Comparison of Immature Platelet Fraction and Factors Associated with Inflammation, Thrombosis and Platelet Reactivity Between Left and Right Atria in Patients with Atrial Fibrillation
Olga Perelshtein Brezinov12, Ziv Sevilya12, Ella Yahud12, Michael Rahkovich12, Yonatan Kogan12, Gergana Marincheva12, Yana Kakzanov12, Eli Lev12, Avishag Laish-Farkash12
1Department of Cardiology, Assuta Ashdod Medical Center, Ashdod.2The Faculty of Health Sciences, Ben-Gurion University of the Negev, Beer Sheva, Israel.
Recent trials found poor temporal relationship between atrial fibrillation (AF) episodes and strokes. Thus, stroke in AF patients probably involves more mechanisms than cardiac embolism. We compared factors of inflammation, thrombosis and platelet reactivity between left (LA) and right atria (RA) and femoral vein (FV) in patients with AF.
Blood samples were collected from patients undergoing AF-ablation from the FV, RA and LA for neutrophil to lymphocyte ratio (NLR), immature platelet fraction (IPF) and count (IPC), CD40 ligand, P-selectin and E-Selectin. IPF was measured by an autoanalyzer; CD40 ligand, P-selectin, and E-Selectin were measured by ELISA and NLR was calculated from complete blood counts.
Sixty-seven patients were included (age 65±10y, 63% male, CHA2DS2-VASc score 2.8±1.8, LA volume index 40±24 mL/m2, 63% paroxysmal AF). There was no difference between FV, RA and LA regarding NLR and CD40 ligand. Factors associated with platelets activity: P-selectin, IPC and IPF% were higher in RA vs LA (60.3 IQR 49.0-76.4 ng/ml vs. 59.3 IQR 49.0-74.7, respectively, p=0.03 for P-selectin, 7.5 IQR 5.2-10 103/μL vs. 7.1 IQR 5-9.8, p<0.01 for IPC, and 3.6 IQR 2.7-5.0 % vs. 3.6 IQR 2.6-4.8, p<0.01 for IPF%). Similar trends were for E-selectin (41.2 IQR 31.1-51.2 ng/mL vs. 38.7 IQR 27.9-50.4 p=0.09). Similar significant differences were found in patients with CHA2DS2-VASC≥2 but not in patients with low score.
Patients with AF, especially those with CHA2DS2-VASc≥2, have higher markers of thrombogenicity in RA compared to LA. There was no difference in inflammatory properties between the atria.
Key Words : Atrial fibrillation, Thrombosis, Inflammation, Platelet activation, P-selectin, IPF, Neutrophil to lymphocyte ratio, Left atrium.
Olga Perleshtein Brezinov
Assuta Ashdod University MC, Ashdod, Israel
Atrial fibrillation (AF) confers a 3-6 fold increase in the risk of ischemic stroke, depending on CHA2DS2-VASc score 1 . Emerging evidence shows that temporal relationship between subclinical AF and stroke occurs only in a minority of patients and that strokes often occur without subclinical AF detected within 30 days before the event 2,3 . Thus, stroke in patients with AF probably involves other mechanisms in addition to cardiac embolism. AF might simply be a marker of stroke risk — possibly indicating myocardial fibrosis or hypertrophy, thrombogenic tendency and platelet hyper-reactivity, as well as a pro-inflammatory state 3. Although in patients with AF, thrombi often originate in the left atrial appendage, anatomical considerations alone cannot completely explain the higher frequency of thrombi in the left heart chambers.
Immature or Reticulated platelets (RPs) are hyper-reactive platelets that are larger platelets with higher dense granules content and contain more RNA compared with mature platelets 4,5 . They are associated with thrombotic propensity and have a greater predilection for thrombus formation. Increased levels of RPs are associated with arterial thrombotic events including acute coronary syndrome and acute stroke 6–8. Measuring the level of RPs is technically difficult. Recently, an automated assay - immature platelet fraction (IPF) - was introduced and correlates directly with reticulated platelets level 9.
Other markers associated with platelets hyper-reactivity, thrombogenicity and inflammation are markers expressed by platelets such as the platelet glycoproteins P-selectin (CD62), CD40 ligand and E-selectin expressed by endothelial cells 10–13. P-selectin and CD40 ligand are glycoproteins translocated to the surface of platelets upon activation, and then cleaved and released into the circulation. They promote platelets activation, aggregation and thrombus formation 14,15. E-selectin is an endothelial surface molecule that acts as an adhesion molecule. E-selectin is enhanced as a result of endothelial activation. High E-selectin levels are found in patients with AF 6,13. Another marker that was recently correlated with thromboembolic events and stroke in non-valvular AF is the neutrophil-to-lymphocyte ratio (NLR) 16,17. NLR is an inflammatory marker that has been correlated with atherosclerosis and coronary cardiac events 18 and was shown to be an independent risk factor for spontaneous echo contrast of LA appendage and LA thrombus formation 17,19.
There is limited data to suggest that there is chamber specific platelet activation that could explain, in part, the propensity for LA thrombus formation in patients with AF 10,12,20. It is still unclear whether there is any difference between right and left atria regarding platelets activation.
The aim of this study is to examine whether inflammatory markers, such as NLR, C-reactive protein (CRP) and CD40 ligand, and/or pro-thrombogenic markers (such as P-selectin, E-selectin and CD-40) differ between systemic circulation and both atria or between LA and RA, thereby potentially explain the higher proportion of stroke in AF patients.
We enrolled consecutive patients with AF who underwent ablation as per European guidelines indications - pulmonary vein isolation and substrate ablation as needed 1. All patients were enrolled during the index hospitalization for the ablation. The study was approved by the ethical review board of our institution (Institutional Helsinki Board) and all subjects provided written informed consent. The study was registered in the Ministry of Health website of clinical trials.
We excluded patients with chronic hemato-oncologic diseases, anemia (hemoglobin < 10 g/dL), or thrombocytopenia (platelet level < 100 K/μL) or any other condition that could affect blood count. Any patient with acute coronary syndrome or an acute infectious or inflammatory disease was excluded as well.
Three blood samples of 5 cc were drawn from each patient during the ablation procedure: one sample from peripheral vein was taken from the femoral vein (FV); one sample was taken from the RA; and a third sample was taken from the LA immediately after trans-septal puncture and prior to the administration of intravenous Heparin. From each sample (three per patient), we measured NLR, IPF, CRP, CD40 ligand, P-selectin and E-selectin. IPF assessment was performed by automated analyzer (Sysmex XN-3000, Sysmex America Inc. Mundelein, Illinois) that uses fluorescent dyes containing polymethine and xazine. This system discriminates between mature and immature platelets and reports the immature platelet fraction. 9. Plasma level of soluble P selectin, E selectin and CD40 ligand were analyzed using enzyme-linked immuno-adsorbent assay (Human P selectin Quantkine ELISA DPSE00, Human E selectin Quantkine ELISA DSLE00 and Human CD40 Ligand Quantkine ELISA DCDL40, R&D systems, Minneapolis).
Statistical analysis of continuous variables between the groups was conducted using student paired t-test for normal distributed variables or Wilcoxon rank test for non-normal distribution as appropriate. Categorical variables were compared by a chi-square (χ2) test. Data are presented as mean ± SD for normally distributed continuous variables, as median with interquartile range (IQR) for continuous variables that are not normally distributed, and as frequency (%) for categorical variables. Statistical significance was accepted at p <0.05. All analyses were performed with the SPSS version 22 statistical software (IBM Inc. Chicago, Illinois).
Baseline patients' characteristics
Sixty-seven patients who underwent AF ablation were enrolled in this study. [Table 1] describes baseline clinical and echocardiographic characteristics. Mean age was 65.1 ± 10.3 years and 42 (62.7%) were men. Forty-two patients (62,7 %) had paroxysmal AF and 25 (37.3%) had persistent AF. Mean CHA2DS2-VASc score was 2.8± 1.8, with 19 patients (28%) having low score and inherent differences in baseline co-morbidities accordingly [Table 1].
Regarding medical treatment, patients with CHA2DS2-VASc ≥2 were treated more with statins and angiotensin converting enzyme inhibitor (ACE-I) or angiotensin II receptor blocker (ARB) than patients with CHA2DS2-VASc<2: 28(58.2%) vs. 5 (26.3%) and 36 (75.0%) vs. 6 (31.6%) respectively (p=0.02 and p=0.001). There was no difference regarding antithrombotic therapy, including aspirin, clopidogrel, ticagrelor or prasugrel. Almost all patients in our cohort - 65 (97.0%) - were treated with anticoagulation according to their CHA2DS2-VASc score. [Table 1]
Table 1. Patients' baseline characteristics
||65.1 ± 10.3
|Male Gender (%)
|Diabetes Mellitus (%)
|Smoker (%) *
|Ischemic heart disease (%) *
|Cerebrovascular disease (%) *
|Peripheral arterial disease (%) *
|Congestive Heart Failure (%) *
|Chronic renal failure (%)
||2.8 ± 1.8
||0.6 ± 0.5
|Arrhythmia Type - Paroxysmal AF (%)
Persistent AF (%)
|Aspirin (%) *
|Other anti-aggregation (Clopidogrel, Ticagrelor, Prasugrel) (%) *
|Beta Blockers (%)
|Anticoagulation (Apixaban, Rivaroxaban, Dabigatran) (%)
Ic (Propafenone, Fecanide)
III (Amiodarone, Dronaderone)
|Ejection Fraction (%)
||55.3 ± 9.1
||55.8 ± 5.8
||47.4 ± 5.3
||48.4 ± 5.3
||47.0 ± 5.3
||32.1 ± 5.6
||32.7 ± 5.0
||31.9 ± 5.9
|Diastolic Dysfunction grade – 0 (%)
|Left Atrial Diameter (mm)
||43.1 ± 5.9
||42.9 ± 6.6
||43.1 ± 5.8
|Left Atrial Area (cm2)
||24.4 ± 5.3
||22.7 ± 4.8
||25.0 ± 5.4
|Left Atrial Volume (mL)
||79.5 ± 26.4
||73.6 ± 28.8
||81.2 ± 25.8
|Left Arial Volume Index (mL/m2)
||40.1 ± 23. 9
||35.7 ± 12.5
||41.3 ± 14.2
|Right Atrial Area (cm2)
||19.2 ± 5.5
||18.9 ± 4.1
||19.3 ± 6.0
|Right Atrial Volume (mL)
||48.8 ± 23.4
||49.7 ± 16.6
||48.6 ± 25.1
|Right Arial Volume Index (mL/m2)
||24.4 ± 10.9
||25.0 ± 6.7
||24.2 ± 11.8
|Estimated Pulmonary Arterial Pressure (mmHg)
||34.7 ± 9.6
||29.7 ± 8.8
||36.3 ± 9.4
* By Fisher Exact Test
Regarding echocardiographic characteristics of the cohort: mean measurements of ejection fraction (EF) were 55.3 ± 9.1 %, LA diameter 43.1± 5.6 mm, LA volume 79.5±26.4 mL and indexed to body surface 40.1± 23.9 mL/m2. Patients with CHA2DS2-VASc ≥2 had higher estimated pulmonary artery wedge pressure (36.3 ±9.4 mmHg vs. 29.7 ±8.8, respectively; p<0.05). There was no difference in LA diameter, volume and indexed volume to body surface area [Table 1].
Platelets reactivity and inflammatory markers
Platelet activation was evaluated by soluble P-selectins, IPF and CD40 ligand levels and activated endothelium was evaluated by soluble E-selectin level [Table 2]. Soluble P-selectin mean levels were significantly higher in the RA compared to LA (60.3 IQR 49.0-76.4 ng/ml vs. 59.3 IQR 49.0-74.7 ng/ml, respectively, p=0.03). CD40 ligand levels did not differ between RA and LA. There was also a trend towards higher soluble E-selectin levels in the RA vs LA (41.2 IQR 31.1-51.2 ng/ml vs. 38.7 IQR 27.9-50.4 ng/ml, respectively, p=0.09) (Table 2). No difference was found in soluble P-selectin, E-selectin and CD40 levels between peripheral blood and both atria.
Table 2. A comparison of platelet and inflammatory markers between peripheral vein and both atria
||P value (left vs. right atria)
|Soluble P-Selectin (ng/mL)
|CD-40 ligand (pg/mL)
Abbreviations: WBC= white blood count; NLR= neutrophils to lymphocytes ratio; CRP= C reactive protein; IPF= immature platelets fraction, IPC=immature platelets count.
The level of immature platelet counts (IPC) was higher in the RA compared to LA, 7.5 IQR 5.2-10.0 103/µL vs. 7.1 IQR 5.0-9.8 103/µL (p <0.01) and 8.2±4.2 103/µL in RA vs. 7.9±4.1 103/µL in LA (p=0.05). IPF% (immature platelets fraction) was higher in the RA compared to LA as well, 3.6 IQR 2.7-5.0 % vs. 3.6 IQR 2.6-4.8 % (p<0.01) and 4.4±2.9 % in RA vs. 4.1±2.5 % in LA (p=0.04).
Inflammatory markers were evaluated by NLR and CRP [Table 2]. There was no difference between RA and LA regarding NLR. However, CRP was lower in the LA vs. RA (2.6 IQR 1.3-5.0 mg/dL, 5.2 mg/dL, p<0.05).
Platelets and inflammatory markers according to CHA2DS2-VASc score
We examined separately patients with CHA2DS2-VASc ≥2 and CHA2DS2-VASc<2 [Table 3]; patients with low CHA2DS2-VASc score had no differences in platelet activation markers such as soluble P-selectin, IPF% or soluble E-selectin [Table 3]. However, patients with CHA2DS2-VASc≥2 had higher levels of IPF% and soluble P-selectin in the RA compared with the LA (3.6 IQR 2.7-9.6 % vs. 3.5 IQR 2.7-5.2 % (p<0.01) and 69.3 IQR 48.5-79.1 ng/mL vs. 59.3. IQR 49.0-75.6 ng/mL (p<0.05), respectively). There was also a trend towards higher soluble E-selectin levels in the RA vs LA in this group of patients. No difference was found between RA and LA regarding NLR in both groups with either CHA2DS2-VASc ≥2 and CHA2DS2-VASc<2 [Table 3].
Table 3. A comparison of platelet and inflammatory markers between atria according to CHA2DS2-VASc Score
|Soluble P-Selectin (ng/mL)
|CD-40 ligand (pg/mL)
Abbreviations: WBC= white blood count; NLR= neutrophils to lymphocytes ratio; CRP= C reactive protein; IPF= immature platelets fraction. IPC=immature platelets count.
The main findings of our study are that in patients with AF, markers of platelet activation are higher in the RA compared with the LA, while most inflammatory markers do not differ between the atria. These differences are especially pronounced in patients with CHA2DS2-VASc≥2.
In this study we examined three microparticles related to platelets reactivity. P-selectin is an important protein in recruitment and aggregation of platelets and is an important marker of activated platelets 21. Soluble P-selectin is a remnant marker of activated pro-thrombotic platelets. Thus, soluble levels of P-selectin correlate with expressed P-selectin within platelet membrane 15,21. Activated platelets also release CD40 ligand. CD40/CD40 ligand activation results in expression of many pro-inflammatory and pro-thrombotic factors, including IL-1, IL6, and TNF-α and correlate with pro-thrombotic states such as cerebrovascular ischemia 21.
Another glycoprotein related to thrombogenicity of platelets is E-selectin. E-selectin is expressed on endothelial cells and is related to activated leukocytes; it is also secreted from intracellular granules and is measurable in soluble form. E-selectin has been shown to be an important factor in neutrophil trafficking and platelets recruitment and therefore has an important key role in thrombus formation. 21,22
Previous studies showed that platelets of patients with AF are more pro-thrombotic compared to non-AF patients 23 . Patients with AF have higher levels of pro-thrombotic markers such as P-selectin, CD40 and mean platelets volume 11,24–26.
Thrombogenicity of the LA was shown to be induced by AF and to return to baseline after return to sinus rhythm 27; however, the reason why thrombi are more prevalent in the LA is still an unanswered question. Temporal relationship between subclinical AF and stroke occurs only in a minority of patients and strokes often occur without subclinical AF detected within 30 days before the event 3,23. Thus, debate still exists whether thrombus formation is related only to the stunning of the atria during and post AF or there are anatomical differences between the atria that cause propensity of LA thrombus (fibrosis, hypertrophy, etc.). Possibly, there are other alternative mechanisms related to cell activation, such as in situ thrombogenic tendency or platelet hyper-reactivity or a pro-inflammatory state that is associated with both AF and stroke 10,28.
There is conflicting evidence regarding the propensity of the LA to form thrombi. Some studies imply that platelets are more activated within the LA. Unexpectedly, other studies suggest the contrary- that there are more pro-thrombotic factors, including elevated platelet activation markers, in the RA, as was shown in our study.
Yamamoto et al. have shown that the coagulation system is more activated in the LA in patients with mitral stenosis 29. Willoughby et al. reported elevated P-selectin levels and elevated ADP-induced platelet aggregation within LA compared to RA in patients who underwent pulmonary vein isolation ablation for AF 10. Lim et al. reported higher platelet activation (assessed by P-selectin levels) in patients with AF compared with a non-AF group, and an increase in thrombin generation in the LA compared with peripheral blood in patients with AF 27.
In contrast to these studies, Schultz et al. did not show any differences between P-selectin, CD40 or endothelial markers in the right and the left atria in patients with either AF or supra-ventricular tachycardias30. Jesel et al. also did not show any atrial specific differences in the levels of pro-coagulant factors, including platelet derived microparticles. They did show, however, that endothelial-derived microparticles, tissue factor activity and collagen-induced platelet aggregation were slightly elevated in the RA of patients with a history of AF 31. Park et al. also did not show any differences between right and left atria regarding pro-thrombotic factors, such as E-selectin, in patients with valvular AF 32.Additionally, Akar et al. showed in AF patients (spontaneously or following atrial pacing), that P-selectin levels were elevated in the coronary sinus compared to the peripheral blood. They also showed increased local thrombin generation in the coronary sinus, decreased nitric oxide production and no change in inflammatory markers. This study suggests increased pro-thrombotic characteristics of platelets especially in the RA, although sampling from the left side of the heart was not performed 20.
In our study we compared directly thrombogenic and inflammatory markers in RA vs. LA and also vs. peripheral blood in AF patients. Our results imply that thrombogenicity is higher in the RA than the LA in the setting of AF. We found an increase in platelet activation markers in the RA, as reflected by elevated P-selectin, E-selectin, IPC and IPF levels in the RA vs. LA, especially in patients with CHA2DS2-VASc≥2. In contrast, we did not find any significant difference between the atria in markers of inflammatory process (CD40 ligand, NLR or CRP levels).
Recently the approach to pathophysiology of AF focuses on atrial cardiomyopathy as the process leading to atrial dysfunction. The fibrillation of atrium is a symptom of this pathology. AF can affect both atria in atrial cardiomyopathy in patients with AF which is not related to rheumatic heart disease 33. Bilge et al. showed that atrial thrombi are formed in the right atrium as well as in the left side in patients with non-valvular AF, as compared with valvular AF 34. Other echocardiographic studies have also shown thrombi formation in the RA as well as in LA 32-35. In a study by Shahin et al. RA dysfunction was demonstrated in addition to left sided atrial dysfunction 35.
Another explanation could stem from the role of the lungs in this process: However, because the lungs are capable of absorbing microthrombi without any clinical significance, as opposed to the left system, the resulting neurological and peripheral consequence of the left side emboli are well expressed 36–39. This explanation suggests that thrombi are formed in the right side of the heart as well as in the left side, with higher thrombogenicity in the RA, but with less clinical consequences. This is probably the reason for the higher soluble P-selectin in RA versus LA found in our study; soluble P-selectin is created by proteolytic degradation of P-selectin on platelets. It was higher in RA either due to secretion from higher IPF in RA (as stated above) or due to more protein sequestration in the lungs. Because of the scarce data available regarding the function and the role of RA in patients with AF, the data from our study that suggest that RA has thrombogenic properties, has a novel additive value to understanding this controversial subject.
Platelets are removed from the circulation in the reticuloendothelial system after a short lifespan of 8-10 days. Much information exists regarding mechanisms of platelets activation. But less is known about the clearance of cells and their death 40. Platelets undergo a process that is similar to intrinsic apoptosis in the liver through Ashwell-Morrell receptor (AMR) 41. Therefore, following this process of apoptosis the venous blood return from the liver likely contains more young platelets. The higher levels of IPF and IPC that we have found in the RA probably reflect the venous blood circulation after the apoptosis process of ‘old’ platelets, contributing to the higher proportion of immature platelets in the RA. According to this theory, it is not surprising to find more young platelets in the RA and pulmonary vasculature.
In this study we found a small difference towards higher level of CRP in the RA compared to LA. CRP is produced in the liver and is distributed to the rest of the circulation through the venous return to the heart 42. This is probably the reason for the difference between the two atria. In our opinion, this difference is too small to have any clinical significance.
One of the limitations of the study is the non-relevance of the FV blood samples regarding thrombogenic factors. Given the sheaths (7F) were already introduced and blood was collected from the FV through the sheaths, it is possible that some of the observed values of the FV are due to introduction of potentially thrombogenic material into the vasculature. Another limitation of the study is the lack of a control group of non-AF patients, in order to better understand the role of AF in these LA-RA differences. However, the above-mentioned previous studies did not find any differences between the atria in non-AF patients 30. Another limitation is a non-uniform rhythm of the cohort during AF ablation: some patients underwent the procedure in sinus rhythm, others were in AF, and some patients started the procedure in AF and converted to sinus.
We found higher markers of thrombogenicity and platelet activation in the right compared to the left atrium in patients with AF. The results were even more pronounced in patients with CHA2DS2-VASc≥2. There does not appear to be a gradient of inflammatory properties between the two atria in these patients. These findings set the basis for further investigation to explain the higher stroke risk in AF patients.