A

C

Fpp

Fl

cP

1d

rchives of Cardiovascular Disease (2010) 103, 394—403

LINICAL RESEARCH

actors influencing the level of circulatingrocoagulant microparticles in acuteulmonary embolism

acteurs infuencant le niveau des microparticules procaogulantes dans’embolie pulmonaire

Laurence Bala, Stéphane Ederhya,Emanuele Di Angelantonioa, Florence Totib,Fatiha Zobairi b, Ghislaine Dufaitrea,Catherine Meulemana, Ziad Mallatc, Franck Boccaraa,Alain Tedguic, Jean-Marie Freyssinetb, Ariel Cohena,∗

a Cardiology Department, Saint-Antoine University and Medical School, AssistancePublique—Hôpitaux de Paris, université Pierre-et-Marie-Curie, 184, rue duFaubourg-St-Antoine, 75571 Paris cedex 12, Franceb Inserm U770, hôpital de Bicêtre, université Louis-Pasteur, faculté de médecine, institutd’hématologie et immunologie, 67400 Strasbourg, Francec Inserm U970, Paris Cardiovascular Research Centre, université Paris-Descartes andAssistance Publique—Hôpitaux de Paris, 75015 Paris, France

Received 12 May 2010; received in revised form 15 June 2010; accepted 17 June 2010Available online 13 August 2010

KEYWORDSMicroparticles;Pulmonary embolism;Cardiovascular risk

SummaryBackground. — Flow cytometry has shown levels of platelet-derived microparticles (PMPs) andendothelial-derived microparticles (EMPs) to be elevated in deep-vein thrombosis. Cardiovas-cular risk factors can also contribute to hypercoagulability due to circulating procoagulant

factors microparticles (CPMPs).Aims. — To investigate in a case-control study the respective contribution of pulmonaryembolism and cardiovascular risk factors to the level of hypercoagulability due to CPMPs.Methods. — CPMP, PMP and EMP levels were measured in 45 consecutive patients (age67.9 ± 11.6 years; 66.7% men) admitted to an intensive care unit for acute pulmonary embolism(APE), 45 healthy control subjects with no history of venous thromboembolism or vascular risk

Abbreviations: APE, acute pulmonary embolism; ControlsCVRFs, controls with cardiovascular risk factors; ControlsnoCVRFs, controls withoutardiovascular risk factors; CPMP, circulating procoagulant microparticle; EMP, endothelial-derived microparticle; IQR, interquartile range;MP, platelet-derived microparticle; VTE, venous thromboembolism.∗ Corresponding author. Fax: +33 1 49 28 28 84.

E-mail address: ariel.cohen@sat.aphp.fr (A. Cohen).

875-2136/$ — see front matter © 2010 Published by Elsevier Masson SAS.oi:10.1016/j.acvd.2010.06.005

Circulating procoagulant microparticles in pulmonary embolism 395

factors (ControlsnoCVRFs), and 45 patients with cardiovascular risk factors (ControlsCVRFs). APEwas diagnosed by spiral computed tomography or scintigraphy. CPMP levels were assessedusing a prothrombinase assay on platelet-depleted plasma (results expressed as nmol/Lequivalent).Results. — CPMP levels were higher in APE patients than in ControlsnoCVRFs (medians 4.7 vs3.2 nmol/L, interquartile ranges [IQRs] 2.9—11.1 vs 2.3—4.6 nmol/L; p = 0.02). Similar resultswere reported for PMPs (medians 2.2 vs 1.9 nmol/L, IQRs 1.7—5.8 vs 1.4—2.4 nmol/L; p = 0.02),whereas EMP levels were not significantly different. However, CPMP procoagulant activity wasnot significantly different in APE patients and ControlsCVRFs.Conclusions. — CPMPs and PMPs were significantly elevated in APE patients vs ControlsnoCVRFs,but this correlation was not significant when APE patients were compared with ControlsCVRFs.Our observations highlight the importance of adjusting for the presence of cardiovascular riskfactors in conditions in which microparticle levels are raised.© 2010 Published by Elsevier Masson SAS.

MOTS CLÉSMicroparticules ;Embolie pulmonaire ;Facteurs de risquecardiovasculaires

RésuméContexte. — La thrombose veineuse profonde est associée à une augmentation du niveau demicroparticules (MP) d’origine endothéliale et plaquettaire mesurée en cytométrie de flux. Lesfacteurs de risque cardiovasculaires (FRCV) ont aussi une influence importante sur le niveaudes microparticules.Objectifs. — Nous avons évalué le niveau des microparticles procoagulantes circulantes chezdes patients admis pour embolie pulmonaire (EP) aiguë et étudié le rôle respectif de la mal-adie veineuse thrombo-embolique et des facteurs de risque cardiovasculaire sur le niveaud’hypercoagulabilité lié aux microparticules.Méthodes. — Les microparticules procoagulantes circulantes, les microparticules plaquettaireset d’origine endothéliale ont été mesurées chez 45 patients consécutifs (âge 67,9 ± 11,6,66,7 % d’homme) admis en unité de soins intensifs pour une embolie pulmonaire aiguë, chez45 patients sans facteur de risque cardiovasculaire et chez 45 patients avec des facteurs derisque cardiovasculaire. L’embolie pulmonaire était documentée soit par angioscanner soit parscintigraphie pulmonaire de ventilation et perfusion. L’activité procoagulante des MP circu-lantes a été mesurée en utilisant un plasma pauvre en plaquettes et un test fonctionnel à laprothrombinase.Résultats. — Les microparticules procoagulantes circulantes étaient plus élevées chez lespatients admis pour une EP (médiane 4,7 nmol/L, interquartile range [IQR] 2,9—11,1) que chezles patients sans FRCV (médiane 3,2 nmol/L, IQR 2.3—4.6 ; p = 0,01). Le niveau des MP d’origineplaquettaire était plus élevé chez les patients présentant une EP comparativement aux patientssans FRCV (médiane 2,2 nmol/L, IQR 1,7—5,8 versus 1,9, 1,4—2,4 ; p = 0,02). Le niveau des MPd’origine endothéliale était, en revanche, comparable dans les deux populations. Cependant,le niveau des MP procoagulantes n’était pas significativement différent des patients avec FRCV,et ce, quel que soit le phénotype considéré.Conclusion. — Les MP procoagulantes totales et plaquettaires sont significativement plusélevées chez les patients admis pour embolie pulmonaire aiguë comparativement à des patientssans facteur de risque cardiovasculaire. Cette relation n’est plus retrouvée lorsque ces patientssont comparés à des sujets témoins avec facteurs de risque cardiovasculaire. Ces donnéesdémontrent l’importance de prendre en compte les facteurs de risque cardiovasculaire dans

es Masso

itfhH

l’interprétation du niveau d© 2010 Publie par Elsevier M

Introduction

According to Virchow’s triad, the pathophysiology ofVTE relies on the presence of blood hypercoagula-bility, stasis of blood flow and vessel-wall damage[1]. The factors involved in venous thrombogenesis

could be defined as soluble coagulant factors, dys-functional endothelium and circulating cells, especiallyplatelets, lymphomonocytes and, potentially, CPMPs[2,3].

ncd[

P.n SAS.

CPMPs are plasma membrane fragments that are releasednto the blood by stimulated cells during activation or apop-osis, and carry procoagulant phosphatidylserine and tissueactor on their surface [4—7]. CPMPs circulate in healthyumans and support low-grade generation of thrombin [8].igh levels of CPMPs have been reported in several sce-

arios, including, for example, in patients with an acuteoronary syndrome [9—11] or atrial fibrillation [12]. Car-iovascular risk factors such as hypertension and diabetes13—15] have been associated with elevated CPMPs and their

3

pgtf

eieassctrtirbcIliac

M

S

B4fs(vaowwppi

AAvTbhpasst

otaHs

hAt

CCoiou

CCivcmrt

C

Bosp

PmAat1pftkkac

QmA(ocw0wtw

96

henotypes. In interpreting the role of CPMPs in thrombo-enesis, we need to take into account potential cofounders,he most prevalent of which appear to be cardiovascular riskactors.

Two previous clinical studies have reported elevated lev-ls of CPMPs — mainly PMPs and EMPs — by flow cytometry,n patients hospitalized for deep-vein thrombosis. How-ver, they had different phenotypic representations, and theuthors did not take into account potential confounders,uch as cardiovascular risk factors [16,17]. Experimentaltudies in vivo are supportive of the involvement of humanell-derived microparticles in venous thrombogenesis in aissue factor-dependent manner, and have described a cor-elation of leukocyte- and PMPs with thrombus weight andissue factor activity [18,19]. Recently, the crucial partic-pation of circulating tissue factor-bearing microparticleseleased by tumour cells in cancer-associated hypercoagula-ility has been emphasized, depending on both the tumourell origin and a critical threshold of microparticles [20—24].n addition, we have shown in a case-control study that CPMPevels, defined by their procoagulant activity, were elevatedn patients without cancer hospitalized with APE, and wenalysed the influence of cardiovascular risk factors on thisorrelation [25].

ethods

tudy subjects

etween November 2004 and December 2005, we included5 consecutive patients admitted to our intensive care unitor APE associated with or without deep-vein thrombo-is, and compared them with 45 healthy controls withoutControlsnoCVRFs) and 45 patients with (ControlsCVRFs) cardio-ascular risk factors, matched for age and sex. ControlsCVRFs

nd patients with APE were also matched for the presencef hypertension. Demographic and clinical characteristicsere recorded prospectively upon enrolment. The studyas approved by the institutional review board and waserformed in accordance with institutional guidelines. Allatients gave written informed consent before participatingn the study.

cute pulmonary embolism casesPE was confirmed by spiral computed tomography (n = 25),entilation-perfusion scintigraphy (n = 20) or both (n = 12).reatment on admission consisted of standard antithrom-otic therapy with low-molecular-weight or unfractionatedeparin. Exclusion criteria were conditions known or sus-ected to increase levels of CPMPs independently, suchs acute coronary syndromes, acute heart failure, stroke,epsis, chronic inflammatory disease, antiphospholipidyndrome, heparin-induced thrombocytopenia, thrombotichrombocytopenic purpura and atrial fibrillation.

Transient VTE risk factors were defined as pregnancy,

estrogen therapy, surgery (< 60 days), trauma, confinedo bed (> 5 days) and recent journey (> 10 hours). Cancernd thrombophilia were defined as chronic VTE risk factors.aemodynamic status was considered over three levels:ubmassive APE (stable haemodynamics with signs of right

tVCwT

L. Bal et al.

eart failure on transthoracic echocardiography), massivePE (unstable haemodynamics with right heart failure onransthoracic echocardiography) and shock.

ontrols with cardiovascular risk factorsontrolsCVRFs comprised patients with no history of VTEr atrial fibrillation who were undergoing routine screen-ng physical examinations for cardiac symptoms at ourutpatient cardiology clinic, with an electrocardiogram doc-menting sinus rhythm.

ontrols without cardiovascular risk factorsontrolsnoCVRF included patients undergoing screening exam-

nation before orthopaedic surgery, with no known cardio-ascular risk factors, history of atrial fibrillation, prior VTE,linical evidence of disease or current cardiovascular treat-ent, and who had an electrocardiogram documenting sinus

hythm. These subjects were assessed by careful examina-ion of their medical histories and by blood tests.

irculating procoagulant microparticles

lood was collected in the acute phase when the diagnosisf VTE was assessed and just before anticoagulation wastarted. Measurement of CPMPs was performed as describedreviously [26], with minor modifications.

reparation of circulating procoagulanticroparticle samples

ll microparticle determinations were performed strictlyccording to Biro et al. [18]. Briefly, citrated blood wasaken soon after admission and centrifuged at 1500 g for5 min at room temperature within the hour after sam-ling. The supernatant was centrifuged again at 13,000 gor 2 min to avoid platelet contamination. Thrombin and fac-or Xa inhibitors (d-phenylalanyl-prolyl-arginyl chloromethyletone and 1,5-dansyl-glutamyl-glycyl-arginyl chloromethyletone, respectively) were added to plasma samples atfinal concentration of 50 �M each, and CaCl2 at a final

oncentration of 50 mM.

uantitation of circulating procoagulanticroparticles

fter capture of microparticles onto annexin V-coated wellsfor 30 min at 37 ◦C), taking advantage of the strong affinityf annexin V for aminophospholipids present in microparti-les at the calcium concentration used, four washing stepsere performed with Tris buffer containing 1 mM CaCl2 and.05% Tween 20, each for 5 min at 20 ◦C, and the last oneithout Tween. The phosphatidylserine content of micropar-

icles, directly responsible for their procoagulant activity,as then measured in a prothrombinase assay. Micropar-

icles were incubated with factor Xa (50 pmol/L), factora (360 pmol/L), prothrombin (1.3 �mol/L) and 2.3 mmol/LaCl2 for 15 min at 37 ◦C, and linear absorbance changesere recorded at 405 nm after the addition of chromozymH (380 �mol/L).

ism 397

Fcv(

ds(fnp

cacfccao

Ap

Ai23s4VC

sp(ta

P

Circulating procoagulant microparticles in pulmonary embol

Quantitation of platelet-derived microparticles andendothelial-derived microparticlesAfter specific capture of PMPs onto anti-glycoprotein Ibantibody-coated wells and of EMPs onto anti-CD31 antibody-coated wells, quantitation was achieved after severalwashing steps using a prothrombinase assay as describedabove. Microparticle levels are expressed as nmol/L of phos-phatidylserine equivalent.

Miscellaneous measurements

Quantification of C-reactive protein was determined byimmunonephelometric tests and circulating brain natri-uretic peptide levels by enzyme immunoassays.

Transthoracic echocardiography

To evaluate right ventricular dysfunction and haemodynamicstatus, transthoracic echocardiography was performed atthe time of admission in all patients with APE. Systolictranstricuspid pressure gradient and left ventricular ejec-tion fraction were also measured.

Statistical analysis

Based on previous studies [11,27,28], we hypothesized thatpatients with APE would have microparticle levels increasedby approximately two standard deviations compared withhealthy controls and by one standard deviation comparedwith subjects without APE but with cardiovascular risk fac-tors. To achieve this with 90% power and p < 0.05 betweenthe three groups, 35 subjects per group were required. Tominimize the risk of a type II error and to account for pos-sible confounders, we recruited in excess of this number ofpatients with APE and controls.

Categorical variables, expressed as percentages, werecompared using the Chi2 test or Fisher’s exact test. After atest for normality, continuous data are expressed as meansand standard deviations or medians with IQRs as appro-priate. Differences between patients and controls wereevaluated using the two-sample t test or the Mann-WhitneyU test. Correlations between annexin V-positive microparti-cles and endothelial and platelet microparticle levels wereevaluated using Spearman’s rank correlation coefficients.

The relationship between CPMP levels and patients’characteristics was estimated using linear regression anal-ysis (after logarithmic transformation of the dependentvariables) and presented using the estimated regressioncoefficient, expressed as percentage increase in micropar-ticle level for the presence of each risk factor or for a unitincrease in continuous variables, such as age. All analyseswere performed using STATA 9 statistical software (STATA,College Station, TX, USA). A probability value of 0.05 wasconsidered statistically significant.

Results

The baseline characteristics of the three groups are given inTable 1. The mean age of patients with APE was 67.9 ± 11.6years and 66.7% were men. Deep-vein thrombosis wasdocumented in 71.1% of patients, 26.7% (n = 12) had haemo-

p

PwC

igure 1. Levels of circulating procoagulant microparticles inontrols with (ControlsCVRFs) and without (ControlsnoCVRFs) cardio-ascular risk factors and in patients with acute pulmonary embolismAPE).

ynamic instability and 4.4% (n = 2) presented in cardiogenichock. APE was submassive in 17 (37.8%) patients. Sixteen35.5%) patients with APE had at least one transient VTE riskactor and eight (17.8%) had a permanent risk factor (cancer,= 3; thrombophilia, n = 5). APE was idiopathic in 21 (46.7%)atients.

There were no significant differences in terms of clini-al cardiovascular risk factors between patients with APEnd ControlsCVRFs, except for current smoking (Table 1). Inontrast, by design, ControlsnoCVRFs were significantly dif-erent from the other two groups with regard to clinicalardiovascular risk factors. However, there was no signifi-ant difference regarding age and sex between ControlsCVRFs

nd ControlsnoCVRF, even if there was a slight predominancef men in the former.

nnexin-positive microparticles in acuteulmonary embolism patients and controls

nnexin V-positive microparticle levels were highern patients with APE (median 4.7 nmol/L, IQR.9—11.1 nmol/L) than in ControlsnoCVRFs (median.2 nmol/L, IQR 2.3—4.6 nmol/L; p = 0.02), but there was noignificant difference compared with ControlsCVRFs(median.9 nmol/L, IQR 3.7—8.4 nmol/L; p = 0.99) (Fig. 1). Annexin-positive microparticle levels were significantly higher inontrolsCVRFs than in ControlsnoCVRFs (p = 0.01; Fig. 1).

Moreover, after adjustment for age, sex and hyperten-ion, CPMPs were significantly correlated with C-reactiverotein (p = 0.02) and brain natriuretic peptide levelsp < 0.01), but not with the echographic haemodynamic sta-us evaluated by systolic transtricuspid pressure gradientnd left ventricular ejection fraction (Table 2).

latelet-derived microparticles in acute

ulmonary embolism patients and controls

MP levels were not significantly different between patientsith APE (median 2.2 nmol/L, IQR 1.7—5.8 nmol/L) andontrolsCVRFs (median 5.5 nmol/L, IQR 2.6—11.3 nmol/L;

398L.

Baletal.

Table 1 Baseline characteristics in patients with acute pulmonary embolism and controls with and without cardiovascular risk factors.

Patient characteristics APE (n = 45) ControlsCVRFs

(n = 45)ControlsnoCVRFs

(n = 45)P value

APE vsControlsCVRFs

APE vsControlsnoCVRFs

ControlsCVRFs vsControlsnoCVRFs

Age (years) 67.9 ± 11.6 67.1 ± 9.9 67.0 ± 9.5 — — —Men 30 (66.7) 30 (66.7) 26 (57.8) — 0.38 0.38Hypertension 25 (55.6) 25 (55.6) 0 — — —Diabetes mellitus 7 (15.6) 9 (20.0) 0 0.58

< 0.01 < 0.01Hypercholesterolaemia 19 (42.2) 14 (31.1) 0 0.27 — —Current smoker 4 (9.1) 16 (35.6) 0

< 0.010.04 —

Coronary artery disease 3 (6.8) 9 (20.0) 0 0.07 0.08< 0.01

History of heart failure 1 (2.2) 0 0 — — —History of TIA or ischaemic stroke 2 (4.4) 1 (2.2) 0 — — —History of atrial fibrillation 6 (13.3) 0 0 — — —

Concomitant treatmentAspirin 8 (17.8) 1 (2.2) — 0.01 — —Beta-blocker 9 (20.5) 10 (22.2) — 0.84 — —Calcium channel blocker 6 (13.3) 4 (8.9) — 0.50 — —Angiotensin-converting enzyme inhibitor 7 (15.9) 7 (15.6) — 0.96 — —Diuretic 8 (17.8) 0 —

< 0.01— —

Nitrate 1 (2.2) 1 (2.2) — — — —Insulin 0 2 (4.4) — — —Oral antidiabetic therapy 2 (4.4) 6 (13.3) — 0.14 — —

Data are mean ± standard deviation or number (%).APE: acute pulmonary embolism; ControlsCVRFs: controls with cardiovascular risk factors; ControlsnoCVRFs: controls without cardiovascular risk factors; TIA: transient ischaemic attack.

Circulating procoagulant microparticles in pulmonary embolism 399

Table 2 Baseline correlates of annexin V-positive microparticles and platelet and endothelial microparticle levels in 45patients with an acute pulmonary embolism.

No. of patients (%) Pearson correlation(95% CI)

% change in MPs(95% CI)

P value

Annexin V-positive MPsClinical presentation

Acute DVT 32 (71.1) 0.07 (−0.02, 0.36) 1.18 (0.59, 2.36) 0.64

Baseline characteristicsAgea 67.9 (11.6) −0.16 (−0.43, 0.14) 0.99 (0.96, 1.01) 0.30Hypertension 25 (55.6) 0.02 (−0.27, 0.31) 1.05 (0.56, 1.97) 0.89Diabetes mellitus 7 (15.6) −0.14 (−0.42, 0.16) 0.66 (0.28, 1.56) 0.35Hypercholesterolaemia 19 (42.2) 0.05 (−0.24, 0.34) 1.12 (0.59, 2.12) 0.73Current smoker 4 (9.1) 0.26 (−0.04, 0.52) 2.57 (0.89, 7.42) 0.09Coronary artery disease 3 (6.8) 0.02 (0.28, 0.32) 1.10 (0.31, 3.93) —Previous atrial fibrillation 6 (13.3) 0.23 (−0.07, 0.49) 2.02 (0.82, 4.06) 0.13Previous DVT 10 (22.2) 0.16 (−0.15, 0.43) 1.48 (0.70, 3.13) 0.31Aspirin 8 (17.8) 0.02 (−0.27, 0.31) 1.06 (0.47, 2.4) 0.89

Biological criteriaC-reactive proteina 59.2 (83.7) −0.36 (−0.59, −0.07) 0.995 (0.99, 0.999) 0.02BNPa,b 279.6 (413.1) 0.45 (0.13, 0.69) 1.001 (1, 1.00) < 0.01

Echographic criteriaLVEF < 40% 3 (6.7) 0.17 (−0.13, 0.44) 2.02 (0.58, 7.01) —STPGa 135 (307.8) 0.08 (−0.23, 0.37) 1.14 (0.69, 1.89) 0.62

Platelet-derived MPs (anti-GP1b)Clinical presentation

Shock — 0.37 (0.09, 0.60) 8.17 (1.70, 39.36) —Acute DVT — 0.16 (−0.14, 0.43) 1.52 (0.71, 3.24) 0.29

Baseline characteristicsAgea — −0.08 (−0.36, 0.22) 0.99 (0.96, 1.02) 0.61Hypertension — 0.15 (−0.15, 0.42) 1.42 (0.71, 2.84) 0.33Diabetes mellitus — −0.19 (−0.46, 0.11) 0.54 (0.21, 1.39) 0.21Hypercholesterolaemia — 0.22 (−0.08, 0.48) 1.67 (0.84, 3.33) 0.15Current smoker — 0.23 (−0.07, 0.49) 2.55 (0.77, 8.43) 0.13Coronary artery disease — 0.10 (−0.20, 0.39) 1.59 (0.39, 6.44) —Previous atrial fibrillation — 0.28 (−0.02, 0.53) 2.57 (0.96, 6.89) 0.07Previous DVT — 0.17 (−0.13, 0.44) 1.61 (0.70, 3.68) 0.27Aspirin — 0.14 (−0.16, 0.42) 1.55 (0.63, 3.82) 0.35

Biological criteriaC-reactive proteina — −0.16 (−0.43, 0.14) 1.0 (0.99, 1.00) 0.30BNPa,b — 0.46 (0.15, 0.69) 1.00 (1, 1.00) < 0.01Ln MPs 1.62 (1.06) 0.88 (0.78, 0.93) —

Echographic criteriaLVEF < 40% — 0.09 (−0.21, 0.38) 1.54 (0.38, 6.19) —STPGa — 0.21 (−0.09, 0.48) 1.47 (0.85, 2.54) 0.17

Endothelial-derived MPs (anti-CD31)Clinical presentation

Shock — 0.28 (−0.02 , 0.53) 8.37 (0.91, 77.2) —Acute DVT — −0.03 (−0.32, 0.27) 0.91 (0.32, 2.61) 0.87

Baseline characteristicsAgea — 0.02 (−0.27, 0.32) 1.00 (0.96, 1.05) 0.88Hypertension — 0.08 (−0.22, 0.36) 1.27 (0.49, 3.31) 0.62Diabetes mellitus — −0.06 (−0.35, 0.24) 0.77 (0.21, 2.86) 0.70Hypercholesterolaemia — 0.08 (−0.22, 0.37) 1.30 (0.50, 3.49) 0.60Current smoker — 0.22 (−0.08, 0.49) 3.30 (0.66, 16.42) 0.15Coronary artery disease — −0.04 (−0.33, 0.26) 0.79 (0.12, 5.31) —Previous atrial fibrillation — 0.19 (−0.11, 0.46) 2.44 (0.62, 9.65) 0.21

400 L. Bal et al.

Table 2 (Suite )

No. of patients (%) Pearson correlation(95% CI)

% change in MPs(95% CI)

P value

Previous DVT — 0.13 (−0.17, 0.41) 1.62 (0.52, 5.05) 0.41Aspirin — 0.02 (−0.28, 0.31) 1.08 (0.31, 3.75) 0.90

Biological criteriaC-reactive proteina — −0.12 (−0.40, 0.18) 0.998 (0.99, 1.00) 0.44BNPa,b — 0.38 (0.05, 0.64) 1.00 (1, 1.00) 0.03Ln MPs 1.62 (1.06) 0.53 (0.28, 0.72)

Echographic criteriaLVEF < 40% — 0.09 (−0.21, 0.37) 1.76 (0.26, 11.78) —STPGa — 0.24 (−0.07, 0.50) 1.84 (0.86, 3.91) 0.12

a Mean.b n = 34.BNP: brain natriuretic peptide; CI: confidence interval; DVT: deep-vein thrombosis; GP: glycoprotein; Ln: natural logarithm; LVEF: leftventricular ejection fraction; MP: microparticle; STPG: systolic transtricuspid pressure gradient.

p = 0.08) (Fig. 2). PMP levels were significantly higher inpatients with APE compared with ControlsnoCVRFs (median1.9 nmol/L, IQR 1.4—2.4 nmol/L of phosphatidylserineequivalent; p = 0.02) and in ControlsCVRFs compared withControlsnoCVRFs (p < 0.001).

After adjustment for age, sex and hypertension, PMPswere associated with brain natriuretic peptide (p = 0.006)but not with C-reactive protein levels (p = 0.302). Neitherpulmonary arterial pressure nor left ventricular systolic dys-function were correlated with PMPs levels (Table 2).

Endothelial-derived microparticles in acutepulmonary embolism patients and controls

EMP levels were not significantly different between patientswith APE (median 0.1 nmol/L, IQR 0.03—0.2 nmol/L) and

Figure 2. Levels of platelet-derived microparticles (PMP) incontrols with (ControlsCVRFs) and without (ControlsnoCVRFs) cardio-vascular risk factors and in patients with acute pulmonary embolism(APE).

Figure 3. Levels of endothelial-derived microparticles (EMP) incv(

CpA0n

w(

Bp

R(ahri

ontrols with (ControlsCVRFs) and without (ControlsnoCVRFs) cardio-ascular risk factors and in patients with acute pulmonary embolismAPE).

ontrolsCVRFs (median 0.2 nmol/L, IQR 0.1—0.2 nmol/L;= 0.44) (Fig. 3). EMP levels were not higher in patients withPE compared with ControlsnoCVRFs (median 0.1 nmol/L, IQR.01—0.1 nmol/L; p = 0.54). EMP levels in ControlsCVRFs wereot significantly higher than in ControlsnoCVRFs (p = 0.06).

After adjustment for age, sex and hypertension, EMPsere only correlated with brain natriuretic peptide levels

p = 0.026) (Table 2).

aseline characteristics and circulatingrocoagulant microparticles

elations between other baseline characteristics and CPMPsannexin V-positive, PMPs and EMPs) were investigated

mong patients with APE after adjustment for age, sex andypertension (Table 2). Current smoking, significantly rep-esented in ControlsCVRFs, (Table 1; p = 0.003), appeared tonfluence CPMP and PMP levels among patients with APE

ism

li

Rat

Otw

wifbcmlotab

icdaaaicsmoRftTbuiqau

Mv

Boutc

oC

Circulating procoagulant microparticles in pulmonary embol

(change in microparticles 2.57%, p = 0.09 and 2.55%, p = 0.13,respectively).

Discussion

To our knowledge, this is the first study to compare CPMPlevels and their phenotypes (using a functional assay basedon prothrombinase) in APE patients. Comparing APE patientswith two different control groups, our study design allowedus to analyse the relative effect of cardiovascular risk fac-tors and APE on microparticle levels. In APE patients, CPMPprocoagulant activity is significantly increased comparedwith the physiological status represented by a populationwith no prior thrombotic events or cardiovascular risk fac-tors (p = 0.02). Nevertheless, procoagulant activity relatedto CPMPs in APE patients was not significantly different tothat in ControlsCVRFs.

Microparticles and venous thromboembolism

VTE results from an imbalance between procoagulant,anticoagulant and fibrinolytic activities. Platelet andendothelial microparticles are considered markers of ongo-ing or recent endothelial cell and platelet activation,or apoptosis. Microparticle procoagulant activity, mainlyrelated to tissue factor activity in the presence of phos-phatidylserine phospholipids, is dependent on the cellularorigin, the initial stimulus and the secondary microparticle-induced cell activation [4,18,29]. Experimental animalstudies with high-resolution online videomicroscopy haverevealed that circulating microparticles mediate the accu-mulation of tissue factor on platelet-rich thrombi, and alsoin the venous thrombosis model [18,30]. Demonstrating theincreased procoagulant activity related to CPMPs in APEpatients vs ControlsnoCVRF, our results argue in favour of theirfunctional participation in venous thrombogenesis in theinterface of soluble coagulants factors, circulating cells anddysfunctional endothelium. As far as microparticle subtypesare concerned, only PMPs appeared to be significantly ele-vated in APE patients vs ControlsnoCVRF. We have to take intoaccount the lower detection limits in the estimation of EMPinvolvement in APE to interpret this result, in terms of thevariability in antibody affinity (CD31 vs CD62E or CD144), asin the availability of the antigen or in the level of endothelialmarkers borne by microparticles [31]. However, this featurehas been used in studies reported by other groups to mea-sure EMPs (CD31) and PMPs (glycoprotein 1b) distinctly [16].In the future, more specific endothelial phenotypes, such asCD62E/CD144/CD146, should be targeted as a priority.

Data emerging from the literature are quite discordantconcerning microparticles and phenotypes involved in APE,mainly because of the different techniques and method-ological approaches used. To our knowledge, there are twoprevious clinical reports on this subject that differ from ourstudy in terms of the design, the method of microparticlequantitation, the population studied and the comparator

used [16,17]. In addition, increasing data correlated therisk of deep-vein thrombosis in patients with cancer totissue factor-bearing tumour cell-derived microparticle lev-els measured by functional test or flow cytometry, with apathophysiological role supported by P-selectin glycoprotein

tutao

401

igand 1 (or other mucinous glycoproteins) and P-selectinnteraction [21,23].

elation between cardiovascular risk factorsnd microparticles in venoushromboembolism

ur observations highlight the importance of adjusting forhe presence of cardiovascular risk factors in conditions inhich microparticle levels are raised.

Diabetes [13,15,32] and hypertension [14] are associatedith endothelial dysfunction and platelet activation, and

ncrease CPMP levels. In a previous report, our group [15]ound a higher level of CPMPs in patients with type 2 dia-etes compared with healthy subjects. A strong positiveorrelation was also found between endothelial and plateleticroparticle levels on the one hand, and the absolute

evel of both systolic and diastolic blood pressures on thether [14]. These findings and the present data emphasizehat cardiovascular risk factors, especially hypertension,re confounders in the previously described relationshipetween circulating microparticles and VTE disease [16,17].

We suggest several hypotheses to account for the intrigu-ng lack of difference in CPMP procoagulant activity betweenases and ControlsCVRFs. First, we could consider the pre-ominant role of cardiovascular risk factors as vascular cellctivators vs venous thrombosis risk factors such as hypox-emia induced by blood stasis. A second hypothesis relies onn expected difference in microparticle consumption kinet-cs between both populations, which would be faster inases of an acute thrombotic event, such as venous thrombo-is [22,33—35]. Indeed, the production and consumption oficroparticles are mainly local processes in the acute phase

f VTE. This is in line with the recent experimental study byamacciotti et al., demonstrating not only variable kineticsor each subtype of microparticle, but also their evolutivehrombogenicity during the thrombotic process itself [19].he sequestration of microparticles in the forming throm-us is also a diluting factor [23]. The last hypothesis is annderestimation of the CPMP effective procoagulant activ-ty due to a lack of sensitivity with our test using a specificuantification of microparticle-linked phosphatidylserine infunctional assay, compared with tissue factor activity eval-ation [36].

icroparticles: cause or consequence inenous thrombogenesis?

y evaluating microparticle-related procoagulant activity innly the acute phase of venous thrombosis, our study isnable to assess whether CPMPs have a causal role in venoushrombogenesis. Nevertheless, we found some interestingorrelations.

The PMP subtype appeared to be an important sourcef procoagulant microparticles, and was correlated withPMPs. This is in line with the hypothesis of platelet activa-

ion and participation in venous thrombogenesis as has beennderlined recently in the literature [2,3], linking venoushrombosis to atherothrombosis. PMPs levels have beenssociated positively with thrombus weight and proteomicsf microparticles after VTE revealed the upregulation of Gal

4

3hda

cala[mb

CwacsbeastIpicnwa[

S

Wrutt[mSefrmabrvpubt[

C

Oto

wdcc

dpvi

C

N

A

SicpAR

R

[

[

[

02

BP, a polypeptide from lectin family that plays a key role inuman platelet aggregation and function, promoting shed-ing of microparticles and generation of leucocyte-plateletggregates [37].

In parallel, we observed a tendency towards an asso-iation of CPMPs and PMPs with severe clinical status, inccordance with brain natriuretic peptide levels, but weaklyinked with C-reactive protein, an inflammatory marker with

poor negative predictive value in pulmonary embolism38,39]. Recently, two studies correlated the severity of pul-onary hypertension with CPMP levels in pulmonary arterylood samples [31,40].

Finally, after adjustment for age, sex and hypertension,PMPs and PMPs appeared to be correlated in our casesith current smoking, a major cardiovascular risk factor intherothrombosis. Moreover, Pomp et al. [41] showed thaturrent smoking remained a risk factor for venous thrombo-is among young people after adjustment for age, sex andody mass index. In the literature, an increasing amount ofvidence suggests the likelihood of a link between arterialnd venous disease [42]. According to results from recenttudies, atherosclerosis and VTE share common risk fac-ors, including age, obesity and current smoking [43,44].n a large population-based study, Sorensen et al. [45]rovided strong evidence that patients with VTE are atncreased risk of subsequent arterial cardiovascular eventsompared with population controls, which is most pro-ounced during the first year of follow-up. This is consistentith underlying common prothrombotic mechanisms suchs thrombogenesis, endothelial damage and inflammation46].

tudy limitations

e did not evaluate the post-VTE prothrombotic stateelated to CPMPs using a second measure of CPMP procoag-lant activity during the first year of follow-up. Moreover,aking into account the limitation of our test related tohe choice of CD31 as single endothelial phenotypic marker47,48], we should also consider the involvement of othericroparticles subtypes, such as leukocyte microparticles.

everal studies demonstrated that platelets, leukocytes andndothelial cells colocalize and interact in the milieu of aorming thrombus [49—52]. By binding the PSGL-1 countereceptor via P-selectin, platelets and PMPs could activateonocytes, enhancing tissue factor expression and proco-

gulant leukocyte microparticle release [53]. As suggestedy experimental studies in vivo, leukocyte microparticleelease might play a key role in thrombus initiation toascular remodelling, and we are going to evaluate theirrocoagulant activity in APE [19,49]. Finally, we might havenderestimated the procoagulant activity related to CPMPy using a prothrombinase assay, whereas the estimation ofissue factor procoagulant activity would be more relevant36].

onclusions

ur study confirms the correlation between APE and CPMPshrough their own procoagulant functionality, relying mainlyn platelet activation. This relation no longer holds true

[

L. Bal et al.

hen patients with APE are compared with those with car-iovascular risk factors, stressing the fact that potentialonfounders should be taken into account when microparti-le levels are analysed in VTE.

Further research is needed to demonstrate in which con-itions (i.e., initial stimuli, cell origin, etc.) the circulatingool of microparticles could be sufficient to precipitate aenous thrombotic event and to evaluate the therapeuticmplications.

onflict of interest statement

one.

cknowledgment

ophie Rushton-Smith, PhD, provided editorial assistancen the final version of this manuscript, including editing,hecking content and language, formatting, referencing andreparing tables and figures, and was funded by Saint-ntoine University hospital and Medical School (associationobert-Debré).

eferences

[1] Virchow R. Gesammalte abhandlungen zur wissenschaftlichenmedtzin. Frankfurt: Medinger Sohn & Co; 1856, pp. 219—732.

[2] Blann AD, Lip GY. Virchow’s triad revisited: the importance ofsoluble coagulation factors, the endothelium, and platelets.Thromb Res 2001;101:321—7.

[3] Sobieszczyk P, Fishbein MC, Goldhaber SZ. Acute pul-monary embolism: don’t ignore the platelet. Circulation2002;106:1748—9.

[4] Freyssinet JM. Cellular microparticles: what are they bad orgood for? J Thromb Haemost 2003;1:1655—62.

[5] VanWijk MJ, VanBavel E, Sturk A, et al. Microparticles in car-diovascular diseases. Cardiovasc Res 2003;59:277—87.

[6] Wolf P. The nature and significance of platelet products inhuman plasma. Br J Haematol 1967;13:269—88.

[7] Zwaal RF, Schroit AJ. Pathophysiologic implications ofmembrane phospholipid asymmetry in blood cells. Blood1997;89:1121—32.

[8] Berckmans RJ, Neiuwland R, Boing AN, et al. Cell-derivedmicroparticles circulate in healthy humans and supportlow grade thrombin generation. Thromb Haemost 2001;85:639—46.

[9] Bernal-Mizrachi L, Jy W, Fierro C, et al. Endothelial micropar-ticles correlate with high-risk angiographic lesions in acutecoronary syndromes. Int J Cardiol 2004;97:439—46.

10] Bernal-Mizrachi L, Jy W, Jimenez JJ, et al. High levels ofcirculating endothelial microparticles in patients with acutecoronary syndromes. Am Heart J 2003;145:962—70.

11] Mallat Z, Benamer H, Hugel B, et al. Elevated levels of shedmembrane microparticles with procoagulant potential in theperipheral circulating blood of patients with acute coronarysyndromes. Circulation 2000;101:841—3.

12] Ederhy S, Di Angelantonio E, Mallat Z, et al. Levels of

circulating procoagulant microparticles in nonvalvular atrialfibrillation. Am J Cardiol 2007;100:989—94.

13] Nomura S, Suzuki M, Katsura K, et al. Platelet-derivedmicroparticles may influence the development of atheroscle-rosis in diabetes mellitus. Atherosclerosis 1995;116:235—40.

ism

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

Circulating procoagulant microparticles in pulmonary embol

[14] Preston RA, Jy W, Jimenez JJ, et al. Effects of severe hyperten-sion on endothelial and platelet microparticles. Hypertension2003;41:211—7.

[15] Sabatier F, Darmon P, Hugel B, et al. Type 1 and type 2 diabeticpatients display different patterns of cellular microparticles.Diabetes 2002;51:2840—5.

[16] Chirinos JA, Heresi GA, Velasquez H, et al. Elevation ofendothelial microparticles, platelets, and leukocyte activationin patients with venous thromboembolism. J Am Coll Cardiol2005;45:1467—71.

[17] Inami N, Nomura S, Kikuchi H, et al. P-selectin and platelet-derived microparticles associated with monocyte activationmarkers in patients with pulmonary embolism. Clin ApplThromb Hemost 2003;9:309—16.

[18] Biro E, Sturk-Maquelin KN, Vogel GM, et al. Human cell-derivedmicroparticles promote thrombus formation in vivo in a tissuefactor-dependent manner. J Thromb Haemost 2003;1:2561—8.

[19] Ramacciotti E, Hawley AE, Farris DM, et al. Leukocyte-and platelet-derived microparticles correlate with throm-bus weight and tissue factor activity in an experimentalmouse model of venous thrombosis. Thromb Haemost2009;101:748—54.

[20] Key NS, Chantrathammachart P, Moody PW, et al. Membranemicroparticles in VTE and cancer. Thromb Res 2010;125(Suppl.2):S80—3.

[21] Tesselaar ME, Romijn FP, van der Linden IK, et al. Microparticle-associated tissue factor activity in cancer patients with andwithout thrombosis. J Thromb Haemost 2009;7:1421—3.

[22] Tesselaar ME, Romijn FP, Van Der Linden IK, et al. Microparticle-associated tissue factor activity: a link between cancer andthrombosis? J Thromb Haemost 2007;5:520—7.

[23] Thomas GM, Panicot-Dubois L, Lacroix R, et al. Cancercell-derived microparticles bearing P-selectin glycoprotein lig-and 1 accelerate thrombus formation in vivo. J Exp Med2009;206:1913—27.

[24] Zwicker JI. Predictive value of tissue factor bearingmicroparticles in cancer associated thrombosis. Thromb Res2010;125(Suppl. 2):S89—91.

[25] Bal L, Ederhy S, Di Angelantonio E, et al. Circulating pro-coagulant microparticles in acute pulmonary embolism: Acase-control study. Int J Cardiol 2009 [Epub ahead of print].

[26] Freyssinet JM, Dignat-George F. More on: measuring circulatingcell-derived microparticles. J Thromb Haemost 2005;3:613—4.

[27] Mallat Z, Hugel B, Ohan J, et al. Shed membrane micropar-ticles with procoagulant potential in human atheroscleroticplaques: a role for apoptosis in plaque thrombogenicity. Cir-culation 1999;99:348—53.

[28] Morel O, Hugel B, Jesel L, et al. Sustained elevated amountsof circulating procoagulant membrane microparticles and solu-ble GPV after acute myocardial infarction in diabetes mellitus.Thromb Haemost 2004;91:345—53.

[29] Hugel B, Martinez MC, Kunzelmann C, et al. Membranemicroparticles: two sides of the coin. Physiology (Bethesda)2005;20:22—7.

[30] Falati S, Liu Q, Gross P, et al. Accumulation of tissue factor intodeveloping thrombi in vivo is dependent upon microparticle P-selectin glycoprotein ligand 1 and platelet P-selectin. J ExpMed 2003;197:1585—98.

[31] Bakouboula B, Morel O, Faure A, et al. Procoagulant membranemicroparticles correlate with the severity of pulmonary arte-

rial hypertension. Am J Respir Crit Care Med 2008;177:536—43.

[32] Diamant M, Nieuwland R, Pablo RF, et al. Elevated numbersof tissue-factor exposing microparticles correlate with com-ponents of the metabolic syndrome in uncomplicated type 2diabetes mellitus. Circulation 2002;106:2442—7.

[

403

33] Combes V, Simon AC, Grau GE, et al. In vitro genera-tion of endothelial microparticles and possible prothromboticactivity in patients with lupus anticoagulant. J Clin Invest1999;104:93—102.

34] Dignat-George F, Camoin-Jau L, Sabatier F, et al. Endothe-lial microparticles: a potential contribution to the thromboticcomplications of the antiphospholipid syndrome. ThrombHaemost 2004;91:667—73.

35] Morel O, Jesel L, Freyssinet JM, et al. Elevated levels of proco-agulant microparticles in a patient with myocardial infarction,antiphospholipid antibodies and multifocal cardiac thrombosis.Thromb J 2005;3:15.

36] Furie B, Furie BC. Mechanisms of thrombus formation. N EnglJ Med 2008;359:938—49.

37] Ramacciotti E, Hawley AE, Wrobleski SK, et al. Proteomicsof microparticles after deep venous thrombosis. Thromb Res2010;125:e269—74.

38] Fox EA, Kahn SR. The relationship between inflammation andvenous thrombosis. A systematic review of clinical studies.Thromb Haemost 2005;94:362—5.

39] Tsai AW, Cushman M, Rosamond WD, et al. Coagulation factors,inflammation markers, and venous thromboembolism: the lon-gitudinal investigation of thromboembolism etiology (LITE). AmJ Med 2002;113:636—42.

40] Amabile N, Heiss C, Real WM, et al. Circulating endothe-lial microparticle levels predict hemodynamic severity ofpulmonary hypertension. Am J Respir Crit Care Med2008;177:1268—75.

41] Pomp ER, Rosendaal FR, Doggen CJ. Smoking increases the riskof venous thrombosis and acts synergistically with oral contra-ceptive use. Am J Hematol 2008;83:97—102.

42] Prandoni P. Links between arterial and venous disease. J InternMed 2007;262:341—50.

43] Ageno W, Becattini C, Brighton T, et al. Cardiovascular riskfactors and venous thromboembolism: a meta-analysis. Circu-lation 2008;117:93—102.

44] Holst AG, Jensen G, Prescott E. Risk factors for venous throm-boembolism: results from the Copenhagen City Heart Study.Circulation 2010;121:1896—903.

45] Sorensen HT, Horvath-Puho E, Pedersen L, et al. Venousthromboembolism and subsequent hospitalisation due to acutearterial cardiovascular events: a 20-year cohort study. Lancet2007;370:1773—9.

46] Goon PK, Lip GY. Arterial disease and venous thromboem-bolism: a modern paradigm? Thromb Haemost 2006;96:111—2.

47] Jimenez JJ, Jy W, Mauro LM, et al. Endothelial cells releasephenotypically and quantitatively distinct microparticles inactivation and apoptosis. Thromb Res 2003;109:175—80.

48] Morel O, Ohlmann P, Epailly E, et al. Endothelial cell activationcontributes to the release of procoagulant microparticles dur-ing acute cardiac allograft rejection. J Heart Lung Transplant2008;27:38—45.

49] Lopez JA, Kearon C, Lee AY. Deep venous thrombosis. HematolAm Soc Hematol Educ Prog 2004:439—56.

50] Mesri M, Altieri DC. Endothelial cell activation by leukocytemicroparticles. J Immunol 1998;161:4382—7.

51] Mesri M, Altieri DC. Leukocyte microparticles stimulateendothelial cell cytokine release and tissue factor inductionin a JNK1 signaling pathway. J Biol Chem 1999;274:23111—8.

52] Myers DD, Hawley AE, Farris DM, et al. P-selectin and leukocyte

microparticles are associated with venous thrombogenesis. JVasc Surg 2003;38:1075—89.

53] Jy W, Mao WW, Horstman L, et al. Platelet microparticles bind,activate and aggregate neutrophils in vitro. Blood Cells Mol Dis1995;21:217—31 [discussion 31a].