ROC) curve analyses of the absolute count of MPAs for predicting LA thrombus formation. AUC indicates area under the curve. The number in parentheses indicates 95 % confidence of intervals

Full size image

CD11b expression on monocytes and granulocytes

A high expression of CD11b on all monocytes, which represents a high extent of activation, was also associated with LA thrombus compared to patients with AF but without LA thrombus formation (21.5 ± 2.6 vs. 44.4 ± 9.6 p < 0.05) (Fig. 2). The extent of CD11b expression on monocytes correlated significantly with the count of MPAs (r = 0.49, p < 0.001). In addition, a high expression of CD11b on granulocytes was seen in patients with LA thrombus (14.1 ± 1.5 vs. 29.7 ± 6.0 p < 0.05), (Fig. 2). Comparable to the MPAs, a multivariate logistic regression analysis validated both, CD11b expression on monocytes (p < 0.05) and CD11b expression on granulocytes (p < 0.01) as independent predictors for thrombus formation in patients with AF after adjustment of age, LVEF, LA-diameter, diabetes, leukocyte count, D-dimer and the IL-6 level (Table 3).

Fig. 2
figure 2

Expression of CD 11b on monocytes and granulocytes in dependence of proven LA thrombus in patients with AF. Expression of CD 11b on monocytes and granulocytes in dependence of proven LA thrombus in patients with AF. Patients with AF but without LA thrombus (no Thromb), patients with AF and attested thrombus formation (Thromb). CD11b expression is expressed as mean fluorescence intensity (MFI) as determined by flow cytometry. Results are presented as mean ± SEM. *p < 0.05

Full size image

Correlation analysis of MPAs and CD 11b expression with echocardiographic parameters for thrombogenicity

Higher counts of MPAs were associated with lower LAA peak emptying flow velocity (LAAEV). There was a significant negative correlation between the content of MPAs and LAAEV (r = −0.42; p < 0.001). Also the CD11b expression on monocytes (r = −0.33; p = 0.001) and on granulocytes (r = −0.27; p = 0.006) correlated negatively with LAAEV. Although the LA-diameter correlated negatively with LAA peak emptying flow velocity (r = −0.40; p < 0.001), both the MPA count as well as the CD11b expression was only associated by trend with the LA diameter. Between the three groups of different LAA flow velocities, those patients with LAAEV <20 cm/s had an increased content of MPAs count (128 ± 13 vs. 284 ± 32 cells/µl, p < 0.001) in comparison with the group of a non thrombogenic LAAEV of greater than 40 cm/s (Fig. 3). Also the group of an intermediate LAAEV (20–40 cm/s) had a lower MPA count in comparison to patients with a thrombogenic LAAEV of <20 (184 ± 19 vs. 284 ± 32 cells/µl, p < 0.05). Both, the CD11b expression on monocytes (17 ± 3 vs. 43 ± 9 p < 0.01) and on granulocytes (12 ± 2 vs. 28 ± 5 p < 0 0.01), were elevated in the group of LAAEV of <20 compared to patients with a non thrombogenic flow of >40 cm/s (Table 4).

Fig. 3
figure 3

Monocyte–platelet aggregates in dependence on the left atrial appendage peak emptying flow velocity. Monocyte–platelet aggregates (MPAs) in dependence on the left atrial appendage (LAA) peak emptying flow velocity (LAAEV). LAAEV below 20 cm/s is regarded as a prothrombotic condition. Results are presented as mean ± SEM. *p < 0.05, ***p < 0.001

Full size image
Table 4 Cross-sectional analysis of MPAs and CD11b expression in in dependence on LAA peak emptying flow velocity (LAAEV)
Full size table

Discussion

Our study revealed three main findings. First, the content of MPAs as a marker for platelet activation was independently increased in patients with echocardiographic proven thrombus formation in the LA. Second, the expression of CD11b on monocytes as well as on granulocytes as indicator for an enhanced inflammatory activation status of these cells were significantly elevated in patients with LA thrombus, compared to those who were in AF but without a LA thrombus. Third, both, the content of MPAs and the expression of CD11b on monocytes and granulocytes correlated negatively with LAA flow velocity.

Predictors of thrombogenicity in patients with AF

Clinical scores, with their modest informative values [29], were commonly used for predicting the thrombogenic risk. Prothrombotic [30, 31] or inflammatory markers [32, 33] have been related to prognostic parameters, like stroke or mortality in AF patients. Platelet activation markers, which are already associated with AF, like P-selectin and CD40 ligand have a short detectability in peripheral blood after activation [12, 34]. In contrast, MPAs are more stable and are described to persist in an elevated state up to several months after activation [18]. Therefore, also for not immediately evolved thrombi, they seem to have advantages for individual thromboembolic risk stratification. In contrast to the modest discriminative capability of the CHA2DS2-VASc Score with a reported value of AUC of about 0.65 [35] the level of MPAs in our study adduces a respectable AUC of 0.87 (0.81–0.94). MPAs with the revealed cut off value of 170 cells/µl showed a high sensitivity (93 %) and simultaneously an acceptable high specificity (73 %) for predicting thrombus formation. A subgroup analysis of our patients without diabetes (S1) revealed a significant elevation of MPAs and CD11b expression on monocytes and granulocytes in patients with detected LA thrombus as well. The presented negative correlation of MPAs and CD11b expression with the established echocardiographic risk marker of low flow velocity in the LAA on the one hand supports the potential of these new biomarkers for individual thromboembolic risk stratification. On the other hand it could indicate a possible causal role of the degree of intra-atrial stasis on the increase of MPAs and CD11b expression. The influence of blood stasis on activation of platelets could be shown previously by Igarashi et al. when they demonstrated a negative correlation between thrombin–antithrombin III complex and the level of D-dimer with low blood flow in the LAA [36]. In our study group the level of D-dimer was increased in patients with proven LA thrombus as well. However, in contrast to the MPAs, the D-dimer did not reach statistical significance after adjustment in the regression analysis. It has to be taking into account, that despite the relatively small number of participants in our study, the translation into clinical process may not seem to be far away. Considering the fact, that TEE, prior catheter ablation, seem to be unnecessary in patients at very low risk anyway [37], the additional evaluation with suitable markers of thrombotic risk, like the content of MPAs have the potential for more safety when TEE before cardioversion or before catheter ablation is spared or when we dare to break off the therapy with anticoagulants after a successful performed pulmonary vein isolation in patients with low risk. As thrombus formation in a low risk population is a very rare event, large randomized trials would be necessary to prove the concept in clinical life adequately.

Platelets and monocytes in the pathophysiology of thrombogenicity in AF

Acute onset of AF not only increases platelet activity, but also led to monocyte–platelet interactions [38]. In platelet thrombi, monocytes constituted about 16 % of platelet thrombus-bound leukocytes, which represent an almost fourfold enrichment as compared with their proportion in circulating blood [39]. Our presented findings of the strong association of MPAs and CD11b expression of monocytes and granulocytes with thrombus formation in patients with AF support the notion of a close link between AF, inflammation and thrombogenesis. Monocytes as well as granulocytes can activate factor X and bind fibrinogen via integrin CD11b (MAC-1) [40]. In this way they are able to modulate fibrin formation, but also thrombus dissolution. Monocytes take part in haemostatic processes via secretion of procoagulant factors and promoting inflammation [41]. Furthermore, platelets are able to activate monocytes and granulocytes which react with rapid translocation of CD11b receptors from an intracellular pool to their surface [17]. The close correlation of MPAs as markers of platelet activation with CD11b expression on monocytes among themselves in our study could demonstrate further evidence of suspected interactions of monocytes and platelets during prothrombotic processes of AF. CD11b may play a key role in the link between inflammation and AF, as the detected leukocyte atrial infiltration in mice could be shown to be dependent on CD11b expression [24]. Therefore, MPA and CD11b seem to be interesting markers in two respects. They are associated with increased thrombogenicity in AF, like our study shows. Inflammation in AF consists in particular of local atrial remodeling processes and therefore biomarkers in system circulation might not detect inflammation in AF at early stages [42]. Because of its pathophysiological impact, the extent of CD11b expression could furthermore reflect a suitable marker for the level of ongoing inflammation in AF. However, this requires further studies. Between some new treatment approaches [43, 44], there is already some evidence for possible beneficial roles of anti-inflammatory therapies for prevention of the development of AF [45] as well. Colchicine was shown to be able to reduce the rate of recurrences of AF after pulmonary vein isolation [46], whereas corticosteroids significantly reduce the risk of postoperative AF [7, 47, 48]. In conclusion, inflammatory targets for modulation of thrombotic processes seem to be worthwhile approaches for further studies.

Study limitations

Several limitations have been noted. First, the study group had a relative small sample size, especially the group with detected LA thrombus. Second, because of the existed close correlation of the content of MPAs and the CD11b expression on monocytes and the concomitant problem of multicollinearity, these parameters could not be integrated together in the logistic regression analysis. Furthermore the patients with AF had a slightly, not significant increased CRP. If the finding of the elevated CRP may influence the formation of MPAs or only reflects the inflammation in AF cannot be answered with certainty.

Conclusion

Our work shows, that the content of MPAs and the level of CD11b expression on monocytes as on granulocytes is associated with echocardiographic-proven thrombogenicity in patients with AF. These findings support the interdependency between inflammation, thrombogenesis and AF. Therefore, new anti-inflammatory treatment strategies for blocking monocyte activation, or blocking monocyte–platelet interactions might be promising therapeutic targets. Both, MPAs and CD11b expression appear appropriate for thromboembolic risk stratification in patients with AF.