is elevated level of soluble endothelial protein c receptor a new risk factor for retinal vein...

Clinical and Experimental Ophthalmology 2006doi:10.1111/j.1442-9071.2006.1212.x

© 2006 Royal Australian and New Zealand College of Ophthalmologists

� Correspondence: Dr Sibel Kadayifcilar, Hacettepe University, School of Medicine, Department of Ophthalmology, TR-06100 Sihhiye, Ankara, Turkey. Email:

[email protected]

The study was presented in part at Association for Research in Vision and Ophthalmology annual meeting, Ft Lauderdale, Florida, USA, 4 May 2005.

Received 14 April 2005; accepted 1 November 2005.

Original Article

Is elevated level of soluble endothelial protein C receptor a new risk factor for retinal vein occlusion?Koray Gumus MD,1 Sibel Kadayifcilar MD,1 Bora Eldem MD,1 Osman Saracbası PhD,2 Osman Ozcebe MD,3 Semra Dundar MD3 and Serafettin Kirazli3Departments of 1Ophthalmology, 2Biostatistics and 3Haematology, Hacettepe University, School of Medicine, Ankara, Turkey

Key words: branch retinal vein occlusion, central retinalvein occlusion, risk factor, soluble endothelial protein Creceptor, venous thrombosis.

INTRODUCTION

Retinal vein occlusion (RVO) is a common cause of visualimpairment among the elderly. The underlying pathogenesisof RVO remains unclear. A number of systemic disorders,such as arterial hypertension, diabetes mellitus, hyperviscos-ity and hyperlipidaemia, as well as ocular risk factors includ-ing glaucoma, hypermetropia and uveitis may alter vascularflow and therefore predispose to RVO.1–3

Haematological factors such as venous stasis and hyper-coagulability may also play a role in the pathogenesis ofRVO.4 Apart from deficiencies of the anticoagulant proteinsantithrombin III (ATIII), protein C and protein S, some muta-tions including factor V Leiden (FVL), prothrombinG20210A and methylene tetrahydrofolate reductase C677Thave been identified as hereditary risk factors.5–7 Besides,high levels of activated factor VII, lipoprotein (a), factor VIII(FVIII) and hyperhomocysteinaemia have been reported asother possible risk factors for patients with RVO.7–9

The protein C anticoagulant pathway is well establishedas an important mechanism for inhibition of the coagulationprocess.10 Endothelial protein C receptor (EPCR), mainlyexpressed on endothelial cells of large vessels, play a criticalrole in this pathway.11–13 Recently, a soluble form of EPCR(sEPCR) has been detected in normal human plasma and hasbeen shown to inhibit the anticoagulant activity of activatedprotein C (APC) by blocking its binding to phospholipidsand by annihilating its ability to inactivate factor Va, therebyleading to thrombotic events.12,14,15

To the best of our knowledge, the relation between levelsof sEPCR and RVO has not been investigated yet. In thisstudy, we addressed whether high levels of sEPCR might

ABSTRACT

Background: To evaluate the systemic and thrombophilicrisk factors for retinal vein occlusion (RVO) and todetermine whether the elevated level of soluble endot-helial protein C receptor (sEPCR) is a risk factor forthrombosis.

Methods: In this case–control study, 56 patients with centralRVO (CRVO), 26 patients with branch RVO (BRVO) and78 healthy sex- and age-matched subjects were enrolled.Following ophthalmological examination, venous blood wasanalysed for glucose, lipid profile, lipoprotein (a), homocys-teine, activated partial thromboplastin time, fibrinogen, fac-tor VIII, protein C activity, protein S activity, activated proteinC resistance, antithrombin III activity, lupus anticoagulant,anti-cardiolipin antibody, anti-phospholipid antibody, sEPCR,factor V Leiden mutation and prothrombin G20210Amutation.

Results: Apart from hypertension, glaucoma, lipoprotein(a), homocysteine and factor VIII, elevated levels ofsEPCR were found to be a risk factor for CRVO (oddsratio, 1.02; 95% confidence interval, 1.007–1.028;P = 0.001). Patients with CRVO had significantly higherlevels of sEPCR than those with BRVO and controls(respectively, 160.1 ± 83.8, 116.8 ± 65.2 and 111.3 ± 60.5;P = 0.005). Moreover, 39% of patients with CRVO hadlevels of sEPCR more than 200 ng/mL, and only 5% ofcontrols and 11% of patients with BRVO had similar highlevels.

Conclusions: Besides known classical risk factors, elevatedlevels of sEPCR seem to be an important candidate riskfactor for especially CRVO.

2 Gumus et al.


play a role as a risk factor for RVO apart from classical riskfactors.

METHODS

Between September 2003 and May 2004, 56 patients withbranch RVO (BRVO) and 26 patients with central RVO(CRVO) were enrolled in this study. The case group com-prised newly diagnosed and prior cases with RVO. The con-trol group consisted of 78 healthy sex- and age-matchedsubjects who presented with refractive errors, presbyopia orcataract and volunteered for the study.

Patients with severe myopia, vasoproliferative retinopa-thy, intraocular inflammatory disease, active systemic infec-tion, elevated liver enzymes and renal dysfunction, a historyof operation within a week or patients taking medicationincluding oral anticoagulants, heparin, vitamins and lipid-lowering drugs were excluded from the study.

According to the interval between the probable time ofocclusion and the examination, the cases were analysed inthree groups. The first group less than 1 month (4 cases –15.4% with CRVO and 8 cases – 14.3% with BRVO); thesecond group 1–4 months (12 cases – 46.1% with CRVOand 26 cases – 46.4% with BRVO); and the third group morethan 4 months (10 cases – 38.5% with CRVO and 22 cases– 39.3% with BRVO).

A detailed medical history for known or suspected hyper-tension, diabetes mellitus, cardiovascular disease, glaucoma,previous thromboembolic events, smoking, family history ofthrombosis, oral contraceptive or hormone replacementtherapy use, previous operations and current medicationswas obtained from all patients.

After Institutional Board approval, a written informedconsent was taken and all patients underwent complete oph-thalmic examination. Venous blood was drawn from theantecubital vein after an overnight fast of at least 8 h foranalysis of fasting blood sugar, liver and renal function tests,serum triglycerides, cholesterol, lipoprotein (a) (LIPOPRO-TEIN (A)X; IMMAGE Immunochemistry Systems) andhomocysteine (fluorescence polarization immunoassay;AXSYM-Homocysteine; AXSYM SYSTEM, Abbott,Wiesbaden, Germany).

Haematological tests included activated partial thrombo-plastin time (aPTT) (STA-C.K.Prest), fibrinogen (clottingmethod of Clauss; STA-Fibrinogen), ATIII activity (chro-mogenic substrate assay; STA-ATIII), protein C activity (syn-thetic chromogenic substrate method; STA-Protein C), freeprotein S activity (immunoturbidimetric method; STA-Liatest Free Protein S), FVIII (the assay consists of the mea-surement of the clotting time; STA-Deficient VIII), lupusanticoagulant (clotting time; partial thromboplastin time-lupus anticoagulant kit) assays. All mentioned tests werecarried out using STA Analyser (Diagnostica Stago Asnieres,France). Anti-cardiolipin and anti-phospholipid antibodies(immunoglobulin M and G) were determined with anenzyme-linked immunosorbent assay (ELISA; EUROIM-MUN, Lübeck, Germany). Functional activated protein C

resistance (APCR) is expressed as the ratio of aPTT in thepresence of APC to aPTT in its absence.

Measurements of sEPCR levels were performed with anewly developed ELISA for plasma sEPCR (AsserachromsEPCR kit, Diagnostica Stago, Asnieres, France). The assayused 10 µL of plasma, and was calibrated with recombinanthuman sEPCR.

Genomic DNA samples were isolated using ‘MagNA PureLC DNA Isolation Kit I’ with a specialized system (MagNAPure LC, Roche Diagnostics GmbH, Penzberg, Germany).Prothrombin G20210A and FVL mutations were analysedwith LightCycler (Roche Diagnostics GmbH).

Data were analysed by SPSS 11.0 for Windows packagesoftware (SPSS, Inc., Chicago, IL, USA). The categoricalvariables were compared by chi-square test and numericalvariables were compared by independent t-test and one-wayANOVA with Tukey comparison test. The analysis of covari-ance (ANCOVA) was performed to assess the effects of someconfounding factors such as age and sex on the plasma levelsof sEPCR. Pearson correlation test was used to examine thepossible relations between plasma sEPCR levels and otherfactors such as protein C activity. A P-value of 0.05 or lesswas considered as being statistically significant.

All potential risk factors were then coded as binary vari-ables so that odds ratios (ORs) represented the odds ofpatients with that factor having an event versus patientswithout that factor having the same event, adjusted for allother variables in the analyses. Initially, we screened eachpossible risk factor separately in a logistic regression model.After screening, ORs with confidence intervals were cal-culated for all discrete variables. Finally, we computed amultiple logistic model including all variables that werestatistically significant (P < 0.05) in the previous stage of theanalysis.

RESULTS

The study group of 82 patients with RVO was composed of46 women and 36 men aged 33–72 years (57.7 ± 9.4). Forty-five women and 33 men aged 34–72 years (57.4 ± 9.7) com-prised the control group. The groups did not differ withregard to age (P = 0.870) or sex (P = 0.874). Patient distri-bution according to the probable time of occlusion did notreveal statistically significant difference among the RVOgroups (P = 0.991).

A detailed list of previous/current systemic diseases(endocrine diseases, pulmonary diseases, cardiovascular dis-eases, gastrointestinal diseases, renal diseases, rheumatolog-ical conditions, infections, neurological diseases andoncological diseases) during or before onset of RVO wascarefully analysed and no statistically significant differencewas detected among the groups except hypertension. Therewas a significantly higher prevalence of hypertension inCRVO and BRVO compared with controls (respectively,P = 0.003 and P = 0.000). RVO groups and the controls didnot reveal any significant difference according to the educa-tional status, presence of diabetes, coronary artery disease

Soluble endothelial protein C receptor level 3


(CAD), cigarette smoking, and the family history of RVOand CAD. However, family history of cerebrovascular dis-ease (CVD) was as high as 38.5% in patients with CRVO(P = 0.001). None of the subjects had experienced deepvenous thrombosis or pulmonary embolism. Seven cases(26.9%) with CRVO, four cases (7.1%) with BRVO and fourotherwise healthy subjects (5.1%) had a history of openangle glaucoma and all were receiving anti-glaucoma medi-cations (P = 0.003). All clinical characteristics of patients aresummarized in Table 1.

The complete blood counts including platelets were withinnormal limits in all participants. None of the cases had ele-vated prothrombin levels and aPTT. No individual has shownpositivity for any of the autoantibodies including anti-phos-pholipid, anti-cardiolipin antibody and lupus anticoagulant.

Mean values of fasting glucose and lipid profiles did notreveal any statistically significant differences among thegroups. Homocysteine was significantly higher in CRVOand BRVO groups (respectively, P = 0.02 and P = 0.001)compared with controls. Lipoprotein (a) was also higher inCRVO and BRVO than those in controls, but this differencewas not statistically significant (P = 0.078). However, it wasfound as a risk factor for CRVO in the logistic regressionmodel (Table 2) . Among the physiological anticoagulants,solely ATIII activity was significantly lower in CRVO.

Marked elevation in FVIII levels in RVO was noted(P < 0.001). There was no statistically significant differencewith respect to fibrinogen, APCR and genetic mutations.The details of these parameters are documented in Table 3.

Soluble EPCR levels were significantly elevated in caseswith CRVO compared with the controls and the cases withBRVO (P = 0.005) (Table 3 and Fig. 1). Neither age(P = 0.436) nor sex (P = 0.619) had effect on levels of sEPCRin CRVO (P = 0.009) (ANCOVA). Correspondingly, theeffects of age and sex on plasma sEPCR levels were analysedand no statistically significant correlation was found (respec-tively, P = 0.199 r = 0.102 and P = 0.815 r = − 0.019). Toappropriately investigate the effect of possible confoundingfactors such as hypertension on elevated levels analysis ofcovariance (ANCOVA) was performed. According to this,hypertension had no effect either on elevated sEPCR levelsor on the difference between the groups (F1,160 = 0.136 andP = 0.713).

According to the probable time of RVO, mean sEPCRlevels did not reveal statistically significant difference amongthe patients both in CRVO (P = 0.230) (respectively,219.0 ± 22.1, 162.7 ± 98.5 and 133.4 ± 71.8) and in BRVO(P = 0.094) (respectively, 147.1 ± 78.5, 125.8 ± 72.7 and95.0 ± 42.8) groups. However, patients who were diagnosedwith RVO in less than a month had slightly higher sEPCR

Table 1. General characteristic of patients and controls

Parameters CRVO (n = 26) n (%) BRVO (n = 56) n (%) Control (n = 78) n (%) P*

Hypertension 17 (65.4) 41 (73.2) 24 (30.8) <0.05Diabetes 2 (7.7) 4 (7.1) 7 (9.0) 0.925CAD 4 (15.4) 5 (8.9) 5 (6.4) 0.411Glaucoma 7 (26.9) 4 (7.1) 4 (5.1) <0.05Cigarette

Ex-smokers 6 (23.1) 16 (28.6) 13 (16.7) 0.480Current smokers 7 (26.9) 12 (21.4) 25 (32.0)

RVO in family 2 (7.7) 2 (3.6) 2 (2.6) 0.545CVD in family 10 (38.5) 6 (10.7) 8 (10.3) <0.05CAD in family 11 (42.3) 10 (17.9) 20 (25.6) 0.062

*The P-values are of Pearson chi-square test. BRVO, branch retinal vein occlusion; CAD, coronary artery disease; CRVO, central retinalvein occlusion; CVD, cerebrovascular disease; n, number; RVO, retinal vein occlusion.

Table 2. Summary of significant risk factors in multivariable logistic regression analysis

Groups Factors P-values Odds ratio 95% confidence interval

CRVO Hypertension (≥5 years) 0.043 4.15 1.044–16.466Glaucoma (yes/no)† 0.041 13.06 1.107–154.129Homocysteine (umol/L) 0.001 1.40 1.156–1.697Lipoprotein (a) (mg/dL) 0.006 1.04 1.011–1.069Factor VIII (%) 0.002 1.02 1.006–1.026sEPCR (ng/dL) 0.001 1.02 1.007–1.028

BRVO Hypertension (≥5 years) 0.001 4.22 1.794–9.918Homocysteine (umol/L) 0.009 1.16 1.037–1.296Factor VIII (%) 0.004 1.01 1.002–1.013

†Only cases whose glaucoma was diagnosed and followed up for more than 5 years before retinal vein occlusion occurred. BRVO, branchretinal vein occlusion; CRVO, central retinal vein occlusion; sEPCR, soluble endothelial protein C receptor.

4 Gumus et al.


levels. Plasma sEPCR levels greater than 200.1 ng/mL wereobserved in 10 cases with CRVO (38.5%), in 6 cases withBRVO (10.7%) and in 4 control subjects (5.1%) (P < 0.001).The distributions of patients according to the levels ofsEPCR are shown in Figure 1. The possible relationshipbetween sEPCR levels and protein C activity was also inves-tigated and a negative correlation was found, especially inpatients with CRVO (r = − 0.597, P = 0.001). This negativecorrelation was present in other groups (BRVO and control),as well, but r-values failed to reach statistical significance.Furthermore, all other variables such as demographic data,clinical characteristics and blood parameters investigated inthis study were analysed with regard to any association withplasma sEPCR levels. As a result, there was no other variableshowing correlation with elevated levels of sEPCR exceptprotein C activity.

Finally, a multiple logistic regression model was used todetermine the association of plasma sEPCR levels and otherpossible risk factors with both CRVO and BRVO. The vari-ables significantly associated with CRVO and BRVO in twoindependent multivariable logistic regressions are summa-rized in Table 2.

DISCUSSION

The exact pathogenesis and risk factors in patients affectedby RVO are still poorly understood. In general, degenerative

Table 3. Laboratory findings of all patients with retinal vein occlusion and controls

Parameters CRVO (n = 26)(a)

BRVO (n = 56)(b)

Control (n = 78)(c)

(P-values)Tukey

a-bTukey

a-cTukey

b-c

sEPCR (ng/mL) 160.1 ± 83.8 116.8 ± 65.2 111.3 ± 60.5 (0.005)(Range) 58.0–412.5 l37.6–398.7 31.2–395.0 0.020 0.003 0.837

Lipoprotein (a) (mg/dL) 31.5 ± 28.1 30.3 ± 29.4 21.5 ± 22.0 (0.078)(Range) (N: 0–30) 2.0–106 2.0–131 1.9–117 0.980 0.200 0.125

Homocysteine (umol/L) 14.6 ± 4.4 14.7 ± 5.7 11.8 ± 3.4 (0.000)(Range) (N: 3.4–20.5) 7.49–24.35 6.43–40.02 5.81–23.16 0.989 0.020 0.001

Logarithmic factor VIII (%) 197.3 177.5 133.4 (0.000)(Range) (N: 53–170) 162.2–240.1 154.6–204.0 120.0–148.4 0.637 0.002 0.003

Protein C activity (%) 101.5 ± 28.0 105.7 ± 16.3 103.6 ± 24.0 (0.716)(Range) (N: 70–130) 47–195 73–132 62–199 0.713 0.915 0.849

Protein S activity (%) 106.6 ± 25.3 105.0 ± 32.0 109.0 ± 24.6 (0.697)(Range) (N: 60–130) 56–150 10–150 14–147 0.964 0.922 0.677

Antithrombin III activity (%) 99.1 ± 18.0 108.5 ± 14.1 103.7 ± 15.6 (0.031)(Range) (N: 80–120) 70–137 73–139 65–132 0.031 0.401 0.178

APCR (second) 149.8 ± 44.8 139.6 ± 42.2 145.4 ± 38.8 (0.533)(Range) (N: 120–300) 49–253 45.7–266.6 47.7–249.3 0.543 0.879 0.701

Factor V Leiden n (%)† 4 (15.4) 11 (19.6) 6 (7.7) (0.121) *Prothrombin 20210A n (%)‡ – 3 (5.4) 3 (3.8) (0.310) **

*Pearson chi-square test; **likelihood ratio test, other P-values are those of one-way ANOVA (Tukey). †One each from control and CRVOis homozygous, others are heterozygous. ‡All are heterozygous. APCR, activated protein C resistance; BRVO, branch retinal vein occlusion;CRVO, central retinal vein occlusion; n, number; N, normal range values according to the laboratory; sEPCR, soluble endothelial protein Creceptor.

Figure 1. Soluble endothelial protein C receptor (sEPCR) levelsin central retinal vein occlusion (CRVO), branch retinal vein occlu-sion (BRVO) and control group. The black lines in the box plotdiagram show the mean values of the groups (�, outlier cases; *,extreme cases).

785626n =

ControlBRVOCRVO

Plas

ma

sEPC

R le

vel (

ng/m

l)

500

450

400

350

300

250

200

150

100

50

0



changes of the vessel wall, abnormal perivascular changesand abnormal haematological factors constitute the primarymechanism of vessel occlusion.16

If the delicate physiological balance between procoagu-lant and anticoagulant systems is shifted towards coagula-tion, a thrombophilic state can result. In this sensitivebalance, the protein C anticoagulant pathway is well estab-lished as a physiologically important mechanism for inhibit-ing the coagulation process.15 The pathway is started whenthrombin binds to thrombomodulin on the endothelial cellsurface.15 This complex rapidly activates protein C to gener-ate the anticoagulant enzyme APC.15 APC functions as ananticoagulant in plasma by inactivating factors Va and VIIIaon membrane surfaces, a process that is potentiated by theplasma vitamin K-dependent factor, protein S.15 EPCR thatis an endothelial cell-specific, type 1 transmembrane proteinthat binds both protein C and APC with high affinity playsa critical role in this complex pathway.12 Binding of EPCR toprotein C enhances the rate of protein C activation by thethrombin–thrombomodulin complex on the endothelial cellsurface.12,15 A soluble form of EPCR has recently beendetected in normal human plasma and has been shown tobind protein C and APC with an affinity to that of intactmembrane-bound EPCR.12,15

Various authors have endeavoured to clarify the possiblerole of sEPCR in the thrombotic events by performing in vivoand in vitro studies. Previous studies revealed that sEPCR isgenerated in vitro from its endothelial-bound parent (EPCR)by metalloproteinase activity, and this activity is inducibleby thrombin and some inflammatory mediators.17 SolubleEPCR binds protein C and APC with similar affinity, but itsbinding to APC inhibits the anticoagulant activity of APCpresumably by blocking its binding to phospholipids and byabrogating its activity to inactivate factor Va.14,15

It might be speculated that thrombin plays a critical rolein the release of sEPCR. This hypothesis was supported byanother study that documented that anticoagulant therapynot only reduced thrombin generation but also decreasedplasma sEPCR levels.18 In that study, patients receiving anti-coagulant therapy had lower sEPCR plasma levels than thoseof healthy volunteers.18 All these findings supported that thereduction in plasma sEPCR levels might be directly relatedto a reduction in thrombin generation.18 It may be assumedthat by reducing the elevated sEPCR levels in patients withRVO through using anticoagulant therapy, the probability offurther vascular occlusion will be reduced. To determine atreatment policy for these patients with elevated levels ofsEPCR, larger prospective studies are needed.

Both genetic and environmental factors may lead to dys-function of the EPCR-mediated coagulation-regulatingmechanism, resulting in increased levels of sEPCR. PlasmasEPCR levels may increase because of several acquired fac-tors, such as sepsis, and systemic lupus erythematosus, as wellas thrombin generation.19 Saposnik et al. documented therelation of genetic control with increased plasma sEPCRlevels and increased risk of venous thrombosis especially inmen.20 This genetic study suggested A3 haplotype, which

was associated with elevated plasma sEPCR levels, as a can-didate risk factor for venous thrombosis.20 Leiden Thrombo-philia Study Group investigated the association of the fourmost common haplotypes of the EPCR gene with plasmasEPCR levels and the risk of deep vein thrombosis.21 Theyreported that the distribution of sEPCR levels in the controlpopulation was trimodal and was genetically controlled byhaplotype 3 (H3).21 This haplotype was found to haveexplained 86.5% of the variation in sEPCR levels. It wasshown that carriers of two H3 alleles had higher sEPCRlevels than carriers of one H3 allele, which had higher levelsthan non-H3 carriers.21 Haplotype 4 was associated with aslightly increased risk. In addition to these findings, it wasclaimed that there was not a strong association betweenEPCR haplotypes and thrombosis risk, but low sEPCR levelsappeared to reduce the risk of deep vein thrombosis. Theyfound that plasma sEPCR levels below 81 ng/mL were asso-ciated with a decrease in risk when compared with sEPCRlevels above this value.21 Another large study by Medina et al.investigated the relationship between two polymorphismsand plasma sEPCR and APC levels and risk of venous throm-bosis.22 Carriers of the 4600AG genotype were found to havesignificantly higher sEPCR levels than those with the AAgenotype, and the 4678CC genotype was associated, to alesser extent, with elevated APC levels.22 As a result, it wasdocumented that individuals carrying the 4600AG genotypehad high sEPCR levels but did not have an increased risk ofthrombosis, whereas individuals carrying the 4678CC gen-otype had higher APC levels and lower risk of venous throm-boembolism.22 These studies suggest that sEPCR levels maybe a reflection of the underlying genotype. However, noneof the haplotypes of the EPCR gene is associated with astrong thrombosis risk, but low sEPCR levels appear toreduce the risk of deep vein thrombosis.21,22

Our main goal was to investigate the levels of sEPCR inRVO patients and healthy subjects and to address whethersEPCR levels are elevated in patients with RVO. Patientswith CRVO were found to have significantly higher levelsof sEPCR than those of control and BRVO subjects(P = 0.005). It is very difficult to evaluate this differencebetween CRVO and BRVO with this limited data. However,the presence of higher levels of EPCR on larger vessels13 mayhelp explain why sEPCR seems to be a risk factor for onlyCRVO, not for smaller vessels such as branch retinal veins.

There are no published data investigating the levels ofsEPCR in patients with RVO. In the literature, there are fewstudies investigating the distribution of sEPCR levels inhealthy subjects and patients with thrombosis. In the healthycontrol subjects, the distribution of the sEPCR levels wasbimodal in three different studies.18,22,23 According to those,although the majority of the subjects had sEPCR levelsbetween 75 and 178 ng/mL, only the 20% of subjects hadvalues more than 200 ng/mL.18,23 Recently, a new study byUitte de Willige et al. investigating the genetic control ofplasma sEPCR levels has documented that sEPCR showstrimodal distribution in healthy subjects.21 In our study, themean plasma sEPCR levels were 160.1 ± 83.8 mg/mL in

6 Gumus et al.


CRVO, 116.8 ± 65.2 mg/mL in BRVO and 111.3 ± 60.5 mg/mL in healthy subjects and no characteristic distribution ofthe plasma sEPCR levels was found. Furthermore, although39% of patients with CRVO had levels more than 200 ng/mL, solely 5–11% of patients in the control and BRVOgroups had this high values. These mean values obtained inthe current study were lower according to the previous pub-lished data.18,23 This discrepancy may be attributed to ethnicvariation and hereditary factors leading to different plasmalevels of sEPCR. To confirm this notion, further geneticstudies are required.

One important limitation of the current study is the factthat sEPCR levels measured at the time the patients werestudied may not reflect those at the time when the CRVOand BRVO occurred. However, as both CRVO and BRVOpatients had similar interval periods between the probabletime of occlusion and the examination, it may not be wrongnot to take this limitation into consideration.

Consequently, in the current series, patients with CRVOdemonstrated elevated levels of sEPCR, which might berelated to increased thrombosis; however, there is an over-riding confusion as to whether these elevated levels ofsEPCR are causative or solely a marker for a hypercoagulablestate. This issue should be clarified with further longitudinalstudies with follow-up sEPCR levels. Additionally, it will bean incorrect approach to claim that elevated levels of sEPCRcan cause RVO individually. RVO is multifactorial diseaseand coincidence of other known risk factors such as systemichypertension, hyperlipidaemia, hyperhomocysteinaemia,elevated levels of FVIII, lipoprotein (a) and some mutationsmay all contribute to the aetiopathogenesis of RVOs.

To the best of our knowledge, this is the first report onthe association between elevated plasma sEPCR levels andRVO. If further studies demonstrate significant role of sEPCRin the aetiopathogenesis of thrombosis in RVO, closer followup and anticoagulation for prophylaxis against thromboem-bolic events at other parts of the body will be advisable.

ACKNOWLEDGEMENT

The study was supported by Hacettepe University ResearchFund (Grant No. 04 D02 101003).

REFERENCES

1. The Eye Disease Case–Control Study Group. Risk factors forcentral retinal vein occlusion. Arch Ophthalmol 1996; 114: 545–54.

2. Hayreh SS, Zimmerman B, McCarthy MJ, Podhajsky P. Sys-temic diseases associated with various types of retinal veinocclusion. Am J Ophthalmol 2001; 131: 61–77.

3. The Eye Disease Case–Control Study Group. Risk factors forbranch vein occlusion. Am J Ophthalmol 1993; 116: 286–96.

4. Williamson TH, Rumley A, Lowe GDO. Blood viscosity, coag-ulation, and activated protein C resistance in central retinalvein occlusion: a population controlled study. Br J Ophthalmol1996; 80: 203–8.

5. Greiner K, Hafner G, Dick B, Peetz D, Prellwitz W, PfeifferN. Retinal vein occlusion and deficiencies in the protein Cpathway. Am J Ophthalmol 1999; 128: 69–74.

6. Hansen L, Kristensen HL, Bek T, Ingerslev J. Markers ofthrombophilia in retinal vein thrombosis. Acta Ophthalmol Scand2000; 78: 523–6.

7. Boyd S, Owens D, Gin T et al. Plasma homocysteine, methyl-ene tetrahydrofolate reductase C677T and factor II G20210Apolymorphisms, factor VIII, and VWF in central retinal veinocclusion. Br J Ophthalmol 2001; 85: 1313–15.

8. Bandello F, D’Angelo SV, Parlavecchia M et al. Hyperco-agulability and high lipoprotein (a) levels in patients withcentral retinal vein occlusion. Thromb Haemost 1994; 72:39–43.

9. Kadayıfçılar S, Özatlı D, Özcebe O, Sener EC. Is activatedfactor VII associated with retinal vein occlusion? Br J Ophthalmol2001; 85: 1174–8.

10. Esmon CT, Ding W, Yasuhiro K et al. The protein C pathway:new insights. Thromb Haemost 1997; 78: 70–4.

11. Fukudome K, Esmon CT. Identification, cloning, and regula-tion of a novel endothelial cell protein C/activated protein Creceptor. J Biol Chem 1994; 269: 26486–91.

12. Fukudome K, Kurosawa S, Stearns-Kurosawa DJ, He X, RezaieAR, Esmon CT. The endothelial cell protein C receptor. Cellsurface expression and direct ligand binding by the solublereceptor. J Biol Chem 1996; 271: 17491–8.

13. Laszik Z, Mitro A, Taylor FB, Ferrell G, Esmon CT. Humanprotein C receptor is present primarily on endothelium of largeblood vessels: implications for control of the protein C path-way. Circulation 1997; 96: 3633–40.

14. Kurosawa S, Stearns-Kurosawa DJ, Hidari N, Esmon CT. Iden-tification of functional endothelial protein C receptor in humanplasma. J Clin Invest 1997; 100: 411–18.

15. Liaw PC, Neuenchwander PF, Smirnov MD, Esmon CT.Mechanisms by which soluble endothelial cell protein C recep-tor modulates protein C and activated protein C function. JBiol Chem 2000; 275: 5447–52.

16. Bhagat N, Goldberg MF, Gascon P, Bell W, Haberman J,Zarbin MA. Central retinal vein occlusion: review of manage-ment. Eur J Ophthalmol 1999; 9: 165–80.

17. Xu J, Qu D, Esmon NL, Esmon CT. Metalloproteolytic releaseof endothelial protein C receptor. J Biol Chem 2000; 275: 6038–44.

18. Stearns-Kurosawa D, Swindle K, D’Angelo A et al. Plasma levelsof endothelial protein C receptor respond to anticoagulanttreatment. Blood 2002; 99: 526–30.

19. Kurosawa S, Stearns-Kurosawa DJ, Carson CW, D’Angelo A,Della Valle P, Esmon CT. Plasma levels of endothelial cellprotein C receptor are elevated in patients with sepsis andsystemic lupus erythematosus: lack of correlation with throm-bomodulin suggests involvement of different pathological pro-cesses. Blood 1998; 91: 725–7.

20. Saposnik B, Reny JL, Gaussem P, Emmerich J, Aiach M,Gandrille S. A haplotype of the EPCR gene is associated withincreased plasma levels of sEPCR and is a candidate risk factorfor thrombosis. Blood 2004; 103: 1311–18.

21. Uitte de Willige S, Van Marion V, Rosendaal FR, Vos HL, deVisser MCH, Bertina RM. Haplotypes of the EPCR gene,plasma sEPCR levels and the risk of deep venous thrombosis.J Thromb Haemost 2004; 2: 1305–10.



22. Medina P, Navarro S, Estelles A et al. Contribution of polymor-phisms in the endothelial protein C receptor gene to solubleendothelial protein C receptor and circulating activated pro-tein C levels, and thrombotic risk. Thromb Haemost 2004; 91:905–11.

23. Stearns-Kurosawa DJ, Burgin C, Parker D, Comp P, KurosawaS. Bimodal distribution of soluble endothelial protein Creceptor levels in healthy populations [Letter]. J Thromb Haemost2003; 1: 855–6.

is elevated level of soluble endothelial protein c receptor a new risk factor for retinal vein...

Documents