intl.elsevierhealth.com/journals/thre

Thrombosis Research (2007) 120, 811–814

BRIEF COMMUNICATION

Influence of sample type on soluble CD40ligand assessment in patients with acutecoronary syndromes

Michael Weber a,⁎, Birgitt Rabenau a, Michael Stanisch a, Holger M. Nef a,Helge Möllmann a, Albrecht Elsässer a, Vesselin Mitrovic a,Christopher Heeschen b, Christian Hamm a

a Kerckhoff Heart Center, Department of Cardiology, Benekestraße 2-8, 61231, Bad Nauheim, Germanyb Klinikum Grosshadern, Ludwig-Maximilians University, Munich, Germany

Received 5 August 2006; received in revised form 24 January 2007; accepted 28 January 2007Available online 2 March 2007

⁎ Corresponding author. Tel.: +49 603E-mail address: M.Weber@kerckhof

0049-3848/$ - see front matter © 200doi:10.1016/j.thromres.2007.01.014

Abstract

Background: Several studies have consistently shown that soluble CD40 ligand (sCD40L)concentrations are elevated in patientswith acute coronary syndromes (ACS) and that itcan be used as a biomarker for risk stratification. However, recently we coulddemonstrate that sampling techniques have impact on sCD40L measurements. Thus, itwas the aim of our prospective study to evaluate the impact of sampling techniques onsCD40L concentrations of patients with acute coronary syndromes compared tocontrols.Methods and results: We included a total of 60 patients, — 30 with an acute coronarysyndrome, 10 with cardiovascular risk factors but no relevant coronary artery diseaseand 20 healthy individuals. Blood samples were collected in gel filled tubes withoutadditives, EDTA filled tubes, or in citrate filled tubes. In EDTA or citrate plasmasamples, sCD40L values were higher in patients with ACS compared to controls. Incontrast, no difference in sCD40L values between ACS patients and controls wasobserved for serum samples.Conclusion: Our data show that only plasma samples, but not serum samples areappropriate for sCD40L measurements. In general, preanalytic conditions are crucialfor the assessment of sCD40L concentration and, thus, should be carefully consideredfor future studies.© 2007 Elsevier Ltd. All rights reserved.

2 9960; fax: +49 6032 9962313f-klinik.de (M. Weber).

7 Elsevier Ltd. All rights reserv

.

ed.

812 M. Weber et al.

Background

CD40 ligand (CD40L) is expressed on a variety ofcell types including activated platelets, vascularendothelial cells, vascular smooth muscle cells,monocytes, and macrophages. Following expressionon the cell surface, CD40L is partly cleaved byproteases and subsequently released into thecirculation as soluble CD40 ligand (sCD40L) whichcan be detected in serum and plasma. Both, surfacebound CD40L and sCD40L are biologically active andare involved in the pathophysiology of acutecoronary syndromes (ACS) [1–4]. It has beenshown consistently in several studies, that sCD40Lis elevated in patients with acute coronary syn-dromes (ACS) and that it provides prognosticinformation with therapeutical implications inde-pendent of established cardiac markers, e.g. cardi-ac troponins [5–7].

However, in a recently published study the im-portance of preanalytic conditions for sCD40L as-sessment has been demonstrated suggesting thatserum is not adequate because of ex vivo release ofsCD40L from platelets [8]. Thus, it was the aim of ourprospective study to evaluate the impact of sam-pling techniques on sCD40L concentrations of pa-tients with acute coronary syndromes compared topatients with cardiovascular risk factors (CVRF) butwithout relevant coronary artery disease and appar-ently healthy individuals.

Methods

We included 60 patients in this study — 30 patientswith an ACS, 10 patients with cardiovascular riskfactors but in whom relevant coronary arterydisease had been ruled out by angiography or stresstest and 20 apparently healthy individuals. An acutecoronary syndrome was diagnosed if the patientshad an episode of chest pain within the last 48 h incombination with elevated troponin T, ST-segmentdepression or ST segment elevation, recurrent orpersisting angina at rest. We considered hyperten-sion, diabetes mellitus, hyperlipidaemia, smoking,and a positive family history as cardiovascular riskfactors. The clinical diagnosis was established bytwo independent cardiologists who were not in-volved in this study and were blinded to sCD40Lvalues. All included patients gave written informedconsent and the study was approved by the localethic board.

Diagnostic left and right coronary angiogramswere performed in all patients by the standardJudkins' technique using a 5 french sheath from afemoral approach. Coronary artery lesions with less

than 50% diameter stenosis were considered nonrelevant.

From all patients with an ACS venous blood wassampled at admission prior to angiography andadministration of clopidogrel or GP IIb/IIIa inhibitors.In patients with an increased cardiovascular riskprofile blood has been sampled after relevantcoronary artery disease has been ruled out byangiography. For blood sampling commercially avail-able plastic tubes (S-Monovette®, Sarstedt Ag, Nüm-brecht, Germany) were used. From each patient bloodhas been sampled in three different types of tubes— ingel filled tubes without any additives, in tubes with0.3 ml 0.106 mol/l trisodium citrate solution and intubes with 1.6 mg tripotassium-EDTA/ml blood.Samples were allowed to clot for 30 min and werecentrifuged at 3000 ×g for 10 min at room tempera-ture. The samples were not put in ice beforecentrifugation. sCD40L values were measured imme-diately after centrifugation with the use of anelectrochemiluminescence-immunoassay (Elecsys®,Roche Diagnostics. Mannheim, Germany). No sampleswere frozen or thawed. Intra-assay imprecision (CV)was determined by measuring samples from 4 differ-ent citrated plasma pools. Intra-assay CV for sCD40L1.8 μg/l 8.9%, for 1.1 μg/l 2.1%, for 0.4 μg/l 1.9%, andfor 0.08 μg/l 1.9%. Inter-assay CV for sCD40L 1.6 μg/l 4.8%, for 1.1 μg/l 1.7%, for 0.4 μg/l 2.0%, and for0.08 μg/l 2.3%.

Values for sCD40L are given as median andinterquartile range (IQR) and for all other variablesas mean and standard deviation (SD). For group wisecomparisons of sCD40L values Mann–Whitney test(2-groups) and Kruskal–Wallis test (n-groups) wereused. For the correlation of sCD40L values assessedby the different sample types the Spearman corre-lation coefficient was calculated. For the compar-ison of dichotomized sCD40L values Fishers exacttest was applied. Linear regression analysis includingthe independent explanatory variables gender, ageand disease state for the dependent variable sCD40Lmeasured from each sampling type was performedAll tests were performed two sided and p valuesbelow 0.05 were considered to indicate statisticalsignificance. For all statistical analysis the statisti-cal software SPSS 10.0 for windows was used.

Results

Sixty patients were included (22 females, aged 49±14 years). There was no difference in age and genderdistribution between ACS patients and patients withcardiovascular risk profile, but healthy individualswere younger and more often female. The detailedbaseline characteristics are shown in Table 1. We didnot observe a correlation of platelet count to sCD40L

Table 1 Baseline characteristics of the patients

ACS Patients with CVRF but no relevant CAD Healthy individuals

n 30 10 20Gender (female/male) 6/24 4/6 12/8Age (years) 57±14 57±8 32±9Creatininekinase MB (U/l) 64±115 nd ndTroponin T (ng/ml) 1.4±3.3 nd ndPlatelets (103/ml) 228±75 242±79 ndsCD40L (ng/ml)- serum 4.64 (2.28–5.82) 3.57 (2.36–6.03) 5.19 (3.75–6.22)sCD40L (ng/ml)- EDTA plasma 0.50 (0.30–1.29) 0.32 (0.12–0.51) 0.16 (0.11–0.28)sCD40L (ng/ml)- citrate plasma 0.23 (0.16–0.35) 0.19 (0.14–0.25) 0.14 (0.11–0.20)

Values are expressed as absolute numbers, as mean±standard deviation or as median and interquartile range in parenthesis. nd = nodata available.

813Influence of sample type on soluble CD40 ligand assessment in patients with ACS

values measured from serum (Rho=0.267; 0=0.110)from EDTA plasma (Rho=−0.51; p=0.766) and fromcitrated plasma (Rho=0.240; p=0.152). The sCD40Lvalues measured in EDTA plasma were higher inpatients with an ACS as compared to the values ofpatients with a cardiovascular risk profile withborderline significance (0.50 (0.30–1.29) ng/ml vs.0.32 (0.12–0.51) ng/ml, p=0.05 and compared tothe values of the healthy individuals (0.16 (0.11–0.28) ng/ml; pb0.001). If sCD40L concentrationwere measured in citrated plasma, values of ACSpatients were higher as compared to those ofhealthy individuals (0.23 (0.16–0.35) ng/ml vs.0.14 (0.11–0.20) ng/ml, p=0.002), but no statisti-cally significant difference was observed betweenACS patients and patients with cardiovascular riskprofiles (0.19 (0.14–0.25), p=0.33). In serum sam-ples, there was no difference in sCD40L valuesbetween patients with ACS and patients withcardiovascular risk profile patients (4.6 (2.3–5.8)ng/ml vs. 3.6 (2.4–6.0) ng/ml; p=0.331). Values ofhealthy individuals (5.19 (3.8–6.2) ng/ml; p=0.002)were even higher as compared to concentrations of

Figure 1 Values of sCD40L of the patients from group 2. The sopredefined cut-off value. A – values attained from EDTA plasma,from serum.

patients with ACS (Fig. 1). We found a moderatecorrelation between sCD40L values in EDTA plasmaand sCD40L values in citrated plasma (r=0.686,pb0.001). In contrast, there was no correlationbetween sCD40L values in serum and sCD40L valuesin EDTA-plasma (r=0.127, p=0.337) (Fig. 2) orcitrated plasma (r=0.158, p=0.28), respectively.sCD40L values were dichotomized at an arbitrarychosen cut-off value of 0.6 ng/ml for EDTAplasmaand0.24 ng/ml for citrated plasma, which was the 90thpercentile of the values observed in patients withcardiovascular risk factors and healthy individuals.Based on sCD40L concentration in EDTA plasmasamples, 11 out of 30 (37%) ACS patients but only 3out of 30 (10%) patients with cardiovascular riskprofile or healthy individuals had values above thiscut-off value (p=0.03). In citrated plasma samples,13 out of 30 (43%) ACS patients and only 3 out of 30(10%) patients with cardiovascular risk profile orhealthy individuals had values above the defined cut-off (p=0.002). Performing linear regression analysisincluding the explanatory variables gender, age anddisease state for the dependent variable sCD40L

lid line indicates the trend line, the dotted line indicates theB – values attained from citrated plasma, C – values attained

Figure 2 Scatter plot of sCD40L values measured fromEDTA plasma to sCD40L values measured from serum.

814 M. Weber et al.

measured from each sample type, we found thatdisease condition (ACS) was the only independentdeterminant for sCD40L values measured from EDTAplasma (p=0.043 for disease state, p=0.681 forgender and p=0.627 for age). Disease state was notan independent determinant for sCD40L measuredfrom citrated plasma and from serum (p=0.182 incitrated plasma and p=0.356 for serum samples), norwas gender (p=0.896 in citrated plasma and p=0.956for serum samples) and age (p=0.264 in citratedplasma and p=0.967 for serum samples).

Discussion

In the present study, we investigated the impact ofthe preanalytic factor sample type on sCD40L valuesin healthy individuals, patients with cardiovascularrisk factors but without relevant coronary arterydisease, and in patients with acute coronarysyndromes. The main finding is that sCD40L valuesmeasured in EDTA or citrate plasma were higher inpatients with an ACS as compared to controls andsignificantly more ACS patients had sCD40L valuesabove a predefined cut-off value. In contrast, nodifference between ACS patients and controls wasobserved for sCD40L values in serum. Therefore, ourfindings together with previously reported findings[8–10] indicate that in vitro platelet activation andsubsequent sCD40L release considerably contributesto elevated values of sCD40L in serum and limitstheir diagnostic utility. In line with this finding is ourobservation, that sCD40L values measured from EDTAor citrated plasmawerehigher in patientswith anACScompared to controls, whereas serum sCD40L mea-surement did not differentiate between ACS patientsand controls. This shows that in serum samples, thepathological in vivo sCD40L release associated withthe onset of the ACS is superposed by the in vitrorelease. Considering these results, it is obvious, thatpreanalytic factors are of paramount importance.

Apparently, serum samples are inadequate for sCD40Lmeasurements and plasma, either EDTA or citrateplasma should be the preferred sample type.

Limitations

A major limitation of the present study is the smallsample size and that patients were not matched withthe control group. However, the aim of the study wasto evaluate the impact of different sample typesunder various disease conditions, especially inpatients with acute coronary syndromes. It was notthe aim to demonstrate differences of sCD40lconcentrations between the different study groups.Therefore, the small sample size and a patients-controls mismatch are acceptable.

Acknowledgments

This study has been sponsored by Roche Diagnosticsand assays have been provided without charge.

References

[1] Schonbeck U, Libby P. The CD40/CD154 receptor/liganddyad. Cell Mol Life Sci 2001;58:4–43.

[2] Henn V, Slupsky JR, Grafe M, Anagnostopoulos I, Forster R,Muller-Berghaus G, et al. CD40 ligand on activated plateletstriggers an inflammatory reaction of endothelial cells. Na-ture 1998;391:591–4.

[3] Henn V, Steinbach S, Buchner K, Presek P, Kroczek RA. Theinflammatory action of CD40 ligand (CD154) expressed onactivated human platelets is temporally limited by coex-pressed CD40. Blood 2001;98:1047–54.

[4] Urbich C, Dernbach E, Aicher A, Zeiher AM, Dimmeler S.CD40 ligand inhibits endothelial cell migration by increasingproduction of endothelial reactive oxygen species. Circu-lation 2002;106:981–6.

[5] Aukrust P, Muller F, Ueland T, Berget T, Aaser E, Brunsvig A,et al. Enhanced levels of soluble and membrane-boundCD40 ligand in patients with unstable angina. Possible re-flection of T lymphocyte and platelet involvement in thepathogenesis of acute coronary syndromes. Circulation1999;100:614–20.

[6] Heeschen C, Dimmeler S, Hamm CW, van den Brand MJ,Boersma E, Zeiher AM, et al. Soluble CD40 ligand in acutecoronary syndromes. N Engl J Med 2003;348:1104–11.

[7] Varo N, de Lemos JA, Libby P, Morrow DA, Murphy SA, Nuzzo R,et al. Soluble CD40L: risk prediction after acute coronarysyndromes. Circulation 2003;108:1049–52.

[8] Weber M, Rabenau B, Stanisch M, Elsaesser A, Mitrovic V,Heeschen C, et al. Influence of sample type and storageconditions on soluble CD40 ligand assessment. Clin Chem2006;52:888–91.

[9] Ahn ER, Lander G, Jy W, Bidot CJ, Jimenez JJ, Horstman LL,et al. Differences of soluble CD40L in sera and plasma:implications on CD40L assay as a marker of thrombotic risk.Thromb Res 2004;114:143–8.

[10] Halldorsdottir AM, Stoker J, Porche-Sorbet R, Eby CS.Soluble CD40 ligand measurement inaccuracies attributableto specimen type, processing time, and ELISA method. ClinChem 2005;51:1054–7.